[Obesity prevalence and associated factors in the elderly in Pelotas, Rio Grande do Sul State, Brazil: obesity classification according to two cutoff points for body mass index].

The aim of this study was to estimate obesity prevalence and associated factors in the elderly, according to two cutoff points for body mass index (BMI). This was a cross-sectional, population-based study of a sample of 596 elderly residents in Pelotas, southern Brazil. Nutritional status was evaluated through BMI based on self-reported weight and stature. Poisson multiple regression with hierarchical analysis was applied, with two dependent variables for definition of obesity: BMI > 27 kg/m(2) and BMI > or =30 kg/m(2), the cutoff points proposed by Lipschitz and the World Health Organization, respectively. Prevalence of obesity was 48.7% (95%CI: 44.6-52.7) for BMI > 27 kg/m2 and 25.3% (95%CI: 21.9-29.0) for BMI > or =30 kg/m(2). Two multiple regression models showed an association between obesity and female gender and age. BMI > 27 kg/m(2) showed an association with age and smoking and BMI > or =30 kg/m(2) with sedentary lifestyle. Obesity prevalence was high in this group of elderly. For the Brazilian elderly population and from a public health perspective, the study suggests the use of a more sensitive cutoff point, namely BMI > 27 kg/m(2).