In patients with mechanical heart valves, thromboembolic events were more frequent with dabigatran, an oral thrombin inhibitor, than with warfarin. This observation raises the possibility that dabigatran may be less effective than conventional anticoagulants in patients with other blood-contacting devices, such as catheters. To address this, we compared the capacity of dabigatran and/or heparin to inhibit catheter-induced thrombin generation in vitro and to attenuate catheter occlusion in rabbits. Using a catheter-induced thrombin generation assay, concentrations of dabigatran over 100 ng/ml prolonged the lag time and time to peak thrombin, and reduced the peak thrombin concentration and endogenous thrombin potential in a concentration-dependent fashion. Compared with saline in a rabbit model of catheter thrombosis, dabigatran prolonged the mean time to catheter occlusion by 2.9- and 1.9-fold when plasma levels were 173 and 140 ng/ml, respectively; values comparable to median peak levels in humans given dabigatran 150 mg twice daily. In contrast, low-dose dabigatran, which produced a level of 60 ng/ml; a value comparable to the trough level of dabigatran in humans, did not prolong the time to occlusion. Whereas a 70 U/kg bolus of heparin prolonged the mean time to occlusion by 3.4-fold, a 15 U/kg bolus had no effect. When low-dose dabigatran was given in combination with 15 U/kg heparin, the mean time to occlusion was prolonged by 2.7-fold. These findings suggest that only peak levels of dabigatran are sufficient to prevent catheter-induced clotting unless supplemented heparin is given.