Antifibrinolytics in the treatment of traumatic brain injury

Purpose of review Traumatic brain injury (TBI) is a leading cause of trauma-related deaths, and pharmacologic interventions to limit intracranial bleeding should improve outcomes. Tranexamic acid reduces mortality in injured patients with major systemic bleeding, but the effects of antifibrinolytic drugs on outcomes after TBI are less clear. We therefore summarize recent evidence to guide clinicians on when (not) to use antifibrinolytic drugs in TBI patients. Recent findings Tranexamic acid is the only antifibrinolytic drug that has been studied in patients with TBI. Several recent studies failed to conclusively demonstrate a benefit on survival or neurologic outcome. A large trial with more than 12 000 patients found no significant effect of tranexamic acid on head-injury related death, all-cause mortality or disability across the overall study population, but observed benefit in patients with mild to moderate TBI. Observational evidence signals potential harm in patients with isolated severe TBI. Summary Given that the effect of tranexamic acid likely depends on a variety of factors, it is unlikely that a ‘one size fits all’ approach of administering antifibrinolytics to all patients will be helpful. Tranexamic acid should be strongly considered in patients with mild to moderate TBI and should be avoided in isolated severe TBI.


INTRODUCTION
Trauma is a leading cause of mortality, particularly in young adults, and accounts for about 4.4 million annual deaths worldwide [1]. Uncontrolled haemorrhage as well as traumatic brain injury (TBI) are the two leading causes of death after trauma [2], and in those patients with TBI, the presence and extent of intracranial bleeding is a strong predictor of mortality [3]. Hence, pharmacologic interventions to limit systemic as well as intracranial bleeding should have a large potential to reduce trauma-associated mortality, both in patients with extracranial injury and with TBI.
Antifibrinolytic agents have been shown to limit blood loss across a wide range of medical conditions, for example in patients with haemoptysis, epistaxis, haematuria or postpartum haemorrhage, as well as in patients undergoing major surgery, including cardiothoracic, abdominal, orthopaedic or obstetric surgery [4][5][6][7][8][9][10][11][12][13]. It therefore seems plausible that antifibrinolytics should also reduce blood loss in trauma patients. Since the publication of the landmark CRASH-2 trial on effects of tranexamic acid (TXA) in injured patients with (risk of) significant bleeding in 2010 [14], this drug has been widely used to prevent trauma-related death and has even been added to the WHO's list of essential medicines in 2011. However, the role of TXA or other antifibrinolytics to improve outcomes after TBI is less clear. Although these drugs should theoretically be beneficial as outlined above, potential harm -for example due to cerebral intravascular microthrombi, dural sinus thrombosis, other thromboembolic complications or promotion of seizure activityis also plausible [15 and trauma-induced coagulopathy. Subsequently, the key pharmacologic properties of the clinically most relevant antifibrinolytic drugs, namely aprotinin, TXA and e-aminocaproic acid (EACA) are described. In the final part, we discuss the role of antifibrinolytics in the treatment of TBI.

FIBRIN, FIBRINOLYSIS AND TRAUMA-INDUCED COAGULOPATHY
At the end of the coagulation cascade, fibrinogen (factor I) is converted to fibrin, which has a pivotal role in haemostasis, as fibrin polymers form a mesh that stabilizes the platelet clot and anchors it to the damaged blood vessel [17]. As a physiologic counterpart to clot formation, fibrinolysis serves to avoid excessive accumulation of intravascular fibrin and to prevent thrombosis, to dissolve existing thrombi and to degrade fibrin once the vascular damage has been repaired [18]. The key step in this process is the conversion of plasminogen to its active form, plasmin. This conversion is catalyzed by tissue plasminogen activator (tPA, main activator in blood), urokinase-type plasminogen activator (uPA) or other proteases [19,20]. Plasmin, in turn, cuts fibrin polymers into soluble elements known as fibrin degradation products, such as D-dimers.
Under physiologic conditions, coagulation as well as fibrinolysis are both tightly regulated by a number of activating or inhibiting enzymes or cofactors to maintain a balance between clot formation and clot dissolvement. However, under a variety of conditions such as (severe) surgical bleeding, cardiopulmonary bypass or trauma, the fragile equilibrium can be readily disturbed. In trauma patients, the bleeding that is initially caused by injury of blood vessels is often sustained and aggravated by an acute coagulopathy, referred to as trauma-induced coagulopathy (TIC) [21 & ]. Although iatrogenic haemodilution by large volumes of crystalloids and other coagulation-factor free solutions as well as hypothermia and acidosis contribute to coagulopathy, TIC is a consequence of the trauma itself and occurs secondary to tissue injury, shock and inflammatory upregulation, independent of the aforementioned exogenous factors [22]. An early phase characterized by impaired coagulation and excessive bleeding (within 6 h after injury) is commonly distinguished from a late phase (onset >24 h after trauma) in which hypercoagulation prevails, but a large heterogeneity exists regarding the timing and clinical presentation of TIC [21 & ]. The cause of TIC is multifactorial and involves endothelial activation, inflammatory upregulation, platelet dysfunction, impaired thrombin generation, fibrinogen depletion and hyperfibrinolysis, as reviewed in detail elsewhere [21 & ,22]. Notably, coagulopathy and hyperfibrinolysis are not only common in patients with systemic injury and massive bleeding but are also regularly observed in patients with (isolated) TBI. This coagulopathy is commonly thought to be triggered by the release of tissue factor (factor III) from injured brain tissue [23] -which is actually an oversimplification given the complex interplay of various factors reviewed in detail elsewhere [24 & ] -and its presence is associated with a markedly increased mortality [25].
In the context of traumatic bleeding and TIC, antifibrinolytic drugs are administered with the primary intention to limit blood loss by targeting (hyper-)fibrinolysis and shifting the balance towards clot stabilization. However, other properties of antifibrinolytic drugs, such as protective effects on the endothelium as well as anti-inflammatory effects, may also mediate beneficial effects in trauma patients [26][27][28]. Interestingly, given these properties, antifibrinolytics have also been increasingly used for indications completely unrelated to haemorrhage, for example for inflammatory skin disorders [29,30].

PHARMACOLOGY OF ANTIFIBRINOLYTIC DRUGS
Two types of antifibrinolytics with different mechanism of action can be distinguished, namely serine protease inhibitors (aprotinin), as well as lysine analogues (TXA and EACA). Aprotinin is a nonspecific, competitive inhibitor that blocks the active sites of a family of enzymes known as serine proteases, which includes plasmin. Similar to the physiological plasmin inhibitor a 2 -antiplasmin, aprotinin primarily

KEY POINTS
Trauma patients are at risk for trauma-induced coagulopathy, which sustains and aggravates bleeding and increases the risk of death.
Antifibrinolytic drugs have been shown to reduce blood loss in a variety of medical and surgical settings and should likewise be useful to target trauma-induced coagulopathy and to reduce mortality after trauma.
Tranexamic acid --the only antifibrinolytic drug that has so far been studied in trauma patients --has been shown to reduce mortality after systemic trauma, but its effects in (isolated) TBI are less clear.
Although strong evidence for effects of TXA in patients with TBI is lacking, a potential benefit has been suggested in patients with mild to moderate TBI and a potential harm in patients with isolated severe TBI. targets free plasmin but has little effect on bound plasmin [31,32]. Aprotinin is administered intravenously, and the typical dosing scheme for adult cardiac surgery, in which aprotinin has been predominantly used, consists of an initial loading dose of 2 million kallikrein inhibitor units (KIU) followed by a continuous infusion of 500 000 KIU per hour until chest closure, with an additional 2 million KIU added to the prime solution of the cardiopulmonary bypass circuit. Aprotinin is degraded by lysosomal enzymes and renally excreted, with a plasma elimination halflife of approximately 5-8 h [33].
After having been the most popular antifibrinolytic drug in the late 1990 s and early 2000 s, safety concerns raised in observational studies as well as in the Blood Conservation Using Antifibrinolytics in a Randomized Trial (BART) study [34] led to withdrawal of aprotinin from the market in November 2007. In the meantime, flaws identified in the BART-trial, as well as re-analyses of the available data have led to a re-evaluation of the risk-benefit profile, and aprotinin was reapproved in Canada in 2011 and in Europe in 2012, but is still unavailable in the United States.
Nowadays, EACA and particularly TXA are the most widely used antifibrinolytics. These synthetic analogues of the amino acid lysine competitively occupy the so-called lysine binding sites of plasminogen, which prevents binding of the fibrin molecule ( Fig. 1) [35]. Both agents can be administered intravenously, orally and topically, and a recent clinical trial showed that TXA is also well tolerated and rapidly resorbed when administered via the intramuscular route [36,37]. In the context of anaesthesia, emergency medicine and critical care, TXA is usually administered to adults in an intravenous loading dose of 15 mg/kg or 1 g, followed by a continuous infusion of 1 g over 8 h. EACA is 6-10 times less potent than TXA [38] and therefore administered in higher doses, with a typical intravenous loading dose of 5 g given over 1 h, followed by a 1 g/h of continuous infusion for 8 h or until the bleeding is controlled. Both drugs are renally excreted unmetabolized, with a plasma half-life of about 2 h [39].

USE OF ANTIFIBRINOLYTIC DRUGS IN TRAUMATIC BRAIN INJURY
Despite the theoretical potential of all antifibrinolytics to target TIC and to reduce bleeding in injured FIGURE 1. Mechanism of action of tranexamic acid and e-aminocaproic acid. The left panel depicts binding of the activator (in blood, this is generally tissue plasminogen activator, tPA) and fibrin to plasminogen, leading to activation of plasminogen to plasmin and breakdown of fibrin into fibrin degradation products. TXA and EACA block the lysine binding sites of plasminogen (right), preventing the binding of fibrin. patients, no clinical data are available on effects of aprotinin or EACA in trauma patients. On the one hand, this can likely be explained by safety concerns about aprotinin and its continuing unavailability in the United States. On the other hand, due to the ubiquitous use of TXA in trauma patients following the CRASH-2 trial and given that TXA is a cheap and readily available drug, there has obviously been little interest to explore the usefulness of other antifibrinolytics for treating trauma-induced bleeding. Although it is not impossible that aprotinin and EACA may have a role in future clinical studies, all current evidence on the effects of antifibrinolytic drugs on outcomes in TBI is limited exclusively to TXA, on which the remainder of this section will focus. Table 1 summarizes the randomized trials specifically investigating the effects of TXA in patients with TBI, as well as recent observational studies (2019-2022).
The CRASH-2 trial was a large-scale study with 200 127 patients in 274 hospitals in 40 countries [14]. Adult trauma patients with significant bleeding, or at risk for significant bleeding, were randomized to receive either 1 g of intravenous TXA over 10 min followed by an infusion of 1 g over 8 h or placebo. The primary outcome was hospital mortality within 28 days of injury. The authors observed an absolute 1.5% lower mortality in patients treated with TXA (14.5 vs. 16%, P ¼ 0.0035). Despite concerns about the external validity and other limitations summarized by Napolitano et al. [40], and despite the fact that TXA is not specifically approved for this indication by major drug agencies including the U.S. Food and Drug Administration (FDA) and the European Medicines Agency (EMA), TXA has since then ubiquitously been used (off-label) in trauma care worldwide. A nested substudy (CRASH-2 Intracranial Bleeding Study, [41]) provided first randomized data on the effects of TXA in patients with TBI [Glasgow Coma Scale (GCS) score 14 and a brain computed tomographic (CT) scan compatible with TBI]. With a sample size of only 270 patients, no significant differences were observed in the primary outcome (haemorrhage growth), or in individual clinical outcomes including death. Only when pooling several adverse outcomes into a 'composite poor outcome', a benefit of TXA was observed.
The CRASH-3 trial, published in 2019, is the first major and up to now the largest clinical trial of the effects of TXA in patients with TBI [42 && ]. The study was performed in 175 hospitals in 29 countries. A total of 12 737 adult patients with a GCS score of 12 or less and any intracranial bleeding on CT scan and no major extracranial bleeding were randomized to receive either TXA or placebo with the same dosing scheme as in the CRASH-2 trial. Although the study initially included patients within 8 h of injury, the protocol was amended during the study period to limit inclusion to patients within 3 h of injury. This change reflected accumulating evidence that TXA should be administered as early as possible, and it has been suggested that administration after 3 h even increases mortality [43]. The amended primary outcome, death from head-injury within 28 days in patients randomized within 3 h of trauma, was analysed in 9127 patients, and no significant difference was found [18.5% in the TXA group vs. 19.8% in the placebo group, relative risk 0.94 (95% confidence interval, 95% CI 0.86-1.02)]. Similarly, there was neither evidence of benefit from TXA on key secondary outcomes, including all-cause mortality and disability, nor of an increase in adverse events and complications. A subgroup analysis suggested a protective effect of TXA in patients with mild to moderate TBI (GCS 9), but not in patients with severe TBI (GCS 8).
Despite actually being a 'negative' study with respect to the primary outcome, the authors of the CRASH-3 trial concluded that treatment with TXA within 3 h reduces head-injury related death. Although it is possible that this is true (a nonsignificant difference does not exclude a beneficial effect), the data actually do not provide strong evidence to claim benefit of TXA, particularly not in the overall cohort as well as in patients with severe TBI. The conclusion may be justified for patients with mild to moderate TBI, but the result found in this subgroup should also be interpreted with caution, as it carries an increased risk for type I error in the absence of a multiplicity adjustment [44]. Moreover, while CRASH-3 was a large and well designed trial, the study results must be viewed in the context of its limitations, such as concerns about the external validity of the results and potential selection bias [45][46][47][48][49][50][51][52][53].
Rowell et al. [54 && ] performed another landmarktrial in which the effect of prehospitally administered TXA was studied. Patients at least 15 years of age with a prehospital GCS 3-12, at least one reactive pupil and a systolic blood pressure of at least 90 mmHg were randomly assigned to two different dosing schemes of TXA (1 g bolus and 1 g maintenance dose or 2 g bolus without maintenance dose) or placebo. Although the researchers initially planned to compare each TXA dosing group to placebo, concerns about the study power led to a protocol change and the two TXA groups were combined for comparison with the placebo group. The primary outcome was the extended Glasgow Outcome Scale score (GOSE) at 6 months after injury, dichotomized as a favourable (GOSE >4) or  an unfavourable (GOSE 4) outcome. Of the 1063 study participants, 966 were analysed. There was neither a significant difference between the groups for the primary outcome nor for key secondary outcomes, including 28-day mortality, 6-month Disability Rating Scale score and progression of intracranial haemorrhage.
In a similar study of prehospital TXA administration [55], trauma patients at risk for haemorrhage were randomized to receive one of three different TXA treatment regimens or placebo, and the TXA groups were pooled for comparison with placebo. No benefit of TXA was observed for the primary outcome, 30-day mortality or for secondary outcomes. Although this study was not specifically designed to address TBI, no benefit was also observed in the subgroup of patients with TBI (n ¼ 168).
Several smaller randomized trials also failed to show beneficial effects of TXA in terms of survival or neurologic outcome [56][57][58][59][60][61][62][63]. Five recent meta-analyses pooled the available evidence across published randomized trials, of which three reported beneficial effects of TXA [64][65][66] and two found no benefit [67,68 && ]. Remarkably, all three studies that report beneficial effects inappropriately included the CRASH-2 trial -not only the patients with TBI from the intracranial bleeding sub-study but also all patients -so that the positive findings are largely attributable to CRASH-2 and do not specifically apply to the population of TBI patients. Hence, the available pooled evidence does not demonstrate a clear benefit of TXA in patients with TBI.
Recent observational studies have shown mixed results, including benefit as well as harm, as summarized in Table 1 [ [69][70][71][72]. Although these studies must be interpreted with caution given their inherent limitations, in particular confounding [73], they allow gauging potential treatment effects of TXA when used in regular clinical practice rather than under controlled trial conditions. Notably, in an analysis of 1827 patients with severe TBI (GCS 8), Bossers et al. [15 ] did not observe an overall relationship between TXA and mortality after thorough adjustment for potential confounders, but found an increased mortality in TXA-treated patients in the subgroup of patients with isolated severe TBI. Considering all available studies jointly, the evidence for using TXA in TBI patients is rather weak. There is no clear evidence of either benefit or harm, and reported effect sizes were generally small. For example, assuming that the point estimate represents a true population effect rather than random sampling error, the overall effect seen in the CRASH-3 trial corresponds to a number needed to treat of 82 patients, that is a large number of patients need to be treated to avert a single death (Fig. 2). TXA is obviously not a magic bullet, and its use should not distract healthcare providers from focusing on those factors that are known to really matter for patient outcome, such as maintaining an adequate blood pressure [74]. Given the various factors that may influence the effect of TXA, such as the timing of administration, type of injury (isolated TBI vs. combined with extracranial haemorrhage) and severity of injury, as well as additional factors such as treatment infrastructure (e.g. access and transport times to definite care), it is unlikely that a 'one size fits all' approach of treating all patients with TXA will be helpful. The question arises as to which patients, if any, may actually benefit from this treatment. Although further studies are needed to provide additional insight, our personal interpretation and treatment recommendation based on the available literature is as follows: (1) In patients with TBI and additional relevant extracranial bleeding, administration of TXA should be strongly considered on the basis of CRASH-2 and other studies demonstrating a lower risk of death in bleeding trauma patients. (2) In patients with isolated mild to moderate TBI (GCS 9), CRASH-3 suggests a beneficial effect of TXA without an evident increase in adverse events. Thus, TXA is potentially lifesaving and should be strongly considered. (3) For patients with isolated severe TBI, there is no evidence from randomized trials that TXA is beneficial, and observational data signal potential harms. TXA should therefore be avoided. (4) Whenever TXA is being considered, it should be administered as early as possible and no later than 3 h after the injury. However, this treatment should not delay prompt treatment of factors known to trigger secondary brain injury.

CONCLUSION
Although TXA appears to improve outcomes in injured patients with (or at risk for) major bleeding, the effects in patients with (isolated) TBI are less clear. Unequivocal evidence for benefit across all patients with TBI is lacking, and further research FIGURE 2. Visual impression of the effect size of tranexamic acid observed for the overall population of patients in the CRASH-3 trial. Assuming that the observed point estimate represents the 'true' effect of tranexamic acid in the population rather than random sampling error, 82 patients need to be treated to avert a single head-injury related death (solid pictogram), that is the vast majority receiving the drug do not benefit (outlined pictograms).
is needed to determine which patients actually do profit from TXA administration. Current literature suggests that TXA may decrease mortality in patients with mild to moderate TBI but may increase mortality in patients with isolated severe TBI.