06 ROQUE_04 LORD_c

Objectives: The objective of this study is to assess the nutritional status, measured by the MNA, in community-dwelling elderly individuals with dementia and to identify clinical risk factors for nutritional risk or malnutrition. Design: Cross-sectional analysis of a cluster randomized clinical trial (Nutrialz). Setting: Community-dwelling individuals attending dementia clinics. Participants: 940 individuals. Measurements: The clinical scales assessed were Mini Nutritional Assessment (MNA), Eating Behaviour Scale (EBS), Charlson comorbidity index, Clinical Dementia Rating (CDR), Mini-Mental State Examination (MMSE), Basic Activities of Daily Living (BADL) score, Instrumental Activities of Daily Living (IADL) score, Neuropsychiatric Inventory Questionnaire (NPI-Q), Cornell depression scale and Zarit Caregiver Burden Interview. Results: 5.2% of participants were classified as being malnourished, 42.6% as being at risk of malnutrition and 52.2% as well nourished. Malnutrition by type of dementia was more frequent in Lewy bodies dementia (18.2%) than in the other types. Worse nutritional status is significantly related to more advanced age and worse cognitive, functional and behavioural profile, as well as increased burden for caregivers. Presence of behavioural symptoms is significantly related to worse nutritional level for all NPI-Q symptoms but depression, exaltation, lack of inhibition and irritability. The items more strongly related to malnutrition are appetite/feeding and hallucinations. Dependence in any basic or instrumental ADL is significantly related to higher risk of malnutrition. Dependence on feeding is a strongly related risk factor, while food preparation is only a moderate one. A logistic regression model to predict at risk/malnutrition kept as significant risk factors EBS (Odds Ratio (OR) 0.84, 95%CI 0.78 to 0.91), Cornell (OR 1.12, 95%CI 1.09 to 1.16), the number of dependent BADL (OR 1.29, 95%CI 1.17 to 1.42), age (OR 1.04, 95%CI 1.02-1.06), MMSE (OR 0.95, 95%CI 0.92 to 0.98) and Charlson (OR 1.18, 95%CI 1.05 to 1.34). A similar model built for prediction of malnutrition retained as significant covariables only EBS, Cornell and the number of dependent BADL. Conclusion: These results will allow a better understanding of the clinical stage previous to malnutrition. An adequate diagnosis and treatment of identified modifiable factors like functional impairment, eating behaviours and depression could delay or avoid malnutrition.