Evaluation of comorbidity scores to predict all-cause mortality in patients with established coronary artery disease

Background: To assess the value of scores based on the presence of comorbid conditions for mortality risk-stratification in patients with coronary artery disease (CAD). Methods: We prospectively followed 305 males with CAD undergoing coronary angiography for 58 months. We correlated the modified Charlson Index (MCI) and the recently proposed CAD-specific index (CSI) with the risk of all-cause mortality. Results: The odds ratio (OR) for death increased by 31% per point increase in the MCI (95% CI = 17-46%; p < 0.0001). The OR for death increased by 16% per point increase in the CSI (95% CI = 8.5-25%; p < 0.0001). In logistic regression models that adjusted for age, left ventricular ejection fraction, and the number of vessels involved with CAD, both the MCI and the CSI were the strongest predictors of mortality according to the chi(2) value for each term, with the MCI having the highest value. The adjusted OR per point increase in the MCI was 1.32 (95% CI = 1. 17-1.48; p < 0.0001); the corresponding adjusted OR per point increase in the CSI was 1.17 (95% CI = 1.09-1.26; p < 0.0001). The model including the MCI had a slightly higher chi(2) value (45.1 vs. 39.1) and area under the receiver operator characteristic curve (0.742 vs. 0.727) than the model including the CSI. Conclusion: The MCI and the newly proposed CSI are powerful tools to predict all-cause mortality in patients with established CAD. Although the CSI was not superior to the MCI, its simplicity might make it useful in populations with a low prevalence of comorbidities not included in this score. (c) 2006 Elsevier Ireland Ltd. All rights reserved.