Conversion to nutrients
Dietary analysis is a two-step process involving the assessment of food consumption (using dietary assessment tools) and the conversion of reported intakes into nutrient data. The increase in research into relationships between diet and health outcomes has led to a need for up to date and reliable food composition databases.
Food composition databases provide detailed information on the nutritional composition of foods. These databases may be either paper-based, commonly referred to as food composition tables (e.g. McCance and Widdowson’s The Composition of Foods) or computer-based food composition databases (e.g. the United States Department of Agriculture National Nutrient Database, USDA, or the electronic versions of McCance and Widdowson, The Composition of Foods integrated dataset, CoF IDS). Currently there are over 150 food composition tables and electronic databases in existence worldwide (EuroFIR). The LanguaL website (an international framework for food description) provides links to food composition databases from various countries which are currently available online.
Food composition databases provide values for energy and nutrients (e.g. protein, vitamins and minerals) for each of the foods listed. These values are traditionally derived from chemical analysis carried out in analytic laboratories and are usually based on either several samples of a food or a composite sample of each food (Gibson, 2005). Nutrient values are usually expressed as the nutrient content of the edible portion of the food per 100g or per common household measure (Gibson, 2005). The number of foods listed in a food composition database can vary widely and will depend on whether the database contains values for composite foods (e.g. lasagne), manufactured products, recipes and dietary supplements in addition to individual foods. Studies of the relationship between diet and health have led to a greater interest in a range of biologically active constituents present in foods and some food composition databases such as the USDA National Nutrient Databank now contain values for some bioactive compounds e.g. isoflavones. Other food components which may be found in some databases include: different types of fatty acids e.g. n-3 fatty acids, trans fatty acids, plant sterol and stanol esters, and carotenoids e.g. lycopene (EuroFIR).
The development of computerised databases has been particularly useful for nutritional epidemiologists. They hold a much greater volume of data compared with printed tables and are more flexible allowing for the manipulation of data such as adding new foods and nutrient values. The most recent version of the USDA’s nutrient database contains information on the nutritional content of over 7000 different foods. The UK Food Standards Agency who are responsible for maintaining the UK Food Composition Tables have brought together all the UK Composition of Foods data into a single file. The UK Composition of Foods Integrated Dataset (CoF IDS) provides nutrient data for 3423 foods (CoF IDS) in a single electronic file divided across 16 Excel worksheets. The CoF IDS will supersede the paper versions of the Composition of Foods tables. These are currently available as the 6th summary edition of McCance and Widdowson’s The Composition of Foods (CoF) which provides nutrient information on more than 1,200 of the most commonly consumed foods in the UK and nine supplements on specific food groups such as vegetable dishes, cereals and cereal products, and a supplement on fatty acids.
Ideally, each country should compile its own food composition database. The level of certain nutrients in some foodstuffs will not differ between countries, but for other nutrients, factors such as soil conditions, climate, cultivars and farming practices can influence concentrations to a much greater degree (Greenfield & Southgate, 2003). Also, recipes for the same composite dish may vary both within and between countries, as might fortification practices (e.g. the fortification of flour and breakfast cereals). Although national food composition databases are needed, international studies such as the European Prospective Investigation into Cancer and Nutrition (EPIC) and the International Study of macro- and micro-nutrients and Blood Pressure (INTERMAP), have highlighted the need for standardizing food composition data produced at the national level (Church, 2008). For example, researchers carrying out the INTERMAP study (a four-country study investigating relationships between individual dietary intakes and blood pressure) utilised existing nutrient databases from each of the four countries, then updated them. Information on new foods were added, and, as far as possible, the analytical methods used to estimate carbohydrate composition data (monosaccharide equivalents) and fibre content, for example, were standardised to increase the comparability between countries in the analyses of nutrient intake. Modifications included the addition of new foods and preparation methods, addition of missing nutrient fields specific to the study objectives and imputation of missing nutrient values, achieved through matching to the most similar food to provide comprehensive nutrient data for food items reported by study individuals (Schakel et al, 2003). Other research groups have also explored the need to revise nutrient databases to accurately estimate nutrient intake in dietary assessments (Ishihara et al, 2006).
The emergence of long term extensive cohort studies such as EPIC and INTERMAP have highlighted the need to increase comparability in food composition information between countries. In recognition of this a number of collaborations and networks have been set up. Examples include the International Network of Food Data Systems (INFOODS) and a number of European initiatives such as the European Food Information Resource Network (EuroFIR), and NORFOODS, a project group established to help co-ordinate work on Nordic food composition data, database systems and nutrient analyses.