PET with 15O-labeled water allows noninvasive quantification of myocardial blood flow (MBF) at baseline and during pharmacologically induced hyperemia to assess the coronary vasodilator reserve (CVR = hyperemic/baseline MBF). Despite widespread use of PET, its reproducibility during one study session has not been tested. Intravenous adenosine (Ado), a powerful coronary vasodilator with a very short decay time, is commonly used for the induction of hyperemia. However, it is not known whether Ado can induce tachyphylaxis after short-term repetitive administration. In this study, we aimed to test the reproducibility of PET assessment of CVR during Ado-induced hyperemia. METHODS In 21 healthy volunteer men, baseline and Ado MBF were measured twice using PET with 15O-labeled water to obtain two CVR assessments within 1 h. RESULTS There was no significant difference between the two baselines (0.89 +/- 0.14 versus 0.99 +/- 0.15 mL/min/g, mean difference 13% +/- 11%) or between the two hyperemic MBFs (3.51 +/- 0.45 versus 3.83 +/- 0.49 mL/min/g, mean difference 10% +/- 14%), resulting in comparable values of CVR (4.05 +/- 0.75 versus 3.93 +/- 0.72, mean difference 2% +/- 15%). The repeatability coefficient for MBF was 0.17 mL/min/g at baseline and 0.94 mL/min/g during hyperemia. The repeatability coefficient of the rate pressure product (RPP) was lower at baseline (1,304 mm Hg x beat/min) than during hyperemia (3,448 mm Hg x beat/min). CONCLUSIONS Repeated measurements of MBF and CVR during the same study session were not significantly different, demonstrating the validity of the technique. The larger variability of hyperemic flow, as indicated by the larger repeatability coefficient, was paralleled by a greater variability of the RPP. This could mean that the greater variability of MBF during stress is more likely due to a variable response to Ado rather than to a measurement error.