Predictive model may help estimate bleeding risk with aspirin use

1. A model was constructed to predict bleeding risk in patients taking aspirin to prevent cardiovascular disease.

2. The predicted data closely match the observed data, supporting the use of this model in future studies and potentially in clinical decision making.

Evidence Rating Level: 2 (Good)

Study Rundown: In 2016, the U.S. Preventative Services Task Force recommending the use of low-dose aspirin in 50- to 59-year-old adults at risk for cardiovascular disease (CVD) or colorectal cancer.  However, they included the important warning that aspirin is not appropriate for patients deemed to be at high risk for bleeding.  In this study, a team from New Zealand developed and evaluated a model to predict the risk of bleeding.  The associated risks of individual demographic factors and clinical measurements were estimated within the study cohort, and then weighted to construct the overall best-fit predictive model.  The final models for men and women closely fit the observed data.  Before it can be applied on a global scale, this clinical tool will need to be validated in other countries with different risk factors and healthcare systems.

Click to read the study in Annals of Internal Medicine

Click to read an accompanying study in Annals of Internal Medicine 

Relevant Reading: Aspirin Use for the Primary Prevention of Cardiovascular Disease and Colorectal Cancer: U.S. Preventive Services Task Force Recommendation Statement

In-Depth [prospective cohort]:  This study developed a predictive model for bleeding risk based on a population of 385,191 patients at New Zealand primary care centers.  All of these patients had a CVD risk greater than 10% based on an established CVD risk model called PREDICT.  Patients were excluded if they already had a known indication or contraindication for aspirin.  Dozens of predictive factors for bleeding risk were drawn from previously published cohort studies and meta-analyses, including basic demographics, systolic blood pressure, lipid labs, and a history of bleeding, peptic ulcer disease, and cancer.  Several key variables like BMI, hemoglobin levels, and platelet counts were excluded because of incomplete data.  All established risk factors were associated with bleeding risk in this study population, but many were not statistically significant.  The final models for men and women constructed using Cox proportional hazards modeling using data from the entire study population.  The slopes of the regression lines were 1.00 (CI, 0.92 to 1.08) and 0.96 (CI 0.90 to 1.02) for men and women, respectively.  Model performance was verified using split-sample model validation.

Image: PD

©2019 2 Minute Medicine, Inc. All rights reserved. No works may be reproduced without expressed written consent from 2 Minute Medicine, Inc. Inquire about licensing here. No article should be construed as medical advice and is not intended as such by the authors or by 2 Minute Medicine, Inc.