EBQ:Halt It Trial

incomplete Journal Club Article

Roberts I, Shakur-Still H, Afolabi A, et al.. "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". Lancet. 2020. 395(10241):1927-1936.
PubMed Full text PDF

Clinical Question

Does high-dose IV tranexamic acid reduce death due to bleeding within five days in adult patients with acute gastrointestinal bleeding compared to placebo?

Conclusion

Tranexamic acid confers no benefit on 5-day mortality due to bleeding in adult patients with acute GI bleed.

Major Points

Tranexamic acid (TXA) is a synthetic lysine derivative that decreases fibrinolysis by inhibiting the conversion of plasminogen to plasmin. It has been used broadly in trauma patients with hemodynamically significant bleeding and appears to confer a mortality benefit in this population. A prior systematic review of TXA for use in acute gastrointestinal bleeding suggested a mortality benefit, although the review was based on a series of small studies prone to bias.

HALT-IT is the first large, pragmatic, international, double-blind RCT to systematically evaluate the use of TXA in acute undifferentiated GI bleed. In this study, patients that received TXA 1g IV loading dose followed by a 3g infusion over 24 hours had the same mortality rate from bleeding at five days as patients receiving a matching normal saline placebo (222 [3.7%] intervention group vs 226 [3.8%] placebo, RR 0.99). All-cause mortality at 28 days was also no different between groups. There was a suggestion of increased venous thromboembolic events in the TXA group (48 [0.8%] vs 26 [0.4%], RR 1.85 [CI 1.15-2.98]).

Study Design

Large, pragmatic, multicenter international placebo-controlled RCT, conducted at 164 hospitals in 15 countries

• 12,009 patients enrolled with acute GI bleeding
  • Tranexamic acid (n=5,994)
    • Placebo (n=6,015)

Population

Patient Demographics

• Patients older (mean age 58), male (65%), and mostly upper GI bleeds (90%)
• Mean time to randomization from onset of bleeding 22h
• Most patients (57%) did not display signs of shock on initial presentation

Inclusion Criteria

• Adult patients with acute gastrointestinal bleeding
• Clinician defined bleeding as significant, meaning possibility of death from bleeding
  • Age of inclusion either 16 or 18 depending on definition of "adult" in home country
• Clinician had to be "uncertain" whether or not to use TXA

Exclusion Criteria

• Clinician-determined clear indication for TXA
• Clinician-determined clear contraindication for TXA

Interventions

• TXA 1g IV over 10 minloading dose followed by a 3g infusion in NS over 24 hours

Outcomes

Primary Outcome

• Death from bleeding at 5 days
  • No difference between TXA (222, 3.7%)and placebo (226, 3.8%)

Secondary Outcomes

• All-cause mortality at 28 days
  • Originally was the primary outcome, replaced by the above (see discussion)
  • No difference between TXA (9.5%) and placebo (9.2%), RR 1·03, 95% CI 0·92– 1.16
• No difference in rebleeding, need for endoscopy/surgery, transfusion requirement, or mortality from other subgroups (i.e. infection, organ failure, etc)
• Thromboembolic events
  • Any: 1.4% in TXA group vs 1.2% in placebo (RR 1.20, 95% CI 0.88-1.64)
  • Venous (DVT and PE): 0.8% in TXA group vs 0.4% in placebo (RR 1.85, 95% CI 1.15-2.98)
• Seizures
  • 0.6% in TXA group vs 0.4% in placebo (RR 1.73, 95% CI 1.03-2.93

Subgroup analysis

4 preplanned subgroups within the primary outcome:

• Time from onset of bleeding, greater or less than 3h
• Site of bleed, upper or lower
• Varices and comorbid liver disease vs other causes of bleeding
• Rockall score

TXA showed no benefit in any subgroup.

Criticisms & Further Discussion

• Previous positive studies of TXA on GI bleeds have been small or observational and subject to bias
• This large RCT is well-designed, pragmatic, and answers a clinically relevant question
• The study authors changed the original primary outcome from all-cause 28-day mortality to death due to bleeding at 5 days
  • During an interim analysis authors noted that >50% of deaths were from non-hemorrhagic causes
  • New primary outcome is disease-specific and at the judgment of the clinician
• Patients for whom TXA was thought to be "clearly indicated" were excluded from the trial
  • Some sicker patients who could have benefited from TXA may have been excluded
• Time to randomization was long (22 hours on average); CRASH-2 signaled that earlier administration of TXA is better
  • However, GI bleed patients tend to present roughly in the timeframe represented in the trial

See Also

• CRASH-2 Trial
• Tranexamic acid
• MATTERs Study

Funding

The UK National Institute for Health Research Health Technology Assessment Programme