Napsagatran, a new synthetic direct thrombin inhibitor, was compared with heparin in a canine model of coronary thrombosis and concomitantly in an ex vivo perfusion chamber model. Occlusive thrombosis of the left circumflex coronary artery was induced by electrical injury. In parallel, arterial subendothelium was exposed to native blood using an annular perfusion chamber for 5, 10 and 20 min at a wall shear rate of 650/s. Dogs received saline, heparin (40 and 70 U/kg/h) or napsagatran (3 and 10 microgram/kg/min). Heparin (40 U/kg/h) and napsagatran (3 microgram/kg/min) delayed or prevented in vivo thrombotic occlusion, but only napsagatran (10 microgram/kg/min) significantly decreased the intracoronary thrombus when compared with saline. High-dose heparin (70 U/kg/h) or napsagatran (10 microgram/kg/min) decreased the platelet-rich thrombus after a 20-min chamber perfusion. Neither heparin nor napsagatran decreased the thrombus volume after a 5-min perfusion. Heparin (70 U/kg/h) and napsagatran (10 microgram/kg/min) prolonged the activated partial thromboplastin time differently (>x6 and x1.4, respectively, P<0.01), whereas the activated clotting time was prolonged equally (x2.5). Thus napsagatran in this model shows arterial antithrombotic effects similar to those of heparin. The chamber experiments suggest that neither compound affects the initiation of platelet thrombus formation. In arterial thrombosis, the activated clotting time has a higher predictive value than the activated partial thromboplastin time when a direct thrombin inhibitor is compared with heparin.