Tranexamic Acid (TXA) in Upper GI Bleeding

Mechanism of Action

Tranexamic acid is a synthetic lysine analogue that competitively inhibits plasminogen activation, thereby suppressing fibrinolysis. By blocking the binding of plasminogen to fibrin, it stabilises clots that have already formed at bleeding sites. It is an antifibrinolytic, not a procoagulant -- it does not promote new clot formation upstream of an existing bleed.

Katzung's Basic and Clinical Pharmacology, 16th Ed.: "Tranexamic acid is an analog of aminocaproic acid and has the same properties. It competitively inhibits plasminogen activation."

Rationale for Use in Upper GI Bleeding (UGIB)

The rationale rests on the observation that fibrinolysis is locally upregulated at mucosal bleeding sites (peptic ulcers, varices, Mallory-Weiss tears). Antifibrinolytics were therefore hypothesised to stabilise haemostatic plugs and reduce rebleeding before definitive endoscopic therapy. In dialysis patients specifically, tranexamic acid has shown benefit as adjunctive therapy for major UGIB by improving platelet function in the context of uremic coagulopathy.

The Pivotal Evidence: HALT-IT Trial (2020)

The HALT-IT trial is the definitive, largest RCT on this question.

Design: 12,009 adults with significant acute upper or lower GI bleeding across 164 hospitals in 15 countries; randomised to TXA (1 g IV loading over 10 min, then 3 g over 24 hrs) vs. placebo.
Primary outcome (death due to bleeding within 5 days):
- TXA group: 3.7% vs. placebo: 3.8% (RR 0.99, 95% CI 0.82-1.18) -- no significant benefit
Thromboembolic events:
- Venous thromboembolism (DVT/PE) was significantly higher with TXA: 0.8% vs. 0.4% (RR 1.85, 95% CI 1.15-2.98)
- Seizures higher with TXA: 0.6% vs. 0.4% (RR 1.73, 95% CI 1.03-2.93)
Conclusion: High-dose, extended-duration IV TXA does not reduce GI bleeding mortality and is not cost-effective.

Meta-Analyses and Systematic Reviews

The literature presents a nuanced picture that depends heavily on dose, route, and whether HALT-IT data dominate the pooled analysis.

1. Dionne et al. (Crit Care Med, 2022) - PMID 34709209

Pooled 12 RCTs and separated by regimen:

High-dose IV TXA (HALT-IT dosing): No mortality reduction (RR 0.98) or bleeding reduction (RR 0.92); significantly increased DVT (RR 2.01), PE (RR 1.78), and seizures (RR 1.73) -- high certainty evidence
Low-dose IV or enteral TXA: No mortality reduction, but reduced rebleeding (RR 0.50, 95% CI 0.33-0.75) and need for surgery (RR 0.58, 95% CI 0.38-0.88) -- moderate certainty
Key takeaway: the dose and route matter enormously.

2. Lee et al. (Am J Emerg Med, 2021) - PMID 33041136

Meta-analysis of 13 RCTs (n = 2,271, predominantly pre-HALT-IT trials):

TXA reduced continued bleeding (RR 0.60), urgent endoscopy need (RR 0.35), and mortality (RR 0.60, 95% CI 0.45-0.80)
Favoured TXA, but the smaller older trials are now outweighed by HALT-IT.

3. Calderon Martinez et al. (Indian J Gastroenterol, 2025) - PMID 40029534

Most recent and largest meta-analysis; 23 studies, >2 million participants:

TXA significantly reduced rebleeding overall (RR 0.81)
Mortality reduction in UGIB specifically: RR 0.72 (95% CI 0.59-0.87) -- upper GI bleeding is a more favourable subgroup
TXA increased mortality in lower GI bleeding (RR 1.67) -- a striking finding
No significant increase in thromboembolic events overall
Need for surgery reduced in low-bias studies (RR 0.85)

4. Djoudjou et al. (Am J Emerg Med, 2025) - PMID 40752050

7 RCTs, 13,608 patients, IV TXA only:

Rebleeding significantly reduced (OR 0.64, p = 0.01)
Failure to control bleeding reduced (OR 0.55, p = 0.03)
Mortality: no significant reduction (OR 0.77, p = 0.12)
Thromboembolic events: non-significant by random-effects, but significant by fixed-effects (OR 1.28) -- signals caution

Summary of Evidence by Outcome

Outcome	High-dose IV TXA	Low-dose / Enteral TXA	UGIB Subgroup
Mortality	No benefit (HALT-IT)	Possible reduction	Possible reduction (2025 MA)
Rebleeding	No significant benefit	Reduced	Reduced
Need for surgery	No benefit	Reduced	Reduced (low-bias studies)
DVT/PE	Increased (RR 1.85)	Uncertain	Uncertain
Seizures	Increased (RR 1.73)	Uncertain	Uncertain

Current Clinical Position

The HALT-IT regimen (high-dose, 24-hour infusion) is not recommended for routine UGIB based on high-certainty evidence of harm without benefit.

However, the story is not entirely closed:

UGIB specifically (as opposed to combined upper + lower GI bleeding) may benefit more than HALT-IT's pooled analysis suggested - subgroup analyses from the 2025 meta-analysis show a mortality benefit in UGIB (RR 0.72).
Low-dose or enteral TXA has shown benefit in older RCTs and moderate-certainty evidence supports rebleeding reduction.
Special populations (e.g., dialysis/CKD patients with uremic platelet dysfunction) may still benefit as adjunctive therapy, where improving platelet function is a separate mechanism.
The standard of care for UGIB remains: early resuscitation, PPI therapy (IV omeprazole/pantoprazole), urgent upper endoscopy within 24 hours, and vasoactive agents (terlipressin/octreotide) for suspected variceal bleeding.

Dosing Reference (where used)

Route	Dose
IV (low-dose)	0.5-1 g over 10 min, then may repeat
IV (HALT-IT, now discouraged)	1 g loading + 3 g over 24 hrs
Oral	15 mg/kg loading dose, then 30 mg/kg every 6 hours

Contraindications to TXA in GI bleeding: upper urinary tract bleeding (risk of ureteric clot obstruction), active DIC, thromboembolic disease, history of seizures (seizure risk is dose-dependent).

Bottom Line

TXA occupies a contested space in UGIB management. The HALT-IT trial effectively ended enthusiasm for high-dose extended IV TXA. However, multiple subsequent meta-analyses, including the most recent 2025 analysis, suggest that for upper GI bleeding specifically, lower-dose regimens may still reduce rebleeding and possibly mortality -- without the excess thrombotic and seizure risk seen at high doses. Most current guidelines do not recommend routine TXA for UGIB pending further evidence, but it remains a reasonable consideration in resource-limited settings or when standard therapy is unavailable, particularly in UGIB patients rather than lower GI bleeding patients.

This is a shared conversation. Sign in to Orris to start your own chat.