Background: The clinical significance of nutritional risk assessment in patients with heart failure with preserved ejection fraction (HFpEF) remains undefined. Geriatric nutritional risk index (GNRI) is a simple nutritional assessment tool for elderly subjects. Its predictive value was evaluated in patients with HFpEF, a common HF phenotype in the elderly population. Methods and Results: The present study enrolled 152 consecutive patients (mean age, 77±11 years; male, 53.9%) who were hospitalized with HFpEF at the authors’ institution. GNRI on admission was calculated as follows: 14.89×serum albumin (g/dl)+41.7×body mass index/22. Characteristics and mortality (median follow-up of 2.1 years) were compared between 2 groups: low GNRI (<92) with moderate or severe nutritional risk; and high GNRI (≥92) with no or low nutritional risk. Patients in the low-GNRI group were more often female, and had lower serum hemoglobin and sodium, but higher serum blood urea nitrogen (BUN), C-reactive protein, and B-type natriuretic peptide (BNP) compared to those in the high-GNRI group (P<0.05, respectively). Physical activity at discharge measured by Barthel index was significantly lower in the low-GNRI group than the high-GNRI group (P<0.05). On Cox hazard analysis, lower GNRI predicted increased mortality independent of age, gender, prior HF hospitalization, and higher BUN and BNP (P<0.01). Conclusions: GNRI may be useful for predicting functional dependency and mortality in patients with HFpEF.  (Circ J 2013; 77: 705–711)