Critical care Topic

Give the TXA? Part 1: major bleeding.

Six bedside calls on tranexamic acid in trauma, postpartum hemorrhage, brain bleeds and surgery. It protects clot that has already formed, so timing decides most of the benefit.

Reviewed October 2026 · verify against current guidelines

TXA I · 1 of 10

Give the TXA?

Part 1 · major bleeding. Make your call on each — give or don't — before reading on. The answers are on card 9.

Your call

Hypotensive after a crash, 90 min in
Same crash, 4 hours in
Postpartum hemorrhage, 40 min after birth
Stab wound, SBP 80, still in the field
Fall, GCS 13, subdural on CT, 2 h in
Total knee arthroplasty

Six decades of TXA

1962	Discovered
2010	CRASH-2
2017	WOMAN
2019	CRASH-3
2020	HALT-IT
2026	TRACTION

TXA I · 2 of 10

It protects the clot.

Tranexamic acid does not build clot. It slows the breakdown of clot that has already formed, so it only helps where fibrinolysis is part of the bleeding.

Where it acts

Plasminogen → (tPA, on fibrin) → plasmin → cuts fibrin clot into fragments. TXA blocks the first step: it occupies the lysine-binding sites, so plasminogen cannot dock on fibrin.

It does	It doesn't
Slow the breakdown of formed clot	Replace factors, fibrinogen or platelets
Work best when fibrinolysis is high: early after injury, birth and bypass	Reverse an anticoagulant or stop arterial bleeding on its own

Pharmacology

Half-life	About 2 hours after an IV dose.
Clearance	Kidney, mostly unchanged. It accumulates in kidney impairment.
Routes	IV, oral, topical, inhaled; IM under study. Never intrathecal.

Why the mechanism predicts the trials — fibrinolysis spikes early: tissue injury and placental separation release tPA within minutes, and that is the window TXA can act in. Later, it often shuts down: once breakdown is suppressed, blocking it further gives no bleeding benefit and may add clot risk.

Not a hemostatic agent — in a patient bleeding from a vessel, TXA buys nothing without surgery, embolization or endoscopy.

TXA I · 3 of 10

The three-hour clock.

In trauma and postpartum hemorrhage, the benefit of TXA depends on how soon it is given. After three hours, no trial shows benefit, and in CRASH-2 bleeding deaths rose.

CRASH-2 · death from bleeding, by time to dose — relative risk, 95% CI

≤ 1 hour	0.68 — favors TXA
1 to 3 hours	0.79 — favors TXA
> 3 hours	1.44 — favors placebo

10% of the survival benefit is lost for every 15 minutes of delay, in a pooled analysis of 40,138 bleeding patients from CRASH-2 and WOMAN.

What the guidelines say

European 2023	Give as soon as possible and within 3 hours of injury; consider giving en route.
NICE NG39	Give as soon as possible; do not start more than 3 hours after injury unless there is evidence of hyperfibrinolysis.

Why the cutoff is debated — it is a subgroup finding: time since injury in CRASH-2 was an estimate in hours, recorded at enrollment. It cannot easily be retested: later trials enrolled almost everyone early, so the clock is a ceiling, not a target.

Start the clock at injury or birth — not at arrival. Transport time counts.

TXA I · 4 of 10

Trauma: fewer deaths, same recovery.

Three placebo-controlled trials. CRASH-2 and PATCH found fewer early deaths; STAAMP did not reach significance. Whether more patients go home well is less clear.

Mortality, TXA against placebo

Trial	Population	Result
CRASH-2 · 2010	20,211, hospital, 28 d	Lower with TXA — significant
STAAMP · 2020	927, prehospital, 30 d	Lower with TXA — not significant
PATCH · 2023	1,310, prehospital, 28 d	Lower with TXA — significant
PATCH · 6 months	Same patients	Lower with TXA

What the guidelines say

European 2023	Recommends TXA for the bleeding patient or one at risk of major bleeding, within 3 hours.
EAST 2025	Conditionally recommends TXA, prehospital or in hospital, if at risk of hemorrhage. No recommendation on dose.

Why PATCH reframed the question — its primary outcome was function: survival with a favorable outcome at 6 months was 53.7% with TXA and 53.5% with placebo. The extra survivors were often severely disabled; fewer deaths did not translate into more patients living independently.

No clot signal in trauma — vascular occlusive events did not differ in CRASH-2 or PATCH.

TXA I · 5 of 10

Before the hospital.

Military teams adopted TXA after CRASH-2. Their experience produced a simpler dose that can be given close to the point of injury.

MATTERS 2012 · Afghanistan · mortality

	TXA	No TXA
All transfused (896 casualties)	17.4%	23.9%
Massive transfusion (10+ units of red cells)	14.4%	28.1%

What TCCC changed in 2020

Dose	A single 2 g, replacing 1 g plus a 1 g infusion. No second dose.
Route	Slow IV or IO push over 1 minute, in the same line as blood if needed.
Indication	Hemorrhagic shock or expected transfusion. Significant TBI was added.
TCCC 2026	The May 2026 edition keeps the single 2 g IV or IO push.

Why military data are not proof — it was observational: TXA was not randomized, and those who got it were more severely injured. More clots were recorded: pulmonary embolism 2.7% vs 0.3%, which the authors linked to injury severity.

TXA I · 6 of 10

Postpartum: treat it early.

For diagnosed postpartum hemorrhage, TXA within 3 hours reduces bleeding deaths. Giving it to every woman in advance is a separate, weaker case.

Treating against preventing — women, %

	TXA	Placebo
WOMAN · treat — bleeding death, ≤ 3 h	1.2	1.7
WOMAN · treat — bleeding death, all	1.5	1.9
WOMAN-2 · prevent — clinical PPH · 15,068	7.0	6.6

Pooled prevention trials · 54,404 women, 2024

Life-threatening bleeding 0.65% vs 0.85% (OR 0.77). No increase in thrombosis. The authors do not recommend TXA for every birth, but suggest considering it before diagnosis in women at high risk of death.

What the guidelines say

WHO/FIGO 2025	TXA for all diagnosed PPH within 3 hours, given without delay as part of the MOTIVE bundle.
ACOG	Consider TXA when initial medical therapy fails.

Where the two guidelines differ — WHO gives it first, alongside uterotonics and massage, because benefit falls with every minute. ACOG gives it second, after uterotonics fail; by then, part of the 3-hour window is gone.

TXA is not a uterotonic — atony still needs oxytocin, massage and a plan for surgery.

TXA I · 7 of 10

Brain bleeds: a narrow signal.

Two large trials. Neither met its primary outcome. One subgroup in traumatic brain injury carries most of the case for giving it.

CRASH-3 · 9,202 treated within 3 h · head injury death — relative risk, 95% CI

All ≤ 3 h	0.94
GCS 9 to 15	0.78
GCS 3 to 8	0.99

Traumatic · CRASH-3 2019	Benefit only in mild to moderate injury, only when given early. Disability and survivors' function did not differ.
Spontaneous · TICH-2 2018	2,325 with primary ICH. Less hematoma growth and fewer deaths by day 7. No difference in function at 90 days.

What the guidelines say

NICE NG232 2023	Consider a 2 g IV bolus within 2 hours in adults with GCS 12 or below and no active extracranial bleeding.
ESO 2025	Spontaneous ICH: benefit and harm uncertain; enroll in trials where possible.

Why the TBI subgroup is contested — it was prespecified, and the mechanism fits: smaller bleeds have more to gain from slowing expansion. It is still a subgroup of a neutral trial, and a prehospital TBI trial of 967 patients found no functional benefit.

Aneurysmal subarachnoid hemorrhage — ULTRA found no better outcome with early TXA. AHA/ASA 2023 advises against routine use.

TXA I · 8 of 10

Surgery: less blood, not fewer deaths.

Given before incision, TXA reliably cuts blood loss and transfusion. No surgical trial has shown it changes mortality.

Bleeding outcomes, TXA against placebo — patients, %

	TXA	Placebo
POISE-3 2022 — major bleeding · 9,535	9.1	11.7
TRACTION 2026 — transfusion · 8,273	7.4	9.8
ATACAS 2017 — reoperation · 4,631	1.4	2.8

Arthroplasty (hip and knee)	Many trials show less blood loss and transfusion, with no rise in thrombosis. Now routine.
Safety (same trials)	ATACAS seizures 0.7% vs 0.1%. POISE-3 did not show TXA was no worse for heart events. TRACTION venous clots 2.1% in both arms.

What the guidelines say

NICE NG24 2026	Offer TXA to adults having theater surgery with any bleeding risk that breaches skin or mucosa. No blood-loss threshold.
EACTS 2024	Recommends antifibrinolytics in cardiac surgery to reduce bleeding, transfusion and reoperation.

Why surgery differs from the ED — it is given before the injury: TXA is on board when surgery and bypass release tPA. The outcome is transfusion: avoiding blood is the goal, so a benefit without a mortality change still counts.

TXA I · 9 of 10

Deciding, and the answers.

Start with what is bleeding. Then ask how long ago it started. Other bleeds are in part 2.

Decision path — what is bleeding?

Trauma with bleeding	Under 3 hours? Yes — IV TXA now. No — don't start.
Postpartum hemorrhage	Give 1 g IV within 3 h of birth.
Surgery with bleeding risk	Give before incision.

Card 1 answers

Give	Crash, 90 min in: inside the window, mortality benefit.
Don't	Same crash, 4 h in: more bleeding deaths after 3 h in CRASH-2.
Give	Stab wound in the field: single 2 g IV or IO push.
Give	Postpartum hemorrhage, 40 min: WOMAN, first-line per WHO.
Consider	GCS 13 subdural, 2 h: the CRASH-3 subgroup that benefited.
Give	Knee arthroplasty: less transfusion, no clot signal.

TXA I · 10 of 10

IV doses by indication.

Adult doses from the trials and guidelines in this post. Local protocols vary.

Intravenous

Trauma, TBI	1 g over 10 min, then 1 g over 8 h, or a single 2 g (ATLS 11th). Within 3 h.
Military, TCCC	Single 2 g slow IV or IO push over 1 min.
Postpartum	1 g over 10 min; repeat 1 g if bleeding continues after 30 min or restarts within 24 h.
Noncardiac surgery	POISE-3: 1 g at the start and 1 g at the end.
Coronary surgery	ATACAS: 50 mg/kg over 30 min after induction.

Before any IV dose — check the route: IV only, never near spinal or epidural drugs. Check the kidneys: keep the loading dose; reduce repeat and infusion doses.

Educational reference only — not medical advice. Verify against local protocol.

Sources

Verify against current guidelines and local protocol before acting.

CRASH-2 (2010) — TXA in trauma with bleeding (n = 20,211).
WOMAN (2017) — TXA for postpartum hemorrhage.
Pooled analysis of CRASH-2 and WOMAN — 40,138 bleeding patients, effect of treatment delay.
MATTERS (2012) — TXA in military trauma, Afghanistan (n = 896).
CRASH-3 (2019) — TXA in traumatic brain injury.
TICH-2 (2018) — TXA in spontaneous intracerebral hemorrhage (n = 2,325).
ULTRA — early TXA in aneurysmal subarachnoid hemorrhage.
HALT-IT (2020).
STAAMP (2020) — prehospital TXA (n = 927).
PATCH (2023) — prehospital TXA in severe trauma (n = 1,310).
WOMAN-2 — prophylactic TXA at birth (n = 15,068).
Pooled prevention trials, 2024 — 54,404 women.
ATACAS (2017) — TXA in coronary surgery (n = 4,631).
POISE-3 (2022) — TXA in noncardiac surgery (n = 9,535).
TRACTION (2026) — TXA and transfusion (n = 8,273).
European trauma guideline 2023.
EAST 2025.
NICE NG39; NICE NG232 (2023); NICE NG24 (2026).
Tactical Combat Casualty Care (TCCC) guidelines, 2020 and May 2026 editions.
WHO/FIGO 2025 postpartum hemorrhage recommendations (MOTIVE bundle).
ACOG postpartum hemorrhage guidance.
ESO 2025 intracerebral hemorrhage guideline.
AHA/ASA 2023 aneurysmal subarachnoid hemorrhage guideline.
EACTS 2024.
ATLS, 11th edition.

Downloads

Every card behind this topic, as published — view full size or download.

Carousel10 slides