The New England Journal of Medicine February 22, 2018

Thrombectomy for Stroke at 6 to 16 Hours with Selection by Perfusion Imaging (DEFUSE 3)

Gregory W Albers, Michael P Marks, Stephanie Kemp, et al. (DEFUSE 3 Investigators)

Bottom Line

In patients with acute ischemic stroke due to large vessel occlusion and salvageable tissue identified on perfusion imaging, endovascular thrombectomy performed 6 to 16 hours after symptom onset significantly improved functional outcomes compared to standard medical therapy alone.

Key Findings

1. Endovascular therapy significantly improved the primary outcome of the overall ordinal distribution of 90-day modified Rankin scale (mRS) scores, with a median score of 3 versus 4 (adjusted OR 3.36; 95% CI, 1.96 to 5.77; P<0.001).

2. Functional independence (mRS score of 0 to 2) at 90 days was achieved by 45% of patients in the endovascular-therapy group compared with 17% in the medical-therapy group (risk ratio 2.67; 95% CI, 1.60 to 4.48; P<0.001).

3. Mortality at 90 days was 14% in the endovascular group and 26% in the medical-therapy group (P=0.05).

4. Rates of symptomatic intracranial hemorrhage were not significantly different between the endovascular and medical therapy groups (7% vs. 4%, P=0.75).

5. The trial was terminated early for efficacy by the data and safety monitoring board after 182 patients were randomized, following the publication of the parallel DAWN trial.

Study Design

Design

RCT

Open-Label

Sample

182

Patients

Duration

90 days

Median

Setting

38 US centers

Population Patients aged 18 to 90 years with acute ischemic stroke presenting 6 to 16 hours after last known well, with a proximal middle-cerebral-artery or internal-carotid-artery occlusion, and a favorable mismatch profile on perfusion imaging (initial infarct volume <70 ml, ischemic tissue to infarct volume ratio ≥1.8, and absolute penumbra volume ≥15 ml).

Intervention Endovascular thrombectomy plus standard medical therapy

Comparator Standard medical therapy alone

Outcome Ordinal score on the modified Rankin scale (mRS, range 0 to 6) assessed at 90 days.

Study Limitations

• The early termination of the trial restricted the final sample size to 182 patients, which limited the statistical power for secondary and subgroup analyses.

• Early stoppage in trials can sometimes lead to an overestimation of the actual magnitude of the treatment effect.

• The stringent inclusion criteria, requiring advanced automated perfusion imaging (CT or MR perfusion) to quantify precise core-penumbra mismatches, limits immediate generalizability to centers lacking this software infrastructure.

• The findings strictly apply to patients with proximal anterior circulation occlusions (internal carotid or M1 segment) and do not address more distal occlusions or posterior circulation strokes in the extended window.

• The trial used an open-label design, which introduces potential bias in medical management, although this was mitigated by blinded endpoint adjudication.

Clinical Significance

DEFUSE 3 provided definitive evidence that the therapeutic window for mechanical thrombectomy could safely and effectively be extended from 6 hours up to 16 hours in patients with a favorable core-to-penumbra mismatch. This trial, alongside the DAWN trial, fundamentally shifted the paradigm of acute stroke management from strict time-based triage to 'tissue-based' triage, leading the 2018 AHA/ASA guidelines to officially recommend extended-window thrombectomy for eligible patients.

Historical Context

Prior to 2018, acute ischemic stroke treatments were heavily bound by strict time windows: intravenous thrombolytics within 4.5 hours and mechanical thrombectomy within 6 hours (established by a series of landmark trials in 2015). This left a major therapeutic gap for patients presenting late, transferred from outside facilities, or suffering 'wake-up' strokes. The parallel DAWN (up to 24 hours using clinical-core mismatch) and DEFUSE 3 (6-16 hours using perfusion mismatch) trials tested the hypothesis that advanced neuroimaging could identify patients who maintain salvageable penumbra beyond standard windows. Their overwhelming success revolutionized neurointerventional care.

Guided Discussion

High-yield insights from every perspective

Med Student

Medical Student

How does the concept of the ischemic penumbra versus the infarct core explain why some patients benefit from thrombectomy up to 16 hours after symptom onset?

Key Response

This tests the understanding of basic stroke pathophysiology. The infarct core represents irreversibly damaged tissue, while the penumbra is hypoperfused but salvageable tissue. DEFUSE 3 demonstrated that by using perfusion imaging to identify patients who still had a significant penumbra compared to the core, late intervention could restore blood flow and rescue brain tissue.

Resident

A patient presents 10 hours after last known well with an MCA occlusion. What specific imaging modalities and parameters are required based on the DEFUSE 3 criteria to determine if they are a candidate for mechanical thrombectomy?

Key Response

Focuses on clinical management. Residents must know that CT perfusion or MR perfusion is needed, specifically looking for an infarct core < 70 mL, a mismatch ratio of > 1.8, and a mismatch volume of > 15 mL to appropriately select candidates for late-window thrombectomy.

Fellow

DEFUSE 3 and DAWN both extended the window for thrombectomy, but they utilized different patient selection criteria. How do the imaging and clinical selection paradigms differ between these two landmark trials, and how might this affect patient eligibility in a comprehensive stroke center?

Key Response

Fellows need to synthesize advanced literature. DAWN used clinical-core mismatch (NIHSS vs. infarct volume) up to 24 hours, while DEFUSE 3 used perfusion-core mismatch (hypoperfused volume vs. infarct volume) up to 16 hours. Understanding these nuances allows fellows to capture patients who might qualify under one trial's criteria but fail the other.

Attending

Given the robust outcomes in DEFUSE 3, how should institutional workflows be adapted for late-presenting or wake-up stroke patients, and what are the resource implications for community versus comprehensive stroke centers?

Key Response

Focuses on systems of care and practice change. Attendings must operationalize evidence, realizing that implementing DEFUSE 3 requires 24/7 access to advanced perfusion imaging and interventional teams, often necessitating robust hub-and-spoke transfer protocols.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD

The DEFUSE 3 trial was halted early for efficacy after an interim analysis. What are the statistical risks of stopping a trial early for benefit, and how might this impact the effect size estimates reported in the final publication?

Key Response

Focuses on biostatistics. Stopping early for efficacy can overestimate the true treatment effect because random highs in data might cross stopping boundaries. Researchers must critically evaluate if the NNT reported is artificially inflated compared to what a fully completed trial would show.

Journal Editor

The trial heavily relied on automated RAPID software to define the perfusion mismatch. As an editor, how does this proprietary software dependency affect the external validity and reproducibility of the study, and what supplementary data would you demand to ensure the results are robust?

Key Response

Evaluates generalizability and methodology. Reviewers would scrutinize whether the results hold up using other perfusion assessment methods or if the success of the trial is inextricably linked to one specific algorithm, limiting global applicability in centers lacking access to this proprietary tool.

Guideline Committee

Based on the findings from DEFUSE 3 and DAWN, how should the AHA/ASA guidelines grade the recommendation for mechanical thrombectomy in the 6-16 hour window, and should the guidelines mandate advanced perfusion imaging for all late-presenting strokes?

Key Response

Directly addresses guideline updates. Following these trials, the 2018 AHA/ASA guidelines updated the recommendation to Class I, Level of Evidence A for EVT in the 6-16 hour window for patients meeting DEFUSE 3 or DAWN criteria, making CTP or DW-MRI a standard of care requirement for selecting late-window patients.

Clinical Landscape

Noteworthy Related Trials

2015

MR CLEAN Trial

n = 500 · NEJM

Tested

Endovascular thrombectomy plus usual care

Population

Acute ischemic stroke patients with proximal anterior circulation occlusion within 6 hours of onset

Comparator

Usual care alone

Endpoint

Modified Rankin scale score at 90 days

Key result: Intraarterial treatment resulted in a significant shift toward better functional outcomes at 90 days.

2015

EXTEND-IA Trial

n = 70 · NEJM

Tested

Endovascular thrombectomy after IV alteplase

Population

Acute ischemic stroke patients with anterior circulation occlusion and salvageable tissue on perfusion imaging within 4.5 hours

Comparator

IV alteplase alone

Endpoint

Reperfusion at 24 hours and early neurologic improvement

Key result: Thrombectomy significantly increased reperfusion rates and improved 90-day functional outcomes in patients selected by perfusion imaging.

2018

DAWN Trial

n = 206 · NEJM

Tested

Endovascular thrombectomy plus standard medical care

Population

Acute ischemic stroke patients with anterior circulation LVO presenting 6 to 24 hours from last known well

Comparator

Standard medical care alone

Endpoint

Utility-weighted modified Rankin scale score at 90 days

Key result: Thrombectomy significantly improved functional outcomes at 90 days compared to medical therapy alone in patients with a clinical-core mismatch.

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