The study was to determine whether a modified MNA (Mini Nutritional Assessment) which adopted population-specific anthropometric cut-points but without BMI could maintain its predicting ability in community-living elderly in Taiwan.

Purposive sampling.

Community-living elderly.

Three hundred and one (138 male and 163 female) >65-year-old outpatients seeking free annual health examination at an area hospital in central Taiwan.

A structured questionnaire elicited personal data, lifestyle information and answers to the MNA. Laboratory results from health checkup provided the needed biochemical data. Each subject's nutritional status was assessed with the MNA in three versions: the original, the MNA-TI (with population-specific cut-points), and the MNA-TII (further eliminated the BMI question and redistributed its score to the MAC and CC questions).

All three versions identified the same 0.7% elderly malnourished. The proportions predicted at risk of malnutrition were 16.6, 12.0 and 10% according to the original, MNA-TI and MNA-TII, respectively. Friedman Test and post hoc analysis indicated that the pattern predicted by the original MNA was different from that predicted by the two modified versions whereas the patterns predicted by the two modified versions were not different from each other.

Adoption of population-specific anthropometric cut-points improves the predicting ability of the MNA in Taiwanese elderly, and the improved functionality is maintained in a version without BMI (but with adjusted MAC and CC scores). A MNA without BMI has greater applicability and can enhance professional efficiency of healthcare workers.