Prediction of Recurrent VTE and Bleeding With Extended Anticoagulation

Quick Takes

  • Prediction of recurrent venous thromboembolism (VTE) and bleeding after an initial 3-6 months of anticoagulation has been challenging.
  • The VTE-PREDICT risk models performed modestly well in external validation cohorts at predicting recurrent VTE and bleeding during secondary VTE prophylaxis.
  • The VTE-PREDICT risk models may have limited utility in facilitating shared decision making about anticoagulation duration after an acute VTE.

Study Questions:

What are the predictors of recurrent venous thromboembolism (VTE) and bleeding in noncancer patients who have completed an initial 3+ month course of anticoagulation?


The authors conducted a competing risk-adjusted model for each recurrent VTE and clinically relevant bleeding (nonmajor and major) using readily available patient characteristics. These were derived from five prior studies of 15,141 patients, 220 recurrent VTE events, and 189 bleeding events. External validation was assessed in five additional studies of 59,257 patients, 2,283 recurrent VTE events, and 3,335 bleeding events. Predictors are specific to type and dose of anticoagulant/antiplatelet agent use.


The recurrent VTE prediction model included age, female sex, body mass index (BMI), type of index event (pulmonary embolism vs. deep vein thrombosis), presence of provoking risk factors, use of estrogen therapy, history of cancer, and prior history of VTE. The bleeding risk model included age, female sex, systolic blood pressure, type of index event, history of cancer, history of bleeding, prior history of stroke, hemoglobin lab value, and concurrent use of nonsteroidal anti-inflammatory medications. In the external validation cohorts, C-statistics ranged from 0.48-0.71 for predicting recurrent VTE and 0.61-0.68 for predicting bleeding. In the Danish VTE cohort, 5-year risks of VTE ranged from 4-19%, while risk of bleeding ranged from 1-19%.


The authors concluded that the VTE-PREDICT risk score can be applied to estimate the effect of extended anticoagulation treatment for individual patients with VTE and to support shared decision making.


While all patients with acute VTE should be treated with at least 3-6 months of anticoagulation, the decision to extend treatment beyond that time can be more challenging. Several prior recurrent VTE risk models have been developed but only a few have been externally validated. There has been even more challenge developing a useful bleeding risk score for patients with VTE. This study attempts to make progress on both fronts, leveraging many of the same clinical elements as are included in previous scores. However, the overall discrimination ability of the VTE-PREDICT risk score is poor-modest in the external validation cohorts, limiting its ability to be definitive. Yet it still may serve as a useful role for clinicians and patients as they engage in shared decision making about the role of secondary VTE prevention after the initial 3- to 6-month acute treatment phase.

Clinical Topics: Anticoagulation Management, Prevention, Vascular Medicine

Keywords: Anticoagulants, Anti-Inflammatory Agents, Non-Steroidal, Blood Pressure, Decision Making, Shared, Estrogens, Hemoglobins, Hemorrhage, Platelet Aggregation Inhibitors, Pulmonary Embolism, Recurrence, Risk, Secondary Prevention, Stroke, Thrombosis, Vascular Diseases, Venous Thrombosis

< Back to Listings