To investigate the in vivo role of the ATP-sensitive potassium channel (K-ATP) in the coronary arteries, we examined the effects of intravenous (i.v.) glibenclamide (GLB, 0.3, 1.0, and 3.0 mg/kg), a specific K-ATP blocker, in chronically instrumented dogs. Epicardial coronary artery diameter (CoD) and coronary blood flow (CBF) were measured continuously. CoD and CBF oscillated in all 6 dogs after injection of 3 mg/kg GLB. CoD oscillated only slightly, with a decrease of 2.3 +/- 0.4%; CBF showed marked oscillation, with a peak flow rate of 21.9 +/- 2.7 ml/min (+26.6%) and a trough flow rate of 10.3 +/- 2.9 ml/min (-46.3%) (baseline flow rate was 17.8 +/- 2.4 ml/min). GLB 1 mg/kg produced a slight decrease in CBF without oscillation, and at 0.3 mg/kg had almost no effect. These oscillations were not associated with a decrease in myocardial blood flow as measured by the hydrogen gas clearance method. Nicorandil (0.2 mg/kg), cromakalim (20 mu g/kg), and diltiazem (0.2 mg/kg) almost completely suppressed the GLB-induced oscillations, but nitroglycerin (NTG 15 mu g/kg) did not. Thus, oscillation of large and small coronary arteries was induced by GLB and was independent of myocardial ischemia. In addition, these findings suggest that K-ATP has an important in vivo role in modulating large and small coronary artery tone through activation of the voltage-dependent Ca2+ channel.