ARTICLE IN PRESS

JOURNAL OF
FOOD COMPOSITION
AND ANALYSIS
Journal of Food Composition and Analysis 21 (2008) S3–S12
www.elsevier.com/locate/jfca

Study Review

Applications of food composition data: Data sources and

considerations for use
Jean A.T. Pennington
Division of Nutrition Research Coordination, National Institutes of Health, Bethesda, MD 20892-5461, USA
Received 2 November 2006; received in revised form 5 February 2007; accepted 6 February 2007

Abstract

Sources of food composition data for nutrition professionals include electronic and hardcopy databases, which may be listed in
database directories, and data from the food industry and the scientific literature. When data are not available from these sources, it may
be necessary to impute or calculate values for a food, and the procedures used should be documented. Considerations for selecting data
from various sources for a particular food include the food name and description, reliability of the source, data quality, and data
variability. Food composition data are used for many purposes—dietary intake surveys/studies, reference work, patient care, student
education, product development, dietary assessment, and clinical research. Understanding of the limitations of the data should be
considered before using them for these various purposes. For example, it may not be appropriate to use data from a reference database
for clinical studies where food component intakes are related to physiological measurements. More precise data from direct analysis of
the food(s) under study may be required. Using food composition data from other countries may be appropriate, but might require
modification for local ingredients or preparation methods. The many and varied sources of food composition data should be used with
consideration for their appropriateness in national surveys, patient care, education, clinical practice, clinical research, and product
development.
r 2007 Elsevier Inc. All rights reserved.

Keywords: Food composition data; Data sources; Data uses; Food databases; Documentation; Quality; Reliability; Variability; Limitations; Food
analysis; Appropriate food data utilization

1. Introduction data. Food composition databases are never complete due

to the constant introduction of new foods into a food
Food composition databases are essential tools for supply (new cultivars/breeds, processed foods, imported
dietitians and other nutrition professionals. Food compo- foods), newly discovered food components that are
sition databases are used to assess the dietary status of associated with health and disease, and continued im-
patients, clients, and students; to assess dietary intakes of provements in methods to analyze foods for food
population groups with defined demographic characteris- components. This paper provides a brief overview of the
tics; to plan and evaluate the dietary adequacy of meals sources of food composition data and considerations for
and diets; and to discern relationships between diet, health, their use in dietetic practice in the United States (US).
and disease from the results of national, clinical, and
epidemiologic studies (Pennington, in press). The collection
and analysis of foods to provide data for food composition
2. Sources of food composition data
databases is very expensive, and there are competing needs
for the available resources to analyze new foods and newer
Food composition data that are available to dietitians
food components and to reanalyze foods that have older
and other nutrition professionals are discussed under the
groupings of US government electronic databases, data-
Tel.: +1 301 594 8824; fax: +1 301 480 3768. base directories, food industry data, and scientific litera-
E-mail address: jp157d@nih.gov. ture; examples of these sources are summarized in Table 1.

0889-1575/$ - see front matter r 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jfca.2007.02.005
 ARTICLE IN PRESS
S4 J.A.T. Pennington / Journal of Food Composition and Analysis 21 (2008) S3–S12

Table 1
On-line information about food composition

Online United States Government databases

ARS, USDA National Nutrient Database for Standard Reference, Release 18, /http://www.nal.usda.gov/fnic/foodcomp/search/S

ARS, USDA Special Interest Databases (added sugars, choline, flavonoids, fluoride, isoflavones, oxalic acid content of selected vegetables,
proanthocyanidins), /http://www.ars.usda.gov/Services/docs.htm?docid=5121S

ARS, USDA Food and Nutrient Database for Dietary Studies (FNDDS), /http://www.ars.usda.gov/Services/docs.htm?docid=7673S
ARS, USDA USDA Database of Vitamin A and vitamin E for NHANES, /http://www.ars.usda.gov/Services/docs.htm?docid=12088S
On-line search tool for 13,000 foods, /http://www.ars.usda.gov/Services/docs.htm?docid=7783S
FDA Total Diet Study (280 foods with data for pesticide residues, radionuclides, moisture, toxic and nutritional elements), /http://www.cfsan.fda.gov/
comm/tds-toc.htmlS
Online Directories
International Nutrient Databank Directory, /http://www.nal.usda.gov/fnic/foodcomp/conf/S (Follow link for PDF file.)
International Food Composition Tables Directory (International/multinational; Asia; Africa; Canada, Caribbean, US; Europe; Latin America; Middle
East; Oceania), /http://www.fao.org/infoods/directory_en.stmS
Links to Food Companies Websites
/http://lancaster.unl.edu/food/ciq-contacts.htmS
Journal of Food Composition and Analysis
/http://www.elsevier.com/wps/find/journaldescription.cws_home/622878/descriptionS

2.1. US Government Electronic Databases Nutrition Examination Survey (NHANES) (National

Center for Health Statistics, 2006). It includes the foods
The national US food composition database is the reported as being consumed by the NHANES participants.
National Nutrient Database for Standard Reference (SR), Compared with SR, the FNDDS contains more brand
which is maintained by the Agricultural Research Service names, fast foods, meats cuts that would be described by
(ARS) of the US Department of Agriculture (US Depart- consumers (less specific than the more anatomical meat cut
ment of Agriculture and Agricultural Research Service, descriptions in the SR), fewer ingredient items, and fewer
2006a). This reference database contains data derived from raw foods that are not usually consumed raw, e.g., meat,
ARS-contracted food analysis, the food industry, and the fish, poultry, grains, and some vegetables. The data for
scientific literature. The data from these sources are FNDDS are derived primarily from the SR, and missing
evaluated, compiled, and entered into the database by values are imputed so that the database is complete for
ARS staff. The SR also contains some values that are making dietary intake assessments. The FNDDS contains
estimated or calculated based on recipe ingredients or 62 food components for approximately 7000 foods.
similar foods. SR, Release 19 contains up to 140 food A separate database containing vitamin A expressed as
components for over 7000 foods. There are some missing micrograms of retinol activity equivalents and vitamin E as
values in the database because not all of the foods have milligrams of alpha-tocopherol has also been developed by
been analyzed for all the food components and not all ARS for use with NHANES (US Department of Agricul-
missing values have been estimated. The SR is routinely ture and Agricultural Research Service, 2006d). The
updated as newer data become available from analysis of FNDDS database has been used to develop the ‘‘What’s
new foods and reanalysis of foods with improved analytical In The Foods You Eat Search Tool’’ for consumers (US
methods or procedures. Department of Agriculture and Agricultural Research
The ARS Special Interest Databases (US Department of Service, 2006e). This online search tool contains nutrient
Agriculture and Agricultural Research Service, 2006b) profiles for about 13,000 foods in typical portion sizes.
provide data for added sugars, choline, flavonoids, Because of the unique features of dietary supplements,
fluoride, isoflavones, proanthocyanidins, and oxalic acid data for them are usually kept in separate databases rather
in selected foods. The added sugars were calculated by than merged with food composition databases. The unique
ARS, the fluoride database was developed under a contract features of dietary supplements (vitamins, minerals, her-
with the University of Minnesota, the isoflavone database bals, and other preparations) include variable potencies;
under a contract with Iowa State University, and the other availability as pills, capsules, tablets, and powders; and
databases contain data culled from the literature by ARS. specific brand name formulations. Dietary supplements
The ARS Food and Nutrient Database for Dietary include thousands of different products, and there are also
Studies (FNDDS) (US Department of Agriculture and different potencies of the same compound with the same
Agricultural Research Service, 2006c) is the database used brand name. A dietary supplement database was developed
for the dietary component of the National Health and by NCHS for use with the NHANES dietary intake data
 ARTICLE IN PRESS
J.A.T. Pennington / Journal of Food Composition and Analysis 21 (2008) S3–S12 S5

(Ervin et al., 1999; Dwyer et al., 2003). The Office of databases and software for commercial sales of electronic
Dietary Supplements (ODS) at the National Institutes of products.
Health (NIH) is currently working on a plan to develop a The International Food Composition Tables Directory
label-based dietary supplement database, and ODS is also maintained by the United Nations International Network
collaborating with ARS to develop an analytical-based of Food Data Systems (INFOODS) is an extensive archival
dietary supplement database (National Institutes of Health collection that lists the names of hardcopy and electronic
and Office of Dietary Supplements, 2006; Dwyer et al., databases and data tables (International Network of Food
2004; Dwyer et al., 2006, Dwyer et al., 2007). These and Data Systems, 2006). The databases and data tables listed
other databases for dietary supplements will likely evolve in the Directory are organized by seven geographic regions
over the next decade as more research is conducted on the (Asia; Africa; Canada, Caribbean, US; Europe; Latin
relationship of dietary supplements to health maintenance America; Middleast; and Oceania), and there is also an
and disease prevention. international or multinational group. Links to some of the
The Food and Drug Administration (FDA) Total Diet electronic databases are provided. The INFOODS Direc-
Study (FDA, 2006) provides annual analytical data on the tory was first compiled in September 1988 and is updated
levels of dietary minerals, folic acid, heavy metals, radio- regularly as new publications become available. Even
nuclides, pesticide residues, industrial chemicals, and other though some of the materials are quite old and out of
chemical contaminants in the core foods of the US food print, the listings provide information about the history of
supply (Food and Drug Administration, 2006; Pennington food data compilation in the various countries and regions.
and Gunderson, 1987; Pennington et al., 1996). The dietary Although the data from other countries may not be strictly
minerals included in this study are calcium, phosphorus, applicable to US foods, this Directory is a valuable source
magnesium, sodium, potassium, iron, zinc, copper, and when data for particular ethnic/regional foods are needed
selenium. In previous years the Total Diet Study also and are not available in US databases.
included analysis for iodine (Pennington and Young, 1990;
Pennington, 1990). The heavy metals in the study are 2.3. Food Industry Data
cadmium, mercury, and lead. The 280 core foods of the
US food supply are identified based on data from The food industry (both food companies and trade
NHANES and are purchased from retail markets, pre- associations) is a valuable source of information on the
pared for consumption, and analyzed individually for composition of foods. Information may be obtained from
nutrients and other food components. The Total Diet the Nutrition Facts panel on food labels, from company
Study assesses dietary intakes of the measured food and trade association websites, from exhibit booths at
components for eight age-sex group diets on an annual conferences, and through personal communication with
basis. industry representatives. A website developed by the
University of Nebraska, Lincoln Cooperative Extension
2.2. Directories of Food Composition Databases in Lancaster County (University of Nebraska-Lincoln
Extension, 2005) provides links to a number of food
Members of the US National Nutrient Databank companies and their product information. Most of the data
Conference maintain an International Nutrient Databank that industry provides on websites, exhibit booths and
Directory, which is available on-line (Braithwaite et al., personal communication is from Nutrition Facts panels for
2006). The 2006 version of this Directory includes detailed nutrition labeling rather than the actual analytical data.
information for databases and/or software developed The label data have been rounded and adjusted to be in
by 33 organizations. All of the organizations listed in compliance with FDA and USDA regulations so the data
the Directory have a food composition database, and may be considered as approximations rather than mean
some have both a database and software for dietary values for most food components. Label data is limited to
analysis, recipe calculations, and/or other functions. The the food components that are required to be on the label
vendors provide the combined database and software unless the product is fortified or makes nutrient claims, in
product or service. Directory listings are obtained by which case, the fortified or ‘‘claimed’’ nutrients must also
asking database/software developers to fill out a ques- appear on the label. Also, because the information for
tionnaire. Not all of the databases available in the US are vitamins A and C and minerals, calcium and iron, are
currently listed in the Directory; however, input is routinely presented as percent Daily Value (DV) on the label,
sought and the Directory is updated periodically. Most of calculations are required to estimate the amount of these
the US databases listed in this Directory use the SR as a food components present in the foods. Beginning in
primary source for selected foods and food components January 2006, food companies were required to add the
and then add additional foods and components to suit levels of trans fat in their products to the Nutrition Facts
the particular needs of the database. These SR-derived panels. Values for trans fat will likely be taken from the
databases are used with dietary assessment systems (for labels and added to existing databases. Although food label
research, patient care, and education) in hospitals, colleges, data has some limitations for use in food composition
universities, and clinics. Private developers create the databases, it is sometimes the only information that is
 ARTICLE IN PRESS
S6 J.A.T. Pennington / Journal of Food Composition and Analysis 21 (2008) S3–S12

available about a product. Procedures for attempting to impute values. It is important to document how the
determine the composition of products based on the label imputation was done and the basis for any calculations.
data and the ingredient listing are provided in Rand et al. This will allow users of the database to understand the
(1991) and Marcoe and Haytowitz (1993). quality of the data and will allow database compilers to
replace the values when better (analytically-based) data
2.4. Scientific Literature become available.

It takes time for the data that appear in scientific

journals to be gathered and compiled and then entered into 3. Considerations for selecting a database
food composition databases. Often journal papers are the
most timely sources of data emerging from analytical Features of food composition databases that may
laboratories. The Journal of Food Composition and Analysis distinguish one database from another include the foods
(2006) is dedicated to publishing papers about the and food components contained, the form of the database
composition of foods and about analytical methods to (hard copy or electronic file), periodicity of update, format
analyze foods for food components. Information about the on the page or computer screen, intended use(s) (e.g.,
composition of foods also appears in other journals, reference database, diet analyses, or product development),
bulletins, and other publications and can be accessed food and food component search capabilities, and asso-
through literature searches. Review articles that summarize ciated software functions. Perhaps one of the main
information about the composition of a particular food or considerations for selecting a database is that it contains
the occurrence of a specific food component in a variety of the types of foods appropriate to the target population of
foods are often very useful for database compilers and interest, whether it be a country; geographic region;
others seeking specific data. Bioactive food components as institutional setting; or a specific age, gender, or ethnic
well as the antioxidant potential of foods are of recent group, and that it contains the food components of
interest because of their association with health and disease interest. One might need a database that contains foods
prevention (Pennington, 2002; Graf et al., 2005). Recent available in school breakfast and lunch programs; foods
publications show considerable interest in the levels of served in hospitals or nursing homes; foods served at
carotenoids, flavonoids, polyphenols, and antioxidant congregate meal sites for older persons; or foods available
potential in foods (Cantos et al., 2002; Choi and Rhee, on tribal reservations or ethnic geographic areas. One
2006; Hulshof et al., 2006; Iwalewa et al., 2005; Lachman might consider if the database contains the types of fruits,
et al., 2003; Lisiewska et al., 2006; Lopez-Amoros et al., vegetables, meats, and grain products that are commonly
2006; Luthria and Pastor-Corrales, 2006; Ninfali and used by the target population and if it contains the snack
Bacchiocca, 2003; Peterson et al., 2004, 2006a, b; Raffo foods, fast foods, and ceremonial/holiday foods eaten by
et al., 2006; Sellapan and Akoh, 2002; Soong and Barlow, the group. One would need to determine if it contains the
2004; Su et al., 2006; Toor et al., 2006a; Toor et al., 2006b; food components that are of interest, i.e., food components
Van Jaarsveld et al., 2006; Vanamala et al., 2006; that might be at risk of inadequate or excess intake or those
Xianquan et al., 2005; Yilmaz and Toledo, 2006; Zhang required for government subsidies or to meet government
and Hamauzu, 2004). standards.
In addition to the foods and food components present in
2.5. Data Imputation the database, consideration should also be given to the cost
of the database and its services, data quality, availability of
Imputation refers to estimating the levels of food software for dietary assessment, and hardware needs. For
components in a food when laboratory data are not some populations, it may be appropriate to use an
available. The imputation might be based on a similar international database or selected foods from an interna-
food, e.g., data for a type of legume might be based on tional database if the foods consumed more closely match
another similar legume or data for a new cultivar of apple that database than a national database. The paper by
or pear might be based on the data for the well-known Stumbo in this issue provides more detail about considera-
cultivars. For multi-ingredient foods, the values might be tions for selecting a dietary assessment system (Stumbo,
imputed from food label data (the Nutrition Facts panel 2007). Other publications on this topic are Buzzard et al.
and the ingredient list) or they may be imputed on the basis (1991) and Probst and Tapsell (2005).
of the nutrient values for the ingredients based on a recipe.
Rand et al. (1991) provides a useful chapter on ‘‘Calcula-
tions for Multi-ingredient Foods.’’ 4. Considerations for using data
Imputation is not a preferred method of obtaining food
composition data, but it is sometimes necessary to fill in When searching for data on a particular food and
missing values in a survey database or to provide looking through several databases and the scientific
information/assistance to a patient, client, student, or literature, some things to consider are the food names
consumer. Most database compilers have had occasion to and descriptions, data quality, and data variability.
 ARTICLE IN PRESS
J.A.T. Pennington / Journal of Food Composition and Analysis 21 (2008) S3–S12 S7

4.1. Food Names and Descriptions various analyses from various countries. The quality of the
data might be gleaned from the sampling design (which
Food names and their descriptors are very important to would help determine if the foods analyzed were repre-
be sure that the associated data corresponds to the food for sentative of all foods by the same name), the number of
which information is sought. Database compilers try to samples analyzed, the analytical methods used, and the
create some consistency for the food names and descriptors reproducibility of the data (i.e., its accuracy). Data quality
to help users find the foods they need; however, in doing has been addressed in great detail in Greenfield and
this, assumptions may be made when food names are Southgate (1992). The ideal food composition database
poorly or vaguely described in the original sources. One would contain food component values from the represen-
might be looking for oranges in general or for a specific tative sampling of each food, and all analyses would be
cultivar (e.g., Valencia, navel) grown in a specific geo- based on aliquots from that representative sample. The
graphic area (e.g., Florida, California, Brazil), and reality is that the data for some foods are based on
harvested during a particular season (e.g., fall, spring). representative sampling; however, for other foods, the data
Sometimes the best one can do is to find the closest match. reflect more of a convenience sampling or an unknown
Even though the foods are somewhat different, the nutrient sampling (i.e., the data may be obtained from sources
values may be similar enough to warrant using the data. where the sample design was not provided or not provided
Sometimes a food name may pertain to more than one in detail). In general, data for any one food may come from
type of food (Table 2) (Ayto, 2002; Rubash, 1990). Because different sources. The data from different sources for the
of many food name synonyms (Table 3), it is possible that same component may be averaged or the ‘‘best’’ data may
listings with different food names and descriptions pertain be used. The data for some food components may be
to the same food (Ayto, 2002; Richter, 2004; Rubash, available from only one source, so different sources may be
1990). merged, e.g., for a food in the SR, data for folates,
The challenge in using the FNDDS and other databases flavonoids, and vitamin E may come from literature
for assessing diets is to appropriately match the foods sources, while the proximate nutrients and major vitamins
described by study participants, students, patients, or and minerals may come from several different USDA-
clients to the foods listed in the databases (that is, to select contract laboratories.
the best fit). This challenge is one reason why food
descriptions are so important. Because survey participants 4.3. Data Variability
are not always able to provide accurate descriptions of the
foods they eat, the FNDDS includes generic foods with the Food composition variability is not an indication of
descriptor ‘‘not further specified’’ (NFS) (USDA, 2006c). error, but a reflection of inherent (genetic), environmental
The nutrient data for NFS foods are based on data for (climate, temperature, soil type), processing (cooking,
similar or representative foods. For example, ‘‘sandwich, preservation method), and analytical factors. Except
NFS’’ might reflect the most commonly consumed perhaps for carefully formulated products (such as medical
sandwich in the survey (perhaps a cheese sandwich on formulas and infant formulas), the food component levels
white bread with mayonnaise), or it might be a composite in databases should be expected to have inherent and
sandwich with nutrient values calculated from weighted acquired variability. With respect to inherent variables, one
data from other sandwiches in the database. might consider the number of cultivars of each fruit and
Scientific names (i.e., Latin genus and species names) for vegetable. There are about 4000 different cultivars of
foods may be helpful when trying to find the same food in tomatoes, e.g., beefsteak, Roma, cherry; some are red,
different databases, especially those from other countries. while others are green or yellow, and there are thousands of
However, scientific names only apply to plants and varieties of apples (Rizza et al., 2002). No two pieces of
animals, not to foods with mixed ingredients, and they fruit or vegetable are exactly the same. For example, the
apply to the entire plant and animal and are not specific for average vitamin C content of a 131 g orange (2 5/800
parts, e.g., turnip roots and turnip tops have the same diameter) may be 69.7 mg (USDA, 2006a), but the actual
scientific name. Also, some of the scientific names are not vitamin C content of an orange depends on factors such as
very specific, for example most starchy beans and snap season, sunlight exposure, cultivar, species, variety, time of
beans are known as Phaseolus vulgaris; most winter and day of harvest, storage length and temperature, and
summer squashs are Cucurbita pepo; and broccoli, Brussels ripeness at harvest. Some food components are more
sprouts, and cabbage are Brassica oleracea. variable than others because they may be susceptible to loss
by heat, light, or other cooking or processing factors. Some
4.2. Data Quality recent papers provide information on food component
variability (Cantos et al., 2002; Davis et al., 2004; Judpra-
It is most helpful if the source of the original analytical song et al., 2006; Kim et al., 2005; Kuman et al., 2006;
data is identified in databases. It would be useful to know Lachman et al., 2003; Lopez-Amoros et al., 2006; Lubetzky
when the foods were analyzed, the country or origin of the et al., 2006; Luthria and Pastor-Corrales, 2006; Milinsh
foods, and/or if the data represent a compilation from et al., 2006; Nikkarinen and Mertanen, 2004; Ninfali and
 ARTICLE IN PRESS
S8 J.A.T. Pennington / Journal of Food Composition and Analysis 21 (2008) S3–S12

Table 2 Table 2 (continued )

Examples of food names with more than one meaninga
Food name Possible meanings
Food name Possible meanings
Parfait US ice cream with sauce and whipped cream;
Ambrosia Italian dessert with oranges and shredded coconut; French fruit layers with syrup, nuts, and whipped cream
Mexican herb with sweet scent used in cold beverages as
Pepper Sweet (bell) pepper (vegetable);
mint is used
Hot/mild/medium chili pepper (vegetable);
Ancho/ Deep red Mexican chili pepper; Spice-black, white, or red (cayenne) pepper
Anchoas Spanish anchovies
Pepperpot US spicy stew/soup made with tripe;
Bar/ French seabass; West Indies spicy stew/soup made with meat, seafood,
Bär Swedish berry vegetables, and cassava juice

Biscuit US dinner roll; Picaddilo Spanish salad of orange, onion, green pepper, and salted
British cookie codfish;
Mexican meat and other ingredients used as a filling/
Calabash US type of squash; stuffing
Spanish passion fruit
Prairie oyster Raw egg in liquid;
Chili Hot/mild/medium chili pepper (vegetable); Calf testicle
Chili powder (dried chili peppers, cumin, oregano,
coriander); Rape/ US source of rapeseed oil (canola oil);
Chili pepper (cayenne pepper); Rapé/ Spanish monkfish or goosefish;
Spicy meat dish (chili con carne; southwestern US Râpé French grated cheese, short for frommage rape
cuisine);
Snowball Shaved ice with sweet syrup in cone/cup;
Spicy meat and bean dish (chili con carne; other US
Snack cake covered with marshmallow and coconut
regions);
Spicy bean dish (other US regions) Squash US vegetable, summer or winter squashes (synonym
marrow in UK);
Chips US potato or tortilla chips;
Sweetened citrus drink in UK, i.e., lemon squash, orange
UK deep-fried potato pieces similar to US French fries
squash
Flan French/UK open fruit tart in sponge cake or pastry crust;
Sweetbread Internal organs (usually pancreas and thymus) of calf or
Spanish baked caramel cream custard
lamb;
Grasshopper US alcoholic beverage with green crème de menthe; Pastry
Insect sometimes used for food
Tortilla Mexican and US flat circular unleavened bread of wheat
Jelly US fruit-based spread for toast; flour or corn meal;
British gelatin dessert Spanish omelette

Kummel/ Swedish hake (fish); Truffle Fungus of genus Tuber;

Kümmel German caraway seeds; Chocolate creams covered with cocoa or chocolate
European sweet colorless liqueur flavored with caraway jimmys
and cumin seeds
Tuna US saltwater fish;
Lady fingers Small elongated sponge cakes; Mexican prickly pear (cactus fruit)
Synonym for okra
Tutti-frutti US brand of chewing gum;
Legumes/ US starchy beans and peas; UK ice cream with mixed candied fruit
Légumes Portuguese synonym for vegetables;
Twister Tropicana drink;
French synonym for vegetables
Kentucky Fried Chicken entrée
Marrow US connective tissue in bones;
a
British summer squash Sources: Ayto, 2002; Richter, 2004; Rubash, 1990.

Melt US sandwich with cheese (patty melt);

UK spleen
Bacchiocca, 2003; Peterson et al., 2004; Raffo et al., 2006;
Mincemeat US chopped fruit (apples, raisins, nuts) used as pie filling;
Robbins et al., 2005; Skupien and Oszmian, 2004; Su et al.,
UK chopped fruit with meat or suet
2006; Toor et al., 2006a, b; Van Jaarsveld et al., 2006;
Mocha US combination of coffee and chocolate; Vanamala et al., 2006; Xianquan et al., 2005; Yilmaz and
UK coffee of high grade (from Mocha seaport in Yemen)
Toledo, 2006; Zhang and Hamauzu, 2004), and Leskova
Muffin Quick bread as in blueberry muffin or bran muffin; et al. (2006) provide a comprehensive review of the effects
English muffin (regional use of term) of heat treatment of foods on vitamin levels.
Noques French gnocchi; The variability of a food composition data point is
Austrian dessert of egg-shaped meringues poached in milk best determined by looking at the standard deviation
and served with custard sauce (AKA snow eggs) (SD) and estimating the coefficient of variation (CV)
Paprika Spice made from the paprika pepper; (CV ¼ SD/mean). Examples are the calcium content of 36
Pepper used as a vegetable samples of whole milk, 101+20 mg/100 g (CV ¼ 20%),
 ARTICLE IN PRESS
J.A.T. Pennington / Journal of Food Composition and Analysis 21 (2008) S3–S12 S9

Table 3 and the iron and zinc content of 37 samples of pan-fired

Examples of food name synonymsa ground beef patty, 2.49+0.40 mg/100 g (CV ¼ 16%) and
More Synonym(s) Notes
5.79+0.56 mg/100 g (CV ¼ 10%), respectively (Pennington
common term et al., 1995). If the food composition database provides
SDs and the number of samples analyzed, users get a better
Balsam pear Bitter melon, bitter gourd Chinese fruit feel for the basis for the data.
Blood sausage Black pudding British The many causes of food component variation are
sausage made compounded in food composition databases because the
with blood data are aggregated from various sources. Because multi-
Blue cheese Roquefort (cheese, salad ingredient foods are made from mixtures of different foods,
dressing) they may have complex mixtures of inherent, environ-
Broccoli raab Rapini mental, and processing variables. Contributing further to
nutrient variation are different analytic methods and
Carbonated Club soda, seltzer, seltzer
techniques, use of different recipes to calculate nutrient
water water, soda, soda water
values for a mixed dish, and the compiler’s unique methods
Celeriac Celery root of aggregating foods and food component values.
Chicory Curly endive

Cilantro Coriander 4.4. Using Data from Other Countries

Eggplant Aubergine (UK)
If a food cannot be found in a national database, then
Frankfurter Hot dog Could be
data from other databases or sources may be used as
meat alone or
meat in a bun approximations. However, there are differences in food
names among countries which can be misleading and
Icing Frosting
confusing (See Table 2). Scientific names may help, but the
Jerusalem Sunchoke cultivars of a particular food are usually different among
artichoke countries, and this may result in different compositional
Kiwi fruit Chinese gooseberry values. The names of foods change from place to place,
both between and within countries. Cuts of beef, lamb,
Ocean perch Redfish
pork, and other meats vary around the world. The variety
Okra Gumbo, lady finger of sausages and cheeses is enormous in places like
Pancake Flapjack, hotcake Denmark and Germany. The parts of foods that are
considered edible (e.g., tops and roots of beets, rind of
Romaine Cos lettuce
cheeses, cores of apples) vary internationally and regionally
Russet potato Brown potato, Idaho potato and these differences will be reflected in the parts of the
Rutabaga Swede (UK) foods that are analyzed for food components.
Analytical methods among countries vary to some
Salsify Vegetable oyster
extent. Hopefully analysts will follow the Association of
Sherbet Sorbet Frozen Official Analytical Chemists (AOAC) procedures and
dessert made
from fruit
quality control measures (Association of Official Analytical
juice or puree Chemists International, 2005). One major difference
between the US laboratory analysis and most other
Shrimp Prawn, scampi (Italian)
countries is that for carbohydrate. The US determines
Snap beans French green beans, green
beans, string beans, wax carbohydrate ‘‘by difference’’ method (extracting out
beans dietary fiber), while most countries analyze foods for
starch and sugar content and sum these to obtain digestible
Soda Pop, soda pop
carbohydrate.
Soybean oil Soy oil; soya oil; vegetable
oil, soy
5. Matching data to intended uses
Soybeans Soya beans, soy beans
Sub Grinder, hero, Hoagie (US Cold meats, For national, regional, or local dietary intake surveys in
east coast), Italian sandwich, cheese on
Italian sub, poor boy, large sliced
the US, the FNDDS or similar survey database may work
submarine, torpedo Italian roll quite well. The FNDDS was developed for US national
with oil surveys, and may need to be adjusted (adding or removing
some foods and/or food components) for regional or local
Stuffing Dressing
surveys. It was developed to assess the dietary intake of
a
Sources: Ayto, 2002; Richter, 2004; Rubash, 1990. population subgroups and compare them with Dietary
 ARTICLE IN PRESS
S10 J.A.T. Pennington / Journal of Food Composition and Analysis 21 (2008) S3–S12

Reference Intakes (RDIs) and to assess dietary patterns rigid. One food should not be recommended over another
and make correlations with disease or health parameters. as a better or lower source of a food component unless the
For a reference in the office or classroom, the SR or a difference between the foods will be of practical importance
similar database derived from the SR is likely the choice. for the patient or client. Small differences in average values
It should be used with the understanding that there for nutrients should not be used to make comparative
are missing values for foods not yet analyzed for certain selections for foods. Rigid dietary recommendations may
components. Patient care may require dietary management override the dietary variety that is necessary to insure
for weight loss, hyperlipidemias, diabetes, low sodium adequate food component intake.
diets, protein restriction for kidney/liver disease, food
allergy, anti-inflammatory diets, etc. Again the SR or
similar database derived from the SR should serve well References
for this purpose. For certain conditions such as celiac
disease and severe sodium restriction, specialized databases Association of Official Analytical Chemists International, 2005. Official
with specific brand name information may be more Methods of Analysis of AOAC International, Eighteenth ed. Associa-
tion of Official Analytical Chemists, Gaithersburg, MD.
appropriate. Ayto, J., 2002. An A–Z of Food & Drink. Oxford University Press,
For research or clinical studies to test the effects of food London.
components on biomarkers, one should use either an Braithwaite, E., Burlingame, B., Chenard, C., Selleey, B., Stumbo, P.
isolated form of the food component, or, if a food is to be (Databank Directory Committee of the National Nutrient Databank
Conference), 2006. International Nutrient Databank Directory, 2006.
used, it should be purchased in bulk and a representative
Retrieved 2007-2-1: /http://www.nal.usda.gov/fnic/foodcomp/conf/S.
sample should be analyzed for the food component(s) of (Follow link for PDF file.)
concern to the study. Reference and survey databases are Buzzard, I.M., Price, K.S., Warren, R.A., 1991. Considerations for
usually not appropriate for feeding studies involving selecting nutrient-calculation software: evaluation of the nutrient
biomarkers. Unfortunately, some clinical researchers may database. American Journal of Clinical Nutrition 54, 7–9.
not have a good understanding of food composition and Cantos, E., Espin, C.C., Tomas-Barberan, F.A., 2002. Varietal differences
among the polyphenol profiles of seven table grape cultivars studied by
may not be knowledgeable of the variation of food LC-DAD-MS-MS. Journal of Agricultural and Food Chemistry 50,
components and rely too heavily on reference food 5696–6591.
composition data in their experiments. This is where the Choi, M.-S., Rhee, K.C., 2006. Production and processing of soybeans
knowledge of dietitians could be especially useful in and nutrition and safety of isoflavone and other soy products for
guiding researchers to the most recent food composition human health. Journal of Medicinal Food 9, 1–10.
Davis, D.R., Epp, M., Riordan, H., 2004. Changes in USDA food
data and pointing out the limitations of using reference composition data for 43 garden crops, 1950–1999. J Am Coll Nutr 23,
data in research protocols where precise food composition 669–682.
data are required. Dwyer, J.T., Picciano, M.F., Raiten, D.J., 2003. Food and dietary
supplement databases for What We Eat in America-NHANES.
Journal of Nutrition 133, 624S–634S.
6. Conclusions
Dwyer, J.T., Picciano, M.F., Betz, J.M., Coates, P.M., 2004. Mission and
activities of the NIH Office of Dietary Supplements. Journal of Food
Food composition databases contain mean values, Composition and Analysis 17, 493–500.
sometimes, but not always, accompanied by the number Dwyer, J.T., Picciano, M.F., Betz, J.M., Fisher, K.D., Saldanha, L.G.,
of samples and some error estimation (SD or Standard Yetley, E.A., Coates, P.M., Radimer, K., Bindewald, B., Sharpless,
Error (SE)). Although the values look very exact, the data K.E., Holden, J., Andrews, K., Zhao, C., Harnly, J., Wolf, W.R.,
Perry, C.R., 2006. Progress in development of an integrated dietary
should be used as approximations, especially if there is no supplement ingredient database at the NIH Office of Dietary
sampling or variation information. Some of the data are Supplements. Journal of Food Composition and Analysis 19,
based on relatively few samples and some of the data are S108–S114.
imputed. Dwyer, J.T., Picciano, M.F., Betz, J.M., Fisher, K.D., Saldanha, L.G.,
Yetley, E.A., Coates, P.M., Milner, J.A., Whitted, J., Burt, V.,
Dietitians should consider food component variability
Radimer, K., Wilger, J., Sharpless, K.E., Holden, J., Andrews, K.,
when making practical use of the data for patients, clients, Roseland, J., Zhao, C., Schweitzer, A., Harnly, J., Wolf, W.R., Perry,
and students. When comparing the food component C.R., 2007. Progress in developing analytical and label-based dietary
content of several foods, take into account the variability supplement databases at the NIH Office of Dietary Supplements.
of the mean levels and whether the differences in levels of Journal of Food Composition and Analysis, this issue, doi:10.1016/
food components are of any practical significance. To j.jfca.2007.07.010.
Ervin, R.B., Wright, J.D., Kennedy-Stephenson, J., 1999. Use of dietary
determine whether one food is a better source than another supplements in the United States, 1988-94. Vital and Health Statistics,
for a particular food component or a range of food Series 11, Number 244.
components requires practical considerations. One might Food and Drug Administration, 2006. Total Diet Study. Retrieved 2006-
look at the difference of the mean concentrations of a food 10-26: /http://www.cfsan.fda.gov/comm/tds-toc.htmlS.
component in two foods with respect to the %DV, the Graf, B.A., Milbury, P.E., Blumberg, J.B., 2005. Flavonols, flavones,
flavanones, and human health: Epidemiological evidence. Journal of
RDIs, and/or cost. Medicinal Food 8, 281–290.
Considering the extent of food component variation, diet Greenfield, H., Southgate, D.A.T., 1992. Food Composition Data:
recommendations for patients and clients should not be Production, Management, and Use. Elsevier Applied Science, London.
 ARTICLE IN PRESS
J.A.T. Pennington / Journal of Food Composition and Analysis 21 (2008) S3–S12 S11

Hulshof, P.J.M., van Roekel-Jansen, T., van de Bovenkamp, P., West, Pennington, J.A.T., 1990. Iodine concentrations in US milk—variation
C.E., 2006. Variation in retinol and carotenoid content of milk and due to time, season, and region. Journal of Dairy Science 73,
milk products in The Netherlands. Journal of Food Composition and 3421–3429.
Analysis 19, 67–75. Pennington, J.A.T., 2002. Food composition databases for bioactive
International Network of Food Data Systems, 2006. International food components. Journal of Food Composition and Analysis 15,
Food Composition Tables Directory. Retrieved 2006-10-26: /http:// 419–434.
www.fao.org/infoods/directory_en.stmS. Pennington, J.A.T., in press. Development and use of food composition
Iwalewa, E.O., Adewunmi, C.O., Omisore, N.O.A., Adebanji, O.A., data and databases. In: Monsen, E. (Ed.), Research, Second ed.
Azike, C.K., Adigun, A.O., Adesina, O.A., Olowoyo, O.G., 2005. Pro- American Dietetic Association, Chicago, Illinois.
and antioxidant effects and cytoprotective potentials of nine edible Pennington, J.A.T., Gunderson, E.L., 1987. History of the Food and
vegetables in Southwest Nigeria. Journal of Medicinal Food 8, Drug Administration’s Total Diet Study—1961–1987. Journal of the
539–544. Association of Official Analytical Chemists 70, 772–782.
Journal of Food Composition and Analysis, 2006. Retrieved 2006-10-26: Pennington, J.A.T., Young, B.E., 1990. Fe, Zn, Cu, Mn, Se, and I in foods
/http://www.elsevier.com/wps/find/journaldescription.cws_home/ from the US total diet study. Journal of Food Composition and
622878/descriptionS. Analysis 3, 166–184.
Judprasong, K., Charoenkiatkul, S., Sungpuag, P., Vasanachitt, K., Pennington, J.A.T., Schoen, S.A., Salmon, G.D., Young, B., Johnson,
Nakjamanong, Y., 2006. Total and soluble oxalate contents in Thai R.D., Marts, R.W., 1995. Composition of core foods of the US food
vegetables, cereal grains and legume seeds and their changes after supply, 1982–1991. Journal of Food Composition and Analysis 8,
cooking. Journal of Food Composition and Analysis 19, 340–347. 129–169.
Kim, H.-K., Ye, S.-H., Lim, T.-S., Ha, T.-Y., Kwon, J.-H., 2005. Pennington, J.A.T., Capar, S.G., Parfitt, C.H., Edwards, C.W., 1996.
Physiological activities of garlic extracts as affected by habitat and History of the Food and Drug Administration’s Total Diet Study (Part
solvents. Journal of Medicinal Food 8, 476–481. II), 1987–1993. Journal of the Association of Official Analytical
Kuman, V., Rani, A., Solanki, S., Hussain, S.M., 2006. Influence of Chemists International 79, 163–170.
growing environment on the biochemical composition and physical Peterson, J.J., Beecher, G.R., Bhagwat, S.A., Swyer, J.T., Gebhardt, S.E.,
characteristics of soybean seed. Journal of Food Composition and Haytowitz, D.B., Holden, J.M., 2006a. Flavanones in grapefruit,
Analysis 19, 188–195. lemons, and limes: A compilation and review of the data from the
Lachman, K., Pronek, D., Hejtmankova, A., Dudjak, J., Pivec, V., analytical literature. Journal of Food Composition and Analysis 19,
Faitova, K., 2003. Total polyphenol and main flavonoid antioxidants S74–S80.
in different onion (Allium cepa L.) varieties. Horticulture Science 30, Peterson, J.J., Dwyer, J.T., Beecher, G.R., Bhagwat, S.A., Gebhardt, S.E.,
142–147. Haytowitz, D.B., Holden, J.M., 2006b. Flavanones in oranges,
Leskova, E., Kubikova, J., Kovacikova, E., Kosicka, M., Porubska, J., tangerines (mandarins), tangors, and tangelos: a compilation and
Holcikova, K., 2006. Vitamin losses: retention during heat treatment review of the data from the analytical literature. Journal of Food
and continual changes expressed by mathematical models. Journal of Composition and Analysis 19, S66–S73.
Food Composition and Analysis 19, 252–276. Peterson, J., Dwyer, J., Jacques, P., Rand, W., Prior, R., Chui, K., 2004.
Lisiewska, Z., Kmiecik, W., Korus, A., 2006. Content of vitamin C, Tea variety and brewing techniques influence flavonoid content of
carotenoids, chlorophylls and polyphenols in green parts of dill black tea. Journal of Food Composition and Analysis 17, 397–406.
(Anethum graveolens L.) depending on plant height. Journal of Food Probst, Y.C., Tapsell, L.C., 2005. Overview of computerized dietary
Composition and Analysis 19, 134–140. assessment programs for research and practice in nutrition education.
Lopez-Amoros, M.L., Hernandez, T., Estrella, I., 2006. Effect of Journal of Nutrition Education and Behavior 37, 20–26.
germination on legume phenolic compounds and their antioxidant Raffo, A., La Malfa, G., Fogliano, V., Maiani, G., Quaaglia, G., 2006.
activity. Journal of Food Composition and Analysis 19, 277–283. Seasonal variations in antioxidant components of cherry tomatoes
Lubetzky, R., Littner, Y., Mimouni, F.B., Dollberg, S., Mandel, D., 2006. (Lycopersicon esculentum cv. Naomi F1). Journal of Food Composi-
Circadian variations in fat content of expressed breast milk from tion and Analysis 19, 11–19.
mothers of preterm infants. Journal of the American College of Rand, W.M., Pennington, J.A.T., Murphy, S.P., Klensin, J.C., 1991.
Nutrition 25, 151–154. Compiling Data for Food Composition Data Bases. United Nations
Luthria, D.L., Pastor-Corrales, M.A., 2006. Phenolic acids content of University Press, Hong Kong.
fifteen dry edible bean (Phaseolus vulgaris L.) varieties. Journal of Richter, H., 2004. Fresh Produce Guide. Try-Foods International,
Food Composition and Analysis 19, 205–211. Incorporated., Apopka, Florida.
Marcoe, K.K., Haytowitz, D.B., 1993. Estimating nutrient values of Rizza, R.A., Go, V.L., McMahon, M.M., Harrison, G.G. (Eds.), 2002.
mixed dishes from label information. Food Technology, 69–73. Encyclopedia of Foods. A Guide to Healthy Nutrition. Academic
Milinsh, M.C., das Gracas Padre, R., Hayashi, C., de Oliveira, C.C., Press, San Diego.
Visentainer, J.V., de souza, N.E., Matsushita, M., 2006. Effects of feed Robbins, R.J., Keck, A.-S., Banuelos, G., Finley, J.W., 2005. Cultivation
protein and lipid contents on fatty acid profile of snail (Helix aspersa conditions and selenium fertilization alter the phenolic profile,
maxima) meat. Journal of Food Composition and Analysis 19, glucosinolate, and sulforaphane content of broccoli. Journal of
212–216. Medicinal Food 8, 204–214.
National Center for Health Statistics, 2006. National Health and Rubash, J., 1990. Master Dictionary of Food and Wine. Van Nostrand
Nutrition Examination Survey (NHANES). Retrieved 2006-10-26: Reinhold, New York.
/http://www.cdc.gov/nchs/nhanes.htmS. Sellapan, S., Akoh, C.C., 2002. Flavonoids and antioxidant capacity of
National Institutes of Health, Office of Dietary Supplements, 2006. Dietary Georgia-grown Vidalia onions. Journal of Agriculture and Food
Supplement Ingredient and Labeling Databases. Retrieved 2006-10-26: Chemistry 50, 5338–5342.
/http://dietary-supplements.info.nih.gov/Health_Information/Dietary_ Skupien, K., Oszmian, J., 2004. Comparison of six cultivars of
Supplement_Ingredient_and_Labeling_Databases.aspxS. strawberries (Fragaria x ananassa Duch) grown in northwest Poland.
Nikkarinen, M., Mertanen, E., 2004. Impact of geological origin on trace European Food Research Technology 219, 66–70.
element composition of edible mushroom. Journal of Food Composi- Soong, Y.Y., Barlow, P.J., 2004. Antioxidant activity and phenolic
tion and Analysis 17, 301–310. content of selected fruit seeds. Food Chemistry 88, 411–417.
Ninfali, P., Bacchiocca, M., 2003. Polyphenols and antioxidant capacity of Stumbo, P., 2007. Considerations for selecting a dietary assessment
vegetables under fresh and frozen conditions. Journal of Agriculture system. Journal of Food Composition and Analysis, this issue,
and Food Chemistry 51, 2222–2226. doi:10.1016/j.jfca.2007.07.011.
 ARTICLE IN PRESS
S12 J.A.T. Pennington / Journal of Food Composition and Analysis 21 (2008) S3–S12

Su, X., Duan, J., Jiang, Y., She, J., Kakuda, Y., 2006. Effects of soaking US Department of Agriculture, Agricultural Research Service, 2006d. USDA
conditions on the antioxidant potentials of oolong tea. Journal of Database of Vitamin A (mcg RAE) and Vitamin E (mg AT) for National
Food Composition and Analysis 19, 348–353. Health and Nutrition Examination survey 1999–2000. Retrieved
Toor, R.K., Savage, G.P., Heeb, A., 2006a. Influence of different types of 2006-10-26: /http://www.ars.usda.gov/Services/docs.htm?docid=12088S.
fertilizers on the major antioxidant components of tomatoes. Journal US Department of Agriculture, Agricultural Research Service, 2006e.
of Food Composition and Analysis 19, 20–27. What’s In The Foods You Eat Search Tool. Retrieved 2006-10-26:
Toor, R.K., Savage, G.P., Lister, C.E., 2006b. Seasonal variations in the /http://www.ars.usda.gov/Services/docs.htm?docid=7783S.
antioxidant composition of greenhouse grown tomatoes. Journal of Vanamala, J., Reddivari, L., Yoo, K.S., Pike, L.M., Patil, B.S., 2006.
Food Composition and Analysis 19, 1–10. Variation in the content of bioactive flavonoids in different brands of
University of Nebraska-Lincoln Extension, 2005. Cook It Quick! orange and grapefruit juices. Journal of Food Composition and
Contacting Food-Related Companies for More Information. Re- Analysis 19, 157–166.
trieved 2006-10-26: /http://lancaster.unl.edu/food/ciq-contacts.htmS. Van Jaarsveld, P.J., Marais, D.W., Harmse, E., Nestel, P., Rodriguez-
US Department of Agriculture, Agricultural Research Service, 2006a. Amaya, D.B., 2006. Retention of beta-carotene in boiled, mashed
Search the USDA National Nutrient Database for Standard Refer- orange-fleshed sweet potato. Journal of Food Composition and
ence, Release 19. Retrieved 2006-04-11 from the Nutrient Data Analysis 19, 321–329.
Laboratory Home Page: /http://www.nal.usda.gov/fnic/foodcomp/ Xianquan, S., Shi, J., Kakuda, Y., Yueming, J., 2005. Stability of lycopene
search/S. during food processing and storage. Journal of Medicinal Food 8,
US Department of Agriculture, Agricultural Research Service, 2006b. 413–422.
Products & Services, Data sets prepared by USDA-ARS’s Nutrient Yilmaz, Y., Toledo, R.T., 2006. Oxygen radical absorbance capacities of
Data laboratory. Retrieved 2006-10-26: /http://www.ars.usda.gov/ grape/wine industry byproducts and effect of solvent types on
Services/docs.htm?docid=5121S. extraction of grape seed polyphenols. Journal of Food Composition
US Department of Agriculture, Agricultural Research Service, and Analysis 19, 41–48.
2006c. USDA Food and Nutrient Database for Dietary Studies Zhang, D., Hamauzu, Y., 2004. Phenolics, ascorbic acid, carotenoids and
(FNDDS). Retrieved 2006-10-26: /http://www.ars.usda.gov/Services/ antioxidant activity of broccoli and their changes during conventional
docs.htm?docid=7673S. and microwave heating. Food Chemistry 88, 503–509.