Predictors of recurrent venous thromboembolism and major bleeding in patients with cancer: A secondary analysis of the CANVAS trial

Abstract

Introduction

Patients with cancer have an increased risk of developing venous thromboembolism (VTE) but also have an increased risk of both recurrent VTE and bleeding with anticoagulation compared to anticoagulated patients without cancer. CANVAS, a randomized pragmatic effectiveness trial, compared the direct oral anticoagulants a class to low molecular weight heparin for treatment of a new VTE in patients with cancer. The aim of this prespecified secondary analysis of the CANVAS trial is to identify predictors of both recurrent VTE and major bleeding in patients with cancer and new VTE.

Methods

Data from the 671 participants in the analysis population were used to identify predictors of recurrent VTE and bleeding during the 6-month treatment period. Significant predictors identified in the univariable models were carried forward in the multivariable models to identify independent predictors of both risks.

Results

Independent predictors of recurrent VTE include ECOG performance status ≥2 (HR, 3.19 [95 % CI, 1.45–7.02]; P < .005), presence of metastatic disease (HR, 2.57 [95 % CI, 1.14–5.80]; P = .023), treatment with bevacizumab (HR, 2.50 [95 % CI, 1.04–5.99]; P = .041), and deep vein thrombosis without pulmonary embolus as index VTE (HR, 1.86 [95 % CI, 1.04–3.33]; P = .037). Independent predictors of major bleeding include serum albumin <3.5 g/dL (HR 1.97 [95 % CI, 1.02–3.79]; P = .044) and metastatic disease (HR 2.80 [95 % CI, 1.08–7.22]; P = .034).

Conclusion

Findings from this pre-specified analysis of the CANVAS trial identified risk factors for recurrent VTE and major bleeding in a population of participants with cancer and new VTE that reflect current oncology clinical practice. Results can be used to identify at risk patients in practice and inform new risk prediction models to improve the care of these patients.