Objectives: to investigate the prevalence of malnutrition and the association between Body Mass Index (BMI), Mini Nutritional Assessment (MNA) and five-year mortality in a representative population of very old (>85 years) people.
Design: A prospective cohort study.
Setting: A population-based study of very old people in northern Sweden and western Finland, living in institutional care or in the community.
Participants: Out of 1195 potential participants, 832 were included (mean age 90.2±4.6 years).
Measurements: Nutritional status was assessed using BMI and MNA and the association of those two variables with five-year mortality was analyzed.
Results: The mean BMI value for the whole population was 25.1±4.5 kg/m2, with no difference between genders (P=0.938). The mean MNA score was 22.5±4.6 for the whole sample, and it was lower for women than for men (P<0.001). Thirteen percent were malnourished (MNA<17) and 40.3% at risk of malnutrition (MNA 17-23.5) according to MNA. Also, 34.8% of those with a MNA score <17 still had a BMI value ≥22.2 kg/m2. A BMI value <22.2 kg/m2 and a MNA score<17 were associated with lower survival. The association with mortality seemed to be J-shaped for BMI, and linear for MNA.
Conclusions: Malnutrition according to MNA was common, but a substantial portion of those with a low MNA score still had a high BMI value, and vice versa. The association with mortality appeared to be J-shaped for BMI, and linear for MNA. The MNA seems to be a good measurement of malnutrition in very old people, and BMI might be misleading and could underestimate the prevalence of malnutrition, especially in women.