The effectiveness and safety of warfarin were compared with those of a low-molecular-weight heparin (dalteparin) for the prevention of deep-vein thrombosis after total hip arthroplasty in a prospective, randomized, multi-institutional trial. Patients who were older than eighteen years of age and were scheduled to have an elective primary or revision total hip arthroplasty were eligible; 580 patients were randomized, 550 had the operation and received prophylaxis, and 382 had evaluable venograms. Prophylaxis was provided either with warfarin beginning the night before the operation or with dalteparin beginning two hours before the operation and was continued until venography was performed. Bleeding was assessed on the basis of intraoperative blood loss, transfusion requirements, a decrease in hematocrit, and clinically identified bleeding complications. The prevalence of deep-vein thrombosis was found to be significantly lower in the patients who had received dalteparin than in those who had received warfarin (twenty-eight [15 per cent] of 192 patients compared with forty-nine [26 per cent] of 190 patients; p = 0.006). Deep-vein thrombosis occurred in the calf veins of twenty-one patients (11 per cent) who had received dalteparin and of forty-three patients (23 per cent) who had received warfarin; this difference was significant (p = 0.003). Proximal deep-vein thrombosis occurred in ten patients (5 per cent) who had received dalteparin and in sixteen patients (8 per cent) who had received warfarin; however, with the numbers available, no significant difference could be detected (p = 0.185). We also could not detect a significant difference with regard to the intraoperative and postoperative blood loss, the decrease in hematocrit, and the prevalence of major bleeding complications between the two groups; however, the patients who had received dalteparin had a significantly higher prevalence of bleeding complications involving the operative site (p = 0.03), and a significantly greater percentage required postoperative transfusions (p = 0.001). We concluded that preoperative prophylaxis with dalteparin is significantly more effective than that with warfarin in preventing deep-vein thrombosis after total hip arthroplasty. The greater effectiveness of dalteparin must be considered, however, in light of an increased need for postoperative transfusions and an increase in the prevalence of wound-related bleeding complications.