We studied whether there was a relationship between the anticoagulant effects of recombinant human soluble thrombomodulin (rhsTM) and activation of protein C in a primate model of acute vascular graft thrombosis in 11 baboons (Papio species). Baboons were pretreated with 0.1, 1 and 5 mg/kg of rhsTM, with or without co-injection of a neutralising monoclonal antibody to protein C (HPC4) in the 1 mg/kg rhsTM group. Subsequently, thrombogenic polyester grafts were deployed for 3 h into chronic exteriorised arteriovenous shunts. Thrombus growth in the graft, plasma-activated protein C (APC) levels, coagulation and thrombosis markers were determined. In untreated baboons, baseline circulating APC levels more than doubled and graft thrombi propagated until reaching equilibrium in about 1 h. Treatment with rhsTM reduced thrombus propagation rates, prolonged the clotting and bleeding times, decreased thrombin-antithrombin complex, beta-thromboglobulin and fibrinopeptide A levels, and, surprisingly, also decreased systemic APC levels, in a dose-dependent manner. In the presence of HPC4 antibody to inhibit APC generation, the acute antithrombotic activity of rhsTM on graft thromboses was not attenuated for up to 80 min, but sustained thrombus accumulation was observed over a 180-min period. These findings suggest that, in contrast to the prevailing hypotheses, the primary antithrombotic activity of rhsTM is independent of protein C, at least in this primate model. Direct inhibition of thrombin's prothrombotic activity upon complex formation with rhsTM might explain the molecular mechanism of the observed antithrombotic effect.