The effects of midazolam on coronary sinus blood flow (CSBF), myocardial oxygen consumption (MVO2), and myocardial lactate balance were investigated in eight patients with stable coronary artery disease undergoing cardiac catheterization. Coronary sinus blood flow was measured by continuous thermodilution. Arterial and coronary sinus blood were analyzed for oxygen and lactate content. The determinants of left ventricular (LV) performance were obtained from the cardiac output measured by thermodilution and from left heart catheterization data. All data were obtained before, and 5 and 15 min after midazolam, 0.2 mg X kg-1 iv. Sleep was induced in all patients after administration of midazolam and persisted throughout the entire study period. Mean aortic and LV end-diastolic pressure were decreased from control values (-15 and -44%, respectively), as well as cardiac index and stroke index (-10 and -15%, respectively). Heart rate increased moderately (+8%), while no change in systemic vascular resistance and maximum velocity of shortening (Vmax) were observed. Midazolam administration was followed by a decrease of CSBF (-24%) and of MVO2 (-26%). Coronary vascular resistance did not change, but coronary sinus oxygen tension increased slightly, suggesting a mild alteration in normal autoregulation. However, no evidence of myocardial ischemia occurred, as judged by the absence of changes in the: 1) ECG, 2) myocardial lactate extraction, and 3) relaxation time constant. These results suggest that midazolam may be used safely in patients with coronary artery disease.