Internationally accepted nutrient-based criteria were applied to analyze data obtained in the first nationwide Brazilian individual dietary survey, showing that a small number of SoFAS food groups account for a large proportion of the intake of total energy, saturated fat, trans-fatty acids, and added and total sugar: beverages, sweets and desserts, meats, milk and dairy, and fats and oils. In general, SoFAS foods have an important contribution to food consumption in Brazil, accounting for a large proportion of the intake of total energy, saturated fat, and sugar.
SoFAS foods are mostly processed foods and mixed dishes cooked with the addition of fats and/or sugar. For example, filled rolls were classified as SoFAS breads while allowable breads included light white and whole wheat loaves and rolls. Also, SoFAS milk and dairy included sugar added flavored milk/soy/yogurt/whey based beverages while allowable milk and dairy included low fat and skimmed milk.
Adolescents presented greater proportionate intake of SoFAS foods and greater consumption of SoFAS beverages and sweets and desserts than was observed for adults and elders. Andrade et al.
analyzed data from a population-based cross-sectional study developed in Rio de Janeiro in the mid-1990 and observed that about one quarter of total energy intake were provided by sodas and high energy-dense foods, including French fries, chocolate flavored milk, cake, cookies, and sweets and desserts. Food habits of Brazilian and US adolescents are comparable, with desserts and sodas contributing most to total energy intake, desserts and pizzas to solid fat intake, and sodas to added sugar intake in US adolescents (10)
The results for the most commonly consumed allowable foods (rice and corn dishes, beverages and legumes) confirm findings of studies that have previously analyzed food consumption patterns in Brazil. Using principal components analysis of data from the 2002–2003 Brazilian Household Expenditures Survey, a common dietary pattern based on rice and beans, caffeinated beverages (coffee and tea), and vegetable oils was observed across all regions of Brazil (24)
. This traditional Brazilian dietary pattern, also identified in other studies (25–27)
, is recognized as healthy and has been associated with favorable weight outcomes in an 18-month randomized trial (28)
and with reduced body mass index and waist circumference in low-income women living on the periphery of Rio de Janeiro (29)
The high contribution of red meat to energy intake indicates a deleterious aspect of the Brazilian diet; this finding is consistent with recent observations on high consumption of red and processed meat in Brazil (30)
. Consumption of red meat has been associated with an increased risk of total, CVD, and cancer mortality in two US prospective studies (31)
, with colorectal cancer and with higher levels of oxidative stress biomarkers in European prospective studies (32, 33)
. Additionally, the increased consumption of red meat negatively impacts the environment as a result of deforestation for cattle grazing, emission of greenhouse gases, increased water pollution, and biodiversity loss (34, 35)
The results in this study suggest that the intake of SoFAS in Brazil is excessive and much higher than the range proposed by Maillot and Drewnowski (36)
, who suggested that a healthy diet can have between 17 and 33% of energy from SoFAS foods (37)
. According to the USDA, 35% of energy intake in the United States is provided by SoFAS foods (38)
, which is lower than the estimated participation of SoFAS foods in Brazil. However, the criteria to classify SoFAS was more conservative in this analysis, as the cut-off limit for SAFA intake was based on 7% of total energy intake instead of the 10% used by the USDA; additionally, the ‘level of insignificance’ criteria were applied to specific food groups.
The importance of processed foods in the Brazilian diet has been evidenced. Monteiro et al
showed that ultra-processed foods, including breads, crackers, cookies, sweets, soft drinks, sausages, cheeses, preserved meat, ready-to-eat meals, mayonnaise and sauces provided 28% of the energy available in the Brazilian households in 2002–03. Additionally, between 2002–03 and 2008–09, possibly as a consequence of the increase in away-from-home food consumption, there has been a reduction in the overall household food availability, except for certain food groups, like beverages, bakery products, and ready-to-eat meals, which include ultra-processed items (39)
This study represents the first examination at the individual level with the use of in-depth dietary intake data of the impact of low nutritional quality foods in the Brazilian diet. The use of individual dietary intake data provides greater ability to create a more detailed analysis of each food and hence a more detailed food categorization system. While the purpose of the broad classification suggested by Monteiro et al
is different and innovative, the criteria proposed in this study are based on internationally accepted and scientifically defined dietary recommendations, which are linked with a large diet and disease literature (11, 22)
. At the same time, it is highly likely that all the ultra-processed high sugar and fatty foods classified by Monteiro are included as SoFAS. The food classification system used in this paper provides guidance on foods that must have a limited consumption or must be included in the diet with caution.
Recommended limits of SAFA, TFA, and AdS intake are defined for the whole diet and not for single foods (11, 22)
. Given the fact that the diet encompasses a variety of foods, and that a considerable proportion of those foods do not contain the restricted components, we elected to allow values which were 30% over the cut-off points. As pointed out by Roodenburg et al.(21)
, this value (30%) is a starting point and future studies should test its consistency and adequacy to the Brazilian dietary pattern. We did not go a step further as was done in the original study (21)
and categorize foods into basic and discretionary as that did not change the categorization in excessive content of saturated fat, trans
-fat and added sugar.
In order to be properly translated into food and nutrition policies and public health messages, the proposed criteria could be improved, for example, by incorporating favorable aspects of foods, like the fiber content, noting that fresh vegetables and fruits are not included in this classification because their consumption should be encouraged. The criteria applied to classify foods in this work are easily understood and can be universally applied as long as the cultural context and the particular public health scenario are considered.
There is another complexity when it relates to nutrient-related cutoffs. Selected foods would be classified as SoFAS if consumed in excessive amounts; however, they represent important sources of beneficial fats, vitamins, minerals and bioactive components, for example, nuts and seeds (41–43)
. Such foods are not extensively consumed in Brazil, yet their consumption could favor individuals adopting diets aiming to reduce weight or to improve health. In Brazil the deficiency of vitamin A is still a significant public health problem (44)
, thus foods like eggs and whole milk, which are important sources of vitamin A, deserve special consideration. Choices of reduced fat milk and cheeses, lean cuts of meats, cooking methods that require less fat/oil (for example: steaming, boiling, baking, grilling) could help to reduce the amount of SoFAS in the Brazilian diet. Therefore, food selection must consider the helpful combination of nutrients and other components of the foods, and should be guided by principles involving the amount and frequency of consumption. By clearly defining the foods that should have limited consumption (and those that should be encouraged) in each food group, the proposed food profiling criteria provide helpful guidance on healthy food and nutrition and offer directions for health oriented food-processing industry.
This study is not without limitations. First, table sugar consumption was not directly obtained. The amount of sugar in coffee, tea, and fruit-based drinks was standardized to 10% for sugar only and 5% for sugar plus artificial sweetener consumption. Levy et al
analyzed household sugar availability data obtained in the Brazilian 2002–2003 HBS and concluded that 75% of the calories from sugars came from “refined sugars and other caloric sweeteners” while 25% came from the sugars added to processed foods. Thus, sugar intake might be based on biased estimates, which is very important given the risen rates of obesity, diabetes, and other metabolic disorders in Brazil.
Second, despite intense efforts to obtain reliable data on food composition, for some foods, the nutritional composition was estimated based on similar foreign foods or preparations. The latest version of the Brazilian Food Composition Table (TACO) contains information on the nutritional composition of about 600 food items (18)
, while approximately 2000 foods and preparations were cited in the Brazilian IDS. Finally, although this analysis is based only on the first day of food records, it is recognized that single 24-h recalls and food records provide decent estimates for population means in extent studies (46)
There are several strengths of this study as well. The Brazilian IDS food record was evaluated and provides an accurate estimation of energy intake (47)
. Additionally, the estimates for the intake of energy and nutrients were comparable with data obtained in similar studies (37, 48, 49)
The present study found very high levels of SoFAS foods consumption in Brazil. Interventions aimed at improving overall diet quality are necessary and should take in consideration foods containing excessive saturated fat, trans-fatty acids, and added sugars, making the consumption of SoFAS foods a major target for Brazilian food and nutrition policies.