In patients with limited exercise capacity and (relative) contraindications to direct vasodilators such as dipyridamole or adenosine, dobutamine stress nuclear myocardial perfusion imaging (DSMPI) represents an alternative, exercise-independent stress modality for the detection of coronary artery disease (CAD). Nondiagnostic test results (absence of reversible perfusion defects with submaximal stress) do occur in approximately 10% of patients. Serious side effects during DSMPI are rare, with no death, myocardial infarction or ventricular fibrillation reported in three DSMPI safety reports for a total of 2,574 patients. On the basis of a total number of 1,014 patients reported in 20 studies, the sensitivity, specificity and accuracy of the test for the detection of CAD were 88%, 74% and 84%, respectively. Mean sensitivities for one-, two- and three-vessel disease were 84%, 95% and 100%, respectively. The sensitivity for detection of left circumflex CAD (50%) was lower, compared with that for left anterior descending CAD (68%) and right CAD (88%). The sensitivity of predicting multivessel disease by multiregion perfusion abnormalities varied widely, from 44% to 89%, although specificity was excellent in all studies (89% to 94%). In direct diagnostic comparisons, DSMPI was more sensitive, but less specific, than dobutamine stress echocardiography and comparable with direct vasodilator myocardial perfusion imaging. In the largest prognostic study, patients with a normal DSMPI study had an annual hard event rate less than 1%. An ischemic scan pattern provided independent prognostic value, with a direct relationship between the extent and severity of the perfusion defects and prognosis. In conclusion, DSMPI seems a safe and useful nonexercise-dependent stress modality to detect CAD and assess prognosis.