Identification of patients with angina but normal coronary arteriograms (syndrome X) using noninvasive means would be desirable. The ability of four established exercise electrocardiographic methods to identify angina patients with and without coronary artery disease was compared with that of a method based on a combination of the above (combined method). A treadmill score, a multivariate method, the ST segment recovery loop, the ST/heart rate adjustment, and the combined method were applied to 112 patients who had typical exertional angina and positive exercise tests (greater than 1 mm ST segment depression); 90 had documented coronary artery disease and 22 had syndrome X. The combined method and the treadmill score had a significantly higher diagnostic accuracy (both 81%, as 91 of the 112 patients were correctly identified by both methods) than the multivariate (66%) and ST segment recovery loop (64%) methods (p less than 0.05). The ST/heart rate adjustment had a lower sensitivity for syndrome X than any other method (1 of 22). Thus methods that involve the assessment of both ST and non ST segment variables have greater accuracy in separating syndrome X and coronary artery disease patients than methods relying more heavily on ST segment changes.