Identification of viable myocardium is possible with 18F-fluorodeoxyglucose (FDG) and positron emission tomography (PET). More recently, the feasibility of cardiac FDG imaging with single photon emission tomography (SPET) has been reported. In patients with diabetes mellitus, poor image quality is frequently obtained with FDG-PET, due to relative or absolute insulin deficiency and peripheral insulin resistance. To improve image quality of the FDG-PET studies, the hyperinsulinaemic glucose clamp has been proposed. We assessed the image quality of cardiac FDG-SPET studies in 10 patients with non-insulin-dependent diabetes mellitus (NIDDM) and compared the results with those obtained in 10 patients without NIDDM. All FDG studies were performed during hyperinsulinaemic glucose clamping. Image quality was expressed as myocardial to blood pool activity (M/B) ratios. Residual viability was assessed in dysfunctional myocardium. The M/B ratios were comparable between patients with and without NIDDM (2.67 +/- 0.8 vs 2.50 +/- 0.7, N.S.). Residual viability was detected in 51% of the dysfunctional segments of the patients with NIDDM and in 49% of the segments of the patients without NIDDM. In the small subset of patients (n = 10) undergoing revascularization, 19 of 20 (95%) segments that had improved wall motion were viable on FDG-SPET. In contrast, 27 of 36 (75%) segments that did not improve were necrotic on FDG-SPET. Thus FDG-SPET during hyperinsulinaemic glucose clamping provides adequate image quality in patients with NIDDM compared with patients without NIDDM, and can be used in the detection of viable myocardium.