RESUMEN
BACKGROUND AND PURPOSE: In addition to lowering stroke risk, warfarin use is also associated with reduced stroke severity in patients with atrial fibrillation and acute ischaemic stroke. It was sought to determine whether the effect of non-vitamin-K oral anticoagulants (NOACs), compared to warfarin, differed by stroke severity. METHODS: Phase III randomized controlled trials with participants who were randomized to receive NOACs or warfarin for stroke prevention in the setting of non-valvular atrial fibrillation were identified. Stroke was classified into two categories, fatal or disabling stroke and non-disabling stroke, and meta-analyses were completed for both outcomes and for comparative case fatality of stroke amongst trials. RESULTS: Five randomized controlled trials met our inclusion criteria. In clinical trials evaluating the NOACs usually prescribed in clinical practice (four trials), acute stroke was reported in 1403 (1.86%) participants, 787 (1.04%) in the NOAC group [386 (0.51%) fatal or disabling, 401 (0.53%) non-disabling] and 616 (0.82%) in the warfarin group [367 (0.49%) fatal or disabling, 249 (0.33%) non-disabling]. On meta-analysis NOACs were significantly superior to warfarin for fatal or disabling stroke (odds ratio [OR] 0.77; 95% confidence interval [CI] 0.66-0.89, I2 = 21%) and non-disabling stroke (OR 0.85; 95% CI 0.73-0.98, I2 = 2%). The case fatality of stroke was no different between groups (OR 0.90, 95% CI 0.75-1.13, I2 = 0%), but the point estimate favoured NOACs. CONCLUSION: In phase III trials of NOACs, for prevention of stroke in atrial fibrillation, NOACs are associated with a lower risk of both fatal/disabling and non-disabling stroke compared to warfarin.