Effects of exercise tolerance, age, and gender on the specificity of radionuclide angiography: sequential ejection fraction analysis during multistage exercise.

We investigated the factors that determine the left ventricular ejection fraction response to exercise in 57 middle-aged, untrained patients with angiographically normal coronary arteries. The ejection fraction was measured by blood pool radionuclide angiography during each stage of exercise. Stepwise regression analysis was applied to 39 variables, and the resulting significant variables were then included in a logistic regression model to determine which of them would predict a normal ejection fraction response. By stepwise regression analysis, the best model (F = 5.6, p = 0.0004) was obtained by combining four variables: achieving 85% of maximal predicted heart rate, peak heart rate, number of exercise stages performed, and resting pulse pressure. Only the latter two variables were significant by logistic regression analysis. The specificity of greater than or equal to 5 ejection fraction unit increase at peak exercise was 42%, 75%, and 100% in patients who exercised one, two, or three stages, respectively. Furthermore, the increase in ejection fraction during each exercise stage was independent of age or gender. Thus the standard criteria of normality during exercise radionuclide angiography have high specificity only when applied to patients with good exercise performance, regardless of age or gender.