Global ejection fraction (EF) from planar gated blood-pool (GBP) imaging is a widely accepted measure of cardiac function. It has been suggested that planar GBP could be replaced by SPECT. In this article, we compare counts-based global EF measured from SPECT and planar images and investigate reasons for discrepancies between the two. METHODS Twenty-three subjects were imaged with both planar and SPECT GBP. SPECT short-axis slices were projected to create reprojected images. Reprojected SPECT (rSPECT) images were created in both the true long-axis view and also in a view typical of planar studies (found to be 60 degrees from the true long-axis). Thus, angle of view effects on global EF could be investigated. In addition, we studied the effects of background and attenuation. RESULTS Long-axis rSPECT EF correlated well with planar EF (r = 0.89) but EF values were significantly higher for rSPECT than for planar (slope = 1.4, intercept = -8 EF units; p < 0.001). We found that background correction may not be necessary with rSPECT, but neither background nor attenuation explained the observed discrepancy between rSPECT and planar EFs. This discrepancy was found to be caused by atrial overlap in the planar image and disappeared when the SPECT slices were reprojected at the same angle of view as the planar images. CONCLUSIONS Global EF can be easily measured from rSPECT GBP images. Long-axis rSPECT EFs are, however, greater than planar EFs by a factor of 1.4 because atrial overlap causes a significant drop in planar EF in planar images. These results suggest that (long-axis) rSPECT EFs may be more accurate than planar EFs.