Journal of Thrombosis and Haemostasis, 13 (Suppl. 1): S195–S199
12878
INVITED REVIEW
Tranexamic acid in trauma: how should we use it?
I. ROBERTS
Clinical Trials Unit, London School of Hygiene & Tropical Medicine, London, UK
To cite this article: Roberts I. Tranexamic acid in trauma: how should we use it?. J Thromb Haemost 2015; 13 (Suppl. 1): S195–S9.
Summary. Tranexamic acid (TXA) reduces blood loss by
inhibiting the enzymatic breakdown of fibrin. It is often
used in surgery to decrease bleeding and the need for
blood transfusion. In 2011, results from a multi-center,
randomized, and placebo-controlled trial (CRASH-2 trial)
showed that TXA (1 g loading dose over 10 min followed
by an infusion of 1 g over 8 h) safely reduces mortality in
bleeding trauma patients. Initiation of TXA treatment
within 3 h of injury reduces the risk of hemorrhage death
by about one-third, regardless of baseline risk. Because it
does not have any serious adverse effects, TXA can be
administered to a wide spectrum of bleeding trauma
patients. Limiting its use to the most severely injured or
those with a diagnosis of ‘hyperfibrinolysis’ would result
in thousands of avoidable deaths. A clinical trial
(CRASH-3 trial) of TXA in patients with traumatic brain
injury is now in progress.
Keywords: clinical trial; fibrinolysis; surgery; tranexamic
acid; wounds and injuries.
Molecular mechanism of action of tranexamic acid
Tranexamic acid (TXA) inhibits the enzymatic break-
down of fibrin blood clots. The process of fibrin break-
down begins when plasminogen, a glycoprotein pro-
enzyme produced by the liver, binds to stands of fibrin.
The plasminogen molecule is folded into loops called
kringles [1,2]. Plasminogen binds to fibrin via lysine
binding sites located at the ends of these loops. If the
lysine residues on fibrin are enzymatically removed, plas-
minogen binding is inhibited [1]. Tissue plasminogen
activator (t-PA) released by the vascular endothelium
also binds to fibrin and converts plasminogen to active
plasmin. Once formed, plasmin cleaves fibrin into small
protein fragments (fibrin degradation products) such as
Correspondence: Ian Roberts, Clinical Trials Unit, London School
of Hygiene & Tropical Medicine, Keppel Street, London WC1E
7HT, UK.
Tel.: +20 7958 8128; fax: +20 7299 4663.
E-mail: Ian.Roberts@lshtm.ac.uk
© 2015 International Society on Thrombosis and Haemostasis
DOI: 10.1111/jth.
d-dimers. This cleavage exposes more lysine residues
which bind more plasminogen thus initiating a positive
feedback that accelerates fibrinolysis [1]. TXA has a
molecular structure similar to lysine. It inhibits fibrinoly-
sis by preventing the binding of plasminogen to fibrin.
TXA has also been shown to inhibit tissue factor
induced fibrinogenolysis [3].
Plasminogen might have other important physiological
and pathophysiological roles apart from fibrinolysis,
including roles in inflammation and tissue remodeling and
so by preventing plasminogen binding, TXA could have
other important biological effects. High concentrations of
plasminogen receptors are found on endothelial cells,
monocytes, lymphocytes, and platelets. When plasmino-
gen binds to cells, its activity is increased and plasmin
linked to the cell surface is less exposed to inhibitors. In
vitro, plasmin has been shown to have pro-inflammatory
effects, including cytokine release, monocyte activation,
and inflammatory cell migration [4–8]. These effects can
be inhibited by TXA [8].
TXA reduces bleeding in surgery
TXA reduces blood loss and the need for blood transfu-
sion in patients undergoing elective and emergency sur-
gery. A 2012 systematic review and meta-analysis of data
from 104 clinical trials of TXA administration in surgical
patients found that TXA reduces blood loss by about
one-third (risk ratio = 0.66, 95% CI 0.65–0.67; P < 0.001)
irrespective of the type of surgery [9,10]. In most of the
included trials, TXA was given immediately prior to inci-
sion. Although the absolute reduction in blood loss with
TXA increased as surgical bleeding increased, the percent-
age reduction was similar [10]. In other words, TXA
appears to reduce blood loss by about one-third irrespec-
tive of whether the operation entails small, moderate, or
large volumes of bleeding. TXA also reduced the proba-
bility of receiving a blood transfusion, again by about a
third (risk ratio 0.62, 95% CI 0.58 to 0.65; P < 0.001) [9].
Although there were fewer deaths in TXA-treated patients
(0.61, 0.38–0.98; P = 0.04) because only one-third of trials
reported mortality data, this estimate could be biased due
to selective reporting of outcomes [9]. There was no
evidence of any increased risk of thromboembolic adverse
events with TXA; however, the effect estimates were
S196 I. Roberts
imprecise and might also be biased due to selective
reporting of outcomes [9].
TXA prevents death due to bleeding in trauma
The CRASH-2 trial was a randomized, placebo-controlled
trial of the effect of tranexamic acid (1 g loading dose
over 10 min followed by an infusion of 1 g over 8 h) on
death and vascular occlusive events in bleeding trauma
patients. A total of 20,211 adult trauma patients, with or
at risk of significant bleeding, who were within 8 h of
their injury, were randomly allocated to receive TXA or a
matching placebo. The primary outcome was death within
4 weeks. TXA significantly reduced death due to bleeding
(RR = 0.85, 95% CI 0.76–0.96) and all-cause mortality
(RR = 0.91, 95% CI 0.85–0.97), with no increase in vas-
cular occlusive events [11]. The CRASH-2 trial recruited
patients from high-, middle-, and low-income settings.
The beneficial effect of TXA on mortality did not vary by
geographical region [12].
Before the start of the trial, we hypothesized that TXA
would be most effective when treatment was initiated
soon after the injury, when bleeding is most profuse. We
therefore prespecified a sub-group analysis of the effect of
TXA on mortality stratified by the time from injury to
the initiation of treatment (<1, 1–3, 3–8 h). We found
strong evidence (Fig. 1) that early treatment was more
effective (P < 0.0001) [7]. When initiated within 1 h of
injury, TXA reduced the risk of death due to bleeding by
nearly one-third (RR = 0.68, 95% CI 0.57–0.82;
P < 0.0001), treatment given between 1 and 3 h reduced
the risk of death due to bleeding by about one-fifth
(RR = 0.79, 0.64–0.97; P = 0.03). However, treatment ini-
tiated after 3 h did not reduce mortality and appeared to
increase the risk of death due to bleeding [13]. Further
analyses of the CRASH-2 trial data have shown that the
effect of TXA on mortality is greatest on the day of the
injury, when early TXA treatment (within 3 h of injury)
reduces the risk of death from all causes by about 20%
and death due to bleeding by about 30% [14]. These
results strongly suggest that TXA improves survival by
reducing bleeding and that there is a short time-window
during which TXA administration can prevent exsangui-
nation.
Despite its mechanism of action, the decrease in surgi-
cal blood loss with TXA and the reduction in hemorrhage
death observed in the CRASH-2 trial, the lack of any
apparent effect of TXA on the receipt of blood transfu-
sion in the CRASH-2 trial has caused some authors to
question its mechanism of action. Why does TXA reduce
blood transfusion in surgery but not trauma? The most
obvious explanation is that traumatic bleeding begins
before hospital admission whereas the patients in the
CRASH-2 trial were given TXA in hospital. The blood
transfusions that trauma patients receive in hospital are
largely in response to bleeding that took place before
TXA was given. Any effect of TXA on transfusion would
also have been diluted by transfusions given later during
the hospital stay, for example, transfusions given after
surgery. Furthermore, because TXA reduced mortality,
more TXA-treated patients had the opportunity to be
transfused. Finally, quantification of bleeding is difficult
in trauma and so the amount of blood transfused might
not accurately reflect the volume of blood lost.
There was no evidence that TXA increased the risk of
vascular occlusive events. Indeed, when given within 3 h
of injury, there was a significant reduction in the odds of
fatal and non-fatal vascular occlusive events (odds
ratio = 0.69, 95% confidence interval 0.53 to 0.89;
P = 0.005) [15]. We have shown that the risk of vascular
occlusive events in trauma patients increases with the
severity of bleeding. Decreasing systolic blood pressure,
longer capillary refill times, and higher heart rates are risk
factors for arterial and venous vascular occlusive events
[16]. By reducing blood loss TXA might therefore reduce
the risk of hemorrhage death and the risk of vascular
occlusive events.
How should we use TXA in trauma?
TXA reduces blood loss in surgery and reduces the risk
of death due to bleeding in trauma patients. It is highly
cost-effective and with no serious side effects [17]. So
which trauma patients should be treated with TXA? All
trauma patients at risk of death due to bleeding should
be treated with TXA. Treatment should be initiated as
soon as possible but not beyond 3 h of injury. The
observation that TXA is most effective when treatment is
initiated within an hour of injury suggests that treatment
should be started by emergency medical services at the
site of the injury or in transit to hospital. If this is not
possible, it should be given immediately on arrival at
hospital. Because TXA does not appear to have any seri-
ous adverse effects, it can be administered safely to a
wide spectrum of patients with traumatic bleeding and
should not be restricted to the most severely injured. Cer-
tainly, any patient requiring a blood transfusion should
be treated with TXA if they are within 3 h of their
injury.
Although the above recommendation may seem obvi-
ous in the light of the evidence for the effectiveness and
safety of TXA in surgery and trauma, the increased
availability of thromboelastography (TEG and ROTEM)
and the burgeoning interest in ‘acute coagulopathy of
trauma’ have led some authors to recommend restricting
TXA use to patients with a diagnosis of ‘hyperfibrinoly-
sis’ or else to the most severely injured patients [18,19].
This misguided recommendation could result in thou-
sands of avoidable deaths. For example, Chapman and
colleagues used thromboelastography (TEG) in an
attempt to define a threshold value for treating fibrinoly-
sis with TXA [18]. They examined the relationship
© 2015 International Society on Thrombosis and Haemostasis

Effect of TXA administration on
haemorrhage death by time since injury
≤1 hour
>1 to ≤ 3 hours
>3 hours
0.85 (0.76–0.96); p<0.0077
.7
Fig. 1. Effect of TXA administration on haemorrhage death by time since injury.
between clot lysis 30 min after maximum clot strength
(LY30) and the risk of hemorrhage death and concluded
that an LY30 of 3% or more defines ‘clinically relevant
hyperfibrinolysis’ that should be treated with TXA.
Although the risk of hemorrhage death in patients with
an LY30 of 3% or more was higher than in those who
did reach this threshold, nevertheless, 5% of patients
with an LY30 lower than the threshold value died from
hemorrhage. If TXA caused serious adverse effects or
was tremendously expensive, we might have to consider
carefully the value of treating patients who face a lower
risk of hemorrhage death. But since TXA is safe and
inexpensive, it is hard to understand why it should not
be offered to patients who face a 5% risk of bleeding to
death. Data from the CRASH-2 trial show that TXA
reduces the risk of hemorrhage death by about one-third
regardless of the baseline risk. Indeed, the risk reduction
might be greater in patients at low risk of death as a
smaller proportion of these patients have unsurvivable
bleeding.
According to the results from Chapman et al., obtained
from a sample of 73 trauma patients, the 3% LY30
threshold has a 64% sensitivity for identifying patients
who will bleed to death. Consequently, large numbers of
patients who could benefit from TXA administration
would go untreated using this threshold. We should not
be surprised by such low sensitivity. Why should the vis-
coelastic properties of venous blood collected from the
arm reflect the extent of fibrinolytic activity in a bleeding
internal organ, let alone accurately predict the overall risk
of bleeding to death? Another important disadvantage of
using thromboelastography to determine which patients
to treat with TXA is that early TXA treatment is essential
and thromboelastography takes time. For these reasons, I
believe that the recommendation that only patients with a
LY30 of 3% or more should be treated with TXA is
flawed. It could result in thousands of avoidable deaths
and should be disregarded.
© 2015 International Society on Thrombosis and Haemostasis
Tranexamic acid in trauma S197
RR (95% Cl) Heterogeneity p=0.000008
RR=0.68 (0.57-0.82) p<0.0001
RR=0.79 (0.64-0.97) p=0.03
RR=1.44 (1.12-1.84); p=0.004
.8 .9 1 1.1 1.2 1.3 1.4 1.5
Treat on the basis of the risk of hemorrhage death
(disregard measures of fibrinolysis)
Other authors have suggested treating only the most
severely injured patients. For example, after reviewing the
incidence and pathophysiology of ‘acute coagulopathy of
trauma,’ Napolitano et al. recommend that TXA is used
only ‘in adult trauma patients with severe hemorrhagic
shock (Systolic Blood Pressure ≤75 mm Hg), with known
predictors of fibrinolysis, or with known fibrinolysis by
TEG (LY30 > 3%).’[19] Considering that TXA reduces
bleeding and that early treatment is much more effective,
it would seem sensible to give TXA before the blood pres-
sure falls to hazardously low levels. Patients with a high
risk of death have the most to gain from TXA use,
because the absolute benefits increase as the baseline risk
increases. However, there are more low-risk trauma
patients than high-risk patients and treating a large num-
ber patients at lower risk could prevent as many or more
deaths than treating a smaller number of patients at high
risk [10].
To reduce the risk of hemorrhage death, it is not the
degree of fibrinolysis that matters but the extent to which
bleeding presents a threat to life and the potential of
TXA administration to reduce bleeding. It does not fol-
low that a biological parameter has to be abnormally
high for there to be a benefit in reducing it. Even patients
with ‘normal’ levels of fibrinolysis could be saved by
TXA administration if this reduces life-threatening bleed-
ing. By analogy, cholesterol reduction is the biological
mechanism through which statins reduce the risk of acute
myocardial infarction but this does not mean that only
patients with abnormally high cholesterol levels will bene-
fit from statin treatment. Statins also reduce the risk of
myocardial infarction in patients with ‘normal’ cholesterol
levels. If fibrinolysis was completely absent throughout
the body, and then provided that we have correctly
inferred the mechanism through which TXA reduces mor-
S198 I. Roberts
tality, we might not expect clinical benefit from TXA use.
But there is no reason to believe that it would ever be
completely absent and whatever the level of fibrinolytic
activity in a bleeding trauma patient, a reduction in fibri-
nolysis could have a benefit.
The decision to treat a patient with TXA should be
based on their estimated risk of death (prognosis) rather
than an arbitrary ‘diagnosis’ based on a single parameter.
Well-validated prognostic models are available for this
purpose and illustrate the value of multivariate prediction
[20]. For example, a 75-year-old with blunt trauma and a
systolic blood pressure of 110 mm Hg (heart rate
80 bpm, respiratory rate 15 bpm, GCS 15) has a similar
risk of death to a 45-year-old patient with exactly the
same parameters but a systolic blood pressure of 60 mm
Hg. Older patients are less able to tolerate blood loss.
Even a slight reduction in bleeding with TXA could pre-
vent death in an older adult at high risk.
Early administration of TXA to a wide range of
trauma patients, including those with normal physiology
in whom bleeding is suspected, should be considered a
preventive measure rather than a treatment for ‘acute
traumatic coagulopathy’. If traumatic bleeding continues
and blood pressure falls this could result in vicious circle
of tissue hypoperfusion and hypoxia leading to defective
coagulation and worse bleeding. Tissue plasminogen acti-
vator is stored within the vascular endothelium in secre-
tory organelles called Weibel Palade bodies. Because the
vascular endothelium contains preformed stores of t-PA,
fibrinolysis can be rapidly initiated in response to fibrino-
lytic triggers. A range of stimuli can induce the release of
t-PA from Weibel Palade bodies, including hypoxia,
trauma, or inflammatory mediators [21]. Tissue hypoper-
fusion can cause massive t-PA release resulting in sys-
temic fibrinolysis. It seems reasonable to take early
measures to reduce bleeding before it leads to a positive
feedback of hypoperfusion and hypoxia leading to sys-
temic fibrinolysis and worsening blood loss. Early TXA
administration by reducing traumatic bleeding could avert
potentially fatal ‘coagulopathic’ bleeding. The challenge
now is to ensure that all bleeding trauma patients whether
they may be in high-, middle-, or low-income settings
receive early TXA treatment.
Uncertainties that should be resolved
Knowing that TXA reduces deaths from extra-cranial
bleeding raises the possibility that it might also improve
outcome in patients with isolated traumatic brain injury.
Intracranial bleeding is common after Traumatic Brain
Injury (TBI) and increases the risk of death and disabil-
ity. Many patients with TBI have an initial lucid period
but then deteriorate due to intracranial bleeding. Three
quarters of those with moderate or severe TBI have intra-
cranial hemorrhage on CT scan and those with a large
bleed have a worse outcome than those with a small
bleed. Intracranial bleeding continues after hospital
admission. Among patients with moderate or severe TBI,
intracranial bleeding continues after admission in 84%.
TXA has the potential to reduce intracranial bleeding
and improve patient outcomes in TBI patients. The
CRASH-3 trial is a multi-center, randomized, and pla-
cebo-controlled trial of the effects of early administration
(within 8 h of injury) of tranexamic acid on death, dis-
ability, and vascular occlusive events in TBI patients [22].
A total of 10 000 adult TBI patients who fulfill the eligi-
bility criteria will be randomized to receive TXA or
matching placebo. We hypothesize that TXA administra-
tion will reduce death and disability by reducing intracra-
nial bleeding. If shown to be safe and effective, TXA
could be a highly cost-effective intervention for a major
global health problem. The trial is ongoing and clinicians
interested in taking part should visit the trial website for
further details http://crash3.lshtm.ac.uk/.
Further research should explore the clinical importance
of any anti-inflammatory effects of TXA. Despite the
results of in vitro studies, there is uncertainty about the
importance of any anti-inflammatory effects of TXA
in vivo [23]. Although analysis of data from the CRASH-
2 trial supports the premise that TXA improves survival
by reducing bleeding, further studies are required to test
the hypothesis that TXA might also have clinically rele-
vant anti-inflammatory effects [14].
Disclosure of Conflict of Interests
The author states that he has no conflict of interest.
References
1 Cesarman-Maus G, Hajjar K. Molecular mechanisms of fibrino-
lysis. Br J Haematol 2005; 129: 307–21.
2 Lucas M, Fretto L, McKee P. The binding of human plasmino-
gen to fibrin and fibrinogen. J Biol Chem 1983; 258: 4249–56.
3 Takada A, Makino Y, Takada Y. Effects of tranexamic acid on
fibrinolysis, fibrinogenolysis and amidolysis. Thromb Res 1986;
42: 39–47.
4 Lighvani S, Baik N, Diggs JE, Khaldoyanidi S, Parmer RJ,
Miles LA. Regulation of macrophage migration by a novel plas-
minogen receptor Plg-RKT. Blood 2011; 118: 5622–30.
5 Gong Y, Hart E, Shchurin A, Hoover-Plow J. Inflammatory
macrophage migration requires MMP-9 activation by plasmino-
gen in mice. J Clin Invest 2008; 118: 3012–24.
6 Busuttil SJ, Ploplis VA, Castellino FJ, Tang L, Eaton JW, Plow
EF. A central role for plasminogen in the inflammatory response
to biomaterials. J Thromb Haemost 2004; 2: 1798–805.
7 Andronicos NM, Chen EI, Baik N, Bai H, Parmer CM, Kiosses
WB, Kamps MP, Yates JR 3rd, Parmer RJ, Miles LA. Proteo-
mics-based discovery of a novel, structurally unique, and devel-
opmentally regulated plasminogen receptor, Plg-RKT, a major
regulator of cell surface plasminogen activation. Blood 2010; 115:
1319–30.
8 Syrovets T, Jendrach M, Rohwedder A, Schu¨le M, Simmet T.
Plasmin-induced expression of cytokines and tissue factor in
human monocytes involves AP-1 and IKKb-mediated NF-kB
activation. Blood 2001; 97: 3941–50.
© 2015 International Society on Thrombosis and Haemostasis

9 Ker K, Edwards P, Perel P, Shakur H, Roberts I. Effect of tra-
nexamic acid on surgical bleeding: systematic review and cumula-
tive meta-analysis. BMJ 2012; 344: e3054.
10 Ker K, Prieto-Merino D, Roberts I. Systematic review, meta-
analysis and meta-regression of the effect of tranexamic acid on
surgical blood loss. B J Surg 2013; 100: 1271–9.
11 CRASH-2 trial collaborators, Shakur H, Roberts I, Bautista R,
Caballero J, Coats T, Dewan Y, El-Sayed H, Gogichaishvili T,
Gupta S, Herrera J, Hunt B, Iribhogbe P, Izurieta M, Khamis
H, Komolafe E, Marrero MA, Mejıa-Mantilla J, Miranda J,
Morales C, et al. Effects of tranexamic acid on death, vascular
occlusive events, and blood transfusion in trauma patients with
significant haemorrhage (CRASH-2): a randomised, placebo-con-
trolled trial. Lancet 2010; 376: 23–32.
12 Ker K, Kiriya J, Perel P, Edwards P, Shakur H, Roberts I.
Avoidable mortality from giving tranexamic acid to bleeding
trauma patients: an estimation based on WHO mortality data, a
systematic literature review and data from the CRASH-2 trial.
BMC Emerg Med 2012; 12: 3.
13 CRASH-2 collaborators, Roberts I, Shakur H, Afolabi A, Brohi
K, Coats T, Dewan Y, Gando S, Guyatt G, Hunt BJ, Morales
C, Perel P, Prieto-Merino D, Woolley T.The importance of early
treatment with tranexamic acid in bleeding trauma patients: an
exploratory analysis of the CRASH-2 randomised controlled
trial. Lancet 2011; 377: 1096–101.
14 Roberts I, Prieto-Merino D, Manno D. Mechanism of action of
tranexamic acid in bleeding trauma patients: an exploratory analy-
sis of data from the CRASH-2 trial. Crit Care 2014; 18: 685.
15 Roberts I, Perel P, Prieto-Merino D, Shakur H, Coats T, Hunt
BJ, Lecky F, Brohi K, Willett K, CRASH-2 Collaborators.
Effect of tranexamic acid on mortality in patients with traumatic
bleeding: prespecified analysis of data from randomised con-
trolled trial. BMJ 2012; 345: e5839.
© 2015 International Society on Thrombosis and Haemostasis
Tranexamic acid in trauma S199
16 Pealing L, Perel P, Prieto-Merino D, Roberts I. Risk factors for
vascular occlusive events and death due to bleeding in trauma
patients; an analysis of the CRASH-2 cohort. PLoS ONE 2012;
7: e50603.
17 Guerriero C, Cairns J, Perel P, Shakur H, Roberts I. Cost-effec-
tiveness analysis of administering tranexamic acid to bleeding
trauma patients using evidence from the CRASH-2 trial. PLoS
ONE 2011; 6: e18987.
18 Chapman M, Moore E, Ramos C, Ghasabyan A, Harr J, Chin
T, Stringham J, Sauaia A, Silliman C, Banerjee A. Fibrinolysis
greater than 3% is the critical value for initiation of antifibrino-
lytic therapy. J Trauma Acute Care Surg 2013; 75: 961–7.
19 Napolitano L, Cohen M, Cotton B, Schreiber M, Moore E. Tra-
nexamic acid in trauma: how should we use it? J Trauma Acute
Care Surg 2013; 74: 1575–86.
20 Perel P, Prieto-Merino D, Shakur H, Clayton T, Lecky F, Bou-
amra O, Russell R, Faulkner M, Steyerberg EW, Roberts I. Pre-
dicting early death in patients with traumatic bleeding:
development and validation of prognostic model. BMJ 2012;
345: e5166.
21 Lowenstein CJ, Morrell CN, Yamakuchi M. Regulation of Wei-
bel-Palade body exocytosis. Trends Cardiovasc Med 2005; 15:
302–8.
22 Dewan Y, Komolafe EO, Mejia-Mantilla JH, Perel P, Roberts I,
Shakur H. CRASH-3 - tranexamic acid for the treatment of sig-
nificant traumatic brain injury: study protocol for an interna-
tional randomized, double-blind, placebo-controlled trial. Trials
2012; 13: 87.
23 Renckens R, Weijer S, De Vos AF, Pater JM, Meijers JC, Hack
CE, Levi M, van der Poll T. Inhibition of plasmin activity by
tranexamic acid does not influence inflammatory pathways dur-
ing human endotoxemia. Arterioscler Thromb Vasc Biol 2004; 24:
483–8.