Use of the Mini Nutritional Assessment to detect frailty in hospitalised older people

The aims of this study were to: (1) determine the prevalence of undernutrition and frailty in hospitalised elderly patients and (2) evaluate the efficacy of both the Mini-Nutritional Assessment (MNA) screening tool and the MNA short form (MNA-SF) in identifying frailty.

A convenient sample of 100 consecutive patients (75.0 % female) admitted to the Geriatric Evaluation and Management Unit (GEMU) at The Queen Elizabeth Hospital in South Australia.

Frailty status was determined using Fried's frailty criteria and nutritional status by the MNA and MNA-SF. Optimal cut-off scores to predict frailty were determined by Youden's Index, Receiver Operator Curves (ROC) and area under curve (AUC).

Undernutrition was common. Using the MNA, 40.0% of patients were malnourished and 44.0% were at risk of malnutrition. By Fried's classification, 66.0 % were frail, 30.0 % were pre-frail and 4.0 % robust. The MNA had a specificity of 0.912 and a sensitivity of 0.516 in predicting frailty using the recommended cut-off for malnourishment (< 17). The optimal MNA cut-off for frailty screening was <17.5 with a specificity of 0.912 and sensitivity of 0.591. The MNA-SF predicted frailty with specificity and sensitivity values of 0.794 and 0.636 respectively, using the standard cut-off of < 8. The optimal MNA-SF cut-off score for frailty was < 9, with specificity and sensitivity values of 0.765 and 0.803 respectively and was better than the optimum MNA cut-off in predicting frailty (Youden Index 0.568 vs. 0.503).

The quickly and easily administered MNA-SF appears to be a good tool for predicting both under-nutrition and frailty in elderly hospitalised people. Further studies would show whether the MNA-SF could also detect frailty in other populations of older people.