Background: Patients undergoing hip or knee joint replacement are at risk for venous thromboembolic complications for up to twelve weeks postoperatively. We evaluated the efficacy and safety of a prolonged post-hospital regimen of enoxaparin, a low-molecular-weight heparin, in this patient population. Methods: Following elective total hip or knee replacement, 968 patients received subcutaneous enoxaparin (30 mg twice daily) for seven to ten days, and 873 were then randomized to receive three weeks of double-blind outpatient treatment with either enoxaparin (40 mg once daily) or a placebo. The primary efficacy end point was the prevalence of objectively confirmed venous thromboembolism or symptomatic pulmonary embolism during the double-blind phase of treatment. Results: Of the 873 randomized patients, 435 underwent elective total hip replacement and 438 underwent elective total knee replacement. Enoxaparin was superior to the placebo in reducing the prevalence of venous thromboembolism in patients treated with hip replacement: 8.0% (eighteen) of the 224 patients treated with enoxaparin had venous thromboembolism compared with 23.2% (forty-nine) of the 211 patients treated with the placebo (p < 0.001; odds ratio, 3.62; 95% confidence interval, 2.00 to 6.55; relative risk reduction, 65.5%). Enoxaparin had no significant benefit in the patients treated with knee replacement: thirty-eight (17.5%) of the 217 patients treated with enoxaparin had venous thromboembolism compared with forty-six (20.8%) of the 221 patients treated with the placebo (p = 0.380; odds ratio, 1.24; 95% confidence interval, 0.76 to 2.02; relative risk reduction, 15.9%). Symptomatic pulmonary embolism developed in three patients, one with a hip replacement and two with a knee replacement; all had received the placebo. There was no significant difference in the prevalence of hemorrhagic episodes or other types of toxicity between the enoxaparin and placebo-treated groups. Conclusions: Prolonging enoxaparin thromboprophylaxis following hip replacement for a total of four weeks provided therapeutic benefit, by reducing the prevalence of venous thromboembolism, without compromising safety. A similar benefit was not observed in patients treated with knee replacement.