We compared the efficacy and safety of low-molecular-weight heparin with that of low-dose unfractionated heparin in the prevention of venous thromboembolism after total knee arthroplasty in a prospective, randomized, multicenter trial. One hundred and eighty-five patients were randomly assigned to two groups: ninety-two received low-molecular-weight heparin (forty milligrams of enoxaparin the evening before the operation and once a day subsequently) and ninety-three received unfractionated heparin (5000 international units the evening before the operation and three times a day thereafter). The prophylaxis was continued until bilateral ascending venography was performed six to nine days after the operation or, if venography was not done, until the eighth postoperative day. Venography revealed a prevalence of deep-vein thrombosis of 27 per cent (twenty-five of ninety-three patients) in the group that received unfractionated heparin and 23 per cent (twenty-one of ninety-two patients) in the group that received low-molecular-weight heparin. The difference was not significant (p = 0.6). Five patients (5 per cent) who received unfractionated heparin and 3 patients (3 per cent) who received low-molecular-weight heparin had a deep-vein thrombosis in the proximal veins. Two patients who received unfractionated heparin and one who received low-molecular-weight heparin had clinical symptoms suggestive of a pulmonary embolism. None of these three patients had a positive ventilation-perfusion scan. There were no deaths, major bleeding episodes, or wound hematomas necessitating operative intervention or discontinuation of the anticoagulation in the series.(ABSTRACT TRUNCATED AT 250 WORDS)