Clinical impact of different scores of the mini nutritional assessment (MNA) in the diagnosis of malnutrition in patients with cognitive impairment.

It was aimed at evaluating the clinical usefulness of the mini nutritional assessment (MNA) to identify malnutrition in elderly patients with cognitive impairment, admitted to a geriatric convalescence unit (intermediate care facility). Sixty-three patients with cognitive impairment were studied. Cognitive impairment was considered when mini mental state examination (MMSE) scores were below 21. MNA and a nutritional evaluation according to the sequential model of the American Institute of Nutrition (AIN) were performed at admission. According to the AIN criteria, malnutrition was considered, if there were abnormalities in at least one of the following parameters: albumin, cholesterol, body mass index (BMI), and branchial circumference. Based on these criteria, 27 patients (42.8%) proved to be undernourished at admission, whereas if taking the original MNA scores, 39 patients (61.9%) were undernourished, 23 (36.5%) were at risk of malnutrition, and 1 (1.5%) was normal. The analyzed population was divided in four categories (quartiles) of the MNA scores: very low (</= 13.5), low (> 13.5 and </= 16), intermediate (> 16 and </= 18.5) and high (> 18.5). Likelihood ratios of each MNA quartile were obtained by dividing the percentage of patients in a given MNA category who were undernourished (according to AIN) by the percentage of patients in the same MNA category who were not undernourished. In the very low MNA quartile, this likelihood ratio was 2.79 and for the low MNA quartile it was 0.49. For intermediate and high MNA categories, likelihood ratios were 1.0 and 0.07 respectively. In the present study, MNA identified undernourished patients with a high clinical diagnostic impact value only, when very low scores (</= 13) are obtained.