Introduction: Body composition changes related to aging alter the capacity of predicting risk through anthropometric parameters. Objective: To discuss methodological aspects of anthropometry in active elderly based on associations between Body Mass Index (BMI) and other nutritional indicators. Methods: Cross-sectional study with active elderly from Macaé, Rio de Janeiro, Brazil (2014/2015). Nutritional status was described according to the BMI (Nutritional Screening Initiative, 1994). Linear regression analysis was performed: the outcome variable was BMI and the dependent ones were circumferences of waist, hip, neck, calf, arm and waist-to-hip ratio (WHR). Results: We assessed 173 people (55.5% female; median 71 years old). Calf and neck circumferences and WHR presented low R2 value. Among women, hip (R2=0.825) and waist circumferences (R2=0.729) individually explained much of the variation in BMI; and among men, waist (R2=0.759) and arm circumferences (R2=0.741) performed better. The cut-off points for waist circumference corresponding to the critical BMI value (27 kg/m2) were 87.9 and 96.8 cm, respectively for women and men. In multiple analysis, the association of waist, hip and arm circumferences with BMI remained significant. Conclusion: Circumferences traditionally used to assess adults had higher linear association with BMI than specific indicators for elderly people. The body composition of active elderly can be more similar to adults’ than that of elderly with other profiles. The waist circumference cut-off points established for adults may not be suitable for elderly populations. We suggest testing the cut-off points obtained by this study on other groups of active elderly.