Caballero B. (ed.) Encyclopaedia of Food Science, Food Technology and Nutrition. Ten-Volume Set

Подождите немного. Документ загружается.

USDA), the food label offers complete, useful and

accurate nutrition information about almost every

food in the grocery store, in easy-to-read formats to

help consumers make healthful food choices.

0052 Among other features, the regulations define cer-

tain nutrient claims (such as light and good source)

and list acceptable health claims about relationships

between a nutrient or food and the risk of a disease or

health-related condition. Some acceptable relation-

ships are: calcium and osteoporosis, fat and cancer,

saturated fat and cholesterol and coronary heart

disease, sodium and hypertension, and folic acid and

neural tube defects.

0053 For dietary supplements, the framework for FDA

regulation is set forth in the 1994 Dietary Supplement

Health and Education Act (DSHEA). (Under

DSHEA’s definition, the term ‘dietary supplement’

includes, with some exceptions, products intended

for ingestion as supplements to the diet. This includes

vitamins; minerals; herbs, botanicals, and other

plant-derived substances; and amino acids and con-

centrates, metabolites, constituents and extracts of

these substances.)

0054 Under DSHEA, FDA approval of supplement

ingredients and products is not required before

marketing, a feature that clearly distinguishes supple-

ment regulation from the agency’s regulation of other

products such as drugs and many additives used in

conventional foods. Manufacturers are responsible

for ensuring that the ingredients are safe and that

the ingredient list is accurate.

0055 Dietary supplements may carry certain claims on

their labels – for example, certain nutrient content

claims like ‘good source’ or ‘high’ are allowed if the

product meets the definition, as are FDA-approved

health claims.

0056 Under DSHEA, too, once a dietary supplement

is marketed, FDA has the responsibility to prove

that the supplement is unsafe before it can take

action to restrict the product’s use. FDA focuses its

resources based on risk, so food products associated

with public health emergencies are assigned the

highest priority, followed by products that may have

caused injury or illness. Then, products thought to be

fraudulent or otherwise in violation of the law are

analyzed.

Information Exchange: Stakeholder Input,

Education

0057 FDA strives to get broad input into its decisions,

through public meetings, advisory committees, and

other sources. By providing the opportunity for ex-

change of ideas and opinions between FDA and

the public, including consumers, patients, health

professionals, and the regulated industry, the agency

is aided in reaching the best policy decisions.

0058Chartered advisory committees, such as the Na-

tional Advisory Committee on Microbiological Cri-

teria for Foods, the FDA Science Board, and CFSAN’s

Food Advisory Committee, play an important role in

providing scientific advice. Internationally, FDA par-

ticipates in Codex Alimentarius, in expert consult-

ations, and on committees of the World Health

Organization and Food and Agricultural Organiza-

tion, to name a few.

0059In addition to the agency’s notice-and-comment

rulemaking process, supplementary outreach efforts

are sometimes necessary to achieve optimal industry

compliance. The following examples illustrate the

types of outreach the agency engages in to inform

audiences.

0060FDA’s response to several sprout-associated food-

borne illness outbreaks can serve as an example of

how the agency reaches out to educate different

groups on their respective food safety roles. Following

the outbreaks in the late 1990s, FDA issued two

guidance documents advising sprout growers and

seed suppliers of steps to reduce microbial contamin-

ation: Reducing Microbial Food Safety Hazards for

Sprouted Seeds and Sampling and Microbial Testing

of Spent Irrigation Water During Sprout Production.

In addition, FDA and the California Department of

Health Services produced an educational video on

good agricultural and manufacturing practices for

sprout producers. Also, to educate the general public

on ways to protect themselves from falling ill from

sprouts, FDA issued advisories directly to consumers

on the potential hazards associated with eating raw

sprouts.

0061To disseminate food-safety messages effectively to

various audiences, including consumers and food

producers and preparers, FDA participates in the

Partnership for Food Safety Education, which

includes representatives from several government

agencies as well as from industry and consumer

organizations.

0062For several years, the Partnership has been under-

taking a Fight Bac (Fight Bacteria) campaign to edu-

cate consumers on four simple principles of preventing

food contamination while preparing food at home:

clean, chill, cook, and combat cross-contamination.

Major corporations have included the ‘Bac’ character

and the food-safety messages in their national con-

sumer-education initiatives.

0063To improve outbreak detection and response when

outbreaks do occur, FDA has developed training

courses. For example, FDA’s satellite courses on

food microbiology, foodborne disease epidemiology,

and traceback in outbreak investigations have been

2598 FOOD AND DRUG ADMINISTRATION

attended by thousands of government and industry

representatives from around the country. FDA has

also conducted presentations at numerous confer-

ences in the USA, Mexico, and Latin America on

how to conduct tracebacks to determine the source

of an outbreak.

006 4 But these examples are a selection of food-safety

efforts undertaken by FDA and its food-safety part-

ners. Perhaps the most comprehensive information on

food-safety efforts in the USA and FDA’s food science

and nutrition roles is available on the internet – at

FDA’s website, www.fda.gov, and the joint federal

food-safety website, www.foodsafety.gov.

Efforts Paying Off

0065 Educational efforts and other food-safety programs

have reaped dramatic rewards in terms of improved

food-safety practices, as evidenced by recent CDC

and FDA data. CDC data suggest that there was a

21% decline in foodborne illness in the USA from the

year 1996 to 1999. This trend continued in 2000 and

2001.

0066 According to two recent phone-tracking surveys,

the incidence of the risky practice of eating raw foods

was significantly reduced in 1998 from 1993 numbers

– down 33% for pink hamburger, 39% for raw

oysters and clams, and 29% for raw egg-containing

foods. The surveys showed similarly encouraging

improvements in consumer behavior in terms of

washing hands and cutting boards to prevent cross-

contamination between raw and other food.

0067Despite the apparent effectiveness of steps to

reduce foodborne illness in the USA, food-related

hazards may never be eliminated altogether, and so

far remain a significant threat to the public health. In

an effort to extend the downward trend in foodborne

illness, FDA must keep up with changes in the science

and technology used in agency-regulated products, in

the ways in which consumers obtain information

about products, in consumer demographics and con-

sumption habits, and in the composition of global

trade and production.

0068By remaining on the cutting edge of science and by

staying true to the tradition of science-based regula-

tory decision-making initiated by Wiley a century

ago, FDA can ensure that its regulatory decisions

are made in the best interests of the public health,

while at the same time not unduly stifling industry

innovation.

See also: Hazard Analysis Critical Control Point;

Legislation: History

Foodborne Diseases See Food Poisoning: Classification; Tracing Origins and Testing; Statistics;

Economic Implications; Microbiology: Classification of Microorganisms; Detection of Foodborne Pathogens and

their Toxins; Diarrheal (Diarrhoeal) Diseases; Colon: Diseases and Disorders; Celiac (Coeliac) Disease;

Clostridium: Occurrence of Clostridium perfringens; Detection of Clostridium perfringens; Food Poisoning by

Clostridium perfringens; Occurrence of Clostridium botulinum; Botulism; Escherichia coli: Occurrence; Detection;

Food Poisoning; Occurrence and Epidemiology of Species other than Escherichia coli; Food Poisoning by Species

other than Escherichia coli

FOOD COMPOSITION TABLES

D A T Southgate, Norwich, Norfolk, UK

Introduction

0001 Knowledge of the composition of foods is essential

for the dietary management of patients and for

most quantitative research on human nutrition. This

information has traditionally been provided by

tables of food composition and over the last 20 or

so years, increasingly by computerized nutritional

databases. The principal features of food composition

tables and nutritional databases are very similar, and

in this article, compilations will be used to describe

both printed tables and computerized databases, and

only when the comments relate to one of the types

specifically will they be distinguished.

FOOD COMPOSITION TABLES 2599

0002 These compilations of the nutritional composition

of foods are principally used for estimating the nutri-

tional composition of diets from measurements of

food consumption, constructing diets in clinical

nutrition, and designing rations and menus. Data on

the composition of foods from these compilations are

also widely used in nutritional surveillance and in

assessments of the adequacy of food supplies and as

authoritative sources for nutritional labeling. Com-

positional data are also widely used in nutrition edu-

cation and the development of nutritional guidance

for public health guidance.

General Description

0003 Food compositional tables and nutritional databases

share many common features. Initially, the databases

were prepared from the printed tables, but now, the

compilation processes are based on computerized

systems that form the starting point for preparing

the printed tables, often in different formats for

different types of use.

0004 The basic format is a matrix of food items against

nutrient data. The formats of printed tables are con-

strained by the dimensions of the printed page and the

size of the tables as a whole. Two formats are used,

one where the values for a number of foods and

nutrients are given on a single page (or a sequence

of pages) and a second, where the values for one food

are given per page. Databases are most frequently

presented as a food/nutrient matrix, the dimensions

of which are constrained only by the database

management system being used.

Introductory Text

0005 Food composition tables have an introductory text.

In databases, this forms part of the documentation of

the database and may be available on line or provided

as hard copy. These provide the user with the ration-

ale adopted by the compilers in selecting the food

items that are included and the range of nutrients

and other constituents for which values are given.

The definitions of the constituents included describe

the modes of expression adopted and the analytical

methods that were used to obtain the values. In most

cases, the values will be derived by a combination of

analytical values measured specifically for the com-

pilation and others obtained from the literature, in-

cluding other databases, and the text will document

the sources.

Coverage of Foods

0006 Most human diets include a very wide range of foods,

especially cooked dishes made up from a number of

ingredients. Most compilers of nutritional data work

within limited budgets, and therefore, the number of

food items in all compilations is restricted. The pri-

mary objective of all compilers is to present compos-

itional values that are representative for the food

items consumed by the population for which the

compilation is prepared.

0007In a national database, priority is given to those

foods that form the major part of the food intake of

the population. These have been identified using na-

tional data on food consumption and a range of other

information from market surveys, information on

food purchasing patterns, and data from the food

industry. The foods consumed by specific groups

such as infant formulas, which provide all the dietary

intake of young infants, are also included. The com-

pilations often include the foods eaten by ethnic

groups within the population and those with different

dietary patterns. Many compilations also include food

items whose composition can be seen as examples of

different types of food. Most national compilations

do not provide information on the composition of

specific brands of foods, unless the brand name is

critical for identifying the food or, more rarely, the

brand is a very important brand leader.

Food Groups

0008The foods are usually arranged in food groups that

bring together foods related to one another. Sub-

groups are often used within these major groupings,

but this varies from compilation to compilation.

Table 1 gives two widely used groupings. The groups

used by the Food and Agricultural Organization

(FAO) of the UN are linked to those use in classifying

foods in commerce and are widely used in tables

prepared for regional use.

Food Identification and Nomenclature

0009Most compilations assign a code number to every

food item. These numbers are widely used when cal-

culations of nutrient intakes are being made when the

numbers are added to the food intake records to

facilitate entry into the database calculations. Systems

of numbering have been developed for assigning

numbers that classify and identify the food for inter-

national use. These coding systems are used in some

national compilations, but not universally.

0010Food names These are given in a formalized way

and include descriptors as facets, which further serve

to identify the foods. Unequivocal identification of a

food item and relating it with its composition are very

important for the users of nutritional databases.

These naming facets include details of the way in

which a foodstuff has been cooked or prepared for

consumption. Even within one country, the naming of

2600 FOOD COMPOSITION TABLES

foods is not uniform, and the same food may be

known by different names and vice versa. Most com-

pilations include a thesaurus giving these alternative

names and frequently include a list of the scientific

(taxonomic) names for those foods for which this is

appropriate.

0011 Descriptions of samples analyzed These give the

sources and numbers of samples on which the data

are based. Most compilations give the nutrient values

on the basis of the edible portion, and these descrip-

tions include details of how the edible material was

prepared and the nature of the inedible matter re-

moved. Many compilations include a factor to con-

vert the values on the edible portion back to a value

on a raw or ‘as purchased’ basis. This information is

used to give an indication of the reliability and quality

of the nutritional compositional values.

Coverage of Constituents

0012 Most compilations include a reasonably comprehen-

sive range of nutrients, but the growth of nutritional

research interests in the role of other biologically

active constituents is increasing pressures for

extending the number of constituents to include

these ‘nonnutrients.’ The coverage is usually deter-

mined by the most demanding of the users, which

tend to be nutritional epidemiologists and nutritional

researchers. The typical coverage in a nutritional

compilation is given in Table 2.

0013 The inclusion of a constituent is conditional on

there being a satisfactory method for analyzing the

nutrient and the existence of a sufficient body of

information on the amounts of the nutrient in a

range of foods to assess its inherent variability. This

latter factor will determine whether or not it is possible

to provide representative data for the nutrient in a

range of foods. Some data for nutrients (or other con-

stituents) may only be available for a limited number of

foods, and in some cases, the distribution of a nutrient

is limited, and many compilations include subsidiary

tables giving this more limited coverage.

Expression of Nutrient Values

0014The nutrient values are usually given per 100 g of

edible matter, and the international preference is for

the nutrient to be given in weight terms. A number of

values given in the compilations are derived values.

0015Derived values Energy values are for metabolizable

energy (energy that can be used by the body) and are

calculated from the protein, fat, carbohydrates, and

alcohol in foods using energy conversion factors.

The most commonly used are those of Atwater, in

which protein is assigned, 4 kcal g

1

(16 kJ g

1

), fat,

9 kcal g

1

(37 kJ g

1

), carbohydrates (total by differ-

ence), 4 kcal g

1

(16 kJ g

1

) and alcohol, 7 kcal g

1

(29 kJ g

1

0016In the UK and some other compilations, a modified

system is used where carbohydrates (as monosac-

charides) are assigned 3.75 kcal g

1

(15 kJ g

1

Some compilations, the US for example, use ‘Specific

Factors’ that are designed to allow for the differences

in the heats of combustion of different food compon-

ents and differing digestibilities.

0017Most compilations discount the small amounts

of metabolizable energy derived from dietary fiber

from the short-chain fatty acids produced by its

tbl0001 Table 1 Examples of major food groups in nutritional compilations

Groups used in FAO tables Groups used in UK tables

Cereals and grain products Cereals and cereal products

Starchy roots, tubers, and fruits (Included within vegetables)

Grain legumes and legume products (Included within vegetables)

Nuts and seeds Nuts

Vegetables and vegetable products Vegetables

Fruits Fruits

Sugars and syrups Sugars, preserves, and confectionery

Meat, poultry, and game Meat and meat products

Eggs Eggs

Fish and shellfish Fish and fish products

Milk and milk products Milk and milk products

Oils and fats Oils and fats

Beverages Beverages

Alcoholic beverages

Miscellaneous Miscellaneous

The FAO database is most widely used for measurements of food

purchases and includes mainly raw foods. The groupings are

linked to classes of foods used to categorize economic aspects

of food production and consumption

The UK database is widely used for studies of individual food

intakes and therefore includes many cooked dishes within

the food group that appears most appropriate

FOOD COMPOSITION TABLES 2601

fermentation in the large intestine. Similarly, the

energy contributed by organic acids is discounted.

0018 Protein values are calculated from the total nitro-

gen values using conversion factors, which allow for

the different amounts of nitrogen in proteins from

different sources.

0019These conversion systems are described in the

introductory documentation.

0020Modes of expression of vitamins Most vitamins are

expressed on a weight basis, except where there are

several vitamers with differing activities when many

tbl0002 Table 2 Coverage of nutrients in nutritional compilations

Constituent Mode of expression Coverage Comments

Water Most

Total nitrogen Many

Protein Calculated from total nitrogen All Some use common 6.25 factor; others

use specific factors for different food

groups

Amino acids mg per amino acid per gram of N or

mg per 100 g of food

Usually in subsidiary

tables

Fat, total All

Fatty acids Fatty acid % total fatty acids or g per

100 g of food

Usually in subsidiary

tables

Compilations may include amounts of

saturated, monounsaturated and

polyunsaturated acids

Carbohydrates total Often ‘by difference’ or by

summation of sugars and starches

All Summation is the preferred approach

Sugars May be expressed as the

monosaccharides

Total values in many Individual sugars given in a few

databases

Starches May be expressed as glucose or as

the polymer

Many

Nonstarch polysaccharides

(NSP)

Usually expressed as polymers Limited

Dietary fiber NSP in UK; total dietary fiber values

in US and many other compilations

Most Data rather limited; some older

compilations give ‘crude fiber’

values

Alcohol g per 100 g or per 100 ml May be restricted to relevant foods

Energy value kcal or kJ, calculated from protein, fat,

carbohydrate, and alcohol

Choice of conversion factors differs

between databases

Inorganic nutrients

Na, K, Ca, Mg, P, Cl Usually mg per 100 g Most

Cu, Fe, Zn Most

F, I, Se, etc. mg per 100 g of food Limited

Vitamins

Fat-soluble

A Retinol mg per 100 g Most

D Cholecalciferol mg per 100 g

E a-Tocopherol mg per 100 g

Carotenes mg per 100 g b-carotene or retinol

equivalents

Most Few give values for different

carotenoids

Vitamin K Phylloquinone Limited

Water-soluble

Vitamin C Usually as mg per 100 g Most

Thiamin

Riboflavin

Niacin

Vitamin B

Limited Few give values for vitamers

Folates mg per 100 g Many Mostly total values

Vitamin B

mg per 100 g of food Most

Pantothenic acid Usually mg per 100 g of food Limited

Biotin

Other constituents

Cholesterol mg per 100 g of food Many

Phytic acid mg per 100 g of food Limited usually in

supplementary tables

Purines mg per 100 g of food Very few

2602 FOOD COMPOSITION TABLES

compilations use a mode of expression designed to

give the vitamin equivalence.

0021 Vitamin A activity is contributed by retional (pre-

formed vitamin A) and the provitamin A carotenoids.

The values may be given as retinol equivalents (mg

RE); b-carotene is usually divided by 6, and other

carotenoids are usually divided by 12 to give retinol

equivalents. These conversion factors are currently

the subject of scientific debate at the present time,

and expression of the carotenoids on a weight basis

is preferred.

0022 Some compilations give vitamin E values as milli-

grams of a-tocopherol equivalents calculated from

the biological equivalence of the different tocopherols

and tocotrienols.

0023 Many compilations include a value for niacin

equivalents, where the tryptophan content (mg) is

divided by 60 and added to the niacin value.

0024 Amino acids and fatty acids Amino acid compos-

ition may be expressed in one of two ways: either as

milligrams of amino acid per gram of total nitrogen

or milligrams of amino acid per 100 g of food. Fatty

acids may be expressed as fatty acid percent of total

fatty acids or milligrams of fatty acid per 100 g of

food.

Compilation of Nutritional Data

0025 As mention before, the compilation of food compos-

ition tables and nutritional databases follow similar

principles. The compilers are faced with three major

decisions: the choice of food items to include, the

choice of constituents, and how to source the data.

0026 Because the compilations are essential tools to be

used, these decisions are taken in consultation with

the range of users of the nutritional data. Analo-

gously, decisions about the format of both tables

and databases are based on how the data will be

used and software developed to meet the users’ re-

quirements. This is usually done by forming a steering

group to guide the compilers in their work. This

group meets regularly during the progress of the

work to monitor progress.

Choice of Food Items

0027 Decisions in this area are the most difficult for the

compiler because of the very wide range of foods

eaten in the human diet. Users want as complete

coverage as possible, and many would like the

inclusion of branded foods. Complete coverage of

the range of foods on sale in a developed country

would require a database with tens of thousands of

food items. Furthermore, complete coverage of all

the foods consumed, which in the main consist of

complex, mixed, cooked dishes, expands this number

possibly by two orders of magnitude. All compilers

therefore have to consider the priorities for inclusion

of food items very critically. The priorities for inclu-

sion are to give coverage of the most important foods

for the population for which the compilation is

intended and the inclusion of foods important in the

diets of specific groups such as infants, ethnic groups,

and other groups whose dietary patterns differ from

that of the population as an whole.

0028The compilers have to develop the systems of

nomenclature used and decide on the descriptions of

the food items in consultation with the users.

Selection of Nutrients

0029This is usually a slightly easier task because the

numbers of nutrients are finite. The number of bio-

logically activity constituents in plants, where there

is growing research interest, is extremely high. The

compilers must be satisfied that reliable analytical

methods exist, and the variation in composition of a

single food is not too great to prejudice the provision

of representative data.

Sources of Compositional Data

0030Compilers have two major sources of data. First, data

that are obtained specifically for the inclusion in the

compilation, which has been called the ‘direct

method.’ This involves the development of sampling

and analytical programs for obtaining the data. This

approach has the advantage in that the compilers

have control over the sampling protocols to insure

the identity of the food and to ensure that the sample

is representative. They can also control the choice of

analytical methods and the quality-assurance schemes

in operation in the analytical laboratory. The com-

pilers are thus completely familiar with the proven-

ance and quality of the data. This approach is,

however, rather expensive. It can also be argued that

this approach discounts the large amount of informa-

tion available on the composition of foods in the

scientific and other literature.

0031Second, the ‘Indirect method.’ In this, the literature

on the composition of foods is searched for suitable

data. This approach involves the compilers in making

judgements about the applicability of the samples

analyzed, their identity, and whether or not the sam-

pling protocol ensured that a representative sample

was analyzed. Judgements also need to be made about

the choice of analytical methods and the control of

their execution. The modes of expression of data may

also need to be converted into that which will be used

in the database. This approach requires the compilers

to make judgements about the data they select from

FOOD COMPOSITION TABLES 2603

the literature, and the assessment process is often

demanding and time-consuming. In practice, most

national compilations are based on a combination

of the two approaches with literature searching typic-

ally forming the first stage in the data-compilation

process. This is followed by the design of sampling

and analytical protocols whenever the literature data

are judged to be inappropriate for use, or the search

shows that few or no data exist for a food item, which

for other reasons, is required in the database.

Using Food-composition Data

0032 The uses of databases can be considered to be of three

major types: nutritional analyses, where the database

is used to calculate nutrient intakes from records of

food consumption; nutritional synthesis, where the

database is used in calculations to generate diets

with specific nutrient contents; and as a primary

source of compositional information.

Nutritional Analyses

0033 In these, the calculations may be based on food con-

sumption measured at many different levels. These

levels have important implications for the compilers

because they set the requirements for compositional

information appropriate to the level of measurement.

These are summarized in Table 3.

0034 Food-supplies level Governments and international

agencies use measurements at this level for assessing

the nutritional adequacy of food supplies. The food-

balance sheets of the Food and Agricultural Organ-

ization of the United Nations are one important

example of this level of use, which uses a special

compilation that gives the composition of raw food

at the wholesale level; e.g., the composition of bulk

cereals and animal carcasses. Many national compil-

ations also include values for this type of data.

0035 Food-purchases level These measurements here are

of two types; household budget studies, usually

undertaken primarily for economic reasons where

food consumption is derived from records of house-

hold purchases in monetary terms, and Food Surveys,

where the amounts of food purchased by a household

are measured, of which the UK National Food Survey

is an example. Both these types of use require data for

the composition of foods ‘as purchased.’ Household

budgetary surveys also require information on

retail prices to convert monetary expenditure into

quantities of food.

0036Individual food-intake level These use records of

food consumption by individuals either in a weighed

intake study or, retrospectively, where the amounts of

food consumed are obtained by dietary recall or a

food-frequency questionnaire. These types of study

provide the most detailed records of foods consumed

and have the most demanding requirements for the

inclusion of food items in a database because the

records include specific branded foods and, more

importantly, cooked mixed dishes, that is foods ‘as

consumed.’

Nutritional Synthesis

0037The compilations are widely used by dieticians and

other health care professionals to calculate diets

that will provide specific amounts of nutrients, for

example, low-energy diets for the obese, controlled

energy, carbohydrate and fats for the diabetic patient,

and low-sodium diets in hypertension. This type of

calculation requires the database to contain the range

of food likely to be eaten by the patients in question,

but the quantitative requirements may not be very

demanding. Where, for example, in metabolic stud-

ies, diets with very precise composition are required,

analyses of the foods are necessary.

0038At the present time, most of these diets are formu-

lated by manual calculation from the database, but

nutritional syntheses programs, which include indi-

vidual food preferences, are becoming available to

computerize the calculation.

tbl0003 Table 3 Requirements for food items and nutrients at different levels of data usage

Level of use Food items Nutrients General features of compilations

Food supplies Foods measured at raw, bulk

or commodity level

Usually restricted to

macronutrients and energy

Relatively small database is

required

Food purchases

Household budget surveys Foods as purchased, retail raw,

prices

Macronutrients, major

inorganic nutrients and vitamins

Information needed on losses of

food during preparation

Food-consumption surveys Foods as purchased retail or

obtained raw

Macronutrients, major

micronutrients

Information needed on losses of

food during preparation

Individual food intakes Foods as consumed, cooked

including mixed dishes

Comprehensive coverage of

macro- and micronutrients

Requirements may extend to

very detailed coverage of

constituents

2604 FOOD COMPOSITION TABLES

Sources of Compositional Information

0039 In addition to their use in providing compositional

data for guiding food choices for health reasons, nu-

tritional databases are now widely used in nutritional

labeling, where it forms part of food regulation. In

most countries, where nutritional labeling is required

(or permitted), the data taken from an authoritative

nutritional database are permitted. The food manu-

facturer or distributor is thus relieved from carrying

out analyses on their products. The regulations have

the requirement that the item in the database must

correspond to the food being labeled; generic labeling

of foods such as meats, fruits, and vegetables is often

permitted.

0040 In the USA, the nutritional labeling regulations

require manufacturers to carry out their own ana-

lyses, using specified methods defined by the Associ-

ation of Official Analytical Chemists, and also

specifies that the sample shall be representative of

production within certain limits. This has provided

the US National Database with a great deal of manu-

facturers’ data on branded foods. The UK database

also uses some manufacturers’ data where the

requirements about choice of analytical method and

that the sampling was representative can be met.

Accuracy in Use

0041 Judgements in this area depend on the use that is

being made of the compilation.

0042 Many users of nutritional databases require a semi-

quantitative guide in order to give general dietetic

advice, but some dietetic uses are more critical, for

example, in formulating diets for diabetics and low-

sodium diets in hypertension; even so, the accuracy

required is not as demanding as for those who con-

duct nutritional analyses with the database in which

nutrient intakes are calculated from records of food

intakes.

0043 Relatively few large formal studies have been

undertaken, but these show that the calculated values

for energy, protein, fat, and available carbohydrates

(sugars and starch), are usually within a few percent

of the analyzed values. The calculated values for

potassium, calcium, phosphorus, and magnesium

show a similar agreement. Analytical values for

sodium are often in excess of calculated estimates

because of the addition of salt in cooking, and

analytical values for iron are similarly slightly higher,

possibly because of iron contamination. Those nutri-

ents for which natural variability is high tend to be

the more difficult to predict with accuracy. Thus, the

micronutrients are usually predicted with a lower

accuracy, say around 10%, except for vitamin C and

folates, where the calculated values may be substan-

tial overestimates.

0044It is important in comparisons of calculated values

with analyzed values that the amounts of foods

making up the analyzed diets be precisely the same

amounts as those used in the calculations; where

collections are not supervised closely, the tendency is

to undercollect.

0045Where a compilation is being used as a source of

compositional data on a specific food, the accuracy

depends on two major factors, first, how closely the

food item in the compilation corresponds to the food

in question, and second, the natural variability of the

food. All users of nutritional data on foods must take

into account that foods are biological materials and

show natural variations in composition. This is true,

even for processed foods that are produced under

closely controlled conditions.

Limitations

0046Many of the limitations have been mentioned briefly

in the earlier sections. They can be considered under

three headings.

Coverage of Foods

0047As is clear from the section on Compilation of Nutri-

tional Data, no existing databases are truly compre-

hensive in the sense of covering all the foods likely to

form part of the human diet. The analytical resources

required to construct such a comprehensive database

are very great indeed, and coverage of all cooked

mixed dishes is virtually impossible. Additionally,

the production of new food products and the reformu-

lating of existing products are a continuous activity of

the food industry. This means that complete coverage

of all food items is a practically impossible aim. Users

of nutritional compilations therefore have to evolve

strategies for dealing with missing foods.

0048In most cases, choosing a food that is biologically

related or similar to the missing food is satisfactory. In

the case of cooked dishes, calculation from the recipe

using data on the loss of weight on cooking provides

an acceptable strategy, but the estimation of vitamin

losses in cooking means that the calculated values

must be regarded as estimates.

Coverage of Nutrients

0049Most compilations provide coverage of all the macro-

nutrients and most of the vitamins and major inor-

ganic nutrients. The principal gaps at present are the

trace elements for which, despite their nutritional

desirability, it is difficult for the compilers to provide

representative values for the reasons discussed earlier.

FOOD COMPOSITION TABLES 2605

The coverage of the carotenoids in most databases is

limited at present, but in the future, one should expect

better coverage. The same will be possible for the

different folate vitamins, although the analytical

methods are possibly less well developed. Many data-

bases give limited coverage of the carbohydrates in

foods. Coverage of the nonnutrient biologically active

components represents a greater problem because of

the great number of potential constituents coupled

with the variety of methods, but nutritional research

interests in these components should eventually

identify the most important constituents to include.

Data Quality

005 0 Quality implies fitness for its purpose, so the uses

made of databases need to be considered in any

discussion of the limitations on the quality of the

data in databases.

0051 Most databases give only mean values, and only a

few include limited estimates of variation. The accur-

acy of databases will always be limited by the natural

variability in the composition of foods, and this

places a limit on the predictive accuracy of any data-

base both for predicting the composition of any one

sample of a food and in the calculations made from

the database. These limitations decline as the number

of food records used in the calculations increases so

that, in effect, the foods eaten approach a representa-

tive sample size.

0052 The major limitation in producing databases is

the documentation of sampling and the analytical

methods used, and particularly the use of quality-

assurance schemes in analytical laboratories. In

many countries, the need for quality assurance

schemes in analytical laboratories producing data on

food composition, particularly for use in databases, is

becoming accepted. Research is needed to develop

objective quality-assurance schemes for sampling

protocols to ensure that this important aspect of

data quality is covered.

0053 However, the quality of the data in most nutritional

databases at present is considerably superior to the

quality of food-intake data, and this is the major

limitation in using nutritional databases to assess

nutrient intakes.

Compatibility of Databases

005 4 The growth in the study of nutritional epidemiology

at the international level has focused special attention

on the compatibility of national databases, and in

this, data quality is an essential aspect. A number of

international initiatives have been undertaken to ad-

dress these issues of these INFOODS (International

Network of Food Data Systems) and FAO have been

very prominent, and a number of European activities

have been undertaken, including Eurofoods and

Cost 99.

Seealso: Bioavailability of Nutrients; Carbohydrates:

ClassificationandProperties; Carotenoids:Occurrence,

Properties,andDetermination; Cereals:Contributionto

theDiet; Dairy Products – Nutritional Contribution;

Dietary Fiber:Determination; Eggs:Structureand

Composition; Fermented Milks:DietaryImportance;

Fruits of Temperate Climates:CommercialandDietary

Importance; Meat:NutritionalValue; Milk:Dietary

Importance; Nutritional Surveillance:InIndustrialized

Countries; Protein:FoodSources; Plant Antinutritional

Factors:Characteristics