OBJECTIVES: Venous thromboembolism (VTE) is one of the major complications after major orthopedic surgeries (MOS). In 2011, FDA approved rivaroxaban for VTE prevention among patients undergoing MOS. The aim of our study is to empirically evaluate the comparative effectiveness of rivaroxaban, warfarin, and low molecular weight heparins (LMWHs) for VTE prevention among MOS patients using  “real world” data. METHODS: A cohort study using IMS Lifelink Plus (2006-2013) data compared the risk of VTE and major bleed events among MOS patients exposed to rivaroxaban, warfarin, LMWHs, or fondaparinux with those who are not anticoagulated within 7 days after their MOS-hospital discharge. Kaplan Meier curves and Cox proportional hazard models were used to assess the risk of VTE and major bleed events and to adjust for potential confounders. RESULTS: A cohort of 35,279 MOS were included which provided 68,340 person years of follow up including 1,004 rivaroxaban, 7,339 warfarin, 5,692 LMWH, 841 fondaparinux exposed patients and 20,403  patients who did not receive an  initial anticoagulant. Risk of VTE was lowest for rivaroxaban (H.R.=0.282; 95%CI:0.156-0.510) followed LMWHs (H.R.=0.671 [95%CI=0.582-0.773]), fondaparinux (H.R.=0.680 [95%CI =0.485-0.951]) and warfarin (H.R.=0.872 [95%CI=0.778-0.978]) when compared to no anticoagulant use in unadjusted cox models.  After adjusting for potential confounders, only rivaroxaban (H.R.=0.395 [95%CI=0.215-0.742]) and LMWHs (H.R.=0.755[95%CI=0.643-0.873]) significantly reduced the risk of VTE. However, these results were not significant in a sensitivity analysis using a more strict definition to detect VTEs in claims data. There were no bleed events for rivaroxaban users and the risk of bleed events were not significantly different among anticoagulants and non-anticoagulant exposure in both the adjusted and the unadjusted models.