Prognostic value of radionuclide angiography in medically treated patients with coronary artery disease. A comparison with clinical and catheterization variables.

To evaluate the usefulness of multiple measures from rest and exercise radionuclide angiography (RNA) in predicting cardiovascular death and cardiovascular events (death or nonfatal myocardial infarction) and to assess the prognostic usefulness of the RNA relative to clinical and catheterization data, we studied 571 stable patients with symptomatic coronary artery disease who had upright rest/exercise first-pass RNA within 3 months of catheterization and were medically treated. With a median follow-up of 5.4 years, 90 patients have died from cardiovascular causes, and 147 patients have either died or suffered a nonfatal myocardial infarction. Using the Cox regression model and a preselected group of RNA variables, the most important RNA predictor of mortality was exercise ejection fraction (chi 2 = 81, p less than 0.00001). Neither rest ejection fraction nor the change in ejection fraction from rest to exercise contributed additional predictive information. Two other RNA study variables, the change in heart rate from rest to exercise and rest end-diastolic volume index, did contribute additional prognostic information to the exercise ejection fraction (chi 2 = 23, p less than 0.0001). Compared with noninvasive clinical data (history, physical examination, electrocardiogram, and chest radiograph), RNA variables were considerably more predictive of mortality (chi 2 = 71 [clinical variables] versus chi 2 = 104 [RNA]). Remarkably, the strength of the relation of RNA variables with mortality was equivalent to that of the set of catheterization variables previously demonstrated in our large angiographic population to be prognostically important (chi 2 = 104 [RNA] versus chi 2 = 102 [catheterization variables]). The RNA contained 84% of the information provided by clinical and catheterization descriptors combined. Furthermore, the RNA contributed significant additional prognostic information to the clinical and catheterization data (chi 2 = 13.6, p = 0.0035). For cardiovascular events, the relative prognostic usefulness of the RNA was similar, although relations with this outcome were generally weaker. Descriptors from the rest/exercise RNA exhibit a powerful relation with long-term outcomes and can be useful in defining risk, even when clinical and catheterization data are available.

Duke Scholars

Author Kerry L. Lee Biostatistics & Bioinformatics, Division of Biostatistics

Author Daniel Benjamin Mark Medicine, Cardiology