Survival in medically treated coronary artery disease.

In 1214 symptomatic medically treated patients with coronary artery disease, 57 noninvasive baseline clinical characteristics and 24 catheterization descriptors were analyzed by a multivariable analysis technique to determine the characteristics that were independent predictors of survival and, in particular, to determine whether noninvasive characteristics contributed prognostic information in addition to catheterization findings. When the noninvasive characteristics were analyzed, 31 characteristics were significant (p < 0.05) univariate predictors of survival, but only 12 contained significant independent prognostic information. Fiveand 7-year survival rates in 197 patients who had none of the independently significant noninvasive characteristics were both 90%. Nineteen variables were significant when the catheterization descriptors were analyzed individually. Only seven were independently significant when they were analyzed jointly. When all 81 baseline characteristics were analyzed jointly, seven noninvasive characteristics (history of peripheral vascular disease, New York Heart Association class IV heart failure, nonspecific intraventricular conduction defect, progressive chest pain, nocturnal pain, premature ventricular complexes on the resting ECG, and left bundle branch block) and six invasive characteristics (left-main stenosis, arteriovenous oxygen difference, number of diseased vessels, abnormal left ventricular contraction, left ventricular end-diastolic pressure and anterior asynergy) were independently significant. Different survival rates may occur in subsets that are uniform with respect to only one or two important characteristics (e.g., coronary anatomy and ventricular function) because of differences in other important baseline characteristics. Both noninvasive and invasive characteristics must be taken into account to define prognosis in coronary disease fully.