← Back

STOP-MSU

Stopping Intracerebral Haemorrhage with Tranexamic Acid for Hyperacute onset Presentation including Mobile Stroke Units

Share Share Copied! Compare

Year of Publication: 2024

Authors: Nawaf Yassi, Henry Zhao, Leonid Churilov, ..., Stephen M Davis

Journal: The Lancet Neurology

Citation: Lancet Neurol 2024;23:577-587

Link: https://doi.org/10.1016/S1474-4422(24)00128-5

Clinical Question

Does intravenous tranexamic acid reduce haematoma growth when administered within 2 hours of spontaneous intracerebral haemorrhage symptom onset compared with placebo?

Bottom Line

Intravenous tranexamic acid did not reduce haematoma growth when administered within 2 hours of ICH symptom onset (43% vs 38%, aOR 1.31, p=0.37). No differences were observed in functional outcomes or safety endpoints. Based on these results, tranexamic acid should not be used routinely in primary ICH, though ongoing phase 3 trials will provide further context.

        Major Points
        Phase 2, double-blind, placebo-controlled RCT testing ultra-early (<2 hour) tranexamic acid in acute ICH
Primary endpoint (haematoma growth >33% or >6 mL at 24h) occurred in 43% of TXA group vs 38% of placebo (aOR 1.31; 95% CI 0.72-2.40; p=0.37)
Achieved fastest reported onset-to-treatment times: median 105 min from onset, 26 min from baseline imaging
Mobile stroke unit paradigm facilitated ultra-early recruitment (44/201 patients enrolled via MSU)
Median absolute haematoma growth was identical in both groups (1.2 mL)
No significant differences in 90-day functional outcomes (mRS <3: 30% TXA vs 32% placebo)
Mortality at 90 days was numerically higher with TXA (18% vs 15%) but not statistically significant
Thromboembolic events were low (3% TXA vs 1% placebo)
Lower than expected haematoma growth in placebo group (38%) compared to STOP-AUST (52%)
Optimal patient selection for antifibrinolytic trials remains challenging

      

Design

Study Type: Phase 2, investigator-led, double-blind, randomized, placebo-controlled, parallel-group trial

Randomization: 1

Blinding: Double-blind; participants, investigators, and treating teams masked; external unmasked person prepared and administered study drug; 90-day mRS assessed by masked telephone assessor

Enrollment Period: March 19, 2018 to February 27, 2023

Follow-up Duration: 90 days

Centers: 25

Countries: Australia, New Zealand, Finland, Taiwan, Viet Nam

Sample Size: 201

Analysis: Intention-to-treat; estimand framework per ICH E9(R1) guidance; binary logistic regression adjusted for baseline haematoma volume; quantile regression for absolute/relative growth; Wilcoxon-Mann-Whitney generalized OR for ordinal mRS; adaptive sample size re-estimation after 144 participants; Stata SE version 18.0 and R version 4.1.1

Arms

Field	Tranexamic acid	Control
Intervention	Intravenous tranexamic acid 1 g over 10 minutes followed by 1 g over 8 hours, commenced within 2 hours of symptom onset	Intravenous normal saline over 10 minutes followed by normal saline over 8 hours (matched dosing regimen)
Duration	Single treatment (bolus + 8-hour infusion)	Single treatment (bolus + 8-hour infusion)

Outcomes

Outcome	Type	Control	Intervention	HR / OR / RR	P-value
Haematoma growth defined as ≥33% relative growth OR ≥6 mL absolute growth on CT at 24 hours (target range 18-30h) from baseline CT, adjusted for baseline haematoma volume	Primary	37/97 (38%)	43/101 (43%)		0.37
Haematoma growth (alternative definition including any IVH increase)	Secondary	48/97 (49%)	53/101 (52%)	1.26 (0.69-2.31)
Absolute ICH growth - mL (median, IQR)	Secondary	1.2 (-0.2 to 7.6)	1.2 (0.1 to 9.3)
Adjusted difference in median absolute ICH growth	Secondary	reference	-0.1 mL (-1.5 to 1.3)
Relative ICH growth - % (median, IQR)	Secondary	14% (-4 to 50)	14% (1 to 55)
Adjusted difference in median relative ICH growth	Secondary	reference	-2% (-14 to 9)
Absolute IVH growth - mL (median, IQR)	Secondary	0 (0 to 0.7)	0 (0 to 0.5)	not estimable
Absolute ICH + IVH growth - mL (median, IQR)	Secondary	2.0 (-0.1 to 9.1)	1.4 (-0.1 to 11.6)
Adjusted difference in median ICH+IVH growth	Secondary	reference	-0.3 mL (-2.0 to 1.4)
mRS <3 or no change from baseline at 90 days	Secondary	31/98 (32%)	31/103 (30%)	0.77 (0.38-1.56)
mRS <4 or no change from baseline at 90 days	Secondary	47/98 (48%)	53/103 (51%)	1.03 (0.53-2.01)
mRS score at 90 days (median, IQR)	Secondary	4 (2 to 5)	3 (2 to 5)	0.83 (0.60-1.14) generalized OR
Utility-weighted mRS at 90 days (mean, SD)	Secondary	0.46 (0.33)	0.43 (0.36)
Mean difference in utility-weighted mRS	Secondary	reference	-0.04 (-0.12 to 0.04)
mRS 0 at 90 days	Secondary	4/98 (4%)	5/103 (5%)
mRS 1 at 90 days	Secondary	9/98 (9%)	8/103 (8%)
mRS 2 at 90 days	Secondary	16/98 (16%)	17/103 (17%)
mRS 3 at 90 days	Secondary	18/98 (18%)	23/103 (22%)
mRS 4 at 90 days	Secondary	26/98 (27%)	14/103 (14%)
mRS 5 at 90 days	Secondary	10/98 (10%)	17/103 (17%)
Death within 7 days	Adverse	8/98 (8%)	8/103 (8%)	1.08 (0.35-3.35)
Death within 90 days (mRS 6)	Adverse	15/98 (15%)	19/103 (18%)	1.61 (0.65-3.98)
Major thromboembolic events at 90 days	Adverse	1/98 (1%)	3/103 (3%)
Risk difference for thromboembolic events	Adverse	reference	0.02 (-0.02 to 0.06)
Mortality hazard ratio (post-hoc Cox regression)	Adverse	reference	HR 1.34 (0.68-2.65)		0.39

Subgroup Analysis

No significant differences observed in any prespecified subgroups or subgroup-by-treatment interactions. Subgroups analyzed: time from onset to treatment (≤1h vs >1h), baseline ICH volume (<30 mL vs ≥30 mL), anatomical location (deep vs cortical; p=0.076 for interaction), baseline IVH (absent vs present), GCS (>12 vs 8-12), age (<70 vs ≥70 years), sex (male vs female), site (MSU vs hospital), and post-hoc baseline systolic BP (≤170 vs >170 mm Hg). Insufficient data for ≤1h subgroup analysis. Trend toward interaction by location (cortical OR 0.35 [0.07-1.76] vs deep OR 1.66 [0.86-3.21], p=0.076).

Criticisms

Sample size did not allow definitive conclusions on secondary functional outcomes
Despite optimization, median 26-minute delay from baseline CT to treatment means some growth likely occurred during this period
Lower than expected haematoma growth in placebo group (38%) compared to STOP-AUST (52%) and power calculations (39%)
Small median absolute growth (1.2 mL in both groups) suggests limited room for treatment effect
High proportion (57-62%) of participants did not have haematoma growth, diluting any observable treatment effect
Participants had relatively small baseline haematomas (median 9.5-11.1 mL) compared to previous trials
Over 80% had deep haematomas limiting conclusions about cortical ICH
Blood pressure data beyond baseline not prospectively collected despite emerging importance of BP control
Resource limitations prevented larger phase 3 study
No detailed screening log available due to pragmatic trial design
29-32% of patients had haematoma regression at follow-up, complicating interpretation
Ongoing phase 3 trial (TICH-3) uses 7-day mortality as primary endpoint based on TICH-2, but STOP-MSU showed no early mortality signal

Funding

Australian Government Medical Research Future Fund (grant MRF1152282)

Based on: STOP-MSU (The Lancet Neurology, 2024)

Authors: Nawaf Yassi, Henry Zhao, Leonid Churilov, ..., Stephen M Davis

Citation: Lancet Neurol 2024;23:577-587

Content summarized and formatted by NeuroTrials.ai.

STOP-MSU

(2024)

Objective

To determine whether intravenous tranexamic acid administered within 2 hours of spontaneous intracerebral haemorrhage symptom onset reduces haematoma growth compared with placebo.

Study Summary

• Haematoma growth occurred in 38% of placebo vs 43% of TXA group (aOR 1.31, 95% CI 0.72-2.40; p=0.37) — no significant difference
• No differences in secondary imaging endpoints (absolute/relative growth, IVH growth) or functional outcomes (mRS at 90 days)
• Mortality at 90 days was numerically higher with TXA (18% vs 15%) but not statistically significant (aOR 1.61, 95% CI 0.65-3.98)
• TXA was safe with no significant increase in thromboembolic events (3% TXA vs 1% placebo); no deaths attributed to study treatment

Intervention

Intravenous tranexamic acid 1 g over 10 min followed by 1 g over 8 h, commenced within 2 hours of symptom onset, versus matched placebo (normal saline).

Inclusion Criteria

Adults >=18 years with acute spontaneous ICH confirmed on non-contrast CT, able to receive treatment within 2 h of stroke onset.

Study Design

Arms: 1:1 randomization — Tranexamic acid arm (IV TXA 1 g bolus over 10 min + 1 g infusion over 8 h) vs Placebo arm (matched IV normal saline over same schedule).

Patients per Arm: Tranexamic acid: 103; Placebo: 98

Outcome

• Primary: Haematoma growth (>=33% relative or >=6 mL absolute on 24 h CT) — 43% TXA vs 38% placebo (aOR 1.31, 95% CI 0.72-2.40; p=0.37)
• mRS <3 at 90 days: 30% TXA vs 32% placebo (aOR 0.77, 95% CI 0.38-1.56)
• mRS <4 at 90 days: 51% TXA vs 48% placebo (aOR 1.03, 95% CI 0.53-2.01)
• Death at 7 days: 8% in both groups (aOR 1.08)
• Death at 90 days: 18% TXA vs 15% placebo (aOR 1.61, 95% CI 0.65-3.98)
• Major thromboembolic events: 3% TXA vs 1% placebo (risk difference 0.02, 95% CI -0.02 to 0.06)

Bottom Line

Major Points

Phase 2, double-blind, placebo-controlled RCT testing ultra-early (<2 hour) tranexamic acid in acute ICH
Primary endpoint (haematoma growth >33% or >6 mL at 24h) occurred in 43% of TXA group vs 38% of placebo (aOR 1.31; 95% CI 0.72-2.40; p=0.37)
Achieved fastest reported onset-to-treatment times: median 105 min from onset, 26 min from baseline imaging
Mobile stroke unit paradigm facilitated ultra-early recruitment (44/201 patients enrolled via MSU)
Median absolute haematoma growth was identical in both groups (1.2 mL)
No significant differences in 90-day functional outcomes (mRS <3: 30% TXA vs 32% placebo)
Mortality at 90 days was numerically higher with TXA (18% vs 15%) but not statistically significant
Thromboembolic events were low (3% TXA vs 1% placebo)
Lower than expected haematoma growth in placebo group (38%) compared to STOP-AUST (52%)
Optimal patient selection for antifibrinolytic trials remains challenging

Study Design

Study Type: Phase 2, investigator-led, double-blind, randomized, placebo-controlled, parallel-group trial
Randomization: Yes
Blinding: Double-blind; participants, investigators, and treating teams masked; external unmasked person prepared and administered study drug; 90-day mRS assessed by masked telephone assessor
Sample Size: 201
Follow-up: 90 days
Centers: 25
Countries: Australia, New Zealand, Finland, Taiwan, Viet Nam

Primary Outcome

Definition: Haematoma growth defined as ≥33% relative growth OR ≥6 mL absolute growth on CT at 24 hours (target range 18-30h) from baseline CT, adjusted for baseline haematoma volume

Control	Intervention	HR/OR	P-value
37/97 (38%)	43/101 (43%)	- (0.72-2.40 (aOR))	0.37

Limitations & Criticisms

Sample size did not allow definitive conclusions on secondary functional outcomes
Despite optimization, median 26-minute delay from baseline CT to treatment means some growth likely occurred during this period
Lower than expected haematoma growth in placebo group (38%) compared to STOP-AUST (52%) and power calculations (39%)
Small median absolute growth (1.2 mL in both groups) suggests limited room for treatment effect
High proportion (57-62%) of participants did not have haematoma growth, diluting any observable treatment effect
Participants had relatively small baseline haematomas (median 9.5-11.1 mL) compared to previous trials
Over 80% had deep haematomas limiting conclusions about cortical ICH
Blood pressure data beyond baseline not prospectively collected despite emerging importance of BP control
Resource limitations prevented larger phase 3 study
No detailed screening log available due to pragmatic trial design
29-32% of patients had haematoma regression at follow-up, complicating interpretation
Ongoing phase 3 trial (TICH-3) uses 7-day mortality as primary endpoint based on TICH-2, but STOP-MSU showed no early mortality signal

Citation

Lancet Neurol 2024;23:577-587

View Original Publication →

STOP-MSU

Stopping Intracerebral Haemorrhage with Tranexamic Acid for Hyperacute onset Presentation including Mobile Stroke Units

Clinical Question

Bottom Line

Major Points

Design

Arms

Outcomes

Subgroup Analysis

Criticisms

Funding

Ask about STOP-MSU