How do calculation method and food data source affect estimates of vitamin A content in foods and dietary intake?

• Compiled vitamin A data were consistent with analytical data from a Brazilian table.
• Vitamin A Brazilian data were little consistent with North American data.
• Vitamin A calculated without β-carotene isomers values can be included in databases.
• Vitamin A data from different countries results in inaccurate dietary intake estimate.

The aim of this work was to evaluate the effect of the calculation method and food data source on estimates of vitamin A content in foods and in dietary intake. The Brazilian Vitamin A Database (BVAD) was elaborated using compiled data on retinol and carotenoids in Brazilian foods. Vitamin A was calculated with and/or without separating β-carotene isomers. Estimates of vitamin A intake and data on 16 plant foods were compared using data from BVAD, USDA National Nutrient Database and Brazilian Food Composition Table (TACO). Percentage difference (D%) was calculated to evaluate data variability. Calculated vitamin A values with and without separating β-carotene isomers (BVAD) were consistent in most of the evaluated foods and did not impact on vitamin A intake estimates. BVAD data were consistent with TACO analytical data in 81% of the selected foods and only in 37.5% with data from the USDA. Estimates of vitamin A intake calculated by Brazilian databases were similar, while by USDA the results were quite different. Therefore, the use of values for β-carotene without isomer separation in databases did not affect vitamin A data consistency, however, the use data from different countries to evaluate vitamin A intake may result in inaccurate values.