To estimate the accuracy of cardiac output (CO) measured by Continuous-wave (CW) Doppler echocardiography from the suprasternal notch, simultaneous thermodilution and Doppler CO were measured at rest in 24 patients with coronary artery disease without aortic valvular lesion. Doppler-derived stroke volume (SV) was calculated from the product of flow velocity integral and cross-sectional area of the aortic annulus or the aortic root just above the sinuses of Valsalva. CO was obtained by multiplying SV by heart rate (HR). A good correlation between the Doppler and thermodilution CO was observed when the aortic diameter was measured just above the sinuses of Valsalva using the inner edge to inner edge method (r=0.88). In addition, the utility of the Doppler method for evaluating the effect of altered left ventricular function during exercise, 21 healthy males were studied using the Doppler method during treadmill exercise testing. Exercise duration ranged 15~18 minutes and CW Doppler recordings were well performed in all cases until the end-stage of exercise (max HR 135~195 bpm). SV increased from 60 ml at rest (upright position) to 102 ml (max SV) at 126 bpm of HR and then decreased to 82 ml at peak exercise. CO increased progressively (from 4.4 to 13.2 l/min) until the peak exercise. Peak velocity of ascending aortic flow also increased from 0.7 m/sec at rest (upright position) to 1.5 m/sec during peak exercise. During low levels of exercise, both the Frank-Starling mechanism and increased contractility might play a role in augmenting SV. During high levels of exercise, CO increased progressively until the end stage of exercise due to further increment of HR, while SV decreased because of the shortening of diastolic filling time. These findings suggest the usefulness of the CW Doppler method for evaluating left ventricular function during upright exercise testing.