For intravenous anticoagulation, heparin has been the mainstay drug, but its use may be contraindicated in heparin-induced thrombocytopenia and thrombosis. Heparin alternatives including direct thrombin inhibitors are available, but clotting assays (e.g. partial thromboplastin time) measure the time required to form fibrin gel when only a small amount of thrombin is generated. It was hypothesized that the extent of thrombin inhibition varies among inhibitors, and thrombin-activated thrombelastography would provide useful data on therapeutic responses to thrombin inhibitors. Thrombin was added (0-100 nmol/l final concentration) to nonrecalcified whole blood to evaluate clot formation on thrombelastography. Effects of direct thrombin inhibitors (argatroban 3.75 microg/ml, bivalirudin 15 microg/ml, and lepirudin 3.0 microg/ml), and heparin cofactor II activator and dermatan disulfate (20 microg/ml) were evaluated in the presence of 100 nmol/l thrombin. The interactions of thrombin and respective inhibitors were also compared by fluorogenic thrombin substrate cleavage. Increasing concentrations of thrombin progressively shortened the lag time and increased viscoelasticity on thrombelastography. Only 20 nmol/l thrombin caused instantaneous clotting, but maximal viscoelastic force was obtained at 50-100 nmol/l thrombin. All thrombin inhibitors prolonged the lag time (lepirudin > bivalirudin > argatroban = dermatan disulfate), but full recovery of thrombelastography viscoelasticity was observed with argatroban and bivalirudin. Lepirudin abrogated clotting, and dermatan disulfate suppressed clot development on thrombelastography. Thrombin substrate cleavage was observed only for bivalirudin, and heparin cofactor II without dermatan disulfate. The modified thrombelastography technique using nonrecalcified whole blood may be useful in evaluating the extent and reversibility of thrombin blockade with direct or indirect thrombin inhibitors.