ABSTRACT
Employing equilibrium-gated radionuclide ventriculography in the left anterior oblique view, six geometric models and five mathematic coefficients of nonuniformity in regional left ventricular emptying were tested for their relative mortality risk-stratifying power and capacity to augment the risk-discriminating potency of the continuous and dichotomized global ejection fraction. Radionuclide ventriculography was performed an average of 7.6 days after acute myocardial infarction. All geometric models significantly separated 20 normal subjects from 137 patients with recent infarction (p less than 0.001). Cumulative mortality data demonstrated that significant independent univariate dichotomizing potency and augmentation of the mortality risk-discriminating power of the global ejection fraction were provided by models of regional emptying that 1) conformed to coronary artery perfusion areas, 2) encompassed total ventricular counts, 3) expressed variability in regional relative to global ejection fraction, and 4) simulated a pattern of emptying directed toward the center of geometry of the left ventricle. The combination of a four quadrant geometric model with axes drawn 45 degrees above the horizontal and a coefficient of variation calculated as square root of sigma(GEF - REF)2/4 x 100/GEF (where GEF = global ejection fraction and REF = regional ejection fraction) proved to be optimal. This coefficient averaged 12.2% in normal subjects and 32.2% in patients with recent acute myocardial infarction (p less than 0.001).(ABSTRACT TRUNCATED AT 250 WORDS)