Standard radionuclide determination of left ventricular ejection fraction assumes a constant attenuation throughout the cardiac cycle, implying a proportional relationship between count rate and volume. In this study we examined patients with different contraction patterns of the left ventricle and the impact of the movement of the left ventricular center on calculation of the ejection fraction. Left ventricular depth at end diastole and end systole was assessed in the left lateral projection during a multigated acquisition in 34 patients with and without ventricular wall motion abnormalities due to coronary artery disease. A corrected left ventricular ejection fraction considering changes in the attenuation during the cardiac cycle was calculated and compared with the standard ejection fraction. In patients without impairment of left ventricular wall motion, the left ventricular center moved slightly posteriorly at end systole, and correction for changes in attenuation reduced the radionuclide ejection fractions by 3.0% (95% confidence interval 1.1-4.9%) as compared with the standard calculation. In patients with anterior wall hypokinesia the corrected ejection fractions were 2.7% higher (0.6-4.8%) and in patients with posterior wall hypokinesia 6.5% lower (3.6-9.4%) than the standard ejection fraction. The change in ventricular depth from end diastole and end systole was significantly different in all three patient groups (p < 0.01). These findings suggest that changes in left ventricular depth during contraction result in slight overestimation of the radionuclide multigated left ventricular ejection fraction in patients with normal ventricular function and that wall hypokinesia in some cases may lead to a significant error in the accuracy of the results.