Reference serving sizes for the Brazilian population: An analysis of processed food labels

OBJECTIVE:To compare serving sizes reported on processed food labels with reference serving sizes according to nutrition labeling legislation and the "Food Guide for the Brazilian Population".METHODS:This cross-sectional study analyzed the labels of 2,072 processed foods in a supermarket of Florianópolis, Santa Caratina, Brazil. The foods were classified according to the Brazilian food labeling legislation. Central tendency and variability values were calculated for the serving sizes and energy values reported on the labels, as well as the ratio between the reported and reference energy value. The Spearman correlation test was performed between the reference serving size and the reference energy density, and also between the reference serving size and energy density of each study food.RESULTS:Nutrition labeling and the Food Guide presented reference servings with different sizes and energy values. The serving sizes reported on the labels did not follow either of the references and presented heterogeneous values, with a maximum range of 55-240 g among ready and semi-ready pre-prepared dishes. The reported energy values were between 0.1 times smaller and 2.4 times larger than the reference values. The reference serving sizes presented a highly inverse correlation with the reference energy density (Spearman coefficient= 0.9) and a very low inverse correlation with the energy density of the foods analyzed (Spearman coefficient= 0.2).CONCLUSION:This study showed the need for standardizing reference serving size information for the Brazilian population as well as reviewing nutrition labeling legislation in order to standardize the serving sizes reported on labels and to update the reference energy density used to calculate serving sizes.