Objective:To investigate the current status of anticoagulant therapy and the incidence of ischemic and bleeding events in hospitalized patients aged 90 years and over with non-valvular atrial fibrillation(NVAF).Methods:We retrospectively collected clinical data, antithrombotic treatment strategies, in-hospital ischemic stroke and bleeding events from NVAF patients(≥90 years)who were admitted to our hospital from June 2014 to August 2018.Based on the CHA 2DS 2-VASc score(2, 3, and ≥4 respectively), patients were divided into three ischemic risk groups, and antithrombotic treatment strategies and in-hospital ischemic stroke events were compared between the three groups.Alternatively, patients were divided into the high bleeding risk group(HAS-BLED score ≥3, )and the non-high bleeding risk group(HAS-BLED score ≤2), and antithrombic treatment strategies and the major bleeding events were compared between the two groups. Results:Among the 223 hospitalized NVAF patients aged 90 years and over, 42.6% of them received anticoagulant treatment, 25.6% received antiplatelet drugs, and 31.8% received non-antithrombotic treatment.With the increase of the CHA 2DS 2-VASc score, there was a trend of declined rates of non-antithrombotic treatment among the three ischemic risk groups(47.4%, 42.9%, 26.4%, P=0.06), and the rates of in-hospital ischemic stroke were similar among groups(10.5%, 12.2%, 15.5%, P=0.75). Moreover, compared with the non-high bleeding risk group, patients in the high bleeding risk group more frequently received anticoagulant treatment(47.2% vs.38.3%)and less frequently received non-antithrombotic therapy(28.7% vs.34.7%). There was no significant difference in antithrombotic treatment strategies( P=0.39)or rate of in-hospital major bleeding events(13.0% vs.10.2%, P=0.51). However, the rate of in-hospital major bleeding events was significantly higher in those with concurrent infections(16.8% vs.6.4%, P=0.02)or respiratory failure(21.3% vs.8.0%, P=0.01). Conclusions:The rate of anticoagulant use in NVAF patients aged 90 years and over is too low during hospitalization, and anticoagulant therapy should be standardized.In addition to the HAS-BLED score, we should consider the complications that increase the bleeding risk, such as infections and respiratory failure, when evaluating the bleeding risk.