To assess the effectiveness and cost-effectiveness of single photon emission computed tomography (SPECT) myocardial perfusion scintigraphy for the diagnosis and management of angina and myocardial infarction (MI). Major electronic databases. Two reviewers independently extracted data and assessed study quality. A decision tree model was used to model the diagnosis decision and a Markov model was developed for the management of patients with suspected coronary artery disease. Costs for the treatments and interventions within strategies were derived from the literature and expressed in 2001-02 pounds sterling. Quality-adjusted life-year (QALY) weights for the different Markov model states were also obtained from the literature. Twenty-one diagnostic and 46 prognostic studies were included, plus two studies comparing SPECT with electrocardiography (ECG)-gated SPECT and one study comparing SPECT with attenuation-corrected SPECT. The diagnostic values of SPECT were generally higher than those of stress ECG, indicating that SPECT provided a better diagnostic performance. SPECT also provided higher positive and lower negative likelihood ratios than stress ECG but heterogeneity was evident among studies. The subgroup of studies including patients with previous MI tended to report a better diagnostic performance for SPECT than stress ECG, but there were too few studies to assess this reliably. The extent and size of the perfusion defect, and whether reversible or fixed, were important factors in predicting future cardiac events such as cardiac death or non-fatal MI. SPECT may be able to identify lower risk patients for whom coronary angiography (CA) might be avoided. Normal SPECT scans were associated with a benign prognosis and the option of medical rather than invasive management. Four studies of patients post-MI reported SPECT to be valuable in stratifying patients into at-risk groups for further cardiac events. The two studies comparing SPECT with ECG-gated SPECT, one diagnostic and the other prognostic, found in favour of gated SPECT. The study comparing SPECT with attenuation-corrected SPECT reported the latter to be more accurate. The systematic review of economic evaluations indicated that strategies involving SPECT were likely either to be dominant or to produce more QALYs at an acceptable cost. There was less agreement about which of the strategies involving SPECT was optimal. The model suggested that, for low prevalence, the incremental cost per unit of output (true positives diagnosed, accurate diagnosis, QALY) for the move from stress ECG-SPECT-CA and from stress ECG-CA to SPECT-CA might be considered worthwhile. Even after allowing for different values for sensitivity or specificity, the least costly and least effective strategy was stress ECG-SPECT-CA. The sensitivity analysis suggested that the cost-effectiveness of SPECT-CA improved if it was assumed that SPECT results allowed for the adoption of a management strategy without recourse to CA. As the time horizon reduced, the incremental cost per QALY increased (as the cost of initial diagnosis and treatment were not offset by survival and quality of life gains). There was a considerable variability in terms of measurement of outcomes, management, setting and patient characteristics, however the direction of evidence tended to favour SPECT in terms of test sensitivity, although these conclusions are based on a relatively small number of diagnostic studies. SPECT, in a variety of settings and patient populations, provided valuable independent and incremental prognostic information to that provided by stress ECG and/or CA that helped to risk-stratify patients and influence the way in which their condition was managed. However, all of the prognostic studies were observational studies and may be biased by unknown confounding factors. Although the ECG-gated and attenuation-corrected SPECT findings seem promising, it is difficult to draw conclusions from so few studies. Further research is needed on the effectiveness and cost-effectiveness, diagnostically and prognostically, of (a) gated and attenuation-corrected SPECT compared with standard SPECT, (b) standard SPECT compared with stress echocardiography and (c) the uncertainty surrounding the results presented in the cost-effectiveness analysis.