Introduction

The nutritional composition of a food is affected by many factors, including, for example, variety, environmental factors (e.g. soil composition, climate), storage, preparation and cooking methods, and processing.  Compilers of food composition databases try to take account of the most important sources of variation, but users need to be aware of the effect that they can have on, for example, estimates of nutrient intake.

As the food supply becomes increasingly diverse, it is more and more difficult for food composition database compilers to keep abreast of the often frequent changes in composition (Gillanders et al., 2002).  One of the challenges faced is the positive trend towards the manipulation of the composition of food, usually to give a more 'healthy' nutrient profile.

The most common ways in which food composition is manipulated include:

·                     reduction in levels of, for example, fat, saturated fatty acids, sugar and salt;

·                     manipulation of animal diets to enhance the nutritional profile of resultant products such as eggs and milk;

·                     enrichment or fortification with nutrients or components with putative health benefits; and

·                     genetic modification.

This article briefly considers each of these approaches and their implications for food composition.

Reductions in levels of macronutrients and salt

'Low-calorie', 'diet' or 'healthy eating' products have been available for many years.  Typically, levels of sugar and/or fat are reduced, aided by the use of alternative sweeteners, fat replacers and other ingredients (e.g. bulking agents or emulsifiers).  Such products are well established and usually easy to identify; the most commonly consumed products may be indicated by a separate entry in food composition databases.  

As potential links between diet and chronic disease become established, there is now an increasing trend towards changes in the nutrient profiles of mainstream prepared and processed food products.  Such changes can benefit consumers without them needing to adjust their dietary habits.  For example, many food manufacturers have changed their formulations or fat sources in order to reduce the trans fatty acid content of products such as fat spreads, biscuits and cakes, as well as fast food within the catering sector.

In the UK, the Food Standards Agency is currently planning a programme to encourage food manufacturers to reduce the content of saturated fatty acids and energy (through reductions in total fat and added sugar) of processed and prepared food products (http://www.food.gov.uk/healthiereating/satfatenergy/).  Food categories being considered include: meat and meat products; milk and milk products (including milk as an ingredient, Cheddar cheese, and ice cream and dairy desserts); fat spreads, including butter; cereals and cereal products (e.g. biscuits, buns, cakes, pastries, fruit pies, pizza, breakfast cereals); potato crisps and other savoury snack-type products; confectionery; soft drinks; potato products (e.g. chips and other roasted or fried potatoes); and pastry.

The same organisation, together with the Department of Health, has also been very active in encouraging food manufacturers to reduce levels of salt in processed foods

(http://www.food.gov.uk/healthiereating/salt/saltprogressstatement/).  There has been positive progress but it is anticipated that changes will be gradual.  Although standard dietary assessment methods are usually considered to underestimate sodium intake, they are often used for convenience.  In such cases, the source of food composition data can have an important effect on assessed intakes, as illustrated below:

Sodium and salt content of a portion of baked beans on toast¹

¹           Medium portion of standard baked beans in tomato sauce on one slice of white bread

²           Holland B. et al. (1991) McCance & Widdowson's The Composition of Foods, 5^th edition.  Cambridge: Royal Society of Chemistry.

³           Food Standards Agency (2002) McCance & Widdowson's The Composition of Foods, 6^th summary edition.  Cambridge: Royal Society of Chemistry.  Sodium content of baked beans unchanged since the 5^th edition; sodium content of bread based on analytical data from 2001.

⁴           http://www.food.gov.uk/science/surveillance/fsisbranch2006/fsis1306

The newer data reflect reductions that have taken place in the sodium content of these foods, and illustrate the differences that can exist between brands.  In this case, the brand leader for baked beans had (at the time of sampling) reduced sodium content to a greater extent than other brands sampled.

Manipulation of animal diets

From a public health perspective, there is interest in reducing dietary intakes of saturated fatty acids and increasing intakes of long-chain n-3 polyunsaturated fatty acids.  This includes the potential of animal-derived foods to increase the intake of very long chain n-3 polyunsaturated fatty acids (Givens and Gibbs, 2006).

Since dairy products are an important source of saturates, the potential for manipulating the diets of cattle to reduce levels of saturates in milk has been a research focus (Demeyer and Doneau, 1999).  The Lipgene project (http://www.lipgene.tcd.ie/), an EU 6^th Framework Integrated Project, which is examining the interaction of food supply and genes in the metabolic syndrome, is considering strategies to modify the fatty acid profile of milk.

Poultry meat (http://www.lipgene.tcd.ie/) and eggs offer a potential route for increasing intakes of long-chain n-3 polyunsaturated fatty acids, through modification of poultry feed.  However, the fatty acid profile of the final product can vary depending on the fatty acid source used in the feed supplement (van Elswyk, 1997).  Eggs described as 'naturally rich in omega-3 from hens fed a vegetarian diet' are available at retail level in a number of countries, including, within Europe: Belgium, France, Germany, Greece, The Netherlands, Spain and the UK (http://www.columbus-concept.com/).

Enrichment and fortification

Both mandatory, and, to a greater extent, voluntary fortification can have an important impact on the accuracy of estimates of nutrient intake (Gillanders et al., 2002).  For example, micronutrient levels in breakfast cereals have changed over time both in terms of levels of fortification used and in the nutrients added.  When the most recent edition of the UK food composition tables was compiled, one major brand was adding calcium to popular children's breakfast cereals and the published values reflect this.  However, since calcium is no longer added to these products, use of the published values will result in an overestimate of calcium intake.

This is also an issue that users need to consider when using 'foreign' food composition databases, since fortification practices do vary between countries, although these will be brought into line with the new European Regulations on the addition of nutrients to foods, which came into force on 1^st July 2007. 

The increasing presence of 'functional foods' and so-called 'superfoods' in the food supply also poses challenges for food composition database compilers and users.  As well as the difficulties in keeping abreast of this rapidly changing sector, appropriate data may be difficult to find since many food composition databases do not contain data for bioactive compounds (e.g. plant stanol esters) and some of the fashionable ingredients sometimes described as 'superfoods' (e.g. goji berries, acai berries).

Genetic modification

There is interest in the enrichment (or 'biofortification') of staple crops with micronutrients using plant breeding and/or transgenic strategies, particularly for their potential in relation to nutritional deficiencies in developing countries.  One of the best-known examples is the development of 'golden rice', a genetically engineered form of rice enriched with beta carotene.  This development has met with some controversy, both because of the use of genetic modification and also owing to concerns that it might interfere with existing vitamin A supplementation and fortification programmes and claims that the seed is too expensive for poor subsistence farmers to afford.  There is also potential to use gene technology to enrich grains with other micronutrients such as vitamin E and folic acid (Sautter et al., 2006).

A cow producing low-fat milk has recently been identified in New Zealand (http://www.vialactia.com/).  It is thought that a natural variation in the cow's genetic make up reduces the amount of fat in her milk, while also reducing the saturated fatty acid content and increasing levels of n-3 fatty acids.  While any commercial development is unlikely in the short term, the finding raises the possibility of breeding cows that produce milk lower in fat or with an altered fatty acid composition.

Conclusion

Users of food composition databases need to keep abreast of developments that may impact on the nutritional composition of food.  In addition, depending on their intended application, they may need to consider whether published food composition data are appropriate for their requirements.  Compilers of food composition databases can help by ensuring that data sources are well-documented (e.g. date of analysis) and that food descriptions are informative.  Ideally, the analytical data on which most databases are based should be updated frequently.  Unfortunately, with the limited financial and personnel resources available and the frequency with which product composition can change, this is likely to remain an aspiration to a large degree.

EuroFIR (www.eurofir.net) is working on a number of relevant activities (see http://www.eurofir.net/workpackages for a description of the individual work packages).  For example:

• establishing and testing a standard food classification and description system for use in European food composition databases (see review at http://www.eurofir.net/technicalreports);
• working with the food industry to consider approaches to acquiring and incorporating (including evaluating) manufacturers data into food composition databases (see (http://www.eurofir.net/technicalreports for a report on 'Pilot Cases on Data Transfer from the Food Industry to European National FCDBs'; also Krines & Finglas, 2006);
• the establishment of a web-based integrated database, EuroFIR-BASIS, which includes data for bioactive components (e.g. glucosinolates, isoflavones and polyphenols) present in major European foods.

Thus, these and other activities within EuroFIR should assist both users and compilers of food composition databases by improving documentation, providing a unified source of information, and strengthening links with industry.

References

Demeyer D & Doreau M (1999) Targets and procedures for altering ruminant meat and milk lipids.  Proceedings of the Nutrition Society, 58, 593-607.

Gillanders L, Steeper A & Watts C (2002) Impact of a dynamic food supply on food composition databases.  Journal of Food Composition and Analysis, 4, 523-526.

Givens DI and Gibbs RA (2006) Very long chain n-3 fatty acids in the food chain in the UK and the potential of animal-derived foods to increase intake.  Nutrition Bulletin, 31, 104-110.

Krines C & Finglas P (2006) The industrial role and perspective in the quest for better nutritional data.  Food Science & Technology, 20 (4), 51-56.

Sautter C, Poletti S, Zhang P & Gruissem W (2006) Biofortification of essential nutritional compounds and trace elements in rice and cassava.  Proceedings of the Nutrition Society, 65, 153-159.

Van Elswyk ME (1997) Comparison of n-3 fatty acid sources in laying hen rations for improvement of whole egg nutritional quality: a review.  British Journal of Nutrition, 78, Supplement 1, S61-S69.