In order to evaluate the efficiency of classical anticoagulant therapy for venous thromboembolic disease in cancer patients, we retrospectively analysed 71 patients treated with intravenous heparin first and then with antivitamin K. After a mean follow-up of 185 +/- 25 days, 23 patients (33%) were dead; nine patients (12%) had suffered from major haemorrhagic complications, which were not fatal, four of which were due to heparin overdosage; 17 patients (24%) showed recurrent venous thromboembolic disease. According to univariate statistical analysis, risk of major bleeding was not associated with the presence of either thrombocytopenia, abnormal blood coagulation, metastases and/or any other hemorrhagic risk factors; recurrence of venous thromboembolic disease was not associated with the presence of other risk factors for venous thromboembolic disease, nor with the presence or absence of metastases and/or of ongoing chemotherapy. Such results suggest that classical anticoagulant therapy for venous thromboembolic disease in cancer patients is neither effective nor safe. The present retrospective study underlines needs for further prospective analyses in order to evaluate potential benefit from other therapeutic strategies, such as use of low molecular weight heparins and/or vena cava filter placement.