Although the control of thrombin in the microvasculature at the endothelial cell surface is crucial to prevent atherothrombosis, the role of antithrombin in arterial thrombosis is unclear. It is widely considered that antithrombin deficiency is unlikely to contribute to arterial thrombosis, but no convincing epidemiological study has been performed because of the low frequency of this deficiency. In this study we evaluated the role in myocardial infarction (MI) of a relatively common mutation affecting antithrombin gene (A384S: Antithrombin Cambridge II) that has functional features that may impair the right control of thrombogenic events caused by injury to the vascular wall. Moreover, this deficiency, which is not detected using common methods to diagnose antithrombin deficiency, also increases the risk of venous thrombosis. We included 1,224 patients with MI (691 consecutive patients and 533 survivors of a premature event), and 1,649 controls. The mutation was identified in 0.3% of controls, but 0.8% of MI patients. After adjusting for sex and other cardiovascular risk factors, the antithrombin Cambridge II significantly increased 5.66-fold the risk of MI (95% CI: 1.53-20.88; p = 0.009). Interestingly, young patients had the highest risk of MI associated with the mutation (OR: 9.98; 95%CI: 1.60-62.24; p = 0.009). This is the first epidemiological study that supports a role for antithrombin deficiency in arterial thrombosis. These results suggest that deficiency of antithrombin may be an independent risk factor for MI that has been underestimated, but larger studies are needed to confirm the relevance of inhibitors of thrombin in arterial thrombosis.