BACKGROUND: Pulmonary venous diastolic flow follows the pattern of mitral flow and is dependent on the pressure difference between the pulmonary vein and the left atrium (LA). The magnitude of the decrease in LA pressure in early diastole depends on both the volume of the blood leaving the LA and the stiffness of the left ventricle (LV) and the LA. Relaxation process is known to govern early diastolic compliance. We hypothesized that in patients with decreased early diastolic compliance due to LV relaxation abnormality, there may be rapid rise in LV and LA pressure, resulting in early peak of pulmonary venous D wave as early LV diastolic filling progress. This study was undertaken to define this hypothesis and to examine the relation of the time interval between E wave peak and D wave peak to mitral doppler indexes. METHOD: Patients with significant mitral or aortic valvular disease, or patients with LV ejection fraction below 60%, or patients who have pseudonormal or restrictive LV filling pattern on mitral and pulmonary venous Doppler, were excluded from this study. Mitral Doppler indexes including peak E velocity, peak A velocity, E wave acceleration time (EAT) and deceleration time (EDT) were measured. E/A ratio was calculated. The isovolumic relaxation time from aortic valve closure (Ac) to the onset of E wave , the time interval from Ac to the peak of E wave (AcE), the time interval from Ac to the peak of D wave, and the diastolic time from Ac to R of electrocardiogram (AcR) were measured by the pulsed wave Doppler and phonocardiography. The time interval from the peak of E wave to the peak of D wave (ED) was calculated by the subtraction of AcE from AcD. RESULTS: 1) ED is significantly shorter in patients with E/A<1 than those with E/A> or =1 (58.9+/-27.4 msec versus 74.7+/-17.2 msec, p<0.05). 2) ED correlated with IVRT (r=-0.400, p<0.01), AcR (r=0.414, p<0.01), but not with E/A ratio, EDT, or EAT. 3) Multivariate linear regression analysis with all the previously mentioned variables showed that IVRT, AcR, and EAT were independent determinants of the ED. CONCLUSION: This study demonstrates that the ED is shortened in patients who are regarded as having LV relaxation abnormality and that ED is affected by IVRT, AcR, and EAT.