PDF: RESCUE-Japan LIMIT
(2022)Objective
Endovascular therapy (EVT) plus medical care versus medical care alone in patients with large ischemic cores (ASPECTS 3–5) due to large vessel occlusion.
Study Summary
Intervention
Endovascular therapy (e.g., stent retriever, aspiration, balloon angioplasty) plus standard medical care vs. standard medical care alone. IV alteplase (0.6 mg/kg) was used when indicated in both groups.
Inclusion Criteria
Acute ischemic stroke, age ≥18, NIHSS ≥6, pre-stroke mRS 0–1, ASPECTS 3–5, ICA or M1 occlusion, <6 hours from LKW or 6–24 hours with DWI-FLAIR mismatch, EVT initiation within 60 min of randomization.
Study Design
Arms: Endovascular Therapy (EVT) + Medical Care vs. Medical Care Alone
Patients per Arm: EVT: 100, Medical Care: 102
Outcome
Bottom Line
In this Japanese trial, patients with acute stroke from a large vessel occlusion and a large ischemic core who were treated with endovascular therapy had significantly better functional outcomes at 90 days than those who received medical care alone. However, endovascular therapy was associated with a higher frequency of any intracranial hemorrhage.
Major Points
- RESCUE-Japan LIMIT was the FIRST positive RCT demonstrating thrombectomy benefit for large ischemic core strokes (ASPECTS 3-5), a population previously considered 'too far gone' for intervention.
- Primary outcome (mRS 0-3 at 90 days): 31.0% vs 12.7% (RR 2.43, 95% CI 1.35-4.37, p=0.002), NNT ≈ 5.5 — a dramatic absolute benefit despite the large established infarcts.
- Published in NEJM 2022, it fundamentally shifted the paradigm: large core was no longer an absolute contraindication to thrombectomy. Used ASPECTS (CT or DWI-MRI) rather than perfusion imaging for selection.
- Favorable ordinal shift across all mRS categories (common OR 2.42, 95% CI 1.46-4.01), confirming benefit was not just in one part of the disability spectrum.
- More ICH in the thrombectomy group (58% vs 31%, p<0.001), but sICH rates were NOT significantly different (9% vs 4.9%, p=0.25) — most hemorrhages were asymptomatic petechial transformation.
- Japanese alteplase dose (0.6 mg/kg) is lower than the international standard (0.9 mg/kg), which may underestimate bleeding risk if extrapolated to international practice.
- Confirmed by subsequent trials: SELECT2 (NEJM 2023, US/international), ANGEL-ASPECT (NEJM 2023, China), and TENSION (NEJM 2024, Europe) all showed large core thrombectomy benefit.
- Notably used ASPECTS 3-5 (moderate-large core) — did NOT include ASPECTS 0-2 (massive core), which remains controversial.
- Median NIHSS was 22 in both arms — these were severe strokes with poor natural history (only 12.7% achieved mRS 0-3 with medical care alone).
- Led to 2023 AHA/ASA guideline update recommending thrombectomy for select patients with large core infarcts (Class 2a recommendation).
Study Design
- Study Type
- Multicenter, open-label, randomized clinical trial.
- Randomization
- Yes
- Blinding
- Open-label for treatment, but the primary outcome (mRS score) was assessed by trained personnel who were unaware of the trial-group assignments.
- Sample Size
- 203
- Follow-up
- 90 days.
- Centers
- 45
- Countries
- Japan
Primary Outcome
Definition: A score of 0 to 3 on the modified Rankin scale at 90 days.
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| 12.7% (13/102) | 31.0% (31/100) | 2.43 (1.35 to 4.37) | 0.002 |
Limitations & Criticisms
- Conducted exclusively in Japan — Japanese stroke patients differ in body habitus, intracranial atherosclerosis prevalence, and treatment patterns (lower alteplase dose of 0.6 mg/kg). Generalizability was uncertain until confirmed by SELECT2, ANGEL-ASPECT, and TENSION.
- Open-label design introduces performance bias — clinicians knew treatment allocation, which could influence post-randomization care (e.g., aggressiveness of ICU management, withdrawal of care decisions).
- Small sample size (n=203) limits power for subgroup analyses and increases susceptibility to chance baseline imbalances.
- ASPECTS scoring is subjective and has moderate inter-rater reliability — some patients with ASPECTS 3-5 may have been misclassified. DWI-MRI was allowed (more sensitive than CT), creating heterogeneity in core assessment method.
- Low IV tPA use (~27%) may not reflect international practice where more patients receive thrombolysis before thrombectomy, potentially affecting the generalizability of bleeding rates.
- The primary endpoint (mRS 0-3) is a more lenient threshold than mRS 0-2 used in most thrombectomy trials — mRS 3 (moderate disability, requires some help) is debatably a 'good' outcome for patients with large cores.
- Stopped early (203 of planned 200 enrolled, but interim analysis prompted early termination) — early stopping tends to overestimate treatment effects.
- No perfusion imaging required — many centers lacked CTP, meaning some patients with favorable penumbral patterns may have been included, inflating results compared to a purely CT-ASPECTS-selected population.
- Mortality was not reduced (18% vs 23.5%, p=0.33) — the benefit was in shifting survivors to less disability, not in saving lives. This raises ethical questions about expanding treatment to large cores.
Citation
N Engl J Med 2022;386:1303-13.