Episode 16- “HALT” the use of TXA in GI bleeds?

Hello and welcome to Episode 16 of ER-Rx. In this episode, we discuss the long-awaited “HALT-IT,” trial, published in June 2020 in the Lancet.

Acute GI bleeding (GIB) is a common cause of death, with a 10% fatality rate for upper GIB and a 3% mortality for lower GIB. Making things worse, patients who re-bleed after initial stabilization are four times more likely to die. Patients with GIB are oftentimes hemodynamically unstable when they present to the ER, and they are usually managed with blood transfusions, medical or endoscopic therapy, or surgery. The “HALT-IT” trial looked at using a 24-hour infusion of TXA (tranexamic acid) in patients with acute GIB. This was an international, randomized, double-blinded, placebo-controlled trial in over 12,000 patients in 164 hospitals and 15 countries. This study looked at a sick population whose patients had to have hypotension, tachycardia, or signs of shock in order to be randomized. 90% of the patients had an upper GIB and the rest had lower GIB. All factors including age, sex, time from randomization to TXA, rates of variceal bleeding, and blood pressures were the same between the groups.

TXA is an antifibrinolytic agent that works by displacing plasminogen from fibrin, resulting in inhibition of fibrinolysis. It also inhibits the activation of plasmin. If you remember from Episode 12, “clot busters” such as alteplase (t-PA) break down clots by activating plasminogen to plasmin, so TXA does the exact opposite of what t-PA does. TXA has been used topically to reduce bleeding during dental procedures, it can be given nebulized for the treatment of non-massive hemoptysis, and going back to Episode 7, TXA can be used topically for nosebleeds. It has also been given IV to reverse intracranial hemorrhage associated with thrombolytic use and perioperative prevention of bleeding during major surgery. In the setting of postpartum hemorrhage and trauma-associated hemorrhage, if given within 3 hours TXA may reduce mortality rates. Therefore, there has been growing interest in using TXA in patients with GIB. In a meta-analysis, TXA for upper GIB showed a reduction in all-cause mortality (RR 0.61, 95% CI 0.42-0.89, P=0.01), but the study was unable to assess the risks of thromboembolic events. Therefore, a much larger trial was needed.

In the HALT-IT trial, patients received either a loading dose of 1 g TXA or placebo over 10 minutes, followed by a maintenance dose of 3 g of TXA or placebo for 24 hours. To put this into perspective, the dose for postpartum hemorrhage is 1 g over 10-20 minutes as studied in the “WOMAN” trial, and the most common dose we use in the ER for trauma-associated hemorrhage is a loading dose of 1 g over 10 minutes followed by 1 g over the next 8 hours as a continuous infusion as studied in the “CRASH-2” trial. The authors chose to give a higher dose over a longer duration as patients with GIB often rebleed after initial hemostasis, particularly within the first 24 hours, so they wanted to cover this high-risk period.

So we have a plausible mechanism and data showing TXA reduces mortality in other bleeding patients. What about in this trial? The primary outcome of death due to bleeding within 5 days occurred at similar rates, 3.7% in the TXA group and 3.8% in the placebo group. All other outcomes, including death due to bleeding within 24 hours (2.1 % vs 2%) and within 28 days (4.2% vs 4.4%), and death from all causes within 28 days (9.5% vs 9.2%) were the same between the groups. When the authors looked at pre-specified subgroup analyses that looked at time to treatment, location of bleed, and cause of bleed they found no differences between the groups. There was also no difference in the rates of interventions including endoscopy, surgery, transfusion, or units of blood administered. All in all, TXA did not perform better than placebo in any measure studied.

However, in terms of safety, they found that the risk of venous events (DVT or PE) was higher in the TXA group than the placebo group (at 0.8% vs 0.4%), but the rates of DVT or PE individually was not different between the groups. This left me slightly confused and I was not able to find more information in the Supplemental Materials. For perspective, the rates of DVT or PE in the WOMAN trial were around 0.3% with no difference between groups, and in the CRASH-2 trial the rates were 0.4% for DVT and 0.7% for PE which was the same in both groups. So I’m not fully convinced of this higher risk of thrombosis.

There was also a higher risk of seizures (0.6% vs 0.4%, CI 1.03-2.93) which was twice the rate compared to The WOMAN trial with its rate of 0.3% in the TXA group and 0.4% in the placebo group. Seizures and thrombotic events are known side effects of TXA found in the package insert. The authors bring up the thought that maybe the higher risk of DVTs and seizures was due to the higher doses used in this trial. Whatever the case, there was a low, but higher rate of thrombosis in the TXA group, but it wasn’t anything that really caused me alarm, and would not be something that I would consider too heavily unless my patient had known vascular disease or hypercoagulopathy.

In conclusion, in the setting of GIB, TXA does not reduce mortality rates or any other meaningful clinical outcomes, but can increase the risk of venous thromboembolism and seizures. Although not without some shortcomings, at the very least this trial can dispel the large mortality reduction that was seen in meta-analyses. At most, this trial can be used as a reason to HALT the use of TXA in these patients. This is how I read the study, and I would not recommend the use of TXA in the setting of GIB outside of a clinical trial.

As always, thank you for your time. Please check out our Instagram page “errxpodcast,” and our YouTube channel with the same name. See you next week.