The predictive ability of bleeding risk stratification models in very old patients on vitamin K antagonist treatment for venous thromboembolism: results of the prospective collaborative EPICA study: comment

Anaemia as an independent key risk factor for major haemorrhage in patients treated with vitamin K antagonists: Results of the SCORE prospective cohort

INTRODUCTION

Risk scores for the prediction of haemorrhage are poorly predictive of major bleeding. The aim of this study was to refine the estimation of bleeding risk by identifying one or several parameters of prognostic significance among these algorithms.

MATERIALS AND METHODS

The SCORE study was a prospective, multicentre cohort study conducted in France in 2009-2010. Patients were eligible if they had received vitamin K antagonist (VKA) for any therapeutic indication for at least 3months. The primary outcome was the occurrence of major bleeding at 1-year follow-up.

RESULTS

In total, 962 patients were included in this study and evaluated at 1year. The incidence of major bleeding at 1-year follow-up (Kaplan-Meier method) was 2.9% [95% confidence interval (CI) 1.9-4.2]. The rate of major bleeding was 8.2% (95 CI 3.4-16.2) per year in patients classified as high risk by at least four scores. In a multivariate Cox analysis, of the risk factors for the different scores, only anaemia <100g/l at inclusion was strongly associated with risk of major bleeding (hazard ratio 6.1, 95% CI 2.7-13.8, P<0.0001). Through an induction tree analysis performed to identify a common parameter in the majority of scores, anaemia was found to be the main predictor of correct classification as high risk by at least four scores (55% of patients classified as high risk by at least four scores vs 3.3% in the absence of anaemia).

CONCLUSION

Anaemia with haemoglobin <100g/l is the most important predictor of high risk of bleeding in patients treated with VKA.