← Back
NeuroTrials.ai
Neurology Clinical Trial Database

SELECT2 1-Year

Endovascular thrombectomy plus medical care versus medical care alone for large ischaemic stroke: 1-year outcomes of the SELECT2 trial

Year of Publication: 2024

Authors: Amrou Sarraj, Michael G Abraham, Ameer E Hassan, ..., Bruce C V Campbell

Journal: The Lancet

Citation: Sarraj A, Abraham MG, Hassan AE, et al. Endovascular thrombectomy plus medical care versus medical care alone for large ischaemic stroke: 1-year outcomes of the SELECT2 trial. Lancet. 2024;403(10428):731-740.

Link: https://doi.org/10.1016/S0140-6736(24)00050-3


Clinical Question

Does endovascular thrombectomy provide long-term (1-year) functional benefit compared with medical care alone in patients with acute ischaemic stroke due to proximal large vessel occlusion and large ischaemic core?

Bottom Line

In patients with large ischaemic stroke (ASPECTS 3-5 or core ≥50mL), thrombectomy plus medical care provided significant functional benefit at 1 year compared with medical care alone. Thrombectomy nearly quadrupled the rate of functional independence (24% vs 6%) with improved quality of life scores and non-significantly lower mortality.

Major Points

  • Prespecified 1-year follow-up analysis of the SELECT2 trial (originally stopped early for efficacy at 90 days)
  • First trial showing 1-year benefit of thrombectomy in large core stroke patients
  • 1-year mRS distribution significantly favored thrombectomy (gOR 1.43, p=0.0019)
  • Functional independence (mRS 0-2) at 1 year: 24% thrombectomy vs 6% medical care (RR 3.17)
  • Independent ambulation (mRS 0-3): 37% vs 18% (RR 1.85)
  • Number needed to treat for 1-point mRS improvement: 6 patients
  • Mortality numerically lower with thrombectomy (45% vs 52%) but not statistically significant
  • Quality of life (Neuro-QOL) scores higher in thrombectomy group across mobility, social health, and cognitive domains
  • 24% of survivors showed ≥1-point mRS improvement between 90 days and 1 year
  • Treatment effect consistent across subgroups including age, NIHSS, time window, ASPECTS, and core volume
  • 93% follow-up rate at 1 year (329/352 patients)

Design

Study Type: Phase 3, open-label, international, multicentre, randomised controlled trial with blinded endpoint assessment (prespecified 1-year follow-up analysis)

Randomization: 1

Blinding: Open-label with blinded outcome assessment. Trained assessors masked to treatment and imaging findings collected outcomes.

Enrollment Period: October 11, 2019 to September 9, 2022

Follow-up Duration: 1 year

Centers: 31

Countries: USA, Canada, Spain, Switzerland, Australia, New Zealand

Sample Size: 352

Analysis: Intention-to-treat. Wilcoxon-Mann-Whitney test for primary outcome. Modified Poisson regression with robust SEs for dichotomous outcomes. Median regression for continuous outcomes. Probabilistic index models for adjusted analyses. Multiple imputation for missing data with pattern-mixture sensitivity analyses. No adjustment for multiplicity for secondary outcomes.


Inclusion Criteria

  • Age 18-85 years
  • Pre-stroke modified Rankin Scale score 0-1
  • Acute ischaemic stroke due to occlusion of internal carotid artery or first segment (M1) of middle cerebral artery
  • Large ischaemic core defined as: ASPECTS 3-5 on non-contrast CT, OR CT perfusion core volume ≥50mL (rCBF <30%), OR MRI diffusion core ≥50mL (ADC <620 × 10⁻⁶ mm²/s)
  • Presenting within 24 hours of last known well

Exclusion Criteria

  • Pre-stroke disability (mRS >1)
  • CT ASPECTS >5 with core <50mL (per protocol exclusion)

Arms

FieldEndovascular thrombectomy plus medical careControl
InterventionEndovascular thrombectomy using stent retrievers, aspiration devices, or both, plus best medical management according to local institutional protocols based on AHA/ASA, European, and Australian guidelines. Anaesthesia choice, access site, and peri-procedural management per local team.Best medical management according to local institutional protocols based on AHA/ASA, European, and Australian guidelines. Hemicraniectomy permitted according to local protocols and patient wishes.
DurationAcute intervention with 1-year follow-up1-year follow-up

Outcomes

OutcomeTypeControlInterventionHR / OR / RRP-value
Ordinal modified Rankin Scale (mRS 0-6) at 1-year follow-up, with scores 5 and 6 merged for analysisPrimaryMedian mRS 6 (IQR 4-6)Median mRS 5 (IQR 3-6)60.0019
Functional independence (mRS 0-2) at 1 yearSecondary9/159 (6%)40/170 (24%)RR 3.17Significant
Independent ambulation (mRS 0-3) at 1 yearSecondary29/159 (18%)63/170 (37%)RR 1.85Significant
Severe disability or death (mRS 5-6) at 1 yearSecondary106/159 (67%)91/170 (54%)RR 0.81Significant
All-cause mortality at 1 yearSecondary83/159 (52%)77/170 (45%)RR 0.89Not significant
Neuro-QOL Mobility score at 1 year (higher = better)Secondary27.2 (19.2-38.3)40.8 (26.2-48.6)Difference 13.37Significant
Neuro-QOL Depression score at 1 year (higher = worse)Secondary52.1 (47.9-57.4)45.3 (36.9-55.1)Difference -6.72Not significant
Neuro-QOL Social health score at 1 year (higher = better)Secondary35.0 (30.5-41.3)40.6 (33.5-44.7)Difference 5.60Significant
Neuro-QOL Cognitive score at 1 year (higher = better)Secondary38.9 (33.0-46.0)46.0 (37.0-59.0)Difference 6.98Significant
mRS distribution at 1 yearSecondarymRS 0: 1%, mRS 1: 3%, mRS 2: 3%, mRS 3: 13%, mRS 4: 15%, mRS 5: 15%, mRS 6: 52%mRS 0: 7%, mRS 1: 14%, mRS 2: 14%, mRS 3: 9%, mRS 4: 8%, mRS 5: 2%, mRS 6: 45%
Evolution of infarction/cerebral oedema with midline shiftAdverse50 (29%)62 (35%)Not reported
Neurological deteriorationAdverse27 (16%)44 (25%)Not reported
Progression of disease/palliative care decisionAdverse38 (22%)31 (17%)Not reported
Symptomatic intracranial haemorrhageAdverse2 (1%)1 (1%)Not reported
Respiratory failureAdverse19 (11%)30 (17%)Not reported
PneumoniaAdverse21 (12%)18 (10%)Not reported
Urinary tract infectionAdverse13 (7%)12 (7%)Not reported
Acute kidney injuryAdverse9 (5%)9 (5%)Not reported
Myocardial infarctionAdverse3 (2%)2 (1%)Not reported

Subgroup Analysis

No significant difference in thrombectomy treatment effect across prespecified subgroups including age (<70 vs ≥70), NIHSS (<20 vs ≥20), time from last known well (<12h vs ≥12h, <6h vs ≥6h), clot location (ICA vs MCA), CT ASPECTS (0-2, 3-5, 6-10), core volume (<70mL vs ≥70mL, <100mL vs ≥100mL, <150mL vs ≥150mL), mismatch status, and affected hemisphere. Patients enrolled at non-US sites had lower point estimates of treatment effect than those at US sites (gOR 1.06 vs 1.54), though trial not powered for subgroup comparisons.


Criticisms

  • 7% missing primary outcome data at 1-year (higher than REVASCAT but lower than MR CLEAN)
  • Trial not powered to evaluate treatment effect differences across subgroups
  • Neuro-QOL scores missing for considerable number of patients with missingness associated with treatment allocation
  • Stroke recurrence data missing for ~50% of patients, primarily due to deaths
  • Adverse event monitoring completed only up to 90-day follow-up
  • Did not collect cause of mortality beyond 90 days
  • Did not collect duration and type of rehabilitation received
  • Only 20% of patients enrolled at non-US sites, limiting generalizability assessment
  • Two protocol crossovers occurred (both at same site)
  • Open-label design (though outcome assessment was blinded)

Funding

Stryker Neurovascular

Based on: SELECT2 1-Year (The Lancet, 2024)

Authors: Amrou Sarraj, Michael G Abraham, Ameer E Hassan, ..., Bruce C V Campbell

Citation: Sarraj A, Abraham MG, Hassan AE, et al. Endovascular thrombectomy plus medical care versus medical care alone for large ischaemic stroke: 1-year outcomes of the SELECT2 trial. Lancet. 2024;403(10428):731-740.

Content summarized and formatted by NeuroTrials.ai.