In view of the potential of low-molecular-weight heparins (LMWH) to simplify initial therapy and allow outpatient treatment of proximal deep-vein thrombosis, we undertook a randomised comparison of fixed-dose subcutaneous LMWH with adjusted-dose intravenous standard heparin in the initial treatment of this disorder. Our main objectives were to compare the efficacy of these regimens for 6 months of follow-up and to assess the risk of clinically important bleeding. Of 170 consecutive symptomatic patients with venographically proven proximal deep-venous thrombosis, 85 received standard heparin (to achieve an activated partial thromboplastin time of 1.5 to 2.0 times the pretreatment value) and 85 LMWH (adjusted only for body weight) for 10 days. Oral coumarin was started on day 7 and continued for at least 3 months. The frequency of recurrent venous thromboembolism diagnosed objectively did not differ significantly between the standard-heparin and LMWH groups (12 [14%] vs 6 [7%]; difference 7% [95% confidence interval -3% to 15%]; p = 0.13). Clinically important bleeding was infrequent in both groups (3.5% for standard heparin vs 1.1% for LMWH; p greater than 0.2). We conclude that fixed-dose subcutaneous LMWH is at least as effective and safe as intravenous adjusted-dose heparin in the initial treatment of symptomatic proximal-vein thrombosis. Since there is no need for laboratory monitoring with the LMWH regimen, patients with venous thrombosis can be treated at home.