Risk Prediction Tool for Assessing the Probability of Death or Myocardial Infarction in Patients With Stable Coronary Artery Disease
- Additional Document Info
- View All
Several risk scores in acute coronary syndromes are available, but few models exist for stable coronary artery disease to guide decision-making and prognosis. A multivariate model was developed using 23 baseline candidate variables from the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Therapy EvaluationTrial (n = 2,287 patients). Discrimination of the model was evaluated by the concordance c-index. The procedure was validated using 100 random half samples. We identified 9 independent predictors of death or myocardial infarction (MI) during a 5-year follow-up. The following predictors and points contributing to the risk score were: heart failure (3), number of diseased coronary arteries (1 for each vessel), diabetes (1), age (1 for each 15 years ≥ age 45), previous revascularization (1), current smoking (1), female (1), previous MI (1), and high-density lipoprotein cholesterol (1: 31 to 40 mg/dL; 2: <30 mg/dL). The risk tool had a potential range from 0 to 15, corresponding to 5-year event rates of 5.8% to 56%. C-indices ranged from 0.67 for the full data set to 0.62 for the validating subsamples. Respective observed versus predicted 5-year event rates for 3 predefined risk strata revealed: 30% had a low-risk score of 0 to 3 (9.3% vs 9.3%, or 1.9%/year); 59% had an intermediate-risk score of 4-6 (18.0% vs 18.1%, or 3.6%/year); and 11% had a high-risk score of 7-11 (36% vs 36.5%, or 7.2%/year). This stable coronary artery disease risk score permitted a prognostic assessment of 5-year probability of death or MI with an approximate 4-fold range in event rates from the lowest (9.3%) to the highest (36%) terciles, thus enabling better clinical practice decisions that allow physicians to tailor the intensity of treatment to the level of risk.
has subject area