Background: Although heart rate (HR) is reportedly associated with major cardiovascular outcomes in the general population, its impact on adverse events in patients with non-valvular atrial fibrillation (NVAF) remains controversial. Thus, we performed post hoc analyses of data from the J-RHYTHM Registry to clarify this in patients with NVAF. Results: Of 7406 outpatients with NVAF from 158 institutions, 6886 (age, 69.8 ± 9.9 years; men, 70.8%), in whom both baseline HR and HR-end (at the time closest to an event or at the last visit of follow-up) were measured during the two-year follow-up period or until the occurrence of an event, constituted the study group. The baseline HR and HR-end values were 72.5 ± 13.3 bpm and 73.3 ± 13.3 bpm, respectively. Thromboembolism, major hemorrhage, all-cause death, and cardiovascular death occurred in 117 (1.7%), 130 (1.9%), 157 (2.3%), and 58 (0.8%) patients, respectively. Baseline HR was not associated with any adverse event, whereas HR-end (per 1-bpm increase) was significantly associated with an increased incidence of all adverse events. Furthermore, the highest quartile of HR-end (≥80 bpm) was independently associated with the incidence of major hemorrhage (adjusted odds ratio [OR], 2.90; 95% confidence interval [CI], 1.69–4.96; P