Methods for evaluating left ventricular diastolic function have been used exclusively in cardiac patients in sinus rhythm and have not been applied to atrial fibrillation because of large cycle variation in flow velocities and absence of atrial contraction. The utility of these rnethods in atrial fibrillation is unknown, especially M-rnode Doppler echocardiography and automatic border detection method.

Left Ventricular Normal group(LVN group, Ejection fraction> 40%) of 21 patients(mean age 67 years old, 12 men, 9 women) and Left Ventricular Dysfunction group (LVD group, EF ( 40%) of 15 patients(mean age 64 years old, 11 men, 4 women) were studied. We measured parameters in pulsed Doppler recordings of rnitral flow velocity and the flow propagation velocity in color M-mode recording of mitral flow, and peak filling rate from automatic border detection system. Measurements were averaged over 10 cardiac cycles.

Mean heart rates were not different in both population(LVN 79±18/min, LVD 83±16/min). Left ventricular diastolic dimension were more dilated LVD group than LVN group(5.7±1.0crn, 4.6±0.7cm p < 0.05) but left atrial dimension were similar(4.4±0.8cm, 4.3±0.7cm). Deceleration time was significantly shortened in LVD group(158.4±24.4ms vs. 180.4+37.6ms, p < 0.05) but other parameters of mitral pulsed Doppler recording such as IVRT, acceleration time, acceleration rate and deceleration rate were not different. Flow propagation velocities were significantly decreased in LVD group(41.7±14.9cm/s vs. 53.0±11. 9cm/s, p < 0.05). Peak filling rates of automatic border detection system were significantly decreased in LVD group(3.4±1.2 EDV/s vs 5.4±1.1 EDV/s, p< 0.001).

Echocardiographic measurements of diastolic function via color M-mode Doppler and autornatic border detection may be useful in presence of atrial fibrillation as sinvs rhythm.