Alimentos mais consumidos no Brasil : Inquérito Nacional de Alimentação 2008-2009 Most consumed foods in Brazil : National Dietary Survey 2008-2009

RESULTADOS: Os alimentos mais frequentemente referidos pela população brasileira foram arroz (84,0%), café (79,0%), feijão (72,8%), pão de sal (63,0%) e carne bovina (48,7%), destacando-se também o consumo de sucos e refrescos (39,8%), refrigerantes (23,0%) e menor presença de frutas (16,0%) e hortaliças (16,0%). Essa confi guração apresenta pouca variação quando se consideram os estratos de sexo e faixa etária; contudo, observa-se que os adolescentes foram o único grupo etário que deixou de citar qualquer hortaliça e que incluiu doces, bebida láctea e biscoitos doces entre os itens mais consumidos. Alimentos marcadamente de consumo regional incluem a farinha de mandioca no Norte e Nordeste e o chá na região Sul. Houve discrepâncias no consumo alimentar entre os estratos de menor e maior renda: indivíduos no quarto de renda mais elevada referiram sanduíches, tomate e alface e aqueles no primeiro quarto de renda citaram os peixes e preparações à base de peixe e farinha de mandioca entre os alimentos mais referidos.

The profound social, economic and cultural changes which have taken place in the country in the last few decades have led to alterations in dietary habits and behavior; for example, the increase in away-from-home eating, 4 increased consumption of processed food 10 and the replacement of traditional meals and dishes with snacks which are high in calories, fat, added sugar and salt. 5Such dietary characteristic are associated with conditions related to nutrition and metabolism, such as obesity, cardiovascular disease, high blood pressure, diabetes and cancer, 9 which are the main causes of death in Brazil. 13is situation leads to the need to diagnose and monitor patterns of food consumption at an individual level, as these trends can be predictors of the population's nutritional and health situation and act as an early warning system for creating health and nutrition policies and campaigns, impossible without the, nationwide dietary surveys regularly carried out.
The aim of this study was to characterize the foods most frequently consumed by the Brazilian population.

METHODS
Data were analyzed for Brazilian individuals of both sexes, aged over ten years old, selected to take part in the INA (National Dietary Survey) which is part of the 2008-2009 Household Budget Survey, carried out by the IBGE (Brazilian Institute of Geography and Statistics).Details on the sample and data collection can be found in IBGE publications.a The Household Budget Survey 2008-2009 (POF) a used a two-stage cluster sample, after the geographical and statistical stratifi cation of the primary sample units based on the 2000 Brazilian Demographic Census.In the fi rst stage, primary sample units were selected with probability proportional to the number of residences in each sector, and a subsample of sectors was selected by simple random sampling in each strata.In the second stage, households were selected using simple random sampling without replacement in each of the primary sample units selected.The primary sample units were evaluated throughout the 12 months of the survey so that, in the four three-month-periods, all of the geographical and socio-economic strata were represented.A subsample of 25% of households from the original 2008-2009 Household Budget Survey sample was randomly selected for investigating individual food consumption.All inhabitants aged over ten resident in the subsample participated Food consumption was estimated using food records from two non-consecutive days, in which individuals reported all of the food and drink consumed on the days in question, the time, the quantities consumed according to portion sizes, cooking method as well as the source of the food (at home or away-from-home).Additional data on sugar and/or sweetener consumption was also collected.When the interviewee was unable to complete the food record, these were completed by another resident in the household, or someone indicated by the interviewee.
The food records were checked by the interviewers in the participants' home, at the same time as the data were stored in a program specially developed for this research.This program contained around 1,500 items (food and drink) from the 2002-2003 POF database.However, the interviewers were able to add other food and drink which did not appear in this database.The program also contained methods of preparation and pre-defi ned portion sizes.Partial analyses were carried out during the data collection, to control the quality by verifying frequencies of response, mean of items consumes in the fi rst and second days' food records, codifying items which were not registered and analyzing items incorrectly entered, among others.Details on the pre-test, training, validation of the instrument and data input were published by the IBGE.b When verifying the reliability of the data, 29 individuals were excluded as their records were considered to be incomplete.In addition, quantities considered to be improbable were entered and the data entered registered in the database.This took into consideration mean portions estimated for each food according to previously defi ned statistical criteria, which included characteristics such as the gender and age group of the participant, the location and the measurement unit reported.b INA participants reported 1,971 food items, which were organized into 21 groups.Prevalence of consumption was estimated for the foods/groups reported on the fi rst day's record, according to gender, age group (from ten to 19: adolescents; 20 to 59: adult; 60 and over: elderly), region (North, Northeast, Midwest, Southeast and South) and income quartiles. 8The fi rst day's food records were as the quality of data on the fi rst day was superior to that of subsequent days' food records. 16The 20 most prevalent foods/groups are described for each strata considered in the analysis.b Income was characterized by estimating total income and mean monthly household variation in assets per capita, which represents the sum of monthly monetary gross income, monthly non-monetary gross income and the variation in household assets, divided by the number of individuals in the household.The variable monthly household income per capita was analyzed per quarter.
Analysis according to gender revealed that men and women had similar rates of prevalence or consumption.Men consumed higher levels of rice, beans, beef and soft drinks, whereas women had higher prevalence of consuming coffee, bread, juice and soft drinks, cake and oil and fat.Potatoes and fried and roasted snacks were among the most prevalent items for men, and soup and broth and sweets were among the most prevalent among women.However, neither potatoes nor soups and broths appeared among the 20 most prevalent items in Brazil (Figure 1).
Analysis stratifi ed by region confi rmed that across the entire country, rice, beans, coffee, bread and beef are the fi ve most commonly consumed foods.Coffee was the most prevalent item in the North (82.8%) and Northeast (83.6%), and rice was the most prevalent  2).
Rice, beans, coffee, bread and beef were the fi ve most commonly consumed food items observed in all household income quartiles.Coffee was the most prevalent item in the fi rst quartile (83.4%) and rice was the most prevalent in the other income quartiles.
Juice occupied sixth position in all income quartiles and soft drink consumption varied from 14.4% in the lowest income quartile (14 th position) to 31.1% in the group with the highest income (eighth position) (Figures 3A and 3B).
The only fruit included among the most prevalent foods for all income categories was the banana, with prevalence rates of consumption varying between 11.7% (lowest income group) and 22.7% (group with the highest income).The 25% with the lowest household income did not report any vegetables among the 20 most commonly consumed foods; in the second and third quartiles salad was reported with prevalence rates of 15.5% and 18.3% respectively.The 25% of the population with the highest levels of income included lettuce and tomato, as well as salad, among the 20 most commonly consumed foods (Figures 3A and 3B).
Prevalence of whole milk consumption varied from 10.9% in the fi rst quartile to 14.1% in the third income quartile; however, milk was not included among the 20 most commonly consumed foods in the highest income group.The same was observed in the case of eggs and sweets.Corn and corn based dishes appeared among the most prevalent foods only in the fi rst and second household income quartiles, whereas cheese and potatoes were among the most prevalent foods in the two higher income quartile (Figures 3A and 3B).
It was only in the fi rst income quartile that the most commonly reported foods consumed included fresh fi sh, dishes made with fresh fi sh and manioc fl our; whereas corn and corn based dishes only appeared among the most prevalent foods in the fi rst and second income quartiles.Fried and roasted snacks did not appear among the most prevalent foods in the quartile with the lowest income, whereas sandwiches were only among the most prevalent in the quartile with the highest income (Figures 3A and 3B).

DISCUSSION
The basic Brazilian diet is characterized by the presence of fruit juices and soft drinks, as well as consumption of coffee and bread, rice, beans and beef, and low levels of fruit and vegetables.Although this confi guration presents few variations when stratifi ed by gender and age group, it was observed that adolescents were the only group which failed to report any vegetables and which included sweets, milk drinks and cookies among the most commonly consumed items.On the other hand, the elderly were the only group to include a greater number of fruit and vegetables among the most prevalent foods.However, it is possible that the prevalence of oil, fat and other ingredients used in food preparation was underestimated, as these items were not reported in isolation when they were part of a dish.The elderly Some foods have a markedly regional character to their consumption, such as manioc fl our in the North and Northeast and teas in the South.In the North, milk was not included among the most prevalent foods, and deep-fried and backed snacks did not appear in the most commonly consumed foods in the Northeast.When income categories are taken into consideration, important discrepancies can be observed between low *fresh fi sh and fi sh dishes and high income quartiles.Individuals in the highest income level reported consuming sandwiches, tomatoes and lettuce and those in the fi rst quartile cited fi sh and manioc fl our among the most prevalent foods, food types which were also reported with high prevalence in the North and Northeast regions.This similarity may be explained by the fact that 44% of individuals in the North and 52% in the Northeast are in the fi rst income quartile, proportions which are much higher than those observed in the Southeast (17%), South (16%) and Midwest (23%) of the country.
It is difficult to compare data on individual food consumption obtained in the INA with data on the availability of foods in the home obtained from the same 2008-2009 POF, as it refers to food acquired for the household and does not include food consumed away-from-home.On the other hand, the INA did not obtain information on children aged under ten.However, we note that foods with the highest prevalence of consumption by individuals also represent an important part of the calories of food acquired for the household, especially rice and beans, which together represent 22% of the total calories available in the home.Beef, the most commonly consumed meat, contributed 4% to the total calories consumed in the home and represented a third of calories provided by meat in the assessment of food available in the home.As for bread, it was verifi ed that its household availability corresponded to 6% of total calories.However, the items which present the highest prevalence rates of consumption: rice, beans, beef and bread, together correspond to approximately 32% of calories available in the home.At the other extreme of prevalence rates, it can be seen that fruit and vegetables contribute only 2.8% of total calories available in the home, which concords with the low prevalence of consumption found for these foods.However, discrepancies were also observed when trying to collate data on availability in the home with those of individual consumption, particularly with relation to soft drinks, which contribute around 2% of calories available in the home.The frequency presented in the Household Budget Survey 2008-2009 suggests that there is signifi cant consumption of these products away-fom-home.a Regional characteristics highlighted in individual consumption confi rm the differentiated profi les of availability of food which characterize the fi ve main regions of the country.Thus, for example, the high prevalence of manioc fl our consumption in the North and Northeast, the high rates of prevalence of fi sh and low rates of milk in the diet in the North, the high prevalence of rice in the Midwest and of soft drinks in the Southeast, found in both surveys.The effect of household income was similar for the two approaches, individual and household.Both in the availability of the foods in the home and in individual consumption, an increased presence of fruit and vegetables, soft drinks and an inverse tendency to consume beans and manioc fl our was observed as income increased.a The 6%) and candy (95.09%), with higher proportions of regular consumption of beans and milk reported in males, and of candies, cookies and sausages among females.Low levels of fruit and vegetable consumption were also reported, with more than 20% of adolescents reporting that they had not consumed these foods in the week prior to the study. 8In contrast to our results, which suggest higher levels of soft drink consumption according to income, the results of the National Pupils' Health Survey 8 found an inverse relationship between soft drink consumption and socio-economic status (mother's level of schooling and range of goods and services) among adolescents.
Population based studies carried out with adolescents in other countries have also demonstrated high levels of fast food consumption and low levels of fruit and vegetable consumption.Bauer et al 3 (2008) verifi ed that on average, 31% of adolescents consume fast food at least three times a week.In Canada, Storey et al 14 (2009), analyzed adolescents in public schools and observed that fewer than 50% met the recommendations for fruit and vegetable and dairy consumption.
It is possible that there is some degree of under reporting in the Household Budget Survey.However, the methods used in this survey were validated in a study which adopted the doubly labelled water method c as the gold standard for estimating energy expenditure.In the validation study, 79 adults (31 men and 48 women) aged between 20 and 59 were investigated.These individuals constituted a subsample of a population based survey developed in Duque de Caxias, Southeastern Brazil.The results indicated that concordance of the reporting varied according to gender, age and nutritional state.Sub-estimation of calorie consumption was, on average, 17%; however, individuals whose weight was within the normal band presented underestimates of around 13% and those who were in the obese category presented underestimates of approximately 30%.c These results were comparable to those obtained in other assessments of dietary records against the doubly labelled water method carried out in Brazil.Among elderly residents in urban areas, the percentage of under reporting of calorie consumption was 17.7%. 6In studies carried out with women from two universities in the city of São Paulo, Southeastern Brazil, underestimating calorie consumption varied between 26%, 28% and 32% when consumption was assessed by 24hr recall, food record and a food frequency questionnaire respectively were compared with the doubly labelled water method data. 12e results presented indicate that the Brazilian diet is becoming characterized by the introduction of processed food of high calorie content and drinks with added sugar, although the traditional dietary habits are still maintained.This pattern of consumption is compatible with the high levels of overweight individuals with metabolic disorders which has marked the epidemiological and nutritional scene in the country.
Our fi ndings reaffi rm the importance of dietary and nutritional policies which involve encouraging the consumption of healthy foods, such as fruit, vegetables and whole grains and maintaining consumption of traditional basic foods, such as rice and beans, while, at the same time, encouraging reduced consumption of processed foods high in sodium, saturated fat and sugar.
The development of proposals for interventions in dietary habits aimed at adolescents could be considered urgent and of top priority, as this age group presented consumption marked by low fruit and vegetable intake and high consumption of foods considered to be unhealthy, such as cookies and soft drinks.

INTRODUCTIONa
Instituto Brasileiro de Geografi a e Estatística.Pesquisa de orçamentos familiares, 2008-2009.Aquisição alimentar per capita.Rio de Janeiro; 2010.b Instituto Brasileiro de Geografi a e Estatística.Pesquisa de orçamentos familiares, 2008-2009.Análise do consumo alimentar pessoal no Brasil.Rio de Janeiro; 2010. in the National Dietary Survey.b From a total of 55,970 households selected for the 2008-2009 POF, 34,003 individuals aged over ten participated in the assessment of individual food consumption.

Figure
Figure 2. Most frequently consumed foods according to age group, excluding rice, beans, coffee and bread.Brazil, 2008-2009.

c
Lopes TS, Ferrioli E, Pfrimer K, Hoffman D, Sichieri R, Pereira RA.Validation of energy intake estimated by the food record applied in a Brazilian national individual dietary survey by the double label water method.In: II World Congress of Public Health Nutrition, 2010, Porto, Portugal.Public Health Nutrition; 2010;13:326.
the Southeast, South and Midwest.Juice was also among the most commonly consumed items in all of the regions, varying between sixth (Northeast, South and Midwest) and seventh (North and Southeast) position.Soft drinks were the eighth most prevalent item in the Southeast, and tenth most prevalent in the South and Midwest; in these regions the prevalence of consumption of this item varied between 245 and 27%.In the North and Northeast, the prevalence rate of consuming soft drinks was 19.7% and 16.1%, corresponding to 11 th and 14 th position respectively.Crackers and eggs were among the most prevalent items in all of the regions except for the South, and milk was among the most prevalent items except in the North.6% of adults (19 th position) and 15.8% of the elderly (14 th position).Soft drinks and deep-fried and backed snacks was among the most commonly consumed items for adolescents and adults only, and salad and cheese were among the most prevalent only for adults and the elderly.Sweets, milk-based fl avored drinks and cookies only appeared among the most commonly consumed foods for adolescents, occupying 11 th , 16 th and 19 th position in the ranking respectively (Figure Women Figure 1.Most commonly consumed foods in Brazil, according to sex.Brazil, 2008-2009.inSome items, which appeared among the most prevalent foods only in some regions, demonstrate local dietary habits: for example, manioc fl our in the North (sixth position, 45.3%) and Northeast (13 th position, 18.2%).Fish and fi sh based dishes appeared among the most prevalent foods in the North (tenth position, 21.6%).Tea was among the most prevalent items in the South (15 th position, 16.7%).Sweets were among the most prevalent foods in the North, Northeast and South, with prevalence rates around 13%. Potatoes were among the most prevalent foods in the South (13.1%) and Southeast (17.2%) and cheese in the South, Southeast and Midwest, with rates of prevalence varying between 13% and 20%.Deep-fried and baked snacks were only not among the most prevalent foods in the Northeast.prevalent fruit in all age groups and it was only among the elderly that another fruit made it into the top 20 most prevalent items: the orange.Adolescents did not have any vegetables at all among the top 20 most commonly consumed foods; adults included salad; the elderly, however, reported lettuce in addition to salad.Whole milk was reported by 12.9% (18 th position) of adolescents, 11.
Most frequently consumed foods according to income quartile, excluding rice, beans, coffee and bread.Brazil, 2008-2009.Most frequently consumed foods according to income quartiles, excluding rice, beans, coffee and bread.Brazil,  2008- results of this study corroborate fi ndings of other localized studies, such as those observed by Andrade et al2 (2009), which reported that rice, beans, coffee, beef and bread occupied fi rst, second, fourth, seventh and 11 th positions, respectively among the most commonly consumed foods for women aged over 35 living in the city of Rio de Janeiro in 2004-2005.Other items most commonly consumed included juice (12 th position), milk (16 th position) and soft drinks (20 th position).