Articles

Effects of a high-dose 24-h infusion of tranexamic acid on

death and thromboembolic events in patients with acute
gastrointestinal bleeding (HALT-IT): an international
randomised, double-blind, placebo-controlled trial
The HALT-IT Trial Collaborators*

Summary
Background Tranexamic acid reduces surgical bleeding and reduces death due to bleeding in patients with trauma. Lancet 2020; 395: 1927–36
Meta-analyses of small trials show that tranexamic acid might decrease deaths from gastrointestinal bleeding. We See Comment page 1885
aimed to assess the effects of tranexamic acid in patients with gastrointestinal bleeding. *Members listed at end of paper
Correspondence to:
Methods We did an international, multicentre, randomised, placebo-controlled trial in 164 hospitals in 15 countries. Clinical Trials Unit, London
Patients were enrolled if the responsible clinician was uncertain whether to use tranexamic acid, were aged above the School of Hygiene & Tropical
Medicine, London WC1 E7HT, UK
minimum age considered an adult in their country (either aged 16 years and older or aged 18 years and older), and haltit@[Link]
had significant (defined as at risk of bleeding to death) upper or lower gastrointestinal bleeding. Patients were
randomly assigned by selection of a numbered treatment pack from a box containing eight packs that were identical
apart from the pack number. Patients received either a loading dose of 1 g tranexamic acid, which was added to
100 mL infusion bag of 0·9% sodium chloride and infused by slow intravenous injection over 10 min, followed by a
maintenance dose of 3 g tranexamic acid added to 1 L of any isotonic intravenous solution and infused at 125 mg/h
for 24 h, or placebo (sodium chloride 0·9%). Patients, caregivers, and those assessing outcomes were masked to
allocation. The primary outcome was death due to bleeding within 5 days of randomisation; analysis excluded patients
who received neither dose of the allocated treatment and those for whom outcome data on death were unavailable.
This trial was registered with Current Controlled Trials, ISRCTN11225767, and [Link], NCT01658124.

Findings Between July 4, 2013, and June 21, 2019, we randomly allocated 12 009 patients to receive tranexamic acid
(5994, 49·9%) or matching placebo (6015, 50·1%), of whom 11 952 (99·5%) received the first dose of the allocated
treatment. Death due to bleeding within 5 days of randomisation occurred in 222 (4%) of 5956 patients in the
tranexamic acid group and in 226 (4%) of 5981 patients in the placebo group (risk ratio [RR] 0·99, 95% CI 0·82–1·18).
Arterial thromboembolic events (myocardial infarction or stroke) were similar in the tranexamic acid group and
placebo group (42 [0·7%] of 5952 vs 46 [0·8%] of 5977; 0·92; 0·60 to 1·39). Venous thromboembolic events (deep vein
thrombosis or pulmonary embolism) were higher in tranexamic acid group than in the placebo group (48 [0·8%] of
5952 vs 26 [0·4%] of 5977; RR 1·85; 95% CI 1·15 to 2·98).

Interpretation We found that tranexamic acid did not reduce death from gastrointestinal bleeding. On the basis of our
results, tranexamic acid should not be used for the treatment of gastrointestinal bleeding outside the context of a
randomised trial.

Funding UK National Institute for Health Research Health Technology Assessment Programme.

Copyright © 2020 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license.

Introduction melaena. Patients are often haemodynamically unstable

Acute severe gastrointestinal bleeding is a common and in need of urgent resuscitation. Acute management
cause of death worldwide.1 Bleeding can occur from of gastrointestinal bleeding includes blood product
the upper or lower gastrointestinal tract, but upper transfusion, medical or endoscopic therapy, and surgery.
gastrointestinal bleeding is more common. The leading Tranexamic acid reduces bleeding by inhibiting blood
causes are peptic ulcer, oesophageal varices, and malig­ clot breakdown (fibrinolysis). Tranexamic acid decreases
nancy. The case fatality rate is approximately 10% for surgical bleeding and reduces death due to bleeding in
upper gastrointestinal bleeding and 3% for lower patients with traumatic and postpartum haemorrhage.5–8 A
gastrointestinal bleeding.2,3 Many patients re-bleed after systematic review and meta-analysis of randomised trials
initial haemostasis and those that do have a four-times of tranexamic acid for upper gastrointestinal bleeding
increased risk of death.4 Patients with acute severe gastro­ included seven trials with a total of 1654 patients.9 There
intestinal bleeding usually present with haematemesis or was a large reduction in all-cause mortality with tranexamic

[Link] Vol 395 June 20, 2020 1927

 Articles

Research in context
Evidence before this study reduce death from gastrointestinal bleeding (RR 0·99, 95% CI
Before this study a Cochrane systematic review and meta- 0·82–1·18) but was associated with an increased risk of venous
analysis of randomised trials of tranexamic acid for upper thromboembolic events (1·85, 1·15–2·98) and seizures (1·73,
gastrointestinal bleeding included seven trials with a total of 1·03–2·93).
1654 patients. There was a large reduction in mortality with
Implications of all the available evidence
tranexamic acid (pooled risk ratio [RR] 0·61, 95% CI 0·42–0·89;
The most recent update of the Cochrane review included eight
p=0·01). However, given the small size of the included trials and
small randomised trials with 1701 participants and showed a
the potential for selection and other biases, we considered this
reduction in mortality with tranexamic acid (RR 0·60, 95% CI
evidence to be hypothesis generating, requiring confirmation
0·42–0·87). Although we cannot entirely rule out a modest
in larger trials. Furthermore, there was substantial uncertainty
increase or decrease in death due to bleeding with tranexamic
about the risk of thromboembolic events with tranexamic acid
acid, we can rule out the large mortality reduction suggested by
(pooled RR 1·86, 95% CI 0·66–5·24).
the Cochrane review. Furthermore, tranexamic acid appears to
Added value of this study increase the risk of venous thromboembolic events in patients
The HALT-IT trial included 12 009 patients from 164 hospitals with gastrointestinal bleeding. On the basis of our results,
in 15 countries. Adult patients with significant upper or lower tranexamic acid should not be used for the treatment of
gastrointestinal bleeding were randomly assigned to receive gastrointestinal bleeding outside the context of a randomised
tranexamic acid (1 g loading dose followed by 3 g maintenance trial. Our results highlight the unreliability of meta-analyses of
dose over 24 h) or matching placebo. Tranexamic acid did not small trials.

acid (risk ratio [RR] 0·61, 95% CI 0·42–0·89; p=0·01). Methods

However, meta-analyses of small trials are prone to Study design and participants
publication and other selection biases, and have a low The HALT-IT trial is an international, randomised,
positive predictive value when compared with results from double blind (participants and trial staff), placebo-
large multicentre trials.10 Furthermore, even in aggregate, controlled trial done in 164 hospitals in 15 countries
the trials included in the meta-analysis were too small to (UK, Pakistan, Nigeria, Egypt, Malaysia, Georgia,
assess the effect of tranexamic acid on thromboembolic Romania, Nepal, Sudan, Saudi Arabia, Spain, Ireland,
adverse events.9 Our objective was to quantify the effects of Albania, Papua New Guinea, and Australia).11 Patients
tranexamic acid on death and thromboembolic events in were enrolled if they were aged above the minimum
acute gastrointestinal bleeding. age considered an adult in their country (either aged
16 years and older or aged 18 years and older) and if
the respon­sible clinician was substantially uncertain
12 009 patients randomly assigned whether to use tranexamic acid. The diagnosis of
significant bleeding was clinical and significant was
defined as a risk of bleeding to death and included
patients with hypotension, tachycardia, or signs of
5994 assigned to tranexamic acid and 6015 assigned to placebo and baseline
baseline data collected data collected shock, or those likely to need transfusion or urgent
endoscopy or surgery.
Severe gastrointestinal bleeding is a frightening
5964 received loading dose 5988 received loading dose
5483 received maintenance dose 5510 received maintenance dose
experience and blood loss can impact on a patient’s
29 received neither dose 23 received neither dose mental and emotional state, impairing their decision
making—the consent procedures considered this, as
well as the need to treat urgently. If the patient was fully
6 consent withdrawn and outcome 11 consent withdrawn and outcome
data unavailable data unavailable competent, written consent was sought. If capacity
3 lost to follow-up was impaired and a personal or professional represen­
tative was available, consent was sought from the
5 consent withdrawn but outcome 7 consent withdrawn but outcome
representative. If neither were able to provide consent,
data provided data provided it was waived and the patient was informed about
the trial and consented for ongoing data collection as
soon as possible afterwards. The trial was approved by
5985 with outcome data 6004 with outcome data
5956 received allocated treatment and 5981 received allocated treatment and the UK NRES Committee East of England (reference
included in the primary analysis included in the primary analysis number 12/EE/0038), and by the national and local
research ethics committees in all participating non-UK
Figure 1: Trial profile countries.

1928 [Link] Vol 395 June 20, 2020

 Articles

Tranexamic acid Placebo Tranexamic acid Placebo

(n=5994) (n=6015) (n=5994) (n=6015)
Age at randomisation, years (Continued from previous column)
Mean (SD) 58·1 (17·0) 58·1 (17·0) Heart rate, beats per min
<40 791 (13%) 779 (13%) <77 812 (14%) 756 (13%)
40–59 2356 (39%) 2333 (39%) 77–91 1546 (26%) 1644 (27%)
60–79 2078 (35%) 2130 (35%) 92–107 1760 (29%) 1720 (29%)
≥80 769 (13%) 773 (13%) >107 1864 (31%) 1885 (31%)
Sex Missing 12 (<1%) 10 (<1%)
Female 2142 (36%) 2124 (35%) Signs of shock
Male 3852 (64%) 3891 (65%) Yes 2574 (43%) 2648 (44%)
Time from onset to randomisation, h No 3420 (57%) 3367 (56%)
Mean (SD) 21·4 (36·4) 22·5 (37·8) Rockall score
≤3 960 (16%) 975 (16%) 1–2 1419 (24%) 1395 (23%)
>3–≤8 1607 (27%) 1551 (26%) 3–4 2306 (38%) 2332 (39%)
>8 3427 (57%) 3488 (58%) 5–7 2269 (38%) 2288 (38%)
Missing 0 1 (<1%) Taking anticoagulants
Suspected location of bleeding Yes 528 (9%) 500 (8%)
Lower 674 (11%) 654 (11%) No 5422 (90%) 5466 (91%)
Upper 5320 (89%) 5361 (89%) Unknown 44 (1%) 49 (1%)
Haematemesis Emergency admission
Yes 4285 (72%) 4240 (71%) Yes 5673 (95%) 5687 (94%)
No 1709 (29%) 1775 (30%) No 321 (5%) 328 (6%)
Melaena or fresh blood per rectum Major comorbidities
Yes 4573 (76%) 4626 (77%) Cardiovascular 1108 (18%) 1132 (19%)
No 1421 (24%) 1389 (23%) Respiratory 337 (6%) 324 (5%)
Suspected variceal bleeding Liver 2432 (41%) 2532 (42%)
Yes 2694 (45%) 2739 (46%) Renal 325 (5%) 310 (5%)
No 3300 (55%) 3276 (54%) Malignancy 417 (7%) 382 (6%)
Suspected active bleeding Other 999 (17%) 968 (16%)
Yes 5247 (88%) 5226 (87%) Any comorbidity 4308 (72%) 4329 (72%)
No 747 (12%) 789 (13%) Data are n (%) or mean (SD).
Systolic blood pressure, mm Hg
Table 1: Baseline characteristics
≥90 5222 (87%) 5216 (87%)
76–89 577 (10%) 577 (10%)
≤75 181 (3%) 201 (3%)
to determine the contents. Block randomisation was
Missing 14 (<1%) 21 (<1%)
used but randomisation was not stratified.
(Table 1 continues in next column)
Procedures
Eligible patients were randomly assigned to get
Randomisation and masking tranexamic acid or placebo as soon as possible and
An independent statistician from Sealed Envelope (London, treatment was started immediately. A loading dose of 1 g
UK) generated randomisation numbers and these were tranexamic acid or placebo (sodium chloride 0·9%) was
given to Sharp Clinical Services UK (Crickhowell, UK), a added to a 100 mL infusion bag of 0·9% sodium chloride
Good Manufacturing Practice certified clinical trial service and infused by slow intravenous injection over 10 min,
provider, to make treatment packs. When a patient was followed by a maintenance dose of 3 g tranexamic acid or
enrolled, the lowest numbered treatment pack was taken placebo added to 1 L of any isotonic intravenous solution
from a box of eight packs. Sharp Clinical Services was and infused at 125 mg/h for 24 h. Every patient was
responsible for masking, which involved removing the assigned a uniquely numbered treatment pack, which
manufacturer’s label and replacing it with the clinical contained eight ampoules of tranexamic acid 500 mg or
trial label and randomisation number. Apart from the placebo, one 100 mL bag of 0·9% sodium chloride (to use
randomisation number, the pack label text was identical with the loading dose), two sterile 10 mL syringes and
for tranexamic acid and placebo. Patients, caregivers, needles, stickers with the trial details and randomisation
and those assessing outcomes were masked to allo­ number (for attaching to infusion bags, forms, and the
cation. We checked the coding by testing each batch of medical records), and instructions. Pfizer, Sandwich, UK
ampoules with high-performance liquid chromatography (PL 00057/0952) manufactured the tranexamic acid

[Link] Vol 395 June 20, 2020 1929

 Articles

and Torbay and South Devon NHS Foundation Trust flagged in the trial database and the investigators were
(MIA [IMP] 13079) manufactured the sodium chloride retrained.
0·9% placebo. We provided information for patients
and representatives, consent forms, and data collection Outcomes
forms. Stickers, instructions, leaflets, and forms were in The primary outcome was death due to bleeding within
local languages. 5 days of randomisation. Cause of death was assigned by
Once randomly assigned, we collected outcome data local principal investigators who provided a narrative of
even if the treatment was not given. Outcome data events leading to death. These were reviewed by the chief
were collected at death, discharge from the randomising investigator (masked to treatment allocation) and queried
hospital, or 28 days after randomisation, whichever if more information was needed to confirm whether death
occurred first. Trial investigators and their institutions was due to bleeding or another cause. Secondary outcomes
provided direct access to the source data for trial-related were death due to bleeding within 24 h and within 28 days
monitoring, audits, and regulatory inspections. Moni­ of randomisation, all-cause and cause-specific mortality
toring was done according to the Sponsor’s Standard at 28 days, rebleeding within 24 h, within 5 days, and
Operating Procedure and the trial protocol. Formal within 28 days of randomisation, surgery or radiological
inspections were carried out by the relevant Regulatory intervention, blood product trans­fusion, thromboembolic
Agencies including the UK Medicines and Healthcare events (deep vein thrombosis, pulmonary embolism,
products Regulatory Agency, Irish Health Products stroke, and myocardial infarction), seizures, other com­
Regulatory Authority, and Nigeria’s National Agency for plications (including other significant cardiac event,
Food and Drug Administration and Control. Adherence to sepsis, pneumonia, respiratory failure, renal failure, liver
allocation sequence was monitored throughout the trial failure), days in an intensive care unit, and functional
and any out of sequence pack use was automatically status. The diagnosis of rebleeding was made by the
clinician based on established criteria. A diagnosis of
thromboembolic events was made using strict definitions
350 Deaths due to bleeding and diagnostic criteria, including a clinical assessment,
Deaths due to any other causes
diagnostic imaging, biomarker tests, and post-mortem
300 examination. Seizures were diagnosed by clinical assess­
ment. Functional status was measured with the Katz
250 Index of Independence in Activities of Daily Living either
at hospital discharge or in-hospital at 28 days.
Number of deaths

200
Statistical analysis
The sample size calculation was initially based on all-
150
cause mortality as the primary outcome since we expected
that most deaths would be due to bleeding.11 However,
100
while the trial was underway, we observed that over half of
all deaths were due to non-bleeding causes. Accumulating
50 evidence from other large trials of tranexamic acid showed
no apparent effect on non-bleeding deaths.12 Furthermore,
0 patients received tranexamic acid (or placebo) only for
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
their initial bleed and because tranexamic acid has a short
Time since randomisation (days)
half-life (approximately 2 h), it will be largely eliminated
Figure 2: Mortality by days from randomisation within 2 days. As such, we did not expect tranexamic acid
to reduce deaths from rebleeding episodes many weeks
after randomisation. The primary outcome was therefore
Tranexamic acid Placebo Risk ratio changed to death due to bleeding within 5 days of
(n=5956) (n=5981) (95% CI)
randomisation on Nov 21, 2018. Based on the amended
Death due to bleeding within 24 h 124 (2·1%) 120 (2·0%) 1·04 (0·81–1·33) primary outcome, assuming a risk of death due to
Death due to bleeding within 5 days 222 (3·7%) 226 (3·8%) 0·99 (0·82–1·18) bleeding of 4%, a study with 12 000 patients has about
Death due to bleeding within 28 days 253 (4·2%) 262 (4·4%) 0·97 (0·82–1·15) 85% power (two-sided α of 5%) to detect a clinically
Rebleeding within 24 h* 41 (0·7%) 41 (0·7%) 1·00 (0·65–1·55) important 25% relative reduction in death due to bleeding
Rebleeding within 5 days* 287 (4·8%) 315 (5·3%) 0·91 (0·78–1·07) from 4% to 3%.
Rebleeding within 28 days* 410 (6·8%) 448 (7·5%) 0·92 (0·81–1·05) We published the statistical analysis plan before
Data are n (%) and risk ratio (95% CI). Death or rebleeding in hospital during follow-up. *Excludes 13 patients missing unblinding.13 The plan gave our reasons for amending
data on rebleed status or rebleed date. the primary outcome measure and for increasing
the sample size. The main analyses compared those
Table 2: Effect of tranexamic acid on death due to bleeding and rebleeding
allocated tranexamic acid with those allocated to placebo

1930 [Link] Vol 395 June 20, 2020

 Articles

on a modified intention-to-treat basis, excluding patients unmasked to treatment (17 patients because the hospital
who received neither dose of the allocated treatment and team wanted to admin­ister tranexamic acid, six because of
those for whom outcome data on death were unavailable. adverse events, three because of clinical concerns, and
We present effect estimates (RRs) with a measure of two as part of post-mortem investigations [13 in the
precision (95% CI). The safety of partici­pants was over­ tranexamic acid group and 15 in the placebo group]).
seen by an independent data monitoring committee, 52 patients received neither dose of the allocated trial
which reviewed four non-masked interim analyses. treatment (29 patients in the tranexamic acid group and
We planned to report four subgroup analyses to examine 23 patients in the placebo group).
the effects of tranexamic acid on the primary outcome 223 patients received antifibrinolytic drugs as part of
stratified by the following baseline characteristics: their clinical care, outside of the trial protocol (105 patients
time to treatment (≤3 h, >3 h), site of bleeding (upper vs in the tranexamic acid group and 118 patients in the
lower gastrointestinal), suspected variceal bleeding and placebo group).
comorbid liver disease compared with other or unknown Baseline characteristics were similar between groups
causes of bleeding, and by clinical Rockall score. We (table 1). Figure 2 shows the number of deaths and cause
modelled an interaction between the treatment effect of death by days since randomisation. There were
and time to treatment with time to treatment as a 1112 deaths. The median time to death was 55 h after
continuous variable. We did post-hoc subgroup analyses randomisation (IQR 18·2–161·8).
to examine the effects of tranexamic acid on the primary
outcome stratified by World Bank country income level
Tranexamic acid Placebo Risk ratio 95% CI
(high vs low and middle income), anticoagulant use, and (n=5956) (n=5981)
systolic blood pressure.
Time since onset
This trial was registered with Current Controlled Trials,
≤3 h 52 (5·4%) 48 (4·9%) 1·10 (0·75–1·61)
ISRCTN11225767, and [Link], NCT01658124.
>3 h 170 (3·4%) 178 (3·6%) 0·96 (0·78–1·18)
p=0·53
Role of the funding source Bleed location
The funders had no role in study design, data collection, Upper 212 (4·0%) 220 (4·1%) 0·97 (0·81–1·17)
data analysis, data interpretation, or writing of the report. Lower 10 (1·5%) 6 (0·9%) 1·61 (0·59–4·40)
The corresponding authors had full access to all the data p=0·34
in the study and had final responsibility for the decision Variceal or liver
to submit for publication. Yes 160 (5·5%) 165 (5·5%) 1·01 (0·81–1·24)
No or unknown 62 (2·0%) 61 (2·1%) 0·99 (0·70–1·40)
Results p=0·94
Rockall score
We enrolled the first patient on July 4, 2013, and the last on
1–2 17 (1·2%) 26 (1·9%) 0·64 (0·35–1·18)
June 21, 2019. We stopped recruiting when the planned
3–4 63 (2·7%) 65 (2·8%) 0·98 (0·70–1·38)
sample size was reached. When the decision to refine the
5–7 142 (6·3%) 135 (5·9%) 1·06 (0·84–1·33)
primary outcome was made in Nov 21, 2018, we had
p=0·32
recruited 10 190 patients. This decision was made blind to Total 222 (3·7%) 226 (3·8%) 0·99 (0·82–1·18)
the accumulating trial data. 12 009 patients were enrolled
and randomly assigned to receive either tranexamic acid 0·35 1·0 1·6
(n=5994, 49·9%) or matching placebo (n=6015, 50·1%),
Figure 3: Effect of tranexamic acid on death due to bleeding within 5 days
of whom 11 952 (99·5%) received the first dose of the Analysis stratified by time since bleeding onset, suspected bleed location, suspected variceal bleeding or comorbid
allocated treatment (figure 1). 29 patients (11 in the liver disease, and Rockall score.
tranexamic acid group and 18 in the placebo group)
withdrew consent after randomisation, but of those,
Tranexamic acid Placebo Risk ratio
12 (five in the tranexamic acid group and seven in the (n=5956) (n=5981) (95% CI)
placebo group) agreed to provide outcome data or had
Bleeding 253 (4·2%) 262 (4·4%) 0·97 (0·82–1·15)
outcome data collected as part of adverse event reporting.
Thromboembolic event 26 (0·4%) 17 (0·3%) 1·54 (0·83–2·83)
We obtained primary outcome data for all but three
Organ failure 109 (1·8%) 114 (1·9%) 0·96 (0·74–1·25)
patients in the tranexamic acid group. There were
Pneumonia 57 (1·0%) 42 (0·7%) 1·36 (0·92–2·03)
14 protocol violations (seven in the tranexamic acid group
and seven in the placebo group), 11 patients did not meet Sepsis 33 (0·6%) 49 (0·8%) 0·68 (0·44–1·05)

the inclusion criteria (ten received tranexamic acid before Malignancy 65 (1·1%) 40 (0·7%) 1·63 (1·10–2·42)
randomisation [six in the tranexamic acid group and four Other 21 (0·4%) 24 (0·4%) 0·88 (0·49–1·58)
in the placebo group], one in the placebo group was All cause 564 (9·5%) 548 (9·2%) 1·03 (0·92–1·16)
younger than 16 years), and there were three consent Data are n (%) and risk ratio (95% CI). Death in hospital during follow-up.
protocol violations (one in the tranexamic acid group and
Table 3: Effect of tranexamic acid on all-cause death
two in the placebo group). A total of 28 patients were

[Link] Vol 395 June 20, 2020 1931

 Articles

Death due to bleeding within 5 days of randomisation open-label antifibrinolytics were removed from the
(table 2) occurred in 222 (3·7%) of 5956 patients in the analysis, the results were similar (0·97, 0·81–1·17). We
tranexamic acid group and in 226 (3·8%) of 5981 patients examined the effect of tranexamic acid on death due to
in the placebo group (RR 0·99, 95% CI 0·82–1·18). bleeding within 5 days of randomisation in prespecified
Similar results were obtained after adjusting for baseline subgroup analyses stratified by time to treatment
covariates (0·98, 0·82–1·17) and in a per-protocol analysis (heterogeneity p=0·53), location of bleeding (p=0·34),
See Online for appendix (0·94, 0·71–1·23). When the 223 patients who received cause of bleeding (p=0·94), and clinical Rockall score
(p=0·32) but recorded no evidence of heterogeneity for
Tranexamic acid Placebo Outcomes these factors (figure 3). When time since bleeding onset
Interventions
was modelled as a continuous variable there was no
Diagnostic endoscopy 4781/5953 (80·3%) 4729/5978 (79·1%) 1·02 (1·00 to 1·03)
evidence of an interaction (heterogeneity p=0·53).
We examined the effect of tranexamic acid on death
Therapeutic endoscopy 2542/5952 (42·7%) 2658/5978 (44.5%) 0·96 (0·92 to 1·00)
due to bleeding within 5 days of randomisation stratified
Diagnostic radiological procedure 1704/5953 (28·6%) 1744/5978 (29·2%) 0·98 (0·93 to 1·04)
by World Bank country income level (high-income vs
Therapeutic radiological procedure 74/5953 (1·2%) 89/5978(1·5%) 0·83 (0·61 to 1·13)
low-income and middle-income), anticoagulant use
Surgical intervention 146/5953 (2·5%) 158/5978 (2·6%) 0·93 (0·74 to 1·16)
and systolic blood pressure. These exploratory analyses
Any surgical, endoscopic, 5216/5956 (87·6%) 5236/5981 (87·5%) 1·00 (0·99 to 1·01)
or radiological intervention
were not prespecified. The relative risks did not appear
Any transfusion 4076/5951 (68·5%) 4129/5978 (69·1%) 0·99 (0·97 to 1·02)
to vary by country income, anticoagulant use, or systolic
Whole blood or red cells 3984/4076 (97·7%) 4018/4129 (97·3%) 1·00 (1·00 to 1·01)
blood pressure (appendix p 6).
Death due to bleeding within 24 h of randomisation
Frozen plasma 910/4076 (22·3%) 993/4129 (24·0%) 0·93 (0·86 to 1·00)
occurred in 124 (2·1%) patients in the tranexamic
Any platelets 219/4076 (5·4%) 255/4129 (6·2%) 0·87 (0·73 to 1·04)
acid group and 120 (2·0%) patients in the placebo
Blood product transfusions
group (RR 1·04, 95% CI 0·81–1·33). Death due to
Units of whole blood or red cells 2·8 (2·4) 2·9 (2·7) –0·06 (0·05 to –0·18)
bleeding within 28 days of randomisation occurred in
Units of frozen plasma 0·9 (2·4) 1·0 (2·6) –0·05 (–0·01 to –0·23)
253 (4·2%) patients in the tranexamic acid group
Units of any platelets 0·2 (0·9) 0·2 (1·0) –0·02 (0·02 to –0·06)
and 262 (4·4%) patients in the placebo group (0·97,
Data for interventions are n/N (%) and risk ratio (95% CI); data for blood product transfusions are mean (SD) and 0·82–1·15). Death from all-causes within 28 days of
difference in means (95% CI).
randomisation occurred in 564 patients (9·5%) in the
Table 4: Effect of tranexamic acid on the need for surgical, endoscopic, and radiological interventions or tranexamic acid group and in 548 patients (9·2%) in the
blood product transfusion placebo group (1·03, 0·92–1·16; table 3). The proportion
of patients with rebleeding was similar in both groups
at 24 h, 5 days, and 28 days after randomisation (table 3).
Tranexamic acid Placebo Outcomes The proportion of patients who had surgery, radiological
 Complications intervention, and blood product transfusion was also
Any thromboembolic event 86/5952 (1·4%) 72/5977 (1·2%) 1·20 (0·88 to 1·64) similar in both groups (table 4).
Venous events (deep vein 48/5952 (0·8%) 26/5977 (0·4%) 1·85 (1·15 to 2·98) The risk of fatal or non-fatal thromboembolic events
thrombosis, pulmonary embolism) and arterial thromboembolic events (myocardial infarc­
Deep vein thrombosis 23/5952 (0·4%) 12/5977 (0·2%) 1·92 (0·96 to 3·86) tion or stroke) was similar in the tranexamic acid
Pulmonary embolism 28/5952 (0·5%) 16/5977 (0·3%) 1·76 (0·95 to 3·24) group and the placebo group (table 5). The risk of
Arterial events (myocardial 42/5952 (0·7%) 46/5977 (0·8%) 0·92 (0·60 to 1·39) venous thromboembolic events (deep vein thrombosis
infarction, stroke)
or pulmonary embolus) was higher in the tranexamic
Myocardial infarction 24/5952 (0·4%) 28/5977 (0·5%) 0·86 (0·50 to 1·48)
group than in the placebo group (table 5) and similar
Stroke 19/5952 (0·3%) 18/5977 (0·3%) 1·06 (0·56 to 2·02) risk was observed after excluding patients who did not
Renal failure 142/5951 (2·4%) 157/5978 (2·6%) 0·91 (0·73 to 1·14) receive the maintenance dose (42 events with tranexamic
Liver failure 196/5952 (3·3%) 184/5977 (3·1%) 1·07 (0·88 to 1·30) acid vs 20 with placebo); RR 2·11, 95% CI 1·24–3·59). In
Respiratory failure 105/5952 (1·8%) 131/5978 (2·2%) 0·81 (0·62 to 1·04) an exploratory subgroup analysis, the risk of venous
Cardiac event 100/5952 (1·7%) 89/5977 (1·5%) 1·13 (0·85 to 1·50) thromboembolic events was higher in patients with
Sepsis 210/5952 (3·5%) 216/5977 (3·6%) 0·98 (0·81 to 1·18) suspected variceal bleeding or liver disease (14 vs two
Pneumonia 193/5952 (3·2%) 174/5978 (2·9%) 1·11 (0·91 to 1·36) events; 7·26, 1·65–31·90) than in patients with other
Seizure 38/5952 (0·6%) 22/5977 (0·4%) 1·73 (1·03 to 2·93) causes of bleeding (34 vs 24 events; 1·38, 0·82–2·32;
Self-care capacity p=0·035 for heterogeneity). The risk of renal, hepatic,
Days in ICU 0·4 (1·8) 0·4 (2·0) –0·06 (0·01 to –0·13) and respiratory failure, cardiac events, sepsis, and
Katz score 5·5 (1·5) 5·5 (1·4) –0·03 (0·02 to –0·09) pneumonia was similar in tranexamic acid and placebo
Data for complications are n/N (%) and risk ratio (95% CI); data for self-care capacity are mean (SD) and difference in treated patients (table 5). Seizures occurred in 38 patients
means (95% CI). Thromboembolic events and complications are not mutually exclusive. ICU=intensive care unit. on tranexamic acid and 22 on placebo (0·6% vs 0·4%;
1·73, 1·03–2·93; table 5), and after excluding patients
Table 5: Complications and self-care capacity in study groups
who did not receive the maintenance dose the

1932 [Link] Vol 395 June 20, 2020

 Articles

corresponding numbers were 33 versus 17 events (1·95, scenarios, the timing of onset is easy to determine, most
1·09–3·50). patients present early, and there are well documented
The mean number of days spent in intensive care was changes in fibrinolysis that provide a biological rationale
similar in both groups (table 5). The mean score on the for tranexamic acid treatment.15,16 However, in gastro­
Katz Index of Independence in Activities of Daily Living intestinal bleeding it is difficult to determine the time of
was also similar in both groups (table 5). bleeding onset, presentation is often delayed (over
80% of patients presented more than 3 h after bleeding
Discussion onset), and the contribution of increased fibrinolysis to
In this trial, tranexamic acid did not reduce death from bleeding is less clear.
gastrointestinal bleeding but was associated with an Almost half of the patients included in our trial had
increased risk of venous thromboembolic events and suspected variceal bleeding due to liver disease and
seizures. The proportion of patients with rebleeding was because these patients had a greater risk of death, they
similar in the tranexamic acid and placebo groups. accounted for nearly three-quarters of deaths. Recent
The randomisation method ensured that participating research shows that acutely ill patients with cirrhosis
clinicians had no foreknowledge of the treatment allo­ have a mixed fibrinolytic phenotype.17 Some have
cation and placebo control ensured outcome assessment increased fibrinolysis, but others have profound hypo­
was blind to treatment group. The inclusion criteria fibrinolysis. The prevalence of hypofibrinolysis appears
were clinical, reflecting the full range of gastrointestinal to be greatest in the most critically ill patients. Using the
bleeding presentations that doctors face in day-to-day same clot lysis assay, reduced fibrinolysis has been
practice. Baseline prognostic factors were well balanced shown to be associated with a small increased risk of
and almost all randomly assigned patients were followed venous thrombosis.18 In our trial, the increased risk of
up. The primary outcome was death due to bleeding venous thromboembolic events with tranexamic acid
within 5 days of randomisation. Our scientific reasons appeared to be more marked in patients with liver
for pre-specifying death due to bleeding as the primary disease, although this was an explo­ ratory subgroup
outcome in the statistical analysis plan are presented in analysis and there was no strong evidence for hetero­
detail elsewhere.12 Although some misclassification of geneity. Nevertheless, reduced fibrinolysis in patients
cause of death is possible, the assessment was masked with liver disease might explain the absence of reduction
to the treatment group. However, because there was no in bleeding deaths with tranexamic acid and the
evidence of a treatment effect for the prespecified increased risk of venous thromboembolic events.
primary endpoint (death due to bleeding at 5 days) or The dose of tranexamic acid used in this trial was
for death from any cause at 28 days, the choice of higher and the duration of treatment was longer (4 g over
endpoint does not influence the interpretation of the 24 h) than in randomised trials of tranexamic acid in
results. Misclassification might also have affected our trauma (2 g over 8 h) or post-partum haemorrhage (1 g
subgroup analyses because at the time of recruitment bolus with a repeat 1 g dose if bleeding continued), which
the site and cause of bleeding cannot be known with did not record any increase in adverse events with
certainty. Our use of the pre-endoscopy Rockall score tranexamic acid. Patients with gastrointestinal bleeding
might have misclassified baseline risk.14 To minimise often rebleed after initial haemostasis, particularly
the risk of false positives, we used strict criteria to within the first 24 h. Because tranexamic acid has a short
diagnose thromboembolic events, including a positive half-life, we used a longer treatment duration to cover
result on imaging (eg, ultrasound) or at post-mortem this high-risk period. Furthermore, previous trials in
examination. Although using this criteria might have gastrointestinal bleeding that appeared to show a large
led to some under reporting, because the diagnostic mortality reduction with tranexamic acid used a high
tests have high specificity, the relative risk estimates dose and a longer duration of treatment than trials in
should be unbiased. Although some patients received trauma and post-partum haemorrhage.9 The longer
anti­fibrinolytics outside of the protocol, the treatment duration of tranexamic acid treatment in this trial might
effect was the same when these patients were excluded. explain the increased risk of venous thromboembolic
Although this is one of the largest randomised trials events and the higher dose could possibly explain the
in gastrointestinal bleeding, we cannot rule out a increased risk of seizures.19
modest increase or decrease in death due to bleeding In summary, we found no evidence that tranexamic
with tranexamic acid. That said, we can rule out the acid decreases the risk of death in patients with gastro­
large mortality reduction suggested by the Cochrane intestinal bleeding. Our results caution against a uniform
systematic review and meta-analysis of previous small approach to the management of patients with major
trials.9 haemorrhage and highlight the need for randomised
Administration of tranexamic acid within 3 h of trials targeted at specific pathophysiological processes.
bleeding onset reduces death due to bleeding in trauma Because gastrointestinal bleeding is a licensed indi­cation
and post-partum haemorrhage without increasing for tranexamic acid, our results could have regulatory
the risk of thromboembolic events. In these bleeding implications. Although this trial can rule out the large

[Link] Vol 395 June 20, 2020 1933

 Articles

mortality reduction suggested by the meta-analysis of (Saudi Arabia), Conor Deasy (Ireland), Joaquin Alvarez Gregori (Spain),
previous small trials, it cannot rule out more modest Bobby Wellsh (Papua New Guinea), Luke Lawton (Australia).
Trial sites and investigators: UK (4751): Royal Stoke University Hospital
treatment effects. Because tranexamic acid reduces (303): Raghavendra Kamath, Adrian Barry, Racquel Carpio, Kay Finney,
bleeding deaths in patients with traumatic and post­ Holly Maguire; Queen’s Medical Centre Campus Nottingham (208):
partum haemorrhage, individual patient data meta- Martin James, Frank Coffey, Chris Gough, Lisa Sawers, Aye-Aye Thi;
analyses should assess the strength of the evidence that Royal Berkshire Hospital (191): Jonathan Simmons, Claire Burnett,
Nicola Jacques, Victoria Murray; St George’s Hospital (173):
the effectiveness and safety of tranexamic acid varies by Richard Pollok, Heather Jarman, Christine Lambe, Sarah Rounding;
the site and cause of bleeding. Blackpool Victoria Hospital (166): Simon Tucker, Romaih Al-Idari,
The HALT-IT Trial Collaborators Samuel Guest, Emma Stoddard; Queen Elizabeth Hospital Birmingham
Writing Committee: Ian Roberts, Haleema Shakur-Still (Co-Chairs), (150): David Yeo, Colin Bergin, Elaine Hardy, Joanne Thunder;
Adefemi Afolabi, Adegboyega Akere, Monica Arribas, Amy Brenner, University Hospital Coventry (128): Paul Jhalli, Edward Hartley,
Rizwana Chaudhri, Ian Gilmore, Kenneth Halligan, Irshad Hussain, Catherine Jarvis, Carly Swann; Royal Infirmary of Edinburgh (125):
Vipul Jairath, Kiran Javaid, Aasia Kayani, Ton Lisman, Raoul Mansukhani, Matthew Reed, Bernadette Gallagher, Julia Grahamslaw, Rachel O’Brien,
Muttiullah Mutti, Muhammad Arif Nadeem, Richard Pollok, Royal London Hospital (125): Timothy Harris, Geoffrey Bellhouse,
Jonathan Simmons, Majid Soomro, Simon Stanworth, Andrew Veitch. Olivia Boulton, Imogen Skene; Glasgow Royal Infirmary (120):
Trial Steering Committee: Christopher Hawkey (Chair), Adefemi Afolabi, Adrian Stanley, Janet Johnstone, Donogh Maguire, Susan Thornton,
Jack Cuzick, Kenneth Halligan (patient representative), David Henry, University College London Hospital (113): Matthew Banks,
Chris Metcalfe, Ian Roberts. Data Monitoring and Ethics Committee: Georgia Bercades, Daniel Marks, Jung Ryu: Whipps Cross University
Richard Gray (Chair), Alan Barkun, Suresh David, Philip Devereaux, Hospital (108): Timothy Harris, Claire Dowty, Jason Pott, Imogen Skene;
Tony Brady (independent statistician). Protocol Committee: Ian Roberts, John Radcliffe Hospital (107): James East, Adam Bailey, Sally Beer,
Haleema Shakur-Still, Timothy Coats, Phil Edwards, Ian Gilmore, Sian Davies; Royal Devon & Exeter Hospital (97): Andrew Appelboam,
Vipul Jairath, Katharine Ker, Daniela Manno, Simon Stanworth, Daisy Mackle, Jennifer Small; Queen Alexandra Hospital Portsmouth
Andrew Veitch. Clinical Trials Unit (CTU): Monica Arribas (trial (88): Christiane Vorwerk, Rachel Atkins, Isobel Bradbury; Leicester Royal
manager/research assistant), Emma Austin (assistant trial manager), Infirmary (84): Timothy Coats, Catriona Bryceland, Lisa McClelland;
Kiran Bal (assistant trial manager), Eni Balogun (trial manager), Salford Royal Hospital (83): Martin Thomas, Kate Clayton,
Collette Barrow (trial administrator), Danielle Beaumont (senior trial Angiy Michael; Great Western Hospital (80): Stephen Haig,
manager/research fellow), Myriam Benyahia (CTU administrator), Saif Al-Nahhas, Tim Godfrey; Southampton General Hospital (80):
Amy Brenner (research fellow), Imogen Brooks (trial assistant 2016–18), Philip Boger, Rachel Comer, Barbara Watkins; Darlington Memorial
Madeleine Cargill (data assistant), Laura Carrington (trial administrator), Hospital (79): Ola Afolabi, Shazad Afzal, Amanda Cowton; St James
Phil Edwards (statistician 2012–16), Lauren Frimley (trial manager/ University Hospital Leeds (79): Simon Everett, Ruth Fazakerley,
research assistant), Amber Geer (assistant data manager), Daniel Gilbert Felicia Onoviran: Poole Hospital (77): Jonathon Snook, Jackie Berry,
(data assistant 2012–13), Catherine Gilliam (trial administrator), Diane Simpson; King’s College Hospital (73): Jeff Keep, Hannah Cotton,
Julio Gil Onandia (clerical assistant), Nayia Golfi (trial manager 2013–15), Sinead Helyar: University Hospital of North Tees (73): Matthew Rutter,
Daniel Hetherington (trial assistant 2012–15), Courtenay Howe Tracey Johnston, Laura O’Rourke; Basingstoke and North Hampshire
(CTU administrator 2015–17), Carolyn Hughes (data assistant 2016–17), Hospital (72): Louisa Chan, Joanna Tambellini, Dawn Trodd; Dorset
David I’anson (assistant trial manager 2016–17), Rob Jackson (data County Hospital (68): James Shutt, Sarah Moreton, Abby Oglesby;
manager 2012–15), Miland Joshi (statistician 2016–17), Sneha Kansagra Addenbrooke’s Hospital (67): Adrian Boyle, Nicola Haeger,
(assistant trial manager 2016–18), Taemi Kawahara (senior trial manager Susie Hardwick; Southmead Hospital (67): Jason Kendall, Beverley
2012–15), Katharine Ker (lecturer), Sergey Kostrov (systems officer Faulkner, Ruth Worner; Royal Victoria Infirmary (64): Sarah Hearnshaw,
2015–19), Daniela Manno (clinical lecturer 2012–15), Raoul Mansukhani Mary Doona, Maria Price; St Thomas’ Hospital (64): Laura Hunter,
(medical statistician 2019–20), Hakim Miah (IT systems manager Maggie Bell, Vania Loureiro; Derriford Hospital (61): Anthony Kehoe,
2013–19), Bernard Ndungu (assistant trial manager 2016–17), Alison Jefferey, Rosalyn Squire; Ipswich Hospital (60): David Hartin,
Kelly Needham (statistician 2018–20), Aroudra Outtandy (trial assistant Stephanie Bell, Alexandra Newman; Musgrove Park Hospital (59):
2013–15), Daniel Pearson (data assistant 2018–19), Tracey Pepple (acting James Gagg, Jayne Foot, Sue Wakeford: Royal Oldham Hospital (58):
senior data manager 2014–19), Danielle Prowse (assistant data manager), Gabrielle May, Thomas Bartram, Paul Cumpstay; Whittington Hospital
Nigel Quashi (data manager 2013–16), Anna Quinn (data assistant (58): Lucy Parker, Rita Das, Sheik Pahary; Basildon University Hospital
2013–15), Maria Ramos (senior project administrator 2012–15), (57): Gavin Wright, Georgina Butt, Natasha Christmas; Wexham Park
Laura Ranopa (trial assistant 2015–20), Mia Reid (clerical assistant Hospital (56): Sarah Wilson, Mohammed Ashfaq, Louise Chandler;
2016–18), Ian Roberts (Chief Investigator/CTU co-director), Royal United Hospital Bath (50): Saif Al-Nahhas, Carrie Demetriou,
Chris Roukas (trial administrator 2013–15), Haleema Shakur-Still Philip Kaye; Manchester Royal Infirmary (48): Simon Carley,
(Project Director/ CTU co-director), Chelci Squires (trial assistant Andrew Brown: Chesterfield Royal Hospital (44): Lucy Jones,
2014–16), Jemma Tanner (clinical trials associate 2013–16), Amanda Whileman; James Cook University Hospital (44):
Andrew Thayne (data assistant), Ruhama Uddin (trial assistant 2018–19). John Greenaway, Julie Tregonning; Newham University Hospital (44):
Rawalpindi Medical University Pakistan National Coordinating Centre Timothy Harris, Geoffrey Bellhouse; Northern General Hospital (44):
Team: Rizwana Chaudhri (Coordinating Centre Director), Avril Kuhrt, Steve Goodacre; Royal Shrewsbury Hospital (43): John Jones,
Muttiullah Mutti (Clinical Lead), Kiran Javaid (Assistant Research Charlotte Owen; Charing Cross Hospital (41): Anu Mitra,
Coordinator), Aasia Kayani (Research Coordinator). Abby Harper-Payne; Sandwell General Hospital (37): Nigel Trudgill,
Nigeria Coordinating Team, University College Hospital Ibadan: Anne Hayes; South Tyneside District Hospital (36): Faheem Butt,
Bukola Fawole (Coordinating Centre Director), Folasade Adenike Bello Gayle Clifford; Victoria Hospital, Fife (35): Andrew Kinnon,
(Coordinating Centre Director), Oladapo Olayemi (Coordinating Centre Susan Fowler, Chelsea and Westminster Hospital (34): Kris Pillay,
Director), Adefemi Afolabi (National Principal Investigator), Shweta Gidwani; Queen Elizabeth Hospital Woolwich (34):
Olujide Okunade (Assistant Trial Coordinator), Olusade Adetayo Alistair McNair, Omer Omer; Gloucestershire Royal Hospital (31):
(Assistant Trial Coordinator). National Coordinators: Rizwana Chaudhri Tanya de Weymarn, Adnan Amin; Royal Hampshire County Hospital
(Pakistan), Muttiullah Mutti (Pakistan), Adefemi Afolabi (Nigeria), (31): Louisa Chan, Jane Martin; Torbay Hospital (31): Nick Mathieu,
Folasade Adenike Bello (Nigeria), Bukola Fawole (Nigeria), Simon Barnes; York Hospital (31): James Turvill, Helen Sweeting;
Oladapo Olayemi (Nigeria), Hussein Khamis (Egypt), University Hospital Crosshouse (29): Morten Draegebo,
Mohammad Shukri Bin Jahit (Malaysia), Tamar Gogichaishvili (Georgia), Marion McNaught, Worthing Hospital (29): Mandy Grocutt,
Radu Bogdan Mateescu (Romania), Ajay Adhikaree (Nepal), Jordi Margalef; Barnsley Hospital (27): Julia Humphrey, Richard Jackson;
Abdelmounem Eltayeib Abdo (Sudan), Mohammad Zaher North Devon District Hospital (27): Fionn Bellis, Jane Hunt; St Mary’s

1934 [Link] Vol 395 June 20, 2020

 Articles

Hospital (24): Anu Mitra; University Hospital Ayr (22): University of Benin Teaching Hospital (74): Rose Ugiagbe,
Alastair Stevenson, King’s Mill Hospital (19): Nicholas Watson; Alexander Atiri, Enadeghe Eghaghe; Olabisi Onabanjo University
Royal Sussex County Hospital (19): Steven Barden; Forth Valley Royal Teaching Hospital (49): Adeleke Adekoya, Adedayo Oluyomi Tade,
Hospital (16): Stuart Paterson, New Cross Hospital (16): Andrew Veitch; Olatunji Shonoiki; University of Ilorin Teaching Hospital (49):
Cumberland Infirmary Hospital (14): Chris Macdonald; Sunderland Samuel Olatoke, Toafiq Raji; Federal Medical Centre, Owerri (46):
Royal Hospital (14): David Hobday; West Cumberland Hospital (13): Christopher Ekwunife, Chigozirim Onyekpere; Ahmadu Bello
Olu Orugun; Yeovil District Hospital (13): Andrew Allison; Northampton University Teaching Hospital (40): Adamu Ahmed,
General Hospital (12): Tristan Dyer; Royal Lancaster Infirmary (12): Daniyan Muhammad; Lagos University Teaching Hospital (37):
Samuel McBride; Royal Liverpool University Hospital (12): Emuobor Odeghe, Olufunmilayo Lesi, Azeberoje Osueni; Aminu Kano
Wojciech Sawicki; Hull Royal Infirmary (10): Ben Rayner; Frimley Park Teaching Hospital (25): Adamu Samaila, Aminu Nahuche; Ekiti State
Hospital (8): Lynsey Flowerdew; Queen Elizabeth Hospital – Gateshead University Teaching Hospital (24): Akande Ajayi; Irrua Specialist
(7): Jamie Barbour; Salisbury District Hospital (7): Jason Klein; Aintree Teaching Hospital (24): Andrew Dongo; University of Nigeria Teaching
University Hospital (6): Stephen Hood; University Hospital of Wales (5): Hospital, Enugu (18): Uchenna Ijoma; Federal Medical Centre,
Nicola Palmer; Northwick Park Hospital (4): Jacob de Wolff; Colchester Abeokuta (8): Ademola Tolulope Adebanjo; Lagos State University
General Hospital (3): Achuth Shenoy; Birmingham City Hospital (2): Teaching Hospital (7): Rufina Igetei; University of Abuja Teaching
Nigel Trudgill; Royal Bournemouth Hospital (1): Peter Swallow; Hospital (7): Monday Yilkudi; Maitama District Hospital (6):
University Hospital Lewisham (1): Rajaventhan Srirajaskanthan. Kehinde Osisanya; Jos University Teaching Hospital (3):
Pakistan (4420): King Edward Medical University, Mayo Hospital Lahore Edith Nonyelum Okeke; Imo State University Teaching Hospital (2):
(539): Irshad Hussain, Hamza Arshad, Naeem Aslam, Anam Bangash, Oguamanam Okezie Enwere. Egypt (709): Kasr Al Aini Internal
Muhammad Qamar, Haroon Zahoor; Rawalpindi Medical University Medicine Hospital, Cairo University (641): Serag Esmat, Omar Ashoush,
Pakistan, Holy Family Hospital, Unit II (501): Muttiullah Mutti, Mazen Naga, Fady Nagy, Mostafa Saiid, Ahmed Shaker; Mataria
Saba Arshad, Quratul ain Ghalib, Tehseen Hameed, Tayyaba Saif, Teaching Hospital (60): Hussein Khamis, Ashraf Helmy, Saafan Saafan;
Wajahat Shafi; Services Institute of Medical Sciences/ Services Hospital Badr Hospital, Helwan University (8): Mohammed Abdel Monem.
GI/ Medical Unit III Lahore (443): Muhammad Arif Nadeem, Abid Ali, Malaysia (464): Hospital Tengku Ampuan Afzan (123): Jiffre Din,
Shehroze Khan, Muhammad Muaaz, Ahmad Taj; Lady Reading Hospital Khairul Azis, Muhyuddin Brukan, Sanjay Singh; Hospital Universiti
(386): Aamir Ghafoor, Aamir Afridi, Mansoor Ahmad, Mujahid Aslam, Sains Malaysia (110): Andee Zakaria, Shaik Farid, Nizam Hashim,
Sandeep Kumar; Asian Institute of Medical Sciences (360): Masykurin Mafauzy; Hospital Sultanah Bahiyah (71): Wan Najmi,
Majid Soomro, Mohsin Ali, Ubedullah Bughio, Adil Chang, Nil Amri, Xin Yi; Hospital Raja Permaisuri Bainun (63):
Sana Shaikh; Jinnah Postgraduate Medical Centre Karachi (296): Mohammad Hisyam, Elaine Ng, Zuhrirahimi Ramli; Pulau Pinang
Syed Ahmad, Zeeshan Ali, Marium Waqar, Aiman Mushir, Sadaf Sattar; General Hospital (55): Shyang Yee Lim, Kelvin Voon, Sir Young Yam;
DHQ Teaching Hospital Sargodha (207): Saifullah Goraya, Sungai Buloh Hospital (42): Mohammad Jahit, Lee Joon.
Sharmeen Aslam, Nighat Fatima, Saadia Noreen, Sheraz Saleem; Georgia (425): High Technology Medical Center, University Clinic (364):
Rawalpindi Medical University Pakistan, Benazir Bhutto Hospital Unit I Besik Melikidze, Davit Kazaishvili, Tamar Gogichaishvili,
(196): Fazal Rahman, Nadeem Iqbal, Mohammad Khalid, Umar Riaz; Nino Grubelashvili, Baadur Mosidze; Centre of Emergency Surgery and
Rawalpindi Medical University Pakistan, Holy Family Hospital, Unit I Traumatology (61): Gia Tomadze, Avto Megreladze. Romania (287):
(169): Muhammad Umar, Tayyab Akhter, Javaria Khan, Noureen Misbah; Clinical Emergency Hospital Bucharest – SCUB Floreasca (128):
Aziz Bhatti Shaheed Teaching Hospital (160): Muhammad Afzal, Ruxandra Oprita, Dorina Pestroiu Calescu, Camelia Chioncel,
Mobeen Kayani, Syed Shah, Shahida Tarar; Bolan Medical Complex Andrei Ragea; Colentina Clinical Hospital (58): Bogdan Mateescu,
Hospital (149): Sherbat Khan, Yasir Iqbal, Essa Khan, Maqbool Reki; Bogdan Busuioc, Andrei Voiosu; Municipal Emergency Hospital Moinesti
Rawalpindi Medical University Pakistan, Benazir Bhutto Hospital Unit (45): Adrian Cotirlet, Iulia Pintilie; Central University Emergency Military
II (120): Tanveer Hussain, Shafqat Iqbal, Muhammad Khurram, Hospital, “Carol Davila” University of Medicine and Pharmacy, Bucharest
Muhammad Shafi; Ghulam Muhammad Mahar Medical College and (38): Mariana Jinga, Daniel Balaban; Regional Institute of
Teaching Hospital Unit I (120): Abrar Shaikh, Aijaz Ahmed, Gastroenterology & Hepatology “Prof Dr Octavian Fodor” (11):
Ameet Kumar, Pinkey Sachdev; Jinnah Hospital Unit I (88): Marcel Tanțău; University Hospital of Emergency Bucharest (5):
Khalid Mahmood Nasir, Zafar Iqbal Chaudhry, Muhammad Zubair; Lucian Negreanu; Emergency County Hospital Mures (2): Simona Bataga.
Lahore General Hospital Medical Unit I (84): Ghias Tayyab, Nepal (50): Lifeline Hospital (16): Khushboo Priya; Gandaki Medical
Junaid Mushtaq, Muhammad Nasir; Mardan Medical Complex & College & Teaching Hospital (12): Shankar Baral; Nepal Medical College &
Teaching Hospital (80): Amir Khan, Amjad Ali, Sajjad Ali; POF Hospital Teaching Hospital (12): Anuj K.C.; Janaki Medical College (4): Vijay Sah;
(76): Wasim Uddin, Sohaib Ahmed, Tazaeen Kazmi; Ghulam National Medical College & Teaching Hospital (4): Vijay Yadav; Lumbini
Muhammad Mahar Medical College Teaching Hospital (58): Medical College & Teaching Hospital (2): Ajay Adhikaree. Sudan (40):
Saleh Channa, Adeeqa Aman, Mouzam Shaikh; DHQ Hospital Ibn Sina Hospital (40): Abdelmounem Abdo, Dalia Ahmed.
Faisalabad (47): Tahir Rizvi, Amjad Hussain; Rawalpindi Medical Saudi Arabia (19): Prince Mohammed Bin Abdulaziz Hospital (19):
University Pakistan, DHQ Hospital Rawalpindi (47): Marzouqah Al Anazi, Areej Al Balkhi. Ireland (17): Cork University
Haider Zaigham Baqai, Zakawat Rasheed; Shifa International Hospital Hospital (17): Conor Deasy. Spain (17): Hospital Nuestra Señora del Prado
(41): Abdus Khan, Adeela Irfan; Allied Hospital Faisalabad Medical Unit (9): Joaquín Álvarez Gregori; Torrecardenas Hospital (8):
III (38): Aamir Husain, Asifa Aslam; Madinah Teaching Hospital Helio Fornieles Pérez. Albania (16): University Hospital Center “Mother
Faisalabad (34): Khalid Yahya, Salman Azhar; Liaquat National Hospital Teresa” (16): Arben Beqiri. Papua New Guinea (13): Port Moresby General
& Postgraduate Medical Centre (33): Mansoor Ul Haq, Adeel Afzal; Hospital (13): Bobby Wellsh. Australia (11): The Townsville Hospital (11):
Services Institute of Medical Sciences/ Services Hospital GI/ Medical Luke Lawton
Unit II Lahore (33): Muhammad Imran, Iram Saeed; Shaukat Khanum
Declaration of interests
Memorial Hospital & Research Centre (29): Aasim Yusuf,
We declare no competing interests.
Mariam Hassan; Hayatabad Medical Complex (25): Mumtaz Marwat,
Muhammad Ishfaq; Sir Ganga Ram Hospital Medical Unit III (20): Data sharing
Tahir Bashir; Liaquat University of Medical and Health Sciences Medical Following publication of the primary and secondary analyses detailed in
Unit II (17): Santosh Kumar; DHQ Hospital, Narowal (16): Sajjad Yaqoob; this statistical analysis plan, individual de-identified patient data,
DHQ Hospital Khuzdar (8): Abdul Wahid. Nigeria (770): University including data dictionary, will be made available via our data sharing
College Hospital Ibadan (143): Adegboyega Akere, Tinuola Fakoya, portal, The Free Bank of Injury and Emergency Research Data (freeBIRD)
Temitope Oke, Edries Tejan; National Hospital Abuja (109): website indefinitely. This will allow for maximum utilisation of the data to For more on freeBIRD see
Oluwole Olaomi, Olawale Badejo, Okafor Nnaemaka, Nancy Ukwu; improve patient care and advance medical knowledge. The trial protocol, [Link]
Obafemi Awolowo University Teaching Hospital, Ile- Ife (99): statistical analysis plan and trial publications will be freely available For the protocol see [Link]
Olukayode Arowolo, Adewale Aderounmu, Funmilola Wuraola; online. [Link]/

[Link] Vol 395 June 20, 2020 1935

 Articles

Acknowledgments 9 Gluud LL, Klingenberg SL, Langholz E. Tranexamic acid for upper
Funding for the HALT-IT trial was provided by the UK National Institute gastrointestinal bleeding. Cochrane Database Syst Rev 2012;
for Health Research Health Technology Assessment Programme 1: CD006640.
(HTA/11/01/04). Funding for this trial covered trial materials, meetings, 10 Sivakumar H, Peyton PJ. Poor agreement in significant findings
and central organisational costs. The study was designed, conducted, between meta-analyses and subsequent large randomized trials in
analysed, and interpreted by the investigators, entirely independently of perioperative medicine. Br J Anaesth 2016; 117: 431–41.
all funding sources. The views and opinions expressed therein are those 11 Roberts I, Coats T, Edwards P, et al. HALT-IT--tranexamic acid for
of the authors and do not necessarily reflect those of the funders. the treatment of gastrointestinal bleeding: study protocol for a
randomised controlled trial. Trials 2014; 15: 450.
References 12 Brenner A, Arribas M, Cuzick J, et al. Outcome measures in clinical
1 van Leerdam ME. Epidemiology of acute upper gastrointestinal trials of treatments for acute severe haemorrhage. Trials 2018;
bleeding. Best Pract Res Clin Gastroenterol 2008; 22: 209–24. 19: 533.
2 Hearnshaw SA, Logan RF, Lowe D, Travis SP, Murphy MF, 13 Brenner A, Afolabi A, Ahmad SM, et al. Tranexamic acid for acute
Palmer KR. Acute upper gastrointestinal bleeding in the UK: gastrointestinal bleeding (the HALT-IT trial): statistical analysis plan
patient characteristics, diagnoses and outcomes in the 2007 UK for an international, randomised, double-blind, placebo-controlled
audit. Gut 2011; 60: 1327–35. trial. Trials 2019; 20: 467.
3 Oakland K, Guy R, Uberoi R, et al. Acute lower GI bleeding in the 14 Stanley AJ, Laine L, Dalton HR, et al. Comparison of risk scoring
UK: patient characteristics, interventions and outcomes in the first systems for patients presenting with upper gastrointestinal
nationwide audit. Gut 2018; 67: 654–62. bleeding: international multicentre prospective study. BMJ 2017;
4 Cook DJ, Griffith LE, Walter SD, et al. The attributable mortality 356: i6432.
and length of intensive care unit stay of clinically important 15 Coats TJ, Morsy M. Biological mechanisms and individual variation
gastrointestinal bleeding in critically ill patients. Crit Care 2001; in fibrinolysis after major trauma. Emerg Med J 2020; 37: 135–40.
5: 368–75. 16 Roberts I, Shakur H, Fawole B, et al. Haematological and
5 Myles PS, Smith JA, Forbes A, et al. Tranexamic acid in patients fibrinolytic status of Nigerian women with post-partum
undergoing coronary-artery surgery. N Engl J Med 2017; haemorrhage. BMC Pregnancy Childbirth 2018; 18: 143.
376: 136–48. 17 Blasi A, Patel V, Adelmeijer J, et al. Mixed fibrinolytic phenotypes in
6 Shakur H, Roberts I, Bautista R, et al. Effects of tranexamic acid on decompensated cirrhosis and acute-on-chronic liver failure with
death, vascular occlusive events, and blood transfusion in trauma hypofibrinolysis in those with complications and poor survival.
patients with significant haemorrhage (CRASH-2): a randomised, Hepatology 2020; 71: 1381–90.
placebo-controlled trial. Lancet 2010; 376: 23–32. 18 Lisman T. Decreased plasma fibrinolytic potential as a risk for venous
7 CRASH-3 trial collaborators. Effects of tranexamic acid on death, and arterial thrombosis. Semin Thromb Hemost 2017; 43: 178–84.
disability, vascular occlusive events and other morbidities in 19 Lecker I, Wang DS, Whissell PD, Avramescu S, Mazer CD,
patients with acute traumatic brain injury (CRASH-3): Orser BA. Tranexamic acid-associated seizures: causes and
a randomised, placebo-controlled trial. Lancet 2019; 394: 1713–23. treatment. Ann Neurol 2016; 79: 18–26.
8 WOMAN Trial Collaborators. Effect of early administration of
tranexamic acid on mortality, hysterectomy, other morbidities in
women with postpartum haemorrhage (The WOMAN trial):
a randomised, placebo-controlled trial. Lancet 2017; 389: 2105–16.

1936 [Link] Vol 395 June 20, 2020