• Population: Adult patients age 18–80 years with at least 30 minutes of AIS symptoms presenting between 3–4.5 hours after onset of symptoms with no significant improvement.
  • Main Exclusion: Multiple exclusions were listed in the manuscript.
• Intervention: tPA 0.9 mg/kg;initial 10 percent bolus, remainder given over 60 min.
• Comparison: Placebo.
• Outcomes:
  • Primary: mRS score 0–1 (favorable) versus 2–6 (unfavorable) at 90 days.
  • Secondary: Global outcome measure combining 90-day outcomes of mRS 0–1, ≥95 Barthel Index, NIHSS score 0–1, score of 1 Glasgow Outcome Scale; mortality at 90 days; any ICH, sICH, symptomatic edema, and other serious adverse events.

Authors’ Conclusions

“Reanalysis of the ECASS III trial data with multiple approaches adjusting for baseline imbalances does not support any significant benefits and continues to support harms for the use of alteplase 3–4.5 hours after stroke onset.”

Key Results

ECASS-III included 821 patients with a mean age of 65 years and 60 percent male. After adjusting for baseline imbalances, multiple methods failed to find statistically significant benefits with thrombolysis given 3–4.5 hours after stroke onset and confirmed the significant increase in harm.

Evidence-Based Medicine Commentary

1. Inter-rater reliability (IRR): The outcome assessment used mRS. The IRR for mRS is moderate at best.^5,6 A clinical trial has internal validity only if imbalances between groups and bias in the assessment of outcome and chance have been excluded as possible explanations for the observed difference in outcomes.

2. Fragility index (FI): The FI is another way to represent the data, and it’s statistically reproducible.⁷ FI is the minimum number of patients who would need to have a different outcome to change the P value from <0.05 to >0.05, although the 0.05 threshold as a measure of statistical significance has its own problems.⁸ A low FI means only a small number of patients would need to have their outcome change for the trial to lose statistical significance. The FI of the original ECASS-III data is 1, meaning only one patient would need to have a different outcome to change the result. This is consistent with the reanalysis study by Alper et al, which found no significant benefit for tPA.

Other data for this time window support the fragility of ECASS-III data. IST-3 was the largest RCT investigating tPA for AIS in treated patients up to 6 hours.9 It didn’t show a benefit for its primary outcome. The pre-specified 3–4.5-hour subgroup was around double that of ECASS-III (n=1,177 versus n=821). IST-3 had a significant decrease in good neurologic outcome in patients randomized to tPA (32 percent) versus placebo (38 percent).

3. Baseline imbalances: A strong predictor of stroke outcome is severity of symptoms at presentation. There was an important baseline imbalance in stroke between the two groups in the ECASS-III trial. Those randomized to placebo had a worse median and mean baseline NIHSS score.

Another difference between the two groups was that double the number of patients with a history of a previous stroke (7.7 percent tPA versus 14.1 percent placebo; P = 0.003) appeared in the placebo arm. Recurrent strokes have a worse outcome than first strokes. The statistical difference in outcome favoring tPA over placebo could be explained by the baseline imbalance.