Background and purpose: We performed a combination of prespecified and exploratory subgroup analyses to detect any treatment differences among various baseline subgroups in the Warfarin-Aspirin Recurrent Stroke Study (WARSS) cohort.
Methods: The WARSS compared the efficacy of adjusted-dose warfarin (INR 1.4-2.8) to aspirin (325 mg/day) for recurrent ischemic stroke or death within 2 years. The effect of warfarin and aspirin was compared among prespecified and exploratory subgroups with respect to sociodemographic and vascular risk factors, stroke subtype, arterial territory, and infarct topography. Hazard ratios and 95% confidence intervals comparing warfarin to aspirin were calculated using Cox proportional hazards models. Differences in hazard ratios were tested using interaction terms.
Results: No treatment differences between warfarin and aspirin were found across multiple prespecified subgroups. In a multivariate model, warfarin was associated with greater hazard among patients with moderate stroke severity (HR 1.63, 95% CI 1.005-2.64, p = 0.047) and a greater benefit among those with posterior circulation location without brainstem infarction (HR 0.54, 95% CI 0.33-0.88, p = 0.013). In post-hoc analyses of the cryptogenic subgroup, warfarin was associated with worse outcomes among patients with moderate stroke severity and better outcomes among those without baseline hypertension or with posterior circulation infarcts sparing the brainstem.
Conclusions: In the WARSS, the majority of subgroup analyses showed no benefit of warfarin over aspirin. Warfarin benefit was limited to brainstem-sparing posterior circulation infarcts and select cryptogenic stroke subgroups. Pending future clinical trial evidence to the contrary, antiplatelets are recommended for survivors of noncardioembolic stroke.
Copyright (c) 2006 S. Karger AG, Basel.