Fig. 2
From: The hazards of chasing subgroups in neutral stroke trials
Effect of clomethiazole on Barthel index ≥ 60. Data were extracted from the CLASS and CLASS-I publications. [24, 27] The Mantel–Haenszel test in Cochrane REVMAN software was used with no adjustments. An odds ratio > 1 implies a good outcome. The overall effect of clomethiazole in CLASS was neutral. The post hoc interaction test between TACS and non-TACS was significant. The component of interest (TACS syndrome) was significant. The CLASS-I follow-up trial was neutral