Relation of stroke onset to start of treatment (OTT) with treatment effect after adjustment for prognostic variables assessed by A) day 90 modified Rankin score 0-1 versus 2-6 (interaction p=0.0269, n=3530 [excluding EPITHET data p=0.0116, n=3431]); B) global test that incorporates modified Rankin score 0-1 versus 2-6, Barthel Index score 95-100 versus 90 or lower and NIHSS score 0-1 versus 2 or more (interaction p=0.0111, n=3535 [excluding EPITHET data p=0.0049, n=3436]); C) mortality (interaction p=0.0444, n=3530 [excluding EPITHET data p=0.0582, n=3431]); and D) parenchymal hemorrhage type 2 (interaction p=0.4140, n=3531 [excluding EPITHET data p=0.4578, n=3431]). Thus, for parenchymal hemorrhage type 2, the fitted line is not statistically distinguishable from a horizontal line. For each graph, the adjusted odds ratio is shown with the 95% CIs. CIs from the models will differ from those shown in the tables because the model uses data from all patients treated within 0-360 min whereas the categorized analyses in the tables are based on subsets of patients: the modeled CIs are deemed to be more reliable.
%: percent.
Lees, KR, Bluhmki, E, von Kummer, R, et al. Time to treatment with intravenous alteplase and outcome in stroke: an updated pooled analysis of ECASS, ATLANTIS, NINDS, and EPITHET trials. Lancet 2010; 375:1695. Illustration used with permission of Elsevier Inc. All rights reserved.