- Email: [email protected]
12th IFDC 2017 special issue – Brazilian Food Composition Table (TBCA): Development and functionalities of the online version
Journal of Food Composition and Analysis 84 (2019) 103287
Contents lists available at ScienceDirect
Journal of Food Composition and Analysis journal homepage: www.elsevier.com/locate/jfca
Original Research Article
12th IFDC 2017 special issue – Brazilian Food Composition Table (TBCA): Development and functionalities of the online version⋆
T
Kristy Soraya Coelhoa,b, Eliana Bistriche Giuntinib, Fernanda Grandea,b, João da Silva Diase, Eduardo Purgattob,c, Bernadette Dora Gombossy de Melo Francob,c, Franco Maria Lajolob,c,d, ⁎ Elizabete Wenzel de Menezesb,d, Applied Human Nutrition Interunits Program (PRONUT), FCF/FEA/FSP – University of São Paulo (USP), São Paulo, SP, Brazil Food Research Center (FoRC/CEPID/FAPESP), São Paulo, Brazil c Department of Food and Experimental Nutrition, Faculty of Pharmaceutical Sciences, University of São Paulo (USP), São Paulo, SP, Brazil d Coordinators of BRASILFOODS, Brazil e Department of Electrical Engineering, Federal University of Paraná (UFPR), Curitiba, PR, Brazil a
b
A R T I C LE I N FO
A B S T R A C T
Keywords: TBCA Food composition data INFOODS Nutrient intake evaluation database Biodiversity database Online food composition table Brazilian foods Brazilian recipes
The first version of the Brazilian Food Composition Table (TBCA) website was created in 1998. Version 6.0, launched in 2017, was developed collaboratively by the Food Research Center, University of São Paulo and the Brazilian Network of Food Data Systems. It is freely available at www.fcf.usp.br/tbca. The updated version comprises two databases: (i) the Biodiversity and Regional Food Database (TBCA B-DB), which contains analytical data on Brazilian food biodiversity; and (ii) the Nutrient Intake Evaluation Database (TBCA NIE-DB), comprising data on most foods consumed in Brazil, including the content of 34 components. The TBCA NIE-DB presents composition data for more than 3400 foods, including raw foods, manufactured products and composite dishes. The data can be searched by food name in Portuguese or English or by the scientific name. Searches by food group or by component can also be performed, and reports are generated according to the data available and database accessed. A new functionality implemented in the website is the evaluation of energy intake for general users. TBCA NIE-DB is a tool for evaluating nutrient intake for individuals and populations and it assists professionals with dietary prescriptions. Computational tools using TBCA NIE-DB can be developed for various applications, such as generating personalized menus.
1. Introduction Food composition tables (FCT) provide detailed information on the chemical composition of fresh and processed foods in a certain country. This information is obtained from chemical analysis or compiled from published data, and is available in print or electronically. Food composition tables (FCTs) are essential for developing programs related to nutrition and health, agriculture, and the food industry, among others (Pennington, 2008; EuroFIR, 2018a). Computer science is essential today, and provides important tools for effective data collection, storage and analysis (Hoggle et al., 2006; Chen et al., 2012). In addition, the use of computational tools to support decision-making is becoming more widespread. The development
of such tools for archiving, management, and the use of FCTs has resulted in the increased use of the internet for dissemination (Bell et al., 2011). Printed FCTs are still produced in several countries, but the electronic versions are increasingly important because they store large amounts of data and allow for easy access, handling, and updating (Church, 2006; Bell et al., 2011). It is important to highlight that FCTs from some countries, including Brazil, are available online as downloadable files (which do not allow online queries) in spreadsheet or PDF format. Lists of electronic FCTs are available on sites such as the FAO/ INFOODS website (FAO/INFOODS, 2018a), ILSI Research Foundation website (ILSI, 2018), EuroFIR website (EuroFIR, 2018b), LanguaL™
⋆ This paper was originally submitted as a poster at the 12th International Food Data Conference (IFDC) held from October 11–13, 2017 in Buenos Aires, Argentina. It received the “Best Poster Award”. ⁎ Corresponding author at: Food Research Center, Rua do Lago, 250 eEd. Semi Industrial, Bloco C – Lab. De Engenharia de Alimentos – Cidade Universitária, 05508-080, São Paulo, SP, Brazil. E-mail address: [email protected] (E.W. de Menezes).
https://doi.org/10.1016/j.jfca.2019.103287 Received 30 March 2018; Received in revised form 12 October 2018; Accepted 8 August 2019 Available online 17 August 2019 0889-1575/ © 2019 Published by Elsevier Inc.
Journal of Food Composition and Analysis 84 (2019) 103287
K.S. Coelho, et al.
Agricultural Organization of the United Nations/International Network of Food Data Systems (FAO/INFOODS). The data record was implemented using an updated version of the form for food composition data compilation proposed by BRASILFOODS (Menezes et al., 2011). This form consists of independent spreadsheets (in Microsoft Excel) to record detailed food identification, nutrient content, data quality (number of samples, sampling plan, sample handling procedures, analytical method, analytical quality control) and other relevant information. The criteria for exclusion from the database were use of unvalidated analytical methodologies, data where adopted conversion factors were not clearly described, data presented in graphs without providing values, and lack of moisture value (for data presented on a dry matter basis).
Thesaurus website, and LanguaL™ (2018). This proliferation may indicate that this method of communication is an efficient way of disseminating FCT information. Online versions also allow for quicker updates and lower costs relative to printed versions. However, certain issues should be considered when presenting online FCTs, to ensure that access to the information is easy and that the information will be used properly. These considerations include (i) access modules (free or with registration of the users), (ii) the number of foods and components available, (iii) data format, (iv) search tools, (v) data documentation, and (vi) tools for recipe calculations (Machackova et al., 2019). The first version of the Brazilian Food Composition Table (TBCA) website was launched in 1998. It represented the first database disseminated on the Internet in Latin America (Menezes et al., 2002). However, the FCTs in TBCA up to version 5.0 did not present a complete nutrient profile for each food entry, as only original analytical data were included. This information was grouped in different databases according to components. These components consisted of proximate composition, dietary fiber (enzymatic-gravimetric methodology), resistant starch, carbohydrate profile (soluble sugars, total, available, and resistant starches), fatty acids and cholesterol, vitamin A and carotenoids, minerals, and glycemic response. Version 5.0 included about 2800 foods across all databases, with 50% of the data related to proximate composition (Giuntini et al., 2006). Despite the importance of this information, it was not useful for evaluating the nutrient intake of either an individual or a population. Version 6.0 of the TBCA website was launched in October 2017. Its development was coordinated by the Brazilian Network of Food Data Systems (BRASILFOODS), at the University of São Paulo (USP), and the Food Research Center (FoRC), at the Center for Research, Innovation and Dissemination of the São Paulo Research Foundation (FAPESP).
2.1.2. TBCA nutrient intake evaluation database The TBCA Nutrient Intake Evaluation Database (TBCA NIE-DB) provides the complete nutrient profile for 34 components, namely proximate composition, fat fractions, 10 vitamins, and 10 minerals. The database includes the most widely consumed foods in Brazil, to enable the nutrient intake evaluation. To develop this database, data were first used to perform the aggregation of foods with a similar description. In addition, data were borrowed from other FCTs to fill in the gaps, resulting in fewer missing values. After aggregation, nutrient data that were not available for food were borrowed from other tables in order to reduce the amount of missing data in the TBCA NIE-BD. The borrowed data were adjusted for the moisture and lipid contents of Brazilian foods. Expanding the number of composite dishes in an FCT is necessary because such dishes represent most of the foods consumed in a country. In this step, priority was given to include the recipes reported in the National Dietary Survey conducted by the Brazilian Institute of Geography and Statistics (NDS/IBGE). After all available data were aggregated, recipes were collected to calculate the nutrient profile for recipes having many ingredients. Details on how composite dishes’ nutrient profiles are calculated can be found in the article on the development of Brazilian Nutrient Intake Evaluation Database by Giuntini et al. (2019). Currently, the TBCA NIE-DB comprises food composition data for 3437 foods. The breakdown of categories is as follows: (i) 432 (12.57%) fresh foods, or foods consumed in their natural form or that require a previous processing method; this group includes fruits, vegetables, raw meat, cereals, eggs, legumes, nuts and seeds; (ii) 215 (6.26%) cooked foods, or foods that are boiled, baked, grilled, or fried but without the addition of other ingredients; this group includes oil and seasonings, vegetables, meat, pasta, and legumes; (iii) 178 (5.18%) processed foods used as ingredients, which includes wheat, corn flours, and dry mix powders (e.g. gelatin, pudding and cake); (iv) 461 (13.41%) ready-to-eat foods, such as cookies, cakes, breads, oils and margarines, milk and dairy products, chocolate, sugar, sauces, and other foods that can also be used in recipes; and (v) 2151 (62.58%) multi-ingredient recipes, from all food groups, including traditional recipes from Brazil. Each component presented in the TBCA NIE-DB contains the information shown in Table 1.
2. Databases on the TBCA website The TBCA version 6.0 consists of two databases: (i) the Nutrient Intake Evaluation Database (TBCA NIE-DB), which provides the complete nutrient profile of the most-consumed foods in Brazil; and (ii) the Biodiversity and Regional Foods Database (TBCA B-DB), which centralizes analytical data on Brazilian food biodiversity and the typical dishes in the various regions of the country. 2.1. Profile and criteria adopted for the elaboration of the databases 2.1.1. General criteria Food identification was performed using the INFOODS multi-faceted naming system (Truswell et al., 1991). The components were described according to the tag names established by INFOODS (FAO/ INFOODS, 2018b). Additionally, the food groups created for Latin America by LATINFOODS were adopted (Morón et al., 1997). The foods were classified according to the following food groups: (A) cereals and cereal products, (B) vegetables and vegetable products, (C) fruit and fruit products, (D) oil and fats, (E) fish and shellfish, (F) meat and meat products, (G) milk and dairy products, (H) beverages, (J) eggs and egg dishes, (K) sugars and sweets, (L) miscellaneous, (M) fast food, (N) foods for special diets, (Q) infant foods, (R) industrialized foods (unprepared), (T) legumes and legume products, and (U) nuts and seeds. Food identification was also performed using a unique alpha-numeric code, which consist of a numeric code for each food and a letter corresponding to the respective food group. The code for foods in TBCA NIE-BD also contains a “C” before the numeric code. For example, in the code C0001A “Rice, white, raw”, “C” signifies the food on the nutrient intake (“consume” in Portuguese”), evaluation database (TBCA NIEDB), while the number identifies the food itself and “A” refers to the cereal group. The process adopted for data compilation, data collection and evaluation was carried out following the guidelines of the Food
2.1.3. TBCA biodiversity and regional foods database The TBCA Biodiversity and Regional Foods Database (TBCA BD-DB) includes only original analytical data. This means that unlike the NIEDB, no data were aggregated, resulting in an incomplete nutrient profile for the listed foods. The aim of the TBCA BD-DB is thus to centralize data on the chemical composition of foods from the Brazilian biodiversity and typical dishes from various regions of the country. Foods were also selected for inclusion in the TBCA BD-DB from the previously compiled data according to indicators proposed by the FAO (FAO/INFOODS, 2016). In the database, the food description includes details on the biodiversity indicators and the place (city or state) where the analyzed samples were collected. This information allows the comparison of foods within the database according to the variety, cultivar or breed, as well as by the origin of the sample. 2
Journal of Food Composition and Analysis 84 (2019) 103287
K.S. Coelho, et al.
Table 1 Information available for each component in the TBCA NIE-DB. ITEM
DESCRIPTION
Nutrient content Standard deviation Minimum value Maximum value Number of data points References Type of data
Defined by a single value or a mean value (for more than two analytical data when available). Represents the variability in the nutrient content; is presented when nutrient content was calculated as the mean value of three or more data points. Lower value for a certain component, presented when the nutrient content was calculated as the mean value of two or more data points. Higher value for a certain component, presented when the nutrient content was calculated as the mean value of two or more data points. The number of analytical values used to calculate the mean nutrient content. Lists the numeric code of all the references used to calculate the mean content of the component. Analytical value: values based on data compiled from publications, laboratories, food industries, and other FCTs; or based on chemical analyses carried out for this purpose. Calculated values: values calculated from other analytical data (e.g. carbohydrates calculated by difference and energy) or values obtained from recipe calculation (nutrient profiles calculated from raw or cooked ingredients, applying appropriate yield and nutrient retention factors). Imputed values: values derived from analytical data for a similar food or different forms of the same food (e.g. vegetable boiled or stewed). The values were also used to identify values for food products in a certain brand, complemented with data on a generic food. The values were corrected for moisture and fat in the complemented food. Part of the imputed values were derived from international FCTs. Assumed value: values were assumed to be “zero” for certain single foods, in their natural forms, based on data from other FCTs or in the following situations (FAO/INFOODS, 2012): (i) alcohol – animal and plant foods that were not fermented; (ii) dietary fiber – animal foods; (iii) carbohydrates – animal foods (except offal and mollusks); (iv) cholesterol, retinol and vitamin B12 – plant foods not fermented (except mushrooms); (v) vitamin C – for oils; (vi) folic acid – for non-fortified foods. Trace value (“tr”) was used when: (i) a component was identified but not quantified, or (ii) when the level of the nutrient was considered not relevant in the food. Not analyzed (“NA”) was used to identify missing data. Usually this occurs when a similar food or preparation was not identified for data input.
an example of this report.
Currently, the database comprises information on 814 foods and 27 Brazilian dishes. This includes data on proximate composition, minerals, and vitamins A, C and E. Most of the data consists of foods from the fruit group; the main components in the available data are vitamins. All other data presented in the TBCA BD-DB correspond to original analytical data obtained from Brazilian food samples.
4.3. “Component search” report (TBCA NIE-DB) Searching by component is a new search option in version 6.0 of the website. After the user chooses the component to be searched, the most important foods are listed. The quantity of the nutrient is ranked according to the food group to which it belongs. A report is generated for online consultation and can be displayed in ascending or descending order of the component content in the food. The food information provided in the “search-by-component” feature in the TBCA NIE-DB is intended for health professionals, such as nutritionists, to use in their professional practice.
3. Data search on the TBCA website Both the TBCA NIE-DB and TBCA BD-DB present the foods in alphabetical order, with their detailed descriptions by type, processing, scientific name, and other details. The food name is given in Portuguese, English and/or the scientific name, as is the respective food group of the item. A data search can be performed in various ways: (i) alphabetical order in whole database, (ii) by specific food – in Portuguese, English or by the scientific name, or (iii) by component within the groups. New search options will be added to the TBCA NIE-DB to allow searches by the following food types: (i) fresh foods or foods that are consumed in their natural state, (ii) food processd or foods that require a processing step prior to consumption (boiled, baked, grilled or fried) without the addition of other ingredients; (iii) processed foods used as ingredients; (iv) ready-to-eat foods; and (v) multi-ingredient recipes, including traditional recipes from Brazil.
5. New functionality of TBCA website One new function available on the TBCA website is called “evaluation of energy intake”. This is a service for the general users of the TBCA website. It allows the calculation of the energy content of one meal, and encourages users to be aware of the foods they consume and the quality of the meal. By inserting information on up to seven foods or composite dishes of one or more meals, the user can determine the total energy intake with an online report based on the TBCA NIE-DB. After selecting the food of interest, the user selects the portion consumed, according to the standardized household measurements, including fractions of a portion (0.5 or 1). Then the amount of energy is shown individually (per food) and as a total (sum of all foods in the meal) for each query performed (Fig. 3). At the end of the evaluation, the user is advised that the data are for information only, and that a nutritionist should be consulted for nutritional guidance.
4. TBCA website reports and profile of the databases The most significant change in the TBCA version 6.0 is the inclusion of the complete nutrient profile for 34 components. These profiles are presented for each food on the NIE-DB with documentation at the component level (mean value, minimum and maximum values, standard deviation, number of data, references, and data type).
6. Perspectives for the future – application in computational tools
4.1. TBCA NIE-DB reports
A computational tool using the TBCA NIE-DB is under development. It will generate personalized menus to assist nutritionists in their professional practice. The nutritional recommendations and preferences of the patient or client are considered when personalized menus are created (Coelho et al., 2019). Other computing tools using the TBCA NIEDB are also under development, as follows: (i) an app to allow nutritionists to consult the TBCA data on tablets and smartphones, and to conduct personalized searches on food composition data; (ii) an app for general users, to enable them to consult the TBCA data to evaluate energy intake; and (iii) a computational tool to assist nutritionists in the
An example of the food composition (statistical information) report is shown in Fig. 1. The codes and the complete reference details are presented at the bottom of the report page. 4.2. TBCA BD-DB reports The TBCA-DB reports present data on the nutrient components available per 100 g of edible portions, and the data source. Fig. 2 shows 3
Journal of Food Composition and Analysis 84 (2019) 103287
K.S. Coelho, et al.
Fig. 1. Food composition (statistical information) report from the TBCA NIE-DB. The details include (A) the food description (Portuguese and English names and scientific name), (B) statistical information and documentation details for 34 components, and (C) the data sources. TBCA NIE-DB presents national data (analytical and calculated) or is borrowed from other FCTs. Analytical data represent the mean of several sources.
4
Journal of Food Composition and Analysis 84 (2019) 103287
K.S. Coelho, et al.
Fig. 2. Food composition report from the TBCA Biodiversity Database and Regional Foods Database (TBCA BD-DB). The report includes (A) a food description (Portuguese and English names and scientific name); (B) data for the components available; and (C) the data source. TBCA BD-DB presents essentially national data; as they deal with the original analytical data are incomplete.
Fig. 3. “Evaluation of energy intake” for one meal (e. g. breakfast). The results include (A) selected foods, (B) individual energy in a food, (C) total energy in a meal, and (D) standardized household measures.
users about the quality and composition of the meal.
nutritional labels calculation for packed foods, taking into consideration yield and retention factors.
Declaration of Competing Interest 7. Conclusion
There are no conflicts of interest to declare.
The TBCA comprises two databases: the TBCA NIE-DB, which contains the nutrient profiles of 34 components and covers foods reported in national food consumption surveys, with the content of recipes included; and the TBCA BD-DB, which collects the analytical data on Brazilian biodiversity in one central location. A total of 3437 foods are included in the TBCA NIE-DB. Unlike the data in the previous versions, 76% of the food composition data now available pertain to ready-to-eat foods. Of these foods, 62.6% are mixed dishes and 13.4% are processed foods. It is important to note that the TBCA is constantly being updated. New food composition data will be included in future versions of the database to provide a wide variety of potential, such as nutrient intake assessment of individuals and populations and dietary prescription, as well information on food biodiversity and regional foods. The TBCA website (version 6.0) has been restructured and it currently presents a responsive template. One new function of the TBCA website, namely “evaluation of energy intake”, provides information about the energy content of one meal and is intended to inform general
Acknowledgments This study was supported by the Food Research Center (FoRC/ CEPID/FAPESP – Project number 2013/07914-8), by BRASILFOODS, and by the post-graduate program in Applied Human Nutrition (PRONUT/FCF/FEA/FSP/USP). Funding was provided by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES). References Bell, S., Colombani, P.C., Pakkala, H., Christensen, T., Møller, A., Finglas, P.M., 2011. Food composition data: identifying new uses, approaching new users. J. Food Anal. 24, 727–731. https://doi.org/10.1016/j.jfca.2011.03.001. Chen, Y., Hsu, C., Liu, L., Yang, S., 2012. Constructing a nutrition diagnosis expert system. Expert Syst. Appl. 39, 2132–2156. https://doi.org/10.1016/j.eswa.2011.07.069. Church, S.M., 2006. The history of food composition databases. Nutr. Bull. 31, 15–20.
5
Journal of Food Composition and Analysis 84 (2019) 103287
K.S. Coelho, et al.
Giuntini, E.B., Coelho, K.S., Grande, F., Marchioni, D.M.L., Carli, E., Sichieri, R., Purgatto, E., Franco, B.D.G.M., Lajolo, F.M., Menezes, E.W., 2019. Brazilian nutrient intake evaluation database: a tool for assessment of the nutrient intake in Brazil. J. Food Compos. Anal. 83. https://doi.org/10.1016/j.jfca.2019.103286. Hoggle, L.B., Michael, M.A., Houston, S., Ayres, E.J., 2006. Nutrition informatics. J. Am. Diet. Assoc. 106 (1), 134–139. https://doi.org/10.1016/j.jada.2005.10.025. ILSI, 2018. ILSI Research Foundation. Food Composition Databases or Tables (FCDB) (2017). [WWW page]. Retrieved January 10, 2018, from: http://www.ilsirf.org/wpcontent/uploads/sites/5/2017/05/StatPlanet/StatPlanet_Cloud_2-2.html.LanguaL™. (2018). Links – Food composition on the Web. [WWW page]. Retrivied January 10, 2018, from:. http://www.langual.org/langual_food_composition.asp. Machackova, M., Holasova, M., Maskova, E., 2019. THINline interactive s.r.o. The new on-line Czech food composition database. Food Chem. 140 (1), 533–538. https://doi. org/10.1016/j.foodchem.2013.01.095. Menezes, E.W., Gonçalves, F.A.R., Giuntini, E.B., Lajolo, F.M., 2002. Brazilian food composition database: internet dissemination and other recent developments. J. Food Anal. 15, 451–462. https://doi.org/10.1006/jfca.2002.1083. Menezes, E.W., Santos, N.C., Giuntini, E.B., Dan, M.C.T., Genovese, M.I., Lajolo, F.M., 2011. Brazilian flavonoid database: application of quality evaluation system. J. Food Anal. 24, 629–636. https://doi.org/10.1016/j.jfca.2010.09.004. Morón, C., Zacarías, I., Pablo, S., 1997. Producción y manejo de datos de composición química de alimentos en nutrición. Organización de las Naciones Unidas para la Agricultura y la Alimentación, Santiago 356p. Pennington, J.A.T., 2008. Applications of food composition data: data sources and considerations for use. J. Food Anal. 21, S3–S12. https://doi.org/10.1016/j.jfca.2007. 02.005. Truswell, A.S., Bateson, D.J., Madafiglio, K.C., Pennington, J.A.T., Rand, W.M., Klensin, J.C., 1991. INFOODS guidelines for describing foods: a systematic approach to describing foods to facilitate international exchange of food composition data. J. Food Anal. 4 (1), 18–38. https://doi.org/10.1016/0889-1575(91)90045-8.
https://doi.org/10.1111/j.1467-3010.2006.00538.x. Coelho, K.S., Giuntini, E.B., Grande, F., Dias, J.S., Lajolo, F.M., Menezes, E.W., 2019. Brazilian Food Composition Table (TBCA): new version and applications in computational tools. 2017, October In: In Poster Session Presented at the 12th International Food Data Conference. Buenos Aires. Argentina. EuroFIR, 2018a. EuroFir – European Food Information Resource Network. Food Information. [WWW Page]. Retrieved January 10, 2018, from:. http://www. eurofir.org/food-information-new/. EuroFIR, 2018b. EuroFir – European Food Information Resource Network. Food Composition Database. [WWW page]. Retrieved January 10, 2018, from:. http:// www.eurofir.org/food-information/food-composition-databases-2/. FAO/INFOODS, 2012. FAO/INFOODS Guidelines for Checking Food Composition Data Prior to the Publication of a User Table/Database - Version 1.0. January 10, 2018, from:. FAO, Rome. http://www.fao.org/docrep/017/ap810e/ap810e.pdf. FAO/INFOODS, 2016. FAO/INFOODS - Food and Agriculture Organization/ International Network of Food Data Systems. Food Composition Database for Biodiversity Version 3.0 – BioFoodComp3.0. [Data file]. Retrieved April 16, 2015, from:. FAO, Rome. http://www.fao.org/infoods/infoods/tables-and-databases/faoinfoods-databases/ en/. FAO/INFOODS, 2018a. FAO/INFOODS - Food and Agriculture Organization/ International Network of Food Data Systems. International Food Composition table/ database Directory. [Data file]. Retrieved January 10, 2018, from:. http://www.fao. org/infoods/infoods/tables-and-databases/en/. FAO/INFOODS, 2018b. FAO/INFOODS - Food and Agriculture Organization/ International Network of Food Data Systems. Tagnames for Food Components. [Data file]. Retrieved January 10, 2018, from:. http://www.fao.org/infoods/infoods/ standards-guidelines/food-component-identifiers-tagnames/en/. Giuntini, E.B., Lajolo, F.M., Menezes, E.W., 2006. Tabela Brasileira de Composição de Alimentos TBCA-USP (Versões 3 e 4) no contexto internacional. Arch. Latinoam. Nutr. 56 (4), 366–374. [Data file]. Retrieved January 10, 2018, from: http://www. scielo.org.ve/scielo.php?script=sci_arttext&pid=S0004-06222006000400009.
6
Contents lists available at ScienceDirect
Journal of Food Composition and Analysis journal homepage: www.elsevier.com/locate/jfca
Original Research Article
12th IFDC 2017 special issue – Brazilian Food Composition Table (TBCA): Development and functionalities of the online version⋆
T
Kristy Soraya Coelhoa,b, Eliana Bistriche Giuntinib, Fernanda Grandea,b, João da Silva Diase, Eduardo Purgattob,c, Bernadette Dora Gombossy de Melo Francob,c, Franco Maria Lajolob,c,d, ⁎ Elizabete Wenzel de Menezesb,d, Applied Human Nutrition Interunits Program (PRONUT), FCF/FEA/FSP – University of São Paulo (USP), São Paulo, SP, Brazil Food Research Center (FoRC/CEPID/FAPESP), São Paulo, Brazil c Department of Food and Experimental Nutrition, Faculty of Pharmaceutical Sciences, University of São Paulo (USP), São Paulo, SP, Brazil d Coordinators of BRASILFOODS, Brazil e Department of Electrical Engineering, Federal University of Paraná (UFPR), Curitiba, PR, Brazil a
b
A R T I C LE I N FO
A B S T R A C T
Keywords: TBCA Food composition data INFOODS Nutrient intake evaluation database Biodiversity database Online food composition table Brazilian foods Brazilian recipes
The first version of the Brazilian Food Composition Table (TBCA) website was created in 1998. Version 6.0, launched in 2017, was developed collaboratively by the Food Research Center, University of São Paulo and the Brazilian Network of Food Data Systems. It is freely available at www.fcf.usp.br/tbca. The updated version comprises two databases: (i) the Biodiversity and Regional Food Database (TBCA B-DB), which contains analytical data on Brazilian food biodiversity; and (ii) the Nutrient Intake Evaluation Database (TBCA NIE-DB), comprising data on most foods consumed in Brazil, including the content of 34 components. The TBCA NIE-DB presents composition data for more than 3400 foods, including raw foods, manufactured products and composite dishes. The data can be searched by food name in Portuguese or English or by the scientific name. Searches by food group or by component can also be performed, and reports are generated according to the data available and database accessed. A new functionality implemented in the website is the evaluation of energy intake for general users. TBCA NIE-DB is a tool for evaluating nutrient intake for individuals and populations and it assists professionals with dietary prescriptions. Computational tools using TBCA NIE-DB can be developed for various applications, such as generating personalized menus.
1. Introduction Food composition tables (FCT) provide detailed information on the chemical composition of fresh and processed foods in a certain country. This information is obtained from chemical analysis or compiled from published data, and is available in print or electronically. Food composition tables (FCTs) are essential for developing programs related to nutrition and health, agriculture, and the food industry, among others (Pennington, 2008; EuroFIR, 2018a). Computer science is essential today, and provides important tools for effective data collection, storage and analysis (Hoggle et al., 2006; Chen et al., 2012). In addition, the use of computational tools to support decision-making is becoming more widespread. The development
of such tools for archiving, management, and the use of FCTs has resulted in the increased use of the internet for dissemination (Bell et al., 2011). Printed FCTs are still produced in several countries, but the electronic versions are increasingly important because they store large amounts of data and allow for easy access, handling, and updating (Church, 2006; Bell et al., 2011). It is important to highlight that FCTs from some countries, including Brazil, are available online as downloadable files (which do not allow online queries) in spreadsheet or PDF format. Lists of electronic FCTs are available on sites such as the FAO/ INFOODS website (FAO/INFOODS, 2018a), ILSI Research Foundation website (ILSI, 2018), EuroFIR website (EuroFIR, 2018b), LanguaL™
⋆ This paper was originally submitted as a poster at the 12th International Food Data Conference (IFDC) held from October 11–13, 2017 in Buenos Aires, Argentina. It received the “Best Poster Award”. ⁎ Corresponding author at: Food Research Center, Rua do Lago, 250 eEd. Semi Industrial, Bloco C – Lab. De Engenharia de Alimentos – Cidade Universitária, 05508-080, São Paulo, SP, Brazil. E-mail address: [email protected] (E.W. de Menezes).
https://doi.org/10.1016/j.jfca.2019.103287 Received 30 March 2018; Received in revised form 12 October 2018; Accepted 8 August 2019 Available online 17 August 2019 0889-1575/ © 2019 Published by Elsevier Inc.
Journal of Food Composition and Analysis 84 (2019) 103287
K.S. Coelho, et al.
Agricultural Organization of the United Nations/International Network of Food Data Systems (FAO/INFOODS). The data record was implemented using an updated version of the form for food composition data compilation proposed by BRASILFOODS (Menezes et al., 2011). This form consists of independent spreadsheets (in Microsoft Excel) to record detailed food identification, nutrient content, data quality (number of samples, sampling plan, sample handling procedures, analytical method, analytical quality control) and other relevant information. The criteria for exclusion from the database were use of unvalidated analytical methodologies, data where adopted conversion factors were not clearly described, data presented in graphs without providing values, and lack of moisture value (for data presented on a dry matter basis).
Thesaurus website, and LanguaL™ (2018). This proliferation may indicate that this method of communication is an efficient way of disseminating FCT information. Online versions also allow for quicker updates and lower costs relative to printed versions. However, certain issues should be considered when presenting online FCTs, to ensure that access to the information is easy and that the information will be used properly. These considerations include (i) access modules (free or with registration of the users), (ii) the number of foods and components available, (iii) data format, (iv) search tools, (v) data documentation, and (vi) tools for recipe calculations (Machackova et al., 2019). The first version of the Brazilian Food Composition Table (TBCA) website was launched in 1998. It represented the first database disseminated on the Internet in Latin America (Menezes et al., 2002). However, the FCTs in TBCA up to version 5.0 did not present a complete nutrient profile for each food entry, as only original analytical data were included. This information was grouped in different databases according to components. These components consisted of proximate composition, dietary fiber (enzymatic-gravimetric methodology), resistant starch, carbohydrate profile (soluble sugars, total, available, and resistant starches), fatty acids and cholesterol, vitamin A and carotenoids, minerals, and glycemic response. Version 5.0 included about 2800 foods across all databases, with 50% of the data related to proximate composition (Giuntini et al., 2006). Despite the importance of this information, it was not useful for evaluating the nutrient intake of either an individual or a population. Version 6.0 of the TBCA website was launched in October 2017. Its development was coordinated by the Brazilian Network of Food Data Systems (BRASILFOODS), at the University of São Paulo (USP), and the Food Research Center (FoRC), at the Center for Research, Innovation and Dissemination of the São Paulo Research Foundation (FAPESP).
2.1.2. TBCA nutrient intake evaluation database The TBCA Nutrient Intake Evaluation Database (TBCA NIE-DB) provides the complete nutrient profile for 34 components, namely proximate composition, fat fractions, 10 vitamins, and 10 minerals. The database includes the most widely consumed foods in Brazil, to enable the nutrient intake evaluation. To develop this database, data were first used to perform the aggregation of foods with a similar description. In addition, data were borrowed from other FCTs to fill in the gaps, resulting in fewer missing values. After aggregation, nutrient data that were not available for food were borrowed from other tables in order to reduce the amount of missing data in the TBCA NIE-BD. The borrowed data were adjusted for the moisture and lipid contents of Brazilian foods. Expanding the number of composite dishes in an FCT is necessary because such dishes represent most of the foods consumed in a country. In this step, priority was given to include the recipes reported in the National Dietary Survey conducted by the Brazilian Institute of Geography and Statistics (NDS/IBGE). After all available data were aggregated, recipes were collected to calculate the nutrient profile for recipes having many ingredients. Details on how composite dishes’ nutrient profiles are calculated can be found in the article on the development of Brazilian Nutrient Intake Evaluation Database by Giuntini et al. (2019). Currently, the TBCA NIE-DB comprises food composition data for 3437 foods. The breakdown of categories is as follows: (i) 432 (12.57%) fresh foods, or foods consumed in their natural form or that require a previous processing method; this group includes fruits, vegetables, raw meat, cereals, eggs, legumes, nuts and seeds; (ii) 215 (6.26%) cooked foods, or foods that are boiled, baked, grilled, or fried but without the addition of other ingredients; this group includes oil and seasonings, vegetables, meat, pasta, and legumes; (iii) 178 (5.18%) processed foods used as ingredients, which includes wheat, corn flours, and dry mix powders (e.g. gelatin, pudding and cake); (iv) 461 (13.41%) ready-to-eat foods, such as cookies, cakes, breads, oils and margarines, milk and dairy products, chocolate, sugar, sauces, and other foods that can also be used in recipes; and (v) 2151 (62.58%) multi-ingredient recipes, from all food groups, including traditional recipes from Brazil. Each component presented in the TBCA NIE-DB contains the information shown in Table 1.
2. Databases on the TBCA website The TBCA version 6.0 consists of two databases: (i) the Nutrient Intake Evaluation Database (TBCA NIE-DB), which provides the complete nutrient profile of the most-consumed foods in Brazil; and (ii) the Biodiversity and Regional Foods Database (TBCA B-DB), which centralizes analytical data on Brazilian food biodiversity and the typical dishes in the various regions of the country. 2.1. Profile and criteria adopted for the elaboration of the databases 2.1.1. General criteria Food identification was performed using the INFOODS multi-faceted naming system (Truswell et al., 1991). The components were described according to the tag names established by INFOODS (FAO/ INFOODS, 2018b). Additionally, the food groups created for Latin America by LATINFOODS were adopted (Morón et al., 1997). The foods were classified according to the following food groups: (A) cereals and cereal products, (B) vegetables and vegetable products, (C) fruit and fruit products, (D) oil and fats, (E) fish and shellfish, (F) meat and meat products, (G) milk and dairy products, (H) beverages, (J) eggs and egg dishes, (K) sugars and sweets, (L) miscellaneous, (M) fast food, (N) foods for special diets, (Q) infant foods, (R) industrialized foods (unprepared), (T) legumes and legume products, and (U) nuts and seeds. Food identification was also performed using a unique alpha-numeric code, which consist of a numeric code for each food and a letter corresponding to the respective food group. The code for foods in TBCA NIE-BD also contains a “C” before the numeric code. For example, in the code C0001A “Rice, white, raw”, “C” signifies the food on the nutrient intake (“consume” in Portuguese”), evaluation database (TBCA NIEDB), while the number identifies the food itself and “A” refers to the cereal group. The process adopted for data compilation, data collection and evaluation was carried out following the guidelines of the Food
2.1.3. TBCA biodiversity and regional foods database The TBCA Biodiversity and Regional Foods Database (TBCA BD-DB) includes only original analytical data. This means that unlike the NIEDB, no data were aggregated, resulting in an incomplete nutrient profile for the listed foods. The aim of the TBCA BD-DB is thus to centralize data on the chemical composition of foods from the Brazilian biodiversity and typical dishes from various regions of the country. Foods were also selected for inclusion in the TBCA BD-DB from the previously compiled data according to indicators proposed by the FAO (FAO/INFOODS, 2016). In the database, the food description includes details on the biodiversity indicators and the place (city or state) where the analyzed samples were collected. This information allows the comparison of foods within the database according to the variety, cultivar or breed, as well as by the origin of the sample. 2
Journal of Food Composition and Analysis 84 (2019) 103287
K.S. Coelho, et al.
Table 1 Information available for each component in the TBCA NIE-DB. ITEM
DESCRIPTION
Nutrient content Standard deviation Minimum value Maximum value Number of data points References Type of data
Defined by a single value or a mean value (for more than two analytical data when available). Represents the variability in the nutrient content; is presented when nutrient content was calculated as the mean value of three or more data points. Lower value for a certain component, presented when the nutrient content was calculated as the mean value of two or more data points. Higher value for a certain component, presented when the nutrient content was calculated as the mean value of two or more data points. The number of analytical values used to calculate the mean nutrient content. Lists the numeric code of all the references used to calculate the mean content of the component. Analytical value: values based on data compiled from publications, laboratories, food industries, and other FCTs; or based on chemical analyses carried out for this purpose. Calculated values: values calculated from other analytical data (e.g. carbohydrates calculated by difference and energy) or values obtained from recipe calculation (nutrient profiles calculated from raw or cooked ingredients, applying appropriate yield and nutrient retention factors). Imputed values: values derived from analytical data for a similar food or different forms of the same food (e.g. vegetable boiled or stewed). The values were also used to identify values for food products in a certain brand, complemented with data on a generic food. The values were corrected for moisture and fat in the complemented food. Part of the imputed values were derived from international FCTs. Assumed value: values were assumed to be “zero” for certain single foods, in their natural forms, based on data from other FCTs or in the following situations (FAO/INFOODS, 2012): (i) alcohol – animal and plant foods that were not fermented; (ii) dietary fiber – animal foods; (iii) carbohydrates – animal foods (except offal and mollusks); (iv) cholesterol, retinol and vitamin B12 – plant foods not fermented (except mushrooms); (v) vitamin C – for oils; (vi) folic acid – for non-fortified foods. Trace value (“tr”) was used when: (i) a component was identified but not quantified, or (ii) when the level of the nutrient was considered not relevant in the food. Not analyzed (“NA”) was used to identify missing data. Usually this occurs when a similar food or preparation was not identified for data input.
an example of this report.
Currently, the database comprises information on 814 foods and 27 Brazilian dishes. This includes data on proximate composition, minerals, and vitamins A, C and E. Most of the data consists of foods from the fruit group; the main components in the available data are vitamins. All other data presented in the TBCA BD-DB correspond to original analytical data obtained from Brazilian food samples.
4.3. “Component search” report (TBCA NIE-DB) Searching by component is a new search option in version 6.0 of the website. After the user chooses the component to be searched, the most important foods are listed. The quantity of the nutrient is ranked according to the food group to which it belongs. A report is generated for online consultation and can be displayed in ascending or descending order of the component content in the food. The food information provided in the “search-by-component” feature in the TBCA NIE-DB is intended for health professionals, such as nutritionists, to use in their professional practice.
3. Data search on the TBCA website Both the TBCA NIE-DB and TBCA BD-DB present the foods in alphabetical order, with their detailed descriptions by type, processing, scientific name, and other details. The food name is given in Portuguese, English and/or the scientific name, as is the respective food group of the item. A data search can be performed in various ways: (i) alphabetical order in whole database, (ii) by specific food – in Portuguese, English or by the scientific name, or (iii) by component within the groups. New search options will be added to the TBCA NIE-DB to allow searches by the following food types: (i) fresh foods or foods that are consumed in their natural state, (ii) food processd or foods that require a processing step prior to consumption (boiled, baked, grilled or fried) without the addition of other ingredients; (iii) processed foods used as ingredients; (iv) ready-to-eat foods; and (v) multi-ingredient recipes, including traditional recipes from Brazil.
5. New functionality of TBCA website One new function available on the TBCA website is called “evaluation of energy intake”. This is a service for the general users of the TBCA website. It allows the calculation of the energy content of one meal, and encourages users to be aware of the foods they consume and the quality of the meal. By inserting information on up to seven foods or composite dishes of one or more meals, the user can determine the total energy intake with an online report based on the TBCA NIE-DB. After selecting the food of interest, the user selects the portion consumed, according to the standardized household measurements, including fractions of a portion (0.5 or 1). Then the amount of energy is shown individually (per food) and as a total (sum of all foods in the meal) for each query performed (Fig. 3). At the end of the evaluation, the user is advised that the data are for information only, and that a nutritionist should be consulted for nutritional guidance.
4. TBCA website reports and profile of the databases The most significant change in the TBCA version 6.0 is the inclusion of the complete nutrient profile for 34 components. These profiles are presented for each food on the NIE-DB with documentation at the component level (mean value, minimum and maximum values, standard deviation, number of data, references, and data type).
6. Perspectives for the future – application in computational tools
4.1. TBCA NIE-DB reports
A computational tool using the TBCA NIE-DB is under development. It will generate personalized menus to assist nutritionists in their professional practice. The nutritional recommendations and preferences of the patient or client are considered when personalized menus are created (Coelho et al., 2019). Other computing tools using the TBCA NIEDB are also under development, as follows: (i) an app to allow nutritionists to consult the TBCA data on tablets and smartphones, and to conduct personalized searches on food composition data; (ii) an app for general users, to enable them to consult the TBCA data to evaluate energy intake; and (iii) a computational tool to assist nutritionists in the
An example of the food composition (statistical information) report is shown in Fig. 1. The codes and the complete reference details are presented at the bottom of the report page. 4.2. TBCA BD-DB reports The TBCA-DB reports present data on the nutrient components available per 100 g of edible portions, and the data source. Fig. 2 shows 3
Journal of Food Composition and Analysis 84 (2019) 103287
K.S. Coelho, et al.
Fig. 1. Food composition (statistical information) report from the TBCA NIE-DB. The details include (A) the food description (Portuguese and English names and scientific name), (B) statistical information and documentation details for 34 components, and (C) the data sources. TBCA NIE-DB presents national data (analytical and calculated) or is borrowed from other FCTs. Analytical data represent the mean of several sources.
4
Journal of Food Composition and Analysis 84 (2019) 103287
K.S. Coelho, et al.
Fig. 2. Food composition report from the TBCA Biodiversity Database and Regional Foods Database (TBCA BD-DB). The report includes (A) a food description (Portuguese and English names and scientific name); (B) data for the components available; and (C) the data source. TBCA BD-DB presents essentially national data; as they deal with the original analytical data are incomplete.
Fig. 3. “Evaluation of energy intake” for one meal (e. g. breakfast). The results include (A) selected foods, (B) individual energy in a food, (C) total energy in a meal, and (D) standardized household measures.
users about the quality and composition of the meal.
nutritional labels calculation for packed foods, taking into consideration yield and retention factors.
Declaration of Competing Interest 7. Conclusion
There are no conflicts of interest to declare.
The TBCA comprises two databases: the TBCA NIE-DB, which contains the nutrient profiles of 34 components and covers foods reported in national food consumption surveys, with the content of recipes included; and the TBCA BD-DB, which collects the analytical data on Brazilian biodiversity in one central location. A total of 3437 foods are included in the TBCA NIE-DB. Unlike the data in the previous versions, 76% of the food composition data now available pertain to ready-to-eat foods. Of these foods, 62.6% are mixed dishes and 13.4% are processed foods. It is important to note that the TBCA is constantly being updated. New food composition data will be included in future versions of the database to provide a wide variety of potential, such as nutrient intake assessment of individuals and populations and dietary prescription, as well information on food biodiversity and regional foods. The TBCA website (version 6.0) has been restructured and it currently presents a responsive template. One new function of the TBCA website, namely “evaluation of energy intake”, provides information about the energy content of one meal and is intended to inform general
Acknowledgments This study was supported by the Food Research Center (FoRC/ CEPID/FAPESP – Project number 2013/07914-8), by BRASILFOODS, and by the post-graduate program in Applied Human Nutrition (PRONUT/FCF/FEA/FSP/USP). Funding was provided by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES). References Bell, S., Colombani, P.C., Pakkala, H., Christensen, T., Møller, A., Finglas, P.M., 2011. Food composition data: identifying new uses, approaching new users. J. Food Anal. 24, 727–731. https://doi.org/10.1016/j.jfca.2011.03.001. Chen, Y., Hsu, C., Liu, L., Yang, S., 2012. Constructing a nutrition diagnosis expert system. Expert Syst. Appl. 39, 2132–2156. https://doi.org/10.1016/j.eswa.2011.07.069. Church, S.M., 2006. The history of food composition databases. Nutr. Bull. 31, 15–20.
5
Journal of Food Composition and Analysis 84 (2019) 103287
K.S. Coelho, et al.
Giuntini, E.B., Coelho, K.S., Grande, F., Marchioni, D.M.L., Carli, E., Sichieri, R., Purgatto, E., Franco, B.D.G.M., Lajolo, F.M., Menezes, E.W., 2019. Brazilian nutrient intake evaluation database: a tool for assessment of the nutrient intake in Brazil. J. Food Compos. Anal. 83. https://doi.org/10.1016/j.jfca.2019.103286. Hoggle, L.B., Michael, M.A., Houston, S., Ayres, E.J., 2006. Nutrition informatics. J. Am. Diet. Assoc. 106 (1), 134–139. https://doi.org/10.1016/j.jada.2005.10.025. ILSI, 2018. ILSI Research Foundation. Food Composition Databases or Tables (FCDB) (2017). [WWW page]. Retrieved January 10, 2018, from: http://www.ilsirf.org/wpcontent/uploads/sites/5/2017/05/StatPlanet/StatPlanet_Cloud_2-2.html.LanguaL™. (2018). Links – Food composition on the Web. [WWW page]. Retrivied January 10, 2018, from:. http://www.langual.org/langual_food_composition.asp. Machackova, M., Holasova, M., Maskova, E., 2019. THINline interactive s.r.o. The new on-line Czech food composition database. Food Chem. 140 (1), 533–538. https://doi. org/10.1016/j.foodchem.2013.01.095. Menezes, E.W., Gonçalves, F.A.R., Giuntini, E.B., Lajolo, F.M., 2002. Brazilian food composition database: internet dissemination and other recent developments. J. Food Anal. 15, 451–462. https://doi.org/10.1006/jfca.2002.1083. Menezes, E.W., Santos, N.C., Giuntini, E.B., Dan, M.C.T., Genovese, M.I., Lajolo, F.M., 2011. Brazilian flavonoid database: application of quality evaluation system. J. Food Anal. 24, 629–636. https://doi.org/10.1016/j.jfca.2010.09.004. Morón, C., Zacarías, I., Pablo, S., 1997. Producción y manejo de datos de composición química de alimentos en nutrición. Organización de las Naciones Unidas para la Agricultura y la Alimentación, Santiago 356p. Pennington, J.A.T., 2008. Applications of food composition data: data sources and considerations for use. J. Food Anal. 21, S3–S12. https://doi.org/10.1016/j.jfca.2007. 02.005. Truswell, A.S., Bateson, D.J., Madafiglio, K.C., Pennington, J.A.T., Rand, W.M., Klensin, J.C., 1991. INFOODS guidelines for describing foods: a systematic approach to describing foods to facilitate international exchange of food composition data. J. Food Anal. 4 (1), 18–38. https://doi.org/10.1016/0889-1575(91)90045-8.
https://doi.org/10.1111/j.1467-3010.2006.00538.x. Coelho, K.S., Giuntini, E.B., Grande, F., Dias, J.S., Lajolo, F.M., Menezes, E.W., 2019. Brazilian Food Composition Table (TBCA): new version and applications in computational tools. 2017, October In: In Poster Session Presented at the 12th International Food Data Conference. Buenos Aires. Argentina. EuroFIR, 2018a. EuroFir – European Food Information Resource Network. Food Information. [WWW Page]. Retrieved January 10, 2018, from:. http://www. eurofir.org/food-information-new/. EuroFIR, 2018b. EuroFir – European Food Information Resource Network. Food Composition Database. [WWW page]. Retrieved January 10, 2018, from:. http:// www.eurofir.org/food-information/food-composition-databases-2/. FAO/INFOODS, 2012. FAO/INFOODS Guidelines for Checking Food Composition Data Prior to the Publication of a User Table/Database - Version 1.0. January 10, 2018, from:. FAO, Rome. http://www.fao.org/docrep/017/ap810e/ap810e.pdf. FAO/INFOODS, 2016. FAO/INFOODS - Food and Agriculture Organization/ International Network of Food Data Systems. Food Composition Database for Biodiversity Version 3.0 – BioFoodComp3.0. [Data file]. Retrieved April 16, 2015, from:. FAO, Rome. http://www.fao.org/infoods/infoods/tables-and-databases/faoinfoods-databases/ en/. FAO/INFOODS, 2018a. FAO/INFOODS - Food and Agriculture Organization/ International Network of Food Data Systems. International Food Composition table/ database Directory. [Data file]. Retrieved January 10, 2018, from:. http://www.fao. org/infoods/infoods/tables-and-databases/en/. FAO/INFOODS, 2018b. FAO/INFOODS - Food and Agriculture Organization/ International Network of Food Data Systems. Tagnames for Food Components. [Data file]. Retrieved January 10, 2018, from:. http://www.fao.org/infoods/infoods/ standards-guidelines/food-component-identifiers-tagnames/en/. Giuntini, E.B., Lajolo, F.M., Menezes, E.W., 2006. Tabela Brasileira de Composição de Alimentos TBCA-USP (Versões 3 e 4) no contexto internacional. Arch. Latinoam. Nutr. 56 (4), 366–374. [Data file]. Retrieved January 10, 2018, from: http://www. scielo.org.ve/scielo.php?script=sci_arttext&pid=S0004-06222006000400009.
6