BackgroundPatients who have had an unprovoked deep venous thrombosis (DVT) or pulmonary embolus (PE) are at a high risk for recurrent venous thromboembolism (VTE). Extended “life-long” anticoagulation has been recommended in these patients. However, the risk benefit ratio of this approach is controversial and the role of the direct oral anticoagulants (DOACs) and aspirin is unclear. Furthermore, in some patients with a “weak provoking factor” there is clinical equipoise regarding continuation or cessation of anticoagulant therapy after treatment of the acute VTE event.ObjectiveA systematic review and meta-analysis to determine the risks (major bleeding) and benefits (recurrent VTE and mortality) of extended anticoagulation with vitamin k antagonists (VKA), DOACs and aspirin in patients with an unprovoked VTE and in those patients with clinical equipoise regarding continuation or cessation of anticoagulant therapy. In addition, we sought to determine the risk of recurrent VTE events once extended anti-thrombotic therapy was discontinued.Data SourcesMEDLINE, Cochrane Register of Controlled Trials, citation review of relevant primary and review articles.Study SelectionRandomized placebo-controlled trials (RCTs) that compared the risk of recurrent VTE in patients with an unprovoked DVT or PE who had been treated for at least 3 months with a VKA or a DOAC and were then randomized to receive an oral anti-thrombotic agent or placebo for at least 6 additional months. We included studies that included patients in whom clinical equipoise existed regarding the continuation or cessation of anticoagulant therapy.Data ExtractionIndependent extraction of articles by both authors using predefined data fields, including study quality indicators. Data were abstracted on study size, study setting, initial event (DVT or PE), percentage of patients where the initial VTE event was unprovoked, the number of recurrent VTE events, major bleeds and mortality during the period of extended anticoagulation in the active treatment and placebo arms. In addition, we recorded the event rate once extended treatment was stopped. Meta-analytic techniques were used to summarize the data. Studies were grouped according to the type of anti-thrombotic agent.Data SynthesisSeven studies which enrolled 6778 patients met our inclusion criteria; two studies evaluated the extended use of Coumadin, three studies evaluated a DOAC and two studies evaluated the use of aspirin. The duration of followup varied from 6 to 37 months. In the Coumadin and aspirin studies 100% of the randomized patients had an unprovoked VTE, while in the DOAC studies between 73.5% and 93.2% of the VTE events were unprovoked. In the control group recurrent VTE occurred in 9.7% of patients compared to 2.8% in the active treatment group (OR 0.21; 95% CI 0.11–0.42, p<0.0001). VKA, DOACs and aspirin significantly reduced the risk of recurrent VTE, with VKA and DOACs being significantly more effective than aspirin. Major bleeding events occurred in 12 patients in the control group (0.4%) and 25 of 3815 (0.6%) patients in the active treatment group (OR 1.64; 95% CI 0.69–3.90, NS). There were 39 (1.3%) deaths in control patients and 33 (0.9%) deaths in the anti-thrombotic group during the treatment period (OR 0.73; 95% CI 0.40–1.33, NS). Patients whose initial VTE event was a PE were more likely to have a recurrent PE than a DVT. The annualized event rate after discontinuation of extended antithrombotic therapy was 4.4% in the control group and 6.5% in the active treatment arm.ConclusionsVKA, DOACs and aspirin significantly reduced the risk of recurrent VTE, with DOACs and VKA being more effective than aspirin. The decision regarding life-long anticoagulation following an unprovoked DVT or PE should depend on the patients’ risk for recurrent PE as well as the patients’ values and preferences.