The exercise-induced changes in left ventricular filling in patients with coronary artery disease are poorly understood. Therefore these changes were studied in relation to a noninvasive indicator of exercise pulmonary venous congestion, the lung-to-heart (L:H) ratio on symptom-limited thallium stress testing. Fifty-six patients undergoing diagnostic treadmill testing were studied; 50 of them had technically adequate Doppler recordings and became the subjects of this study. Doppler left ventricular filling was assessed with patients in the supine position both before and after exercise. Measurements included early (E) and late (A) filling velocities, their ratio, the diastolic time-velocity integral, and the diastolic filling time. The L:H ratio was considered abnormal if it was greater than the upper 95% confidence limit for a separate group of normal subjects. Twelve subjects had a documented prior myocardial infarction, 16 had stress-induced ischemia, and 20 had abnormal L:H ratios. A greater E and a longer diastolic filling time in the group with an abnormal L:H ratio were the only postexercise measurements that differed; however, E was the only filling parameter that both differed between groups after exercise (abnormal L:H group 87 +/- 25 cm/sec; normal 68 +/- 20 cm/sec; p < 0.01) and whose change from rest to after exercise was significantly different (p < 0.05). Since Doppler velocities are directly related to instantaneous gradients, the higher E in patients with evidence of exercise pulmonary congestion suggests a higher exercise early diastolic left atrial pressure.