In a randomized, prospective, double-blind trial, the effect of conventional low-dose heparin (5,000 units every 8 hours) was compared with that of a low molecular weight fragment (2165 LMWH, Kabi Vitrum AB, starting 2,500 units 2 hr before surgery, and then 5,000 units subcutaneously every morning for 9 days). A total of 90 patients admitted because of hip fracture fulfilled the inclusion criteria and were analyzed for development of pulmonary embolism and deep vein thrombosis: 46 patients were included in the low molecular weight heparin (LMWH) group, and 44 in the conventional heparin group. Two and three, respectively, died before diagnostic tests were performed. In the remaining patients a ventilation-perfusion lung scan was performed 8 days after intervention. In the first 57 patients studied a bilateral ascending venography was performed on the ninth day only if clinical symptoms suggested a deep venous thrombosis. Because of the rate of venous thrombosis detected in those patients was unexpectedly low, venography was requested in the remaining 33 patients, even if the screening tests were negative. Pulmonary embolism occurred in six patients, all in the LMWH group. Deep vein thrombosis occurred in 14 patients in the LMWH group and in six patients in the conventional heparin group. Both differences are statistically significant. Mortality did not differ between the groups, nor did haemorrhagic complications. Our findings suggest that, in patients with hip fracture, LMWH is not useful at the dosage used.