← Back
NeuroTrials.ai
Neurology Clinical Trial Database

AURORA

Thrombectomy for anterior circulation stroke beyond 6 h from time last known well (AURORA): a systematic review and individual patient data meta-analysis

Year of Publication: 2022

Authors: Jovin TG, Nogueira RG, Lansberg MG, ..., Albers GW

Journal: The Lancet

Citation: Lancet 2022; 399: 249–58. Published Online November 11, 2021.

Link: https://doi.org/10.1016/S0140-6736(21)01341-6


Clinical Question

Does endovascular thrombectomy improve outcomes in patients with anterior circulation large vessel occlusion stroke randomly assigned 6–24 hours from time last known well?

Bottom Line

Endovascular thrombectomy provides substantial benefit in patients with anterior circulation large vessel occlusion stroke across the 6–24 hour time window when imaging evidence of reversible cerebral ischemia is present. Thrombectomy should not be withheld on the basis of advanced age, moderate infarct size, mode of presentation, or point in time within the 6–24 h window.

Major Points

  • Individual patient data meta-analysis of 6 RCTs including 505 patients randomized beyond 6 h from time last known well
  • Primary outcome (mRS ordinal shift at 90 days) favored thrombectomy: adjusted cOR 2.54 (95% CI 1.83–3.54; p<0.0001); NNT = 3
  • Functional independence (mRS 0–2, secondary) more than doubled: 45.9% vs 19.3% (adjusted OR 3.88, 1.94–7.78; p<0.0001)
  • No increase in mortality (16.5% vs 19.3%; adjusted OR 0.96, p=0.88) or symptomatic ICH (5.3% vs 3.3%; adjusted OR 1.74, p=0.23)
  • Treatment effect was stronger in the 12–24 h window (cOR 5.86) than 6–12 h (cOR 1.76); p-interaction=0.0087
  • No heterogeneity of effect across age, gender, NIHSS, occlusion site, ASPECTS, or mode of presentation (wake-up, witnessed, unwitnessed)
  • Findings support broadening eligibility for thrombectomy across the 6–24 h window in patients with evidence of salvageable brain tissue

Design

Study Type: Systematic review and individual patient data meta-analysis of randomized controlled trials

Randomization: 1

Enrollment Period: Trials published/searched Jan 1, 2010 through March 1, 2021

Follow-up Duration: 90 days

Centers: 0

Countries: USA, Canada, France, Spain, Ireland, South Korea, UK, Brazil

Sample Size: 505

Analyzed: 505

Analysis: Mixed-effects ordinal logistic regression for the primary outcome (mRS shift at 90 days); unadjusted and adjusted common odds ratios (adjusted for age, gender, baseline NIHSS, baseline ASPECTS, location of occlusion [intracranial ICA vs MCA], and time from onset to random assignment). Between-trial variance modeled by including trial and trial*treatment as random effects.


Inclusion Criteria

  • Adult patients enrolled in one of six qualifying RCTs (DAWN, DEFUSE 3, ESCAPE, RESILIENT, POSITIVE, REVASCAT)
  • Acute ischemic stroke due to anterior circulation large vessel occlusion (intracranial ICA, M1 MCA, or both)
  • Random assignment beyond 6 h from time last known well (up to 8–24 h depending on source trial)
  • Imaging evidence of salvageable brain tissue per source-trial protocol (varying: CT perfusion / MRI mismatch, ASPECTS threshold, or infarct volume threshold)

Exclusion Criteria

  • Largest infarcts at baseline (excluded across all source trials by ASPECTS or infarct-volume thresholds)
  • Posterior circulation occlusions
  • Presentation beyond 24 h from time last known well
  • Mild neurological deficits (NIHSS <6)
  • Substantial pre-stroke disability (mRS ≥2)

Baseline Characteristics

Overall:

  • N: 505
  • Mean Age (years): 68.6 (SD 13.7)
  • Women n (%): 259 (51.3%)
  • Median baseline NIHSS: 16 (IQR 13–20)
  • Median baseline ASPECTS: 8 (IQR 7–9)
  • Intracranial ICA occlusion: 128/505 (25.3%)
  • M1 MCA occlusion: 362/505 (71.7%)
  • IV alteplase given: 66/505 (13.1%)
  • Median onset to randomization (min): 625 (IQR 472–808)
  • Thrombectomy N: 266
  • Control N: 239

Arms

FieldEndovascular ThrombectomyControl
N266239
InterventionMechanical thrombectomy with modern-generation devices (stent-retrievers such as Trevo or Solitaire, and/or aspiration catheters such as Penumbra) plus best medical therapyBest medical therapy alone (no endovascular intervention)
DurationSingle procedure; 90-day outcome assessment90-day outcome assessment

Outcomes

OutcomeTypeControlInterventionHR / OR / RRP-value
Disability on the modified Rankin Scale (mRS) at 90 days, analyzed by mixed-effects ordinal logistic regression (shift analysis across the full 0–6 mRS distribution)Primaryn=239; mRS shift distribution at 90 days (see Figure 1)n=266; mRS shift distribution at 90 days (see Figure 1)2.54<0.0001
Functional independence (mRS 0–2) at 90 daysSecondary46/238 (19.3%)122/266 (45.9%)Rate ratio 2.37 (95% CI 1.69–3.33); adjusted rate ratio 2.19 (1.44–3.34); OR 2.99 (1.70–5.25); adjusted OR 3.88 (1.94–7.78)<0.0001
Excellent outcome (mRS 0–1) at 90 daysSecondary30/238 (12.6%)75/266 (28.2%)Rate ratio 2.24 (1.46–3.42); adjusted rate ratio 1.99 (1.15–3.45); OR 2.42 (1.27–4.60); adjusted OR 2.41 (1.07–5.43)0.0002 (unadjusted RR); 0.034 (adjusted OR)
NIHSS 0–2 at 24 hSecondary10/234 (4.3%)41/262 (15.6%)Rate ratio 3.66 (1.83–7.32); adjusted rate ratio 3.69 (1.82–7.48); OR 4.15 (2.03–8.51); adjusted OR 4.85 (2.21–10.61)<0.0001
Early neurological recovery (ENR: ≥8-point NIHSS reduction or NIHSS 0–1) at 24 hSecondary33/234 (14.1%)106/262 (40.5%)Rate ratio 2.86 (1.94–4.24); adjusted rate ratio 2.87 (1.93–4.26); OR 4.23 (2.08–8.61); adjusted OR 4.23 (2.08–8.61)<0.0001
NIHSS at 24 h (median)Secondary15.08.0Median difference favoring thrombectomy
Change in NIHSS from baseline to 24 h (median)Secondary–2.0–6.0Greater median NIHSS reduction with thrombectomy
90-day mortalitySafety46/238 (19.3%)44/266 (16.5%)Adjusted OR 0.96 (0.58–1.60); adjusted rate ratio 0.97 (0.64–1.48)0.88 (adjusted OR); 0.90 (adjusted rate ratio)
Symptomatic intracerebral hemorrhage within 90 daysSafety8/239 (3.3%)14/266 (5.3%)Adjusted OR 1.74 (0.70–4.31); adjusted rate ratio 1.68 (0.70–4.04)0.23 (adjusted OR); 0.25 (adjusted rate ratio)

Subgroup Analysis

No heterogeneity of treatment effect noted across subgroups defined by age, gender, baseline stroke severity (NIHSS), vessel occlusion site, baseline ASPECTS, or mode of presentation (wake-up, witnessed onset, unwitnessed onset). Treatment effect was stronger in patients randomly assigned within 12–24 h (common OR 5.86, 95% CI 3.14–10.94) than those randomly assigned within 6–12 h (common OR 1.76, 95% CI 1.18–2.62); p-interaction=0.0087.


Criticisms

  • Heterogeneous imaging selection criteria across included trials — some required advanced perfusion imaging (DAWN, DEFUSE 3), others relied on ASPECTS alone
  • Small contributions from four of the six trials (ESCAPE n=49; RESILIENT, POSITIVE, REVASCAT each ~21–26) — pooled estimates driven largely by DAWN and DEFUSE 3
  • Findings apply only to patients with imaging evidence of reversible cerebral ischemia; generalizability to unselected late-window patients cannot be inferred
  • Four included trials (DEFUSE 3, ESCAPE, POSITIVE, REVASCAT) were stopped early, raising possibility of treatment-effect overestimation
  • Funded by Stryker Neurovascular, a manufacturer of thrombectomy devices — potential for sponsorship bias

Funding

Stryker Neurovascular

Based on: AURORA (The Lancet, 2022)

Authors: Jovin TG, Nogueira RG, Lansberg MG, ..., Albers GW

Citation: Lancet 2022; 399: 249–58. Published Online November 11, 2021.

Content summarized and formatted by NeuroTrials.ai.