Background: Left ventricular diastolic dysfunction (LVDD) was reported as a significant predictor of mortality, mainly in patients with heart failure. However, prospective data are scarce in the hospitalized elderly population. Methods: We studied the association of severe LVDD, defined by conventional echocardiographic parameters, with cardiovascular and all-cause mortality in a population of 331 hospitalized elderly patients with a history of cardiovascular disease (mean age ± SD, 87 ± 7 years). After a mean follow-up of 378 days, 110 deaths occurred. Results: Compared with left ventricular systolic dysfunction (LVSD), subjects with severe diastolic dysfunction had a similar prevalence (12% vs 10%) and similar cardiovascular and all-cause mortality (18% vs 19%, 49% vs 50%). Both cardiovascular and all-cause mortality increased progressively and significantly with increasing number of diagnostic criteria of LVDD (P = .035, P = .013) and reached 48.7% for all-cause mortality when at least 2 criteria were met. In addition to cardiovascular risk factors and LVSD, severe LVDD provided incremental and independent prognostic information of all-cause mortality with increased χ2 value of Cox regression model (48.1 vs 43.5, P = .022). Conclusions: Severe LVDD, diagnosed by conventional echocardiography, has similar prevalence and prognosis as LVSD and provides incremental prognostic value, which highlights the clinical significance of routine evaluation of LVDD in risk assessment strategies of the hospitalized elderly. © 2010 Mosby, Inc. All rights reserved.