Efficacy and safety of tissue plasminogen activator 3 to 4.5 hours after acute ischemic stroke: a metaanalysis

Abstract

Background and purpose: The Third European Cooperative Acute Stroke Study (ECASS-3) demonstrated a benefit of treatment with intravenous tissue plasminogen activator (tPA) for acute stroke in the 3- to 4.5-hour time-window. Prior studies, however, have failed to demonstrate a significant benefit of tPA for patients treated beyond 3 hours. The purpose of this study was to produce reliable and precise estimates of the treatment effect of tPA by pooling data from all relevant studies.

Methods: A metaanalysis was undertaken to determine the efficacy of tPA in the 3- to 4.5-hour time-window. The effect of tPA on favorable outcome and mortality was assessed.

Results: The metaanalysis included data from patients treated in the 3- to 4.5-hour time-window in ECASS-1 (n=234), ECASS-2 (n=265), ECASS-3 (n=821) and The Alteplase Thrombolysis for Acute Noninterventional Therapy in Ischemic Stroke (ATLANTIS) (n=302). tPA treatment was associated with an increased chance of favorable outcome (odds ratio 1.31; 95% CI: 1.10 to 1.56; P=0.002) and no significant difference in mortality (odds ratio 1.04; 95% CI: 0.75 to 1.43; P=0.83) compared to placebo treated patients.

Conclusions: Treatment with tPA in the 3- to 4.5-hour time-window is beneficial. It results in an increased rate of favorable outcome without adversely affecting mortality.

Comparison of tPA versus placebo in the 3-4.5 hour time-window after acute ischemic stroke. Three 90-day endpoints were assessed: favorable outcome using a global outcome measure (panel A), favorable outcome defined as a modified Rankin Scale Score of 0-1 (panel B), and mortality (panel C). *The global outcome measure is a global odds ratio derived from a statistical model that takes into account the outcome on three individual outcome scales at day 90: mRS 0-1, NIHSS 0-1, and Barthel Index>=95. Number of event ratios are not listed for this outcome measure because these cannot be derived from the model.