Mitral flow velocity patterns are frequently ''normalized'' by the alteration in the loading condition even in the presence of left ventricular (LV) diastolic dysfunction. In addition, a simple index, the ratio of mitral peak early diastolic flow velocity to mitral peak flaw velocity at atrial contraction, is not obtainable in patients with atrial fibrillation (Af). Thus these limitations hamper the value of analyzing the mitral flow velocity pattern in the assessment of abnormal LV diastolic characteristics. This study was designed to elucidate the hypothesis that peak early diastolic flow velocity decreases progressively from the base to the apex in patients with LV diastolic dysfunction. Regional diastolic flow velocity patterns at 1, 2, or 3 cm from the mitral tip toward the apex were simultaneously recorded with the mitral flow velocity pattern by using multigate pulsed Doppler echocardiography in 42 subjects with normal LV function (31 normal volunteers and 11 patients with Af only), 17 patients with hypertensive heart disease, and 22 patients with dilated cardiomyopathy. In the normal subjects early diastolic flow velocity at the mitral tip was maintained at the positions 1 to 3 cm away from the tip into the LV cavity. In contrast, regional peak early diastolic flow velocity progressively decreased toward the apex in patients with hypertensive heart disease and dilated cardiomyopathy. These findings were observed even in patients with a normalized mitral flow velocity pattern or those with Af. Thus the assessment of the intraventricular decrease in peak early diastolic flow velocity may be useful in detecting LV diastolic dysfunction, particularly in patients with Af or a ''normalized'' mitral flow velocity pattern.