The mechanism of resting or nocturnal angina relative to primary alterations in myocardial oxygen supply or demand (MVO2) is still uncertain. We studied 23 patients (15 males, eight females, ages 45–83 years) with coronary artery disease (CAD) rest angina using electrocardiographic hemodynamic monitoring by intra-arterial line Swan-Ganz thermodilution catheter. Data were recorded continuously on magnetic tape for 11–43 hours. Cardiac index (CI) was measured every 2 hours at onset of after relief of pain. In the absence of pain during sleep, mean arterial pressure (MAP) heart rate fell significantly (p ≤ 0.01), with minor decreases in right atrial pressure, pulmonary artery systolic (PASP) pulmonary artery diastolic (PADP) or wedge pressures. CI stroke index (SI) varied only slightly. Five patients did not develop pain during study; seven others experienced episodes of pain acceptable as ischemic, but not associated with significant electrocardiographic or hemodynamic changes, showed poor, delayed or variable responses to nitroglycerin (NTG). In the remaining 11 patients, reproducible episodes of pain were always preceded (average 8 minutes, range 3–22 minutes) by ischemic electrocardiographic changes increases in PASP (25.0 ± 1.1 to 47.6 ± 4.0 mm Hg, p ≤ 0.001), in PADP (10.9 ± 0.8 to 26.6±1.1 mm Hg, p ≤ 0.001) with decreases in SI (32.1 ± 2.5 to 27.7 ± 1.9 ml/beat/m2, p ≤ 0.001). All 11 patients responded promptly predictably to NTG. Two patients had transient hemodynamic ischemic electrocardiographic changes without pain, usually at night. In seven of 11 patients, PASP was the first hemodynamic variable to change; the remaining four had concomitant increases in MAP. Subsequent increases in MAP occurred in six patients with no change or a decrease in levels with the onset of pain in two. In three patients, MAP increased during some episodes, remained unchanged in some, decreased in others, especially in those associated with ST-segment elevation during pain. At the onset of electrocardiographic hemodynamic changes before the onset of pain, the double product (heart rate × systolic arterial pressure) was unchanged in seven patients increased slightly in four. The overall data — indicating the development of acute left ventricular failure associated with ischemic electrocardiographic changes, but without significant alterations in the major determinants of MVO2 at the onset of the ischemic episode — raise the possibility of a spontaneous reduction in myocardial perfusion as a primary mechanism of rest angina in many patients with CAD.