In-hospital major bleeding during ST-elevation and non-ST-elevation myocardial infarction care: derivation and validation of a model from the ACTION Registry®-GWTG™.

Bleeding, a common complication of acute myocardial infarction (AMI) treatment, is associated with worse outcomes. A contemporary model for major bleeding associated with AMI treatment can stratify patients at elevated risk for bleeding and is needed to risk-adjust AMI practice and outcomes. Using the Acute Coronary Treatment and Intervention Outcomes Network Registry-Get With the Guidelines (ACTION Registry-GWTG) database, an in-hospital major bleeding risk model was developed in a population of patients with ST-segment elevation myocardial infarction and non-ST-segment elevation myocardial infarction. The model used only baseline variables and was developed (n = 72,313) and validated (n = 17,960) in patients with AMI (at 251 United States centers from January 2007 to December 2008). The 12 most statistically and clinically significant variables were incorporated into the final regression model. The calibration plots are shown, and the model discrimination is demonstrated in derivation and validation cohorts, as well as across key subgroups. The rate of major bleeding in the overall population was 10.8%. The 12 factors associated with major bleeding in the model were heart rate, baseline hemoglobin, female gender, baseline serum creatinine, age, electrocardiographic changes, heart failure or shock, diabetes, peripheral artery disease, body weight, systolic blood pressure, and home warfarin use. The risk model discriminated well in the derivation (C-statistic = 0.73) and validation (C-statistic = 0.71) cohorts. A risk score for major bleeding corresponded well with observed bleeding: very low risk (3.9%), low risk (7.3%), moderate risk (16.1%), high risk (29.0%), and very high risk (39.8%). In conclusion, the ACTION Registry-GWTG in-hospital major bleeding model stratifies risk for major bleeding using variables at presentation and enables risk-adjusted bleeding outcomes for quality improvement initiatives and clinical decision making.

Duke Scholars

Author Robin Mathews Medicine, Cardiology

Author Matthew Todd Roe Medicine, Cardiology

Author Karen Patton Alexander Medicine, Cardiology