In this paper we demonstrate that among four phase III trials comparing warfarin and non-vitamin K antagonist oral anticoagulants (NOACs) vs warfarin in patients with non-valvular atrial fibrillation there was heterogeneity in the thromboembolic and bleeding risk of the included populations, variability in the definitions of primary endpoints, especially for safety, and different criteria for reducing NOAC dosage according to renal function. Therefore, findings from these four studies are not directly comparable and it is incorrect to conduct comparative evaluations between various NOACs. In each patient the antithrombotic strategy should be tailored taking into account NOAC characteristics, the individual thromboembolic and bleeding risk, presence of comorbidities, mental status and socio-economic conditions.