There is no universally accepted approach to the initiation of systemic anticoagulant therapy. In an open, randomized study, two anticoagulant regimens that differed only in the timing of warfarin therapy after the start of heparin were compared. We randomized 119 patients with acute thromboembolic events to receive warfarin either within 48 hours of the start of heparin (early group, n = 63) or 96 hours or later after the start of heparin (late group, n = 56). Heparin was given as a 5000 IU bolus as a constant infusion titrated to maintain the activated partial thromboplastin time at 1.5 to 2 times control values. Warfarin was started at 10 mg daily for 3 days and the dose was titrated to maintain the prothrombin time at 1.2 to 1.5 times control values. There were no significant differences between the early and late warfarin groups with regard to age, sex, indication for anticoagulation, heparin dose, mean activated partial thromboplastin time during heparin, warfarin dose at discharge, length of warfarin therapy before discharge, bleeding, recurrent thromboembolic events, or mortality rates. Time to the start of warfarin after heparin was 31 hours and 108 hours in the early and late groups, respectively. Length of hospitalization, hospital costs, and the incidence of heparin-induced infusion phlebitis and thrombocytopenia were significantly less in the early group compared with the late group. Early initiation of warfarin after heparin is safer, less expensive, and as effective as the late initiation of warfarin.