Whether the factor V Leiden mutation increases the risk of recurrent venous thromboembolism (VTE) remains controversial, and homozygous carriers have not been extensively evaluated. In the present cohort study, we have assessed the incidence of recurrent VTE in a subset of non-treated patients previously included in the Procare Register who suffered a first objectively diagnosed VTE. We compared the incidence rates in homozygotes (n = 32) and heterozygotes (n = 108). The median observation time was 4.5 years (range 1.0-6.9 years) in the homozygous group, and 4.15 years (range 0.3-5.9 years) in the heterozygous group. The incidence of recurrence was 5.75/100 patient-years (95% confidence interval = 4.9-7.4) and 3.2/100 patient-years (95% confidence interval = 2.9-3.5) for the homozygotes and heterozygotes, respectively. Homozygotes had a higher overall risk of recurrence than did heterozygotes (relative risk, 1.8; 95% confidence interval = 1.0-6.17). Since short-term prophylaxis was encouraged in risk situations, idiopathic VTE (76%) occurred more frequently than provoked VTE (24%). The precipitating condition was pregnancy in all cases (three homozygotes, one heterozygote). None were receiving prophylaxis. Balancing the risk of recurrence against the risk of major bleeding from oral anticoagulation therapy, it appears that factor V Leiden homozygotes with a first VTE are unlikely to benefit from long-term full-dose oral anticoagulant treatment. All these patients should receive short-term prophylaxis during risk situations, particularly during pregnancy.