OBJECTIVE To compare the efficacy and safety of recombinant human tissue-type plasminogen activator (rt-PA, supplied as Activase) with heparin alone or rt-PA plus heparin in the treatment of venographically documented proximal deep venous thrombosis (DVT) of the leg. METHODS Sixty-four patients underwent 65 randomizations to rt-PA alone (n = 36), rt-PA plus heparin (n = 17), or heparin alone (n = 12) in a prospective, multicenter, randomized, open-label trial, with efficacy assessed by a radiology panel unaware of treatment assignment. Patients randomly assigned to rt-PA received 0.05 mg/kg/hour for 24 hours via a peripheral vein, with a maximum dose of 150 mg. All patients then received heparin and warfarin for the remainder of the hospitalization. Follow-up venography was performed 24 to 36 hours after initiation of therapy. RESULTS Complete or more than 50% lysis occurred in 10 (28%) patients treated with rt-PA, five (29%) patients with rt-PA plus heparin, and no patient treated with heparin. No lysis occurred in 16 (44%) patients treated with rt-PA plus heparin, and 10 (83%) patients who received heparin alone (p = 0.04). There was one major complication, a nonfatal intracranial hemorrhage in a patient who received rt-PA alone. At 7 to 10 days after initiation of treatment, the level of serum glutamic oxaloacetic transaminase nearly doubled among all patients, including those assigned to receive heparin alone. CONCLUSIONS (1) rt-PA and rt-PA plus heparin cause more clot lysis than heparin alone; (2) the addition of heparin to rt-PA does not improve the lysis rate; (3) DVT treated with heparin is commonly associated with a rise in the transaminase level; (4) heparin does not increase the risk of bleeding from rt-PA therapy; and (5) alternative dosing regimens and modes of administration of rt-PA should be investigated to improve further its efficacy and safety in the treatment of acute DVT.