Food consumed does not account for the higher prevalence of obesity among stunted adults in a very-low-income population in the Northeast of Brazil (Maceio, Alagoas)

Objective: To study the food pattern of stunted and nonstunted, obese and nonobese individuals in a very- low-income population. Design: A household survey. Setting: Slum set up by the 'Homeless Movement', city of Maceio (Alagoas), Brazil. Subjects and methods: A total of 532 adults classified by sex, stature (Zless than or equal to and Z > - 2s. d. of the NCHS curves), and body mass index (BMI) were compared using the following variables: waist circumference, waist - hip circumference ratio (W/H), percentage body fat (skinfold thickness and bioelectrical impedance), and food intake (24-h recall). Results: The prevalence of stunting was 22.6%. In all, 30% of the stunted subjects were overweight or obese, compared with 23% for the nonstunted individuals (P<0.05). In women, logistic regression analysis showed a strong association among weight, abdominal fat, and stunting (r = 0.81). No significant differences were observed in the values of W/H or in the qualitative menu of the different categories. Energy intake was below the RDA figures ( about 63%). There was similarity among the groups regarding the proportion of macronutrients, except for the fact that stunted obese women ingested less fat and protein than nonstunted obese women. Stunted obese individuals consumed less energy ( 5962 kJ) than the population as a whole ( 6213 kJ), an amount far lower than their average needs, which were calculated on the basis of their shorter stature ( 8109 kJ). Conclusion: The observed energy consumption seems compatible with the panorama of undernutrition present in the population, but it does not explain the high prevalence of obesity detected. Sponsorship: Alagoas State Foundation for the Support of Research (FAPEAL) and the National Agency for the Advancement of College Graduates ( Ministry of Education and Sports), Brazil.