A. Trichopoulou
Several laboratories in Greece carry out analyses of foods which are produced or are commonly consumed in Greece. The results of these analyses are published both in national and international journals, but have not been put together and published in a national food composition table. Food composition tables which are published in Greece are based mainly on data from other countries, and a systematic comparision of such data and analyses done nationally has not as yet been carried out. An analysis of the foods commonly consumed by Greek households and in a number of national institutions has been carried out by the Department of Nutrition and Biochemistry, Athens School of Hygiene. This information has been published in a booklet [1] and circulated to hospitals and other institutions. The data have now been expanded and computerized.
Under the auspices of the Greek Society for Food and Nutrition, an attempt is being made to contact various government departments and agencies, laboratories, and other organizations interested in nutrition in order to promote cooperation and to identify areas in which research is needed.
Reference
1 Trichopoulou, A.: Food composition tables (Lenis, Athens 1982).
E. Carnovale
The first food composition tables were compiled in Italy in 1946 by the National Institute of Nutrition [1]. The tables were comprised of foods most commonly eaten in Italy and provided information on the content of protein, lipid, and carbohydrate and energy value and waste coefficient. The data were obtained from analyses carried out by the institute and from the literature. Later, the tables were extended to include the mineral and vitamin content of foods. In 1976, the tables were completely revised and extended according to the guidelines prepared by Southgate [2]. The revised tables [3] include in total 375 food items which were subdivided as follows: cereals and cereal products; fresh and dried legumes; vegetables; fresh and dried fruit; meat, raw, processed, and offals; fish, fresh, frozen, and processed; milk and milk products; eggs; fats and oils; sugar, preserves and confectionery; miscellaneous foods and prepared meals. Recipes represent only 10% of the items included.
The nutrients and other constituents given for each food item were: water, proteins (total N × the specific factor), lipids, available carbohydrates (soluble sugars plus starch, expressed as monosaccharides), crude fibre, iron, calcium, phosphorus, thiamine, riboflavin, niacin, retinol equivalents, vitamin C, and energy value expressed as kilo-calories and kilo-joules. Edible matter was also included. The conversion factors used for proteins, lipids, and available carbohydrates (expressed as monosaccharides) were 4.0, 9.0, and 3.75, respectively. In addition, information was given on the amino acids, fatty acids, and cholesterol content of specific food items. Revised data from previous tables, data from analyses carried out at the institute, and data taken from the literature after careful evaluation were used as sources for the tables.
In addition to their use in research and teaching and for food surveys, these tables are the official source used by the Ministry of Agriculture for the calculation of nutrients for the national food balance sheets. They are now also used by the Central Institute for Statistics (ISTAT) for the same purpose. These tables have been used by the National Institute of Nutrition as a first approach to calculate the nutrient content of food items, in terms of the daily recommended allowances of nutrients for the Italian population [4].
Food composition tables were also compiled by Fidanza and Liguori [5] based on the guidelines proposed by Southgate [2]. These tables, which were published as an appendix to a textbook on human nutrition, were revised and extended in 1981. Other food composition tables in Italy are derived largely from the above-mentioned tables.
At present, there is no national nutrient data bank in Italy. However, a number of computerized systems have been developed by food companies and research institutions. In 1982, the food composition tables of the National Institute of Nutrition were computerized, using an IBM series 1 computer. The system used for storage, retrieval, updating, and calculations was EDL, and the work was done on time sharing with a Disk, Diskette 8 system. Food items were coded with 6 digits following the groupings reported above. Within each group the food items were organized randomly. Until now, the system has been used only by the National Institute of Nutrition, mainly for research on dietary intake and heart disease.
References
5 Fidanza, F.; Liguori, G.: Lineamenti di nutrizione umana (Idelson, Napoli 1974).
A.B. Cramwinckel
At present there are two food composition tables in The Netherlands [1, 2]: The Netherlands Food Composition Table (NVT) and the table from the Committee on the Uniform Coding of Nutrition Questionnaires (UCV table).
The first edition of the NVT was prepared in 1941 from available analytical data in The Netherlands and abroad. Since then the table has been revised many times, the latest edition being the 34th which was published in 1983 [1]. Data from abroad have been gradually replaced by data from various institutes in The Netherlands. Since 1972, this table has been administered by the Committee for Food Composition Tables set up under the auspices of the Nutrition Council which is an advisory body to the Minister for Agriculture and Fisheries and to the Minister for Welfare, Public Health and Culture. The table, which is published by the Nutrition Education Office, is designed basically for use by those with a professional interest in nutrition such as dietitians, medical practitioners, and nutritionists. The latest edition contains information on the energy value and the content of 15 nutrients for 426 food items. In addition, there are also tables giving data for particular foods on their fat content, fatty acid composition, cholesterol content, maximum level of sodium expected, and alcohol content. A table with data on the sodium, potassium, magnesium, and calcium content of mineral waters is also included. In addition to the NVT, a food composition table especially for use at the household level [3] has been prepared by a working group under the auspices of the Nutrition Education Office.
The UCV table [2] is prepared and administered by a group of research workers. In 1972, a working group was set up with the object of making studies on food composition as efficient as possible by adopting uniform methods for the collection and analysis of data [4]. The results obtained from such studies would then be more comparable with one another. The UCV table or, to be more precise, the UCV nutrient data bank was established primarily with information from the NVT together with other data principally from foreign sources. In the beginning, the first priority was to extend the table to include as much data as possible with the intention of avoiding missing values. As the table developed, more attention was paid to such aspects as reliability and to recording sources of the data. In addition, information on processed foods was expanded. Since its establishment, the UCV table has developed to a stage where it is used in practically all nutrition studies in The Netherlands. This has come about for a number of reasons: (1) the aim of the table is quite clear: it is to provide a data bank on the nutrient content of foods and dishes which is as complete and accurate as possible; (2) the data in the table are checked by computer for mistakes and missing values: (3) quick and efficient mechanisms have been established for the collection and assessment of new data for inclusion in the table and sending such data to users of the table, and (4) the availability within the GVO project (development of health education in kindergarten and primary schools) in Nijmegen of computer facilities and of personnel for the coordination and provision of secretarial assistance. The UCV table is now administered from the Institute of Nutrition and Toxicology (CIVO-TNO) in Zeist.
The 1984 edition of the table [2] has data on about 1,400 foods, 35,000 nutrient values, and 70,000 references and other pieces of background information. The UCV table is thus larger than the NVT and also differs from it in a number of ways. The UCV table has few missing values, and it includes brand names and information on the origin of the data. A handbook giving portion sizes and further descriptions of foods included in the table has also been prepared [5]. Two series of computer programs have been developed for use with the UCV table. The first series enables an updated table to be brought out each year. Existing values in the table can be compared with new data proposed for inclusion, and, if the new data are accepted, the original data can be replaced by the new data. Each new piece of nutrient data is accompanied by codes indicating its source, the year it was included in the table, and the institute that delivered the proposal. The second series of computer programs enables the table to be used. Programs are available to calculate the nutrient composition of complex dishes from the proportion of the various constituent foods in a dish, to calculate the energy value from the proportion of macronutrients, and to check that the sum of the weights of the nutrients, dietary fibre, and water is equal to the weight of a food. In addition, several programs have been developed to carry out calculations on data collected from nutrition studies to assist in the interpretation of the nutrient consequences of the results obtained.
The objective in The Netherlands is to produce one common food composition table. To this end, the Nutrition Council [6] has made recommendations to the Minister for Agriculture and Fisheries and the Minister for Welfare, Public Health and Culture concerning the development and administration of a central nutrient data bank and to encourage the use of such a data book. Regardless of what is finally decided, a number of important questions will have to be addressed with respect to such a table. These include: (1) the basis on which foods are listed in the table; (2) which foods should be listed separately, and which foods should be listed together on the basis of their composition; (3) what criteria should be used for the acceptance of analytical data on foods into the table; (4) what procedures should be followed for the incorporation of new analytical data into the table; (5) how can finance be obtained to carry out food analyses on a regular basis; (6) how should the accuracy and variation of analytical data in the table be expressed, and (7) whether standard portion sizes of foods should be included in the table and, if so, how such data should be expressed.
References
G. Dybowska
Studies on the composition and nutritional value of foods in Poland began after the Second World War, and the first food composition tables were prepared in 1948 mainly from data derived from foreign sources. In 1954, new tables containing information based in part on research carried out in Poland were prepared [1]. A total of 300 food items was included, and information was provided on the content of: energy, protein, fat, total carbohydrates and fibre, ash, calcium, phosphorus, iron, vitamin A, thiamine, riboflavin, niacin, vitamin C, and the proportion of waste.
Additional food composition tables using data both from within Poland and other countries were compiled by Szabuniewicz and Kierst [2], Bielińska et al. [3], and Szczygieł et al. [4, 5]. The magnesium content was added to the list of nutrients.
In 1975 the tables were again revised [6], and data were included concerning the content of the following: sodium, potassium, copper, sulphur, chlorine, fluorine, manganese, molybdenum, iodine, cobalt, vitamin D, total tocopherols, α-tocopherol, pantothenic acid, vitamin B6, biotin, folacin (including free folic acid), vitamin B12, choline, selected fatty acids, cholesterol, and essential amino acids. Some data were obtained as the result of research carried out in Poland. In 1976 new food composition tables were compiled which included about 750 food items [7]. Food composition tables for use at the household level have also been compiled by Piekarska and Szczygieł [8]. These tables include about 300 basic foods and provide information on the main sources of energy and nutrients and give the recommended daily intake of each.
The most recently published tables [9] contain 991 food items and include data on energy content (expressed in kilo-calories and kilo-joules), water, protein, fat, total carbohydrates, fibre, ash, calcium, phosphorus, iron, magnesium, retinol, carotene, retinol equivalent, thiamine, riboflavin, niacin, and ascorbic acid. Data on each food item are provided in two ways: firstly, values are calculated per 100 g of edible part of food and secondly, for 100 g food as purchased. In addition, the proportion of waste is estimated. The previous system of dividing foods into twelve groups [10] has been abandoned. In this table, foods are divided into 38 groups on the basis of, for example, origin, main nutrients, consumption level, and, for the first time, energy and nutrient content of alcoholic and non-alcoholic beverages.
At present, several studies to update food composition tables are being coordinated and carried out by the Institute of Food and Nutrition, under the supervision of the Ministry of Health and Welfare. More detailed information on essential amino acids, fatty acids, trace elements, etc. will be included. From the above-mentioned food composition tables it has been possible to formulate recommended daily intakes for four economic levels [10]. While the recommended intakes are the same, foods are selected for each economic level on the basis of price.
Computers have been introduced to carry out a number of tasks related to food composition tables. About 10 years ago, they were used for designing balanced meals and menus with a minimum cost [11]. Electronic data processing has also been used to develop various menus in large-scale catering units and cafeterias. In 1982, the Institute of Human Nutrition, Warsaw Agricultural University, initiated a program called ‘Dieta’ which is the first attempt to develop a nutrient data bank in Poland. However, the basic purpose of this program is not to collect data, but to analyze the fate of nutrients from the raw food to the consumers' table and to compare the nutritive values of consumed foods with the recommended daily allowances or intakes. It is intended that this program be used for individuals as well as for specific groups and the population as a whole.
References
1 Rudowska-Koprowska, J.: Tablice wartości od
ywczych
produktów spoyczych (PZWL, Warszawa
1954).
7 Sadowska, H.: Warunki zdrowotne zywności i zywienia (PZWL, Warszawa 1976).
8 Pickarska, J.; Szczygieł, A.: Popularne tabele wartości
odywczych ywności (PZWL, Warszawa 1979).
9 Piekarska, J.; Łoś-Kuczera, M.: Skład i wartość odywcza
produktów spoyczych (PZWL, Warszawa
1983).
I. Martins
The first Portuguese food composition table was published by the National Institute of Health in 1961. The analytical work was carried out by the Department of Food and Chemistry in the National Institute of Health, in Lisbon and in Oporto, under the supervision of Prof. Gonçalves Ferreira. The second extended edition was published in 1963 and a third revised edition in 1977 [1]. This table is divided into three sections. The first section includes an introduction with a brief definition of the terminology used, the food groups and a description of each, and the methods of analysis used. In the second and third sections the foods are arranged according to six basic groups as follows: group 1: milk and dairy products (excluding butter which is included in the third group); group 2: meat and meat products, fish and fish products, crustaceans, molluscs, and eggs; group 3: fats and oils; group 4: cereals and cereal products, dry legumes, cocoa and confectionery; group 5: vegetables, potatoes, fruits, and nuts, and group 6: non-alcoholic and alcoholic beverages.
Information is given on the composition of more than 700 foods including raw foods, canned foods, and fruit preserves. The foods are listed in alphabetical order according to their common names; their systematic names are also given. In general, manufacturers' names have been used only where otherwise the designation of the food would have been difficult.
Samples of foods were purchased in various shops in several regions of the country so as to be as representative as possible. All the values in the tables were obtained by direct analysis of the food samples and were examined in relation to published data wherever such data were available. Values are calculated per 100 g of edible part of the food and are given for the following: water and proportion of edible matter; energy expressed in kilo-calories; carbohydrates; for fruits, values for saccharose and reducing sugars are given; fats; proteins; dietary fibre; ash; sodium chloride; carotene and vitamin A; thiamine, riboflavin and niacin; ascorbic acid; calcium, phosphorus, iron and copper, and also alcohol content in alcoholic beverages.
In the third section, values for additional constituents in a more limited range of foods are given. These constituents include: sodium, potassium, and magnesium; trace elements such as manganese, zinc, cobalt, fluorine, and iodine; cholesterol; essential fatty acids (linoleic, linolenic, arachidonic, and penta- and hexaenoic acids); essential amino acids (arginine, phenylalanine, histidine, isoleucine, leucine, lysine, methionine, threonine, tryptophan, and valine), and titratable acidity, organic acids, and pectic acid (fruits).
A revised and extended edition of the food composition table is being prepared. Changes will be introduced mainly because of changes in methods of food analysis, of food production, and of food technology. As a result of an increase in awareness of their importance, some nutrients not included in earlier editions will be added. It is also planned to include some new foods in the new edition because of their importance in the national diet.
Two new sections will also be introduced. One consists of foods consumed by children and infants (milk, baby cereals, and other baby foods) and is almost ready. The other section is still in the planning stage. It will contain values obtained by direct analysis of a range of cooked foods, individually cooked items, and also some traditional Portuguese dishes.
The Portuguese food composition table has a very important role to play in the sectors of health, agriculture, industry, and education. As in most countries, this table has been used largely for: calculation of nutrient intakes from records of food consumption, formulation of diets, calculation of desirable food supplies that will provide a specified intake of the nutrients, calculation of the nutrient composition of a manufactured food from its ingredients, and calculation of the composition of cooked dishes from recipes.
At present, in Portugal, there is no nutrient data bank. Data from the Food and Nutrition Documentation Service in the Nutrition Studies and Research Centre have been used largely by the government ministeries of Health, Agriculture and Education and also by private entities. The Nutrition Studies and Research Centre of the National Institute of Health is organizing a nutrient data base which, in a first phase, will include 104 coded foods and 16 nutrients. It will be in operation very soon.
Reference
J. Kevany
No national food composition tables are published in the Irish Republic, and few routine analyses are carried out to determine the composition of Irish foods. Some information on composition can be acquired from food manufacturers and has been included in data banks where appropriate.
McCance and Widdowson's Composition of Foods [1] forms the basis of all nutrient data banks in Ireland. These data banks are linked to academic and research institutions. At Trinity College, Dublin, attempts have been made to expand these tables and to make computer analysis of foods for cholesterol values more complete. Such modifications have also been made by other institutions with computer data banks. The Dublin Institute of Technology has developed values for lignin, cellulose, non-cellulose polysaccharide, oxalic acid, phytic acid, purine, and animal protein for appropriate foods.
At present there is no national data bank in Ireland. Dietitians involved in institutional services generally do not have access to a computerized data bank and rely on hand calculations based on the tables of Paul and Southgate [1]. No attempt has been made to coordinate nutrient data banks used in the Irish Republic, and the Government has no formal policy on food composition tables or nutrient data banks.
Reference
O. Moreiras-Varela
In 1980, the Institute of Nutrition of the Consejo Superior de Investigaciones Científicas published a review of the recommended daily intakes of energy and nutrients for various strata in the population. Thus, it has been considered necessary to provide an up-to-date analysis of the composition of those foods most commonly consumed. In the selection of foods to be included in the food composition tables, attention was directed towards those which were considered suitable for planning an ideal diet or necessary for mantaining good nutritional status.
In spite of the limitations on the use of data derived from food composition tables, such data are essential in planning nutrition programs and in implementing food policies.
These tables have been compiled from data collected from various sources within the country, and in addition data from the literature have also be incorporated. Such data were selected on the basis of the similarity of the food and the ecological conditions of the country they came from, or because of the reputation of the institution which produced the data.
Considerable attention has been given to what is referred to as the edible portion. Estimation of this is sometimes subjective and difficult to express exactly because of the great number of sources of variation to which it is subjected. Considerable error is made in the estimation of actual intake of a particular food when only the quantity bought in the market is known, whereas logically the composition refers to the edible part.
The Spanish Nutrition Society in collaboration with the Institute of Nutrition has set up a committee to revise the food composition tables on the basis of the guidelines of Southgate [1]. Contact has been made with various laboratories in Spain to undertake analyses in order to provide data which are at a present missing from the tables.
These tables are used on a national basis, and recently the Interministerial Commission for Food Planning adopted them officially as a working document. They are also being used in the National Food Survey which the Institute of Nutrition is carrying out jointly with the National Institute of Statistics in a random sample of 25,000 families.
All data incorporated in the food composition tables are stored on computer in the Institute of Nutrition and form the basis of a national nutrient data bank.
Reference
A. Blumenthal
‘Official Methods of Food Analyses in Switzerland’ [1] contains food composition tables. However, only relatively few foods are included, and many food constituents such as dietary fibre, magnesium, and zinc have not been incorporated. A new edition of this publication is being prepared, but as yet it has not been decided whether food composition tables will again be included.
Various food manufacturers and distributors have prepared leaflets setting out the nutritional composition of their products. Data are derived partly from data banks and partly from food composition tables, but mostly from the manufacturers' own analyses of their products.
There is no national data bank or specific-purpose data bank generally available to interested parties. On a national basis, reference is made to internationally approved food composition tables such as Souci/Fachmann/Kraut [2], and McCance and Widdowson [3]. The Federal Office of Public Health is not considering the establishment of a national nutrient data bank.
At present, except for foods for special diets, manufacturers participate in the nutrition labelling of their products on a voluntary basis, and the Federal Office of Public Health has no plans to make this obligatory in the near future. There are, however, special regulations with regard to the labelling of the vitamin content of foods. Manufacturers are permitted to display the vitamin content of their products only after approval has been obtained from the Federal Office of Public Health. Such approval is only granted after samples have been analyzed by the National Vitamin Institute.
Guidelines are being prepared for the nutrition labelling of foods. It is considered that the nutrient content of foods such as milk and eggs can be obtained from internationally approved food composition tables. However, the nutrient content of prepared foods, such as ice cream and canned foods, will have to be determined by analysis.
References
D.A.T. Southgate
The first food composition tables were published in the United Kingdom in 1940 by McCance and Widdowson [1] and entitled ‘The chemical composition of foods’. This was the first comprehensive series of data for foods eaten in the UK, and it was unique in several respects. First, it gave the composition of cooked foods. Secondly, it gave values for a large range of inorganic constituents. Thirdly, it included values for carbohydrates based on direct measurement, and moreover it separated the carhohydrates into two classes: available - the sugars and starches - and unavailable - those not digested in the small intestine.
The third edition of these tables was published in 1960 [2] and formed the basis of the first computerized nutrient data bank in the UK. For the first time, analyses carried out in other laboratories were included. This change was necessary because of the large volume of published work on the vitamins and amino acids in foods in which it was considered inappropriate to duplicate. Thus, the preparation of the third edition called for a very detailed and extensive examination of the literature, particularly for the vitamins and amino acids in foods.
Computerized versions of the third edition, including data from other sources, were prepared by two research groups. The Department of Health and Social Security also developed a nutrient data bank based on the third edition and supplemented with data derived by calculation from recipes and with more recent analytical data. While this computerized table was used extensively by the Department of Health and Social Security in studies on the nutrient intake of selected groups and was also used by a number of associated research groups, it was never published or made generally available.
In 1968, work began on revising the third edition of the McCance and Widdowson table. The Ministry of Agriculture, Fisheries and Food provided the organizational framework for the revision and initially with other government departments, but later alone, funded the analytical work. The work was carried out in a Medical Research Council Unit with also a substantial financial commitment. It had been decided to prepare a computerized version of the tables which would be made available and sold alongside the published book.
These data form the basis of the national nutrient data bank. The revision and updating of this data bank, which is now the responsibility of the Ministry of Agriculture, Fisheries and Food was guided initially by the advice of the Subcommittee on Food Composition Tables of the Committee on the Composition of Foods and Food Products and now by a Working Party on Nutrition. A continuous program of revision is in progress based on a detailed assessment of the literature and supplemented by a substantial analytical program. As the work on the various food groups is completed, it is proposed to publish it as supplements. Each supplement will include new data and such old data as are considered to be relevant and reliable. Eventually, the 5th edition will be published. There are plans for issuing the supplements in a computer-readable form, and these supplements will be incorporated into existing data banks. The data base is sold as a punched tape, and individual users have constructed their own programs and data management systems. Most users are involved in research studies and have added data from other sources. Several entrepreneurial groups have produced computer packages for the use of the data bank. While data are not directly accessible on-line, selected portions of data were entered into the Prestel system by the Ministry of Agriculture, Fisheries, and Food, so that these are generally accessible. This mode of operation in relation to the use of computerized nutritional information parallels the early development of food compositional data in the UK. The original impetus for the work of McCance and Widdowson arose from work in clinical nutrition in which food composition data were essential in the dietary management of diabetes. The work was developed and expanded for use in the assessment of dietary intake of individuals and populations. Thus, the UK interest in this area has always been primarily nutritional, and the major users of tables and data banks are dietitians, nutrition research groups, and government departments.
An outline of the nutrient data bank reveals the organisation of printed data as follows. The foods items are arranged into the 14 groups: cereals and cereal products; milk and milk products; eggs; fats and oils; meat and meat products; fish and fish products; vegetables; fruit; nuts; sugars, preserves and confectionery; non-alcoholic beverages; alcoholic beverages; sauces, soups, and miscellaneous foods. Within each group the arrangement is usually alphabetically within subgrouping of related types of food, although some groups are not divided in this way. The foods have code numbers and a description of the sample is given and the proportion of edible material in the food as purchased. Scientific (taxonomic) names, description of analytical methods and sources of the analytical data are given in appendices, but these do not form part of the computerized data bank. The proximate nutrients, inorganic constituents, and vitamins are given in the first section of the tables. Values for amino acids, fatty acids, and cholesterol are given in separate sections. Amino acids are quoted as milligrams per g N, and fatty acids as g per 100 g total fatty acids. A printed supplement gives amino acids and fatty acids per 100 g food [3]. The nutrients included are: total nitrogen; protein factor (specific for different food groups) × N (adjusted to allow for presence of non-protein, non-amino acid N where known); fat; sugars (lactose in milk products); starch, available carbohydrate (sum of sugars and starch); dietary fibre (sum of non-starch polysaccharides and lignin); alcohol where present; inorganic constituents, Na, K, Ca, Mg, Fe, Cu, Zn, P, Cl (S values for many foods); vitamin A, carotenes, vitamin D, thiamine, riboflavin, nicotinic acid, vitamin C, vitamin E, vitamin B6, vitamin B12, folate, pantothenic acid, and biotin; amino acids, protein amino acids (no values for hydroxypyroline); fatty acids coverage varies between food groups, individual fatty acids are given, grouped into saturated, mono-unsaturated, and polyunsaturated fatty acids.
Most values are derived by direct analysis on pooled representative samples of the food in question. The samples are based on consumption data for the UK and for proprietary products, different brands are combined according to their share of the market. Literature values are also considered and evaluated according to the criteria given by Southgate [4] and by Paul and Southgate [5]. The printed tables include a key giving the sources of all values and references to the literature where appropriate.
References
L. Arab
In order to obtain an overview of computerized nutrient data banks in Europe, a questionnaire was sent to 70 research centres in 20 countries. Information was sought on when the system was established, its design, technical specifications, and about number of foods and nutrients included, and the uses to which the system is put. In addition, each centre was asked to calculate the mean intake of various nutrients from a 2-day dietary protocol. In total, 32 centres completed and returned the questionnaire (see table II), and seven also carried out the task as requested. From the responses it is clear that there are three distinct groups of people involved with food composition tables as follows:
food analysts who carry out analysis of nutrients and contaminants in foods: their work ranges from measuring one substance in many foods to measuring many substances in one food group;
food table constructors who are middlemen collecting and assessing data for inclusion in food composition tables which are usually published on a national basis;
users of food composition tables who often have constructed nutrient data bases and developed computer programmes to enable them to be used; usually they want to calculate the nutrient or contaminant content of foods for such purposes as nutritional assessment surveys, toxicological surveys, recipe calculation, and patient counselling.
In many cases, respondents were involved in more than one of these functions as shown in tables II and III. In summary, most of the respondents are working in institutes carrying out research mainly on cancer and heart disease. However, considerable use is also made of data banks in nutritional surveys and for teaching purposes. At least one quarter of those centres surveyed offered the system to others on a contract basis.
The major criticism of the existing data systems was that no indication is given on the variability in nutrient composition and on the reliability of the chemical analyses. In general, most computer systems do not give a measure of the variability in analytical results, and it is often assumed that results are accurate without due consideration being given to the variation between food samples and between analyses from different laboratories. Concern was also expressed that most of the tables lacked information on specific nutrients for many foods included in the table. However, specific information on missing values was not sought in the questionnaire.
Table II. Information on food composition tables and nutrient data banks in Europe derived from questionnaire survey
| Country | Institute and contact person | Name of system | Data stored | Main use of table/data bank | Computer | Date established | ||
| number of foods | number of recipes | number of nutrients and contaminants | ||||||
| Austria | Forschungsinstitut der Ernährungswirtschaft Blaasstrasse 29 A-1190 Wien Mr. Pfannhauser | - | ? | 0 | 4 | food analyses | HP 1000 | 1978 |
| Ludwig-Boltzmann-Institut für Stoffwechseler krankungen und Ernährung Wolkersbergenstrasse 1 A-1130 Wien Mr. V. Veitl | - | 1,138 | 101 | 23 | - | Prime | 1981 | |
| Belgium | Department of Nutrition and Dietetics Faculty of Medicine University Hospital De Pintelaan 185 B-9000 Ghent Prof. G. Verdonk | - | 200 | - | 21 | food composition tables | - | 1970 |
| Ecole Santé Publique Campus Erasme ULB Route de Lennik 808 B-1070 Brussels Mr. Kornitzer | - | 185 | 0 | 20 | other | CDC Cyper | 1981 | |
| Institut Paul Lambin Clos Chapelle-Aux-Champs B-1200 Brussels Mr. J.M. Pyke | - | 700 | 100 | 21 | food analyses, other | Data | 1980 | |
| Denmark | National Food Institute Mørkhøj Bygade 19 DK-2860 Søborg Mr. A. Møller | Levnedsmiddeltabel | 370 | 0 | 83 | food analyses, foodcomposition tables, other | MPS 3000/ PDP 11 | 1983 |
| Federal Republic of Germany | Klinisches Institut für Herzinfarktforschung an der Medizinischen Universitätsklinik Bergheimerstrasse 58 D-6900 Heidelberg Dr. L. Arab | BLS-Bundeslebensmittelschlüssel | 5,000 | 1,000 | 140 | other | Apple II/ IBM | 1983 |
| Institut für Sozialmedizin und Epidemiologie des Bundesgesundheitsamtes Postfach 33 00 13 D-1000 Berlin 33 Prof. H. Rottka | as above | other | ||||||
| Dokumentationsstelle der Universität Hohenheim Postfach 700562 Paracelsusstrasse 2 D-7000 Stuttgart Dr. H. Haendler | Datendokumentation des Fachgebietes Futtermittelkunde (feed data documentation) | 18,000 | 0 | 800 | feed composition tables, other | ICL 2900 | 1950 | |
| Zentrale Erfassungs-und Bewertungsstelle für Umweltchemikalien des Bundesgesundheitsamtes Thielallee 88–92 D-1000 Berlin 33 | Datenbank ZEBS | ? | 0 | 18 | food composition tables | Siemens | 1974 | |
| Deutsche Forschungsanstalt für Lebensmittelchemie Lichtenbergstrasse 4 D-8046 Garching Dr. H. Scherz | Lindas | 655 | 0 | 232 | food analyses, food composition tables | Cyber 175 | 1981 | |
| ENSIG-Institut für Agrarsoziologie Bismarckstrasse 4 D-6300 Giessen Mr. H. Böing | Glanz | 2,000 | 600 | 38 | other | Cyber 174 | 1982 | |
| Finland | Department of Nutrition University of Helsinki SF-00710 Helsinki 71 Ms. M. Ahola | Ravinto-ohilmat (nutrient data processing programs) | 1,220 | 770 | 60 | food analyses, food composition tables, other | Burroughs 7800 | 1972 |
| Department of Food Chemistry and Technology University of Helsinki SF-00710 Helsinki 71 Dr. P. Varo | Mineral Composition Tables1 | 400 | 80 | 29 | food analyses, food composition tables | Burroughs 7800 | 1971 | |
| France | Fondation Française pour la Nutrition 71, av. Victor-Hugo F-75116 Paris Dr. J. C. Favier | Compaliné | 1,000 | 50 | 90 | food analyses, other | Apple II/ Sirius | 1983 |
| Inserm. Unité 63 22, av. du Doyen-Lépine F-69500 Bron Dr. S. Renaud | - | 600 | 10 | 23 | food analyses, food composition tables, other | IBM 360/ 168 | 1979 | |
| ADE 9, bd des Capucines F-75002 Paris Dr. Z. L. Ostrowski | Krystell | 1,650 | 660 | 18 | food composition tables | IBM PC XT IBM 5110 | 1976 | |
| Greece | Department of Nutrition and Biochemistry Athens School of Hydiene Hof Alexandra 196 Athens 602 Prof. A. Trichopoulou | Food composition tables2 | 280 | 110 | 4 | food composition tables | Hewlett-Packard | 1984 |
| Ireland | Department of Community Health Trinity College 196 Pearse Street Dublin Ms. C. Wickham | - | 1,000 | 100 | 51 | food composition tables, other | - | 1981 |
| Italy | Istituto Nazionale della Nutrizione Via Ardeatina 546 I-00179 Roma Dr. E. Carnovale | In, Compo | 375 | 40 | 40 | food analyses, other | IBM | 1982 |
| Istituto Scienza Alimentazione San Costanzo Casella Postale 333 I-06100 Perugia Prof. F. Fidanza | Alime Giuli-Pri | 1,072 | 59 | 32 | food analyses, food composition tables, other | Prime 550 | 1982 | |
| Esna, room no. C228 FAO, Rome Mr. W. Potaemi | Food Composition Data Management and Food Name International File | 848 | 100 | 53 | food composition tables | IBM 370/168 | 1982 | |
| The Netherlands | Voedingsraad (Netherlands Nutrition Council) Prinses Margrietplantsoen 20 Postbus 95945 2509 CX's Gravenhage Mr. B.C. Breedveld | Nederlandse Voedingsmiddeltable | 450 | 65 | 22 | food composition tables | - | 1942 |
| Institute for Nutrition and Toxlcology (CIVO-TNO) PO Box 360 3700 A.I Zeist Ms. W.G.M. Boeijen | UCV | 1,400 | 50 | 30 | other | IBM 4043 | 1973 | |
| Norway | Section for Dietary Research University of Oslo POB 1117 Blindern N-317 Oslo 3 Ms. K. Trygg | AKFPRG | 1,500 | 800 | 28 | other | DEC 10 | 1971 |
| National Society for Nutrition and Health Elisenbergveien 12 N-0265 Oslo 2 Mrs. A.H. Rimstead | Statens Ernaringsrads Mat Varetabell | 696 | 130 | 14 | food composition tables | - | 1962 | |
| Poland | Institute of Human Nutrition Agricultural University SGGW-AR Nowoursynowska 166 02–766 Warsaw Mrs. B. Kowrygo | Dieta | 700 | 200 | 87 | other | RIAD 10–35 | 1983 |
| Spain | Instituto de Nutrición Facultad de Farmacia CSIC Departamento de Fisiologia Ciudad Universitaria Madrid 28040 Dra. O. Moreiras-Varela | Tablas de Composición de Alimentos | 231 | 29 | 18 | food analyses, food composition tables. other | CBM 4032 CBM 8292 | 1981 1983 |
| Sweden | The Swedish National Food Administration Box 622 S-75126 Uppsala Mrs. L. Bergström | KOST | 1,400 | 530 | 35 | food analyses. food composition tables, other | Nord 100 CE | 1981 |
| Department of Clinical Nutrition Annedalklinikerna Sahlsgren's Hospital S-41345 Gothenburg | Calculation of Dietary Record and Dietary History | 1,300 | 360 | 33 | food analyses, other | IBM 3033 Guts | 1978 | |
| UK | MRC Dunn Nutrition Unit Milton Road Cambridge CB4 IXI Miss A.A. Paul | MWIN: Recipe | 1,300 | 300 | 68 | other | HP 1000 | 1979 |
| Food Research Institute Colney Lane Norwich NR4 7VA Norfolk Dr. D.A.T. Southgate | McCance and Widdowson Composition of Foods 4th ed. plus 1,000 series | 1,147 | 140 | 69 | food analyses, food composition tables, other | PDP 11/34 | 1982 | |
Table III. Functions of respondents to survey of those involved with food composition tables
| Function | Number of respondents |
| A | 14 |
| C | 12 |
| U | 24 |
| A+C+U | 6 |
| A+C | 2 |
| A+U | 5 |
| C+U | 1 |
A = Food analysts;
C = food table constructors;
U = users of food composition tables.
Many respondents felt that more software should be produced which would make nutrients data banks easier to use for inexperienced people. It was also considered desirable that error-checking programmes should be built in. As shown in table II the number of food items in the various data banks ranged from 21 to 5,000, the average number being about 1,000. Most data banks contained less than 100 recipes, and two contained no recipes at all. The number of nutrients or constituents ranged from 4 to 140, with an average of less than 50 per food.
No information was obtained on the initial cost of establishing a food composition table or nutrient data bank, and few people provided information on the cost of maintaining them. However, the answers obtained for the food composition tables ranged from a quarter to 3 full-time members of staff, the average being equivalent to 1 full-time member of staff. Nutrient data banks were reported to be more expensive to maintain and involved on average 2 full-time members of staff.
The most commonly cited source of data on nutrients was McCance and Widdowson's The Composition of Foods [1], although the Souci/Fachmann/Kraut tables [2] are also widely used in Europe. All the centres surveyed used a variety of other sources including their own food analyses and data supplied from food manufacturers.
A major task in the development of a nutrient data bank is the coding of data. In all centres surveyed, data are coded using numbers or an alphanumeric code with 3–12 digits, but mostly a 4-digit code is used. In the various systems, the numbers are randomly assigned, organized alphabetically by food group or alphabetically within a food group. Mostly, foods are coded according to food group. In some cases the code also contains information on the method of storage, brand names of foods, or country of origin. Generally, the amounts of foods and their codes are entered directly or using special coding forms, but occasionally, use is still made of punch cards. No centre has an optical reader. Of the 32 centres surveyed, 12 makes of computer are in use: IBM was represented by at least six different types of machine. Clearly, it will not be possible to standardize the larger computer hardware. However, this is not a serious problem, because data tapes can be made compatible with most large systems, and all centres have the machines required to read these tapes. Although the use of microcomputers is becoming increasingly attractive with their increasing storage capabilities at low prices, it will be more difficult to standardize microcomputers. The transfer of information is difficult between microcomputers, and programming is practically not interchangeable between languages. Twelve computer languages were listed as being used for storage, retrieval, updating, and calculation. The most frequently mentioned were Fortran, Cobol, and Basic. As at present few centres are using microcomputers, it would be worthwhile considering the development of standardized system for their use with nutrient data.
While it is difficult to generalize from the survey about he quality of the food analysis data included in the tables, several generalizations can be made. In the centres conducting analyses, 1–3 samples of a food item are taken, and analyses are done on 1–3 aliquots of each sample. Internal standards are not used regularly, but when used, they are run with every batch. Data based on average values, after elimination of extreme values, are entered in the tables. Much of the data in the food composition tables are obtained using unpublished methods, and data on the nutrient content of foods and obtained for different laboratories. Thus, it is highly likely, that data within the various tables are not always strictly comparable.
Each centre was asked to carry out the task of calculating the mean intake of a number of nutrients from a 2-day dietary protocol, taken from the paper by Hoover and Perloff [3]. Seven centres carried out the task which represented a considerable effort on their part from language translation to updating of systems and printout of results in the required format. The calculation of energy, macronutrients, cholesterol, fibre, four minerals and five vitamins were compared. The results obtained are shown in table IV. The most stable parameter was energy consumption, for which the standard deviation between centres was 10% of the mean. The estimations of calcium and fibre varied considerably, with deviations from the mean of 44 and 42% respectively. Differences between extreme values were greater than 50% for iron, vitamin A and the B vitamins, fibre and calcium. Reference values based on calculations using USDA Handbook No. 8 are also given in table IV.
In general, most data banks are relatively new and undeveloped, most only being in existence since 1980. As yet sophisticated programmes for nutrient interactions have not been developed, and a satisfactory way of dealing with missing values has not been found. Thus, at this early stage, the investment in the development of these nutrient data banks is not so great that compromises in the interest of standardization would be at the sacrifice of great personal time and effort. However, as progress is rapid, the issue of standardization is urgent. Those surveyed showed a great deal of interest in developing a high degree of compatibility between nutrient data banks.
Table IV. Daily intake of energy and various nutrients calculated from a 2-day dietary protocol by seven research centres
| Centre | Energy | Protein | Fat | Carbohydrate | Fibre | Cholesterol | Potassium | Calcium | Phosphorus | Iron | Vitamin A | Vitamin B1 | Vitamin B2 | Vitamin B6 | Vitamin C |
| kJ | g | g | g | g | mg | mg | mg | mg | mg | μg | μg | μg | μ | g | |
| A | 4,455 | 48 | 50 | 99 | 3.5 | 218 | 1,092 | 172 | 517 | 6.2 | 223 | 700 | 600 | - | 34 |
| B | 4,757 | 43 | 55 | 115 | - | 240 | 1,230 | 181 | 542 | 10.0 | 390 | 800 | 600 | 800 | 55 |
| C | 5,003 | 54 | 65 | 106 | 11.6 | - | 1,383 | 406 | 760 | 6.3 | 312 | 621 | 577 | 757 | 50 |
| D | 5,227 | 54 | 67 | 113 | 10.6 | - | 1,396 | 384 | 778 | 6.5 | 312 | 640 | 580 | 737 | 44 |
| E | 5,354 | 55 | 69 | 115 | 11.9 | - | 1,431 | 295 | 711 | 6.8 | 278 | 630 | 600 | 732 | 40 |
| F | 5,443 | 40 | 71 | 125 | - | 269 | 1,377 | 225 | 625 | 8.8 | 297 | 586 | 541 | 697 | 49 |
| G | 6,163 | 66 | 70 | 139 | - | 296 | - | 111 | - | 4.7 | 242 | 453 | 433 | - | 46 |
| Mean | 5,200 | 51 | 64 | 116 | 9.4 | 256 | 1,318 | 253 | 655 | 7.0 | 293 | 633 | 561 | 745 | 45 |
| SD | 547 | 8.7 | 8.1 | 13 | 4.0 | 34 | 131 | 112 | 111 | 1.8 | 55 | 106 | 60 | 38 | 7 |
| Coefficient of variation, % | 10 | 17 | 13 | 11 | 42 | 13 | 10 | 44 | 17 | 28 | 19 | 17 | 11 | 5 | 15 |
| Maximum difference | 1,708 | 26 | 20 | 40 | 8.4 | 78 | 304 | 295 | 261 | 5.3 | 167 | 347 | 167 | 103 | 21 |
| Reference1 | 5,138 | 56 | 59 | 121 | 85 | 283 | 1,272 | 272 | 626 | 7.7 | 527 | 827 | 795 | - | 45 |
1 Values based on data in reference 4.
References
4 United States Department of Agriculture: Composition of foods, Agricultural Handbook No. 8 (United States Department of Agriculture, Washington 1976).
5 Koivistoinen, P.: Mineral element composition of Finnish foods: Na, K, Ca, Mg, P, S, Fe, Cu, Mn, Zn, Mo, Co, Ni, Cr, F, Se, Si, Rb, Al, B, Br, Hg, As, Cd, Pb and ash. Acta agr. scand. 1980: suppl. 22.
6 Trichopoulou, A.: Food composition tables (Lenis, Athens 1982).
