← Back
NeuroTrials.ai
Neurology Clinical Trial Database

SELECT Late

Association of Endovascular Thrombectomy vs Medical Management With Functional and Safety Outcomes in Patients Treated Beyond 24 Hours of Last Known Well

Share Share Copied! Compare

Year of Publication: 2023

Authors: Amrou Sarraj, Timothy J. Kleinig, Ameer E. Hassan, ..., Bruce C. V. Campbell

Journal: JAMA Neurology

Citation: JAMA Neurol. 2023;80(2):172-182

Clinical Question

Is endovascular thrombectomy associated with better functional independence compared with medical management in patients treated beyond 24 hours after they were last known well?

Bottom Line

EVT beyond 24 hours of last known well was associated with significantly higher functional independence (aOR 4.56) and lower mortality (aOR 0.49) compared with medical management, though with increased symptomatic ICH risk. Benefits persisted across propensity-matched cohorts.

Major Points

  • EVT achieved functional independence in 38% vs 10% with medical management (IPTW aOR 4.56, 95% CI 2.28-9.09, P<0.001)
  • Mortality was lower with EVT (26% vs 41%, IPTW aOR 0.49, 95% CI 0.27-0.89, P=0.02)
  • sICH was significantly higher with EVT (10.1% vs 1.7%, IPTW aOR 10.65, P=0.003)
  • Results consistent across three propensity-matched cohorts using clinical characteristics, ASPECTS, and perfusion parameters
  • In patients with perfusion mismatch, EVT showed aOR 4.82 for functional independence (P=0.049)
  • 81% of EVT patients achieved successful reperfusion (mTICI 2b-3)
  • Time from last known well to procedure and low ASPECTS (0-5) independently associated with sICH
  • No significant interaction between EVT benefit and ASPECTS or ischemic core size

Design

Study Type: Retrospective observational cohort study

Randomization:

Blinding: None (retrospective observational study)

Enrollment Period: July 2012 to December 2021

Follow-up Duration: 90 days

Centers: 17

Countries: United States, Spain, Australia, New Zealand

Sample Size: 301

Analysis: Propensity score-weighted multivariable logistic regression with inverse probability treatment weighting (IPTW); 1:1 propensity score matching without replacement using nearest-neighbor method; Stata version 15

Inclusion Criteria

  • Adults aged ≥18 years
  • Acute ischemic stroke due to large-vessel occlusion
  • Occlusion in internal carotid artery or M1/M2 segments of middle cerebral artery
  • Treatment beyond 24 hours of last known well

Exclusion Criteria

  • No ICA/M1/M2 occlusion
  • Unknown or unavailable time parameters

Arms

FieldControlEndovascular Thrombectomy
InterventionBest medical management according to local guidelinesEndovascular thrombectomy using stent retriever, aspiration device, or combination, plus best medical management
Duration90 days follow-up90 days follow-up

Outcomes

OutcomeTypeControlInterventionHR / OR / RRP-value
Functional independence defined as modified Rankin Scale score 0-2 at 90-day follow-upPrimary10.4% (12/115)38.1% (69/181)4.56<0.001
Nondisabled (mRS 0-1)Secondary3.5%18.2%<0.001
Ambulatory or bodily-needs capable (mRS 0-3)Secondary23.5%50.8%<0.001
Requiring constant care or dead (mRS 5-6)Secondary49.6%32.6%0.004
Functional independence - PS matched (clinical + ASPECTS)Secondary19%35%3.140.047
Functional independence - PS matched (clinical + perfusion)Secondary17%35%4.170.03
Functional independence - PS matched (clinical + ASPECTS + perfusion)Secondary21%45%4.390.04
Functional independence in patients with perfusion mismatchSecondary11%32%4.820.049
Functional independence - witnessed stroke onset subgroupSecondary13%34%7.610.03
90-day MortalityAdverse40.9%26.0%0.490.02
Symptomatic ICH (SITS-MOST definition)Adverse1.8%10.1%10.650.003
Asymptomatic ICHAdverse9.1%21.1%0.008
Neurological worsening (≥4 points NIHSS)Adverse18.6%23.6%0.31

Subgroup Analysis

EVT benefit consistent across ASPECTS 6-10 (EVT 39% vs Control 21%, P=0.04) and ASPECTS 0-5 (EVT 21% vs Control 5%, P=0.02) with no significant interaction (P=0.18). Similarly consistent across small (<50mL) ischemic core (EVT 37% vs Control 19%, P=0.04) and large (≥50mL) core (EVT 17% vs Control 0%, P=0.15) without significant interaction (P=0.49). Time from LKW to procedure not significantly associated with functional independence (aOR 0.99 per hour, P=0.2). ASPECTS 0-5 (aOR 4.58, P=0.01) and time from LKW (aOR 1.02 per hour, P=0.02) independently associated with sICH.

Criticisms

  • Retrospective observational design with inherent selection bias
  • Treatment allocation not randomized - physicians selected patients perceived more likely to benefit for EVT
  • Significant baseline imbalances between groups (EVT patients had lower NIHSS, better ASPECTS, smaller ischemic cores)
  • Unable to adjust for unmeasured confounders
  • Imaging protocols varied by site and at discretion of local investigators
  • Limited sample sizes in propensity-matched cohorts reduced precision of estimates
  • Six centers reported EVT cases only, potentially introducing additional selection bias
  • No follow-up infarct volumes available to assess impact on final infarct size
  • Perfusion imaging not available for all patients (62% overall)
  • Low event rates for sICH limited statistical power for safety analyses

Funding

Not explicitly stated; author disclosures include grants and personal fees from Stryker Neurovascular, Medtronic, Microvention, Penumbra, Cerenovus, Genentech, and others

Based on: SELECT Late (JAMA Neurology, 2023)

Authors: Amrou Sarraj, Timothy J. Kleinig, Ameer E. Hassan, ..., Bruce C. V. Campbell

Citation: JAMA Neurol. 2023;80(2):172-182

Content summarized and formatted by NeuroTrials.ai.