Abstract 16443: Development of a New Risk Stratification Scheme to Predict Warfarin-Associated Hemorrhage: the AnTicoagulation and Risk Factors In Atrial Fibrillation (ATRIA) Study
Background: Warfarin is highly effective in preventing atrial fibrillation-related thromboembolism but significantly raises hemorrhage risk. While risk schemes for stroke are widely employed to guide the anticoagulation decision, hemorrhage risk is less objectively quantified. Our goal was to develop and validate a new risk stratification scheme to predict warfarin-associated hemorrhage in patients with atrial fibrillation.
Methods: We followed 9,186 patients with atrial fibrillation contributing 32,888 person-years of follow-up on warfarin as part of the Anticoagulation and Risk factors in Atrial Fibrillation (ATRIA) cohort. Clinical data and incident hospitalizations for major hemorrhage were obtained from databases and hemorrhage events validated through medical record review. We developed a risk index to predict warfarin-associated hemorrhage using Cox regression with time-varying predictors. We selected from among demographic, clinical, pharmacy, and laboratory variables using bootstrapping approaches, then internally validated the model via split sample testing.
Results: We observed 461 first major hemorrhages during follow-up (1.4% per year). Five independent variables were identified in the final risk model and assigned weights according to regression coefficients: anemia, severe renal disease (e.g., glomerular filtration rate < 30 ml/min or dialysis-dependent), age >= 75, prior hospitalization for bleeding, and hypertension (Table). The c-statistic for the continuous risk score was 0.74. Collapsing points into a 3-category risk index, the major hemorrhage rate was 0.8% in the low risk group (0–3 points), 2.6% in medium risk (4 points), and 5.8% in high risk (5–10 points).
Conclusions: A simple 5 variable risk score was effective in quantifying the risk of warfarin-associated hemorrhage in a large, community-based cohort of patients with atrial fibrillation and can be used to estimate risks associated with anticoagulation.
- © 2010 by American Heart Association, Inc.