Three dosages (0.3, 0.7, and 1.0 mg/kg) of recombinant hirudin, a specific inhibitor of thrombin, were compared with heparin (50 units/kg) and placebo for reducing thrombus formation in the carotid arteries of 50 pigs after deep injury by balloon dilatation. Each drug was administered as a bolus followed immediately by a continuous infusion of the same dose per hour. Major end points were quantitative indium-111-labeled platelet and iodine-125-labeled fibrinogen deposition and the incidence of mural thrombosis. This study showed that heparin, at a dose that prolonged the activated partial thromboplastin time (APTT) to twice the control time, did not prevent mural thrombosis or significantly reduce platelet deposition compared with placebo but did reduce fibrinogen deposition. Recombinant hirudin markedly reduced platelet and fibrinogen deposition in a dose-related manner and totally eliminated mural thrombosis at an APTT of two to three times that of control. Platelet deposition (x 16/cm2, mean±SEM) in areas of deep arterial injury for the placebo, heparin, and 0.3, 0.7, and 1.0 mg/kg hirudin groups was 54+±21, 33+9, 22+f6, 8±1, and 7±1, respectively; electron microscopy showed a single layer (or less) of platelets at the two highest hirudin dosages. The incidence of macroscopic mural thrombosis was 76% with placebo, 57% with heparin, 46% with 0.3 mg/kg hirudin; there were no thrombi with 0.7 or 1.0 mg/kg hirudin (p<0.01). The APTT is a valuable index of the plasma hirudin concentration (r=0.89, p<0.001) and correlated inversely with quantitative platelet (r= -0.67, p<0.0001) and fibrinogen deposition (r= -0.42, p=0.005). The thrombin time is overly sensitive and plays no role in monitoring heparin or hirudin therapy for arterial thrombosis. Heparin and all dosages of hirudin abolished platelet aggregation to thrombin and prolonged the bleeding time to a similar degree; neither test reflects the in vivo antithrombotic efficacy of heparin or hirudin. Thrombin plays a critical role in arterial thrombosis. Heparin in conventional doses does not significantly reduce arterial platelet thrombosis. Hirudin produces a dose-dependent reduction in platelet and fibrinogenfibrin deposition, which correlates with the prolongation of the APTT and the hirudin plasma level. Hirudin totally eliminates macroscopic mural thrombus formation at dosages that prolong the APTT to at least twice control and is a promising therapeutic agent for platelet-rich arterial thrombi. (Circulation 1990;82:1476-1484 Ba alloon dilatation produces mechanical disruption of the arterial wall with injury into the media that releases tissue thromboplastin (tissue factor) and exposes very thrombogenic deep arterial structures (mainly smooth muscle cells and