PDF: ISPASM
(2021)Objective
To assess the safety of a 7-day intravenous infusion of tirofiban compared with placebo in patients with aneurysmal subarachnoid hemorrhage treated with ventriculostomy and coiling
Study Summary
• No difference in intracranial hemorrhage, death, or serious adverse events between groups
• Tirofiban reduced clinical vasospasm (6% vs 42%, P=0.01) and radiographic vasospasm (6% vs 42%, P=0.01)
Intervention
7-day continuous intravenous tirofiban infusion (0.10 µg/kg/min) started within 24 hours from angiogram and at least 12 hours after EVD placement, compared with placebo (0.9% normal saline)
Inclusion Criteria
Age 18-85 years, aSAH documented by angiogram, EVD placed in operative room, Hunt-Hess Grade ≤4 after correction, modified Fisher grade ≤4, aneurysm secured with coiling <48 hours from rupture and <24 hours from admission
Study Design
Arms: Tirofiban (n=18): IV tirofiban 0.10 µg/kg/min × 7 days; Placebo (n=12): IV normal saline × 7 days
Patients per Arm: Tirofiban: 18, Placebo: 12
Outcome
• DCI significantly reduced: 6% vs 33% (P=0.04), NNT=3.7
• Clinical vasospasm reduced: 6% vs 42% (P=0.01)
• Radiographic vasospasm reduced: 6% vs 42% (P=0.01)
• No difference in death (6% vs 0%, P=0.40), VPS requirement (33% vs 25%, P=0.62), or thrombocytopenia (6% vs 0%, P=0.40)
Bottom Line
Seven-day continuous intravenous tirofiban infusion did not increase the risk of intracranial hemorrhage in aSAH patients treated with coiling and EVD placed in the operative room. Tirofiban significantly reduced the incidence of delayed cerebral ischemia by 27% (absolute risk reduction) with a number needed to treat of 3.7, and also reduced clinical and radiographic vasospasm. The results support proceeding with further multicenter testing of tirofiban's safety and efficacy in a larger, more pragmatic population.
Major Points
- Phase 1/2a single-center double-blind placebo-controlled randomized trial (2:1 randomization)
- 30 patients enrolled from February 2019 to July 2020 (59 screened, 29 excluded)
- 18 patients received tirofiban, 12 received placebo
- Primary endpoint: any intracranial hemorrhage during hospital stay - no significant difference (11% vs 8%, P=0.80)
- Tirofiban significantly reduced DCI: 6% (1/18) vs 33% (4/12), P=0.04
- Absolute risk reduction for DCI: 27% with NNT of 3.7
- Clinical vasospasm significantly reduced: 6% (1/18) vs 42% (5/12), P=0.01
- Radiographic vasospasm significantly reduced on DSA or CTA: 6% vs 42%, P=0.01
- CTP changes suggesting vasospasm significantly reduced: 6% vs 33%, P=0.01
- No difference in deaths: 6% (1/18) vs 0% (0/12), P=0.40
- No difference in all hemorrhages: 11% vs 8%, P=0.80
- No difference in EVD/VPS-related hemorrhage: 6% vs 0%, P=0.40
- No difference in serious adverse events: 58.82% (10/18) vs 58.33% (7/12), P=0.88
- No difference in thrombocytopenia: 6% vs 0%, P=0.40
- No difference in VPS requirement: 33% vs 25%, P=0.62
- Trend toward decreased ICU length of stay: 11±3.9 vs 13±5.26 days, P=0.17
- Three patients unblinded: 1 large EVD hematoma (placebo), 1 asymptomatic thrombocytopenia (tirofiban), 1 femoral artery dissection (placebo)
- Independent DSMB met 3 times (after each 10 patients enrolled) to review safety
- All patients completed 6-week follow-up
- More males in tirofiban arm (33% vs 0%, P=0.02)
- Tirofiban infusion: 0.10 µg/kg/min for 7 days (reduced to 0.05 µg/kg/min if CrCl <30 mL/min)
- EVD placed in operative room using Ghajar guide for single-pass placement only
- Patients with craniotomy, craniectomy, stent deployment, or on antiplatelet therapy excluded
- Study drug started within 24 hours from angiogram but at least 12 hours after EVD placement
Study Design
- Study Type
- Phase 1/2a, double-blind, placebo-controlled, randomized controlled trial
- Randomization
- Yes
- Blinding
- Double-blind. Research pharmacy prepared drugs in black bags. Patients, nurses, physicians, treating neurosurgeons, and data analysts blinded. Only research pharmacy unblinded. Three patients unblinded per safety protocol for emergent issues
- Sample Size
- 30
- Follow-up
- 6 weeks
- Centers
- 1
- Countries
- United States
Primary Outcome
Definition: Any intracranial hemorrhage during hospital stay (expected average 3 weeks), including EVD/VPS-related hemorrhage, new hemorrhage, or change in existing hemorrhage. Assessed by CT scans obtained for EVD weaning, after EVD removal, after VPS placement, when clinically indicated, and by comparing baseline MRI with discharge MRI
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| 8% (1/12), 95% CI 0.01%-37% | 11% (2/18), 95% CI 1.86%-34% | - (Tirofiban: 1.86%-34%; Placebo: 0.01%-37%) | 0.80 |
Limitations & Criticisms
- Single-center study limits generalizability
- Small sample size (30 patients) - designed as Phase 1/2a safety study, not powered for efficacy
- Imbalanced sex distribution between arms (more males in tirofiban: 33% vs 0%, P=0.02)
- EVD placement restricted to operative room with Ghajar guide and single-pass only - not pragmatic, limits real-world applicability
- Excluded patients with craniotomy/craniectomy - cannot generalize to surgical clipping patients
- Excluded patients with stent-assisted coiling or on antiplatelet therapy
- Excluded poor-grade patients (Hunt-Hess 5 persisting after EVD)
- Only included patients who could undergo MRI - excludes patients with pacemakers, etc
- Three patients unblinded during study (10% of cohort) which could introduce bias
- No adjustment for multiple comparisons in secondary outcomes
- Wide confidence intervals on primary outcome due to small sample size
- Single outcome assessor bias possible despite blinding of treating physicians
- Did not measure platelet function, D-dimers, or fibrinogen-degradation products to confirm mechanism
- Follow-up limited to 6 weeks - longer-term outcomes unknown
- Higher baseline abnormal CTP in placebo group (33% vs 5.6%, P=0.05)
- Study stopped after enrolling 30 patients (target achieved), but relatively low screening-to-enrollment ratio (30/59, 50.8%) suggests selection bias
- DSMB review after every 10 patients may have influenced enrollment decisions
- No standardized treatment protocol for DCI rescue therapy detailed
- Cannot determine optimal duration of tirofiban therapy (7 days chosen based on retrospective data)
- Dose selection (0.10 µg/kg/min) based on prior protocols but not formally dose-optimized
- Timeline constraints (EVD within hours, coiling <48 hours, drug start <24 hours from angiogram) may limit enrollment in multicenter settings
Citation
Stroke. 2021;52:3750-3758