SummaryInhibition of arterial thrombus formation by ancrod, a fibrinogen depleting agent isolated from a snake venom, r-hirudin, an inhibitor of thrombin-mediated fibrinogen cleavage, or the glycoprotein (GP)IIb/IIIa-receptor antagonist Ro 43-8857 interfering with fibrinogen binding to platelets, was evaluated in two canine models. As a marker of platelet-dependent thrombus formation, cyclic blood flow reductions (CFR) were induced in the left coronary artery (LAD) of mongrel dogs by mechanical injury of the endothelium combined with critical stenosis. In the second model CFRs were induced by thrombolysis of a copper coil-induced thrombus in the carotid artery. Blood flow rate during the reocclusion phase was used as an additional parameter of efficacy.The frequency of CFRs used as indicator of platelet aggregation and adhesion was significantly diminished by all treatments in both the carotid artery- and the LAD-model. In the LAD-model, following ancrod treatment, CFRs were correlated with plasma fibrinogen concentrations. Carotid artery blood flow after reperfusion, used as indicator of occlusive thrombus formation, rapidly declined to zero in the control group but remained at a high level after treatment with ancrod or r-hirudin. Ro 43-8857 at the selected dose improved blood flow rate only to a minor degree but prolonged the bleeding time from a mean value of 87.2 ± 10.9 s (n = 24) to values >300 s in 50% of the animals.Our results indicate that CFRs as indicator of platelet aggregation and adhesion are inhibited by either treatments. Blood flow as indicator of occlusive thrombus formation, however, is effectively improved by ancrod and r-hirudin only. Inhibition of fibrinogen binding to platelet GPIIb/IIIa receptors alone was found to be a less potent antithrombotic principle in this model.