A review based on 17 randomised studies on low molecular weight heparin (LMWH) versus unfractionated heparin (UFH) in the treatment of objectively verified deep venous thrombosis (DVT) is presented. Long-term treatment with LMWH was compared with long-term UFH in two studies and with warfarin in one study. In the rest of the studies LMWH and UFH were used during initiation of oral anticoagulant therapy, and these studies were included in a meta-analysis. The relative risk of progression of DVT during LMWH treatment compared with UFH was 0.63 (95% confidence interval: 0.39-1.00) and the relative risk of major bleeding was 0.41 (95% confidence interval: 0.24-0.70). There was no significant difference in the reduction of Marder score during treatment (LMWH 5.0 versus UFH 3.8) or in the frequency of new symptomatic, scintigraphically or angiographically verified pulmonary embolism (LMWH 0.6% versus UFH 1.1%). The frequency of complications seemed independent of whether LMWH was administered once or twice daily. Monitoring of LMWH treatment is not considered necessary but determination of anti-factor Xa in plasma is recommended if bleeding occurs during treatment with LMWH.