To determine the individual reproducibility of radionuclide ventriculography over an extended period of time, 33 patients with stable coronary artery disease were studied at rest and during three stages of exercise on two occasions separated by 1 year. The individual interstudy variability of ejection fraction (EF), end-diastolic volume (EDV), end-systolic volume, and cardiac output was determined by calculating the mean and standard deviation of the difference between the individual studies (initial – 1 year). Despite high correlations between an EF measured at study 1 and study 2 of 0.96 at rest and 0.87 during maximal exercise, the individual interstudy difference was 0.01 ± 0.04 and-0.02 ± 0.09, respectively. The correlation of percent change in EF from rest to maximal exercise was 0.49 and the individual interstudy differences was -1.2 ± 19%. Correlations of the EDV were 0.81 at rest and 0.72 during maximal exercise while the individual difference was 0.7 ± 38 and-0.8 ± 49 ml, respectively. Considering two standard deviations as the confidence limits for a true change, an EF change of 8 EF units (0.08) at rest and 18 (0.18) during exercise, and EDV changes of approximately 100 ml are needed in an individual to state with confidence that the observed difference between the two studies are true changes and not the result of technologic variability. Because of the large individual interstudy variability in EF and volume measurement, caution must be taken in assuming that any change over a year is due to more than technique variability.