Protein-energy malnutrition is thought to be widespread in hospitalized patients. However, the specificity of indexes used to assess malnutrition is uncertain. We therefore assessed the rate of false-positive diagnoses of malnutrition when biochemical-anthropometric indexes were applied to healthy subjects. Nutritional status was assessed in 175 healthy blood donors (aged 44.2 +/- 13.4 y) and in 34 highly fit elderly volunteers (aged 74.7 +/- 3.6 y) participating in the Nijmegen Four Days Walking March. We investigated both the Nutritional Risk Index [(1.489 x albumin) + (41.7 x present/usual weight)] and the Maastricht Index [20.68-(0.24 x albumin, g/L)-(19.21 x serum transthyretin, g/L)-(1.86 x lymphocytes, 10(6)/L)-(0.04 x ideal weight)]. We found previously that 52-64% of nonsurgical hospitalized patients were malnourished according to these indexes. In the present study, 1.9% of the 209 volunteers had apparent malnutrition according to the Nutritional Risk Index and 3.8% according to the Maastricht Index. The prevalence of apparent malnutrition in the elderly volunteers was 5.9% and 20.6%, respectively. The rate of false-positive diagnoses was acceptably low in those aged < 70 y with both the Nutritional Risk Index and the Maastricht Index; therefore, the use of these indexes will not cause a clinically significant increase in the prevalence of malnutrition because patients who are not malnourished are included. The high percentage of spurious malnutrition in the elderly limits the use of the Maastricht Index to subjects aged < 70 y.