Résumé : Early diastolic mitral inflow and annulus velocities ratio (E/e') is the parameter used to estimate left ventricular (LV) filling pressure, but their use for patients with reduced ejection fraction (EF) (<50 %) is controversial. LV endocardial area change rate with 3-dimensional (3-D) speckle-tracking can be used to evaluate the entire LV diastolic function, we tested the hypothesis that the early diastolic area change rate (E-ACR) represents LV relaxation, and E and E-ACR ratio (E/E-ACR) can accurately predict LV filling pressure in patients with reduced and preserved EF (≥50 %). We studied 125 patients with EF of 40 ± 17 %. E-ACR of all patients was measured using the global area change rate, and LV relaxation time constant (τ, n = 53) and pulmonary capillary wedge pressure (PCWP, n = 125) were measured by cardiac catheterization. E-ACR significantly correlated with τ (r = -0.60, p < 0.001). In the retrospective group (n = 101), E/E-ACR correlated with PCWP for patients with reduced EF as well as those with preserved EF (n = 79 and 22, r = 0.62 and 0.81, both p < 0.001). In overall patients, E/E-ACR ≥94 cm was a better predictor of elevated mean PCWP (>12 mmHg) (sensitivity: 69 %, specificity: 87 %, areas under the curves = 0.82; p < 0.001) and was a more accurate predictor than E/e' (p = 0.001) for the retrospective group. For the validation group (n = 24), E/E-ACR ≥94 cm was similarly predictive for elevated mean PCWP (sensitivity: 60 %, specificity: 100 %, positive predictive value: 100 %, negative predictive value: 78 %). In conclusions, global E-ACR by 3-D speckle-tracking proved to be dependent on LV relaxation. Moreover, E/E-ACR was found to be a powerful predictor of LV filling pressure regardless of EF.