Medical devices, such as catheters and heart valves, are an important part of patient care. However, blood-contacting devices can activate the blood coagulation cascade to produce factor (f) Xa, the clotting enzyme that induces thrombin generation. By activating platelets and converting soluble fibrinogen to fibrin, thrombin leads to blood clot formation. Blood clots that form on medical devices create problems because they may foul the device and/or serve as a nidus for infection. In addition, clots can break off from the device, travel through the circulation and lodge in distant organs; a process known as embolization. This is particularly problematic with central venous catheters because clots that form on them can break off and lodge in pulmonary arteries, thereby producing a pulmonary embolism. Similarly, clots that form on heart valves can break off and lodge in cerebral arteries, thereby producing a stroke. Therefore, anticoagulants, blood thinning drugs, are frequently used to prevent clotting on medical devices. Conventional anticoagulants, such as heparin and warfarin, target multiple clotting factors. Heparin binds to antithrombin in plasma and accelerates the rate at which it inhibits fXa, thrombin and many other clotting enzymes. Warfarin, which is a vitamin K antagonist, attenuates thrombin generation by interfering with the synthesis of the vitamin K-dependent clotting factors, which include fX and prothrombin, the precursor of thrombin. In contrast to heparin and warfarin, more recent anticoagulants inhibit only a single clotting enzyme. For example, fondaparinux, a synthetic heparin fragment, only inhibits fXa and dabigatran, an oral thrombin inhibitor, only targets thrombin. Although effective for many indications, fondaparinux was less effective than heparin for preventing clotting on catheters in patients undergoing heart interventions and dabigatran was less effective than warfarin for preventing strokes in patients with mechanical heart valves. The failure of these new anticoagulants highlights the need for a better understanding into the drivers of clotting on medical devices. Therefore, the overall purpose of this thesis is to gain this understanding so that more rational approaches to its prevention can be identified. In the classical model of blood coagulation, clotting is triggered via two distinct pathways; the tissue factor (TF) pathway or extrinsic pathway and the contact pathway or intrinsic pathway; pathways which are initiated by fVIIa and fXIIa, respectively. The mechanism by which medical devices initiate clotting is uncertain. Platelet and complement activation and microparticle formation have been implicated, which would drive clotting via the TF pathway. Alternatively, medical devices can bind and activate fXII, thereby initiating the contact pathway. We hypothesized that medical devices trigger clotting via the contact pathway and induce the local generation of fXa and thrombin in concentrations that exceed the capacity of fondaparinux and dabigatran to inhibit them. To test this hypothesis, we used catheters as a prototypical medical device and we used a combination of in vitro and rabbit models. Several lines of evidence indicate that catheters initiate clotting via the contact pathway. First, catheter segments shortened the clotting time of human plasma, and this activity was attenuated in fXII- or fXI-deficient plasma, which are key components of the contact pathway, but not in fVII-deficient plasma, which is the critical component of the extrinsic pathway. Second, corn trypsin inhibitor (CTI), a potent and specific inhibitor of fXIIa, attenuates catheter thrombosis. Third, selective knockdown of fXII or fXI with antisense oligonucleotides attenuated catheter-induced thrombosis in rabbits, whereas knockdown of fVII had no effect. Therefore, these results revealed the importance of the contact pathway in device-associated thrombosis, and identified CTI or fXII or fXI knockdown as novel strategies for preventing this problem. Focusing on fXIIa as the root cause of medical device associated clotting, we coated catheters with CTI using a polyethylene glycol (PEG) spacer. In addition to unmodified catheters, other controls included catheters coated with albumin via a PEG spacer or catheters coated with PEG alone. Compared with unmodified catheters or with the other controls, CTI-coated catheters attenuated clotting in buffer or plasma systems and were resistant to occlusion in rabbits. These findings support the concept that catheter-induced clotting is driven via the contact pathway and identify CTI coating as a viable strategy for its prevention. We next set out to test the hypothesis that fondaparinux and dabigatran, which inhibit fXa and thrombin, respectively, are less effective than heparin, which inhibits multiple clotting enzymes. Fondaparinux and dabigatran were less effective than heparin at preventing catheter induced clotting and thrombin generation, respectively. Likewise, in a rabbit model of catheter thrombosis, fondaparinux was less effective than heparin and dabigatran was only effective when administered at doses that yielded plasma dabigatran levels similar to those found at peak in human given the drug; at trough levels, dabigatran was no better than placebo. Finally, we also showed synergy between heparin and either fondaparinux or dabigatran. Thus, when co-administered to rabbits in doses that on their own had no effect, the combination of fondaparinux or dabigatran plus heparin extended the time to catheter thrombosis. These findings support the hypothesis that when catheters trigger clotting via the contact pathway, fXa and thrombin are generated in concentrations that overwhelm the capacity of fondaparinux or dabigatran to inhibit them. Furthermore, the synergy between heparin and fondaparinux or dabigatran has clinical implications because it explains why supplemental heparin attenuated the risk of catheter thrombosis in patients treated with fondaparinux who underwent cardiac procedures and it identifies the potential role of supplemental heparin in dabigatran-treated patients who require such interventions. In summary, we have shown that catheters trigger clotting via the contact pathway and have identified CTI coating or fXII or fXI knockdown as viable strategies for prevention of this problem. In addition, for prevention of catheter thrombosis, we also have shown that heparin, which inhibits multiple coagulation enzymes, is more effective than fondaparinux or dabigatran, which only inhibit fXa or thrombin, respectively; findings consistent with the clinical observations. Moreover, the synergy that we observed between fondaparinux or dabigatran and heparin identifies supplemental heparin as strategy for preventing catheter thrombosis in patients receiving these drugs. Taken together, these studies provide insight into the mechanisms of catheter thrombosis and potential strategies for its prevention. / Thesis / Doctor of Philosophy (PhD)