DX-9065a is an antithrombin III (AT III)-independent and selective inhibitor of activated blood coagulation factor X (FXa). We evaluated the effects of DX-9065a and warfarin on bleeding time and blood loss in rat tail transection model and on blood loss in hydrochloride (HCl)-induced rat gastrointestinal haemorrhage model. The blood loss was determined by measuring the haemoglobin content in saline immersed with transected tail or hematin chloride content in the gaster after HCl administration. DX-9065a or warfarin was administered orally at 1 h or 15-21 h before the haemorrhagic stimuli, respectively. The dose required for 50% inhibition of thrombus formation (ID50) was 21 mg/kg for DX-9065a and 0.75 mg/kg for warfarin in a copper wire-inserted arteriovenous (AV) shunt model. In contrast to DX-9065a (10 or 30 mg/kg), warfarin (0.75 mg/kg) significantly prolonged the bleeding time. In rat tail transection model, the blood loss for the control group was 102+/-41 microl at 20 min after the transection. While warfarin (0.75 mg/kg) facilitated the blood loss about 5 times as much as the control, DX-9065a (10 or 30 mg/kg) did not. In rat gastrointestinal model, the blood loss for the control group was 15.9+/-5.6 microl at 15 min after HCl administration. In contrast to DX-9065a (10 or 30 mg/kg), warfarin (0.75 mg/kg) increased the blood loss about twice as much as the control. Thus, compared with warfarin, DX-9065a only increased bleeding time or blood loss to a minor extent in the doses tested. These observations suggest that direct inhibition of FXa could be preferable to warfarin in the suppression of thrombosis without haemorrhagic complications.