Atrial pacing has been used to assess the physiologic impact of coronary artery disease (CAD). Several variables have served as markers of pacing-induced myocardial ischemia, but their specificities and sensitivities are unknown. Accordingly, in 28 patients, incremental atrial pacing was performed. Of the 28, 10 had no CAD. The left ventricular ejection fraction (LVEF) (by gated equilibrium blood pool scintigraphy) increased in this group (0.60 +/- 0.11 [mean +/- standard deviation] before pacing to 0.67 +/- 0.13 at peak-pacing, p = 0.002). In no patient did left ventricular end-diastolic pressure increase by greater than 5 mm Hg. No patient had lactate production, and 2 (20%) had electrocardiographic S-T segment depression greater than or equal to 0.1 mV. Four (40%) had chest pain with atrial pacing. In the remaining 18 patients with CAD, atrial pacing caused a decrease in LVEF greater than or equal to 0.05 (0.46 +/- 0.10 to 0.33 +/- 0.09, p less than 0.001) and new segmental wall motion abnormalities in all, indicating pacing-induced myocardial ischemia. Only 8 (44%) had an increase in left ventricular end-diastolic pressure of greater than 5 mm Hg, and only 9 (50%) had lactate production. Ten (56%) had ischemic electrocardiographic changes, and 12 (67%) had chest pain. Thus, the electrocardiographic, metabolic, and hemodynamic alterations that may accompany pacing-induced ischemia are specific but relatively insensitive markers of ischemia. In contrast, chest pain during atrial pacing is a nonspecific occurrence, appearing with similar frequency in normal subjects and patients with CAD and pacing-induced ischemia.