Summary Background Prophylaxis for venous thromboembolism is recommended for at least 10 days after total knee arthroplasty; oral regimens could enable shorter hospital stays. We aimed to test the efficacy and safety of oral rivaroxaban for the prevention of venous thromboembolism after total knee arthroplasty. Methods In a randomised, double-blind, phase III study, 3148 patients undergoing knee arthroplasty received either oral rivaroxaban 10 mg once daily, beginning 6–8 h after surgery, or subcutaneous enoxaparin 30 mg every 12 h, starting 12–24 h after surgery. Patients had mandatory bilateral venography between days 11 and 15. The primary efficacy outcome was the composite of any deep-vein thrombosis, non-fatal pulmonary embolism, or death from any cause up to day 17 after surgery. Efficacy was assessed as non-inferiority of rivaroxaban compared with enoxaparin in the per-protocol population (absolute non-inferiority limit −4%); if non-inferiority was shown, we assessed whether rivaroxaban had superior efficacy in the modified intention-to-treat population. The primary safety outcome was major bleeding. This trial is registered with ClinicalTrials.gov , number NCT00362232. Findings The primary efficacy outcome occurred in 67 (6·9%) of 965 patients given rivaroxaban and in 97 (10·1%) of 959 given enoxaparin (absolute risk reduction 3·19%, 95% CI 0·71–5·67; p=0·0118). Ten (0·7%) of 1526 patients given rivaroxaban and four (0·3%) of 1508 given enoxaparin had major bleeding (p=0·1096). Interpretation Oral rivaroxaban 10 mg once daily for 10–14 days was significantly superior to subcutaneous enoxaparin 30 mg given every 12 h for the prevention of venous thromboembolism after total knee arthroplasty. Funding Bayer Schering Pharma AG, Johnson & Johnson Pharmaceutical Research & Development.