Any bleeding complication
Measured Factor Detail
This calculator is a model that predicts the 90-day bleeding risk in patients who are on Warfarin therapy. The model takes into consideration three variables, age 60 years or older, and female sex. A score of 0, 1.3-3, and greater than 3 represented low, intermediate or a high risk of bleeding, respectively. This calculator helps identify a subgroup of patients who have a high risk of developing any bleeding complications during anticoagulation therapy.
bleeding risk score = [1.6xAge] + [1.3xSex] + [2.2xMalignancy]
Measured Factor Low Impact
Measured Factor High Impact
- Critical High: > 3 points
- Normal: 0 point
- Normal Adult Male: 0 point
- Normal Adult Female: 0 point
- Normal Geriatric Male: 0 point
- Normal Geriatric Female: 0 point
This model for predicting the risk of bleeding has not been validated in large clinical studies. Additionally, it does not take into consideration important risk factors such as anemia, history of bleeding, or the use of antiplatelet agents. As a result, it is not recommended for clinical use at this time.
Study Validation 1
A multivariate analyses aimed at investigating different predictors of bleeding in a cohort of anticoagulated patients and evaluated the predictive value of several bleeding risk stratification schemes. Study population consisted of a cohort of 7,329 patients with atrial fibrillation (AF) participating in the SPORTIF (Stroke Prevention Using an Oral Thrombin Inhibitor in Atrial Fibrillation) III and V clinical trials. All patients were anticoagulated orally with adjusted-dose warfarin or fixed-dose ximelagatran. Major bleeding was defined as fatal or clinically overt bleeding associated with transfusions of ≥ 2U of blood or ≥20 g/l decrease in hemoglobin or bleeding involving a critical anatomic site other than the brain parenchyma. The predictive value of several bleeding risk schemes were tested and these included; HEMORR2HAGES, HAS-BLED, Kuijer et al., Beyth et al., and Shireman at al. From the analyses, significant predictors of bleeding included aspirin use, renal impairment, age 75 years or older, diabetes and heart failure or left ventricular dysfunction. Of the tested schemas, the HAS-BLED score performed the best (p<0.0001). Among all the schemes, the HAS-BLED score had a marginally better c-statistic value or goodness of fit value (0.67) than all others and the Kuijer et al. exhibited the lowest c-statistic value (0.50). The Kuijer et al. stratification scheme had 11, 120 and 5 patients in the low, moderate and high risk bleeding groups, respectively (95% confidence interval of 0.48-0.56). Researches concluded that of the previously published schemas for risk of bleeding the HAS-BLED scheme offers the most useful predictive capacity.
Study Validation 2
This systematic review article evaluated and reviewed the clinical prediction rules (CPRs) available for estimating the bleeding risk in patients starting warfarin therapy. A systematic review of PubMed was performed and a total of four studies were included in this review article. The primary outcome measured was the ability to distinguish between patients at high risk and low risk of experiencing major bleeding while on warfarin therapy. The predictive ability of each CPR was defined in terms of likelihood ratios (LR). Various CPRs were identified, including the Kuijer et al. which takes into consideration age, sex and known malignancy in order to determine the odds ratio for bleeding in patients receiving a three-month course of anticoagulation for treatment of venous thromboembolism (VTE) or pulmonary embolism (PE). Results found that 17% of patients classified as being at low risk for a bleed experienced no bleeds, while 26% of those classified as high risk experienced a bleeding complication. The modified Outpatient Bleeding Risk Index exhibited moderate predictive ability for major bleeding in two studies. The HEMORRA2HAGES prediction tool was also reviewed and found to have greater predictive accuracy than the others (p<0.001). Despite these results and others, none of the available CPRs exhibited sufficient predictive accuracy or have trials evaluating the impact of their use on patient outcomes. As a result, no existing CPR can be recommended for widespread use in practice.
Study Additional 1
This review article described the risk of bleeding in patients taking anticoagulants used for the prevention and treatment of thromboembolism. Goals of this article includes reviewing the epidemiology and individual risk factors for anticoagulation-related hemorrhagic complications, compare features of various risk tools, and highlight situations where bleeding risk tools may be the most useful. Analyses of clinical studies have shown that vitamin K antagonist (VKA) treatment in atrial fibrillation (AF) increases the risk of major bleeding by 0.3-0.5% per year. Of note, risk factors for major hemorrhage includes history of congestive heart failure, advanced age, hepatic or renal disease, diabetes mellitus and cerebrovascular disease. A variety of bleeding risk assessment tools were evaluated, including the bleeding risk assessment tool developed by Kuijer et al, HEMORR2HAGES, Outpatient Bleeding Risk Index (OBRI), ATRIA and HAS-BLED among others. Each risk scheme was developed in very different patient populations or clinical settings; therefore, the risk scores vary widely. Additionally, validation studies for these schemes have shown that these risk schemes are modestly predictive in patients with AF. The bleeding risk assessment tool developed by Kuijer et al found 170 patients (22%), 460 patients (59%), 150 patients (19%) having low, intermediate, and high risk for bleeding. No single risk score was consistently superior to others in predicting bleeding complications. Despite this, the article highlights there are certain situations when bleeding risk tools are useful in clinical practice. For the most part, bleeding risk tools seem to be the most useful for patients at the lower end of thrombotic risk, where the net benefits of anticoagulation are smaller and the risk of bleeding may be more influential. Bleeding risk tools can also be useful in identifying low-risk groups of patients who can be reassured that they are unlikely to have significant bleeding complications. In conclusion, the article recognizes that tools to predict bleeding risk can provide valuable information about potential risks, however none of the available risk tools are highly predictive and able to effectively predict hemorrhage.
Study Additional 2
This retrospective study determined whether the concomitant use of warfarin and antidepressants increase the risk of bleeding outcomes compared with the use of warfarin alone. Additionally, the study aimed to characterize the risk of bleeding in warfarin-treated patients taking selective serotonin reuptake inhibitors (SSRIs). A chart review was performed that looked at records over 6 months with focus on international normalized ratio values (INR), medical history, bleeding type and incidence, and hospitalizations due to bleeding. Eligible patients for the antidepressant group included those taking warfarin and antidepressants consistently for 6 months, while the control group consisted of patients not taking an antidepressant with warfarin. Bleeding risk was assessed by calculating the Beyth Outpatient Bleeding Risk Index and the Bleeding Risk Score developed by Kuijer et al. The Beyth index provides a risk of bleeding based on age, history of gastrointestinal (GI) bleed, history of stroke, recent myocardial infarction, low hematocrit value, low creatinine value, or a diagnosis of diabetes. The Bleeding Risk Score developed by Kuijer et al provides an estimate of minor and major bleeding depending on age, sex and known malignancy. The Kuijer index was used to compare the differences in bleeding risk between the 2 groups. The Beyth index was used to compare the risk of major bleeding and identify clinical variables related to bleeding outcomes. Results showed that there was no significant difference between treatment groups for risk of any bleeding as measured by the Kuijer index. However, patients in the antidepressant group had a higher risk for major bleeding as calculated by the Beyth index (p<0.01). Overall, the use of any antidepressant with warfarin did not show an association with a higher incidence of bleeding during the 6-month period. However, the use of an SSRI with warfarin was associated with an increase in bleeding (P=0.04).
Study Additional 3
This review article described different bleeding risk stratification schemes, including the Outpatient Bleeding Risk Index (OBRI), Kuijer Bleeding Risk Index and the Shireman model. The OBRI assigns one point for each of four variables; age, history of stroke, history of gastrointestinal bleed (GI), and recent myocardial infarction, severe anemia, diabetes or renal impairment. Although proven to be a good predictor of bleeding, a recent larger study showed that it is not helpful in identifying patients at high risk for major bleeding. The Kuijer index takes into consideration only 3 variables; age, female sex and the presence of malignancy in order to predict the 90-day bleeding risk of 1% in low-risk patients and 7% in high-risk patients. This index, however, was not validated in a larger study and does not include important risk factors such as anemia, history of bleeding, or use of other antiplatelet agents. Therefore, it is not recommended for clinical use. Lastly, and more recently, the Shireman model takes into consideration 8 clinical variables and identifies patients at low, intermediate and high risk of major bleeding within 90 days of hospital discharge. This model has good face validity and was developed and validated in a large group of patients. However, this model is complex to calculate and is limited to patients that are 65 years or older with atrial fibrillation. Overall, the OBRI and the Shireman model can be used confidently in combination with other predictors in order to help make decisions, but unfortunately the Kuijer index is not recommended.