New oral anticoagulants in patients with atrial fibrillation (2)

摘要

Christian Ruff and colleagues' meta-analysis (March 15, p 955) raises several questions. When calculating the numbers of outcome events on the basis of event rates, sample size, and duration of follow-up, how were patients who discontinued the study prematurely considered? The rates of discontinuation in the new-oral-anticoagulant (NOAC)-investigating studies were 34% with edoxaban, 17-21% with dabigatran at 2 years, 22-24% with rivaroxaban and 25-28% with apixaban. Since rates of discontinuation were higher in patients randomised to dabigatran or rivaroxaban than to warfarin, patients treated with warfarin might have been exposed to anticoagulant therapy longer and thus might have had a higher frequency of cerebral haemorrhages.