PDF: ECASS II
(1998)Objective
To evaluate the efficacy and safety of tPA (0.9mg/kg) given within 6 hours in acute ischemic stroke .
Study Summary
• ECASS II did not demonstrate a statistically significant benefit of alteplase over placebo for the primary outcome (mRS 0–1 at 90 days).<br>• A post-hoc analysis showed a significant benefit for mRS 0–2 (54.3% vs. 46.0%; p=0.024).
Intervention
tPA administered IV within 6 hours after the onset of stroke.
Inclusion Criteria
Adults (18-80 years) with moderate to severe ischaemic hemispheric stroke, no or only minor early signs of infarction on initial CT, treatment within 6 h of symptom onset.
Study Design
Arms: tPA group vs. placebo group.
Patients per Arm: Alteplase: 409, Placebo: 391
Outcome
• Favorable trend in primary outcome (90 day mRS 0-1): Alteplase 40.3% vs. placebo 36.6% (however it didn't reach statistical significance', p=0.277); • Post-hoc analysis of mRS (0-2): Alteplase 54.3% vs. placebo 46.0% (absolute difference 8.3%, p=0.024); • Mortality at 90 days: 10.6% in both groups (no significant difference); • Symptomatic intracranial hemorrhage (sICH): Alteplase 8.8% vs. placebo 3.4%.
Bottom Line
Alteplase did not significantly improve the predefined primary outcome (mRS 0–1), but a post hoc analysis showed a significant benefit when mRS 0–2 was considered. Symptomatic intracranial hemorrhage was more common with alteplase, but mortality was similar.
Major Points
- ECASS II was the largest European thrombolysis trial of its era (800 patients, 108 centers, 16 countries) — it failed to show benefit with the CORRECT tPA dose (0.9 mg/kg) within 6 hours, unlike the successful NINDS trial (within 3 hours) and the failed ECASS I (which used 1.1 mg/kg).
- Primary outcome (mRS 0–1 at 90 days) NOT significant: 40.3% alteplase vs 36.6% placebo (p=0.277). This was the predefined primary endpoint — ECASS II was a technically negative trial.
- Post hoc analysis (mRS 0–2 at 90 days) WAS significant: 54.3% vs 46.0% (p=0.024) — suggesting a real but modest benefit that the more stringent primary endpoint missed. This post hoc result helped motivate ECASS III.
- sICH significantly higher: 8.8% alteplase vs 3.4% placebo — double the hemorrhage rate. PH2 (large parenchymal hemorrhage, the clinically relevant type): 8.1% vs 0.8%. This highlighted the bleeding cost of extending treatment beyond 3 hours.
- Mortality identical: 10.5% vs 10.3% (p=0.816) — the excess sICH was offset by fewer large infarcts, creating a 'mortality wash' similar to NINDS. The net mortality balance suggested treatment was not futile despite the primary endpoint failure.
- Critical historical context: ECASS I (1995) had failed partly because the 1.1 mg/kg dose caused excessive hemorrhage. ECASS II corrected this to 0.9 mg/kg (the NINDS dose) but extended the window to 6 hours — the results showed that the dose correction worked but the time extension diluted efficacy.
- Used the more stringent mRS 0–1 as primary endpoint — unlike NINDS which used mRS 0–1 as one of multiple co-primary endpoints and also showed benefit on the Barthel Index and NIHSS. The choice of a single strict endpoint reduced power to detect benefit.
- NIHSS improvement at day 30 was statistically better with alteplase (p=0.035) — showing a consistent signal of neurological improvement across multiple outcome measures, even when the primary endpoint was missed.
- Together with NINDS (positive, 0–3h), ECASS I (negative, wrong dose), and ATLANTIS (negative, 3–5h), ECASS II defined the thrombolysis evidence landscape that led to the 3-hour treatment window as the initial standard of care.
- Directly motivated ECASS III (2008), which successfully demonstrated alteplase benefit in the 3–4.5 hour window using lessons learned from ECASS II — ultimately expanding the treatment window by 90 minutes and saving thousands of lives annually.
Study Design
- Study Type
- Randomized, double-blind, placebo-controlled trial
- Randomization
- Yes
- Blinding
- Double-blind
- Sample Size
- 800
- Follow-up
- 90 days
- Centers
- 108
- Countries
- Austria, Australia, Belgium, Denmark, Finland, France, Germany, Italy, Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, UK
Primary Outcome
Definition: Proportion with mRS 0–1 at 90 days
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| 36.6% | 40.3% | - (1.17 (0.9–1.6)) | 0.277 |
Limitations & Criticisms
- Technically a NEGATIVE trial — primary endpoint (mRS 0–1) not significant (p=0.277). The significant mRS 0–2 result was POST HOC and cannot carry the same evidentiary weight as a prespecified analysis.
- 6-hour treatment window was too wide — diluted the benefit seen in the 0–3h subgroup with the diminishing returns in the 3–6h window. ECASS III later confirmed that 3–4.5h was the correct extended window.
- sICH rate of 8.8% (vs 3.4% placebo) with PH2 at 8.1% vs 0.8% — one of the highest hemorrhage rates in the thrombolysis trial program, raising safety concerns about extending the treatment window.
- Choice of mRS 0–1 as the sole primary endpoint was overly stringent — mRS 0–2 (functional independence) may be more clinically meaningful and was the endpoint used successfully in ECASS III.
- Protocol violations: some patients had early CT changes >1/3 MCA territory — though fewer than in ECASS I, protocol violations still occurred and may have contributed to sICH.
- European-only enrollment — US and Canadian patients were not included, and the NINDS trial's different population may explain some of the discrepancy between trial results.
- No ordinal mRS analysis — if the full mRS distribution had been analyzed (as became standard later), the trial might have shown a significant treatment effect across all disability levels.
- High placebo response rate (36.6% achieving mRS 0–1) — reflects enrollment of milder strokes that recover spontaneously, reducing the detectable treatment difference.
- Underpowered for clinically important subgroups — the 0–3h stratum could not be definitively analyzed, leaving uncertain whether earlier treatment within the 6h window had clear benefit.
Citation
Lancet. 1998;352(9136):1245–1251