Número de replicações de inquéritos dietéticos para estimativa da ingestão de nutrientes em gestantes Brasileiras

Abstract

Objectives: to determine the number of replicates of dietary measurements required to the usual nutrient intake and to classify Brazilian pregnant women into categories of intake. Methods: a prospective study was carried out enrolling 82 pregnant women, in which data of total energy and 18 nutrients were obtained by three 24hour food recalls, one in each gestational trimester. Different equations were used to estimate the number of replicates required to classify pregnant women into categories of intake, based on the variances ratio of within-person/between person, and to estimate usual intake, established by within-person variance. Results: to classify the women into categories of intake between 11 and 51 replicates were required, considering a correlation coefficient of 0.9. Considering a correlation coefficient of 0.7, the number of replicates ranged from four to 19. To estimate usual intake between two and 33 would be required, replicates, admitting an error of 10%. Considering an error of 20%, from one to seven replicates of dietary measurements would be necessary. Conclusions: a high number of replicates of dietary measurements are required to estimate nutrient intake during pregnancy. A reduced number of dietary replicates among pregnant women could attenuate the associations between diet and maternal and child health outcomes.