← Back
NeuroTrials.ai
Neurology Clinical Trial Database

TENSION

Endovascular thrombectomy for acute ischaemic stroke with established large infarct: multicentre, open-label, randomised trial

Share Share Copied! Compare

Year of Publication: 2023

Authors: Bendszus M, Fiehler J, Subtil F, ..., Thomalla G; TENSION Investigators

Journal: Lancet

Citation: Lancet 2023; 402: 1753-63. Published Online October 11, 2023.

Link: https://doi.org/10.1016/S0140-6736(23)02032-9

Clinical Question

Does endovascular thrombectomy improve functional outcome in acute ischaemic stroke with large established infarct (ASPECTS 3-5) when patient selection relies on standard non-contrast CT?

Bottom Line

In patients with acute ischaemic stroke from anterior circulation large vessel occlusion and large established infarct (ASPECTS 3-5), endovascular thrombectomy up to 12 hours from onset improved functional outcomes (adjusted common OR 2.58) and reduced mortality (HR 0.67) compared with medical treatment alone, using non-contrast CT-based selection in the majority of patients.

Major Points

  • First trial to randomise large-infarct stroke patients using non-contrast CT as the predominant imaging modality (82% CT, 18% MRI)
  • Trial stopped early for efficacy after the first pre-planned interim analysis
  • Primary outcome: significant shift in mRS distribution at 90 days favoring thrombectomy (median mRS 4 [3-6] vs 6 [4-6]; adjusted common OR 2.58, 95% CI 1.60-4.15; p=0.0001)
  • Mortality reduction with thrombectomy (49/122 [40%] vs 63/123 [51%]; HR 0.67, 95% CI 0.46-0.98; p=0.038) — a novel finding not seen in prior large-core thrombectomy trials
  • No safety concerns: symptomatic ICH similar between arms (5% vs 5%)
  • Extended time window up to 12 hours from symptom onset/last known well
  • Supports thrombectomy for large-core stroke selectable by non-contrast CT alone

Design

Study Type: Prospective multicentre, open-label, blinded-endpoint, two-arm, investigator-initiated, post-market randomised controlled trial

Randomization: 1

Blinding: Open-label with blinded outcome assessment (PROBE design); outcome assessors and core lab imaging readers masked to treatment assignment

Allocation: 1:1 central web-based randomisation using a permuted block design (random block sizes of two and four), stratified by time from symptom onset or last known well (<6 h and 6-11 h) and stroke severity (NIHSS ≤18 and NIHSS 19-25)

Enrollment Period: July 17, 2018 to February 21, 2023

Follow-up Duration: 90 days

Centers: 41

Countries: Austria, Czech Republic, Denmark, France, Germany, Norway, Slovakia, Spain, Canada

Sample Size: 253

Analyzed: 253

Analysis: Intention-to-treat for primary efficacy analysis; safety analysis in the safety population based on treatment received

Power Calculation: Assuming a proportional-odds alternative with an odds ratio of 1.5 for the primary mRS shift analysis, a total of 620 patients were required to achieve 80% power for a one-sided test at the 0.025 level. Assuming a 5% dropout rate for the 90-day primary endpoint, an effective sample size of 665 was necessary. Two interim analyses (for futility and early efficacy) were planned after one-third and two-thirds of patients had completed 90-day follow-up.

Registration: ClinicalTrials.gov NCT03094715

Inclusion Criteria

  • Age ≥18 years
  • Acute ischaemic stroke due to focal occlusion in the M1 segment of the middle cerebral artery or intracranial segment of the distal internal carotid artery on CTA or MRA
  • ASPECTS 3-5 on unenhanced CT or DWI (assessed locally)
  • NIHSS <26
  • Premorbid mRS 0-2 (premorbidly independent)
  • Randomisation within 11 h of symptom onset or last known well
  • Expected completion of endovascular thrombectomy within 12 h

Exclusion Criteria

  • Known vascular disease preventing endovascular thrombectomy
  • High-grade extracranial stenosis expected to require acute stent placement
  • Any acute intracranial bleeding
  • Mass effect on imaging

Baseline Characteristics

0:

  • Characteristic: Median age, years (IQR)
  • Intervention (n=125): 73 (65-81)
  • Control (n=128): 74 (64-80)

1:

  • Characteristic: Female sex, n (%)
  • Intervention (n=125): 56 (45%)
  • Control (n=128): 67 (52%)

2:

  • Characteristic: Median NIHSS at hospital arrival (IQR)
  • Intervention (n=125): 19 (16-22)
  • Control (n=128): 18 (15-22)

3:

  • Characteristic: Median pre-stroke mRS (IQR)
  • Intervention (n=125): 0 (0-1)
  • Control (n=128): 0 (0-1)

4:

  • Characteristic: Hypertension, n/N (%)
  • Intervention (n=125): 99/123 (80%)
  • Control (n=128): 98/121 (81%)

5:

  • Characteristic: Diabetes, n/N (%)
  • Intervention (n=125): 27/119 (23%)
  • Control (n=128): 29/121 (24%)

6:

  • Characteristic: Atrial fibrillation, n/N (%)
  • Intervention (n=125): 31/117 (26%)
  • Control (n=128): 48/118 (41%)

7:

  • Characteristic: ASPECTS 3 (local read), n (%)
  • Intervention (n=125): 36 (29%)
  • Control (n=128): 48 (38%)

8:

  • Characteristic: ASPECTS 4 (local read), n (%)
  • Intervention (n=125): 45 (36%)
  • Control (n=128): 39 (30%)

9:

  • Characteristic: ASPECTS 5 (local read), n (%)
  • Intervention (n=125): 44 (35%)
  • Control (n=128): 41 (32%)

10:

  • Characteristic: Internal carotid artery occlusion, n/N (%)
  • Intervention (n=125): 41/125 (33%)
  • Control (n=128): 37/127 (29%)

11:

  • Characteristic: M1 MCA occlusion, n/N (%)
  • Intervention (n=125): 83/125 (66%)
  • Control (n=128): 88/127 (69%)

12:

  • Characteristic: CT used for enrolment imaging, n (%)
  • Intervention (n=125): 104 (83%)
  • Control (n=128): 104 (81%)

13:

  • Characteristic: IV alteplase administered, n (%)
  • Intervention (n=125): 49 (39%)
  • Control (n=128): 44 (34%)

Arms

FieldEndovascular thrombectomy + medical treatmentControl
N125128
InterventionEndovascular thrombectomy in addition to medical treatment (standard of care), initiated up to 12 h from stroke onsetMedical treatment (standard of care) alone
DurationSingle procedure with 90-day follow-up90-day follow-up

Outcomes

OutcomeTypeControlInterventionHR / OR / RRP-value
Functional outcome across the entire range of the modified Rankin Scale (shift analysis)PrimaryMedical treatment alone (n=128); median mRS 6 (IQR 4-6)Endovascular thrombectomy (n=125); median mRS 4 (IQR 3-6)2.580.0001
Mortality at 90 days (secondary efficacy)Secondary63/123 (51%)49/122 (40%)HR 0.670.038
Independent functional outcome (mRS 0-2) at 90 daysSecondary3/122 (2%)21/124 (17%)adjusted OR 7.160.0016
Moderate functional outcome (mRS 0-3) at 90 daysSecondary16/122 (13%)39/124 (31%)adjusted OR 2.840.0018
Decompressive craniotomy during in-hospital treatmentSecondary9 (7%)11 (9%)0.65
Successful recanalisation (mTICI 2b or better) — thrombectomy arm onlySecondarynot applicable104 (83%); 83% (95% CI 76-89)
Distal embolisation — thrombectomy arm onlySecondarynot applicable4 (3%); 3% (95% CI 1-8)
Mean infarct volume at 24 (±6) h after randomisation, mL (SD)Secondary227.7 (107.2)205.8 (139.1)adjusted ratio 0.930.35
EQ-5D index at 90 (±14) days, median (IQR)Secondary0.4 (0.2-0.6)0.6 (0.3-0.9)0.0060
EQ-5D visual analogue scale at 90 (±14) days, median (IQR)Secondary40 (20-58)50 (30-70)0.13
PROMIS-10 physical health T-score at 90 (±14) days, median (IQR)Secondary34.9 (29.6-37.4)39.8 (34.9-50.8)0.0008
PROMIS-10 mental health T-score at 90 (±14) days, median (IQR)Secondary38.8 (32.6-43.5)43.5 (36.3-50.8)0.025
PHQ-4 anxiety at 90 (±14) days, n/N (%)Secondary18/35 (51%)12/52 (23%)0.011
PHQ-4 depression at 90 (±14) days, n/N (%)Secondary11/35 (31%)14/52 (27%)0.81
Symptomatic intracranial haemorrhage (safety population)Safety6/125 (5%)7/128 (5%)1.00
Mortality (censored at 90 days) (safety population)Safety63/123 (51%)49/122 (40%)HR 0.670.038
Death or dependency (mRS 4-6) at 90 (±14) daysAdverse103/119 (87%)88/127 (69%)adjusted OR 0.340.0011
Mortality at 7 days or dischargeAdverse36 (29%)30 (23%)OR 0.750.31
Stroke-related deathAdverse36/119 (30%)33/127 (26%)OR 0.800.44
New ischaemic strokeAdverse20 (16%)17 (13%)OR 0.800.52
Space-occupying infarctionAdverse17 (14%)21 (16%)OR 1.260.52
Parenchymal haemorrhage type 2Adverse10/116 (9%)11/124 (9%)0.95
Fatal symptomatic intracranial haemorrhageAdverse4 (3%)3 (2%)0.72
At least one serious adverse eventAdverse88 (70%)71 (55%)0.014

Subgroup Analysis

Prespecified subgroup analyses were generally consistent with the primary analysis and favoured endovascular thrombectomy in almost all prespecified subgroups; the trial was not powered for subgroup analyses. An exploratory (not prespecified) analysis adjusting for baseline ASPECTS was also consistent with the primary outcome.

Criticisms

  • Open-label design (endpoints assessed blinded, but treatment allocation known to clinicians and patients)
  • Trial stopped early for efficacy — early stopping can overestimate effect size
  • Small sample size (n=253) limits precision of subgroup and secondary outcome estimates
  • Predominantly European population; single Canadian site — generalisability to other settings requires confirmation
  • Imbalance in baseline ASPECTS favoured medical treatment group (lower ASPECTS 0-2 more common in the medical group per core lab)

Funding

EU Horizon 2020 research and innovation programme (grant 754640)

Based on: TENSION (Lancet, 2023)

Authors: Bendszus M, Fiehler J, Subtil F, ..., Thomalla G; TENSION Investigators

Citation: Lancet 2023; 402: 1753-63. Published Online October 11, 2023.

Content summarized and formatted by NeuroTrials.ai.