Biomarker-Based ABC-Bleeding Risk Score for Atrial Fibrillation

Study Questions:

Can a new risk score for major bleeding be developed using biomarkers in patients with atrial fibrillation (AF)?


Using the 14,537-patient ARISTOTLE trial (apixaban vs. warfarin in AF), a new biomarker-based risk score for major bleeding was developed and internally validated. The score was then externally validated using the 8,468-patient RE-LY trial (dabigatran vs. warfarin in AF). All major bleeding events were centrally adjudicated and the predictive values of biomarkers and clinical variables were assessed using Cox regression models.


The most important biomarker predictors for major bleeding were growth differentiation factor-15 (GDF-15), high-sensitivity cardiac troponin T (cTnT-hs), and hemoglobin. Additionally, age and a prior history of bleeding were also important predictors of major bleeding events. The ABC-bleeding score (age, biomarkers [GDF-15, cTnT-hs, and hemoglobin], and clinical history [prior bleeding]) demonstrated a higher c-index than the conventional HAS-BLED and newer ORBIT scores for major bleeding (0.68; 95% confidence interval [CI], 0.66-0.70 vs. 0.61; 95% CI, 0.59-0.63 vs. 0.65; 95% CI, 0.62-0.67, respectively) in the derivation cohort. A higher c-index was also found for the ABC-bleeding score (0.71; 95% CI, 0.68-0.73) in the external validation cohort as compared to the HAS-BLED (0.62; 95% CI, 0.59-0.64) and ORBIT (0.68; 95% CI, 0.65-0.70) bleeding risk scores.


The authors concluded that the use of a novel bleeding risk score that incorporates biomarker data performed better than the HAS-BLED and ORBIT scores, and should be used for decision support on anticoagulation treatment for AF patients.


Predicting bleeding risk has always been more challenging than predicting stroke risk in AF patients. This new bleeding risk score has shown good predictive and discriminatory ability in two large cohorts of patients, outperforming the more widely endorsed HAS-BLED score and the newer ORBIT bleeding risk score. However, the routine availability of GDF-15 and cTnT-hs biomarkers may significantly limit the usability of the ABC-bleeding risk score. Additionally, the need for a nomogram or electronic calculator to calculate major bleeding risk also limits the use of the adoption of this newer score in daily clinical practice. While the ABC-bleeding risk score has been derived and validated to predict major bleeding events, it has yet to be tested in a clinical decision-+making study.

< Back to Listings