It is difficult to draw conclusions from published meta-analysis on the treatment of proximal deep vein thrombosis by low molecular weight heparins in view of methodological problems related to selection of the studies and the addition of early and late complications. To determine the efficacy and safety of subcutaneous low molecular weight heparins in the initial treatment of proximal deep vein thrombosis we reviewed all published randomised controlled studies comparing this treatment to the standard intravenous unfractionated heparin regimen. Adverse events were taken into account up to 48 hours after cessation of heparin treatment. Results point towards equivalent or greater efficacy, with reduction in thrombosis size, and safety of low molecular weight heparin. The mortality rate was low (0-1.2%). Incidence of recurrent venous thromboembolism (0-2%) or major bleeding (0-2.4%) was also low, though such events were linked to a high mortality rate (9-16%). Thrombocytopenia occurred in 0 to 2.5% of cases.