← Back

DAWN

Thrombectomy 6 to 24 Hours after Stroke with a Mismatch between Deficit and Infarct

Share Share Copied! Compare

Year of Publication: 2018

Authors: Nogueira RG, Jadhav AP, Haussen DC, ..., Budzik RF

Journal: New England Journal of Medicine

Citation: N Engl J Med 2018;378:11–21

Link: https://www.nejm.org/doi/full/10.1056/NEJMoa1706442

PDF: https://www.nejm.org/doi/pdf/10.1056/NEJMoa1706442

Clinical Question

Does endovascular thrombectomy provide benefit in patients with acute ischemic stroke 6–24 hours after last known well, with a mismatch between clinical deficit and infarct volume?

Bottom Line

Endovascular thrombectomy plus standard care improved 90-day functional outcomes compared with standard care alone in patients presenting 6–24 hours after stroke onset with a clinical–core mismatch.

        Major Points
        First RCT to show benefit of thrombectomy in the 6–24 hour window using clinical-core mismatch (RAPID automated perfusion software).
Used Bayesian adaptive enrichment design with 3 mismatch groups: Group A (age ≥80, NIHSS ≥10, core <21 mL), Group B (age <80, NIHSS ≥10, core <31 mL), Group C (age <80, NIHSS ≥20, core 31–<51 mL).
Eligible vessels: intracranial ICA or proximal MCA (M1 segment) on CTA or MRA.
Trial stopped early for efficacy after 206 of up to 500 planned patients at prespecified interim analysis (significance threshold posterior probability ≥0.986; observed posterior probability of superiority >0.999).
First co-primary end point (utility-weighted mRS at 90 days): 5.5 vs 3.4 (adjusted difference 2.0 points, 95% credible interval 1.1–3.0, posterior probability >0.999).
Second co-primary end point — functional independence (mRS 0–2) at 90 days: 49% vs 13% (adjusted difference 33 percentage points, 95% credible interval 21–44, NNT ≈ 2.8). Elevated from secondary to co-primary at FDA request 30 months into the trial while still blinded.
Successful reperfusion (mTICI 2b/3): 84%. Median time from last known well to randomization: 12.2 hours (thrombectomy arm). Median onset-to-reperfusion: 13.6 hours.
No significant increase in sICH (6/107 vs 3/99, P=0.50) or 90-day all-cause mortality (20/107 vs 18/99, P=1.00). Stroke-related death 17/107 vs 18/99. Procedure-related complications 7/107.
~55% of patients presented with wake-up stroke — DAWN validated treatment of wake-up strokes using perfusion mismatch selection.
Results, together with DEFUSE 3, led to AHA/ASA 2018 guideline update extending thrombectomy window to 24 hours for selected patients.

      

Design

Study Type: Randomized controlled trial

Randomization: 1

Blinding: Blinded endpoint assessment

Enrollment Period: September 2014 – February 2017

Follow-up Duration: 90 days

Centers: 26

Countries: USA, Canada, Europe, Australia

Sample Size: 206

Analysis: Bayesian adaptive enrichment design with pre-specified interim analysis

Inclusion Criteria

Age ≥18 years
Last known well 6–24 hours before randomization
Pre-stroke mRS 0–1
Occlusion of intracranial ICA or M1 MCA on CTA/MRA
Mismatch between NIHSS and infarct volume based on RAPID imaging
Group A: Age ≥80, NIHSS ≥10, infarct <21 ml
Group B: Age <80, NIHSS ≥10, infarct <31 ml
Group C: Age <80, NIHSS ≥20, infarct 31–<51 ml

Exclusion Criteria

Evidence of intracranial hemorrhage on CT/MRI.
Large infarct core exceeding mismatch group thresholds (≥21 mL for Group A, ≥31 mL for Group B, ≥51 mL for Group C).
Pre-stroke mRS ≥2 (functional dependence prior to index event).
Known life expectancy <6 months.
Posterior circulation occlusion (vertebral, basilar, PCA).
Isolated M2 or more distal MCA occlusion.
Rapidly improving symptoms.
Known allergy to contrast dye, nickel, or titanium.
Pregnancy.
Baseline CT/MRI showing infarct involving >1/3 of the MCA territory.
Concurrent participation in another interventional trial.

Baseline Characteristics

Characteristic	Control	Active
N	99	107
Age (mean±SD)	70.7±13.2	69.4±14.1
Female (%)	48%	61%
Hypertension (%)	76%	78%
Diabetes (%)	31%	24%
Atrial Fibrillation (%)	24%	40%
Prior Stroke or TIA (%)	11%	11%
NIHSS (median)	17	17
Ischemic Core Volume (median, mL)	8.9	7.6
Occlusion Site - ICA (%)	19%	21%
Occlusion Site - M1 MCA (%)	78%	78%
Wake-Up Stroke (%)	47%	63%
IV tPA received (%)	13%	5%
Time from LKW to randomization (median, h)	13.3	12.2

Arms

Field	Thrombectomy + Standard Care	Control
Intervention	Mechanical thrombectomy with the Trevo Retriever (Stryker Neurovascular) as the primary device plus guideline-based standard medical care including IV alteplase if eligible. Thrombectomy performed by neurointerventionalists meeting site qualification criteria (≥40 mechanical thrombectomy procedures annually). Rescue reperfusion therapy with other devices or pharmacologic agents was not permitted. Concomitant cervical ICA stenting at the time of thrombectomy was not permitted; carotid angioplasty was permitted only if necessary to allow intracranial access for the retriever device.	Standard medical care provided in accordance with local guidelines, without endovascular treatment. Patients who had not received intravenous alteplase could receive antiplatelet agents, which could be started within 24 hours after randomization. Standard stroke unit or intensive care unit admission; rescue thrombectomy was not permitted.
Duration	Single procedure within 24 hours of last known well	NA

Outcomes

Outcome	Type	Control	Intervention	HR / OR / RR	P-value
First co-primary end point: Score on the utility-weighted modified Rankin Scale at 90 days (range 0 [death] to 10 [no symptoms/disability]). Second co-primary end point: Functional independence (mRS 0–2) at 90 days = 49% vs 13% (adjusted difference 33 percentage points, 95% credible interval 21–44, posterior probability >0.999) — see Secondary Outcomes for full row.	Primary	3.4	5.5		posterior probability of superiority >0.999
[Second co-primary end point per FDA request] Functional independence (mRS 0–2) at 90 days \| Adjusted difference 33 percentage points, 95% credible interval 21–44; NNT 2.8	Secondary	13% (13/99)	49% (52/107)		posterior probability of superiority >0.999
Early response (NIHSS decrease ≥10 or NIHSS 0–1 by day 5/6/7 or discharge) \| Absolute difference 29 percentage points (95% CI 16–41); Risk Ratio 3 (95% CI 2–4)	Secondary	19% (19/99)	48% (51/107)	RR 3 (2–4)	<0.001
Recanalization at 24 h \| Absolute difference 40 percentage points (95% CI 27–52); Risk Ratio 2 (95% CI 2–4)	Secondary	39% (39/99)	77% (82/107)	RR 2 (2–4)	<0.001
Change from baseline in infarct volume at 24 h (median, mL; IQR control 0–42, intervention 0–28)	Secondary	13	1		0.003 (Wilcoxon)
Infarct volume at 24 h (median, mL; IQR control 8–68, intervention 0–48)	Secondary	22	8		<0.001 (Wilcoxon)
Successful reperfusion (mTICI 2b/3)	Secondary	NA	84% (90/107)		NA
Stroke-related death at 90 days	Adverse	18/99 (18%)	17/107 (16%)	RR 1 (1–2)
Death from any cause at 90 days	Adverse	18/99 (18%)	20/107 (19%)	RR 1 (1–2)	1.00
Symptomatic intracranial hemorrhage at 24 h	Adverse	3/99 (3%)	6/107 (6%)	RR 2 (1–7)	0.50
Neurologic deterioration	Adverse	26/99 (26%)	15/107 (14%)	RR 1 (0–1)	0.04
Procedure-related complications	Adverse	NA	7/107 (7%)

Subgroup Analysis

Consistent treatment benefit across all prespecified subgroups: age (<80 vs ≥80), baseline NIHSS (10–17 vs >17), occlusion site (ICA vs M1 MCA), time from last known well (6–12h vs >12–24h), and mismatch group (A, B, C). The largest absolute adjusted difference was seen in Group C (age <80, NIHSS ≥20, core 31–<51 mL: 2.5 points) but the credible interval was wide (−0.6 to 5.5) due to small numbers. Wake-up stroke patients (~55% of cohort) benefited similarly to witnessed-onset patients. Posterior probability of benefit was >0.99 in most subgroups and ≥0.93 in all subgroups (lowest: unwitnessed stroke 0.93; Group C 0.95).

Criticisms

Imbalance in baseline atrial fibrillation (40% thrombectomy vs 24% control) — AF associated with larger infarcts and worse outcomes.
Strict exclusion of large ischemic cores (>51 mL) limits generalizability — later addressed by SELECT2, ANGEL-ASPECT, and RESCUE-Japan LIMIT.
RAPID perfusion imaging software required — not universally available, particularly in lower-resource settings.
Industry-sponsored (Stryker) — may influence device selection and reporting.
Low IV tPA use (5% thrombectomy, 13% control, due to late presentation) — limited data on combined IV tPA + late-window thrombectomy.
Stopped early after 206 of up to 500 planned patients — may overestimate treatment effect.
Bayesian adaptive design is complex and less intuitive than traditional frequentist analysis.
Posterior circulation occlusions were excluded — DAWN results cannot be extrapolated to basilar artery occlusion.

Funding

Stryker Neurovascular

Based on: DAWN (New England Journal of Medicine, 2018)

Authors: Nogueira RG, Jadhav AP, Haussen DC, ..., Budzik RF

Citation: N Engl J Med 2018;378:11–21

Content summarized and formatted by NeuroTrials.ai.

DAWN

(2018)

Objective

To evaluate the efficacy of endovascular thrombectomy 6 to 24 hours after stroke onset in patients with a clinical–imaging mismatch.

Study Summary

• Thrombectomy in patients with anterior circulation large-vessel occlusion and imaging-deficit mismatch within 24 hours from last known well was associated with better outcome.

Intervention

Endovascular thrombectomy with the Trevo device plus standard care vs standard care alone. Imaging-based eligibility required mismatch between NIHSS and infarct volume (measured by MRI or perfusion CT with RAPID).

Inclusion Criteria

Acute anterior circulation large-vessel occlusion, 6–24 hours from last known well, infarct core <21 mL (age ≥80), <31 mL (age <80, NIHSS ≥10), or 31–<51 mL (age <80, NIHSS ≥20).

Study Design

Arms: Thrombectomy + Standard Care vs. Standard Care Alone

Patients per Arm: Thrombectomy: 107, Standard Care: 99

Outcome

• Utility-weighted mRS at 90 days (first co-primary): 5.5 vs 3.4 (adjusted difference 2.0 points, 95% credible interval 1.1–3.0, posterior probability of superiority >0.999). • Functional independence (mRS 0–2 at 90 days, second co-primary): 49% vs 13% (adjusted difference 33 percentage points, 95% credible interval 21–44, posterior probability of superiority >0.999). • Stroke-related death at 90 days: 17/107 (16%) vs 18/99 (18%). • 90-day mortality: 20/107 (19%) vs 18/99 (18%) (P=1.00). • Symptomatic ICH: 6/107 (6%) vs 3/99 (3%) (P=0.50). • Procedure-related complications: 7/107 (7%). • Trial stopped early for efficacy at prespecified interim analysis (206 of up to 500 planned patients).

Bottom Line

Major Points

First RCT to show benefit of thrombectomy in the 6–24 hour window using clinical-core mismatch (RAPID automated perfusion software).
Used Bayesian adaptive enrichment design with 3 mismatch groups: Group A (age ≥80, NIHSS ≥10, core <21 mL), Group B (age <80, NIHSS ≥10, core <31 mL), Group C (age <80, NIHSS ≥20, core 31–<51 mL).
Eligible vessels: intracranial ICA or proximal MCA (M1 segment) on CTA or MRA.
Trial stopped early for efficacy after 206 of up to 500 planned patients at prespecified interim analysis (significance threshold posterior probability ≥0.986; observed posterior probability of superiority >0.999).
First co-primary end point (utility-weighted mRS at 90 days): 5.5 vs 3.4 (adjusted difference 2.0 points, 95% credible interval 1.1–3.0, posterior probability >0.999).
Second co-primary end point — functional independence (mRS 0–2) at 90 days: 49% vs 13% (adjusted difference 33 percentage points, 95% credible interval 21–44, NNT ≈ 2.8). Elevated from secondary to co-primary at FDA request 30 months into the trial while still blinded.
Successful reperfusion (mTICI 2b/3): 84%. Median time from last known well to randomization: 12.2 hours (thrombectomy arm). Median onset-to-reperfusion: 13.6 hours.
No significant increase in sICH (6/107 vs 3/99, P=0.50) or 90-day all-cause mortality (20/107 vs 18/99, P=1.00). Stroke-related death 17/107 vs 18/99. Procedure-related complications 7/107.
~55% of patients presented with wake-up stroke — DAWN validated treatment of wake-up strokes using perfusion mismatch selection.
Results, together with DEFUSE 3, led to AHA/ASA 2018 guideline update extending thrombectomy window to 24 hours for selected patients.

Study Design

Study Type: Randomized controlled trial
Randomization: Yes
Blinding: Blinded endpoint assessment
Sample Size: 206
Follow-up: 90 days
Centers: 26
Countries: USA, Canada, Europe, Australia

Primary Outcome

Definition: First co-primary end point: Score on the utility-weighted modified Rankin Scale at 90 days (range 0 [death] to 10 [no symptoms/disability]). Second co-primary end point: Functional independence (mRS 0–2) at 90 days = 49% vs 13% (adjusted difference 33 percentage points, 95% credible interval 21–44, posterior probability >0.999) — see Secondary Outcomes for full row.

Control	Intervention	HR/OR	P-value
3.4	5.5	- (1.1–3.0 (95% credible interval))	posterior probability of superiority >0.999

Limitations & Criticisms

Imbalance in baseline atrial fibrillation (40% thrombectomy vs 24% control) — AF associated with larger infarcts and worse outcomes.
Strict exclusion of large ischemic cores (>51 mL) limits generalizability — later addressed by SELECT2, ANGEL-ASPECT, and RESCUE-Japan LIMIT.
RAPID perfusion imaging software required — not universally available, particularly in lower-resource settings.
Industry-sponsored (Stryker) — may influence device selection and reporting.
Low IV tPA use (5% thrombectomy, 13% control, due to late presentation) — limited data on combined IV tPA + late-window thrombectomy.
Stopped early after 206 of up to 500 planned patients — may overestimate treatment effect.
Bayesian adaptive design is complex and less intuitive than traditional frequentist analysis.
Posterior circulation occlusions were excluded — DAWN results cannot be extrapolated to basilar artery occlusion.

Citation

N Engl J Med 2018;378:11–21

View Original Publication →

DAWN

Thrombectomy 6 to 24 Hours after Stroke with a Mismatch between Deficit and Infarct

Clinical Question

Bottom Line

Major Points

Design

Inclusion Criteria

Exclusion Criteria

Baseline Characteristics

Arms

Outcomes

Subgroup Analysis

Criticisms

Funding

Ask about DAWN