A new model to predict the risk of major gastrointestinal bleeding in patients on direct oral anticoagulants (dabigatran and rivaroxaban)

Aims The aim of this study was to identify factors associated with gastrointestinal bleeding (GIB) in patients on direct oral anticoagulants (DOACs) and develop a risk score that would provide an effective tool for the clinical assessment of GIB. Methods This was a multicentre retrospective analysis of clinical and follow-up data of patients treated with DOACs. The score was developed through logistic regression. The performance of score was evaluated using the area under the receiver operating characteristic curve (AUC), sensitivity, specificity and Hosmer–Lemeshow test. Results The 11 903 patients had a mean age of 65.1 years. In multivariate analysis, age ≥65 years, alcohol use, history of peptic ulcer, history of major bleeding, abnormal liver function or renal function, cancer, platelet count <100 × 109/L, anaemia, and concurrent antiplatelet agent or non-steroidal anti-inflammatory drug treatment were independent risk factors for GIB, and concurrent treatment with gastrointestinal protective agents were a protective factor. The Alfalfa-DOAC-GIB score was constructed using these 12 factors. The AUC of the Alfalfa-DOAC-GIB score was 0.77 (95% CI 0.74–0.81), which was higher than that of the HAS-BLED score (0.69; 95% CI 0.65–0.72) and the New score (0.65; 95% CI 0.61–0.68). Conclusions Based on 12 factors, we developed a gastrointestinal bleeding risk score. The newly developed Alfalfa-DOAC-GIB score has better predictive value than the HAS-BLED score and the New score, and might be an effective tool to help reduce the occurrence of GIB in patients using DOACs.