The optimal duration of oral anticoagulant therapy for patients with acute proximal deep vein thrombosis (DVT) is uncertain. Based on the hypothesis that a normal impedance plethysmogram (IPG) following DVT defines a group of patients at low risk of recurrent venous thromboembolism (VTE), a trial was conducted to evaluate the efficacy of only four weeks of warfarin. Patients with venographically confirmed acute proximal DVT who had received four weeks of warfarin after initial heparin and whose four week IPG was normal were allocated to either continue warfarin (targeted International Normalized Ratio 2.0 to 3.0) for a further eight weeks or receive placebo. Patients with an abnormal four week IPG received warfarin for a further eight weeks. Based on clinical characteristics at the time of the qualifying thrombosis, all patients were classified as having either continuing or transient risk factors for recurrent VTE. During the eight weeks following randomization, nine (8.6%) of the 105 placebo patients developed recurrent VTE compared to one (0.9%) of the 109 warfarin patients, P = 0.009. Over the entire 11 months of follow-up, 12 placebo patients developed recurrence compared to seven warfarin patients, P = 0.3. Nineteen of the 192 patients with an abnormal four week IPG experienced recurrence during the nine months after discontinuing warfarin. In the 301 patients who received three months of warfarin in the randomized trial or in the cohort study, all 26 recurrent events were in the 212 patients with continuing risk factors.(ABSTRACT TRUNCATED AT 250 WORDS)