PDF: SKIP
(2021)Objective
To evaluate whether mechanical thrombectomy alone is noninferior to combined IV thrombolysis (0.6 mg/kg alteplase) plus thrombectomy in LVO stroke.
Study Summary
Intervention
Mechanical thrombectomy alone vs. IV alteplase (0.6 mg/kg) plus mechanical thrombectomy. IV tPA initiated within 4.5 hours; thrombectomy performed with approved devices.
Inclusion Criteria
Acute ischemic stroke with ICA or M1 occlusion, NIHSS ≥6, ASPECTS ≥6 on CT or ≥5 on DWI, age 18–85, eligible for IV alteplase (0.6 mg/kg) within 4.5 hours.
Study Design
Arms: Thrombectomy Alone vs. IV Alteplase + Thrombectomy
Patients per Arm: Thrombectomy: 101, Combined: 103
Outcome
Bottom Line
This trial failed to demonstrate the noninferiority of mechanical thrombectomy alone compared to combined IV alteplase plus mechanical thrombectomy for achieving a favorable functional outcome at 90 days. However, the confidence intervals were wide, and the results also did not establish the inferiority of thrombectomy alone. The rate of any intracerebral hemorrhage was significantly lower in the thrombectomy-alone group.
Major Points
- SKIP was the first of several 'direct-to-thrombectomy' trials testing whether IV thrombolysis can be safely omitted in patients going straight to mechanical thrombectomy for LVO stroke.
- Primary outcome (mRS 0-2 at 90 days): 59.4% thrombectomy-alone vs 57.3% combined — FAILED to demonstrate noninferiority (p=0.18 for noninferiority with margin of 0.74). The result was INCONCLUSIVE, not negative.
- Key safety signal: any ICH was significantly LOWER with thrombectomy alone (33.7% vs 50.5%, p=0.02), supporting the hypothesis that skipping tPA reduces hemorrhagic complications.
- Used the Japanese alteplase dose of 0.6 mg/kg (vs international 0.9 mg/kg) — this LOWER dose means the bleeding reduction from skipping tPA may be LESS dramatic than in settings using the full dose.
- The noninferiority margin (OR 0.74) was derived from trials using 0.9 mg/kg alteplase — applying this margin to a 0.6 mg/kg dose is methodologically questionable.
- Part of a wave of direct-to-thrombectomy trials: SKIP (Japan 2021, inconclusive), DIRECT-MT (China 2020, noninferior), MR CLEAN-NO IV (Netherlands 2021, noninferior), SWIFT DIRECT (Europe 2022, failed noninferiority).
- Both arms achieved excellent reperfusion rates (90-93% TICI ≥2b), reflecting high thrombectomy quality in Japanese centers. The high success rate may have masked any incremental benefit of pre-treatment tPA.
- Mortality was identical (7.9% vs 8.7%) and sICH was similar (5.9% vs 7.8%, p=0.78) — no safety concern with either approach.
- Current consensus: IV thrombolysis should NOT be withheld to expedite thrombectomy (AHA/ASA guidelines), but may be skipped in specific scenarios (e.g., direct presentation to thrombectomy-capable center with minimal delay).
- The 2023 AHA/ASA Acute Stroke guidelines gave a Class IIb recommendation that thrombectomy without IV tPA 'may be reasonable' in select patients at thrombectomy-capable centers.
Study Design
- Study Type
- Investigator-initiated, multicenter, randomized, open-label, noninferiority clinical trial.
- Randomization
- Yes
- Blinding
- Open-label for the intervention, but the 90-day modified Rankin Scale score was assessed by personnel blinded to treatment group assignment.
- Sample Size
- 204
- Follow-up
- 90 days.
- Centers
- 23
- Countries
- Japan
Primary Outcome
Definition: Favorable outcome, defined as a modified Rankin Scale (mRS) score of 0 to 2 at 90 days.
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| 57.3% (59/103) | 59.4% (60/101) | 1.09 (0.63 to ∞ (1-sided 97.5% CI)) | 0.18 for noninferiority |
Limitations & Criticisms
- Open-label design — knowledge of tPA administration could influence post-randomization decisions (e.g., aggressiveness of BP management, timing of antiplatelet initiation).
- Noninferiority margin of 0.74 was borrowed from trials using 0.9 mg/kg alteplase — applying this to 0.6 mg/kg dose is methodologically problematic, as the treatment effect of lower-dose tPA may be smaller.
- Inconclusive result (neither noninferior nor inferior) is the worst possible outcome for clinical practice — provides no actionable guidance.
- Small sample size (n=204) with wide confidence intervals — the study was likely underpowered for the true effect size in a population with high baseline good outcomes.
- Conducted exclusively in Japan with Japanese alteplase dose (0.6 mg/kg) — results cannot be directly applied to international practice using 0.9 mg/kg dose.
- Higher-than-expected favorable outcomes in both arms (~58-59%) — baseline event rate assumptions were wrong, further reducing statistical power.
- Sex imbalance between groups (70% male in control vs 55% in thrombectomy-alone) despite randomization — this could confound results given sex-based differences in stroke outcomes.
- Time from onset to groin puncture was NOT significantly shortened by skipping tPA — undermines the key rationale that direct thrombectomy saves time.
- No assessment of clot characteristics or clot burden — some clots may be more amenable to pre-treatment with tPA (e.g., fresh red clots vs organized white clots).
Citation
JAMA. 2021;325(3):244-253.