Eligible randomised controlled trials (RCTs) compared warfarin with an alternative antithrombotic regimen for the primary or secondary prevention of stroke. At least 400 patients with non-valvular atrial fibrillation were included in the warfarin arm. The primary outcome was the occurrence of ischaemic or hemorrhagic stroke or non–central nervous system embolism.
Across the warfarin arms of the trials, the proportion of male patients ranged from 55% to 70%, the mean age from 70 to 81 years, the proportion with hypertension from 53% to 91% and, where reported, the proportion with diabetes from 14% to 40%. Where reported, the mix of type of atrial fibrillation (paroxysmal, persistent or permanent) varied across trials. Time within therapeutic range for the warfarin treated patients ranged from 55% to 68%. The comparator antithrombotic regimens varied across trials; only two trials used the same comparator (ximelagatran). Mean/median follow-up ranged from 11 to 32 months.
One reviewer screened titles and abstracts; it was unclear how many reviewers screened full papers.