Abstract

Summary

Background & aims

This cohort study aimed to investigate and compare the ability to predict malnutrition in a group of frail older hospital patients in the United Kingdom using the nutritional risk screening tools, MUST (malnutrition universal screening tool), MNA-SF® (mini nutritional assessment-short form) and bioelectrical impedance assessment (BIA) of body composition.

Methods

MUST and MNA-SF was performed on 78 patients (49 males and 29 females, age: 82 y ± 7.9, body mass index (BMI): 25.5 kg/m2 ± 5.4), categorised by nutritional risk, and statistical comparison and test reliability performed. BIA was performed in 66 patients and fat free mass (FFM), fat mass (FM) and body cell mass (BCM) and index values (kg/m2) calculated and compared against reference values.

Results

MUST scored 77% patients ‘low risk’, 9% ‘medium risk’ and 14% ‘high risk’, compared to MNA-SF categorisation: 9%, 46% and 45%, respectively (P < 0.000001). Reliability assessment found poor reliability between the screening tools (coefficient, r = 0.4). Significant positive correlations were found between most variables (P < 0.05–<0.001); although females exhibited greater variation. FFM index analysis found 40% of males low/depleted, 21% borderline/at risk with 96% categorised by MNA-SF as either malnourished or at risk (MUST-35%). 29% males had low FM index and all appropriately classified by MNA-SF. 30% females had low FFM index or borderline, MNA-SF screening appropriately categorised 86% (compared to MUST-29%).

Conclusions

This preliminary data may have significant clinical implications and highlights the potential ability of the MNA-SF and BIA to accurately assess malnutrition risk over MUST in frail older hospital patients.