Introduction: The American College of Physicians guidelines for venous thromboembolism (VTE) prophylaxis stresses bleeding risk assessment before VTE prophylaxis. Rivaroxaban has recently been accepted for VTE prevention. However most of the publications heavily focus on the thrombosis rate. Hypothesis: The rate of major bleeding in patients receiving Rivaroxaban is higher than those receiving LMWH for VTE prophylaxis. Methods: MEDLINE, EMBASE, Cochrane Register of Randomized Controlled Trials as well as abstracts presented at national meetings [AHA, ACC, ASH]. We included all prospective experimental trials which declared and counted major bleeding criteria and utilized Rivaroxaban in primary VTE prophylaxis, published from 1-1-06 to 12-21-11. The data was abstracted by 2 independent reviewers in accordance with PRISMA guidelines. We used R [R Meta package v 0.8-2] for analysis. The OR estimates were done with Mantel-Hanzel fixed estimates method and with random effect model by DerSimonian and Liard. Results: We identified 335 records, screened 15 and 9 trials (n=21,035) were included in the analysis. There were 3 knee replacement, 5 hip replacement, and 1 medically ill prophylaxis trials. All compared to LMWH prophylaxis. Although no heterogeneity was detected [I2=0%, p=0.9], we present a conservative random effect model. The odds of major bleed were 1.98 [95%CI 1.3-3.0] times higher for Rivaroxaban compared to LMWH. Individual omission of each of the publications had no significant effect on the overall significance of the findings. When only those manuscripts using a Rivaroxaban dose of 10 mg daily were analyzed, the major bleeding OR was 1.98 [95%CI 1.3-3.1]. Conclusion: Rivaroxaban is associated with a higher incidence of major bleeding compared to LMWH when used for VTE prophylaxis. This information will be instrumental in calculating risks and benefits of patients who need VTE prophylaxis.