SciELO - Scientific Electronic Library Online

vol.30 issue10Social class, psychosocial occupational risk factors, and the association with self-rated health and mental health in ChileBASTA APLICAR UMA INJEÇÃO? DESAFIOS E CONTRADIÇÕES DA SAÚDE PÚBLICA NOS TEMPOS DE JK (1956-1961). Muniz ES. Belo Horizonte: Fino Traço/Campina Grande: EDUEPB/Rio de Janeiro: Editora Fiocruz; 2013. 104 p. author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand




Related links


Cadernos de Saúde Pública

Print version ISSN 0102-311X

Cad. Saúde Pública vol.30 no.10 Rio de Janeiro Oct. 2014 


Association between eating patterns and body mass index in a sample of children and adolescents in Northeastern Brazil

Associação entre padrões alimentares e índice de massa corporal em amostra de crianças e adolescentes do Nordeste brasileiro

Asociación entre patrones de alimentación e índice de masa corporal en una muestra de niños y adolescentes en el noreste de Brasil

Nadya Helena Alves dos Santos 1  

Rosemeire Leovigildo Fiaccone 2  

Maurício Lima Barreto 3  

Luce Alves da Silva 1  

Rita de Cássia Ribeiro Silva 1  

1Escola de Nutrição, Universidade Federal da Bahia, Salvador, Brasil.

2Instituto de Matemática, Universidade Federal da Bahia, Salvador, Brasil.

3Instituto de Saúde Coletiva, Universidade Federal da Bahia, Salvador, Brasil.


The aim of this study was to assess the relationship between eating patterns and body mass index (BMI) in children and adolescents. This is a cross-sectional study of 1,247 male and female students, aged between 6 and 12, from public elementary schools in São Francisco do Conde, Bahia State, Brasil. BMI was used to analyze the children’s nutritional status. Food consumption frequencies, in addition to demographic and socioeconomic information, were collected for each participant. Dietary patterns were identified through a factor analysis. The prevalence of overweight and obesity was 17.3% (10.2% overweight and 7.1% obese). Two eating patterns, “obesogenic” and “prudent”, were identified. The former is characterized by sweets and sugars, typical Brazilian dishes, pastries, fast food, oils, milk, cereals, cakes, and sauces, and was positively associated with increased BMI (ßi = 0.244; p = 0.018). An “obesogenic” dietary pattern was associated with increased BMI.

Key words: Feeding Behavior; Body Mass Index; Adolescent; Child


O objetivo deste estudo foi identificar a associação entre padrões alimentares e índice de massa corporal (IMC) em crianças e adolescentes. Estudo transversal realizado em amostra de 1.247 estudantes entre 6 a 12 anos, de ambos os sexos, matriculados na rede pública de ensino de São Francisco do Conde, Bahia, Brasil. Para avaliar o estado nutricional foi utilizado o IMC. Informações de frequência de consumo alimentar, além das demográficas e socioeconômicas foram obtidas para cada participante. Os padrões alimentares foram obtidos a partir de análise fatorial. A prevalência de excesso ponderal foi de 17,3% (10,2% de sobrepeso e 7,1% de obesidade). Foram encontrados dois padrões alimentares: padrão “obesogênico” e “prudente”. O primeiro, caracterizado pelo consumo de doces, pratos típicos brasileiros, pastelarias, fast food, óleos, leite, cereais, bolos e molhos, esteve positivamente associado ao aumento do IMC (ßi = 0,244; p = 0,018). Os resultados apontaram associação do padrão obesogênico com aumento do IMC.

Palavras-Chave: Comportamento Alimentar; Índice de Massa Corporal; Adoslecente; Criança


El objetivo de este estudio fue identificar la asociación entre patrones dietéticos e índice de masa corporal (IMC) en niños y adolescentes. Se trata de un estudio transversal, realizado en muestra de 1.247 estudiantes entre 6 a 12 años de edad, de ambos los sexos, inscritos en la red pública de enseñanza de São Francisco do Conde, Bahía, Brasil. Para evaluar el estado nutricional se utilizó el IMC. La información sobre la frecuencia de consumo alimentario, además de la demográfica y socioeconómica, se consiguió con cada participante. Los patrones dietéticos fueron obtenidos a partir del análisis factorial. La prevalencia de exceso ponderal fue de 17,3% [10,2% de sobrepeso y 7,1% de obesidad]. Fueron encontrados dos patrones dietéticos: “obesogénico” y “prudente”. El primero, caracterizado por el consumo de azúcares, platos típicos brasileños, pastelerías, fast food, aceites, leche, cereales, pasteles y salsas estuvo asociado al aumento del IMC (ßi = 0,244; p = 0,018). Los resultados apuntaron asociación del patrón dietético “obesogénico” y aumento del IMC.

Palabras-clave: Conducta Alimentaria; Índice de Masa Corporal; Adoslecente; Niño


Obesity is a chronic, multifactorial disease characterized by the accumulation of fat tissue either locally or throughout the body as a result of the positive difference between energy consumption and expenditure 1. Excess weight in childhood and adolescence has been associated with adverse short-term and long-term health effects. These effects include an increased risk of cardiovascular disease and factors related to metabolic abnormalities, such as dyslipidemia, glucose intolerance and an increased likelihood of being obese as an adult 2. Other symptoms such as sleep apnea 3,4 as well as psychological and social repercussions have been reported in the literature 5.

Overweight and obesity are becoming the major nutritional problems in the modern world and are rapidly increasing in many developed and developing countries 6. In the United States, the prevalence of obesity among 6- to 11-year-old children has increased over the last 30 years from 7% to 20%, and among 12- to 19-year-old adolescents, the prevalence has increased from 5% to 18% 7. In Brazil, national surveys conducted over the past few decades have also shown an increase in overweight and obesity among Brazilian children and adolescents, suggesting an epidemic trend. The Household Budget Survey (HBS) conducted in 2008/2009 estimated that 30% of children between 5 and 9 years of age and approximately 20% of 10- to 18-year-olds had excess weight 8.

Genetic, physiological, psychosocial and metabolic factors may be associated with excess weight. However, the factors that explain the increase in the number of obese children appear to be specifically related to lifestyle changes 9,10. According to the literature, an obesity epidemic in youth appears to be the result of changes in modern society. One of the most commonly mentioned changes is the increased availability of highly energy-dense manufactured foods, especially those rich in saturated fat and simple carbohydrates, at the expense of plant-based foods, combined with few opportunities for physical activity 11. These changes, mainly imposed by the new “lifestyle” experienced in the last few decades, have increased the risk factors associated with non-transmissible chronic diseases 12.

Most studies use food, alone or associated with micronutrient ingestion to study the association between diet and chronic disease; nutrients and foods are not consumed in isolation but in combination. As a result, the World Health Organization (WHO) 13 has suggested that assessments of food intake in population-based studies on nutrition should be based on eating patterns. Dietary patterns can summarize the combined and potentially synergistic effects of a repertoire of foods contributing to usual dietary intake in a defined population. Several nutritional epidemiology researchers have investigated the link between eating patterns and excess weight and have found a positive correlation between inadequate eating patterns and higher body mass index gain in children and adolescents 14,15,16,17. However, this relationship was not observed in other studies 18,19,20. Overall, studies on the relationship between eating patterns and weight status during this phase of life are not consensual.

In order to support and strengthen the knowledge in this field, this study sought to explore the relationship between eating patterns and weight status, based on an exploratory factor analysis, with the main component estimation method used to measure eating patterns. Our hypothesis is that a higher adherence to dietary patterns that are characterized by foods that are energy-dense, high in fat and low in fiber predispose young people to an increased body mass index. These results may contribute to the creation and implementation of strategies for promoting healthy eating and better health in children and adolescents.


Study design/population/sample

A cross-sectional design was used to study 6- to 12-year old children living in São Francisco do Conde, a municipality located in the metropolitan region of Salvador, Northeast Brazil. This municipality has 33,183 inhabitants and a high urbanization rate (80.2%). Economically, the municipality has the third highest development index in the state of Bahia, and the city council is the largest local employer. However, this region experiences challenges in certain areas, including its social development index (30th), level of education (139th) and health indicators (178th) 21.

We used data from the São Francisco do Conde Municipal Education Department for the year 2010 to calculate the sample size. Of 3,734 registered students, 2,649 were from rural areas and 1,085 were from urban areas. Considering the prevalence of respiratory allergies and overweight of 25% 22,23, a sample size calculation with a 3% error and a 95% level of confidence resulted in samples of 531 and 775 students in urban and rural areas, respectively. However, these students were distributed across 22 schools in the county school system. Thus, to minimize travel costs and time for subject recruitment, we chose (not at random) to include in our study all students from schools that had at least 150 or more students. Therefore, only nine schools attended this requirement. All students aged 6-12 years in each school were eligible for the study. In addition to this, we added 15% to the total sample size to account for students who chose not to participate in the study, resulting in a total of 1,500 students being enrolled in the study.

Response variable: body mass index

• Anthropometric data

Each participant’s weight was obtained using a Master portable digital scale, and height was measured using a Leicester Height Measure portable stadiometer (Seca, Hamburg, Germany). The weight of the uniform (100g) was subtracted during the analysis. To assess anthropometric status, tables from the WHO 24 with percentile values of body mass index [BMI = weight (kg)/height (m)2] according to age and sex were used as reference. For classification to an anthropometric status, we used the WHO 2006 25 proposal: underweight (< 3rd percentile); normal weight (≥ 3rd percentile and < 85th percentile, reference category); overweight (≥ 85th percentile and < 97th percentile) and obese (≥ 97th percentile). The overweight and obese categories were aggregated. Therefore, children with excess BMI were situated on or above the 85th percentile.

Principal independent variable

• Eating patterns

The items of the quantitative food frequency questionnaire (FFQ) were included based on a study developed by Borges et al. 26. The FFQ was validated recently by Mascarenhas 27. The average energy and nutrient values from the FFQ were compared with those from a three-day food record through the paired t test and Pearson correlation coefficients. The Pearson correlation coefficients, after being adjusted and corrected for variability, ranged from 0.27 to 0.99 [carbohydrates (0.41), proteins (0.62), lipids (0.44), zinc (0.99), magnesium (0.45), calcium (0.33), fiber (0.27) iron (0.40)]. This questionnaire was applied to the students’ parents or guardians by trained nutritionists and nutrition students in meetings together with the students. Information missing in the parents’ or guardians’ answers was completed by the students.

The FFQ contained 97 food items. This questionnaire evaluated the quantity of foods and nutrients consumed during the preceding 6 month period. The portions or home-cooking measurements reported were converted into grams or milliliters. To minimize possible sources of error (memory bias) in the food consumption information, an album with drawings of foods and utensils of different dimensions was used, and, in addition, standard liquid measurements were presented to the mothers at the time of the interview.

For the analysis, the foods were grouped into 22 food clusters according to the coded nutritional content, including: sugars and sweets; typical Brazilian dishes (feijoada: a stew of black beans with beef and pork; feijão-tropeiro: a dish made with beans, cassava flour, sausage, garlic, onion, bacon and eggs; acarajé: a dish made from peeled black-eyed peas, formed into balls, and deep fried in palm oil); soft drinks; pastries; fast food; oils; milk; beef; chicken; fish; eggs; processed meat products; breads; cereals (rice, cassava flour, and pasta); cakes; roots; baked beans, legumes; fruits; leafy vegetables; sauces; and artificial sweeteners.

Frequencies of consuming foods or food types were summarized in one unique value for each student using the method described by Neumann et al. 28. To do so, foods were initially codified according to individual consumption frequencies: never = 0; from 1 to 3 times a month = 1; once a week = 2: 2 to 3 times a week = 3; 4 to 7 times a week = 4. Then codified frequencies corresponding to foods actually consumed by the individual were summed in each food type group, resulting in the numerator of the summary measure. The denominator has been defined as the maximum number of foods that the individual could eat in each food type group, multiplied by 4.

As an example, for a particular individual, the sum of codified frequencies for the sugars and sweets group was 20. In this food type group, the denominator would be 40 (the maximum consumption is 10 food products, a number that was multiplied by 4). In this way, the score of sweets consumption for this particular individual was 20/40 = 0,5. This is how summary measures were obtained for each individual belonging to the sample.

The food intake registered by the FFQ was converted into energy and nutrients by the Virtual Nutri program 29. The daily intake frequency was calculated based on the weekly and monthly intake frequency of each food. The values, multiplied by the concentration of nutrients in the food, resulted in the daily intake value 30,31.

Confounding variables

• Physical activity level

Physical activity levels were measured by applying the short version of the International Physical Activity Questionnaire (IPAQ) in an interview 32. Guedes et al. 32 has demonstrated the reproducibility and validity of the questionnaire. The student’s parents or guardians were asked about the time and the frequency of moderate and intense activities and walking during the previous week. Information missing in the parents’ or guardians’ answers were completed by the students. For this study, the final results were divided into two groups using a cutoff of 300 minutes/week of moderate or intense activity. The children with ≥ 300 minutes of activity per week were considered active (reference category), and the children with < 300 minutes/week were classified as insufficiently active 33.

• Sedentary behavior

In addition, a questionnaire about the average time that the children spend watching television, playing video games, playing or surfing on a computer on a typical weekday and on a typical weekend day was applied 34. Time spent was summed into the total screen time. We chose to record the total time spent on this activity for the entire week. In this study, the results were divided into two categories using the median as a cutoff. Thus, this variable was dichotomized as non-sedentary (< 3.35 hours/day; reference category) or sedentary (≥ 3.35 hours/day).

• Sexual maturation

Evaluation of the stage of sexual maturity was based on breast and pubic hair characteristics in girls and genitals and pubic hair in boys. Based on this staging, adolescents were grouped according to categories described by Marshall & Tanner 35,36 into pre-pubescent (reference category) and pubescent. The identification of these stages was made by means of self-description that was supported by drawings provided by the interviewers.

Other variables

The other variables included in the study were gender (male, female – reference category), age (< 10 years old, ≥ 10 years old – reference category). Questions concerning socioeconomic characteristics were answered by the student’s parents or guardians: caregiver education (≤ 4th grade, > 4th grade – reference category), family income (1.5 times the Brazilian minimum wage, ≥ 1.5 times the Brazilian minimum wage – reference category) and household location (urban, rural – reference category). The economic classification was defined using the Brazilian Economic Classification Criteria 37, which includes the possession of domestic goods and the education level of the head of the family. The families were divided into the following economic categories (starting with the greatest purchasing power): A1, A2, B1, B2, C1, C2, D and E. Only the categories B2 reference category, C2, D and E were found.

Data processing

The data processing and database construction were performed using Epi Info, version 6.04 (Centers for Disease Control and Prevention, Atlanta, USA). The data were entered in duplicate after the questionnaires were revised, and coding errors made in the fields were corrected. Simple frequencies were verified and analyses were conducted to determine the consistency between the questions and the answers to clean up the database.

Statistical analysis

In the first instance, descriptive analyses were performed to characterize the study population using proportions for the categorical data. Before proceeding to the exploratory factor analysis and to assure that this method is adequate in order to explain the maximum variation in food intakes, the Kaiser-Mayer-Olkin (KMO) coefficient and Bartlett’s test of sphericity were used. To assess the degree of inter correlations between variables we adopted a value greater than 0.60 for the KMO 38. Factor analysis with principal component estimation method followed by an orthogonal rotation (varimax) was used for the exploratory factor structure (pattern) analysis. To identify the number of principal components to be retained, the following criteria were used: The criterion of eigenvalues exceeding 1, the scree plot (which is a graphical presentation of eigenvalues) and the interpretability of each component. The derived factors were labeled on the basis of data and a review of the literature. The factor score for each pattern was calculated by summing intakes of food groups weighted by their factor loadings, and each participant received a factor score for each identified pattern. Food groups with factor loadings greater than 0.40 and communality over 0.20 were retained in the patterns identified 38.

All component scores were approximately normally distributed and Spearman’s correlations coefficients were calculated to measure the association between pattern scores and nutrient intakes. Additionally, partial correlation coefficients were calculated, adjusting for energy intake, and thus represented the association between dietary patterns and relative nutrient intake.

Multiple linear regression analysis was used to evaluate the associations between food consumption factor scores and body mass index. The models were adjusted for gender, age, sexual development, sedentarism, physical activity level, household location, economic conditions and for energy consumption. The effects of the dietary patterns were also adjusted for each other since rotations can introduce correlations between them, although principal components are uncorrelated. Confounding variables were chosen based on current knowledge drawn from the literature review. We tested the data for interactions between dietary patterns and gender in the regression analysis.

The statistical tests were two-tailed with a 5% level of significance. The statistical analyses were performed using SPSS, version 17.0 (SPSS Inc., Chicago, USA).


Ethical approval was provided by the Ethics Research Committee of the School of Nutrition at the Federal University of Bahia, Brazil. The study was conducted in accordance with the Declaration of Helsinki, as revised in 2000. Written and informed consent detailing all procedures to be carried out was signed by a parent or legal guardian of each participant.


Of the students initially selected (1,500 students), 193 (12.8%) did not participate in the study (because they refused, because their families moved to another city or because they were transferred to another school). Ultimately, a total of 1,307 students of both genders between 6 and 12 years old participated in the study. A total of 1,247 students had complete data for FFQ which was included in the analyses. There were no statistically significant differences between the original sample and the sample used in this study in terms of socioeconomic, anthropomorphic and demographic characteristics (data not shown).

There were higher percentages of male students (53.1%) and students 10 years old or older (50.2%). The other characteristics are shown in Table 1. The prevalence of weight excess was 17.3% (10.2% overweight and 7.1% obese).

Table 1 Characteristics of the study population from São Francisco do Conde, Bahia State, Brazil, 2010. 

Variable n %
Age (years) *    
 < 10 609 49.8
 ≥ 10 614 50.2
 Male 662 53.1
 Female 585 46.9
Caregiver education *    
 Illiterate 49 4.2
 Elementary school-cycle I 306 25.9
 Elementary school-cycle II 302 25.6
 Middle school and above 523 44.3
Family income (minimum wage **) *    
 ≥ 1.5 506 42.3
 < 1.5 691 57.7
Economic classification *    
 Class B 661 56.3
 Classes C, D, E 514 43.7
Household location    
 Urban 434 34.8
 Rural 813 65.2
Puberty *    
 Pre-pubescent 953 76.4
 Pubescent 294 23.6
Sedentarism *    
 Sedentary 311 25.1
 Non-sedentary 926 74.9
Physical activity level *    
 Active 985 79.2
 Inactive 258 20.8

* Different totals because of no available data;

** Value in 2010: BRL 510.00, equivalent to US$ 290.70.

Both the KMO index (0.864) and Bartlett’s test (χ2 = 4335,619; p < 0.001) indicated that the correlation among the variables was sufficiently strong for a factor analysis. To reduce bias as a result of multiple testing and to better identify common dietary patterns, only the dietary patterns with an eigenvalue of > 1.5 were extracted. This cut-off was made on the basis of the scree plots, which indicated a clear break after the second factor (i.e. dietary pattern) with an eigenvalue of 1.6.

The factor analysis revealed two eating patterns that were responsible for 47.89% of the total variance. These eating patterns were designated “obesogenic” and “prudent”. The factor loadings of the eating patterns for each component are shown in Table 2. The first component was positively correlated with the intake of sugars, typical Brazilian dishes, pastries, fast food, oils, milk, cereals, cakes and sauces. The second component was positively correlated with roots, legumes, fruits and leafy vegetables. Some foods and food groups were not included in the analyses because of the low communalities (h2 < 0.20) (i.e. the proportion of variance of each variable that could be explained by the factors): soft drinks, beef, chicken, fish, eggs, processed meat products, bread, baked beans, and artificial sweeteners.

Table 2 Distribution of the factor loadings for the food consumption patterns of 6- to 12-year-old children and adolescents enrolled in the public school system of São Francisco do Conde, Bahia State, Brazil, 2010. 

Foods and food groups “Obesogenic” pattern “Prudent” pattern h2
Sugars 0.697 0.085 0.498
Typical Brazilian dishes 0.601 0.115 0.378
Pastries 0.586 0.113 0.351
Fast food 0.767 0.012 0.590
Oils 0.619 0.265 0.457
Milk 0.656 0.165 0.470
Cereals (rice, cassava flour, and pasta) 0.424 0.319 0.326
Cakes and filled cookies 0.531 0.017 0.290
Roots and tubers 0.253 0.641 0.460
Legumes 0.052 0.825 0.684
Fruits 0.463 0.503 0.449
Leafy vegetables 0.073 0.747 0.568
Sauces 0.640 0.100 0.423

Eigenvalues 4.090 1.658  
% Variance explained 30.25 17.63  
% Variance accumulated 47.89    

Table 3 presents the correlations between the dietary patterns score and estimated nutrient intake, both absolute and energy-adjusted. Positive correlations (Spearman coefficient) were observed between “obesogenic” dietary pattern and intakes of total fat (r = 0.52), polyunsaturated fatty acids (PUFA) (r = 0.41), monounsaturated fatty acid (MUFA) (r = 0.495), saturated fatty acid (SFA) (r = 0.436) points that were attenuated but not removed by energy adjustment. The majority of correlations with protein, fiber and micronutrients were reversed after energy adjustment. The micronutrients rather than the macronutrients were more strongly correlated with the ‘prudent’ dietary pattern. Energy adjustment did not greatly attenuate the correlations between this pattern and fiber or most micronutrients. However, correlations with some macronutrients were reversed; notably total fat, PUFA, MUFA and SFA.

Table 3 Correlation coefficients between dietary pattern scores and corresponding absolute nutrient intakes and partial correlation coefficients between dietary pattern scores and nutrient intakes adjusting for energy intake. Children and adolescents enrolled in the public school system of São Francisco do Conde, Bahia State, Brazil, 2010. 

Nutrient “Obesogenic” pattern “Prudent” pattern
Absolute Adjusted Absolute Adjusted
Energy 0.491 * - 0.224 *  
Carbohydrate 0.459 * -0.023 0.219 * 0.007 *
Total fat 0.523 * 0.083 * 0.193 * -0.118 *
PUFA 0.411 * 0.086 * 0.167 * -0.020
MUFA 0.495 * 0.121 * 0.148 * -0.062 *
SFA 0.436 * 0.132 * 0.128 * -0.115 *
Protein 0.432 * -0.143 * 0.273 * 0.123 *
Fiber 0.305 * -0.120 * 0.309 * 0.257 *
Fe 0.420 * -0.009 0.195 * 0.114 *
Zn 0.358 * -0.005 0.174 * 0.090 *
Vitamin B6 0.327 * -0.029 0.344 * 0.321 *
Vitamin B12 0.120 * -0.011 0.352 * 0.327 *
Niacin 0.351 * -0.052 0.197 * 0.120 *
Folate 0.348 * -0.057 * 0.304 * 0.276 *
Retinol 0.219 * -0.001 0.566 * 0.557 *
Na 0.417 * -0.041 0.088 * -0.082 *
Ca 0.360 * -0.069 * 0.181 * 0.094 *

Ca: calcium; Fe: iron; MUFA: monounsaturated fatty acid; Na: sodium; PUFA: polyunsaturated fatty acids; SFA: saturated fatty acid; Zn: zinc.

* p < 0.001.

Table 4 shows the regression coefficients (βi) from the multiple linear regression analysis for each food consumption factor scores and the body mass index values. There was a significant positive association between “obesogenic” dietary pattern and body mass index. This relationship was statistically significant even after adjusting for age, gender, sexual development, sedentarism, physical activity, economic variable (ABEP), household location and total energy intake (βi = 0.244; p = 0.018).

Table 4 Multiple linear regression for evaluating the relationship between eating patterns and body mass index in 6- to 12-year-old children and adolescents enrolled in the public school system of São Francisco do Conde, Bahia State, Brazil, 2010. 

  “Obesogenic” pattern “Prudent” pattern
β * p-value β * p-value
Model 0 0.232 0.008 0.003 0.977
Model 1 0.229 0.007 -0.003 0.976
Model 2 0.242 0.005 -0.006 0.941
Model 3 0.231 0.009 -0.009 0.913
Model 4 0.244 0.018 -0.025 0.776

Model 0: effects of the dietary patterns adjusted for each other since; Model 1: Model 0 + adjusted by gender, age, sexual maturation; Model 2: adjusted by gender, age, sexual maturation, sedentarism and physical activity; Model 3: adjusted by gender, age, sexual maturation, sedentarism, physical activity and economic classification; Model 4: adjusted by gender, age, sexual maturation, sedentarism, physical activity, economic classification, household location and energy intake.

* Regression coefficient.

There was no association between body mass index and the food consumption score for the students with the “prudent” dietary pattern, even after the appropriate adjustments have been made.

There was no interaction between gender and dietary patterns (“prudent” dietary pattern p = 0.373, “obesogenic” dietary pattern p = 0.507).


In this study, the influence of eating patterns on body mass index was investigated in a population of 6- to 12 year-old children living in the city of São Francisco do Conde. The prevalence of excess weight found here (17.3%) is higher than the one found in other studies in Salvador 39, or in studies conducted in Corumbá in Mato Grosso do Sul State (6.5% obese; 6.2% overweight) 40 and Belo Horizonte, Minas Gerais State (3.1% obese; 8.4% overweight) 41. However, the prevalence is lower than the one found in studies conducted in Fortaleza in Ceará State (19.5% obese and overweight) 42, and Santos in São Paulo State (18% obese; 15.7% overweight) 43.

With factor analysis, two food consumption patterns were identified: (1) the “obesogenic” pattern composed of sweets and sugars, typical Brazilian dishes, pastries, fast food, oils, milk, cereals, cakes, and sauces; and (2) the “prudent” pattern composed of roots, legumes, fruits, and leafy vegetables. From the results we can observe that the “obesogenic” pattern shows an increase in the average BMI. Thus, this pattern has a negative effect on the healthy growth of children and adolescents, especially because it includes foods with fat, saturated fat and simple carbohydrates. While it is difficult to make direct comparisons among dietary pattern studies, findings from our study were similar to those found in other studies on Brazilian population food consumption patterns 44,45; these studies have shown increased consumption of manufactured products, insufficient consumption of fruits and vegetables, and increased amounts of overall fat and saturated fat in the diet.

The WHO has suggested that epidemiologic nutritional analyses of population food consumption should be based on eating patterns rather than on nutrient intake 13. In order to evaluate the association between diet and diseases, particularly for chronic diseases, the use of dietary patterns has several advantages compared with methods focusing on nutrients and food alone. Indeed, this approach reduces the chance of finding spurious associations between exposure (dietary) and outcome (chronic diseases), and it incorporates both the complex interactions between nutrients (synergistic or antagonistic) and their correlations, which can modify bioavailability.

After necessary adjustments, the results of the multiple linear regression show a significant positive association between the “obesogenic” food consumption pattern and BMI. Cross sectional studies that analyze the relationship between eating patterns and overweight/obesity, especially in youth, showed that junk food eating patterns, characterized by high intake of sweets such as chocolates, ice cream, foods with added sugar, fried foods (French fries, hamburgers, “empanados”, popcorn, bacon), sodas and alcoholic drinks, were related to excess weight in Korean 46,47 and Spanish 48 youth populations. However, this association was not found in studies of Scottish 19 and American adolescents 20. Therefore, our results corroborated the previous findings, showing that dietary patterns that are high in energy-dense, high-fat and low-fiber foods predispose young people to overweight later. Ambrosini 49 points out in his revision paper that studies reporting positive associations between this type of dietary pattern and later obesity risk were of consistently higher quality than those reporting null associations.

Inadequate eating habits may have a large effect on the occurrence of overweight/obesity, especially in children and adolescents. It is possible that the effect of fatty foods on the BMI gain is due to their high energy density, associated with low nutrient contents. In addition, the flavor of these foods may lead to weight gain by promoting excessive consumption. Some studies have also shown that the power of satiety of fats is lower compared to carbohydrates and proteins, making it more difficult to adjust compensation after a meal rich in fat, which is directly related to excessive energy consumption 50.

Clear positive correlations, which were fairly robust to energy adjustment, were evident between the scores on the “obesogenic” pattern and fats intake. Furthermore, intakes of almost all of the micronutrients as well as fiber and protein were negatively correlated with this pattern after energy adjustment. Surprisingly an inverse relationship was seen with carbohydrate, although it was not significant. In total, this result is not surprising given that high scores on the “obesogenic” pattern are associated with a high consumption of high-fat and nutrient-poor processed foods. This pattern, characterized by an elevated intake of fat, is likely to contribute to the increasing prevalence of childhood obesity and the fact that the correlations of this pattern with energy were strong and those with micronutrients were reversed after energy adjustment suggests that it is a marker for an energy-dense, nutrient-poor diet. The “prudent” pattern showed positive relationships with fiber and many micronutrients that were robust to energy adjustment. This result was again not unexpected given that higher pattern scores are related to higher intakes of nutrient-dense foods such as fruit, vegetables and leafy vegetables. The results observed in the current study were consistent with results obtained Cribb et al. 51 and Patel et al. 52, which found a “processed” pattern that was positively associated with fats and sugar, as well as “health conscious” and “traditional” patterns that both showed positive linear relationships with most nutrients. Results from the current study and others investigating dietary patterns and nutrient intakes have the potential to inform new public health initiatives, as this method of examining dietary patterns as a whole can be more useful than focusing on single foods and/or nutrients. It is evident from the results presented here that the “obesogenic processed” pattern is the least ideal in terms of nutrient intake.

Limitations of the study

There are several limitations to this study: as a cross-sectional study, it is unable to determine causal relationships because it does not account for the temporal progression from exposure to effect. Many studies are still limited to cross-sectional designs, which only allows for the investigation of association between variables. However, several randomized controlled intervention studies lead to the conclusion that changes towards healthy eating habits positively affect weight loss in all age groups 53. It is important to highlight that semi-quantitative FFQ was validated only for adolescents. Furthermore, the nutritional intakes of children estimated via this questionnaire might not represent their true intakes and may depend on recall bias and social desirability 54, however, it is an appropriate instrument for ranking dietary intakes. Moreover, misclassification of the children’s dietary intake, resulting from administration of the FFQ to the student’s mothers, might happen especially considering foods eaten away from home. Thus, according to Drewnouski 55, these dietary data may reflect the mother’s dietary image rather than the real diet of the child. This approach was still necessary, given that cognitive skills required to estimate and recall usual intakes, as well as the knowledge of how foods are prepared, are limited in elementary school children 56. The limitation resulting from interviewing parents or guardians rather than the students was mitigated by applying the interviews in the presence of the students, asking them to complement or correct any missing or incomplete information. In addition, data driven by factor analysis depends on subjective or arbitrary decisions, for example, when grouping the food items or labeling the factors 57. Also, the decision about how many factors to retain and which correlation matrix rotation method to use in the factor analysis could represent limitations of this study, which may contribute to the inconsistency and considerably limit the ability to generalize the results. However, these decisions were made taking the aims of the study and the interpretability of the data into account, as recommended by Hearty & Gibney 58.


Despite the limitations, the results of this study support the hypothesis that the “obesogenic” pattern composed of sweets and sugars, typical Brazilian dishes, pastries, fast food, oils, milk, cereals, cakes, and sauces are correlated with increased body mass index, after adjusting for confounders. Improving diet quality may reduce the risk of obesity in young people.


This work was financially supported by FAPESB (project n. 7638/2009).


. World Health Organization. Obesity: preventing and managing the global epidemic. Geneva: World Health Organization; 2000. (Technical Report Series). [ Links ]

. Herouvi D, Karanasios E, Karayianni C, Karavanaki K. Cardiovascular disease in childhood: the role of obesity. Eur J Pediatr 2013; 172:721-32. [ Links ]

. Moraleda-Cibrian M, O’Brien LM. Sleep duration and body mass index in children and adolescents with and without obstructive sleep apnea. Sleep Breath 2014; 18:555-61. [ Links ]

. Burt J, Dube L, Thibault L, Gruber R. Sleep and eating in childhood: a potential behavioral mechanism underlying the relationship between poor sleep and obesity. Sleep Med 2013; 15:71-5. [ Links ]

. Perez LM, Garcia K, Herrera R. Psychological, behavioral and familial factors in obese cuban children and adolescents. MEDICC Rev 2013; 15:24-8. [ Links ]

. Malik VS, Willett WC, Hu FB. Global obesity: trends, risk factors and policy implications. Nat Rev Endocrinol 2013; 9:13-27. [ Links ]

. Ogden CL, Carroll MD, Flegal KM. High body mass index for age among US children and adolescents, 2003-2006. JAMA 2008; 299:2401-5. [ Links ]

. Instituto Brasileiro de Geografia e Estatística. Antropometria e estado nutricional de crianças, adolescentes e adultos no Brasil. Pesquisa de Orçamento Familiar – POF 2008-2009. Rio de Janeiro: Instituto Brasileiro de Geografia e Estatística; 2010. [ Links ]

. Gonzalez-Gross M, Melendez A. Sedentarism, active lifestyle and sport: impact on health and obesity prevention. Nutr Hosp 2013; 28 Suppl 5:89-98. [ Links ]

. Iannotti RJ, Wang J. Trends in physical activity, sedentary behavior, diet, and BMI among US adolescents, 2001-2009. Pediatrics 2013; 132:606-14. [ Links ]

. Kourlaba G, Panagiotakos DB, Mihas K, Alevizos A, Marayiannis K, Mariolis A, et al. Dietary patterns in relation to socio-economic and lifestyle characteristics among Greek adolescents: a multivariate analysis. Public Health Nutr 2009; 12:1366-72. [ Links ]

. Gubbels JS, van Assema P, Kremers SP. Physical activity, sedentary behavior, and dietary patterns among children. Curr Nutr Rep 2013; 2:105-12. [ Links ]

. World Health Organization. Report of a Joint FAO/WHO Consultation. Preparation and use of food-based dietary guidelines. Geneva: World Health Organization; 1998. [ Links ]

. Bahreynian M, Paknahad Z, Maracy MR. Major dietary patterns and their associations with overweight and obesity among Iranian children. Int J Prev Med 2013; 4:448-58. [ Links ]

. Shang X, Li Y, Liu A, Zhang Q, Hu X, Du S, et al. Dietary pattern and its association with the prevalence of obesity and related cardiometabolic risk factors among Chinese children. PLoS One 2012; 7:e43183. [ Links ]

. Rodriguez-Ramirez S, Mundo-Rosas V, Garcia-Guerra A, Shamah-Levy T. Dietary patterns are associated with overweight and obesity in Mexican school-age children. Arch Latinoam Nutr 2011; 61:270-8. [ Links ]

. Wosje KS, Khoury PR, Claytor RP, Copeland KA, Hornung RW, Daniels SR, et al. Dietary patterns associated with fat and bone mass in young children. Am J Clin Nutr 2010; 92:294-303. [ Links ]

. Oellingrath IM, Svendsen MV, Brantsaeter AL. Eating patterns and overweight in 9- to 10-year-old children in Telemark County, Norway: a cross-sectional study. Eur J Clin Nutr 2010; 64:1272-9. [ Links ]

. Craig LC, McNeill G, Macdiarmid JI, Masson LF, Holmes BA. Dietary patterns of school-age children in Scotland: association with socio-economic indicators, physical activity and obesity. Br J Nutr 2010; 103:319-34. [ Links ]

. Cutler GJ, Flood A, Hannan PJ, Slavin JL, Neumark-Sztainer D. Association between major patterns of dietary intake and weight status in adolescents. Br J Nutr 2012; 108:349-56. [ Links ]

. Superintendência de Estudos Econômicos. Índice de desenvolvimento econômico e social dos municípios baianos. Salvador: Superintendência de Estudos Econômicos; 2008. [ Links ]

. Souza CO, Silva RCR, Assis AMO, Fiaccone RL, Pinto EJ, Moraes LTLP. Associação entre inatividade física e excesso de peso em adolescentes de Salvador, Bahia – Brasil. Rev Bras Epidemiol 2010;13:468-75. [ Links ]

. Cooper PJ, Rodrigues LC, Cruz AA, Barreto ML. Asthma in Latin America: a public heath challenge and research opportunity. Allergy 2009; 64:5-17. [ Links ]

. de Onis M, Onyango AW, Borghi E, Siyam A, Nishida C, Siekmann J. Development of a WHO growth reference for school-aged children and adolescents. Bull World Health Organ 2007; 85:660-7. [ Links ]

. WHO Child Growth Standards based on length/height, weight and age. Acta Paediatr Suppl 2006; 450:76-85. [ Links ]

. Borges CQ, Silva RCR, Assis AMO, Pinto EJ, Fiaccone RL, Pinheiro SMC. Fatores associados à anemia em crianças e adolescentes de escolas públicas de Salvador, Bahia, Brasil. Cad Saúde Pública 2009; 25:877-88. [ Links ]

. Mascarenhas JMO. Padrão de consumo alimentar e a ocorrência de sintomas de asma em adolescentes de Salvador-Ba. Salvador: Universidade Federal da Bahia; 2013. [ Links ]

. Neumann AICP, Martins IS, Marcopito LF, Araujo EAC. Padrões alimentares associados a fatores de risco para doenças cardiovasculares entre residentes de um município brasileiro. Rev Panam Salud Pública 2007; 22:329-39. [ Links ]

. Philippi ST, Szarfarc SC, Laterza CR. Virtual Nutri – versão 1 for Windows. Sistema de análise nutricional. São Paulo: Departamento de Nutrição, Faculdade de Saúde Pública, Universidade de São Paulo; 1996. [ Links ]

. Galante AP, Colli C. Desenvolvimento e aplicação de um questionário semiquantitativo de freqüência alimentar on-line para estimar a ingestão de cálcio e ferro. Rev Bras Epidemiol 2008; 11:402-10. [ Links ]

. Slater B, Philippi ST, Fisberg RM, Latorre MR. Validation of a semi-quantitative adolescent food frequency questionnaire applied at a public school in Sao Paulo, Brazil. Eur J Clin Nutr 200; 57:629-35. [ Links ]

. Guedes DP, Lopes CC, Guedes JERP. Reprodutibilidade e validade do Questionário Internacional de Atividade Física em adolescentes. Rev Bras Med Esporte 2005; 11:151-8. [ Links ]

. Global Strategy on Diet, Physical Activity and Health. (accessed on 07/Sep/2013). [ Links ]

. Matusitz J, McCormick J. Sedentarism: the effects of Internet use on human obesity in the United States. Soc Work Public Health 2012; 27:250-69. [ Links ]

. Marshall WA, Tanner JM. Variations in pattern of pubertal changes in girls. Arch Dis Child 1969; 44:291-303. [ Links ]

. Marshall WA, Tanner JM. Variations in the pattern of pubertal changes in boys. Arch Dis Child 1970; 45:13-23. [ Links ]

. Associação Brasileira de Empresas de Pesquisa. Critério de classificação econômica Brasil 2008. (accessed on 02/Dec/2013). [ Links ]

. Olinto MTA. Padrões alimentares: análise de componentes principais. In: Kac G, Sichieri R, Gigante DP, organizadores. Epidemiologia nutricional. Rio de Janeiro: Editora Fiocruz/Atheneu; 2007. p. 213-25. [ Links ]

. Matos SM, Jesus SR, Saldiva SR, Prado MS, D’Innocenzo S, Assis AM, et al. Overweight, asthma symptoms, atopy and pulmonary function in children of 4-12 years of age: findings from the SCAALA cohort in Salvador, Bahia, Brazil. Public Health Nutr 2011; 14:1270-8. [ Links ]

. Baruki SBS, Rosado LEFPL, Rosado GP, Ribeiro RCL. Associação entre estado nutricional e atividade física em escolares da Rede Municipal de Ensino em Corumbá - MS. Rev Bras Med Esporte 2006; 12:90-4. [ Links ]

. Ribeiro RQC, Lotufo PA, Lamounier JA, Oliveira RG, Soares JF, Botter DA. Fatores adicionais de risco cardiovascular associados ao excesso de peso em crianças e adolescentes: o estudo do coração de Belo Horizonte. Arq Bras Cardiol 2006; 86:408-18. [ Links ]

. Campos LA, Leite ÁJM, Almeida PC. Prevalência de sobrepeso e obesidade em adolescentes escolares do município de Fortaleza, Brasil. Rev Bras Saúde Matern Infant 2007; 7:183-90. [ Links ]

. Costa RF, Cintra IP, Fisberg M. Prevalência de sobrepeso e obesidade em escolares da cidade de Santos, SP. Arq Bras Endocrinol Metab 2006; 50:60-7. [ Links ]

. Silva RCR, Assis AMO, Szarfarc SC, Pinto E, Costa LCC, Rodrigues LC. Iniquidades socioeconômicas na conformação dos padrões alimentares de crianças e adolescentes. Rev Nutr 2012; 25:451-61. [ Links ]

. Leal GVS, Philippi ST, Matsudo SMM, Toassa EC. Consumo alimentar e padrão de refeições de adolescentes, São Paulo, Brasil. Rev Bras Epidemiol 2010; 13:457-67. [ Links ]

. Li SJ, Paik HY, Joung H. Dietary patterns are associated with sexual maturation in Korean children. Br J Nutr 2006; 95:817-23. [ Links ]

. Song Y, Park MJ, Paik HY, Joung H. Secular trends in dietary patterns and obesity-related risk factors in Korean adolescents aged 10-19 years. Int J Obes (Lond) 2010; 34:48-56. [ Links ]

. Aranceta J, Perez-Rodrigo C, Ribas L, Serra-Majem L. Sociodemographic and lifestyle determinants of food patterns in Spanish children and adolescents: the enKid study. Eur J Clin Nutr 2003; 57 Suppl 1:S40-4. [ Links ]

. Ambrosini GL. Childhood dietary patterns and later obesity: a review of the evidence. Proc Nutr Soc 2013; 27:1-10. [ Links ]

. Stubbs RJ, van Wyk MC, Johnstone AM, Harbron CG. Breakfasts high in protein, fat or carbohydrate: effect on within-day appetite and energy balance. Eur J Clin Nutr 1996; 50:409-17. [ Links ]

. Cribb V, Emmett P, Northstone K. Dietary patterns throughout childhood and associations with nutrient intakes. Public Health Nutr 2013; 16:1801-9. [ Links ]

. Patel S, Murray CS, Simpson A, Custovic A. Dietary patterns of 11-year-old children and associations with nutrient intakes. J Hum Nutr Diet 2011; 24:399-400. [ Links ]

. Reynolds KD, Franklin FA, Binkley D, Raczynski JM, Harrington KF, Kirk KA, et al. Increasing the fruit and vegetable consumption of fourth-graders: results from the high 5 project. Prev Med 2000; 30:309-19. [ Links ]

. Klesges LM, Baranowski T, Beech B, Cullen K, Murray DM, Rochon J, et al. Social desirability bias in self-reported dietary, physical activity and weight concerns measures in 8- to 10-year-old African-American girls: results from the Girls Health Enrichment Multisite Studies (GEMS). Prev Med 2004; 38 Suppl:S78-87. [ Links ]

. Drewnowski A. Diet image: a new perspective on the food-frequency questionnaire. Nutr Rev 2001; 59:370-2. [ Links ]

. Livingstone MB, Robson PJ. Measurement of dietary intake in children. Proc Nutr Soc 2000; 59:279-93. [ Links ]

. Martinez ME, Marshall JR, Sechrest L. Invited commentary: factor analysis and the search for objectivity. Am J Epidemiol 1998; 148:17-9. [ Links ]

. Hearty AP, Gibney MJ. Comparison of cluster and principal component analysis techniques to derive dietary patterns in Irish adults. Br J Nutr 2009; 101:598-608. [ Links ]

Received: October 15, 2013; Revised: February 14, 2014; Accepted: February 17, 2014

Correspondence. R. C. R. Silva. Escola de Nutrição, Universidade Federal da Bahia. Rua Araújo Pinho 32, Salvador, BA 40150-260, Brasil.


N. H. A. Santos participated in the interpretation of results and writing of the manuscript. R. L. Fiaccone designed the work and reviewed the manuscript. M. L. Barreto participated in the interpretation of results and revision of the manuscript. L. A. Silva participated in the field work and reviewed the manuscript. R. C. R. Silva participated in the study design, data collection, interpretation of results, and writing of the manuscript.

Conflicts of interest

The authors declare no conflict of interest.

Creative Commons License This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License, which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.