To assess the efficacy and safety of dermatan sulphate (MF 701) in preventing postoperative deep vein thrombosis (DVT), 324 patients aged 40 years or over undergoing elective major general surgical operations were included in a randomized trial comparing MF 701 (100 mg intramuscularly once a day) with unfractionated calcium heparin (UFH, 5000 units subcutaneously three times daily). Both treatments were initiated before operation and continued until discharge. In all, 316 patients were included in the analysis (MF 701, 157; UFH, 159). Serial impedance plethysmography was performed in all patients; a 125I-radiolabelled fibrinogen uptake test was added to impedance plethysmography in a randomized subsample of 62 patients (MF 701, 28; UFH, 34). Positivity in either test was confirmed where possible by venography. DVT was diagnosed by venography or, when this could not be performed, by positivity of either impedance plethysmography or fibrinogen uptake test. The incidence of DVT was 3.1 per cent (patients receiving MF 701) and 1.6 per cent (those receiving UFH) in patients undergoing impedance plethysmography alone, and 7.1 and 11.8 per cent, respectively, in those undergoing both impedance plethysmography and fibrinogen uptake test; in neither case was the difference between treatments statistically significant. There were five in-hospital deaths, two in patients receiving MF 701 and three in patients on UFH. The incidence of clinically overt haemorrhage was 5.7 per cent in patients on MF 701 and 17.6 per cent in those on UFH (P less than 0.01). Postoperative transfusions and reoperations due to bleeding were significantly less frequent in patients receiving MF 701. Mortality rates at 3 months were similar for the two treatment groups. Compared with standard prophylaxis using UFH, MF 701 showed a similar efficacy with a significantly greater safety.