Prospective Evaluation of a Clinical and Exercise-test Model for the Prediction of left Main Coronary Artery Disease

Publisher Website
Google Scholar
Abstract
In a multivariate logistic regression analysis of data from 508 patients, only two clinical factors, age and typicality of pain, were independently significant predictors of left main coronary artery disease. The resulting multivariate equation was prospectively applied to another 370 patients to derive pre-exercise-test (ETT) probabilities of left main coronary artery disease, and these pre-ETT probabilities were combined with literature-derived like lihood ratios for various ETT findings to derive post-ETT probabilities. This model, which can be displayed in simple graphic form, accurately predicted the probability of left main coronary artery disease when prospectively evaluated in this independent validation set of patients. The likelihood of left main coronary artery disease was 16% when the ETT increased the probability, and 4% when it decreased the probability (p <0.001 ). While 48% of patients had mid-range (5-15%) probabilities of left main coronary artery disease before the ETT, only 24% fell into this range of probabilities after the ETT (p <0.0001 ), as ETT results moved patients into higher and lower probability ranges. Thus, probability of left main coronary artery disease can be calculated from clinical and ETT data with this model. These estimated pre- and post-ETT probabilities of left main coronary artery disease may aid in the selection of patients for noninvasive testing or for cardiac catheterization.