Background Although intravenous heparin is routinely used in the treatment of patients with acute coronary syndromes, this anticoagulant requires antithrombin III as a cofactor, has no affinity to clot-bound thrombin, and is bound or inactivated by several plasma proteins and platelet factor 4. Recombinant hirudin, the prototypic direct thrombin inhibitor, has been demonstrated in pilot studies to yield improved angiographic and clinical outcomes compared with heparin. We compared these two antithrombins in a large-scale randomized trial.Methods and Results At 275 participating hospitals in 12 countries, patients within 12 hours from the onset of ischemic chest discomfort with an abnormal ECG were randomly assigned to receive a 72-to 120-hour infusion of heparin (5000-U bolus and 1000- to 1300-U/h infusion, adjusted to activated partial thromboplastin time (APTT) of 60 to 90 seconds or hirudin (0.6-mg/kg bolus and 0.2-mg/kg per hour infusion without aPTT adjustment) on a double-blind basis. Although recruitment of 12 000 patients was planned, the trial was stopped earlier because of an excess of intracerebral hemorrhagic events after 2564 patients were enrolled. The overall incidence of hemorrhagic stroke tended to be higher for patients receiving hirudin (1.3%) compared with heparin (0.7%), P=.11, but the incidence was significantly higher in patients receiving thrombolytic therapy (1264 patients, 1.8%) compared with those who did not (1168 patients, 0.3%), P<.001. The hemorrhagic stroke rate varied by the thrombolytic and antithrombin combination: tissue-type plasminogen activator and heparin, 0.9%; with hirudin, 1.7%; streptokinase with heparin, 2.7%; with hirudin, 3.2%. All these rates are higher than the overall incidence of hemorrhagic stroke in the patients receiving thrombolytic therapy and intravenous heparin in the GUSTO I trial (30 892 patients with rate of 0.7%, 95% CI of 0.6 to 0.8%). Among the 26 patients who had intracerebral hemorrhages, the aPTT was significantly elevated compared with the event-free patients (110+-46 versus 87+-36 seconds at 12 hours of therapy, respectively), P=.03.Conclusions At the dose of hirudin tested, there was a trend of an excess of hemorrhagic stroke compared with heparin. Heparin, at a slightly higher dose than previously used in a large-scale trial (approximately 20% increase) was accompanied by a twofold risk of hemorrhagic stroke in patients receiving thrombolytic therapy. With both thrombin inhibitors, the aPTT appears to be a useful index for predicting risk of hemorrhagic stroke in patients receiving thrombolytic therapy. (Circulation. 1994;90:1631-1637.)