HALT IT – TXA for GI Bleeds

For this month’s physically distanced Journal Club – a review of the HALT IT trial looking at TXA for GI Bleeds.

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 Lancet. 2020; 395(10241):1927-1936.  https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)30848-5/fulltext

TL;DR:

• Tranexemic acid (TXA) did NOT reduce 5-day mortality in upper and lower GI bleed patients.
• It did show small increase risk of VTE.
• Well designed, large trial. Best evidence to date for TXA in GI bleed.

Background:

• GI bleeds have a high mortality rate (10%)
• TXA is an antifibrinolytic agent that has been shown to be effective at preventing bleeding complications in a variety of settings (surgery, trauma, epistaxis)
  • We love it for its low cost, minimal SE profile, and numerous indications in the ER
• A 2012 Cochrane Review for UGIB showed a reduction in all-cause mortality with TXA for GI bleeds
  • However, individual trials were small and prone to biases making it difficult to draw definitive conclusions… but was the best evidence we had until now

Clinical Question:

• Does IV tranexamic acid reduce 5-day death due to bleeding?

Methods:

• International, multi-centre, randomized, double-blind, placebo controlled trial (15 countries, 164 hospitals)
• Intervention:
  • 1g TXA IV over 10 min the 125mg/hr x 24 hours (3g) vs. Placebo

Patients:

• Adults (>16/18 yo depending on country)
• With “significant” acute GI bleeding
  • Risk of bleeding to death:  
    • Hypotension
    • Tachycardia
    • Signs of shock
    • Likely to need transfusion, urgent endoscopy or surgery
• Treating clinician had to be “substantially uncertain” about whether to use TXA

Outcome:

• Primary outcome: death due to bleeding at 5 days
• Secondary outcomes:
  • Death due to bleeding at 24 h and 28 d randomization
  • All cause and cause specific mortality at 28d
  • Rebleed (24h, 5d, 28d)
  • Blood product transfusion
  • MI/CVA
  • VTE (DVT/PE)
  • Seizures
  • Days in the ICU
  • Functional status in-hospital or at 28d
  • Other (cardiac events, sepsis, pneumonia, resp failure, renal failure, liver failure)

Results:

• Patients:
  • 12 009 patients enrolled
  • 65% male, mean age 58
  • the mean time from bleeding onset to randomization was 22 hours
    • only 16% of patients presented within 3 hours
  • 89% had upper GI bleeding
  • 45% suspected to have variceal bleeds
  • 60% showed NO signs of shock at enrolment
• NO BENEFIT
• Primary outcome:
  • Death due to bleeding at 5 days = 3.7% (n=222) of TXA group and 3.8% (n=226) of placebo group (RR 0.99, 95% CI 0.82-1.18)
• Secondary outcomes:
  • All cause mortality at 28 days (9.5% of TXA group and 9.2% of placebo group, RR 1.03, 95% CI 0.92-1.16)
  • No difference in rebleeding, surgery, endoscopy, need for transfusion, or total blood products transfused
  • RISKS – doubling of venous thromboembolic events (0.4% placebo and 0.8% TXA (RR 1.85, 95% CI 1.15-2.98)
  • NNH = 250

Limitations:

• The authors changed the primary outcome from “all-cause mortality” to “death due to bleeding at 5d”
  • Not entirely sure why, as “all cause mortality” is clinically more relevant to us, “death” vs “death from bleeding” are one and the same when talking to family members
  • However, the change of this primary outcome forced the authors to increase their sample size by 4000, and still powered the study to detect a difference in the original outcome
• The majority of patients in the trial had variceal bleeding due to liver disease and accounted for 75% of deaths, but also increased risk of VTE was more pronounced in patients with liver disease
• Only 9% of patients were on anticoagulants, so unsure of applicability of data to that group

Thoughts:

• Overall really well done study, very few protocol violations, excellent follow up
• Time to randomization was high (mean ~ 22hrs)
  • Looking back at CRASH-2 showed mortality benefit (>20 000 pts, 4.9% vs 5.7% (p = 0.0077)) when 1g TXA given <3 hrs, but not after that.
  • Perhaps we would have been benefit it TXA was administered earlier in GI bleed patients
  • However, if these patients don’t present to us within that time frame, there is no point seeing if TXA works for GI bleeds if given in <3hrs
• Since CRASH2 there has been lots of interest in TXA, but multiple studies since this positive publication have failed to show similar benefit from TXA
  • TICH -2 – no difference in mortality or neurological outcomes with TXA and ICH (Sprigg 2018)
  • WOMAN no difference in mortality or hysterectomy in PPH (WOMAN 2017) (*revised primary outcome, “death due to PPH” showed small benefit, NNT 267, but fragility index 0…, so overall not a resounding “positive study” in my opinion)
  • CRASH3 – no difference in mortality or neurologic outcomes in TBI (CRASH 2019)
  • Important to have these negative studies published given the well known publication bias of only positive studies
• Not entirely clear why they had the inclusion criteria of “clinician substantially uncertain whether to use TXA”
  • How do we know if our gestalt is correct if we eliminated those patients from the trial?
  • Perhaps this is mostly to include obvious criteria such as allergy, or for ethical reasons for clinicians to be able to treat patients to the best of their ability, but it is hard to know how many patients were not included and if that would have had any effect on the outcome

Take Home Thoughts:

• Despite some of the study weaknesses, this is a very well-designed, large trial, and a “negative” trial of this magnitude is clinically important.
• Based on the current evidence, I will not be using TXA in the management of the GI bleed patients
• What are you thoughts? Does this change your clinical practice?

References:

1. Crash-2 Collaborators. (2011). The importance of early treatment with tranexamic acid in bleeding trauma patients: an exploratory analysis of the CRASH-2 randomised controlled trial. The Lancet, 377(9771), 1096-1101.
2. Dewan, Y., Komolafe, E. O., Mejía-Mantilla, J. H., Perel, P., Roberts, I., & Shakur, H. (2012). CRASH-3-tranexamic acid for the treatment of significant traumatic brain injury: study protocol for an international randomized, double-blind, placebo-controlled trial. Trials, 13(1), 1-14.
3. Gluud, L. L., Klingenberg, S. L., & Langholz, E. (2012). Tranexamic acid for upper gastrointestinal bleeding. Cochrane Database of Systematic Reviews, (1).
4. 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 Lancet. 2020; 395(10241):1927-1936. 
5. Justin Morgenstern, “TXA for GI bleeds”, First10EM blog, March 9, 2020. Available at: https://first10em.com/txa-for-gi-bleeds/.
6. Roberts, I., Shakur-Still, H., Afolabi, A., Akere, A., Arribas, M., Brenner, A., … & Jairath, V. (2020). 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 Lancet, 395(10241), 1927-1936.
7. Salim Rezaie, “REBEL Cast Ep85: The HALT-IT Trial – TXA in Acute GI Bleeds”, REBEL EM blog, June 27, 2020. Available at: https://rebelem.com/rebel-cast-ep85-the-halt-it-trial-txa-in-acute-gi-bleeds/.
8. Shakur, H., Elbourne, D., Gülmezoglu, M., Alfirevic, Z., Ronsmans, C., Allen, E., & Roberts, I. (2010). The WOMAN Trial (World Maternal Antifibrinolytic Trial): tranexamic acid for the treatment of postpartum haemorrhage: an international randomised, double blind placebo controlled trial. Trials, 11(1), 40.
9. Sprigg, N., Flaherty, K., Appleton, J. P., Salman, R. A. S., Bereczki, D., Beridze, M., … & Dineen, R. A. (2018). Tranexamic acid for hyperacute primary IntraCerebral Haemorrhage (TICH-2): an international randomised, placebo-controlled, phase 3 superiority trial. The Lancet, 391(10135), 2107-2115.