Tissue Doppler echocardiography provides a reliable noninvasive estimation of left ventricular (LV) filling pressures irrespective of LV ejection fraction. However, the diagnostic accuracy of E/Ea ratio has not been adequately determined in patients with suspected heart failure (HF) with preserved LV systolic function in the acute care setting. To investigate the accuracy of E/Ea ratio in the emergency diagnosis of decompensated HF with preserved LV systolic function. Seventy patients with a LV ejection fraction >or=45%, 32 with decompensated HF (77 +/- 12 years of age), and 38 with noncardiac cause of acute dyspnea (74 +/- 12) were enrolled. B-type natriuretic peptide (BNP) was measured on admission; lateral, septal and average E/Ea ratios were calculated within 24 hours. Using receiver-operating characteristic curves to evaluate diagnostic performance, BNP (AUC of 0.875, P < 0.0001) and E/Ea ratios (AUC of 0.90-0.92, P < 0.0001) provided similar accuracy for predicting decompensated HF. Optimal cutoffs were 146 pg/ml for BNP (sensitivity and specificity of 90.6% and 76.3%), 9.8 for lateral E/Ea (83.3% and 88.9%), 12.7 for septal E/Ea (76.7% and 91.4%), and 11.5 for average E/Ea ratio (80% and 94.3%). On multivariate logistic regression analysis, average E/Ea ratio yielded independent additional information to a model based on the clinical judgment and BNP level according to the standard cutoff of 100 pg/ml. Tissue Doppler echocardiography is accurate for predicting decompensated HF with preserved LV systolic function and may be used as a diagnostic complement to inconclusive BNP level in this setting.