The study included patients receiving hospitalization in Zadar General Hospital from September 1^st 2016 until August 31^st 2017 with diagnoses (MKB-10) I63 and K92 and atrial fibrillation (FA) at the time of inoculation. In the observed period of the year, 378 patients with cerebrovascular insult (CV) and 63 patients with gastrointestinal (GI) bleeding were hospitalized. The study included 77 patients with CVI. The first group consisted of 3 (4%) of the patients who had some of the new oral anticoagulants (NOAC) therapy in the FA therapy. The second group consisted of 22 (29%) of the patients who had warfarin therapy, the third group consisted of 15 (19%) of patients receiving acetylsalicylic acid (ASA) while the fourth group consisted of 37 (48%) patients who did not have any of the aforementioned drugs in the therapy. Also, in the observed period of the year, 63 patients with GI bleeding were hospitalized. The study included 10 patients with this diagnosis. Two (20%) patients who did not have FA at the time of admission were FA. 5 (50%) patients had NOAC and 3 (30%) warfarin. No patient from the FA was hospitalized to have ASA therapy. In the group of 22 patients who had warfarin in therapy, well-regulated anticoagulation therapy with INR values ranging from 2 to 3.5 had 4 patients (18%). As it is known that in the Zadar County in the observed period, 2/3 of the patients taking anticoagulant therapy took warfarin and 1/3 of NOAC, and the published data show that the ratio of hospitalized patients with FA and CVI was 1: 8 compared to those who took NOAC against warfarin, and from the literature data it is known that their efficacy in preventing CVI alike, suggests that warfarin-treated patients were not optimally anticoagulated so they could not avoid the occurrence of thromboembolic CVI. At the same time, in hospitalized patients with GI bleeding, there were more those taking NOAC in therapy, which is consistent with the literature data, but also the known data on the nonoptimal INR regulation in patients taking warfarin. It can be concluded that, in conditions of non-optimal regulation of INR, patients with chronic FA do not need to use warfarin, but optimal CVI prevention will have NOAC, with an acceptable risk of gastrointestinal bleeding. (1)