ORBIT Bleeding Risk Score for Atrial Fibrillation

Study Questions:

Can routinely available clinical information be used to predict bleeding risk in anticoagulated patients with atrial fibrillation (AF)?


The authors used data from the prospective Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT-AF) of incident and prevalent AF patients at 176 US sites. Cox proportional hazards were used to identify all factors (12) independently associated with major bleeding among patients taking oral anticoagulants (OACs) over a median follow-up of 2 years. A simplified numerical bedside risk score of five predictive factors was developed. The full and simplified models were compared to two existing bleeding risk scores (HAS-BLED and ATRIA) in the ORBIT-AF population and the external ROCKET-AF clinical trial population.


Among 7,411 ORBIT-AF patients taking OACs, the rate of major bleeding was 4.0/100 person-years. The 12-variable model exhibited modest predictive ability for major bleeding (C-index 0.69). The simplified five-variable model, which includes one point for older age (75+ years), two points for anemia, two points for bleeding history, one point for renal insufficiency, and one point for concurrent antiplatelet use, also had modest clinical predictive ability for major bleeding (C-index 0.57). Both the full and simplified ORBIT models have similar discrimination, but better calibration for major bleeding risk prediction as compared to the HAS-BLED and ATRIA scores in the external validation cohort.


The authors concluded that the simple five-item ORBIT bleeding risk model has better predictive ability for major bleeding among anticoagulated AF patients when compared to the HAS-BLED and ATRIA scores.


While clinical prediction scores are commonly used to estimate stroke risk and determine treatment regimens among AF patients, use of bleeding risk scores has not been widely adopted. This is primarily limited by their relatively poor predictive ability and the lack of meaningful integration in clinical decision making. The ORBIT bleeding risk score does not significantly improve the discriminatory activity as compared to the existing HAS-BLED and ATRIA scores, but it does have better major bleeding risk calibration. However, until prospective data demonstrate a role of bleeding risk scores in clinical decision making, clinicians should remain cautious about using bleeding risk scores to determine when OAC therapy is appropriate. Instead, using the ORBIT bleeding risk score (or other bleeding risk scores) to screen for reversible bleeding risk factors, such as concurrent antiplatelet use, is very appropriate given currently available data.

< Back to Listings