ATTRACTION
(2026)Objective
To assess the efficacy and safety of adjunctive tirofiban after successful endovascular reperfusion in patients with acute ischaemic stroke due to anterior-circulation large-vessel occlusion.
Study Summary
• Symptomatic intracranial haemorrhage at 48h: 12% (82/687) tirofiban vs 9% (65/691) placebo (numerically higher, not significant)
• Any intracranial haemorrhage at 48h: 34% tirofiban vs 32% placebo (not significant)
• 90-day mortality: 18% (126) tirofiban vs 19% (131) placebo (not significant)
Intervention
Tirofiban intra-arterial bolus 5 μg/kg (max 0.5 mg) followed by IV infusion 0.1 μg/kg/min for 24 h, initiated immediately after successful endovascular reperfusion.
Inclusion Criteria
Adults ≥18 years with acute ischaemic stroke due to anterior-circulation large-vessel occlusion (ICA, M1, or M2) within 24 h of last known well; NIHSS 6-30; pre-stroke mRS 0-1; ASPECTS ≥6; successful reperfusion (mTICI 2b/3) after EVT or spontaneous improvement.
Study Design
Arms: Tirofiban (n=689) vs Placebo/normal saline (n=691)
Patients per Arm: Tirofiban: 689; Placebo: 691
Outcome
• Symptomatic ICH at 48h: 12% vs 9% (numerically higher with tirofiban, not significant)
• 90-day mortality: 18% vs 19% (no significant difference)
• Exploratory subgroup analyses raised a possible safety concern in patients with cardioembolic stroke
Bottom Line
In patients with acute ischaemic stroke due to anterior-circulation large-vessel occlusion who achieved successful reperfusion (mTICI 2b/3) after endovascular thrombectomy, adjunctive tirofiban (5 μg/kg IA bolus + 0.1 μg/kg/min IV for 24h) increased 90-day functional independence (49% vs 43%, adjusted RR 1.15) without a statistically significant increase in symptomatic intracranial haemorrhage, though sICH was numerically higher and bleeding risk warrants caution—especially in cardioembolic stroke.
Major Points
- Adjunctive tirofiban significantly increased 90-day functional independence (mRS 0-2): 49% vs 43% (adjusted RR 1.15, 95% CI 1.03-1.27, p=0.0092).
- Absolute risk difference favored tirofiban by 6.1 percentage points (95% CI 0.8-11.3, p=0.023).
- Symptomatic intracranial haemorrhage at 48h was numerically higher with tirofiban (12% vs 9%) but the difference was not statistically significant.
- No significant difference in any ICH at 48h (34% vs 32%) or 90-day mortality (18% vs 19%).
- Exploratory subgroup analysis raised a possible safety concern in cardioembolic stroke patients.
- Conducted in a near-exclusively Han Chinese population (99%), limiting generalizability.
- Supports consideration of tirofiban as a post-reperfusion adjunctive strategy, with bleeding risk-benefit weighing.
Study Design
- Study Type
- Multicentre, double-blind, randomised, placebo-controlled trial
- Randomization
- Yes
- Blinding
- Double-blind (patients, treating clinicians, investigators, and outcome assessors masked; identical-appearing study kits)
- Sample Size
- 1380
- Follow-up
- 90 days
- Centers
- 82
- Countries
- China
Primary Outcome
Definition: Functional independence (modified Rankin Scale score 0-2)
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| 299/691 (43%) | 340/689 (49%) | - (1.03-1.27 (adjusted RR); 0.8-11.3 (absolute risk difference)) | 0.023 (unadjusted); 0.0092 (adjusted) |
Limitations & Criticisms
- Conducted exclusively in China with 99% Han Chinese population, limiting generalizability to other ethnic groups and healthcare systems.
- Symptomatic ICH was numerically higher in tirofiban group (12% vs 9%); although not statistically significant, the clinical relevance of this trend warrants caution.
- Possible safety concern in cardioembolic stroke subgroup may limit applicability in this important subpopulation.
- Tirofiban dosing regimen (IA bolus + 24h IV infusion) is more complex than simple oral antiplatelet strategies and may be harder to implement in some settings.
- Trial used a combined IA + IV route; results may not generalize to IV-only or other dosing regimens.
Citation
Huang H, et al. Lancet. Published online June 24, 2026. https://doi.org/10.1016/S0140-6736(26)00983-9