Comparison of left ventricular ejection fractions determined in healthy anesthetized dogs by echocardiography and gated equilibrium radionuclide ventriculography.

Left ventricular ejection fractions (LVEF) of 8 pentobarbital-anesthetized dogs were calculated by gated equilibrium radionuclide ventriculography (RVG) and by M-mode and two-dimensional echocardiography (2-DE) prior to and during constant IV infusion of isoproterenol. Mean LVEF (+/- SD), determined with RVG by use of an automatic edge detection algorithm (RVG-auto) to define the left ventricular region of interest, increased from a resting value of 53.5% (+/- 4.9%) to 71.9% (+/- 6.8%) during isoproterenol infusion. Mean LVEF, determined with RVG by use of visual inspection (RVG-manual) to define the left ventricular region of interest increased from a resting value of 51.6% +/- 3.8% to 67.0% +/- 5.6% during isoproterenol infusion. Using 2-DE and the bullet formula to calculate left ventricular volume (LVV = 5/6 x cross-sectional area x length), mean LVEF increased from 52.3% (+/- 3.50) to 74.7% (+/- 5.0%). Using 2-DE area measurements and Teicholz formula, mean LVEF increased from 48.9% (+/- 5.1%) to 69.5% (+/- 6.0%). Using M-mode echocardiographic left ventricular diameter measurements and Teicholz formula, mean LVEF increased from 52.3 (+/- 9.0%) to 78.3% (+/- 8.1%). Before and during isoproterenol infusion, the mean LVEF values calculated by RVG agreed closely with mean LVEF values calculated from M-mode and 2-DE. Correlation coefficients determined from linear regression analysis of LVEF by echocardiography vs LVEF by radionuclide ventriculography ranged from 0.79 to 0.88. Correlation coefficients were higher and SEM were lower when LVEF was determined by RVG-manual, rather than by RVG-auto methods and when LVEF was calculated from 2-DE measurements, rather than from M-mode measurements.