New England Journal of Medicine June 11, 2015

Stent-Retriever Thrombectomy after Intravenous t-PA vs. t-PA Alone in Stroke (SWIFT PRIME)

Jeffrey L. Saver, Mayank Goyal, Alain Bonafe, Hans-Christoph Diener, Elad I. Levy, Vitor M. Pereira, et al.

Bottom Line

In patients with acute ischemic stroke caused by proximal anterior circulation occlusions, early mechanical thrombectomy using a stent retriever in addition to intravenous t-PA significantly improved functional outcomes at 90 days compared to t-PA alone.

Key Findings

1. Thrombectomy with the stent retriever plus intravenous t-PA significantly reduced global disability across the entire range of modified Rankin Scale (mRS) scores at 90 days compared to t-PA alone (P<0.001).
2. The rate of functional independence (defined as an mRS score of 0 to 2) was significantly higher in the endovascular intervention group than in the control group (60% vs. 35%, P<0.001).
3. Substantial reperfusion was achieved in 88% of patients in the intervention arm at the end of the procedure, with a median time of 57 minutes from qualifying imaging to groin puncture.
4. There were no significant differences between the intervention and control groups regarding 90-day mortality (9% vs. 12%, P=0.50) or the rate of symptomatic intracranial hemorrhage (0% vs. 3%, P=0.12).

Study Design

Design
Randomized Controlled Trial
Open-Label
Sample
196
Patients
Duration
90 days
Median
Setting
Multicenter, US and Europe
Population Adult patients receiving intravenous t-PA for acute ischemic stroke due to a proximal anterior intracranial circulation occlusion (ICA or M1) within 6 hours of symptom onset, with moderate-to-severe neurologic deficits and a small ischemic core on advanced imaging.
Intervention Mechanical thrombectomy using a stent retriever (Solitaire FR) in addition to standard intravenous t-PA therapy.
Comparator Standard medical therapy with intravenous t-PA alone.
Outcome Severity of global disability at 90 days, assessed by an ordinal shift analysis of the modified Rankin Scale (mRS) scores.

Study Limitations

The trial was terminated early for efficacy following the positive results of the MR CLEAN trial, which may lead to an overestimation of the treatment effect size.
Stringent neuroimaging inclusion criteria, requiring a small ischemic core and confirmed salvageable penumbra, limits the immediate generalizability to patients with larger baseline infarcts.
The open-label design regarding treatment assignment is an inherent limitation, though the primary endpoint was assessed by blinded adjudicators (PROBE design).

Clinical Significance

SWIFT PRIME provided incontrovertible evidence that mechanical thrombectomy using modern stent retrievers, when added to standard intravenous t-PA, yields a profound 25% absolute increase in the rate of functional independence for patients with anterior circulation large vessel occlusions. This mandated a systemic reorganization of acute stroke care to emphasize rapid imaging and expedited transfer to comprehensive stroke centers capable of performing endovascular procedures.

Historical Context

Prior to 2015, endovascular stroke therapy struggled to prove superiority over standard IV t-PA, as demonstrated by the neutral results of the IMS III, MR RESCUE, and SYNTHESIS Expansion trials in 2013. Those trials suffered from older-generation devices, delayed reperfusion times, and a lack of advanced imaging selection. In a historic turning point for neurology, SWIFT PRIME was one of five landmark trials published in 2015 (alongside MR CLEAN, ESCAPE, EXTEND-IA, and REVASCAT) that utilized second-generation stent retrievers (Solitaire), optimized workflow metrics, and strict perfusion imaging criteria. Together, these trials definitively established mechanical thrombectomy as the standard of care for large vessel occlusion strokes.

Guided Discussion

High-yield insights from every perspective

Med Student
Medical Student

SWIFT PRIME targeted patients with an acute proximal anterior circulation occlusion (like the MCA M1 segment). Physiologically, why is rapid mechanical reperfusion so critical in these specific patients compared to those with small vessel lacunar strokes?

Key Response

Large vessel occlusions (LVOs) block a massive amount of blood flow, creating a rapidly expanding central core of infarction and a large surrounding 'penumbra' (electrically silent but salvageable tissue). IV t-PA alone has notoriously low recanalization rates for LVOs due to the massive clot burden. Rapid mechanical removal saves this extensive penumbra before it progresses to irreversible infarction. In contrast, small vessel lacunar strokes do not have large, targetable clots or massive penumbras amenable to macrovascular extraction.

Resident
Resident

You are evaluating a patient with an MCA M1 occlusion 3 hours from symptom onset who received IV t-PA at an outside hospital. Based on the inclusion criteria of SWIFT PRIME, what specific clinical and imaging parameters must this patient meet to be an ideal candidate for immediate endovascular thrombectomy?

Key Response

Residents must master EVT inclusion criteria. Key SWIFT PRIME criteria included: a proximal anterior circulation LVO (distal ICA or M1), moderate-to-severe neurological deficit (NIHSS 8-29), ability to initiate EVT within 6 hours of symptom onset, and critically, a favorable imaging profile proving salvageable tissue (a small core infarct, such as ASPECTS > 5 or specific CT/MR perfusion mismatch criteria). Recognizing these triage parameters is essential for acute stroke management.

Fellow
Fellow

SWIFT PRIME utilized advanced neuroimaging (CT or MRI perfusion with RAPID software) to select patients with a 'target mismatch.' How does this imaging-based selection differ from the purely non-contrast CT time-based selection used in MR CLEAN, and what are the implications for the treatment effect sizes observed across these trials?

Key Response

MR CLEAN used standard non-contrast CT (time-based selection), resulting in a broader, less homogenous cohort and an absolute risk reduction (ARR) of ~13% for functional independence. By using perfusion imaging to ensure a small core and large penumbra, SWIFT PRIME heavily enriched its treatment arm with patients highly likely to benefit from reperfusion, resulting in a much larger ARR of ~25%. Fellows should understand how trial selection criteria (perfusion vs. NCCT) directly modulate the Number Needed to Treat (NNT).

Attending
Attending

The SWIFT PRIME trial, along with the rest of the 2015 thrombectomy 'tsunami', demonstrated a Number Needed to Treat (NNT) of around 4 for functional independence. How does an effect size of this magnitude fundamentally alter systems of care, particularly regarding EMS routing and hub-and-spoke stroke networks?

Key Response

Attendings must understand how major evidence shapes systems of care. The overwhelming efficacy of EVT shifted pre-hospital protocols toward the 'mothership' or optimized 'drip-and-ship' models. EMS protocols were altered in many regions to bypass closer primary stroke centers (which can only administer IV t-PA) in favor of transporting patients with suspected LVOs directly to Comprehensive Stroke Centers capable of performing thrombectomy, as the delay in EVT significantly diminishes the profound benefits seen in these trials.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD
PhD

SWIFT PRIME was halted early for efficacy after an interim analysis of 196 patients, partly influenced by the publication of MR CLEAN. What are the statistical and methodological risks of early termination for benefit in clinical trials, and how might this have impacted the reported effect size?

Key Response

Truncated trials run a high risk of exaggerating treatment effects because they often stop at a 'random high' in the data trajectory. While stopping early was ethically sound once EVT efficacy became undeniable across concurrent trials, PhDs and trialists must recognize that the reported 25% absolute difference in mRS 0-2 might be an overestimation of the true long-term population effect. This 'winner's curse' must be accounted for when powering future non-inferiority trials in neuro-interventions.

Journal Editor
Journal Editor

Given that endovascular thrombectomy is a procedural intervention that cannot be blinded to treating physicians or patients, what specific methodological safeguards must the authors implement to ensure the primary endpoint (mRS at 90 days) is not affected by ascertainment bias, and as a reviewer, what flaws would you look for in this process?

Key Response

Editors scrutinize subjective endpoints in unblinded trials. SWIFT PRIME used a PROBE (Prospective Randomized Open, Blinded End-point) design. The mRS assessors were ostensibly blinded to treatment allocation, often via standardized video or phone interviews by independent personnel. A rigorous reviewer would question the success of this blinding—for example, whether assessors could deduce treatment arms through incidental unblinding (e.g., patient mentions of a groin puncture, ICU stay length, or visible groin scars) and whether inter-rater reliability of the mRS was formally tracked.

Guideline Committee
Guideline Committee

Following SWIFT PRIME and contemporary trials, the AHA/ASA updated their acute ischemic stroke guidelines to give mechanical thrombectomy a Class I recommendation. Specifically regarding the 'bridging' approach, how do the current guidelines position the administration of IV t-PA prior to EVT, and does the SWIFT PRIME design support this?

Key Response

The 2015/2018 AHA/ASA guidelines (Class I, LOE A) mandate that eligible patients receive mechanical thrombectomy with a stent retriever if they have an LVO in the anterior circulation within 6 hours. Crucially, the guidelines state that patients eligible for IV alteplase should receive it even if EVT is being considered, and EVT should not be delayed to assess the clinical response to IV t-PA. SWIFT PRIME explicitly studied this bridging approach (t-PA + EVT vs. t-PA alone), perfectly validating the guideline's stance that IV thrombolytics should not be withheld in EVT-eligible patients.

Clinical Landscape

Noteworthy Related Trials

2015

MR CLEAN

n = 500 · NEJM

Tested

Intraarterial treatment (mostly stent retrievers) plus usual care

Population

Patients with acute ischemic stroke and proximal anterior circulation occlusion within 6 hours

Comparator

Usual care alone (including IV t-PA)

Endpoint

Modified Rankin Scale (mRS) score at 90 days

Key result: Intraarterial treatment significantly shifted the distribution of 90-day mRS scores in favor of better functional outcomes compared to usual care alone.
2015

EXTEND-IA

n = 70 · NEJM

Tested

Endovascular therapy (Solitaire stent retriever) after IV t-PA

Population

Patients with acute ischemic stroke and large vessel occlusion selected by CT perfusion imaging

Comparator

IV t-PA alone

Endpoint

Reperfusion at 24 hours and early neurologic improvement

Key result: Endovascular therapy resulted in significantly better reperfusion, early neurologic improvement, and improved functional independence at 90 days.
2015

ESCAPE

n = 315 · NEJM

Tested

Rapid endovascular therapy (stent retrievers) up to 12 hours after symptom onset

Population

Patients with acute ischemic stroke, proximal occlusion, and small infarct core with good collaterals on CT

Comparator

Standard medical care (including IV t-PA if eligible)

Endpoint

Modified Rankin Scale (mRS) score at 90 days

Key result: Rapid endovascular treatment significantly reduced mortality and increased the rate of functional independence at 90 days compared to standard care.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis