D-dimer assay is used to stratify patients with unprovoked venous thromboembolism (VTE) for the risk of recurrence. However, this approach was never evaluated since direct oral anticoagulants are available. With this multicenter, prospective cohort study we aimed at assessing the value of an algorithm incorporating serial D-dimer testing and administration of reduced dose apixaban (2.5 mg twice daily) only to patients with positive test. 732 outpatients aged 18 to 74 years, anticoagulated for at least 12 months after a first unprovoked VTE were included. Patients underwent D-dimer testing with commercial assays and pre-established cutoffs. If the baseline D-dimer, during anticoagulation, was negative, anticoagulation was stopped and testing repeated after 15, 30, and 60 days. Patients with serially negative results (286; 39.1%) were left without anticoagulation. At first positive result, the remaining 446 patients (60.9%) were given apixaban for 18 months. All patients underwent follow-up planned for 18 months. The study was interrupted after a planned interim analysis for the high rate of primary outcomes (7.3%; 95% CI, 4.5 to 11.2), including symptomatic proximal DVT or PE recurrence, death for VTE, major bleeding, occurring in patients off anticoagulation, versus that in those receiving apixaban (1.1%; 95% CI, 0.4 to 2.6; adjusted HR, 8.2; 95% CI, 3.2 to 25.3). In conclusion, in patients anticoagulated for at least one year after a first unprovoked VTE, the decision whether to further extend anticoagulation should not be based on D-dimer testing. The results confirmed the high efficacy and safety of reduced dose apixaban against recurrences. ClinTrials.gov: NCT03678506.