BACKGROUND The main objective of the study presented was to test if thrombus regression can be improved by treatment with an intravenously or subcutaneously administered low molecular weight heparin (LMWH). Patients with acute deep vein thrombosis were randomly assigned to receive either intravenous UFH (131 patients), intravenous (i.v.) LMWH (128 patients), or 8000 IU of the same LMWH bid subcutaneously (s.c.) (128 patients). All patients were treated with heparin for 14 to 16 days. Vitamin-K-antagonist prophylaxis was started between Day 12 and Day 14 after enrollment into the study. METHODS Phlebographies and perfusion/ventilation lung scans were performed at baseline and on Days 12 to 16. Primary endpoint of the study was a reduction of the phlebographic Marder score. Secondary endpoints were recurrent thrombosis and pulmonary embolism (PE), major and minor bleedings and the rate of PE at inclusion and at the end of the study assessed by ventilation/perfusion scans. RESULTS The Marder score improved by at least 30% in 32.4% (95% CI: 22.6 ... 42.2) of the patients receiving UFH, in 34.0% (95% CI: 24.9 ... 44.0) receiving LMWH i.v. and in 42.6% (95% CI: 32.8 ... 52.8) treated with the low molecular weight heparin s.c. The difference between LMWH s.c. and UFH was 10.2% (95% CI: -3.7% ... +24.5%) (p = 0.11). PE with clinical signs confirmed by objective methods occurred in three patients of the UFH group, one of whom died and was not observed in patients of the i.v. or s.c. LMWH-groups. During the first 15 days no patient receiving UFH or i.v. LMWH, and one patient on s.c. LMWH had a recurrent thrombosis. Major bleedings were observed in four patients receiving i.v. UFH compared to nine patients on i.v. LMWH (one of these patients died) and one patient on s.c. LMWH. Perfusion ventilation lung scans were obtained from 287 patients at baseline and from 246 patients on Days 12-16. PE, defined according to PIOPED-criteria as intermediate or high probability scans, was observed in 38.0% of the patients entering the study and in 18.3% on Days 12 to 16. New asymptomatic PE occurred less frequently in the groups on LMWH (7.1%, 7.5%, respectively) than in the UFH-group (12.6%) (not significant). CONCLUSIONS S.c. treatment with a LMWH (certoparin) (b.i.d.) is at least as effective as UFH i.v. The hypothesis of increased efficacy of subcutaneous LMWH in resolving venous thrombi will have to be confirmed by an independent study comparing s.c. LMWH with UFH. The i.v. continuous infusion of the LMWH for 12 to 16 days does not result in a higher venous re-opening rate than intravenous standard heparin.