To determine whether platelet-activating factor is a specific mediator of cyclic flow variations in damaged stenotic arteries and whether it contributes to reocclusion after thrombolysis, femoral arteries in anesthetized dogs were subjected to mural injury and high grade stenosis to induce cyclic flow variations (28 ± 4/h) or methods selected to elicit platelet-rich and fibrin-rich thrombosis. Oral administration of a novel triazolobenzodiazepine (U46,195 [10 mg/kg]) that selectively inhibits platelet-activating factor abolished cyclic flow variations within 120 min and far ≥2 h thereafter compared with persistent flow variations in dogs given saline solution. Platelet aggregation induced ex vivo with platelet-activating factor was inhibited in parallel with in vivo inhibition of cyclic flow variations after administration of U46,195. However, buccal macosa bleeding time was not affected. After thrombosis, administration of U46,195 before thrombolysis was induced with human recombinant tissue-type plasminogen activator (1.7 mg/kg intravenously over 60 min) prevented reocclusion within 120 min in six of eight and six of seven arteries by platelet-rich and fibrin-rich thrombosis, respectively. In contrast, in dogs given saline solution, reocclusion occurred in eight of eight (p = 0.007 compared with U46,195) and five of eight arteries by platelet-rich and fibrin-rich thrombosis, respectively. Thus, both cyclic flow variations and reocclusion after thrombolysis appear to be mediated in part by platelet-activating factor. The results suggest that inhibition of platelet-activating factor with specific antagonists may be useful in reducing platelet-mediated occlusion of coronary arteries without eliciting bleeding.