Limitations in the comparison of the Brazilian National Dietary Surveys of 2008–2009 and 2017–2018

OBJETIVE: To present particular characteristics of two Brazilian National Dietary Surveys (Inquéritos Nacionais de Alimentação - INA) and the methodology used to better compare their data. METHODS: This study details the differences between both INA conducted by the Brazilian Institute of Geography and Statistics (IBGE) in 2008–2009 and 2017–2018. We present the alterations in data collecting methods and food composition tables as well as the analysis strategies recommended to obtain such data. A validation study with 95 participants of the third wave of the Longitudinal Study of Adult Health assessed the measurement error associated with the procedures adopted in the 24-hours dietary recall of INA 2017–2018. The reference standards were urinary protein recovery, sodium, and potassium biomarkers. Different strategies were used in the analysis of INA to compare two essential dietary items that had their collection method changed: fats and sugars. RESULTS: The validation study indicated lower underreport in the most recent survey with higher means of energy intake. The correlation of means for the 24-hours recalls with their respective biomarkers was 0.58 for proteins, 0.31 for potassium, and 0.30 for sodium. Comparing the food composition tables used in both surveys with the data obtained by INA 2008-2009, the mean variation of energy, macronutrients, and minerals was lower than 15%, except for trans fats and selenium, which had means 40% and 52% lower in the Tabela Brasileira de Composição de Alimentos (TBCA - Brazilian Food Composition Table). INA 2017–2018 presents lower means for added sugar, using a generic question about the frequency of sugar consumption as a measure for sugar as an additional item. CONCLUSION: The methodological changes promoted in the most recent INA enhanced food groups and nutrients intake estimation, adding detailed and specific data in dietary habits reports.