ABSTRACT
The objective of this study was to identify the discriminating criterion for body mass index (BMI) in the prediction of low fat free mass and high body fat percentage according to sex among older people. Observational analytical study with cross-sectional design was used for this study. All institutionalized older people from the city of Uberaba (Minas Gerais, Brazil) who fit within the inclusion and exclusion criteria were approached. Sixty-five institutionalized older people were evaluated after signing a Free and Informed Consent Form. Descriptive and inferential statistical procedures were employed for the analysis, using Student's t-test and multiple linear regression. Receiver Operating Characteristic (ROC) curves were constructed to determine the BMI (kg/m(2)) cut-off points. The study complied with all the ethical norms for research involving human beings. In comparing the anthropometric measurements obtained via bioimpedance, elder male had higher mean height and body water volume than females. However, women had higher mean triceps skinfold and fat free mass than men. The BMI cut-off points, as discriminators of low fat free mass percentage and high body fat percentage in women, were ≤22.4 kg/m(2) and >26.6 kg/m(2), respectively; while for men they were ≤19.2 kg/m(2) and >23.8 kg/m(2). The results of this study indicate the need for multicenter studies aimed at suggesting BMI cut-off points for institutionalized older people, taking into account specific sex characteristics.