Abstract

Background: The body mass index (BMI) >= 30 kg/m2 is the universally accepted cut-off point for defining obesity; however, its accuracy in classifying obesity in older adults is poorly understood.Objectives: To assess the performance of the BMI cut-off point >= 30 kg/m2 in classifying obesity in older adults, using the fat mass index (FMI) and fat mass percentage (FM%) as reference criteria; and to establish region- and sex-specific BMI-based cut-off points to classify obesity in older adults.Methods: The present study is a secondary analysis derived from a cross-sectional project that included a sample of 1463 older adults from ten Latin American and Caribbean countries. Volunteers underwent total body water measurements using the deuterium dilution technique to determine FMI and FM%. Accuracy of the BMI and derived cutoff points was assessed by the area under the receiver operating characteristic curve (AUC). Results: The BMI cut-off point >= 30 kg/m2 had low sensitivity for classifying obesity in these older adults compared to the FMI and FM%. The AUC values for the optimal BMI-derived cut-off points showed an acceptable-to-outstanding discriminatory capacity in diagnosing obesity defined by the FMI. There was also a better balance between sensitivity and specificity than with the values obtained by a BMI >= 30 kg/m2 in older subjects in both regions. Conclusion: The BMI cut-off point >= 30 kg/m2 had poor sensitivity for accurately diagnosing obesity in older adults from two regions. The region-and sex-specific BMI-derived cut-off points for defining obesity using the FMI are more accurate in classifying obesity in older men and women subjects from both regions.