BACKGROUND A direct comparison of the efficacy and safety profiles of left atrial appendage occlusion (LAAO) devices and novel oral anticoagulants (NOACs) for stroke prevention in atrial fibrillation is warranted but currently unavailable. OBJECTIVE The aim of this study was to compare the >1-year efficacy and safety of LAAO devices and NOACs for stroke prevention in patients with atrial fibrillation. METHODS We performed a systematic review on randomized controlled trials (RCTs) and observational studies. RCTs were analyzed by means of a network meta-analysis method using warfarin as a bridge to compare LAAO to individual NOAC or all NOACs as a whole. Observational studies were analyzed with the meta-proportion function where pooled event rates were compared. RESULTS A total of 6 RCTs and 27 observational studies were included. A network meta-analysis of RCTs indicated that LAAO was less effective than NOACs for stroke prevention (odds ratio 0.86), but had a lower rate of hemorrhagic events during follow-up. However, a meta-proportion analysis of observational studies revealed that LAAO devices were associated with a lower rate of both thromboembolic events (1.8 events per 100 patient-years vs 2.4 events per 100 patient years) and major bleeding events during follow-up (2.2 events per 100 patient-years vs 2.5 events per 100 patient-years) as compared with NOACs. With prolonged follow-up duration after LAAO implantation, the rate of thromboembolic events decreased (2.1, 1.8, and 1.0 events per 100 person-years for 1, 1-2, and >2 years, respectively). CONCLUSION Although superiority of LAAO over NOACs was not demonstrated by RCTs in terms of stroke prevention, LAAO was found to be consistently associated with a lower rate of both thromboembolic and hemorrhagic events as compared with NOACs in observational studies.