SWIFT
(2012)Objective
Compare the efficacy and safety of the Solitaire Flow Restoration stent retriever versus the Merci Retriever for mechanical thrombectomy in acute ischaemic stroke.
Study Summary
• Good 90-day neurological outcome: 58% Solitaire vs 33% Merci (OR 2.78, 95% CI 1.25–6.22; p=0.02).
• 90-day mortality was lower with Solitaire: 17% vs 38% (OR 0.34, 95% CI 0.14–0.81; p=0.02); symptomatic ICH 2% vs 11%.
Intervention
Mechanical thrombectomy with the Solitaire Flow Restoration stent retriever (Covidien/ev3) vs the Merci Retrieval System (Stryker) within 8 h of acute ischaemic stroke onset.
Inclusion Criteria
Age 22–85 with acute ischaemic stroke, NIHSS 8–30, angiographically confirmed proximal cerebral artery occlusion, treatable by thrombectomy within 8 h of onset, and ineligible for or failed IV rt-PA.
Study Design
Arms: Solitaire Flow Restoration device vs Merci Retriever (1:1 randomisation)
Patients per Arm: 58 Solitaire vs 55 Merci (113 randomised; 31 additional Solitaire roll-in patients analysed separately)
Outcome
• Good 90-day neurological outcome (mRS ≤2 or equivalent): 58% vs 33% (OR 2.78, 95% CI 1.25–6.22, p=0.02).
• 90-day mortality: 17% vs 38% (OR 0.34, 95% CI 0.14–0.81, p=0.02); symptomatic ICH 2% vs 11% (p=0.057).
Clinical Question
In patients with acute ischaemic stroke due to large vessel occlusion, is mechanical thrombectomy with the Solitaire stent retriever non-inferior (and potentially superior) to the Merci Retriever for successful recanalisation without symptomatic intracranial haemorrhage?
Bottom Line
Solitaire stent-retriever thrombectomy achieved substantially higher rates of successful recanalisation without symptomatic ICH, better 90-day neurological outcomes, and lower mortality than the Merci Retriever, prompting early trial termination.
Major Points
- First head-to-head randomised comparison of two neurothrombectomy devices in acute ischaemic stroke, stopped early for efficacy after prespecified interim analysis.
- Primary composite endpoint of TIMI 2/3 recanalisation in all treatable vessels without symptomatic ICH was met in 61% of Solitaire vs 24% of Merci patients (OR 4.87, 95% CI 2.14–11.10; p-superiority=0.0001).
- 90-day good neurological outcome occurred in 58% Solitaire vs 33% Merci (OR 2.78, p=0.02), and 90-day mortality was halved with Solitaire (17% vs 38%, OR 0.34, p=0.02).
- Symptomatic ICH was numerically lower with Solitaire (2% vs 11%, p=0.057); all radiological ICH 17% vs 36% (p=0.02).
- Established stent-retriever technology as the preferred approach for endovascular stroke treatment and set the stage for later positive stroke thrombectomy trials (MR CLEAN, ESCAPE, SWIFT PRIME, EXTEND-IA, REVASCAT).
Study Design
- Study Type
- Randomized Controlled Trial (non-inferiority with nested superiority)
- Randomization
- Yes
- Blinding
- Open-label with blinded core-lab angiographic assessment and blinded clinical events adjudication
- Sample Size
- 113
- Follow-up
- 90 days
- Centers
- 18
- Countries
- USA, France
Primary Outcome
Definition: Successful recanalisation (TIMI 2 or 3 flow in all treatable vessels) with the assigned study device without rescue treatment and without symptomatic intracranial haemorrhage
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| 24% (13/54) | 61% (34/56) | 4.87 (2.14–11.10) | <0.0001 non-inferiority; 0.0001 superiority |
Limitations & Criticisms
- Small sample size (n=113) due to prespecified early stopping for efficacy, limiting the precision of treatment-effect estimates.
- Open-label design (unavoidable given device-specific procedures) with imbalance in a few baseline covariates (notably atrial fibrillation more common in Merci group), although effects persisted after multivariable adjustment.
- Comparator was the Merci Retriever, a first-generation device that was later superseded; no comparison to aspiration thrombectomy (Penumbra) or other stent retrievers, and no comparison to IV thrombolysis or best medical therapy alone.
- Enrolled patients were largely ineligible for or failed IV rt-PA, and 74% of Solitaire roll-in patients were rt-PA failures, limiting generalisability to bridging-therapy populations.
- 17 of 18 sites had no prior experience with Solitaire (brief roll-in phase only), which the authors argue conservatively biased against Solitaire but could also reflect a learning effect.
- Predominantly US enrolment (17 of 18 sites) may limit generalisability to non-US practice.
Citation
Lancet 2012;380(9849):1241-1249. DOI: 10.1016/S0140-6736(12)61384-1