PURPOSE: To assess differing mechanisms of thrombolysis determining time to reperfusion, completeness of thrombus dissolution, and embolic potential, MATERIALS AND METHODS: An in vitro perfusion model designed to mimic arterial flow conditions was created, Bifurcated limbs allowed continuous flow through one channel and the placement of radiolabeled (iodine-125) thrombus housed in a 5-cm segment of polytetrafluoroethylene graft in the other, The three experimental groups consisted of a standard continuous urokinase infusion, a pulsed pressurized injection of saline, and a similar injection with urokinase, A continuous infusion of 5% dextrose served as a control group, Time to reflow (as assessed with ultrasonic flow monitoring), completeness of thrombus dissolution (I-125 liberated into solution), and the number and size of embolic particles produced (detected by a series of graduated filter sizes) were analyzed, RESULTS: Time to reflow was significantly faster for both groups when pressurized injections were used (P < .001), There was no reflow in the control arm at 90 minutes, Completeness of thrombus dissolution was higher when a continuous infusion of urokinase was used in comparison to either of the power injection groups or the control (P < .05), The amount of embolic debris produced was significantly lower with a continuous infusion of urokinase compared with either of the power lysis groups (P < .05), but significantly greater than the control arm (P < .001). The size of the embolic particles in the power pulsed lysis groups was significantly decreased by the addition of urokinase (P < .05), CONCLUSIONS: Reflow is more rapidly established by the use of mechanical means, However; a less complete dissolution of thrombus in conjunction with a greater amount of embolic debris is achieved with this approach, The size of the embolic particles produced is reduced by the addition of a thrombolytic agent.