See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/5854957

Carbohydrate Terminology and Classification

Article in European Journal of Clinical Nutrition · January 2008

DOI: 10.1038/sj.ejcn.1602936 · Source: PubMed

CITATIONS READS

406 31,945

2 authors, including:

John H Cummings
University of Dundee
346 PUBLICATIONS 50,198 CITATIONS

SEE PROFILE

All content following this page was uploaded by John H Cummings on 04 December 2015.

The user has requested enhancement of the downloaded file.

 European Journal of Clinical Nutrition (2007) 61 (Suppl 1), S5–S18
& 2007 Nature Publishing Group All rights reserved 0954-3007/07 $30.00
www.nature.com/ejcn

REVIEW
Carbohydrate terminology and classification
JH Cummings1 and AM Stephen2

1
Gut group, Division of Pathology and Neuroscience, Ninewells Hospital and Medical School, Dundee, UK and 2Population Nutrition
Research, MRC Human Nutrition Research, Elsie Widdowson Laboratory, Cambridge, UK

Dietary carbohydrates are a group of chemically defined substances with a range of physical and physiological properties and
health benefits. As with other macronutrients, the primary classification of dietary carbohydrate is based on chemistry, that is
character of individual monomers, degree of polymerization (DP) and type of linkage (a or b), as agreed at the Food and
Agriculture Organization/World Health Organization Expert Consultation in 1997. This divides carbohydrates into three main
groups, sugars (DP 1–2), oligosaccharides (short-chain carbohydrates) (DP 3–9) and polysaccharides (DPX10). Within this
classification, a number of terms are used such as mono- and disaccharides, polyols, oligosaccharides, starch, modified starch,
non-starch polysaccharides, total carbohydrate, sugars, etc. While effects of carbohydrates are ultimately related to their primary
chemistry, they are modified by their physical properties. These include water solubility, hydration, gel formation, crystalline
state, association with other molecules such as protein, lipid and divalent cations and aggregation into complex structures in cell
walls and other specialized plant tissues. A classification based on chemistry is essential for a system of measurement, predication
of properties and estimation of intakes, but does not allow a simple translation into nutritional effects since each class of
carbohydrate has overlapping physiological properties and effects on health. This dichotomy has led to the use of a number of
terms to describe carbohydrate in foods, for example intrinsic and extrinsic sugars, prebiotic, resistant starch, dietary fibre,
available and unavailable carbohydrate, complex carbohydrate, glycaemic and whole grain. This paper reviews these terms and
suggests that some are more useful than others. A clearer understanding of what is meant by any particular word used to
describe carbohydrate is essential to progress in translating the growing knowledge of the physiological properties of
carbohydrate into public health messages.
European Journal of Clinical Nutrition (2007) 61 (Suppl 1), S5–S18. doi:10.1038/sj.ejcn.1602936

Keywords: carbohydrate; sugars; oligosaccharides; starch; dietary fibre; classification

Introduction density, large bowel cancer, constipation and resistance to

The dietary carbohydrates are a diverse group of substances
with a range of chemical, physical and physiological proper-
ties. While carbohydrates are principally substrates for
energy metabolism, they can affect satiety, blood glucose
and insulin, lipid metabolism and, through fermentation,
exert a major control on colonic function, including bowel
habit, transit, the metabolism and balance of the commensal
flora and large bowel epithelial cell health. They may also be
immunomodulatory and influence calcium absorption.
These properties have implications for our overall health;
contributing particularly to the control of body weight,
diabetes and ageing, cardiovascular disease, bone mineral
Correspondence: Professor JH Cummings, Division of Pathology and
Neuroscience, Ninewells Hospital and Medical School, Dundee DD1 9SY, UK.
E-mail: j.h.cummings@dundee.ac.uk
gut infection.
Classification
As for other macronutrients, the primary classification of
dietary carbohydrates, as proposed at the Joint Food and
Agriculture Organization (FAO)/World Health Organization
(WHO) Expert Consultation on Carbohydrates in human
nutrition convened in Rome in 1997 (FAO, 1998), is by
molecular size, as determined by degree of polymerization
(DP), the type of linkage (a or non-a) and character of
individual monomers (Table 1). This classification is analo-
gous to that used for dietary fat, which is based on carbon
chain length, number and position of double bonds and
their configuration as cis or trans. A chemical approach is
necessary for a coherent and enforceable approach to
measurement and labelling forms the basis for terminology
 Carbohydrate terminology and classification
JH Cummings and AM Stephen
S6
Table 1 The major dietary carbohydrates Two principal approaches to total carbohydrate are used,
a
first, that derived ‘by difference’ and second, the direct
Class (DP ) Subgroup Principal components
measurement of the individual components that are then
Sugars (1–2) Monosaccharides Glucose, fructose, galactose combined to give a total. Calculating carbohydrate ‘by
Disaccharides Sucrose, lactose, maltose, difference’ has been used since the early 20th century and
trehalose is still widely used around the world (Atwater and Woods,
Polyols (sugar Sorbitol, mannitol, lactitol,
1986; United States Department of Agriculture, 2007). The
alcohols) xylitol, erythritol, isomalt,
maltitol moisture, protein, fat, ash and alcohol content of a food are
Oligosaccharides Malto- Maltodextrins determined, subtracted from the total weight of the food and
(3–9) (short-chain oligosaccharides the remainder, or ‘difference’, is considered to be carbohy-
carbohydrates) (a-glucans)
drate. There are, however, a number of problems with this
Non-a-glucan Raffinose, stachyose, fructo and
oligosaccharides galacto oligosaccharides, approach in that the ‘by difference’ figure includes non-
polydextrose, inulin carbohydrate components such as lignin, organic acids,
Polysaccharides Starch Amylose, amylopectin, modified tannins, waxes and some Maillard products. In addition to
(X10) (a-glucans) starches
this error, it combines all the analytical errors from the other
Non-starch Cellulose, hemicellulose, pectin,
polysaccharides arabinoxylans, b-glucan, analyses. Also, a single global figure for carbohydrates in
(NSPs) glucomannans, plant gums and food is uninformative because it fails to identify the many
mucilages, hydrocolloids types of carbohydrates and thus to allow some under-
a standing of the potential health benefits of those foods.
Degree of polymerization or number of monomeric (single sugar) units.
Based on Food and Agriculture Organization/World Health Organization Direct analysis of carbohydrate components and summa-
‘Carbohydrates in Human Nutrition’ report (1998), and Cummings et al. tion to obtain a total carbohydrate value has been the basis
(1997). of carbohydrate analysis in the UK since 1929, when the first
values were published by McCance and Lawrence (1929).
and an understanding of the physiological and health effects Those countries that use McCance and Widdowson’s, The
of these macronutrients. Composition of Foods (Food Standards Agency/Institute
A chemical approach divides carbohydrates into three of Food Research, 2002) also express carbohydrate using
main groups, sugars (DP1–2), oligosaccharides (short-chain this approach. The total figure obtained is for what McCance
carbohydrates) (DP3–9) and polysaccharides (DPX10). and Lawrence called ‘available carbohydrate’ and therefore
Sugars comprise (i) monosaccharides, (ii) disaccharides and differs from carbohydrate by difference in that it does not
(iii) polyols (sugar alcohols). Oligosaccharides are either (a) contain the plant cell wall polysaccharides (fibre). In
malto-oligosaccharides (a-glucans), principally occurring addition, it is not complicated by analytical difficulties with
from the hydrolysis of starch and (b) non-a-glucan such as other food components. Dietary intake of total carbohydrate
raffinose and stachyose (a galactosides), fructo- and galacto- and its components using direct analysis enables examina-
oligosaccharides and other oligosaccharides. Polysaccharides tion of geographic variations and changes in intake over
may be divided into starch (a-1:4 and 1:6 glucans) and non- time of individual carbohydrate types and their relationship
starch polysaccharides (NSPs), of which the major compo- with health outcomes. Total carbohydrate by direct measure-
nents are the polysaccharides of the plant cell wall such as ment is preferable and simplified methods to do this should
cellulose, hemicellulose and pectin but also includes plant be developed.
gums, mucilages and hydrocolloids. Some carbohydrates, Figures obtained for carbohydrate by difference and
like inulin, do not fit neatly into this scheme because they carbohydrate analysed directly are not always the same,
exist in nature in multiple molecular forms. Inulin, GFN, particularly for complex mixtures, and foods containing fibre
from plants may have from 2 to 200 fructose units and so or certain types of starch, like pasta (Stephen, 2006). This
crosses the boundary between oligosaccharides and poly- results in apparently different carbohydrate intakes for the
saccharides (Roberfroid, 2005). same list of foods consumed, as shown in Table 2. Fifty-two
A variety of methodologies are available for the measure- dietary records from a study conducted in Canada, where
ment of the carbohydrate content of food and the compo- carbohydrate by difference is used (Health Canada, 2005)
nents are listed in Table 1 (Englyst et al., 2007). were subsequently analysed in the UK using values based on
McCance and Widdowson’s The Composition of Foods (Hol-
land et al., 1991b, 1992). In this study, energy intake was 12%
higher and carbohydrate intake 14% higher when measured
Terminology ‘by difference’ (Stephen, 2006). Comparison of carbohydrate
intake among different countries should therefore be viewed
Total carbohydrate with caution if the method of carbohydrate determination is
Although the individual components of dietary carbohy- not the same. Worldwide variations in carbohydrate intake
drate are readily identifiable, there is some confusion as to assumed to be due to differences in types of foods consumed,
what comprises total carbohydrate as reported in food tables. are also, in part, due to methodology.

European Journal of Clinical Nutrition


Table 2 Energy and macronutrient intakes for 52 weighed records
analysed using Canadian and UK food tables
Energy Protein Fat CHO
(kcal) (g) (%) (g) (%) (g) (%)
Analysis using Canadian nutrient file
Mean of 52 records 2265 95.9 (16.9) 81.4 (31.5) 294.2 (52.9)
Analysis using ‘The composition of foods’
Mean of 52 records 1992* 89.8 (17.9) 74.5 (32.7) 252.4* (48.5)*
*Po0.001.
From Stephen, 2006.
Sugars
The term ‘sugars’ is conventionally used to describe the
mono- and disaccharides in food.
The three principal monosaccharides are glucose, fructose
and galactose, which are the building blocks of naturally
occurring di-, oligo- and polysaccharides. Free glucose and
fructose occur in honey and cooked or dried fruit (invert
sugar), in small amounts, and in larger amounts in fruit and
berries where they are the main energy source (Holland et al.,
1992). Corn syrup, a glucose syrup produced by the
hydrolysis of cornstarch, and high fructose corn syrup,
containing glucose and fructose, are increasingly used by the
food industry in many countries. Fructose is the sweetest of
all the food carbohydrates. Sugars are used as a sweetener to
improve the palatability of many foods and beverages, and
are also used for food preservation and in jams and jellies.
Sugars confer functional characteristics to foods, like visco-
sity, texture, body and browning capacity. They increase
dough yield in baked goods, influence starch and protein
breakdown, and control moisture thus preventing drying out
(Institute of Medicine, 2001).
The polyols, such as sorbitol are alcohols of glucose and
other sugars. They are found naturally in some fruits and are
made commercially by using aldose reductase to convert the
aldehyde group of the glucose molecule to the alcohol.
Sorbitol is used as a replacement for sucrose in the diet of
people with diabetes.
The principal disaccharides are sucrose (a-Glc(1-2)b-Fru)
and lactose (b-Gal(1-4)Glc). Sucrose is found very widely in
fruits, berries and vegetables, and can be extracted from
sugar cane or beet. Lactose is the main sugar in milk. Of the
less abundant disaccharides, maltose, derived from starch,
occurs in sprouted wheat and barley. Trehalose (a-Glc(1-4)
a-Glc) is found in yeast, fungi (mushrooms) and in small
amounts in bread and honey. It is used by the food industry
as a replacement for sucrose where less sweet taste is desired
but with similar technological properties.
Because of the perceived negative impact of sugars on
health, a number of terms have been used to categorize them
more specifically, mainly to highlight their origin and
identify them for labelling purposes, for example total
Carbohydrate terminology and classification
JH Cummings and AM Stephen
S7
sugars, added sugar, free sugars (WHO, 2003), refined sugars
(Nordic Council, 2004), discretionary sugar (New Zealand
Nutrition Foundation, 2004) and intrinsic sugars, milk
sugars and non-milk extrinsic sugars (Department of Health,
1989).
Total sugars
For labelling purposes, the category of total sugars has been
proposed. This includes all sugars from whatever source in a
food, and is defined as ‘all monosaccharides and disaccha-
rides other than polyols’ (European Communities, 1990).
This term is now accepted by the European Union, Australia
and New Zealand and may well be adopted by other
countries. It is probably the most useful way to describe,
measure and label sugars.
Free sugars
Traditionally ‘free sugars’ referred to any sugars in a food that
were free and not bound (Holland et al., 1992), and included
all mono- and disaccharides present in a food, including
lactose (Southgate, 1978). This term was also used analyti-
cally to describe when the carbohydrate in a food was
hydrolysed and components detected by chromatography or
colorimetric methods (Southgate, 1978). In recent years, the
use of the term ‘free sugars’ has changed, to refer to all
‘monosaccharides and disaccharides added to foods by the
manufacturer, cook and consumer, plus sugars naturally
present in honey, syrups and fruit juices’ and was the
preferred term for the WHO/FAO Expert Consultation on
‘Diet, Nutrition and the Prevention of Chronic Diseases’
(WHO, 2003). This new meaning of the term reflects the
same sources as those captured in the term ‘non-milk
extrinsic sugars’ outlined below. However, it is entirely
different from the traditional use of the term by the analyst,
which is a potential source of confusion.
Added sugars
In the United States, ‘added sugars’ is a commonly used term
and comprises sugars and syrups that are added to foods
during processing or preparation (Institute of Medicine,
2001). In the new United States Department of Agriculture
food composition tables, added sugars are defined as those
sugars added to foods and beverages during processing or
home preparation (Pehrsson et al., 2005). This would include
sugars listed in the ingredient list on a food product,
including honey, molasses, fruit juice concentrate, brown
sugar, corn sweetener, sucrose, lactose, glucose, high-fructose
corn syrup and malt syrup.
Extrinsic and intrinsic sugars
These terms had their origin in a United Kingdom (UK)
Department of Health Committee report in 1989 (Department
European Journal of Clinical Nutrition
 Carbohydrate terminology and classification
JH Cummings and AM Stephen
S8
of Health, 1989), which examined the role of sugars in the
diet. The terms were developed ‘to distinguish sugar as
naturally integrated into the cellular structure of a food
(intrinsic) from those that are free in the food or added to it
(extrinsic)’. These were defined in the report as:
Intrinsic sugars. Sugars forming an integral part of certain
unprocessed foodstuffs, that is enclosed in the cell, the most
important being whole fruits and vegetables (containing
mainly fructose, glucose and sucrose). Intrinsic sugars are
therefore naturally occurring and are always accompanied by
other nutrients.
Extrinsic sugars. Sugars not located within the cellular
structure of a food. Extrinsic sugars are mainly found in fruit
juice and are those added to processed foods. Lactose in milk
is extrinsic in that it is not found within the cellular
structure of food and has important nutritional benefits, so
the term non-milk extrinsic sugars was introduced to
indicate the group of sugars, other than intrinsic and milk
sugars, that should be restricted in the diet.
Non-milk extrinsic sugars. All extrinsic sugars, which are
not from milk, that is excluding lactose. This includes fruit
juices and honey and those sugars added to foods as a
sweetener in cooking or at the table, as in hot drinks and
breakfast cereal, or during processing. This terminology has
remained popular among nutritionists in the UK, and is used
in dietary surveys and other reports where intakes are
described (Gibson, 2000; Kelly et al., 2005). However, it is
not well understood by the public and is not used in public
communications about sugars.
Dividing sugars into intrinsic and extrinsic creates pro-
blems for the analyst and, therefore, for food labelling.
While ingredient lists can be used to identify the source of
sugars in foods, analytically it is not readily possible to
distinguish their origin in a processed food.
Other terms in use include ‘sugars’, ‘sugar’, ‘discretionary
sugar’, ‘refined sugars’, ‘refined sugar’, ‘natural sugar’ and
‘total available sugars’ (Stephen and Thane, 2007). Some of
these appear to equate to sucrose only, and within the EU
‘sucrose’ may be designated as ‘sugars’ on food labels
(European Communities, 2000). Many of the terms are used
in publications about intakes, often with little reference to
what components they include. This has the result of
making intake comparisons very difficult and points to the
need for a uniform terminology. There is little justification
for most of these terms apart from total sugars and their
subdivision into mono- and disaccharides. The relation of
sugars to health is determined more by the food matrix
in which they are contained and more thought should be
given to characterizing this because it also affects the other
nutrients in the food, and these many alternative terms do
not really describe a property of sugars per se.
Oligosaccharides, short-chain carbohydrates
‘Oligosaccharides are compounds in which monosaccharide
units are joined by glycosidic linkages’. Their DP has been
European Journal of Clinical Nutrition
variously defined as including anything from 2 to 19
monosaccharide units (http://www.britannica.com/eb/
article-9057022/oligosaccharide; British Nutrition Foundation,
1990; Food and Drug Administration, 1993). However, the
disaccharides (DP2) are thought of as sugars by nutritionists
(Roberfroid et al., 1993; Asp, 1995; Cummings and Englyst,
1995; Southgate, 1995), although a disaccharide composed
of two fructose residues, for example inulobiose, is consid-
ered a fructan (Roberfroid, 2005).
The dividing line between oligo- and polysaccharides is
also arbitrary since there is a continuum of molecular size
from simple sugars to complex polymers of DP 100 000 or
more in food. Most authorities recommend a DP of 10 as
the dividing point between oligo- and polysaccharides (IUB–
IUPAC and Joint Commission on Biochemical Nomencla-
ture, 1982), although in the most recent International Union
of Pure and Applied Chemistry–International Union of
Biochemistry Nomenclature Recommendation the issue is
not really addressed and a polysaccharide is just considered
to be ‘a macromolecule consisting of a large number of
monosaccharide (glycose) residues joined to each other by
glycosidic linkages’ (IUPAC–IUB Joint Commission on
Biochemical Nomenclature, 1996).
In practice, precipitation from aqueous solutions with
80%v/v ethanol is the step used in many carbohydrate
analysis procedures to separate these two groups (Southgate,
1991; Prosky et al., 1992; Englyst et al., 1994). However, some
branched-chain carbohydrates of DP between 10 and 100
remain in solution in 80% v/v ethanol so there is no clear
and absolute division. Furthermore, carbohydrates such as
inulin and polydextrose contain mixtures of polymers of
different chain lengths that cross the oligosaccharide/poly-
saccharide boundary. In categorizing oligosaccharides found
normally in the diet, alcohol precipitation would seem to be
the most practical way of delineating them from polysac-
charides. For novel oligosaccharides, such as that are now
being developed by the food industry as ingredients, the
average DP for that particular substance, as determined by
the manufacturer, should provide the basis on which to put
it into the appropriate carbohydrate class. In the light of the
lack of clarity surrounding the definition of oligosaccharides,
the Paris carbohydrate group suggested calling this group
‘short-chain carbohydrates’ (Cummings et al., 1997).
Food oligosaccharides fall into two groups: (i) maltodex-
trins, which are mostly derived from starch and include
maltotriose and a-limit dextrins that have both a1–4 and
a1–6 bonds and an average DP8. Maltodextrins are widely
used in the food industry as sweeteners, fat substitutes and to
modify the texture of food products. They are digested and
absorbed like other a-glucans and (ii) oligosaccharides that
are not a-glucans. These oligosaccharides include raffinose
(a-Gal(1-6)a-Glc(1-2)b-Fru), stachyose ((Gal)2 1:6 Glu 1:2
Fru) and verbascose ((Gal)3 1:6 Glu 1:2 Fru). They are in
effect, sucrose joined to varying numbers of galactose
molecules and are found in a variety of plant seeds, for
example peas, beans and lentils. Also important in this group

are inulin and fructo oligosaccharides (a-Glc(1-2)b-Fru(2-1)
b-Fru(N) or b-Fru(2-1)b-Fru(N)). They are fructans and are the
storage carbohydrates in artichokes and chicory with small
amounts of low molecular weight found in wheat, rye,
asparagus and members of the onion, leek and garlic family.
They can be produced industrially. The chemical bonds
linking these oligosaccharides are not a-1,4 or 1,6 glucans
and, therefore, they are not susceptible to pancreatic or
brush border enzyme breakdown (Oku et al., 1984; Hidaka
et al., 1986; Cummings et al., 2001). They have become
known as ‘non-digestible oligosaccharides’ (Roberfroid et al.,
1993). Some of them, mainly the fructans and galactans,
have unique properties in the gut and are known as
prebiotics (see later).
Milk oligosaccharides
Milk, especially human milk, contains oligosaccharides that
are predominantly galactose containing, although great
diversity of structure is found (Kunz et al., 2000). Almost
all carry lactose at their reducing end and are elongated by
addition of N-acetylglucosamine-linked b1–3 or b1–6 to a
galactose residue, followed by further galactose with b1–3 or
b1–4 bonds. Other monomers include L-fucose and sialic
acid. The principal oligosaccharide in milk is lacto-N-
tetraose. Total oligosaccharides in human milk are in the
range 5.0–8.0 g/l, but only trace amounts are present in cow’s
milk (Ward et al., 2006).
The oligosaccharides of breast milk have long been
credited with being the principal growth factor for bifido-
bacteria in the infant gut and thus primarily responsible for
these bacteria dominating the microbiota found in breast-fed
babies. In this context, milk oligosaccharides are acting as
prebiotics. Bifidobacteria can grow on milk oligosaccharides
as their sole carbon source while lactobacilli may not be able
to do so (Ward et al., 2006). The similarities between milk
oligosaccharide structure and epithelial cell surface carbohy-
drates in the gut suggest that milk oligosaccharides may act
as soluble receptors for gut pathogens and thus form an
essential part of colonization resistance. They may also be
immunomodulatory.
Starch and modified starch
Starch, the principal carbohydrate in most diets, is the
storage carbohydrate of plants such as cereals, root vegeta-
bles and legumes and consists of only glucose molecules. It
occurs in a partially crystalline form in granules and
comprises two polymers: amylose (DPB103) and amylopec-
tin (DPB104–105). Most common cereal starches contain
15–30% amylose, which is a non-branching helical chain of
glucose residues linked by a-1,4 glucosidic bonds. Amylo-
pectin is a high-molecular-weight, highly branched polymer
containing both a-1,4 and a-1,6 linkages. Some starches from
maize, rice, sorghum and barley contain largely amylopectin
and are known as ‘waxy’. The crystalline form of the amylose
Carbohydrate terminology and classification
JH Cummings and AM Stephen
S9
and amylopectin in the starch granules confers on them
distinct X-ray diffraction patterns, A, B and C. The A type is
characteristic of cereals (rice, wheat and maize), the B type of
potato, banana and high amylose starches while the C type is
intermediate between A and B and found in legumes. In their
native (raw) form, the B starches are resistant to digestion by
pancreatic amylase. The crystalline structure is lost when
starch is heated in water (gelatinization), thus permitting
digestion to take place. Recrystallization (retrogradation)
takes place to a variable extent after cooking and is in the B
form (Galliard, 1987; Hoover and Sosulski, 1991).
Modified starch
The proportions of amylose and amylopectin in a starchy
food are variable and can be altered by plant breeding.
Different cultivars of common species such as rice, have a
wide range of amylose to amylopectin ratios (Kennedy and
Burlingame, 2003). Techniques are rapidly emerging,
enabling starches to be produced for specific purposes by
genetically modifying the crop used for their production
(Regina et al., 2006). High amylose cornstarch and high
amylopectin (waxy) cornstarch have been available for some
time, and display quite different functional as well as
nutritional properties. High amylose starches require higher
temperatures for gelatinization and are more prone to
retrograde and to form amylose–lipid complexes. Such
properties can be utilized in the formation of foods with
high-resistant starch (RS) content.
Starches can also be modified chemically to impart func-
tional properties needed to produce certain qualities in
foodstuffs such as a decrease in viscosity and to improve gel
stability, mouth feel, appearance and texture, and resistance to
heat treatment. Various processes are used to modify starch,
the two most important being substitution and crosslinking.
Substitution involves etherification or esterification of a
relatively small number of hydroxyl groups on the glucose
units of amylose and amylopectin. This reduces retrograda-
tion, which is part of the process of staling of bread, for
example. Substitution also lowers gelatinization temperature,
provides freeze–thaw stability and increases viscosity. Cross-
linking involves the introduction of a limited number of
linkages between the chains of amylose and amylopectin. The
process reinforces hydrogen bonding, which occurs within the
granule. Crosslinking increases gelatinization temperature,
improves acid and heat stability, inhibits gel formation and
controls viscosity during processing. Altering the chemical
nature of starch can lead to it becoming resistant to digestion.
NSPs
NSPs are the non-a-glucan polysaccharides of the diet
(Table 1). They are essentially ‘macromolecules consisting
of a large number of monosaccharides (glycose) residues
joined to each other by glycosidic linkages’ (IUB-IUPAC and
Joint Commission on Biochemical Nomenclature, 1982) and
European Journal of Clinical Nutrition
 Carbohydrate terminology and classification
JH Cummings and AM Stephen
S10
are principally found in the plant cell wall. The term NSP was
first proposed at a meeting sponsored by the European
Economic Community Committee on Medical Research in
Cambridge in December 1978. The meeting was convened to
discuss the results of the analysis of nine foods for ‘dietary
fibre’ by a number of different methods in use in laboratories
around the world. The proposal was made because ‘An
accurate chemical identification of polysaccharides in the
diet is the first priorityy’ (James and Theander, 1981). At
the meeting, the first NSP values, measured as the sum of
constituent sugars, were presented for a selection of the
test samples (James and Theander, 1981). NSPs are the most
diverse of all the carbohydrate groups and comprise a
mixture of many molecular forms, of which cellulose, a
straight chain b1–4-linked glucan (DP 103–106) is the most
widely distributed. Because of its linear, unbranched nature,
cellulose molecules are able to pack closely together in a
three-dimensional latticework forming microfibrils. These
form the basis of cellulose fibres, which are woven into the
plant cell wall and give it structure. Cellulose comprises
between 10 and 30% of the NSP in foods (Holland et al.,
1988, 1991a, 1992).
In contrast, the hemicelluloses are a large group of
polysaccharide hetero polymers, which contain a mixture
of hexose (6C) and pentose (5C) sugars, often in highly
branched chains. Mostly, they comprise a backbone of xylose
sugars with branches of arabinose, mannose, galactose and
glucose and have a DP of 150–200. Typical of the hemi-
celluloses are the arabinoxylans found in cereals. About half
of hemicelluloses contain uronic acids, which are carboxy-
lated derivatives of glucose and galactose. They are impor-
tant in determining the properties of hemicelluloses,
behaving as carboxylic acids and are able to form salts with
metal ions such as calcium and zinc.
Common to all cell walls is, pectin, which is primarily a
1–4b-D galacturonic acid polymer, although 10–25% other
sugars such as rhamnose, galactose and arabinose, may also
be present as side chains. Between 3 and 11% of the uronic
acids have methyl substitutions, which improve the gel-
forming properties of pectin, as used in jam making. Some
residues are acetylated. Calcium and magnesium complexes
with uronic acids are characteristic of pectins.
Chemically related to the cell wall NSP, but not strictly cell
wall components, are the plant gums and mucilages. Plant
gums are sticky exudates that form at the sites of injuries to
plants. Many are highly branched complex uronic acid
containing polymers, such as Gum Arabic, named after the
Arabian port from which it was originally exported to Europe.
It comes from the Acacia tree and is one of the better known
plant gums, being sold commercially as an adhesive and used
in the food industry as a thickener and to retard sugar
crystallization. Other plant gums include karaya (sterculia),
guar, locust bean gum, xanthan and tragacanth, all of which
are licensed food additives (Saltmarsh, 2000).
Plant mucilages are botanically distinct in that they are
usually mixed with the endosperm of storage carbohydrates
European Journal of Clinical Nutrition
of seeds. Their role is to retain water and prevent desiccation.
They are neutral polysaccharides like the hemicelluloses, of
which guar gum, from the cluster bean (Cyamopsis tetra-
gonolobus), and carob gums are similar 1–4b-D galactoman-
nans with 1–6a-galactose single-unit side chains. Again, they
are widely used in the food and pharmaceutical industries as
thickeners and stabilizers in mayonnaise, soups and tooth-
pastes.
The algal polysaccharides, which include carageenan, agar
and alginate, are all NSP extracted from seaweeds or algae.
They replace cellulose in the cell wall and have gel-forming
properties. Carageenan and agar, are highly sulphated and
the ability of carageenan to react with milk protein has led to
its use in dairy products and chocolate.
Up-to-date values for the NSP content of foods are
published (Food Standards Agency/Institute of Food
Research, 2002).
Terminology based on physiology
In classifying dietary carbohydrate by its chemistry, the
principal challenge is to reconcile the various chemical
divisions with those that reflect physiology and health. A
classification based purely on chemistry does not allow a
simple translation into nutritional benefits since each of
the major chemical classes of carbohydrate has a variety of
overlapping physiological effects (Table 3). Terminology
based on physiological properties helps to focus on the
potential health benefits of carbohydrate, and identify foods
that are likely to be part of a healthy diet. But, as can be seen
from Table 4, each physiological or health benefit of
carbohydrate is attributable to several subgroups from the
main classification (Table 1). Moreover, this approach is
always open to the possibility of extensive revision, as new
physiological properties of dietary carbohydrates become
known. For example, the concept of prebiosis has added a
new dimension to understanding of how carbohydrates
behave in both the small and large intestines.
The physiology of carbohydrate can vary among indivi-
duals and populations. The classic example is lactose, which
is poorly hydrolysed by the small bowel mucosa of all adults
except Caucasians, most of whom retain the ability to digest
lactose into adult life. Additionally, within any simple
chemical group of carbohydrate, for example polyols, wide
variation in absorption may occur, ranging from almost
complete absorption of erythritol, to complete lack of
absorption of lactitol (Livesey, 2003). Similarly, starch may
have a variety of fates in the gut depending on granule
structure, whether raw or cooked and subsequent processing,
for example freezing (Stephen et al., 1983; Englyst et al.,
1992; Silvester et al., 1995). Furthermore, terminology based
on physiological properties alone provides the analyst with
an impossible target.
This dichotomy has led to the introduction of a number of
terms to describe various fractions and subfractions of
 Carbohydrate terminology and classification
JH Cummings and AM Stephen
S11
Table 3 Principal physiological properties of dietary carbohydrates

Provide Increase Glycaemica Cholesterol Increase calcium Source of Alter balance Increase stool Immunomodulatory
energy satiety lowering absorption SCFAb of microflora output
(prebiotic)

Monosaccharides | |
Disaccharides | | |
Polyols | |c |
Maltodextrins | |
Oligosaccharides | | | | |
(non-a-glucan)
Starch | | |d |d
NSP | | |e | |
a
Provides carbohydrate for metabolism (FAO, 1998).
b
Short chain fatty acids.
c
Except erythritol.
d
Resistant starch.
e
Some forms of non-starch polysaccharide (NSP) only.

Table 4 Physiological/health groupings of dietary carbohydrate Table 5 Preferred terminology of dietary carbohydrates

Glycaemica Glucose, fructose, galactose, sucrose, lactose, Chemical Physiological/botanical

maltose, trehalose, maltodextrins, starch
Non-glycaemic Polyols, oligosaccharides (non-a-glucan), resistant Useful Monosaccharides Prebiotic
and modified starches, NSP Disaccharides Resistant starch
Increase stool Polyols (except erythritol), some starches, NSP, Polyols Dietary fibrea
output lactose (in some populations), fructose (if taken in Total sugars Glycaemic
large amounts) Short-chain
No effect on stool Glucose, galactose, sucrose, maltose, trehalose, carbohydrates
weight maltodextrins, oligosaccharides, most starches Oligosaccharides
Polysaccharides
NSP, non-starch polysaccharide. Starch
a
Defined as in Table 2. Non-starch
polysaccharides
Total carbohydrate
carbohydrate and these are listed in Table 5. However, the Less Sugars Non-digestible oligosaccharides
useful Sugar Soluble and insoluble fibre
problems are by no means insurmountable in reconciling
Free sugars Available and unavailable
these different objectives for classification. With a sound Refined sugars carbohydrate
chemical identification for all the carbohydrates, it is then Added sugars Complex carbohydrate
possible to group them according to their health and Extrinsic and intrinsic
sugars
physiological effects.
a
Intrinsic plant cell wall polysaccharides.

Prebiotics
‘A prebiotic is a non-digestible food ingredient that benefi-
cially affects the host by selectively stimulating the growth
and/or activity of one of a limited number of bacteria in the
colon, and thus improves host health’ (Gibson and Roberfroid,
1995; Gibson et al., 2004).
As a group, prebiotics are thus defined by a single
physiological parameter, although this is by no means itself
clearly established (Macfarlane et al., 2006). Analytically
they cross the boundaries between disaccharides and
polysaccharides (DPX10) (van Loo et al., 1995). It is likely
in the future that a wider spectrum from the point of DP and
molecular form of carbohydrates will be shown to be
prebiotic. Prebiotic carbohydrates have unexpected proper-
ties in the gut in that they alter the balance of the gut
microflora towards what is considered to be a more healthy
one (Macfarlane et al., 2006). They have been shown to
increase calcium absorption and bone mineral density in
adolescents. (Elia and Cummings, 2007; prebiotics are dealt
with in more detail in the accompanying paper on
Physiology).
Resistant starch
RS is the sum of starch and products of starch digestion (such
as maltose, maltotriose and a-limit dextrins) that are not
absorbed in the small bowel (Englyst et al., 1992; Champ
et al., 2003). All unmodified starch, if solubilized, can be
hydrolysed by pancreatic a-amylase. However, the rate and
extent to which starch is broken down is altered by a number
of physical and chemical properties. This has led to a
classification of RS that is now widely used (Englyst and
Cummings, 1987).
RS can be fractionated into four types:
 RS I: physically inaccessible starch mostly present in
whole grains

European Journal of Clinical Nutrition

 Carbohydrate terminology and classification
JH Cummings and AM Stephen
S12
Table 6 Some currently proposed definitions/descriptions of dietary fibre

Dietary fibre consists of nondigestible carbohydrates and lignin that are intrinsic and intact in plants
Functional fibre consists of isolated, nondigestible carbohydrates that have beneficial physiological effects in humans
Total fibre is the sum of Dietary fibre and Added fibre (Institute of Medicine, 2001).
‘Dietary fibre means carbohydrate polymers with a degree of polymerization (DP) not lower than 3 which are neither digested nor absorbed in the small
intestine. A degree of polymerization not lower than 3 is intended to exclude mono- and disaccharides. It is not intended to reflect the average DP of a
mixture. Dietary fibre consists of one or more of:
K Edible carbohydrate polymers naturally occurring in the food as consumed;
K carbohydrate polymers, which have been obtained from food raw material by physical, enzymatic or chemical means,
K synthetic carbohydrate polymers.’ http://www.ccfnsdu.de/fileadmin/user_upload/PDF/ReportCCFNSDU2005.pdf
‘Dietary fibre should be defined to include all non-digestible carbohydrates (NDC). Lignin and other non-digestible but quantitatively minor components
that are associated with the dietary fibre polysaccharides and may influence their physiological properties should be included as well’ European Food
Standards Agency Draft Paper on Carbohydrates and Dietary Fibre (Becker and Asp, 2006).
‘Dietary fibres are the endogenous components of plant material in the diet which are resistant to digestion by enzymes produced by humans. They are
predominantly non-starch polysaccharides and lignin and may include, in addition, associated substances’ (Health and Welfare Canada, 1985).
‘Dietary fibre is the edible parts of plants or analogous carbohydrates that are resistant to digestion and absorption in the human small intestine with
complete or partial fermentation in the large intestine. Dietary fibre includes polysaccharides, oligosaccharides, lignin, and associated plant substances.
Dietary fibres promote beneficial physiological effects including laxation, and/or blood cholesterol attenuation, and/or blood glucose attenuation’
(American Association of Cereal Chemists, 2001).

 RS II: RS granules (Type B)
 RS III: retrograded starch (after food processing)
 RS IV: modified starches.
The observation that the rate and extent of starch
digestion can vary has been one of the most important
developments in our understanding of carbohydrates in the
past 30 years. There are implications of this for the glycaemic
response to foods, for fermentation in the large bowel and
for conditions such as diabetes and obesity. Methods have
been devised to measure these starch fractions in the
laboratory (Champ et al., 2003).
Dietary fibre
The term fibre, or dietary fibre, has many different meanings
in the nutrition world. It is not a precise reference to a
chemical component, or components, of the diet, but is
essentially a physiological concept as embodied in the
original definition by Trowell, ‘the proportion of food which
is derived from the cellular walls of plants which is digested
very poorly in human beings’ (Trowell, 1972).
The dietary fibre hypothesis was one of the most compel-
ling in nutrition and public health in the latter half of the
twentieth century. It provided the stimulus to a great deal of
research, for example epidemiological, physiological, analy-
tical and technical. It has been the catalyst for progress in
our understanding of the cause of a number of common
diseases, especially those of the large bowel and has given
governments and the food industry valuable targets for
healthy eating. However, in the 30 years since Trowell,
Walker, Burkitt and others first proposed the fibre hypo-
thesis, nutritional science has progressed, especially our
understanding of dietary carbohydrates and with this the
apparently unique role of fibre, essentially the plant cell wall,
in many physiological processes and in disease prevention
(Cummings et al., 2004).
Were it not for the perceived public perception that fibre is
good for you, and therefore the need to provide a value for
fibre on food labels, the term would be best consigned to the
history books. However, various national and international
bodies continue to struggle with the definition and the latest
versions of their deliberations on fibre are given in Table 6.
Common to these definitions is the concept of non-
digestibility in the small intestine.
Non-digestibility needs to be defined. If it is carbohydrate
that passes across the ileo-caecal valve, then to define it
requires complex physiological studies in humans. More-
over, it will vary widely from person to person (Stephen et al.,
1983; Englyst et al., 1992; Silvester et al., 1995; Molis et al.,
1996; Ellegard et al., 1997; Langkilde et al., 2002) and be
affected by the cooking of food, storage, chewing, ripeness
and the presence of other foods (Englyst and Cummings,
1986; Champ et al., 2003). It will include many dietary
components, for example lactose in some populations, some
polyols, some starches (RS) and NSP. There is no enforceable
method that can be used to measure this physiological
fraction of the diet.
As can be seen in the accompanying paper on the
physiology of carbohydrates (Elia and Cummings, 2007),
digestibility has an entirely different context when it is used
in the description of energy metabolism. Here it is defined as
‘the proportion of combustible energy that is absorbed over
the entire length of the gastrointestinal tract’. It would be
useful to have these different concepts of digestion aligned.
If there is a desire to use the word ‘fibre’, then it should
always be ‘qualified by a statement itemizing those carbohy-
drates and other substances intended for inclusion’, by
which is meant carbohydrates identified in the chemical
classification table (Table 1).
At a meeting of the authors of the scientific update papers,
and other experts, convened by WHO/FAO and held in
Geneva on 17–18 July 2006, the definition of dietary fibre
was discussed, including the one suggested by the US

European Journal of Clinical Nutrition


National Academy of Sciences in 2001 and that currently
proposed by Codex (http://www.ccnfsdu.de/fileadmin/user_
upload/PDF/ReportCCNFSDU2005.pdf) (Table 6).
The experts agreed that the definition of dietary fibre
should be more clearly linked to health and, after discussion,
the following definition was proposed.
‘Dietary fibre consists of intrinsic plant cell wall poly-
saccharides’.
The established epidemiological support for the health
benefits of dietary fibre is based on diets that contain fruits,
vegetables and whole-grain foods, for which the intrinsic
plant cell wall polysaccharides are a good marker. Although
isolated or extracted fibre preparations have been shown to
have physiological effects experimentally, these cannot be
translated into health benefits directly because the epide-
miological evidence points to fruits, vegetables and whole-
grain foods as beneficial, and in a normal diet, these
polysaccharides are part of the plant cell wall complex and
do not exist individually.
Soluble and insoluble dietary fibre
These terms arose out of the early chemistry of NSPs, which
showed that the fractional extraction of NSP could be
controlled by changing the pH of solutions. They proved
very useful in the initial understanding of the properties of
dietary fibre, allowing a simple division into those which
principally had effects on glucose and lipid absorption from
the small intestine (soluble) and those which were slowly
and incompletely fermented and had more pronounced
effects on bowel habit (insoluble). However, the separation of
soluble and insoluble fractions is very pH dependent, making
the link with specific physiological properties less certain.
Much insoluble fibre is completely fermented and not all
soluble fibre has effects on glucose and lipid absorption. Many
of the early studies were done with isolated gums or extracts
of cell walls, whereas these various forms of fibre exist
together mostly in intact cell walls of plants.
Nevertheless, certain fibre-rich foods effect glycaemic
control and lipid levels and have been widely used in the
management of diabetes particularly the legumes and pulses
rather than high bran products (Kiehm et al., 1976; Simpson
et al., 1979, 1981; Rivellese et al., 1980; Mann, 1984;
Chandalia et al., 2000; Giacco et al., 2000; Mann et al.,
2004). Work on the variability in glycaemic response of
different types of foods has led to the concept of the
glycaemic index (Crapo et al., 1977; Jenkins et al., 1981) and
a new area of nutritional science has been developed,
concerned with glycaemic responses to carbohydrate-
containing foods (Foster-Powell et al., 2002).
Available and unavailable carbohydrate
A major step forward conceptually in our understanding of
carbohydrates was made by McCance and Lawrence (1929)
Carbohydrate terminology and classification
JH Cummings and AM Stephen
S13
with the division of dietary carbohydrate into available and
unavailable. In an attempt to prepare food tables for diabetic
diets, they realized that not all carbohydrates could be
‘utilized and metabolized’, that is provide the body with
‘carbohydrates for metabolism’. Available carbohydrate was
defined as ‘starch and soluble sugars’ and unavailable as
‘mainly hemicellulose and fibre (cellulose)’. This concept
proved useful, not the least because it drew attention to the
fact that some carbohydrate is not digested and absorbed in
the small intestine but rather reaches the large bowel where
it is fermented or even excreted in faeces.
An FAO technical workshop in Rome in 2002 on ‘Food
energy—methods of analysis and conversion factors’ (FAO,
2003) defined available carbohydrate as ‘that fraction of
carbohydrate that can be digested by human enzymes, is
absorbed and enters into intermediary metabolism’. (It does
not include dietary fibre, which can be a source of energy
only after fermentation).
It is somewhat misleading to talk of carbohydrate as
‘unavailable’ because carbohydrate that reaches the colon is
able to provide the body with energy through fermentation
and absorption of short-chain fatty acids. There are many
properties of carbohydrate of which site of digestion is only
one. An alternative to the terms ‘available’ and ‘unavailable’
today would be to describe carbohydrates either as glycaemic
(that is providing carbohydrate for metabolism) or non-
glycaemic, which is closer to the original concept of
McCance and Lawrence. However, the FAO Technical work-
shop on Food Energy in its recommendations said, ‘Available
carbohydrate is a useful concept in energy evaluation and
should be retained. This recommendation is at odds with the
view of the expert consultation in 1997 on carbohydrates,
which endorsed the use of the term ‘glycaemic carbohydrate’
to mean ‘providing carbohydrate for metabolism’. The group
expressed concerns that ‘glycaemic carbohydrate’ might be
confused or even equated with the concept of ‘glycaemic
index’, which is an index that describes the relative blood
glucose response to different ‘available carbohydrates’. The
term ‘available’ seems to convey adequately the concept of
‘providing carbohydrate for metabolism’, while avoiding this
confusion’. Furthermore, in the discussion of the energy
value of carbohydrate, the terms ‘available’ and ‘unavailable’
are extensively used (Elia and Cummings, 2007). On balance,
however, we would recommend ‘glycaemic’ as a more precise
and measurable fraction.
Glycaemic carbohydrate
A more recently developed distinction with regard to human
health, although arising out of the original McCance and
Lawrence concept, is whether or not the carbohydrate source
does or does not directly provide carbohydrate as an
energy source following the process of digestion and
absorption in the small intestine. Carbohydrate, which
provides glucose for metabolism is referred to as ‘glycaemic
European Journal of Clinical Nutrition
 Carbohydrate terminology and classification
JH Cummings and AM Stephen
S14
carbohydrate’, whereas carbohydrates that pass to the
large intestine prior to being metabolized, is referred to
as ‘non-glycaemic carbohydrate’. Most mono- and di-
saccharides, some oligosaccharides (maltodextrins) and
rapidly digested starches may be classified as glycaemic
carbohydrate. Slowly digested starches are also considered
to be glycaemic carbohydrate though glucose is less
rapidly generated. The remaining oligosaccharides, NSPs
and RS are considered to be non-glycaemic carbohydrates.
Most carbohydrate-containing unprocessed foods contain
both glycaemic and non-glycaemic carbohydrate. The extent
to which carbohydrate in foods raises blood glucose
concentration compared with an equivalent amount of
reference carbohydrate has also been used as a means of
classifying dietary carbohydrate and is known as the
glycaemic index (Venn and Green, 2007). Many factors
affect the glycaemic response to carbohydrate, including the
intrinsic properties of the food and also extrinsic factors such
as the composition of the meal, the overall diet and
biological variations of the host.
Complex carbohydrates
This term was first used in the McGovern report, ‘Dietary
Goals for the United States’ in 1977 (Select Committee on
Nutrition and Human Needs, 1977). It was coined largely to
distinguish sugars from other carbohydrates and in the
report denotes ‘fruit, vegetables and whole-grains’. The term
has never been formally defined and has since come to be
used to describe either starch alone, or the combination of all
polysaccharides (British Nutrition Foundation, 1990). It was
used to encourage consumption of what are considered to be
healthy foods such as whole-grain cereals, etc., but becomes
meaningless when used to describe fruits and vegetables,
which are low in starch. As a substitute term for starch it
would seem to have little merit and, in principle, it is better
to discuss carbohydrate components by using their common
chemical names.
Physical effects of carbohydrates
Aside from their chemistry, the physiological properties of
carbohydrates are also affected by the physical state of the
food. In this context, there may be a unique role for NSP in
the control of carbohydrate metabolism. The early studies of
viscosity on glucose responses pointed to the physical
properties of NSP as being important (Jenkins et al., 1978).
In a different context, the classic study of Haber et al. (1977)
with apples shows clearly that where carbohydrate, in this
case mainly glucose and fructose, is entrapped intracellularly
in plant foods, its release in the gut is slowed and blood
glucose moderated and insulin responses lowered. A unique
property of NSP, therefore, is that it forms the plant cell wall
and thus a physical structure to foods. Numerous studies
European Journal of Clinical Nutrition
have now shown that the physical structure of starchy foods
influences the glycaemic response.
The physical structure of a food has, therefore, a role to
play in the regulation of carbohydrate metabolism. The use
of viscous-soluble NSP isolates in this context will probably
be seen as a stepping stone to understanding fibre but not
the ultimate goal. From a nutritional and analytical view
points, the intrinsic polysaccharides of the plant cell wall,
known as dietary fibre or NSP, should be viewed as a single
entity that uniquely provides physical structure to foods.
There is however a major problem in characterizing and
measuring the key physical attributes of a food that
contributes to modifying its effects.
Whole grain
The essence of the dietary recommendations in many
countries is to eat a diet high in whole grains, fruits and
vegetables and this is embodied in the WHO/FAO report on
‘Diet, Nutrition and the Prevention of Chronic Diseases’
(WHO, 2003) in the description of population nutrient
intake goals. However, the term ‘whole grain’ has several
meanings from ‘whole of the grain’ through to physically
intact structures. A precise definition is clearly needed for
labelling purposes.
Whole grain comprises whole wheat, whole-wheat flour,
wheat flakes, bulgar wheat, whole and rolled oats, oatmeal,
oat flakes, brown rice, whole rye and rye flour, whole barley
and popcorn. Cornmeal is not included as it is generally
dehulled, de-branned and de-germed. Sweet corn has been
included in some analyses (Harnack et al., 2003), but is not
included in analyses from the UK, where it is considered a
vegetable (Henderson et al., 2002; Thane et al., 2005, 2007).
Foods containing added bran, but not including endosperm
or germ, are included in some analyses (Jacobs et al., 2001)
but not in others, but are not whole grains and should be
excluded. Similarly, foods which are largely whole grain, but
not entirely, such as puffed wheat, where the puffing and
toasting causes some of the outer layers to drop off are not
truly whole grain. Such foods are difficult to consider, since
they may be consumed by a considerable proportion of the
population and are indicated as whole wheat on the label of
some products but not others.
There are also problems with the definition of a whole
grain food. For labelling purposes, the Food and Drug
Administration (1999) in the United States has defined a
whole grain food as ‘a product containing 451% whole
grain by weight per reference amount customarily consumed
per day’ (Seal, 2006), and this standard has been used in
some surveys to assess intake (Lang et al., 2003; Jensen et al.,
2004). However, many foods contain less whole grain than
this, but can still make a substantial contribution to total
whole-grain intake (Thane et al., 2005, 2007).
In studies by Jacobs et al. (1998) in the United States,
breakfast cereals were ‘considered to be whole grain if the

product contained X25% whole grain or bran y’. In a
more complete analysis of the United Kingdom, foods with
whole-grain contents of 10% or more were used, rather than
the cutoffs of 25 or 51%. In the UK National Diet and
Nutrition Surveys, it was found that for young people aged
4–18 years, intake of whole grains was underestimated by
28% if only those products with 451% whole grains were
included and underestimated by 15% if only those with
whole-grain content greater than 25% were included. More
recently for adults, the 51% cutoff would have under-
estimated intake by 18% for the 1986–87 Dietary Survey of
British Adults (Gregory et al., 1990), and 27% in the NDNS
survey of adults in 2000–01 (Henderson et al., 2002),
indicating not only the underestimation but that the
manner of consuming whole grains may also be changing,
with more foods with lower amounts now being consumed.
It is a considerable amount of work to determine whole-grain
content down to the 10% level, but this would include foods
like porridge, which when cooked is 90% water but is
consumed in substantial quantities in parts of the world
(Thane et al., 2007), and which may have important health
benefits.
In his editorial comment on the 1998 Jacobs’ paper,
Willett (1998) remarks ‘The physical form of whole grains
can vary from intact kernels (for which we should probably
reserve the term whole grain) to finely milled flour (whole
grain flour)’. This distinction between intact whole grains or
physically disrupted, although present in the food in its
entirety, is important. Grain structure affects the glycaemic
response to food (Foster-Powell et al., 2002) and high intakes
of whole-grain foods protect against the development of type
II diabetes (Venn and Mann, 2004) and cardiovascular
disease (Seal, 2006), yet we know little about the physical
form of grains in these studies.
It is also worth noting that the type of grain contributing
to whole grains may vary from country to country. For the
UK, about 90% of the whole-grain intake is wheat (Thane
et al., 2005), while in the United States, a larger proportion is
likely to be from oats, given the popularity of whole-grain
oat cereals. With the different physical and physiological
properties of these two grains, such differences need to be
taken into account when interpreting health impacts of
whole-grain consumption.
Conclusions
(1) Dietary carbohydrates should be classified according to
their chemical form, as recommended at the 1997 FAO/
WHO Expert Consultation.
(2) The physiological and health effects of carbohydrates are
dependent not only on their primary chemical form but
also on their physical properties, which include water
solubility, gel formation, crystallization state, association
with other molecules and aggregation into the complex
structures of the plant cell wall.
Carbohydrate terminology and classification
JH Cummings and AM Stephen
S15
(3) Total carbohydrate in food should be determined by
direct measurement rather than ‘by difference’.
(4) Many terms exist to describe sugars in the diet. The most
useful are total sugars and their division into mono- and
disaccharides. The use of other terms creates difficulties
for the analyst, confusion for the consumer and suggests
properties of foods that are not related to sugars
themselves, but to the food matrix.
(5) Because neither chemical nor physical description of
carbohydrates directly reflects their physiological pro-
perties and health benefits, a number of terms to
describe carbohydrates, based on their physiology, have
been created. Of these prebiotic, glycaemic, RS and
dietary fibre are useful.
(6) Dietary fibre should be defined to reflect the health
benefits of a diet rich in fruits, vegetables and whole
grains and not the variable physiological properties or
health effects of the various carbohydrate types. The
definition proposed by the group was ‘intrinsic plant cell
wall polysaccharides’.
(7) The effects of foods containing different types of fibre on
glycaemic control and lipid levels should be investigated
further to determine the exact properties needed for
their effects. The distinction between soluble and
insoluble forms of fibre is inappropriate since the
separation is pH dependent and does not reflect the
physiological properties of whole foods in the gut.
(8) The term whole grains should be defined more clearly
and the role of intact versus milled grains established.
The whole-grain concept, along with fresh fruits and
vegetables is central to a healthy diet message.
Acknowledgements
The authors thank Professor Ingvar Bosaeus, Dr Barbara
Burlingame, Professor Jim Mann, Professor Timothy Key,
Professor Carolyn Summerbell, Dr Bernard Venn and Dr
Martin Wiseman for the valuable comments they provided
on the earlier manuscript.
Conflict of interest
During the preparation and peer review of this paper in 2006,
the authors and peer reviewers declared the following
interests.
Authors
Professor John H Cummings: Chairman, Biotherapeutics
Committee, Danone; Member, Working Group on Foods
with Health Benefits, Danone; funding for research work at
the University of Dundee, ORAFTI (2004).
Dr Alison M Stephen: Contract with World Sugar Research
Organization on trends in intakes of sugars and sources in
the diet (contract is with MRC-Human Nutrition Resource);
contracts with Cereal Partners UK on whole-grain intakes in
the UK and relationship to adiposity (contract was with
European Journal of Clinical Nutrition
 Carbohydrate terminology and classification
JH Cummings and AM Stephen
S16
MRC-Human Nutrition Resource); Adviser to Audrey Eyton
on scientific content on book ‘F2 Diet’; Scientific Advisory
Panel of Canadian Sugar Institute (not for profit but funded
by sugar industry) (1995–2002).
Peer-reviewers
Professor Ingvar Bosaeus: none declared.
Dr Barbara Burlingame: none declared.
Professor Jim Mann: none declared.
Professor Timothy Key: none declared.
Professor Carolyn Summerbell: none declared.
Dr Bernard Venn: none declared.
Dr Martin Wiseman: none declared.
References
American Association of Cereal Chemists (2001). The definition of
dietary fibre. Cereal Foods World 46, 112–126.
Asp N-G (1995). Classification and methodology of food carbohy-
drates as related to nutritional effects. Am J Clin Nutr 61 (Suppl 4),
S980–S987.
Atwater WO, Woods CD (1986). The Chemical Composition of
American Food Materials. Experiment Station Bulletin no. 28. US
Official Experiment Stations: Washington, DC.
Becker W, Asp N-G (2006). Carbohydrates and dietary fibre.
British Nutrition Foundation (1990). Complex Carbohydrates in Foods:
Report of the British Nutrition Foundation’s Task Force. Chapman and
Hall: London.
Champ M, Langkilde AM, Brouns F, Kettlitz B, Le Bail-Collet Y
(2003). Advances in dietary fibre characterization. 2. Consump-
tion, chemistry, physiology and measurement of resistant starch;
implications for health and food labelling. Nutr Res Rev 16,
143–161.
Chandalia M, Garg A, Lutjohann D, von Bergmann K, Grundy SM,
Brinkley LJ (2000). Beneficial effects of high dietary fiber intake in
patients with type 2 diabetes mellitus. N Engl J Med 342, 1392–
1398.
Crapo PA, Reaven G, Olefsky J (1977). Postprandial plasma-glucose
and insulin responses to different complex carbohydrates. Diabetes
26, 1178–1183.
Cummings JH, Englyst HN (1995). Gastrointestinal effects of food
carbohydrate. Am J Clin Nutr 61, 938S–945S.
Cummings JH, Edmond LM, Magee EA (2004). Dietary carbohydrates
and health: do we still need the fibre concept? Clin Nutr Suppl 1,
5–17.
Cummings JH, Macfarlane GT, Englyst HN (2001). Prebiotic diges-
tion and fermentation. Am J Clin Nutr 73, 415S–420S.
Cummings JH, Roberfroid MB, Andersson H, Barth C, Ferro-Luzzi A,
Ghoos Y et al. (1997). A new look at dietary carbohydrate:
chemistry, physiology and health. Eur J Clin Nutr 51, 417–423.
Department of Health (1989). Dietary Sugars and Human Disease. Her
Majesty’s Stationery Office: London.
Elia M, Cummings JH (2007). Physiological aspects of energy
metabolism and gastrointestinal effects of carbohydrates. Eur J
Clin Nutr 61 (Suppl 1), S40–S74.
Ellegard L, Andersson H, Bosaeus I (1997). Inulin and oligofructose
do not influence the absorption of cholesterol, or the excretion of
cholesterol, Ca, Mg, Zn, Fe, or bile acids but increases energy
excretion in ilesotomy subjects. Eur J Clin Nutr 51, 1–5.
Englyst HN, Cummings JH (1986). Digestion of the polysaccharides
of banana (Musa paradisiaca sapientum) in man. Am J Clin Nutr 44,
42–50.
Englyst HN, Cummings JH (1987). Resistant starch, a ‘new’ food
component: a classification of starch for nutritional purposes. In:
European Journal of Clinical Nutrition
Morton ID (ed). Cereals in a European Context. Ellis Horwood:
Chichester. pp 221–233.
Englyst HN, Kingman SM, Cummings JH (1992). Classification and
measurement of nutritionally important starch fractions. Eur J Clin
Nutr 46, S33–S50.
Englyst HN, Quigley ME, Hudson GJ (1994). Determination of
dietary fibre as non-starch polysaccharides with gas–liquid
chromatographic, high-performance liquid chromatographic or
spectrophotometric measurement of constituent sugars. Analyst
119, 1497–1509.
Englyst K, Liu S, Englyst H (2007). Nutritional characterisation and
measurement of dietary carbohydrates. Eur J Clin Nutr 61 (Suppl 1),
S19–S39.
European Communities (1990). Council directive 90/496/EEC on
nutrition labelling for foodstuffs.
European Communities (2000). Council directive 2000/13/EC on the
approximation of the laws of the member states relating to the
labelling, presentation and advertising of foodstuffs.
FAO (1998). Carbohydrates in human nutrition. Food and Agriculture
Organization of the United Nations: Rome. Report of a Joint FAO/
WHO Expert Consultation. FAO Food and Nutrition Paper no. 66.
FAO (2003). Food Energy—Methods of Analysis and Conversion Factors.
Food and Agriculture Organization of the United Nations: Rome.
Report of a Technical Workshop. FAO Food and Nutrition Paper
no. 77.
Food and Drug Administration (1993). Food labelling: mandatory
status of nutrition labelling and nutrient content revision, format
for nutrition label. Fed Regist 58, 2079–2228.
Food and Drug Administration (1999). Health claim notification for
whole grain foods. http://www.cfsan.fda.gov/Bdms/flgrains.html.
Food Standards Agency/Institute of Food Research (2002). The
Composition of Foods Sixth Summary Edition. Royal Society of
Chemistry: Cambridge.
Foster-Powell K, Holt SHA, Brand-Miller JC (2002). International
table of glycemic index and glycemic load values: 2002. Am J Clin
Nutr 76, 5–56.
Galliard TE (1987). Starch: Properties and Potential. John Wiley & Sons:
Chichester.
Giacco R, Parillo M, Rivellese AA, Lasorella G, Giacco A, D’Episcopo L
et al. (2000). Long-term dietary treatment with increased amounts
at fiber-rich low-glycemic index natural foods improves blood
glucose control and reduces the number of hypoglycemic events
in type 1 diabetic patients. Diabetes Care 23, 1461–1466.
Gibson GR, Roberfroid M (1995). Dietary modulation of the human
colonic microbiota: introducing the concept of prebiotics. J Nutr
125, 1401–1412.
Gibson GR, Probert HM, Van Loo J, Rastall RA, Roberfroid M (2004).
Dietary modulation of the human colonic microbiotia: updating
the concept of prebiotics. Nutr Res Rev 17, 259–275.
Gibson SA (2000). Breakfast cereal consumption in young children:
associations with non-milk extrinsic sugars and caries experience:
further analysis of data from the UK national diet and nutrition
survey of children aged 1.5–4.5 years. Public Health Nutr 3, 227–232.
Gregory J, Foster K, Tyler H, Wiseman M (1990). Dietary and
Nutritional Survey of British Adults. Her Majesty’s Stationery Office:
London.
Haber GB, Heaton KW, Murphy D, Burroughs LF (1977). Depletion
and disruption of dietary fibre: effects on satiety, plasma-glucose
and serum insulin. Lancet ii, 679–682.
Harnack L, Walters SA, Jacobs Jr DR (2003). Dietary intake and food
sources of whole grains among US children and adolescents: data
from the 1994–1996 continuing survey of food intakes by
individuals. J Am Diet Assoc 103, 1015–1019.
Health Canada (2005). Canadian nutrient file. http://www.hc-sc.
gc.ca/fn-an/nutrition/fiche-nutri-data/index_e.html.
Health and Welfare Canada (1985). Report of the Expert Advisory
Committee a Dietary Fibre.

Henderson L, Gregory J, Swan G (2002). The National Diet & Nutrition
Survey: Adults Aged 19 to 64 Years. Vol. 1: Types and Quantities of
Food Consumed. Her Majesty’s Stationery Office: London.
Hidaka H, Eida T, Takizawa T, Tokunaga T, Tashiro Y (1986). Effects of
fructooligosaccharides on intestinal flora and human health.
Bifidobacteria Microflora 5, 37–50.
Holland B, Unwin ID, Buss DH (1992). Fruit and Nuts. First Supplement
to 5th Edition of McCance and Widdowson’s the Composition of Foods.
Her Majesty’s Stationery Office: London.
Holland B, Unwin ID, Buss DH (eds) (1988). Cereals and Cereal
Products. Royal Society of Chemistry: Cambridge.
Holland B, Unwin ID, Buss DH (eds) (1991a). Vegetables, Herbs and
Spices. Royal Society of Chemistry: Cambridge.
Holland B, Welch AA, Unwin ID, Buss DH, Paul AA, Southgate DAT
(1991b). McCance & Widdowson’s the Composition of Foods. Her
Majesty’s Stationery Office: London.
Hoover R, Sosulski FW (1991). Composition, structure, functionality,
and chemical modification of legume starches: a review. Can J
Physiol Pharmacol 69, 79–92.
http://www.britannica.com/eb/article-9057022/oligosaccharide.
http://www.ccnfsdu.de/fileadmin/user_upload/PDF/ReportCCNFSDU
2005.pdf.
Institute of Medicine (2001). Dietary Reference Intakes for Energy,
Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein and Amino
Acids. National Academy Press: Washington, DC.
IUB–IUPAC Joint Commission on Biochemical Nomenclature (1982).
Abbreviated terminology of oligosaccharide chains. J Biol Chem
257, 3347–3351.
IUPAC–IUB Joint Commission on Biochemical Nomenclature (1996).
Nomenclature of carbohydrates. www.chem.qmul.ac.uk/iupac/
2carb/.
Jacobs Jr DR, Meyer HE, Solvoll K (2001). Reduced mortality among
whole grain bread eaters in men and women in the Norwegian
county study. Eur J Clin Nutr 55, 137–143.
Jacobs Jr DR, Meyer KA, Kushi LH, Folsom AR (1998). Whole-grain
intake may reduce the risk of ischemic heart disease death in
postmenopausal women: the Iowa women’s health study. Am J
Clin Nutr 68, 248–257.
James WPT, Theander O (eds) (1981). The Analysis of Dietary Fibre in
Food. Marcel Dekker: New York. p 259.
Jenkins DJ, Wolever TM, Leeds AR, Gassull MA, Haisman P, Dilawari J
et al. (1978). Dietary fibres, fibre analogs, and glucose-tolerance-
importance of viscosity. Br Med J 1, 1392–1394.
Jenkins DJ, Wolever TM, Taylor RH, Barker H, Fielden H, Baldwin JM
et al. (1981). Glycemic index of foods: a physiological basis for
carbohydrate exchange. Am J Clin Nutr 34, 362–366.
Jensen MK, Koh-Banerjee P, Hu FB, Franz M, Sampson L, Gr+nbæk M
et al. (2004). Intakes of whole grains, bran, and germ and the risk of
coronary heart disease in men. Am J Clin Nutr 80, 1492–1499.
Kelly SA, Summerbell C, Rugg-Gunn AJ, Adamson A, Fletcher E,
Moynihan PJ (2005). Comparison of methods to estimate non-
milk extrinsic sugars and their application to sugars in the diet of
young adolescents. Br J Nutr 94, 114–124.
Kennedy G, Burlingame B (2003). Analysis of food composition data
on rice from a plant genetic resources perspective. Food Chem 80,
589–596.
Kiehm TG, Anderson JW, Ward K (1976). Beneficial effects of a high
carbohydrate, high fiber diet on hyperglycemic diabetic men. Am J
Clin Nutr 29, 895–899.
Kunz C, Rudloff S, Baier W, Klein N, Strobel S (2000). Oligosaccha-
rides in human milk: structural, functional, and metabolic aspects.
Annu Rev Nutr 20, 699–722.
Lang R, Thane CW, Bolton-Smith C, Jebb SA (2003). Consumption of
whole-grain foods by British adults: findings from further analysis
of two national dietary surveys. Public Health Nutr 6, 479–484.
Langkilde AM, Champ M, Anderson H (2002). Effects of high-
resistant-starch banana flour (RS2) on in vitro fermentation and
the small-bowel excretion of energy, nutrients, and sterols: an
ileostomy study. Am J Clin Nutr 75, 104–111.
Carbohydrate terminology and classification
JH Cummings and AM Stephen
S17
Livesey G (2003). Health potential of polyols as sugar replacers, with
emphasis on low glycaemic properties. Nutr Res Rev 16, 163–191.
Macfarlane S, Macfarlane GT, Cummings JH (2006). Prebiotics in the
gastrointestinal tract. Aliment Pharmacol Ther 24, 701–714.
Mann J (1984). Lines to legumes: changing concepts of diabetic diets.
Diabet Med 1, 191–198.
Mann J, De Leeuw I, Hermansen K, Karamanos B, Karlström B,
Katsilambros N et al. (2004). Evidence-based nutritional ap-
proaches to the treatment and prevention of diabetes mellitus.
Nutr Metab Cardiovasc Dis 14, 373–394.
McCance RA, Lawrence RD (1929). The Carbohydrate Content of Foods.
Her Majesty’s Stationery Office: London.
Molis C, Flourie B, Ouarne F, Gailing MF, Lartigue S, Guibert A et al.
(1996). Digestion, excretion and energy value of fructooligosac-
charides in healthy humans. Am J Clin Nutr 64, 324–328.
National Academy of Sciences (2001). Dietary References Intakes.
Proposed Definition of Dietary Fiber. National Academy of Sciences:
Washington DC.
New Zealand Nutrition Foundation (2004). Sugar and our diet.
http://www.chelsea.co.nz/shared/documents/report.pdf.
Oku T, Tokunaga T, Hosoya N (1984). Non-digestibility of a new
sweetener ‘Neosugar’ in the rat. J Nutr 114, 1574–1581.
Pehrsson PR, Cutrufelli RL, Gebhardt SE, Lemar LE, Holcomb GT,
Haytowitz DB et al. (2005). USDA database for the added sugars
content of selected foods. www.ars.usda.gov/nutrientdata.
Prosky L, Asp N-G, Schweizer TF, De Vries JW, Furda I (1992).
Determination of insoluble and soluble dietary fiber in foods and
food products: collaborative study. J AOAC Int 75, 360–367.
Regina A, Bird A, Topping D, Bowden S, Freeman J, Barsby T et al.
(2006). High-amylose wheat generated by RNA interference
improves indices of large-bowel health in rats. Proc Natl Acad Sci
USA 103, 3546–3551.
Rivellese A, Riccardi G, Giacco A, Pacioni D, Genovese S, Mattioli PL
et al. (1980). Effect of dietary fibre on glucose control and serum-
lipoproteins in diabetic-patients. Lancet 2, 447–450.
Roberfroid M (2005). Inulin-type Fructans. Functional Food Ingredients.
CRC Press: Boca Raton.
Roberfroid M, Gibson GR, Delzenne N (1993). The biochemistry of
oligofructose, a non-digestible fiber: an approach to calculate its
caloric value. Nutr Rev 51, 137–146.
Saltmarsh ME (2000). Essential Guide to Food Additives. Leatherhead
Publishing: Leatherhead.
Seal CJ (2006). Whole grains and CVD risk. Proc Nutr Soc 65, 24–34.
Select Committee on Nutrition and Human Needs (1977). Dietary
Goals for the United States. US Government Printing Office:
Washington, DC.
Silvester KR, Englyst HN, Cummings JH (1995). Ileal recovery of
starch from whole diets containing resistant starch measured
in vitro and fermentation of ileal effluent. Am J Clin Nutr 62,
403–411.
Simpson HC, Simpson RW, Lousley S, Carter RD, Geekie M,
Hockaday TD et al. (1981). A high-carbohydrate leguminous fiber
diet improves all aspects of diabetic control. Lancet 1, 1–5.
Simpson RW, Mann JI, Eaton J, Carter RD, Hockaday TD (1979).
High-carbohydrate diets and insulin dependent diabetics. Br Med J
2, 523–525.
Southgate DAT (1978). Free sugars in foods. J Hum Nutr 32, 335–347.
Southgate DAT (1991). Determination of Food Carbohydrates. Elsevier
Science Publishers Ltd: Barking.
Southgate DAT (1995). Digestion and metabolism of sugars. Am J Clin
Nutr 62 (Suppl 1), S203–S211.
Stephen AM (2006). Impact of Carbohydrate Methodology on
Dietary Intake. Copenhagen, April 2006.
Stephen AM, Thane CT, Bramweh CF (2007). Sugars in the diet:
definitions, intakes and sources—an update. (in preparation).
Stephen AM, Haddad AC, Phillips SF (1983). Passage of carbohydrate
into the colon. Direct measurements in humans. Gastroenterology
85, 589–595.
European Journal of Clinical Nutrition
 Carbohydrate terminology and classification
JH Cummings and AM Stephen
S18
Thane CT, Jones AR, Stephen AM, Seal CJ, Jebb SA (2005). Whole-
grain intake of British young people aged 4–18 years. Br J Nutr 94,
825–831.
Thane CW, Jones AR, Stephen AM, Seal CJ, Jebb SA (2007).
Comparative whole-grain intake of British adults in 1986–7 and
2000–1. Br J Nutr 97, 987–992.
The Nordic Council (2004). Nordic Nutrition Recommendations
2004, Integrating Nutrition and Physical Activity. The Nordic
Council of Ministers, The Nordic Food Policy Co-Operation.
Copenhagen.
Trowell H (1972). Dietary fibre and coronary heart disease. Eur J Clin
Biol Res 17, 345–349.
United States Department of Agriculture (2007): http://www.ars.usda.
gov/nutrientdata.
van Loo J, Coussement P, de Leenheer L, Hoebregs H, Smits G
(1995). On the presence of inulin and oligofructose as
European Journal of Clinical Nutrition
View publication stats
natural ingredients in the western diet. Crit Rev Food Sci Nutr 35,
525–552.
Venn BJ, Green TJ (2007). Glycemic index and glycemic load:
measurement issues and their effect on diet–disease relationships.
Eur J Clin Nutr 61 (Suppl 1), S122–S131.
Venn BJ, Mann JI (2004). Cereals grains, legumes and diabetes. Eur J
Clin Nutr 58, 1443–1461.
Ward RE, Niñonuevo M, Mills DA, Lebrilla CB, German JB (2006). In
vitro fermentation of breast milk oligosaccharides by Bifidobacter-
ium infantis and Lactobacillus gasseri. Appl Environ Microbiol 72,
4497–4499.
WHO (2003). Diet, Nutrition and the Prevention of Chronic Diseases.
Report of a Joint WHO/FAO Expert Consultation. WHO Technical
Report Series no. 916. World Health Organization: Geneva.
Willett WC (1998). The dietary pyramid: does the foundation need
repair? Am J Clin Nutr 68, 218–219.