Reversible perfusion defects on ^99mTc-sestamibi imaging during hyperemia are thought to occur due to myocardial blood flow (MBF) "mismatch" between regions with and without stenosis. We have recently shown that myocardial blood volume (MBV) distal to a stenosis decreases during hyperemia, resulting in a reversible perfusion defect on myocardial contrast echocardiography (MCE). In this study, we hypothesized that a reversible perfusion defect on ^99mTc-sestamibi imaging during hyperemia results from the same mechanism. We tested our hypothesis under the following conditions: 1) increases in MBF in the absence of changes in MBV by using direct intracoronary infusion of adenosine (group I, n = 10 dogs); 2) decrease in MBV despite an increase in MBF by left main infusion of adenosine proximal to a noncritical coronary stenosis placed on either coronary artery (group II, n = 13 dogs); and 3) reduction in both resting MBF and MBV by placement of a severe stenosis (group III, n = 7 dogs). In group I dogs, no difference in MBV or ^99mTc-sestamibi uptake was found between the two coronary beds despite an up to fourfold increase in MBF in one bed with adenosine. In group II dogs, MBV distal to the stenosis decreased during hyperemia despite a twofold increase in mean MBF. A good correlation was found between ^99mTc-sestamibi uptake and MBV ratios from the stenosed versus normal bed (r = 0.91, P < 0.001). In group III dogs, both MBF and MBV were decreased in the stenosed bed at rest with a good correlation noted between ^99mTc-sestamibi uptake and MBV ratios from the stenosed versus normal bed (r = 0.92, P = 0.004). We conclude that reversible defects on ^99mTc-sestamibi during vasodilator stress imaging are related to decreases in MBV distal to a stenosis and not to "flow mismatch" between beds. The decrease in MBV results in reduced ^99mTc-sestamibi uptake during hyperemia.