PURPOSE: In this study, we aimed to estimate recurrence risk after incident venous thromboembolism, stratified according to unprovoked, provoked, and cancer-related venous thromboembolism in a prospective cohort of inpatients and outpatients receiving routine care. METHODS: We linked nationwide Danish health registries to identify all patients with incident venous thromboembolism from January 2000 through December 2015. Rates of recurrence were calculated and Cox regression was used to compute hazard ratios (HRs) with 95% confidence intervals (CIs) by incident venous thromboembolism type after adjusting for coexisting risk factors. RESULTS: The study included 73,993 patients with incident venous thromboembolism (54.1% females; mean age, 62.3 years). At 6-month follow-up, rates per 100 person-years were 6.80, 6.92, and 9.06 for provoked, unprovoked, and cancer-related venous thromboembolism, respectively. At 10-year follow-up, corresponding rates were 2.22, 2.84, and 3.70, respectively. Additionally, at 6-month follow-up, hazard rates of recurrence were comparable for patients with unprovoked venous thromboembolism 1.01 (95% CI, 0.92-1.11) and provoked. At 10-year follow-up, unprovoked venous thromboembolism (HR, 1.17; 95% CI, 1.12-1.23) and cancer -related venous thromboembolism (HR, 1.21; 95% CI, 1.12-1.32) were associated with higher risk of recurrence compared with that found in provoked venous thromboembolism. CONCLUSIONS: In this nationwide cohort, patients with cancer -related venous thromboembolism had the highest risk of recurrence. At 6-month follow-up, there were similar risks of recurrence for patients with unprovoked and provoked venous thromboembolism. At 10-year follow-up, recurrence risks were similar for patients with unprovoked venous thromboembolism and patients with cancer -related venous thromboembolism. High recurrence risks in all categories indicate that further research is needed to optimize duration of extended anticoagulation for these patients. (C) 2018 Elsevier Inc. All rights reserved.