Eighty-five subjects with stable angina pectoris and proven obstructive coronary disease were followed prospectively (mean follow-up 4.2 ± 2.0 years) to assess the value of various predictors of longevity. After patients with congestive heart failure, hypertension, left bundle branch block, valvular heart disease or recent propranolol therapy were excluded, subjects were followed until a major cardiac event (new acute myocardial infarction, cardiac surgery or death) occurred. During follow-up, 22 patients died, 49 survived without events, nine underwent coronary bypass surgery, and five had nonfatal myocardial infarctions. The measurements made at the onset of the study, including cardiothoracic ratio (C/T) on chest x-ray, resting electrocardiographic abnormalities, maximum exercise tolerance testing (METT) data, systolic time interval (STI) measurements (before exercise and 34 minutes after METT), and results of cardiac catheterization (55 patients), were analyzed at its conclusion to determine the best predictor of subsequent mortality. Of these measurements, left ventricular ejection fraction, endurance time on METT and C/T were shown to be useful prognostic indicators of subsequent mortality. However, the pre-ejection phase-to-left ventricular ejection time (PEP/LVET) ratio (0.40 ± 0.05 in survivors vs 0.50 + 0.09 in nonsurvivors) in the resting pre-exercise state wa1s significantly more predictive of mortality than the other measurements. On life-table analysis, the difference in survival between subjects with a resting PEP/LVET < 0.50 and those with a resting PEP/LVET 2 0.50 was highly significant. The measurement of the STIs after maximal exercise testing failed to improve upon the prognostic ability of the simple determination of PEP/LVET in the resting, supine state. STIs provided highly specific noninvasive prognostic information in this group of patients with stable angina pectoris.