OBJECTIVE: To determine the effect of plasma glucose lowering on coronary circulatory function in type 2 diabetes mellitus. METHODS: Twenty patients with type 2 diabetes and 18 weight-matched controls were studied. At baseline, myocardial blood flow (MBF) was measured with [(13)N]ammonia and positron emission tomography at rest, during cold pressor testing (CPT), and during adenosine hyperaemia. In diabetic patients, MBF and blood chemistry were analysed again after 3 months of glucose-lowering treatment with glyburide and metformin. RESULTS: Although hyperaemic MBF did not differ significantly between the patients and controls (1.81 (0.38) v 1.97 (0.43) ml/min/g; mean (SD)), the CPT-induced MBF increase (DeltaMBF) was significantly less in diabetic patients than in controls (0.07 (0.07) v 0.25 (0.12) ml/min/g; p<0.001). Treatment with glyburide and metformin significantly decreased plasma glucose concentrations from 207 (76) to 134 (52) mg/dl (p<0.001). This decrease in plasma glucose was paralleled by a significant increase in DeltaMBF in response to CPT (0.20 (0.16) from 0.07 (0.07) ml/min/g; p<0.001), which tended to be lower than in controls at baseline (0.20 (0.16) v 0.25 (0.12) ml/min/g; p = NS). The decrease in plasma glucose concentrations correlated significantly with the improvement in DeltaMBF in response to CPT (r = 0.67, p<0.01). CONCLUSIONS: Type 2 diabetes mellitus is associated with abnormal MBF response to CPT, which can be significantly improved by euglycaemic control with glyburide and metformin. The close association between the decrease in plasma glucose concentration and the improvement in coronary vasomotor function in response to CPT suggests a direct adverse effect of raised plasma glucose concentration on diabetes-related coronary vascular disease.