Low molecular weight heparin (LMWH), unfractionated heparin (UFH) and warfarin were compared with respect to efficacy and safety in the prevention of thrombo-embolism in general surgery. Meta-analysis (MA) with a priori definition of the MA protocol was used to combine the results from randomised trials with patients who underwent general surgery and deep-vein thrombosis (DVT) prophylaxis with LMWH, UFH or warfarin. Forty-four studies were identified for assessment and 33 were included, however, none for warfarin. For efficacy (DVT and pulmonary embolism) and major bleeding, no significant difference between the LMWH- and UFH-treated groups was demonstrated. The relative risk of minor bleedings for LMWH versus UFH was 0.75 (0.64-0.88; 95% confidence interval) and is significant (p < 0.05). Within the limitations of the MA, LMWH and UFH did not differ significantly in terms of prevention of thrombo-embolism, but LMWH had a significantly better safety profile. On this basis, LMWH may be preferable to UFH in the prevention of thrombo-embolism in general surgery.