Nutritional risk score is not sensitive enough to predict weight loss in diseased elderly subjects.

Objectives: malnutrition is highly prevalent in diseased elderly people, especially in hospital departments, and weight loss also occurs during hospital stays. Among the tools proposed to define malnutrition and to screen persons at risk of malnutrition, the nutritional risk score (NRS) has been suggested as a simple tool.

Design And Participants: the aim of the present study was to test the validity of the NRS to predict weight changes 3 months after discharge. NRS and visual analogue scales for hunger, early satiety, thirst, fatigue and pain were performed in 106 patients aged 65 yrs and over, 67 treated for cancer, without overt malnutrition.

Results: forty six patients lost more than 1 kg. None of the parameters tested was correlated with weight changes, except NRS which was weakly correlated (r=-0.22, P=0.037). However a 0 score at NRS was associated with weight changes ranging -9 to +5 kg. Furthermore, the 11 patients with a high NRS score and weight loss were in a clinical situation that made it very likely that they would lose weight. In conclusion, NRS is not a valid score to predict weight change.