OBJECTIVE: This study investigates the efficacy of radionuclide myocardial perfusion imaging (MPI) in the evaluation of cardiac outcome in optimally treated diabetic patients without manifestations of coronary artery disease (CAD) in relation to the inherent clinical risk.METHODS: Follow-up data were collected from 86 diabetic patients who had undergone adenosine stressing MPI. These patients either had no symptoms or had noncardiac chest discomfort, had a normal resting electrocardiogram, had no known CAD or prior positive stress test results, and were receiving currently recommended therapy. Endpoints were cardiac death, myocardial infarction, new-onset heart failure, and CAD diagnosed by angiography at least 2 months from the MPI, irrespective of subsequent revascularization.RESULTS: Twenty-six (30%) diabetic patients had abnormal perfusion and the remaining had a normal scan. Over a median follow-up period of 32.5 months 14 cardiac events occurred. In patients with normal MPI, the annual cardiac event rate was 4.0% compared with 12.2% in those with abnormal MPI (P=0.008). In multivariate analysis, myocardial ischemia (hazard ratio 5.3; P=0.006), obesity (hazard ratio 6.8; P=0.005), the ALFEDIAM/SFC risk (hazard ratio 6.8; P=0.002), and type 2 diabetes (hazard ratio 5.3; P=0.035) were found to be independent predictors of cardiac events. The former two variables remained independent determinants of the outcome, together with peripheral arterial disease, when a different clinical risk classification system was applied. MPI provided incremental prognostic information over both clinical models formed.CONCLUSION: Adenosine MPI can effectively risk-stratify optimally treated diabetic patients without manifestations of CAD. In this subset, clinical variables can also determine the outcome independently, but MPI adds incremental predictability over them.