Untitled

Comision Federal de la mejora regulatoria
Coordinación general de la mejora regulatoria
22 octubre de 2009
Por este conducto nos estamos permitiendo, en tiempo y forma, hacer llegar a esa
autoridad, las propuestas, debidamente documentadas, a la modificación de la Norma
Oficial Mexicana NOM-051-SCFI-1994, Especificaciones generales de etiquetado
para alimentos y bebidas no alcohólicas preenvasado cuyo Proyecto de Modificación
se encuentra en fase de consulta pública para comentarios de los interesados en el
tiempo concedido para ello.
Es de aclarar que, como también lo acreditamos con la copia de nuestros escritos que
para tal efecto exhibimos, las propuestas y bibliografía de que se habla, la hemos hecho
llegar en diversas ocasiones a esa autoridad, buscando una solución para el problema
discriminatorio que resiente el producto ISOMALT, decisión que a pesar del tiempo
transcurrido nunca se ha dado, limitándose a expresarnos que nuestra solicitud se
evaluará en su momento exhibiendo copia de la documentación oficial que lo comprenda.
Ahora, con motivo de la revisión a la Norma de que se habla y de su Proyecto publicado
en el Diario Oficial el miércoles 26 de agosto de 2009 para comentarios al respecto,
esperamos que se haga efectiva la evaluación con la que se nos ha respondido hasta la
fecha.
Estimamos que ahora es la oportunidad para que quede reconocida nuestra posición en la
NOM 051-SCFI-SSA1-2009 el dispositivo de que se habla y poder comercializar el
producto en igualdad de circunstancias con los que lo hacen en el mercado de esa
manera y que son similares al nuestro.
1/2

Mucho agradeceremos dar el alcance que merece nuestra propuesta que formalmente
solicitamos con este escrito que se basa en la evidencia documentada que de antemano
hemos exhibido y la que anexo enviaremos nuevamente.
Anke Sentko
Vice President Regulatory Affairs
& Nutritional Communication
Wim Caers
Manager Regulatory Affairs
2/2

Subject: MEXICO – Comments on PROY‐NOM‐051‐SCFI/SSA1‐2009
SSS/WCS/IML
Regulatory Affairs
October 22 nd , 2009, page 1 / 4
On the 26th of August 2009, the proposal for PROY‐NOM‐051‐SCFI/SSA1‐2009 was published in Mexico’s
Official Journal. Comments can be submitted during the 60 days of public consultation.
Item of PROY‐
NOM‐051
3.3 Definition:
“Sugars”
Proposal Justification
“All mono‐ and
disaccharides present in a
food or non‐alcoholic
beverage, excluding
polyols and isomaltulose”
‐ Exclusion of polyols:
Because the physiological properties of polyols are
different from sugars: polyols are low glycemic, do not
promote dental caries and are reduced in calories (for
isomalt: Ziesenitz, S.C., 1996, Review ‘Basic Structure and
metabolism of isomalt’, Advances in Sweetness: 109‐133
[annex 1.1]).
It will bring the Mexican definition in line with Codex
Alimentarius (cfr. The joint FAO/WHO Scientific Update on
Carbohydrates in Human Nutrition (2007) excludes polyols
from the total sugar content [Annex 1.2]. and the European
Legislation ( 90/496/EEC; [annex 1.3]).
Polyols are mainly used for sugarfree products, no added
sugars and/or energy reduced type of products.
‐ Exclusion of isomaltulose:
Although this carbohydrate is a disaccharide from a
chemistry point of view, its nutritional and physiological
properties significantly differ from those of sugars as it is
e.g., non‐cariogenic, slow release and low glycaemic so that
it would be misleading to the consumer to have them
classified as sugars (see e.g. US‐FDA GRAS Notification No.
GRN 000184 (isomaltulose)[Annex 1.4]; US Health Claims
Regulation on dietary noncariogenic carbohydrate
sweeteners and dental caries [Annex 1.5]; ‘Palatinose TM
(isomaltulose) – a new innovative carbohydrate’ [Annex
1.6]; ‘Dossier for the scientific substantiation of claims
related to Palatinose TM and its nutritional physiological
properties’ [Annex 1.7]; Report from Imfeld (University of
Zürich): ‘Pilot Study on the Dental Care Properties of
Isomaltulose (Palatinose TM )’ [Annex 1.8];
It will also support correct information to the consumer.
Isomaltulose will be included under the “carbohydrate”
listing, like e.g. starch or maltodextrins, which is where it
really belongs, rather than under the “sugar” section.

Item of PROY‐
NOM‐051
3.19
Definition:
“Dietary fibre”
5.1.1
“Energy
calculation”
5.1.4 “Energy
content
polyalcohols
and
polydextrose”
Proposal Justification
We support the Dietary
Fibre definition of Codex
Alimentarius including DP 3
– 9 as stated in Footnote 2.
Carbohydrates: 4 kcal/g –
17 kJ
Proteins: 4 kcal/g – 17 kJ
Fats: 9 kcal/g – 37 kJ
Alcohol: 7 kcal/g – 29 kJ
Organic acids: 3 kcal/g – 13
kJ
Isomalt: 2.0 kcal/g – 8 kJ
Lactitol: 2.0 kcal/g – 8 kJ
Maltitol: 2.1 kcal/g – 8 kJ
Mannitol: 1.6 kcal/g – 6 kJ
Sorbitol: 2.6 kcal/g – 10 kJ
Xylitol: 2.4 kcal/g – 10 kJ
HSH (maltitol syrup): 3.0
kcal/g – 12 kJ
SSS/WCS/IML
Regulatory Affairs
October 22 nd , 2009, page 2 / 4
Regarding this definition we want to stress that we support
it, provided that as foreseen in footnote 2, carbohydrates
with a DP (degree of polymerisation) of 3 to 9 monomeric
units are included into the definition.
Non‐digestible oligosaccharides, e.g. oligofructose do
behave as dietary fibres, and these properties do not
change at DP =10 or higher. It will also harmonise the
Mexican DF definition with almost all other definitions, as
proposed worldwide, including EU, EFSA, Institute of
Medicin (US), FSANZ, ILSI, AACC…
For those ingredients, for which the scientific basis for an
individual energy value is established, this individual value
should be used.
Scientific evidence is available for the following energy
conversions factors:
‐ Inulin/oligofructose: in the US, a value of 1.5 kcal/g is
used. Further support for this value is given in Annex 2.1.
and Annex 2.2.
For those ingredients, for which the scientific basis for an
individual energy value is established, this individual value
should be used.
Having the NAFTA in mind, where US and Canada accepted
deviated caloric values for polyols, if necessary scientific
substantiation is available, it should also be appropriate as
well for Mexico to handle the same caloric values:
‐ US: each polyol has its own FDA notified energy
conversion factor, based upon a scientific evaluation by
FASEB (Federation of American Societies for Experimental
Biology; 1994; “The Evaluation of the Energy of Certain
Sugar Alcohols used as Food Ingredients”) [Annex 3.1].
“Letter of no objection” received for caloric value of
2.0 kcal/g for isomalt [Annex 3.2].
‐ Canada: has confirmed the use of 2 kcal/g (http://www.hc‐
sc.gc.ca/fn‐an/securit/addit/sweeten‐edulcor/polyols_polydextose_factsheet‐
polyols_polydextose_fiche‐eng.php)

Item of PROY‐
NOM‐051
A.3.1 & A.3.2
Proposal Justification
A definition and examples
for nutrition claims, are
given in PROY‐NOM‐051‐
SCFI/SSA1‐2009
(e.g. “Source of calcium”,
“High fibre content”, “low
in fats”)
SSS/WCS/IML
Regulatory Affairs
October 22 nd , 2009, page 3 / 4
Laying down fixed conditions for the nutrition claims,
makes it easier for producers and for regulators to control
the products. It avoids misleading communication and it
supports harmonization. We suggest these claims to be in
line with the Codex Alimentarius (see Annex 4.1 )

List of annexes:
SSS/WCS/IML
Regulatory Affairs
October 22 nd , 2009, page 4 / 4
Annex 1.1: Ziesenitz, S.C.; 1996; “Basic Structure and Metabolism of Isomalt”; Advances in Sweeteners.
Annex 1.2: “Joint FAO/WHO Scientific Update on Carbohydrates in Human Nutrition”; Eur. J. Clin. Nutr.
61(S1); 2007.
Annex 1.3: Council Directive 90/496/EEC on Nutrition Labelling of Foodstuffs.
Annex 1.4: US‐FDA GRAS Notification No. GRN 00184 (Isomaltulose).
Annex 1.5: US Health Claims Regulation on Dietary Noncariogenic Carbohydrate Sweeteners and Dental
Caries, Fed. Reg. Vol. 73, No. 102, May 27, 2008, 30299‐30301.
Annex 1.6: “Palatinose TM (isomaltulose) – a new innovative carbohydrate”
Annex 1.7: “Dossier for the Scientific Substantiation of Claims related to Palatinose TM and its Nutritional
Physiological Properties”
Annex 1.8: Report from Imfeld (University of Zürich): “Pilot Study on the Dental Care Properties of
Isomaltulose (Palatinose TM )”.
Annex 2.1: “The Determination of a Caloric Value for Inulin and Oligofructose”, Cantox Inc., 1999.
Annex 2.2: Roberfroid, M.B., 1999; “Caloric Value of Inulin and Oligofructose”.
Annex 3.1: Scientific Evaluation by FASEB, 1994; “Evaluation of the Energy of Certain Sugar Alcohols used
as Food Ingredients”
Annex 3.2: “Letter of no objection” from USA for the caloric value of 2.0 kcal/g for isomalt.
Annex 4.1: Codex Alimentarius CAC/GL 23‐1997 “Guidelines for Use of Nutrition and Health Claims”, Table
of Conditions for Nutrient Content.

Annex 1.1: Ziesenitz, S.C.; 1996; “Basic Structure and
Metabolism of Isomalt”; Advances in Sweeteners.

Annex 1.2: “Joint FAO/WHO Scientific Update on
Carbohydrates in Human Nutrition”;
Eur. J. Clin. Nutr. 61(S1); 2007.

Volume 61 Supplement 1 December 2007 www.nature.com/ejcn
Joint FAO/WHO Scientifi c
Update on Carbohydrates in
Human Nutrition
Guest Editors:
Chizuru Nishida, Frank Martinez Nocito
and Jim Mann

I Brouwer
Vrije Universiteit Amsterdam
Netherlands
B Bistrian
Harvard Medical School
USA
N Butte
University of Houston, Texas
USA
J Chen
Chinese Center for Disease
Control & Prevention
China
T Cole
University of London
UK
M Gibney
University College Dublin
Ireland
Editor-in-Chief
PS Shetty
University of Southampton, Southampton, UK
A Dangour
University of London
UK
J Hautvast
University of Wageningen
Netherlands
R Hurrell
Swiss Federal Institute of Technology
Switzerland
A Kurpad
Institute for Population Health Research
India
Associate Editors
D Lobo
University of Nottingham
UK
Editorial Board
M Lean
University of Glasgow
Scotland
J Mann
University of Otago
New Zealand
T McMichael
National Centre for
Epidemiology & Population Health
Australia
M Muller
University of Kiel
Germany
S Olsen
Danish Epidemiology Science Centre
Denmark
F Pasanisi
University of Naples Federico II
Italy
P Ritz
INSERM
France
M Soares
Curtin University
Australia
A Stephen
University of Cambridge
UK
HPS Sachdev
Sitaram Bhartia Institute of
Science and Research
India
A Sawaya
Federal University of Sao Paulo
Brazil
K Tontisirin
Mahidhol University
Thailand
J Tuomilehto
University of Helsinki
Finland
M Valencia
University of Yucatan
Mexico
H Vorster
North West University,
Potchefstroom
South Africa
K Westerterp
University of Limburg
Netherlands
W Willett
Harvard School of Public Health
USA

Volume 61, Supplement 1, December 2007
Joint FAO/WHO Scientific Update on Carbohydrates in
Human Nutrition
Guest Editors:
Chizuru Nishida, Frank Martinez Nocito and
Jim Mann

www.nature.com/ejcn
EJCN is published by Nature Publishing Group, a division of Macmillan
Publishers Ltd.
Scope EJCN is published monthly. It covers theoretical aspects of nutrition;
relations of function to nutritional status; nutritional causes and effects of
disease; the epidemiology of disease; community nutrition and education; the
determinants of eating behaviour. Papers on animal nutrition will not be
accepted unless they include parallel studies on human subjects.
This journal is covered by Current Contents, Current Contents Clinical Medicine,
Current Contents Life Sciences, EMBASE, Excerpta Medica, Index Medicus,
BIOSIS, CABS, CAB Abstracts, CAB Health and Nutrition Research Newsletter.
The Editor of the EJCN requires all contributors to read the following statement
in conjunction with the ‘Directions to Contributors’.
Ethics of investigation on human subjects. Authors must state in the text of their
paper the steps which they have taken to conform with accepted procedures
for the maintenance of clinical standards. Consent must be obtained from
each patient or subjects after full explanation of the purpose, nature and risks
of all procedures used. The fact that such consent has been given should be
recorded in the paper. If an Ethical Committee is in operation in the institution
where the work was done, it should be stated in the paper that the protocol
has been approved by that Committee. If no such Committee exists, this fact
should also be stated in the paper. The Editor has the right and duty not to
accept papers if, in his opinion, the ethical aspects are open to doubt.
Editorial To expedite the review process, authors should submit their
manuscript online at http://mts-ejcn.nature.com. Detailed ‘Directions to
Contributors’ are available at this website, or in print in the first issue of the year.
Publisher All business correspondence and enquiries about sponsorship
opportunities should be addressed to EJCN, Nature Publishing Group, The
Macmillan Building, 4 Crinan Street, London, N1 9XW, UK. Tel: +44 20 7833
4000. Fax: +44 20 7843 4640.
Executive Editor: Joyce-Rachel John
Production Controller: Steve Chillmaid
EJCN is online at www.nature.com/ejcn. Visit the journal’s home pages for
details of the aims and scope, readership, instructions to authors and how to
contact the Editor and publishing staff. Use the website to order a subscription,
reprints, a sample copy or individual articles.
Free to all readers: The complete text of all articles in the back archive 4 years
after publication; table of contents and abstracts for all articles published
within the last 4 years and the complete text of the January 2007 issue.
Register to receive the table of contents by e-mail as each issue is published.
2008 Subscriptions
Institutional Subscriptions
New institutional policy
From 2005 on, NPG has moved to a site licence policy for institutional online access,
using prices based on Full-Time Equivalents (FTE) or Research and Development
(R&D) staff. Institutions may also purchase a separate print subscription.
Subscribing to a site licence
Contact your local sales representative for a tailored price quote for your
institution. You will be required to complete a NPG site licence agreement.
More information, contact details and FTE/R&D definitions are available at the
http://www.nature.com/libraries.
Institutional print subscriptions
Orders can be placed with your regular subscription agent or through NPG –
either online at www.nature.com/ejcn or by contacting our customer services
department. Prices (excluding Japan) are as follows. For prices applicable to
Japan, please contact institutions@natureasia.com
The Americas $1022.00 Europe &812.00 UK/Rest of World d524.00
Personal Subscriptions
Personal customers who pay by personal cheque or credit card can either
purchase a combined print plus online subscription or an online-only
subscription. Prices (excluding Japan) are as follows. For prices applicable to
Japan, please contact institutions@natureasia.com
Combined (print plus online)
The Americas $462.00 Europe &367.00 UK/Rest of World d237.00
Online only
The Americas $415.00 Europe &330.00 UK/Rest of World d213.00
Contact information
SITE LICENCES
The Americas:
Tel: +1 800 221 2123. Fax: +1 212 689 9711.
E-mail: institutions@natureny.com
Asia Pacific:
Tel: +81 3 3267 8769. Fax: +81 3 3267 8746.
E-mail: institutions@natureasia.com
Australia and New Zealand:
Tel: +61 3 9825 1160. Fax: +61 3 9825 1010.
E-mail: nature@macmillan.com.au
India:
Tel: +91 124 288 1054. Fax: +91 124 288 1053.
E-mail: npgindia@nature.com
Rest of the World:
Tel: +44 (0) 20 7843 4759. Fax: +44 (0) 20 7843 4998.
E-mail: institutions@nature.com
PRINT SUBSCRIPTIONS
All customers:
Customer Service Department, Nature Publishing Group, Houndmills,
Basingstoke, Hants, RG21 6XS, UK. Tel: +44 (0) 1256 329242. Fax: +44 (0)
1256 812358. E-mail: subscriptions@nature.com
Prices are applicable in the following region: US dollars ($) for North, Central,
South America and Canada; Euros (&) for all European countries (excluding the
UK); Yen for Japan; Sterling (d) for UK and rest of world. Please ensure you use the
appropriate currency. All prices, specifications and details are subject to change
without prior notification. Single issues of EJCN are available. For information,
please contact: Subscriptions Dept, Nature Publishing Group, Brunel Road,
Houndmills, Basingstoke, Hampshire RG21 6XS, UK. Tel: +44 (0) 1256 329242.
Fax: +44 (0) 1256 812358. E-mail: subscriptions@nature.com
Advertising Enquiries concerning print and web advertisements should
be addressed to: Alf Anderson, Advertising Manager. Tel: +1 617 475 9231;
Fax: +1 617 494 4960. E-mail: a.anderson@boston.nature.com
Supplements Enquiries concerning supplements should be addressed to:
Sue Deeley, Supplements Editor. Tel: +1 212 726 9230; Fax: +1 646 563 7003.
E-mail: s.deeley@natureny.com
Reprints and permissions For detailed information on reprint and
permission requests, including instructions for obtaining these online, please
visit http://www.nature.com/reprints. Alternatively, for reprints, please
e-mail: reprints@nature.com. For North America: Tel: +1 212 726 9631.
Fax: +1 212 679 0843. Outside North America: Tel: +44 (0) 20 7843 4967.
Fax: +44 (0) 20 7843 4791. For permissions, please e-mail: ajpermissions@
nature.com. Tel: +44 (0) 20 7843 4613.
Copyright r 2007 Nature Publishing Group
Copyright r 2007, in the articles in this supplement, World Health Organization
ISSN 0954-3007
EISSN 1476-5640
All rights of reproduction are reserved in respect of all papers, articles, illustrations,
etc., published in this journal in all countries of the world. All material
published in this journal is protected by copyright, which covers exclusive
rights to reproduce and distribute the material. No material published in this
journal may be reproduced or stored on microfilm or in electronic, optical or
magnetic form without the written authorisation of the publisher.
Authorisation to photocopy material for internal or personal use, or internal or
personal use of specific clients, is granted by Nature Publishing Group, to
libraries and others registered with the Copyright Clearance Center (CCC)
Transaction Reporting Service, provided the relevant copyright fee is paid
directly to CCC, 222 Rosewood Drive, Danvers, MA 01923, USA. Identification
Code for EJCN: 0954-3007/07.
Apart from any fair dealing for the purposes of research or private study, or
criticism or review, as permitted under the Copyright, Designs and Patent Act
1988, this publication may be reproduced, stored or transmitted, in any form
or by any means, only with the prior permission in writing of the publishers, or
in the case of reprographic reproduction, in accordance with the terms of
licences issued by the Copyright Licensing Agency.
European Journal of Clinical Nutrition (ISSN 0954-3007) is published 12 times
a year by Nature Publishing Group, a division of Macmillan Publishers Ltd. US
Mailing Agent: Mercury Airfreight International Inc., 365 Blair Road, Avenel,
NJ 07001, USA. Periodicals postage paid at Rahway, NJ, USA.
POSTMASTER: Send address changes to EJCN Nature Publishing Group,
Mercury Airfreight International Inc., 365 Blair Road, Avenel, NJ 07001, USA.
Typeset by Macmillan India Ltd, Bangalore, India
Printed on acid-free paper, effective with Volume 48, Issue 1, 1994

S1
S5
S19
S40
S75
S100
S112
S122
S132
INTRODUCTION
Volume 61, Supplement 1, December 2007
FAO/WHO Scientific Update on carbohydrates in human nutrition: introduction
C Nishida and F Martinez Nocito
REVIEWS
Carbohydrate terminology and classification
JH Cummings and AM Stephen
Nutritional characterization and measurement of dietary carbohydrates
KN Englyst, S Liu and HN Englyst
Physiological aspects of energy metabolism and gastrointestinal effects of carbohydrates
M Elia and JH Cummings
Carbohydrate intake and obesity
RM van Dam and JC Seidell
Dietary carbohydrate: relationship to cardiovascular disease and disorders of carbohydrate metabolism
J Mann
Carbohydrates and cancer: an overview of the epidemiological evidence
TJ Key and EA Spencer
Glycemic index and glycemic load: measurement issues and their effect on diet--disease relationships
BJ Venn and TJ Green
CONCLUSIONS
FAO/WHO Scientific Update on carbohydrates in human nutrition: conclusions
J Mann, JH Cummings, HN Englyst, T Key, S Liu, G Riccardi, C Summerbell, R Uauy, RM van Dam,
B Venn, HH Vorster and M Wiseman
Copyright r 2007 Nature Publishing Group
Copyright r 2007, in the articles in this supplement, World Health Organization
Disclaimer: The named authors alone are responsible for the views expressed in this publication. All
reasonable precautions have been taken by WHO to verify the information contained in this
publication. However, the published material is beingdistributed without warranty of any kind, either
express or implied. The responsibility for the interpretation and use of the material lies with the reader.
In no event shall the World Health Organization be liable for damages arising from its use.
This issue is nowavailable at:
www.nature.com/ejcn

Beginning January 2008, Nature Publishing Group (NPG) will publish
Obesity. Th e preeminent journal in its fi eld, Obesity is intended to increase
knowledge, foster research and promote better treatment for people with
obesity and their loved ones.
With a newly redesigned front section, the journal will feature editorials,
commentaries, reviews, information about articles of interest in the issue
as well as timely and relevant coverage of research that defi nes the fi eld.
Covering the breadth of obesity research, the journal covers topics such as:
• clinical nutrition medicine
• pharmacology
• nutrition medicine
• genetics
• adipogenesis
• behavioral epidemiology
• biophysics and lipid metabolism
• exercise and human physiology
• nutritional epidemiology
Joining Nature Publishing Group not only means changes to the journal’s
look in print, but online as well. Obesity’s homepage will off er robust
online features unique to Nature Publishing Group, like:
• Table of Contents E-alerts is a free service which delivers the table of
contents page directly to your inbox on the day the issue publishes. Th e
quickest and easiest way to stay abreast of the latest research and reviews,
free monthly article promotions, special issues and focuses from Obesity.
• Free sample issue available all year long online at www.nature.com/obesity
• Integrated searches across all 70+ Nature Publishing Group journals and
products
• Reference linking via CrossRef and Medline
www.nature.com/obesity
ANNOUNCING A NEW LOOK
& NEW PUBLISHER FOR 2008
• phenotyping
• fat cell physiology
• aging, neuroscience
• transgenic models
• metabolic syndrome
• nutrition behavior
• pediatric obesity
• adipocyte cell biology
ISSN:
1930-7381
2005 IMPACT FACTOR:
3.491*
RANKING:
30/93
Endocrinology & Metabolism
8/55 Nutrition & Dietetics
EDITOR:
Dr. Richard Bergman
Offi cial Journal of
The Obesity Society
*Journal Citation Reports, Thomson, 2007

FAO/WHO Scientific Update on carbohydrates in
human nutrition: introduction
C Nishida 1 and F Martinez Nocito 2
1 Department of Nutrition for Health and Development, World Health Organization (WHO), Geneva, Switzerland and 2 Nutrition and
Consumer Protection Division, Food and Agriculture Organization of the United Nations (FAO), Rome, Italy
European Journal of Clinical Nutrition (2007) 61 (Suppl 1), S1–S4; doi:10.1038/sj.ejcn.1602935
Keywords: carbohydrates; human nutrition; expert consultation; FAO; WHO; recommendations
Carbohydrates in human nutrition
Since the 1950s, FAO and WHO have regularly held joint
expert meetings to review the state of scientific knowledge
on the role of various nutrients in the human diet, that is,
proteins, fats and oils, and most vitamins, minerals and trace
elements, to provide guidance on their requirements and
recommended intakes (Weisell, 2002).
The Joint FAO/WHO Expert Meeting on Carbohydrates in
Human Nutrition held in Geneva from 16 to 26 September
1979 was the first to focus on the topic of carbohydrates
(FAO, 1980) and aimed to review the role of carbohydrates as
determinants of human health and diseases. It was wide
ranging in scope and addressed the important role of
carbohydrates (1) as sources of energy contributing to the
improvement of human nutritional status; (2) as determinants
of the sensory qualities of foods, that is, flavour and
texture and the acceptability of foods; (3) and their influence
on the physiology and pathology of the large intestine,
particularly through a deeper understanding of the role of
dietary fibre; (4) as potential determinants of dental caries,
obesity, cardiovascular disease and diabetes; and (5) as they
relate to the nutrition of infants and young children,
including the role of lactose and its inclusion in weaning
foods.
The Expert Meeting recognized, however, that it was
inappropriate to focus on the nutritional effects of a single
dietary component due to the wide range of carbohydrates
consumed in the diet, and the fact that the overall
Correspondence: Dr C Nishida, Department of Nutrition for Health and
Development, World Health Organization (WHO), Avenue Appia 20, CH-1211
Geneva 27, Switzerland.
E-mail: nishidac@who.int
European Journal of Clinical Nutrition (2007) 61 (Suppl 1), S1–S4
& 2007 Nature Publishing Group All rights reserved 0954-3007/07 $30.00
www.nature.com/ejcn
nutritional consequences of any dietary pattern represents
the integration of a wide range of interactive effects. Further
studies were, therefore, considered to be necessary for
understanding the interactions between carbohydrates and
many other components of the diet to contribute to the
improvement of the health and nutritional well-being of the
world’s population (FAO, 1980).
Eighteen years later, the Joint FAO/WHO Expert Consultation
on Carbohydrates in Human Nutrition was convened in
Rome from 14 to 18 April 1997 (FAO, 1998). Much progress
was made in understanding the role that carbohydrates
play in human nutrition and health. These included the
following: (1) additional understanding of the role of dietary
fibre, and in particular, its role in moderating the process of
digestion in the small intestine and its potential as a major
substrate for fermentation in the colon; (2) increased
understanding of the diverse physiological roles that carbohydrates
have, dependent upon the site, rate and extent of
digestion and fermentation in the gut; (3) the potential
of carbohydrates to enhance physical performance through
glycogen loading; and (4) further understanding of the
relationship between the diet and various noncommunicable
diseases, including obesity, type II diabetes, coronary
heart disease and some forms of cancer. Thus, it was
again confirmed that carbohydrates are not only an energy
source, but also have important impacts on the maintenance
of health. The following recommendations were derived: (1)
the terminology used to describe dietary carbohydrates
should be based primarily on molecular size (degree of
polymerization), with additional terms used to define
nutritional groupings based on physiological properties;
(2) the total carbohydrate in the diet should be measured
as the sum of the individual carbohydrates rather than
‘by difference’, as was also recommended by the 1979

S2
Expert Meeting; (3) the analysis and labelling of dietary
carbohydrate should be based on chemical divisions;
(4) at least 55% of total energy should be provided
from a variety of carbohydrate sources, regardless of
the nature of the dietary pattern; and (5) the glycaemic
index—which measures the impact of foods on the
integrated response of blood glucose—be used to compare
foods of similar composition within food groups
(FAO, 1998).
Two other joint FAO/WHO meetings convened in early
2000 also contributed in part to recommendations relevant
to carbohydrates. These were the Joint WHO/FAO Expert
Consultation on Diet, Nutrition and the Prevention of
Chronic Diseases convened in Geneva, from 28 January to
1 February 2002 (WHO, 2003; Nishida and Shetty, 2004)
and the Technical Workshop on Food Energy—Methods of
Analysis and Conversion Factors held in Rome, from 3 to 6
December 2002 (FAO, 2003).
The overall objective of the 2002 Joint WHO/FAO Expert
Consultation was to update current international recommendations
on diet, nutrition and the prevention of chronic
diseases. The Consultation evaluated the latest scientific
evidence and lessons learned from implementing national
intervention strategies to reduce the burden of noncommunicable
diseases. On the basis of more recent evidence, the
earlier recommendations on population nutrient intake
goals to prevent diet- and nutrition-related chronic diseases
formulated in 1989 by the WHO Study Group (WHO, 1990)
were updated. Industrialization, urbanization, economic
development and market globalization have resulted in
rapid changes in dietary patterns described as ‘nutrition
transition’ which reflects both quantitative and qualitative
changes in the diet. The adverse dietary changes include
shifts in the structure of the diet towards a higher energy
density with a greater role for fat and sugars in foods, greater
saturated fat intake (mostly from animal sources), reduced
intakes of complex carbohydrates and dietary fibre and
reduced fruit and vegetable intakes (WHO, 2003). This,
together with a decline in energy expenditure associated
with a sedentary lifestyle, has significantly impacted the
health and nutritional status of the population. This
phenomenon has been most marked in developing countries
and those undergoing rapid socioeconomic transition, and
has contributed to the increasing burden of diet- and
nutrition-related noncommunicable diseases, such as obesity,
type II diabetes, cardiovascular disease, including
hypertension and stroke, and some forms of cancer. Population
nutrient intake goals updated by this Consultation have
provided the basis for dietary recommendations for the
prevention of these diseases when formulating national
dietary guidelines and national food and nutrition policy.
The outcomes and recommendations of the Consultation
also provided the scientific basis for the WHO Global
Strategy on Diet, Physical Activity and Health endorsed by
the 57th World Health Assembly (WHA 57.17) in 2004
(WHO, 2004).
European Journal of Clinical Nutrition
FAO/WHO Scientific Update on carbohydrates
C Nishida and F Martinez Nocito
The range of population nutrient intake goals for carbohydrates
recommended by the Consultation was 55–75% of
total energy (WHO, 2003), the same as that recommended
by the 1989 WHO Study Group (WHO, 1990). The wide
range was based on the consideration of protein and fat
requirements as well as the observation that intakes of
carbohydrates over this range were not always compatible
with optimal human health. Furthermore, the Consultation
emphasized that the aim of the recommendation was to
maximize the intake of minimally processed carbohydrates
and minimize the intake of free sugars (o10% of energy
intake). The Consultation further indicated that regular
consumption of whole-grain cereals, fruits and vegetables,
which are preferred sources of nonstarch polysaccharides,
were likely to reduce the risk of diet- and nutrition-related
noncommunicable diseases. The Consultation agreed that
the best definition of dietary fibre remains to be established,
given the potential benefits of resistant starch (WHO, 2003).
The 2002 FAO Technical Workshop on Food Energy—
Methods of Analysis and Conversion Factors (FAO, 2003) was
organized as a follow-up to the 2001 Joint FAO/WHO/UNU
Expert Consultation on Human Energy Requirements convened
in Rome, 17–24 October 2001 (FAO, 2004; Shetty and
Martinez Nocito, 2005) to review the issue of how best to
match energy requirements with food intakes, given the new
energy requirement values based on energy expenditure. The
Technical Workshop also addressed the request made by the
Codex Committee on Nutrition and Foods for Special
Dietary Usages (CCNFSDU) for harmonizing energy conversion
factors, and thus enabled uniformity in labelling and
information provided to consumers (CCNFSDU, 2001,
2002). The workshop reviewed the commonly used analytical
methods for protein, fat and carbohydrate, and made
recommendations regarding the preferred state-of-the-art
methods and the most appropriate technology, as well as
existing acceptable methods used in the absence of preferred
methods.
FAO/WHO Scientific Update on carbohydrates in
human nutrition
As part of the normative work and the complimentary
mandates of the two organizations to periodically update
nutrient requirements and regularly develop related global
guidelines, FAO and WHO have been exploring the possibility
of holding an expert consultation to update the work of
the 1997 Expert Consultation. Considered necessary given
the developments and other relevant recommendations
made during the intervening period, including those from
the 2002 Joint WHO/FAO Expert Consultation (WHO, 2003),
FAO and WHO agreed in 2005 to undertake a scientific
update on some of the key issues related to carbohydrates in
human nutrition. The key issues identified included terminology
and classification, measurement, physiology, carbohydrates
and diseases (that is, obesity, diabetes mellitus,

cardiovascular diseases and cancer), and glycaemic index
and glycaemic load. This update of existing knowledge and
evidence relating to the current recommendations was
viewed as essential in the process leading up to a forthcoming
Expert Consultation on Carbohydrates in Human
Nutrition.
Process of undertaking the scientific update and
criteria for selecting experts
The process and criteria for selecting experts to be invited to
prepare the scientific papers on each identified issue relating
to carbohydrates were discussed and agreed upon by FAO
and WHO. The identification of the issues to be reviewed
and possible experts to prepare these papers began in June
2005. The names of possible experts who might prepare or
peer-review each scientific review paper were identified after
consulting various other nutrition experts on the basis of
their competency and expertise in each of the identified
areas of work, as well as their independence. The consultation
and review of selecting possible experts continued until
September 2005. It was agreed that the review papers would
be a thorough scientific update of those identified issues
related to carbohydrates and of topics for further consideration,
both of which are presented in this supplement. It was
further agreed that the outcomes of this review process
should be seen as the conclusions of the scientific update,
not as updated recommendations.
By June 2006, the scientific review papers had been
completed and peer-reviewed. The papers were then further
examined, together with peer reviewer comments, at an
authors’ meeting held in Geneva from 17 to 18 July 2006, to
identify any gaps or issues needing further consideration
before being finalized. The authors’ meeting was also
attended by selected expert peer reviewers to ensure a high
level of critical review and analysis of each paper. Taking the
critical comments received from the peer-review process and
the discussions at the authors’ meeting, the scientific review
papers were further revised and sent for a second round of
peer review before their finalization.
Transparency of the process
Forty experts were involved in this scientific update, serving
either as an author of a review paper or as an expert peer
reviewer. Before being officially invited to take part in this
work, all experts were requested to declare any possible
conflict of interest to ensure the integrity of each expert’s
contribution. These declarations of interest, which were
carefully assessed by FAO and WHO, would be publicly
disclosed after obtaining the expert’s agreement in writing to
do so. Public disclosure of experts’ declaration of interest
involved (1) announcing the declarations at the authors’
meeting; and (2) appropriately disclosing the declaration in
FAO/WHO Scientific Update on carbohydrates
C Nishida and F Martinez Nocito
the subsequent publication of the papers prepared for the
scientific update, that is, this supplement. The primary
purpose of this transparency was to ensure open and
productive debate on the key issues selected for the scientific
update by providing insight into the differing perspectives of
all participating experts.
The review papers published in this supplement provide
the rationale and scientific basis supporting the conclusions
and proposals presented by the Scientific Update. In
addition, they provide the scientific community with a
valuable resource relating to several important nutrition
topics. Rapid progress is taking place in a number of
scientific fields that affect issues related to human nutrient
requirements. Evidence derived from a range of different
scientific approaches has helped to clarify the role of diet,
some individual dietary components and even particular
nutrients in the aetiology of various diseases. Thus, changes
in diet have strong effects, both positive and negative, on the
health and nutritional status of people throughout the life
course, and the potential to promote human health and
reduce the risk of a number of chronic diseases. FAO and
WHO are committed to continue to provide scientifically
sound, evidence-based advice and guidelines on human
nutrient requirements and other related topics through a
transparent and neutral process.
Acknowledgements
Special acknowledgement and deep appreciation are expressed
by FAO and WHO to Dr Kraisid Tontisirin, the former
Director, Nutrition and Consumer Protection Division in
FAO; and Dr Prakash Shetty, the former Chief, Nutrition
Planning, Assessment and Evaluation Service in the Nutrition
and Consumer Protection Division, FAO, for their
tremendous support and invaluable contributions in undertaking
the Joint FAO/WHO Scientific Update on Carbohydrates
in Human Nutrition. FAO and WHO also wish to
express special appreciation to the authors of the review
papers prepared for the Scientific Update, as well as the
expert peer reviewers, who critically evaluated the review
papers and provided valuable comments and contributions.
We also thank Dr Denise Costa Coitinho, former Director,
Department of Nutrition for Health and Development,
WHO; Dr Ezzeddine Boutrif, Director, Nutrition and Consumer
Protection Division, FAO; and Dr Jorgen Schlundt,
Director, Department of Food Safety, Zoonoses and Foodborne
Diseases and Acting Director, Department of Nutrition
for Health and Development, for their sustained support in
carrying out and completing this scientific work.
Conflict of interest
The authors, Dr Chizuru Nishida and Mr Frank Martinez
Nocito, declare no conflict of interest.
S3
European Journal of Clinical Nutrition

S4
References
FAO/WHO Scientific Update on carbohydrates
C Nishida and F Martinez Nocito
CCNFSDU (2001). 23rd Session, November 2001 (ALINORM 03/26,
paragraphs 133-137). Codex Committee on Nutrition and Foods
for Special Dietary Uses (CCNFSDU).
CCNFSDU (2002). 24th Session, November 2002 (ALINORM 03/26A,
paragraphs 113-115). Codex Committee on Nutrition and Foods
for Special Dietary Uses (CCNFSDU).
FAO (1980). Carbohydrates in human nutrition. Report of a Joint FAO/
WHO expert meeting. FAO Food and Nutrition Paper No. 15 Food
and Agriculture Organization of the United Nations: Rome.
FAO (1998). Carbohydrates in human nutrition. Report of a Joint FAO/
WHO Expert Consultation. FAO Food and Nutrition Paper No. 66
Food and Agriculture Organization of the United Nations: Rome.
FAO (2003). Food energy—methods of analysis and conversion factors.
Report of a technical workshop. FAO Food and Nutrition Paper No.
77 Food and Agriculture Organization of the United Nations: Rome.
FAO (2004). Human energy requirements Report of a Joint FAO/WHO/
UNU Expert Consultation. FAO Food and Nutrition Technical
European Journal of Clinical Nutrition
Report Series No.1 Food and Agriculture Organization of the
United Nations: Rome.
Nishida C, Shetty PS (eds). (2004). Diet, nutrition and the prevention
of chronic diseases. Background papers of the Joint WHO/
FAO Expert Consultation. Public Health Nutr 7 (Suppl 1A),
S99–S250.
Shetty PS, Martinez Nocito F (eds). (2005). Human energy requirements.
Background papers of the Joint FAO/WHO/UNU expert
consultation. Public Health Nutr 8 (Suppl 7A), S929–S1228.
Weisell R (2002). The process of determining nutritional requirements.
Food Nutr Agric 30, 14–21.
WHO (1990). Diet, nutrition and the prevention of chronic diseases.
Report of a WHO Study Group. WHO Technical Report Series No.
797 World Health Organization: Geneva.
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.
WHO (2004). Global Strategy on Diet, Physical Activity and Health.
World Health Organization: Geneva.

REVIEW
Carbohydrate terminology and classification
JH Cummings 1 and AM Stephen 2
1 Gut group, Division of Pathology and Neuroscience, Ninewells Hospital and Medical School, Dundee, UK and 2 Population 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
The dietary carbohydrates are a diverse group of substances
with a range of chemical, physical and physiological properties.
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
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
density, large bowel cancer, constipation and resistance to
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 analogous
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

S6
Table 1 The major dietary carbohydrates
Class (DP a ) Subgroup Principal components
Sugars (1–2) Monosaccharides Glucose, fructose, galactose
Disaccharides Sucrose, lactose, maltose,
Oligosaccharides
(3–9) (short-chain
carbohydrates)
Polysaccharides
(X10)
and an understanding of the physiological and health effects
of these macronutrients.
A chemical approach divides carbohydrates into three
main groups, sugars (DP1–2), oligosaccharides (short-chain
carbohydrates) (DP3–9) and polysaccharides (DPX10).
Sugars comprise (i) monosaccharides, (ii) disaccharides and
(iii) polyols (sugar alcohols). Oligosaccharides are either (a)
malto-oligosaccharides (a-glucans), principally occurring
from the hydrolysis of starch and (b) non-a-glucan such as
raffinose and stachyose (a galactosides), fructo- and galactooligosaccharides
and other oligosaccharides. Polysaccharides
may be divided into starch (a-1:4 and 1:6 glucans) and nonstarch
polysaccharides (NSPs), of which the major components
are the polysaccharides of the plant cell wall such as
cellulose, hemicellulose and pectin but also includes plant
gums, mucilages and hydrocolloids. Some carbohydrates,
like inulin, do not fit neatly into this scheme because they
exist in nature in multiple molecular forms. Inulin, GFN,
from plants may have from 2 to 200 fructose units and so
crosses the boundary between oligosaccharides and polysaccharides
(Roberfroid, 2005).
A variety of methodologies are available for the measurement
of the carbohydrate content of food and the components
are listed in Table 1 (Englyst et al., 2007).
Terminology
Polyols (sugar
alcohols)
Maltooligosaccharides
(a-glucans)
Non-a-glucan
oligosaccharides
Carbohydrate terminology and classification
JH Cummings and AM Stephen
trehalose
Sorbitol, mannitol, lactitol,
xylitol, erythritol, isomalt,
maltitol
Maltodextrins
Raffinose, stachyose, fructo and
galacto oligosaccharides,
Starch
polydextrose, inulin
Amylose, amylopectin, modified
(a-glucans) starches
Non-starch Cellulose, hemicellulose, pectin,
polysaccharides arabinoxylans, b-glucan,
(NSPs)
glucomannans, plant gums and
mucilages, hydrocolloids
a Degree of polymerization or number of monomeric (single sugar) units.
Based on Food and Agriculture Organization/World Health Organization
‘Carbohydrates in Human Nutrition’ report (1998), and Cummings et al.
(1997).
Total carbohydrate
Although the individual components of dietary carbohydrate
are readily identifiable, there is some confusion as to
what comprises total carbohydrate as reported in food tables.
European Journal of Clinical Nutrition
Two principal approaches to total carbohydrate are used,
first, that derived ‘by difference’ and second, the direct
measurement of the individual components that are then
combined to give a total. Calculating carbohydrate ‘by
difference’ has been used since the early 20th century and
is still widely used around the world (Atwater and Woods,
1986; United States Department of Agriculture, 2007). The
moisture, protein, fat, ash and alcohol content of a food are
determined, subtracted from the total weight of the food and
the remainder, or ‘difference’, is considered to be carbohydrate.
There are, however, a number of problems with this
approach in that the ‘by difference’ figure includes noncarbohydrate
components such as lignin, organic acids,
tannins, waxes and some Maillard products. In addition to
this error, it combines all the analytical errors from the other
analyses. Also, a single global figure for carbohydrates in
food is uninformative because it fails to identify the many
types of carbohydrates and thus to allow some understanding
of the potential health benefits of those foods.
Direct analysis of carbohydrate components and summation
to obtain a total carbohydrate value has been the basis
of carbohydrate analysis in the UK since 1929, when the first
values were published by McCance and Lawrence (1929).
Those countries that use McCance and Widdowson’s, The
Composition of Foods (Food Standards Agency/Institute
of Food Research, 2002) also express carbohydrate using
this approach. The total figure obtained is for what McCance
and Lawrence called ‘available carbohydrate’ and therefore
differs from carbohydrate by difference in that it does not
contain the plant cell wall polysaccharides (fibre). In
addition, it is not complicated by analytical difficulties with
other food components. Dietary intake of total carbohydrate
and its components using direct analysis enables examination
of geographic variations and changes in intake over
time of individual carbohydrate types and their relationship
with health outcomes. Total carbohydrate by direct measurement
is preferable and simplified methods to do this should
be developed.
Figures obtained for carbohydrate by difference and
carbohydrate analysed directly are not always the same,
particularly for complex mixtures, and foods containing fibre
or certain types of starch, like pasta (Stephen, 2006). This
results in apparently different carbohydrate intakes for the
same list of foods consumed, as shown in Table 2. Fifty-two
dietary records from a study conducted in Canada, where
carbohydrate by difference is used (Health Canada, 2005)
were subsequently analysed in the UK using values based on
McCance and Widdowson’s The Composition of Foods (Holland
et al., 1991b, 1992). In this study, energy intake was 12%
higher and carbohydrate intake 14% higher when measured
‘by difference’ (Stephen, 2006). Comparison of carbohydrate
intake among different countries should therefore be viewed
with caution if the method of carbohydrate determination is
not the same. Worldwide variations in carbohydrate intake
assumed to be due to differences in types of foods consumed,
are also, in part, due to methodology.

Table 2 Energy and macronutrient intakes for 52 weighed records
analysed using Canadian and UK food tables
Energy
(kcal)
Protein
(g) (%)
Fat
(g) (%)
CHO
(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 viscosity,
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
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 disaccharides
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 analytically
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
S7
European Journal of Clinical Nutrition

S8
Carbohydrate terminology and classification
JH Cummings and AM Stephen
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 problems
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 considered
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 Nomenclature,
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/polysaccharide
boundary. In categorizing oligosaccharides found
normally in the diet, alcohol precipitation would seem to be
the most practical way of delineating them from polysaccharides.
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) maltodextrins,
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-Ntetraose.
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 bifidobacteria
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 carbohydrates
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 vegetables
and legumes and consists of only glucose molecules. It
occurs in a partially crystalline form in granules and
comprises two polymers: amylose (DPB10 3 ) and amylopectin
(DPB10 4 –10 5 ). 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. Amylopectin
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
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 functional
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 retrogradation,
which is part of the process of staling of bread, for
example. Substitution also lowers gelatinization temperature,
provides freeze–thaw stability and increases viscosity. Crosslinking
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
S9
European Journal of Clinical Nutrition

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 10 3 –10 6 ) 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 hemicelluloses
are the arabinoxylans found in cereals. About half
of hemicelluloses contain uronic acids, which are carboxylated
derivatives of glucose and galactose. They are important
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 gelforming
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
Carbohydrate terminology and classification
JH Cummings and AM Stephen
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 tetragonolobus),
and carob gums are similar 1–4b-D galactomannans
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 toothpastes.
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 individuals
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

Table 3 Principal physiological properties of dietary carbohydrates
Provide
energy
Increase
satiety
Glycaemic a
Cholesterol
lowering
carbohydrate and these are listed in Table 5. However, the
problems are by no means insurmountable in reconciling
these different objectives for classification. With a sound
chemical identification for all the carbohydrates, it is then
possible to group them according to their health and
physiological effects.
Prebiotics
‘A prebiotic is a non-digestible food ingredient that beneficially
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 properties
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
Increase calcium
absorption
Source of
SCFA b
Alter balance
of microflora
(prebiotic)
Increase stool
output
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
Glucose, fructose, galactose, sucrose, lactose,
maltose, trehalose, maltodextrins, starch
Non-glycaemic Polyols, oligosaccharides (non-a-glucan), resistant
Glycaemic a
Increase stool
output
No effect on stool
weight
NSP, non-starch polysaccharide.
a Defined as in Table 2.
and modified starches, NSP
Polyols (except erythritol), some starches, NSP,
lactose (in some populations), fructose (if taken in
large amounts)
Glucose, galactose, sucrose, maltose, trehalose,
maltodextrins, oligosaccharides, most starches
Carbohydrate terminology and classification
JH Cummings and AM Stephen
Table 5 Preferred terminology of dietary carbohydrates
Chemical Physiological/botanical
Useful Monosaccharides
Disaccharides
Polyols
Total sugars
Short-chain
carbohydrates
Oligosaccharides
Polysaccharides
Starch
Non-starch
polysaccharides
Total carbohydrate
Less
useful
Sugars
Sugar
Free sugars
Refined sugars
Added sugars
Extrinsic and intrinsic
sugars
a Intrinsic plant cell wall polysaccharides.
Prebiotic
Resistant starch
Dietary fibre a
Glycaemic
Immunomodulatory
Non-digestible oligosaccharides
Soluble and insoluble fibre
Available and unavailable
carbohydrate
Complex carbohydrate
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
S11
European Journal of Clinical Nutrition

S12
Carbohydrate terminology and classification
JH Cummings and AM Stephen
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 compelling
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, analytical
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 hypothesis,
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).
European Journal of Clinical Nutrition
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 nondigestibility
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. Moreover,
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 carbohydrates
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

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 polysaccharides’.
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 epidemiological
evidence points to fruits, vegetables and wholegrain
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 carbohydratecontaining
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
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 nonglycaemic,
which is closer to the original concept of
McCance and Lawrence. However, the FAO Technical workshop
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
S13
European Journal of Clinical Nutrition

S14
carbohydrate’, whereas carbohydrates that pass to the
large intestine prior to being metabolized, is referred to
as ‘non-glycaemic carbohydrate’. Most mono- and disaccharides,
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
Carbohydrate terminology and classification
JH Cummings and AM Stephen
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 underestimated
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
(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 properties
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
S15
European Journal of Clinical Nutrition

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
Carbohydrate terminology and classification
JH Cummings and AM Stephen
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 carbohydrates
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. Consumption,
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 digestion
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-toleranceimportance
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 nonmilk
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). Oligosaccharides
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 highresistant-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
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 approaches
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 fructooligosaccharides
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 serumlipoproteins
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.
S17
European Journal of Clinical Nutrition

S18
Carbohydrate terminology and classification
JH Cummings and AM Stephen
Thane CT, Jones AR, Stephen AM, Seal CJ, Jebb SA (2005). Wholegrain
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
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 Bifidobacterium
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.

REVIEW
Nutritional characterization and measurement of
dietary carbohydrates
KN Englyst 1 , S Liu 2 and HN Englyst 1
1 2
Englyst Carbohydrates, University of Southampton Science Park, Southampton, UK and Departments of Epidemiology and Medicine,
University of California, Los Angeles, CA, USA
Dietary carbohydrate characterization should reflect relevant nutritional and functional attributes, and be measured as
chemically identified components. A nutritional classification based on these principles is presented, with a main grouping into
‘available carbohydrates’, which are digested and absorbed in the small intestine providing carbohydrates for metabolism, and
‘resistant carbohydrates’, which resist digestion in the small intestine or are poorly absorbed/metabolized. For the available
carbohydrates, the chemical division into the starch and total sugars categories does not adequately reflect the physiological or
nutritional attributes of foods. Characterizing carbohydrate release from starchy foods provides insight into some of the inherent
mechanisms responsible for the varied metabolic effects. Also, a pragmatic approach to product signposting consistent with
guidelines to limit free (or added) sugars is proposed. The most prominent of the resistant carbohydrates are the non-starch
polysaccharides (NSP) from plant cell walls, which are characteristic of the largely unrefined plant foods that provide the
evidence base for the definition and measurement of dietary fibre as ‘intrinsic plant cell-wall polysaccharides’ as proposed in
conjunction with this paper and endorsed by the scientific update. Indigestibility in the small intestine was not considered to be
an adequate basis for the definition of dietary fibre, as there is insufficient evidence to establish public health policy by this
approach and concerns have been raised about potential detrimental effects of high intakes of rapidly fermentable resistant
carbohydrates. Functional ingredients such as resistant starch and resistant oligosaccharides should therefore be researched and
managed separately from dietary fibre, using specific health or function claims where appropriate. This structured approach to
the characterization of nutritionally relevant features of dietary carbohydrates provides the basis for establishing population
reference intakes, nutrition claims and food labelling that will assist the consumer with properly informed dietary choices.
European Journal of Clinical Nutrition (2007) 61 (Suppl 1), S19–S39; doi:10.1038/sj.ejcn.1602937
Keywords: available carbohydrates; resistant carbohydrates; dietary fibre; non-starch polysaccharides; resistant starch; free
sugars; glycaemic index
Introduction
The classification and measurement of dietary carbohydrates
requires a systematic approach that describes both the
chemical and functional properties of carbohydrates in
foods. The objective of this paper is therefore to provide an
in-depth examination of the issues essential to the achievement
of a suitable approach to the nutritional characterization
of dietary carbohydrates. As summarized in Figure 1,
there are a number of reasons why information on the type
and amounts of carbohydrates present in foods is required
including, nutrition labeling, food composition tables,
Correspondence: Dr K Englyst, Englyst Carbohydrates, University of
Southampton Science Park, Southampton, UK.
E-mail: klaus@englyst.co.uk
European Journal of Clinical Nutrition (2007) 61 (Suppl 1), S19–S39
& 2007 Nature Publishing Group All rights reserved 0954-3007/07 $30.00
www.nature.com/ejcn
product development and nutrition research. These applications
all have an impact on public health, which must be
considered the ultimate purpose of providing appropriate
carbohydrate characterizations.
Nutritional considerations for the classification and
measurement of carbohydrates
Foods can contain a range of chemically distinct carbohydrate
substances, which have varied gastrointestinal and
metabolic properties. In addition, biological origin and food
processing have an important role in determining the overall
attributes of the food matrix and the physico-chemical
properties of carbohydrates in foods, which can have a major
impact on their physiological handling (Figure 2). This

S20
Food
Composition
Nutrition Labelling
Food Tables
Nutrition
Research
Metabolic
Epidemiological
Product
Development
Processing
Ingredients
Public Health
Dietary guidelines, Population reference intake values, Nutrition & health claims
Figure 1 Requirements for carbohydrate measurements. A range of
carbohydrate measurements are used in diverse but interrelated
fields within nutrition and food technology. Carbohydrate characterizations
should evolve to describe specific functional attributes
of carbohydrate containing foods as these are identified by nutrition
research. This then informs the development of appropriate
measurements that can be applied in food composition and product
development, which in turn stimulates further research. These
activities contribute to the scientific evidence base on which dietary
guidelines are formulated. Appropriate carbohydrate characterizations
can assist the consumer with informed diet selections and
thereby can make a significant contribution to public health.
Chemical Identity
(Sugar identity, linkage type, size)
Food Matrix
(Biological origin, food processing)
Carbohydrate Food Properties
(Physico-chemical characteristics)
Meal Factorss Subject Factors
Gastrointestinal Handling
(Rate and extent of digestion and absorption)
(Absorbed in small intestine) (Entry to large intestine)
Physiology and Utilisation of
Available Carbohydrates
(Glycaemic index, substrate metabolism)
Physiology and Utilisation of
Resistant Carbohydrates
(Fermentability, prebiotic effect, SCFAs)
Figure 2 Determinants of gastrointestinal fate of dietary carbohydrates.
Largely depending on their physiochemical properties,
different carbohydrate foods can exert a range of physiological
effects, including varied rates of digestion and absorption in the
small intestine and varied fermentability and profile of fermentation
products in the large intestine. Even though there will be a degree
of variation in gastrointestinal handling due to meal and subject
factors, the emphasis from a nutritional perspective must to be on
describing the inherent physico-chemical characteristics of the
foods, reflecting both the chemical identity of carbohydrates and
the influence of the food matrix on functional attributes.
variability in functionality needs to be considered in the
nutritional characterization of dietary carbohydrates, taking
into account issues of both food matrix and chemical
identity (Englyst and Englyst, 2005). Based on current
knowledge of the mechanisms by which dietary carbohydrates
exert their influence on physiology and health, it
is possible to describe these characteristics and incorporate
them into an overall classification scheme (Table 1) that can
evolve as new evidence becomes available. Other papers in
this report have dealt with physiology and health aspects
of dietary carbohydrates in detail. The key nutritional issues
connected with the characterization of carbohydrates are
summarized here.
European Journal of Clinical Nutrition
Nutritional characterization and measurement of dietary carbohydrates
K Englyst et al
Gastrointestinal fate and metabolizable energy
In terms of providing the body with metabolizable substrates,
the digestion of carbohydrates should be considered
as an event of both the upper (absorption of metabolizable
carbohydrate) and lower (fermentation providing shortchain
fatty acids (SCFA)) gastrointestinal tract. Therefore,
for calculation of the contribution to energy, there is a need
to know the gastrointestinal and metabolic fate of carbohydrates
(which is discussed in more depth in the accompanying
physiology paper by Elia and Cummings (2007).
Although a number of essentially synonymous terms have
been used to describe this division, the terms available
carbohydrate and resistant carbohydrate should probably be
considered the most informative and these can be defined as
follows.
Available carbohydrates are those that are absorbed in the
small intestine and provide carbohydrate for metabolism.
This definition is equivalent to the glycaemic carbohydrate
term used in the 1998 Food and Agriculture Organization/
World Health Organization report on carbohydrates, but
which has not been widely used. The ‘available carbohydrate’
term is long established and has been widely
adopted (McCance and Lawrence, 1929; FAO, 2003).
Resistant carbohydrates are those that resist digestion in
the small intestine, or are poorly absorbed and/or metabolized.
This definition is based on a similar one proposed to
distinguish gastrointestinal fate (Rumessen, 1992), but
includes the aspect of poorly metabolized to more effectively
encompass the polyols. In essence, the resistant carbohydrate
definition is equivalent to the indigestible, unavailable
and non-glycaemic terms, but provides a relatively more
accurate description of this grouping of carbohydrates.
Metabolism of different types of sugars
The available carbohydrates are absorbed and metabolized
as the monosaccharides glucose, fructose and galactose,
although lactase deficiency can result in malabsorption of
lactose (Gudmand-Hoyer, 1994). While glucose is utilized by
all tissues, the majority of fructose and galactose metabolism
occurs in the liver, with an estimated 50–70% hepatic
extraction of fructose from the portal vein. Although it is
difficult to draw firm conclusions regarding the nutritional
implications of the metabolic effects of ingesting different
types of sugar, enough concern has been raised, particularly
regarding fructose, to warrant monitoring the intake of
individual sugars (Daly, 2003; Fried and Rao, 2003; Gross
et al., 2004).
Rate of digestion and absorption
The influx of exogenous carbohydrate for metabolism is
determined by the rate that carbohydrates become available
for absorption at the epithelium of the small intestine. This
is influenced by numerous gastrointestinal factors, including
the rate that carbohydrates leave the stomach and the

Table 1 Nutritional characterization of dietary carbohydrates. Modified from Englyst and Englyst 2005
Main categories Chemical
components
Available
carbohydrates
Sugars
Starch
Resistant
carbohydrates NSP
Nutritional grouping Physiology and health
Lactose Malabsorbed by those with lactase deficiency
Fructose (including from sucrose) Largely metabolized by liver. Possible detrimental effect on lipid metabolism
Available glucose from sugars,
maltodextrins and starch. Rate of
release measured as RAG and SAG
diffusion of released sugars from the alimentary food bolus.
The rate that carbohydrates are released from food, through
the disruption of the food matrix and the action of
endogenous amylases on starch, is therefore an important
determinant of carbohydrate entry to the portal vein.
Carbohydrate type, biological origin and food processing
all contribute to the food properties that can influence the
rate of carbohydrate release (Figure 2). The nutritional
significance of the rate of carbohydrate digestion and
absorption is the impact it has on postprandial blood
glucose homeostasis and the associated metabolic and
endocrine responses. Although the glycaemic index (GI)
has been the subject of some controversy, the majority of the
metabolic and epidemiological evidence lends support to an
increase in the consumption of slow-release carbohydrates in
place of their rapidly absorbed high GI counterparts (Jenkins
et al., 2002; Willett et al., 2002; Brand-Miller et al., 2003).
Functional properties of resistant carbohydrates
Nutritionally, the most prominent resistant carbohydrates
are the intrinsic plant cell-wall polysaccharides, which, as
described in later sections, provide the only definition of
dietary fibre consistent with the plant-rich diet. There are
numerous other sources of resistant carbohydrates that occur
naturally in small amounts or that have been developed
as functional ingredients. These include extracted polysaccharides
such as gums, oligosaccharides such as fructans,
polydextrose, resistant maltodextrins and high resistant
starch (RS) ingredients. Depending on their physico-chemical
properties, these heterogeneous resistant carbohydrates have
a range of properties that include viscosity in the upper
RAG and SAG reflect the rate of glucose release from food, which is a main
determinant of the GI. Evidence to suggest that metabolic response
associated with slow-release carbohydrates are most conducive to optimal health
RS Varied rate and extent of fermentation. Insufficient knowledge of effect on health
Dietary fibre (intrinsic plant cell
wall polysaccharides)
Nutritional characterization and measurement of dietary carbohydrates
K Englyst et al
Marker for minimally refined plant foods that are rich in micronutrients and shown
to be beneficial to health
Added NSP Varied rate and extent of fermentation. Some have specific functional properties
RSCC Present naturally and added Varied rate and extent of fermentation. Some have specific functional properties
Sugar alcohols Present naturally and added Partly absorbed and metabolized, and partly fermented
Abbreviations: GI, glycaemic index; NSP, non-starch polysaccharides; RAG, rapidly available glucose; RS, resistant starch; RSCC, resistant short-chain carbohydrates;
SAG, slowly available glucose.
gastrointestinal tract (Ellis et al., 1996), fermentation and
fermentation products (Wong et al., 2006), prebiotic effects
(Roberfroid, 2005; Macfarlane et al., 2006) and mineral
absorption (Abrams et al., 2005). However, as functional
ingredients can be incorporated into foods in high amounts,
there has also been concern about potentially detrimental
effects of large amounts of easily fermentable carbohydrate
reaching the large intestine (Goodlad, 2007). There is
therefore a requirement to research and evaluate how these
substances should be managed from a health promotion
perspective.
Food properties
The nutritional role of dietary carbohydrates cannot be
adequately addressed without consideration of the overall
characteristics of the foods themselves. Therefore, although
fruit, vegetables and whole grains are considered carbohydrate-rich
foods, their health benefits can often be attributed
to their low-energy density and high content of micronutrients
and phytochemicals (Southgate and Englyst, 1985;
Liu et al., 2000a, b, 2001; Liu, 2002; Englyst and Englyst,
2005). These overall nutritional attributes of foods need to be
recognized within dietary guidelines and perhaps, just as
importantly, they ought to be supported by consistent public
health messages relating to dietary carbohydrate consumption.
One approach is to use a prefix identifying the food source
of the carbohydrate, with the most recognized example
being the division between intrinsic and extrinsic (free)
sugars, which describes whether or not they are contained
within cellular structures. At face value, it seems rather
contradictory that consumption of the intrinsic sugars from
S21
European Journal of Clinical Nutrition

S22
fruit and vegetables should be promoted, while the chemically
identical extrinsic sugars are restricted. Of course, it is
not the intrinsic sugars themselves that are being promoted,
but rather the overall health benefits associated with the
fruit and vegetable food group.
Measurement of dietary carbohydrates
Carbohydrate determinations should describe chemical
composition accurately, and provide information of nutritional
relevance, thereby complementing dietary guidelines.
The traditional calculation of carbohydrate ‘by difference’
does not conform to either of these criteria as (i) it combines
the analytical uncertainties of the other macronutrient
measures as well as any unidentified material present and
(ii) a single value for carbohydrate cannot reflect the range of
carbohydrate components or their diverse nutritional properties.
For this reason, carbohydrates should instead be
categorized based on relevant nutritional properties, and
measured as the sum of chemically identified components
(Table 1). The analytical challenge is to apply chemical,
physical and enzymatic approaches to exploit these characteristic
differences to achieve determinations of each
carbohydrate fraction. The measurement principles for the
main carbohydrates are described in Table 2. It is always
preferable to apply rational methods (which specifically
measure the component of interest), rather than empirical
methods (which are defined by the methodology) or
proximate analysis, which are either prone to errors or
limited in their interpretation.
The basic requirements of analytical methodologies for
the determination of dietary carbohydrates as the sum of
their constituent sugars can be described as follows.
Sample preparation. Preparation techniques should ensure
sample homogeneity and facilitate the extraction of the
nutrients of interest. For compositional analysis, this is
typically achieved by freeze-drying and milling the food. The
exception is for measurements of carbohydrates that reflects
their digestibility (for example, RS). Such samples need to be
prepared and analysed ‘as eaten’.
Isolation of specific fractions. The first stage of the analysis is
to ensure that the fraction of interest is completely extracted
from the food matrix in its native form (for example, sugars),
or dispersed to such an extent that it can be hydrolysed and
measured as its component parts (for example, total starch).
Interfering compounds can be accounted for by a sample
blank measurement, although it is preferable to remove
these by enzymatic or physical approaches, especially when
the sample blank is high or the compound of interest is
present in only small amounts.
Hydrolysis to constituent sugars. Once appropriately isolated,
oligosaccharides and polysaccharides may be subjected to
European Journal of Clinical Nutrition
Nutritional characterization and measurement of dietary carbohydrates
K Englyst et al
enzymatic (adds specificity) or acidic (when appropriate
enzyme unavailable) hydrolysis to release their constituent
sugars.
Detection. Separation by gas chromatography or highperformance
liquid chromatography (HPLC) can be used to
measure specific monosaccharaides, disaccharides and small
oligosaccharides. Colorimetric assays can be used to measure
sugars with reducing groups. Enzyme-linked colour reactions
can be used for the determination of individual sugar species
(for example, glucose oxidase-linked assays).
The following sections describe the specific methodological
issues in the measurement of individual fractions of
available carbohydrates and resistant carbohydrates.
Determination of available carbohydrates
Sugars. The common available sugars are glucose, fructose,
galactose, maltose, sucrose and lactose. Aqueous extraction
of these sugars is readily achieved from disrupted food
matrices by a short period of heating and mixing. The
solubility of sugars and the insolubility of proteins and
polysaccharides in 80% ethanol can be used to further isolate
them. When several sugars are present in a sample,
quantification of the individual mono- and di-saccharides
is best achieved by chromatography.
Maltodextrins. The short-chain a-glucan maltodextrins
occur naturally in plants in only small amounts, but can
be manufactured from starch by hydrolysis with acid, heat or
enzymes. Analytically, maltodextrins are usually included
within the total starch value, but a separate value can be
obtained by measuring the glucose released by amyloglucosidase
(EC 3.2.1.3) in the supernatant fraction of an 80%
aqueous ethanol extraction, with a correction for glucose
and maltose content.
Starch and starch digestibility. Starch, the storage polysaccharide
of many plants, consists of a 1–4 linked glucose
monomers and occurs as linear polymers (amylose) or as
macromolecules of shorter chains with a-1–6 branch
linkages (amylopectin). Historically, starch has been determined
by approaches including polarimetry and the formation
of starch–iodine complexes, but their use is generally
considered inappropriate for complex food systems. Therefore,
for quantitative determination, starch should be
hydrolysed and measured as the component glucose
monosaccharide units released, applying a 0.9 hydration
factor to convert them to the polysaccharide on a weight
basis. This is achieved most conveniently with a combination
of amylolytic enzymes (typically amylase (EC 3.2.1.2)
and amyloglucosidase (EC 3.2.1.3)) that hydrolyse the starch
polymer, including the a-1–6 branch linkages, and the final
cleavage of maltose and isomaltose to glucose. Procedures for
the accurate measurement of total starch need to include a

Table 2 Principles of carbohydrate measurement
Carbohydrate Types and dietary occurrence Extraction and isolation Hydrolysis and quantification
Sugars The main dietary sources are
Fruit and vegetables: sucrose, glucose,
fructose
Milk and dairy: lactose, galactose
Added sugars: mainly sucrose, glucose,
fructose
Starch Starch: consists of a-1–4-, a-1–6-linked
glucose, as amylose (short linear chains)
and amylopectin (larger, more branched
molecules). Principle sources are cereal
grain, legumes and tubers.
Maltodextrins: Mainly from hydrolysed
starch.
RS: defined as ‘the starch and starch
degradation products that on average resist
digestion in the small intestine’.
Starchy foods have a range of physicochemical
characteristics, influencing their
rate and extent of digestion.
NSP NSP are a grouping of several types of
polysaccharide that do not have the a 1–4
glucosidic linkage characteristic of starch.
Intrinsic plant cell-wall NSP: this component
is defined as Dietary Fibre as it consistent
reflect the health benefits of plant-rich
diets.
Other NSP: usually added extracts.
RSCC Fructans: Natural inulin (e.g., onions) and
fructooligosaccarides from hydrolysed
inulin or synthesized.
a-Galactosides: Sucrose with galactose
units, raffinose ( þ 1), stachyose ( þ 2),
verbascose ( þ 3).
Other RSCC: Mainly manufactured by
synthesis or by polysaccharide hydrolysis.
Examples are galacto- and xylooligosaccharides,
resistant maltodextrins,
polydextrose.
Polyols (sugar alcohols) Occur naturally in small amounts.
Manufactured and added to foods.
heating step at 1001 C for the gelatinization of starch
granules and treatment with either dimethyl sulphoxide or
sodium/potassium hydroxide for the dispersion of RS
(Englyst et al., 1982, 1992).
It is the physico-chemical characteristics that make the
enzymatic digestion of starch nutritionally interesting, and
more challenging to deal with appropriately from an
analytical perspective. The rate and extent that starch is
hydrolysed is determined by the accessibility of the amylolytic
enzymes, which explains why food processing, sample
Nutritional characterization and measurement of dietary carbohydrates
K Englyst et al
Recovered by aqueous extraction and
can be further isolated from other
macronutrients in an 80% ethanol
fraction.
Total starch including maltodextrins:
Majority dispersed in aqueous
conditions. Some high amylose and
retrograded starch requires chemical
dispersion.
Maltodextrins: Conveniently isolated as
the a-glucans (DP43) soluble in 80%
ethanol.
RS: isolated from samples prepared ‘as
eaten’ with digestible starch removed
by conditions that correlate with in vivo
data.
For all starch fractions glucose from
sugars must be accounted for.
NSP is isolated by the dispersion and
enzymatic hydrolysis of starch, which is
then removed along with sugars by
precipitating the NSP in 80% ethanol.
Fructans: Aqueous extraction. Need to
account for glucose and fructose,
including that from sucrose.
a-Galactosides: Aqueous extraction.
If determined as monosaccharide
components need to account for
glucose, fructose and galactose.
Other RSCC: Aqueous extraction and
isolation from polysaccharides in an
80% ethanol fraction. Sugars, sugar
alcohols, maltodextrins, fructans and
raffinose family must be accounted for.
Polyols are easily extracted in aqueous
conditions.
Monosaccharides and disaccharides can
be determined specifically by
chromatography.
Enzyme linked colorimetric assays can
measure individual monosaccharides.
Starch and its components are
determined as glucose released by
enzymatic hydrolysis applying a
hydration factor of 0.9.
Total starch including maltodextrins:
Measured as the sum of glucose
released following complete dispersion
of starch
Maltodextrins: Measured as glucose
released from the a-glucans (DP43)
soluble in an 80% ethanol.
RS: measured as glucose released from
RS fraction following physical and alkali
dispersion.
Following acid hydrolysis NSP
constituent sugars are determined
individually by chromatography or as a
total by a colorimetric assay.
Dietary fibre: For the majority of
products total NSP provides a measure
of dietary fibre
Other NSP: Should be accounted for
separately from dietary fibre. Identified
by NSP sugar profiles.
Fructans: Hydrolysed by fructanase and
measured as fructose (and glucose)
components.
a-galactosides: Determined specifically
by chromatography or alternatively
hydrolysed enzymatically and measured
as their components.
Other RSCC: Determined as the
monosaccharide components released
by enzymatic or acid hydrolysis. Intact
species can be determined by
chromatography, but is less specific.
Polyols can be determined directly by
chromatography.
Abbreviations: DP, degree of polymerization; NSP, non-starch polysaccharides; RS, resistant starch; RSCC, resistant short-chain carbohydrates.
preparation and analytical methodology can all influence
various aspects of starch determination.
The influence of the physico-chemical characteristics of
foods on starch digestibility can be described by measuring
the rate and extent of glucose released by amylolytic
enzymes under in vitro conditions controlled for pH,
temperature, viscosity and mixing (Englyst et al., 1992).
The terms rapidly available glucose and slowly available
glucose are used to describe rate of release characteristics and
their physiological relevance has been confirmed by the
S23
European Journal of Clinical Nutrition

S24
Table 3 Carbohydrate digestibility fractions for a selection of foods
Food g/100 g as eaten
SAG in %
Fru RAG SAG RS available CHO
Long grain rice (brown) 0.0 17.2 12.5 1.7 42.2
Rice pudding (canned) 2.5 10.9 0.2 0.2 1.7
Spaghetti 0.2 17.4 13.0 1.8 42.4
Wholemeal spaghetti 0.3 18.1 9.4 0.9 33.8
Brown bread 0.3 41.9 1.3 2.3 3.0
Wholemeal bread 0.4 35.5 1.5 1.9 3.9
Rye bread 1.7 27.5 5.1 2.8 15.0
Swiss roll 16.4 41.5 1.0 1.5 1.7
Crispbread-rye 1.6 59.5 5.4 2.5 8.1
Oatcakes 0.7 47.8 8.4 2.6 14.7
Water biscuit crackers 0.9 72.7 5.1 0.2 6.5
Digestive biscuits 4.8 44.2 11.4 1.1 18.9
Shortbread 8.2 31.1 21.9 1.8 35.7
Corn flakes 3.8 81.7 2.5 3.6 2.8
Muesli 17.1 38.6 4.7 1.3 7.8
Shredded wheat 1.2 68.7 2.8 2.4 3.9
Potato salad 1.7 8.0 2.1 1.0 17.8
Boiled potatoes 0.6 13.9 0.4 0.5 2.6
Potato crisps 0.2 50.7 1.8 1.0 3.4
Chick peas 0.5 4.2 11.7 3.5 71.0
Green split peas 0.5 6.9 8.7 3.4 54.2
Red kidney beans 0.8 6.4 8.3 3.4 53.3
Haricot beans 0.4 4.6 7.9 3.6 61.1
Baked beans 5.7 11.9 1.8 1.8 9.5
Sweetcorn 1.5 13.2 4.0 0.6 21.4
demonstration of strong correlations between their content
in foods with glycaemic response and GI values (Englyst
et al., 1999, 2003). Table 3 shows the carbohydrate-release
characteristics for a range of products.
Determination of resistant carbohydrates
Polyols (sugar alcohols). The polyols are hydrogenated
carbohydrates, which include sorbitol, mannitol, xylitol
and maltitol. Typically, there is only a small intake of
naturally occurring polyols, principally in the form of
sorbitol from apples and pears. However, the various types
of polyols can be manufactured and are used as sugar
replacers, which can be present in large amounts in some
products, particularly ‘sugar-free’ confectionery. Different
polyol types can be absorbed and metabolized to varying
extents, although a proportion of most polyols enters the
large intestine as fermentation substrates (Livesey, 2003). As
it is not practical to encompass the varied fate of polyols
within a classification scheme, polyols are usually considered
within the resistant carbohydrate grouping.
The sugar alcohols are easily extracted in aqueous or
ethanol fractions and can be measured directly by HPLC.
They are more stable than sugars in alkali conditions, a
feature that can be utilized to isolate and measure sugar
alcohols specifically by gas chromatography without interference
from sugars (Quigley et al., 1999).
European Journal of Clinical Nutrition
Nutritional characterization and measurement of dietary carbohydrates
K Englyst et al
Resistant short-chain carbohydrates. This fraction encompasses
all fructans, and the resistant carbohydrates that are
soluble in 80% ethanol, other than the sugar alcohols. The
resistant short-chain carbohydrate term was developed in
order not to be constricted by the chemical definition of
oligosaccharides as degree of polymerization (DP) 3–9, as
molecules of considerably higher DP may be included,
depending on branching (Englyst and Hudson, 1996;
Quigley et al., 1999). The terms non-digestible oligosaccharides
and resistant oligosaccharides are synonymous with the
RSCC term and in practice all describe the same substances.
This diverse group of substances are typically consumed in
only small amounts from naturally occurring sources,
principally as fructans present in onions, Jerusalem artichoke,
wheat and chicory, and as the small a-galactosides
(raffinose family) from legumes, with more complex galactooligosacchides
found in breast milk. A range of resistant
oligosaccharides have also been developed as functional
ingredients. There is a need to have a consistent approach to
the determination of all RSCCs, so that their dietary content
and relevance can be identified, and to prevent the potential
for gaps in carbohydrate classification and measurement.
There are several approaches to the determination of RSCC
depending on the chemical characteristics of the individual
species.
It is convenient to measure fructans as their fructose and
glucose constituents after specific hydrolysis with fructanase
(EC 3.2.1.7) (McCleary et al., 2000). The a-galactosides can
either be determined individually by chromatography or by
their monosaccharide constituents after enzymatic hydrolysis
with a-galactosidase (EC 3.2.1.22) and a-glucosidase (EC
3.2.1.20) (Vinjamoori et al., 2004).
Other RSCC can be measured by a single method based
on the acid hydrolysis of RSCC isolated in an 80% ethanol
extract and determination of constituent sugars by gas
chromatography (Quigley et al., 1999). This approach is
suitable for the determination of a wide range of substances,
including, but not restricted to isomalto-oligosaccharides,
xylo-oligosaccharides, galacto-oligosaccharides and resistant
maltodextrins.
There are also methods for the determination of RSCC by
the chromatographic separation of intact species. However,
as discussed later, this type of method can lack specificity
and there is the possibility of overlap between different
methods, which would lead to double counting and an
overestimate of the total RSCC content.
Non-starch polysaccharides. This group of carbohydrates is
defined as the polysaccharides that do not contain the a-1–4linked
glucose that is characteristic of starch. There are
various types of non-starch polysaccharides (NSP) that differ
in their sugar composition and glycosidic linkages, which are
important features in determining their physico-chemical
properties. The NSP present in plant cell walls have a
structural function in defining the integrity of plant cells
and tissues. NSP can also occur as gums and mucilages, some

Table 4 NSP in a selection of foods
of which are extracted and used as food additives for their
technological properties. Biological variation between plant
species means that different food groups have characteristic
profiles of NSP, as identified by their spectrum of constituent
sugars (Table 4). The NSP glucose is present in all food types,
occurring mainly in the form of cellulose, but some cereal
products, such as oats and barley, can contain considerable
amounts of the more soluble b-glucan. Galacturonic acid is
the main component of pectin found in fruit and vegetables.
Xylose is found predominantly as arabinoxylans in cereals.
Arabinose, mannose and galactose are present in all food
types, with rhamnose and fucose being present in only small
amounts in some fruits and vegetables.
As NSP is a chemically defined substance, its occurrence in
foods can be measured directly and specifically. This is
achieved by the extraction and isolation of these polysaccharides
from other carbohydrate components, with their
subsequent hydrolysis to constituent sugars for colorimetric
or chromatographic determination. Some individual NSP
such as b-glucans, can be determined as their constituent
sugars released by hydrolysis with the relevant substratespecific
enzymes.
Currently, the most convenient determination of total NSP
is achieved by hydrolysing the extracted NSP by an acid
treatment with measurement of the released monosaccharides
(Englyst et al., 1982, 1994). Extraction involves the
complete dispersal of starch with dimethylsulphoxide and
aqueous gelatinization, making it susceptible to hydrolysis
Nutritional characterization and measurement of dietary carbohydrates
K Englyst et al
Food NSP g/100 g as eaten NSP constituent sugars
Rha Fuc Ara Xyl Man Gal Glu Uronic
acid
Bran, wheat 37.2 0.0 0.0 8.8 16.8 0.1 0.6 9.8 1.1
Bread, wholemeal 5.2 0.0 0.0 1.5 2.0 0.1 0.2 1.3 0.2
Bread, white 1.6 0.0 0.0 0.5 0.7 0.1 0.1 0.3 0.0
Bread, Rye 7.3 0.0 0.0 1.9 3.2 0.1 0.2 1.9 0.1
Cornflakes 0.9 0.0 0.0 0.1 0.3 0.0 0.0 0.4 0.1
Meal, oats 7.0 0.0 0.0 0.8 1.1 0.0 0.1 4.9 0.2
Spaghetti, White 1.2 0.0 0.0 0.4 0.5 0.0 0.1 0.2 0.0
Spaghetti, Wholewheat 3.5 0.0 0.0 1.0 1.3 0.1 0.1 1.0 0.1
Beans, French 3.1 0.0 0.0 0.2 0.2 0.1 0.4 1.2 0.9
Peas, Cooked 5.2 0.1 0.0 1.1 0.2 0.0 0.1 3.0 0.7
Potato 1.2 0.0 0.0 0.1 0.0 0.0 0.4 0.5 0.2
Cabbage, Winter 3.7 0.2 0.0 0.8 0.2 0.1 0.4 1.2 0.9
Cauliflower, Cooked 1.6 0.1 0.0 0.3 0.1 0.0 0.2 0.6 0.4
Broccoli 3.0 0.1 0.0 0.4 0.2 0.1 0.3 1.1 0.9
Carrots, Cooked 2.5 0.1 0.0 0.3 0.0 0.1 0.4 0.9 0.8
Tomato 1.1 0.0 0.0 0.1 0.1 0.1 0.1 0.5 0.3
Cucumber 0.5 0.0 0.0 0.0 0.0 0.0 0.1 0.2 0.1
Apple, Cox 2.0 0.1 0.0 0.3 0.1 0.1 0.1 0.7 0.6
Peach 1.5 0.0 0.0 0.3 0.1 0.0 0.1 0.5 0.5
Banana 3.3 0.1 0.0 0.4 1.0 0.1 0.2 0.4 1.2
Orange 2.1 0.0 0.0 0.3 0.1 0.1 0.3 0.5 0.9
Melon, Honeydew 0.5 0.0 0.0 0.0 0.0 0.0 0.0 0.3 0.2
Kiwi fruit 1.7 0.0 0.0 0.0 0.2 0.1 0.1 0.7 0.6
Hazelnuts 6.5 0.2 0.0 1.4 0.5 0.0 0.5 2.3 1.6
by a combination of amylolytic enzymes. The NSP is isolated
by precipitation in acidified 80% ethanol, which removes
hydrolysed starch and any sugars present in the sample. If
information on NSP solubility is required, then the insoluble
fraction can be extracted by precipitation in pH 7 phosphate
buffer, which together with a total NSP value can be used to
calculate the soluble fraction. While 2 M sulphuric acid is
sufficient for the hydrolysis of most types of NSP, an initial
12 M sulphuric acid step is needed for the hydrolysis of
cellulose, and indeed omitting this step can be used to
calculate the cellulose content of a sample. Pectin is difficult
to hydrolyse, requiring a subsequent incubation with
pectinase. The sugars released may be measured by gas
liquid chromatography or HPLC to obtain values for
individual monosaccharides. A single value for total sugars
may be obtained by colorimetric determination of reducing
groups (Englyst et al., 1994). Further issues relating to NSP
are considered in the sections on the definition and
measurement of dietary fibre.
Resistant starch. Starch digestion by endogenous enzymes is
a continuous process during passage through the small
intestine, which, for a number of different reasons may not
go to completion. Starch can escape digestion in the small
intestine because it is physically inaccessible, within the food
matrix (RS1), or within starch granules (RS2), or because it is
present as retrograded starch (RS3) produced during food
manufacture and preparation. For many foods, a small
S25
European Journal of Clinical Nutrition

S26
proportion of the starch present will be RS (typically 0–5% of
starch in most cereal products) although for some foods such
as legumes this is higher (typically 10–20% of starch for some
beans), and food processing influences the amount present.
In addition, starch can be chemically modified to a form that
is resistant to enzymatic digestion (sometimes termed RS4),
which includes starch that has been etherized, esterified or
cross-bonded. Various high RS preparations have become
available, through either plant breeding to increase amylose
contents, or by physical and chemical manufacturing
processes that increase the RS fraction of starches. As each
RS preparation has its own specific physico-chemical
characteristics that can influence the rate and site of
fermentation in the colon, it is appropriate that each
preparation should be considered on its own merits, and
the properties associated with one preparation cannot
necessarily be extrapolated to others.
RS is physiologically defined on the basis of in vivo studies.
Quantifying the amount of starch entering the human colon
presents considerable technical problems (Champ et al.,
2003). Hydrogen excreted in breath has been used as an
indicator of fermentation in the colon, but it is considered
to lack the sensitivity required for quantification. Intubation
has been used to sample digesta from the ileum, but this
technique is restricted to liquidized meals. Several studies
have used human ileostomy subjects as a model to
investigate the digestive physiology of the small intestine.
Carbohydrate analysis of the ileostomy effluent allows
determination of the amount of starch escaping digestion
in the small intestine, which was found to vary between
individuals by 720% around the mean (Englyst and
Cummings, 1985, 1986, 1987; Silvester et al., 1995). It is
from these studies that the currently accepted definition of
RS is derived as ‘the sum of starch and starch degradation
products that, on average, reach the human large intestine’
(Englyst et al., 1992). The challenge for in vitro methods that
attempt to quantify the RS content of foods is to reflect the
mean values obtained by the in vivo studies.
Except for the modified RS4, resistant starch is not
chemically different from starch digested and absorbed from
the small intestine although products high in amylose have
a higher propensity to form RS. The amounts of RS in foods is
largely dependent on the degree of food processing, which
can result in an increase or a decrease in the RS values from
those found in the raw product. Therefore, RS needs to be
measured in foods as they would normally be eaten, and
values cannot be derived by summing the RS contents of raw
ingredients, or indeed be measured in samples that have
undergone laboratory preparation (freeze-drying/milling)
before analysis, as this can influence the RS content.
Numerous RS methods have been proposed which have
been reviewed (Champ et al., 2003). The basis of methods
for the determination of RS is the measurement of starch
remaining unhydrolysed after a defined period of incubation
with amylolytic enzymes. Measurements of total RS must
include RS1, RS2 and RS3. However, due to inappropriate
European Journal of Clinical Nutrition
Nutritional characterization and measurement of dietary carbohydrates
K Englyst et al
sample preparation and gelatinization of starch during the
procedure, many methods measure only RS3, or only RS2
and RS3 fractions. In addition, few of the methods have been
developed and validated in conjunction with in vivo studies.
Consequently, highly variable RS values have been reported
for similar foods.
In summary, methods designed to investigate the rate and/
or the extent of starch digestion in the human gut should
incorporate the following principles; analysis of foods
prepared ‘as eaten’, reproducible disruption of the physical
structure of food, standardized conditions of amylolytic
hydrolysis, development and validation of methods in
conjunction with in vivo studies. The analytical procedure
for rapidly available glucose, slowly available glucose and RS
determination conforms to these principles, and has been
tuned to yield values for RS that match the mean proportion
of starch recovered in ileostomy studies for both single foods
and mixed meals (Englyst et al., 1992; Silvester et al., 1995).
Dietary fibre
The term dietary fibre has been applied by different
researchers and in varied disciplines within the field of
nutrition to describe a diverse range of substances. This has
resulted in often disparate interpretations of what is meant
by the term, a situation that has not been helped by the fact
that the phrase ‘dietary fibre’ does not in itself provide an
unambiguous description of what it consists of. At best,
‘dietary’ infers that this is a food component, and ‘fibre’
implies a fibrous, coarse or structural nature. It is agreed that
it was originally used as shorthand for plant cell walls and
that it was intended as a nutritional term describing a
potentially beneficial characteristic of the diet.
In the intervening period, a large number of definitions
have been proposed. Suggestions have included one or more
of the following characteristics: chemical identity; intrinsic
material occurring naturally in foods; resistance to digestion
in the small intestine; demonstration of a specific physiological
effect; and material recovered by a particular methodology.
Where inconsistencies between definitions have
emerged, these can be related directly to which inclusion
criteria have been applied. In striving for a usable approach,
it is important that any limitations associated with proposed
definitions are identified so that any potential conflict with
dietary guidelines and health issues can be assessed.
In essence, the current situation can be summarized by
two contrasting approaches. One firmly retains the link with
plant foods with the definition ‘dietary fibre consists of
intrinsic plant cell-wall polysaccharides’, which remains true
to the original concept. The other approach has ‘indigestibility
in the small intestine’ as its central feature and
encompasses a wider range of substances from diverse
sources. In November 2006, the Codex committee on
nutrition and foods for special dietary uses (CCNFSDU) was

asked to consider both the ‘plant-rich diet’ and the
‘indigestibility’ approaches to dietary fibre definition.
As the substances included within these two approaches
to definition is not the same, the potential public health
implications associated with each of them will also differ. Its
prominent position in guidelines and the associated health
messages has led to good consumer recognition of the
dietary fibre term. Therefore, dietary fibre is very much a
public health term, and a foremost consideration must be to
ensure that consumers can interpret dietary fibre values and
any associated nutrition claims in a manner that assists them
with informed choice in diet selection, and that does not present
the opportunity for the misrepresentation of products.
An important aspect of nutrition is the requirement to
describe clearly defined nutritional components. To this end,
methods of analysis are a secondary issue, where suitability
should be assessed on how well they measure the intended
component. This principle should apply to the definition
and measurement of every nutrient, but for dietary fibre
there has been an inappropriate emphasis on methodology,
to the extent that some proposed definitions have been
based on the material recovered by a particular analytical
procedure. This is an unacceptable situation with respect to
describing food composition. Analytical methods should be
‘fit for purpose’, which for dietary fibre can be assessed by
the following criteria: (1) whether the material described in
the respective stated aims of methods are suitable as a
measure of dietary fibre; (2) the degree to which the methods
actually measure the material described within their respective
stated aims.
The sections that follow evaluate the rationales behind
the ‘plant-rich diet’ and ‘indigestibility’ approaches to the
definition of dietary fibre, including discussion of their
perceived limitations (summarized in Table 5). This includes
an assessment of the methodological approaches available
for the determination of the substances included in each
definition, with a comparison of the main NSP and
enzymatic–gravimetric methods provided in Table 6.
The plant-rich diet approach to dietary fibre definition
Associated definition. By this approach, dietary fibre is a
characterized component of plant foods, providing a consistent
indicator of the minimally refined plant-rich diet
promoted by the food-based guidelines for dietary fibre
consumption. As part of the Food and Agriculture Organization/World
Health Organization scientific update on carbohydrates
in human nutrition, held in Geneva July 2006, it
was agreed that the definition of dietary fibre should
maintain this clear link to fruits, vegetables and whole grain
cereals and the following definition was subsequently
endorsed on behalf of the carbohydrate scientific update
for consideration by CCNFSDU.
‘Dietary fibre consists of intrinsic plant cell-wall polysaccharides’
This definition, together with its rationale and associated
measurement, was proposed at the Geneva meeting in the
Nutritional characterization and measurement of dietary carbohydrates
K Englyst et al
presentation relating to this paper, and is described in the
following sections. The definition cannot easily be misinterpreted,
as demonstrated when evaluating its component
parts:
‘Intrinsic’—This emphasizes that the health benefits of
plant-rich diets is not restricted just to the plant cell wall
(or its polysaccharide component), but may relate as well
to the overall profile of associated micronutrients and
phytochemicals.
‘Plant cell wall’—This identifies the food component of
interest and therefore specifies that it is the structural
polysaccharides of the plant cell wall that should be
determined.
‘Polysaccharides’—This establishes that dietary fibre is a
carbohydrate term, providing the required chemical
element that should form an essential part of any
definition.
Rationale and implications of the plant-rich diet approach. The
modern concept of dietary fibre as a protective nutritional
component has stemmed largely from observations that
diets rich in unrefined plant foods were associated with a
lower incidence of certain diseases including diverticular
disease, colon cancer and diabetes (Burkitt, 1969; Trowell,
1972; Trowell et al., 1985). When compared with their
refined counterparts, the most prominent identifying characteristic
of these foods and diets was the presence of largely
unprocessed plant cell-wall material, which is composed
predominantly of structural polysaccharides. The need to
distinguish this carbohydrate fraction on the basis that it did
not provide the same energy as starch and sugars was already
recognized (McCance and Lawrence, 1929). However, rather
than the issue of energy alone, it has been the prospect of an
association with more direct health benefits that has driven
the demand for a dietary fibre term. The prominent public
health status of dietary fibre and the positive nutritional
message it conveys is largely the result of the consistent
advice within dietary guidelines to increase consumption of
dietary fibre in the form of fruits, vegetables and whole
grains (Department of Health, 1991; WHO, 2003; USDA/
DHHS, 2005). Furthermore, the reference intake values and
nutrition claims relating to dietary fibre have been established
from the health benefits that have been associated
with the intake of these naturally high fibre foods. For
example, the current American reference intake values are
based mainly on three prospective studies on the association
with cardiovascular disease (Pietinen et al., 1996; Rimm et al.,
1996; Wolk et al., 1999; IOM, 2002).
There are important distinctions between advice that
specifies increased intake of dietary fibre from specific food
groups, as opposed to a solely nutrient-based approach. For
instance, their high water content means that fruits and
vegetables may not at first appear to be particularly good
sources of dietary fibre when they are considered in terms of
g/100 g as consumed. In fact, it is precisely this quantitatively
small amount of plant cell wall material that confers the
S27
European Journal of Clinical Nutrition

S28
Table 5 Comparison of the plant-rich diet and indigestibility approaches to the definition of dietary fibre
Plant-rich diet approach Indigestibility approach
Definition: ‘Intrinsic plant cell-wall polysaccharides’. Definition: ‘Indigestible carbohydrate (DP43) and lignin.’
Rationale: This definition is targeted specifically at the fruits, vegetables
and whole grain products that are consistently linked with health
benefits.
These foods have the characteristic feature of containing plant cell walls,
which mainly consist of structural polysaccharides, which can be
quantified in chemical terms. Other non-carbohydrate components are
not included as they can neither be determined specifically nor would
their inclusion enhance the definition as an indicator of these foods.
The definition recognizes that the benefits of a natural fibre-rich diet are
not due to any single component, but rather the effect of synergistic
elements including micronutrients, phytochemicals and low energy
density.
Scientific evidence for rationale: This is a food-based rationale, which is
strongly supported by the epidemiological evidence for the health
benefits of fruits, vegetables and whole grain products.
Retaining a distinct dietary fibre term identifying plant-rich diets with
their unique health benefits reinforces the food-based dietary guidelines.
This distinction allows the properties of other resistant carbohydrates to
be researched and if appropriate promoted in their own right.
Nutrition labelling: A dietary fibre value describing intrinsic plant cell-wall
polysaccharides would guide consumers to the selection of plant-rich
foods.
If other sources of resistant carbohydrates are present, then there would
be scope for these to be labelled specifically.
Nutrition and health claims: The claims for dietary fibre are largely based
on the epidemiological evidence, which relates to fibre from plant-rich
diets.
When appropriate, specific health claims should be established for
individual resistant carbohydrate functional ingredients, thereby
acknowledging their specific properties and taking account of variations
in their effective and safe dosages.
Population reference intakes: The population reference intake values for
dietary fibre are largely based on the epidemiological evidence that
minimally refined plant-rich diets are associated with a lower incidence of
several diseases.
The intrinsic plant cell-wall polysaccharide definition ensures that dietary
fibre intakes contributing towards the reference value would consistently
reflect both the epidemiological evidence and the intended message of
the dietary guidelines.
Impact on food industry: Although values for ‘intrinsic plant cell-wall
polysaccharides’ are generally lower than those for the indigestibility
approach, this should not make a difference to the marketing of the
majority of products, as population reference intakes and claims would
be based on the plant-based approach.
The emphasis would be on manufacturers to incorporate minimally
refined plant ingredients into products to achieve nutrition claims for
dietary fibre.
There would be a positive opportunity to market other types of resistant
carbohydrates with respect to their specific functional properties.
For food labelling purposes, there would be cost savings with the analysis
of NSP compared to the enzymatic-gravimetric and supplementary
analysis.
Impact on nutrition research: Maintaining ‘intrinsic plant cell-wall
polysaccharides’ as a distinct definition of dietary fibre facilitates research
European Journal of Clinical Nutrition
Nutritional characterization and measurement of dietary carbohydrates
K Englyst et al
Rationale: There are numerous versions of this definition, which have the
common feature of placing the emphasis on escaping digestion in the
small intestine. The definition is not restricted to carbohydrates as it
encompasses lignin and other substances associated with the plant cell
wall.
In addition to the plant cell-wall polysaccharides, the indigestibility
criterion has the implication of including RS and other extracted or
synthesized carbohydrates, including resistant oligosaccharides.
However, as this grouping can include a wide range of substances it has
been suggested that there should also be a demonstrated physiological
effect for a specific material to be included.
Scientific evidence for rationale: For the existing epidemiological evidence
relating to the last few decades this definition provides a reasonable
indicator of plant-rich diets, as supplementation with resistant
carbohydrate preparations was uncommon. However, this is not always
the case for manufactured products developed recently.
Specific physiological properties have been associated with individual
supplements, but these vary depending on type, making it difficult to
consider them within a single definition. The long term health effects/
safety remains to be established.
Nutrition labelling: By the indigestibility approach the fibre label would
not provide a consistent indicator of plant-rich foods that may mislead
consumers who have this expectation. By grouping all indigestible
carbohydrates within a single undifferentiated nutrition label, there is less
opportunity to identify any added functional ingredients.
Nutrition and health claims: The epidemiological evidence for dietary fibre
cannot be extrapolated to a definition that includes enzymaticgravimetric
values of unknown composition, as well as a range of
supplemented materials with varied functional properties. There is the
potential for inappropriate nutrition claims for materials with either no
effect or detrimental properties, which would undermine the position of
dietary fibre as a beneficial food component.
Population reference intakes: The use of this definition could result in a
situation where the consumer selects supplemented products on the
basis that they will contribute towards the reference intake value,
although in reality this would not be a true reflection of the intention of
the dietary guidelines. This raises two concerns (1) that the
supplemented product is unjustly promoted on the back of the
epidemiological evidence; and (2) that if direct substitution of products
occurs, then the consumption of the intended target of plant-rich food
groups may be diminished.
Impact on food industry: With this definition, there would be less impetus
for the manufacturer to incorporate unrefined plant ingredients, as it
would be possible to elevate the dietary fibre content through processing
or supplementation instead. However, it would be difficult for the
consumer to distinguish between these different types of product if they
carried identical nutrition claims. This may be perceived as conflicting
with the intended aim of reference intake values and dietary guidelines
which are targeted at plant-rich diets.
Grouping varied functional ingredients together limits the opportunities
for manufacturers to promote the specific properties of individual
products. As gravimetric values are influenced by food processing, food
labelling cannot be based on food table values of component
ingredients.
Impact on nutrition research: The indigestibility approach groups diverse
substances including plant cell-wall material, retrograded starch,

Table 5 Continued
Plant-rich diet approach Indigestibility approach
into the benefits of plant-rich diets, and encourages specific research into
types of resistant carbohydrate preparations.
Only with detailed information on distinct substances will it be possible
for future epidemiological studies to establish the intakes and effects of
different types of resistant carbohydrates.
DP, degree of polymerisation.
supplements and non-carbohydrate artifacts in unknown proportions.
This single undifferentiated grouping will not provide the detailed
information required by future epidemiology studies to establish the
intakes and health effects of different types of resistant carbohydrates.
Nutrition research is better served by detailed information on specific
food components.
Table 6 Comparison of the principles and analytical issues relating to the principal methods associated with the plant-rich diet approach (NSP method)
and the indigestibility approach (enzymatic–gravimetric methods) to the definition of dietary fibre
NSP Method Enzymatic–gravimetric methods (AOAC 985.29 & 991.43)
General principles General principles
Stated aim: To measure polysaccharides that do not contain the a 1–4
glucosidic linkages characteristic of starch (i.e., NSP)
Analytical principle: Complete dispersion and enzymatic hydrolysis of
starch.
Precipitate residue in 80% ethanol and isolate by centrifugation.
Hydrolyse and measure NSP as constituent sugars by colorimetry or
chromatography.
Information provided: Values for total, soluble and insoluble NSP, with the
option of detailed information on constituent sugars by the GC version.
Effect of food processing: As a chemically distinct food component, NSP is
minimally affected by normal food processing.
Is stated aim achieved: Yes. The procedure completely removes starch and
sugars and provides a specific determination of NSP.
Analytical issues Analytical issues
Specific reagents and equipment:
Enzymes: Heat stable amylase, (EC 3.2.1.1), pullulanase (EC 3.2.1.41),
pancreatin (these enzymes should be devoid of NSP hydrolytic activities),
pectinase (EC 3.2.1.15).
Analysis vessels: screw cap test tubes.
Equipment: Centrifuge and either spectrophotometer or GC system.
Practical issues: All the steps of this procedure are conducted in test tubes,
which makes it well suited to the analysis of large batch sizes.
It is important to ensure complete starch dispersion and hydrolysis,
which is achieved by a combination of physical, chemical and enzymatic
steps.
The chemical end-point determination techniques are those used in the
measurement of other carbohydrates (e.g., sugars, starch).
The procedure takes 1 day with colorimetric measure or 1.5 days for
GC measure.
Environmental impact: Only small amounts of solvent waste generated.
Suitability for use in developing countries: The NSP procedure only requires
standard laboratory equipment including a spectrophotometer for the
colorimetric version.
Traceability: The primary standard is a representative mixture of the
individual monosaccharides of NSP.
Method specificity: Only NSP is measured, with no interference from other
substances.
Method reproducibility: A range of certified reference materials are
available (e.g., BCR). Method CV o5%.
Nutritional characterization and measurement of dietary carbohydrates
K Englyst et al
Stated aim: To measure the sum of indigestible polysaccharides and
lignin.
Analytical principle: Partial enzymatic hydrolysis of starch and protein.
Precipitate residue in 80% ethanol and isolate by filtration.
Record total residue weight and then determine and subtract ash and
protein contents.
Information provided: Weight of total, soluble and insoluble residue
containing carbohydrate and non-carbohydrate material in unknown
proportions.
Effect of food processing: A range of materials are recovered in the residue,
which is highly dependent on food processing (e.g., retrograded starch,
Malliard reaction products).
Is stated aim achieved: No, not consistently. In addition to NSP, this
procedure measures a variable amount of RS, which may not relate to the
true extent of physiological starch digestion. In addition to lignin, the
non-carbohydrate part can include food processing artifacts.
Specific reagents and equipment:
Enzymes: Heat stable amylase, (EC 3.2.1.1), protease, amyloglucosidase
(EC 3.2.1.3). These enzymes should be devoid of NSP hydrolytic
activities.
Analysis vessels: 400 ml beakers and fritted glass crucibles.
Equipment: vacuum manifold, muffle furnace and Kjeldahl equipment.
Practical issues: Batch sizes are limited by the difficulties of handling large
numbers of 400 ml beakers.
The selective removal of starch other than RS is difficult or impossible to
achieve within this procedure.
The method is labour intensive due to: preparation and repeated
weighing of the crucibles; numerous pH checks; manual transfer and
filtration of residues; subsidiary ash and Kjeldahl methods.
The procedure takes 1.5–2 days or more with longer filtration times.
Environmental impact: Large amounts of solvent waste are generated.
Suitability for use in developing countries: The gravimetric procedure
requires specialist glassware, muffle furnace and Kjeldahl equipment for
the measurement of nitrogen.
Traceability: No primary standard is available as the procedure does not
measure a chemically distinct component.
Method specificity: Any added material or food processing artefacts
recovered in the residue are a potential source of interference.
Method reproducibility: A range of certified reference materials are
available (e.g., BCR). Method CV o5%.
S29
European Journal of Clinical Nutrition

S30
Table 6 Continued
NSP Method Enzymatic–gravimetric methods (AOAC 985.29 & 991.43)
Suitability as a measure of dietary fibre Suitability as a measure of dietary fibre
Potential discrepancies with definitions: For plant foods, the NSP content is
a measure of ‘intrinsic plant cell-wall polysaccharides’.
In a few plants NSP can occur as gums and alginates, but these are not
typical foods and are more likely to occur as ingredient extracts.
When extracted or synthesized NSP are present in products then these
will be known by the manufacturer and can be deducted from the NSP
measurement to obtain a value for the intrinsic plant cell-wall
polysaccharides. The presence of specific extracts can often be identified
by their NSP constituent sugar profile.
With the plant cell-wall polysaccharide definition, resistant
oligosaccharides and RS are separate groupings. Their content in foods is
measured specifically and they do not conflict with the NSP
measurement.
Evaluation of method: The intrinsic plant cell-wall polysaccharide
definition provides a clear link to the plant-rich diet shown to be
beneficial to health. The NSP procedure provides measurements that are
suitable for this definition.
Abbreviation: GC, gas chromatography.
high water-holding ability of fruit and vegetables, which in
turn is responsible for their low energy density. In addition,
the plant cell walls have a central role in defining the high
nutrient density with respect to vitamins, minerals and
phytochemicals, which are considered as closely associated
companion nutrients. These are unique nutritional properties
associated with the dietary fibre in these food groups, and
therefore food-based guidelines for dietary fibre are always
applicable even if, as is the case with fruits and vegetables,
their contribution to dietary fibre intake is often modest
compared with that derived from whole grain products.
The benefits of plant cell-wall-rich foods is supported by
prospective observational studies that identified significant
inverse relationships between intake of fruits, vegetables
and whole grains and incidence of cardiovascular disease,
diabetes and some cancers (Jacobs et al., 1998; Liu et al.,
1999, 2003; van Dam et al., 2002; Bazzano et al., 2003, 2005;
Rissanen et al., 2003; Slavin, 2003; Steffen et al., 2003; WHO,
2003). To ensure consistency for the public health message
being conveyed, it is essential that any measure of dietary
fibre is a true representation of the unrefined plant foodbased
diet endorsed by the epidemiological evidence and
dietary guidelines.
The nutritional relevance of ‘intrinsic plant cell wall
polysaccharides’ can be considered at various levels (1) as a
distinct carbohydrate component of the food, (2) as a
provider of cell wall structures and (3) as a marker of a diet
rich in micronutrients. When present as an intrinsic part of
plant foods, these elements connected with cell-wall polysaccharides
cannot be disassociated from one another, with
the implication that it is not possible to assign the benefits of
dietary fibre-rich diets to just one of these attributes. In other
words, the proposal to define dietary fibre as the ‘intrinsic
plant cell-wall polysaccharides’ is based on the fact that this
European Journal of Clinical Nutrition
Nutritional characterization and measurement of dietary carbohydrates
K Englyst et al
Potential discrepancies with definitions: As the AOAC gravimetric
procedure measures a range of indigestible materials of varied
composition and origin it does not provide a consistent measure of plant
cell-wall material.
It can include non-carbohydrate food processing artifacts (e.g., Maillard
reaction products) that are not part of any dietary fibre definition.
The residual starch recovered can be misleading, as it does not relate to
physiologically RS, for which separate measurement is required.
It does not recover resistant oligosaccharides, resistant maltodextrins or
all RS, and therefore by itself does not provide a measure of the
indigestible carbohydrates proposed for inclusion. These substances
require separate analysis if they are to be included.
Evaluation of method: The indigestible carbohydrate and lignin definition
does not consistently identify plant-rich diets. Neither does the AOAC
gravimetric procedure provide a consistent measurement of the material
included in this definition.
is the only component that is consistently associated with
the plant-rich diet linked with reduced disease incidence.
Determination of intrinsic plant cell-wall polysaccharides. This
approach describes a chemically defined food component
that can be determined by an enzymatic–chemical method.
The stated aim of this method is to measure the polysaccharides
that do not have the a-(1–4) glucosidic linkages
characteristic of starch. Therefore, the method is designed to
disperse and remove all starch, with NSP measured as the
sum of chemically identified NSP constituent sugars (Englyst
et al., 1994).
The enzymatic–chemical method for the analysis of NSP is
an extension of the pioneering work of McCance and
Lawrence (1929) and later of Southgate (1969), which
recognized the importance of the direct measurement of
the various types of carbohydrates for nutrition composition
purposes. NSP forms part of the unified scheme to classify
and measure all food carbohydrates (Table 1). To evaluate it
as a measure of dietary fibre, the NSP method is assessed here
in terms of its suitability to measure ‘intrinsic plant cell wall
polysaccharides’.
In typically consumed unsupplemented foods, the entire
NSP component will be derived from the intrinsic plant cell
wall. The advantage of NSP as a chemically distinct
substance is that it is not in itself created or destroyed by
normal food preparation or storage techniques, which
means that NSP can be used as a fairly consistent indicator
of plant cell-wall material. When added preparations of NSP
are present in foods, these too are measured as their
carbohydrate components, and will contribute to the total
NSP value. Manufacturers’ data on the amount and type of
carbohydrate preparation used will normally be sufficient to

account for any supplemented material present. However, for
the purpose of traceability and authenticity checks, it would
in most cases be possible to identify the presence of specific
preparations by their profile of constituent NSP sugars. For
example, the presence of guar gum in a product can be
identified by higher galactose and mannose compared with
the NSP sugar profile of the unsupplemented food.
Unfortunately, a lack of understanding of the practical
issues involved has led to inaccurate statements about the
complexity of the NSP method. The actual situation is that
the enzymatic–gravimetric method, promoted as part of the
‘indigestibility approach’, is more time consuming, resource
demanding and subsequently more expensive to perform
than the NSP procedure. The NSP method has been subjected
to successful collaborative trials (Wood et al., 1993;
Pendlington et al., 1996) and for routine purposes, including
food labelling, NSP can be determined by colorimetry with a
simple spectrophotometer. Furthermore, the NSP method
is well suited to the analysis of large batch sizes as it uses test
tubes as the reaction vessel, compared with cumbersome
400 ml beakers and filtration crucibles used in the enzymaticgravimetric
methods.
The indigestibility approach to dietary fibre definition
Associated definition. By this approach, the primary defining
characteristic is indigestibility in the small intestine, thereby
grouping together diverse substances. In addition to cell-wall
polysaccharides, such a grouping would include non-structural
carbohydrates that are normally absent, or present only
in small amounts in most foods (for example, inulin), and
in the case of RS is largely dependent on food processing.
Also included would be extracted, synthesized, or otherwise
manufactured polysaccharides and oligosaccharides that
could be added to individual foods in considerable amounts.
A proposed definition based on this approach has been
considered by CCNFSDU in the context of providing guidelines
for the use of nutrition claims. The proposed definition
states as follows:
Dietary fibre means carbohydrate polymers with a DP not
lower than 3, which are neither digested nor absorbed in the
small intestine. A DP 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 edible carbohydrate
polymers naturally occurring in the food as consumed,
carbohydrate polymers, which have been obtained from
food raw material by physical, enzymatic or chemical means,
synthetic carbohydrate polymers.
In addition, this definition is associated with a lengthy
footnote included to justify the use of specific enzymatic–
gravimetric methods, which are acknowledged to recover a
wide range of non-carbohydrate materials that would
otherwise fall outside the stated definition.
Also linked with the definition is a statement relating to
the physiological properties generally considered to be
Nutritional characterization and measurement of dietary carbohydrates
K Englyst et al
associated with dietary fibre and the recommendation that
‘where a declaration or claim is made with respect to dietary
fibre, a physiological effect should be scientifically demonstrated’,
with the exception of naturally occurring polymers
for which no such justifying criteria were deemed necessary.
With respect to the application of this definition, it goes on
to raise issues as to the food safety requirements, diverse
efficacies of different substances purporting to be dietary fibre,
and the consumer perception of fibre as being of plant origin.
Due to the diverse nature of the substances included, there
is no single analytical method currently available that will
provide an accurate and comprehensive determination of
the material encompassed within the indigestibility approach.
Instead, 10 methods of analysis are stated in
connection with this definition, with one of two versions
of an enzymatic–gravimetric technique considered to be the
principal method.
Rationale and implications of the indigestibility approach. As
evidenced by the length of the above definition and
associated conditions, the rationale for this indigestibility
approach is necessarily more complex as it tries to amalgamate
issues of food composition, analytical methodologies
and physiological attributes. The primary basis for the
indigestibility approach is the fundamental difference in
the physiological handling of carbohydrates depending on
their gastrointestinal fate.
However, although it may be possible to group diverse
substances by a shared attribute such as indigestibility, this
does not mean that such groupings should necessarily form
the basis for dietary advice. The relation between the
amount and type of fermentable substrate reaching the
colon and related physiological parameters is incompletely
understood, with both beneficial and potentially adverse
effects having been reported. There is insufficient evidence
to suggest that all sources of resistant carbohydrates should
be actively promoted, or that it would be desirable to set a
single population reference value for total resistant carbohydrate
intake. Nevertheless, this would in essence be the
prospect with the indigestibility based dietary fibre definition.
As the characteristic of ‘being neither digested nor
absorbed in the small intestine’ does not in itself equate to
a health benefit, it is implied that a further level of justifying
criteria are needed for functional ingredients to be considered
as dietary fibre by the indigestibility approach. This
therefore relies on an evidence base of specific physiological
properties being associated with individual substances.
Although a range of physiological parameters have been
investigated, it is not always clear to what extent these
translate into actual health benefits. What is apparent is the
diversity of the substances and their efficacies with respect
to physiological outcomes varies widely. For example, the
varied impact of different resistant carbohydrates on stool
weight and prebiotic effects are reviewed in the physiology
paper (Elia and Cummings, 2007).
S31
European Journal of Clinical Nutrition

S32
The common attribute of the substances included in the
indigestibility approach is that they provide potential
substrates for colonic fermentation, stimulating bacterial
growth and the production of SCFA, which have a range of
physiological effects (Macfarlane et al., 2006; Wong et al.,
2006). Different amounts and types of substrate vary in the
rate, site and extent of fermentation and the profile of SCFA
produced. Although butyrate has been proposed as protective
against colon cancer, the effects it has are complex and
somewhat contradictory (Sengupta et al., 2006). Some
studies have found butyrate providing substrates have had
adverse effects (Burn et al., 1996; Wacker et al., 2002), and it
seems that the amount, site, whether promixal or distal, and
underlying conditions all influence the effect of butyrate.
The desirability of providing large amounts of easily
fermented resistant carbohydrate substrates has been questioned
(Wasan and Goodlad, 1996; Goodlad, 2007), with
concern about the impact of a feast or famine scenario on
gut health. The epidemiological evidence base for a protective
effect of resistant carbohydrates against colon cancer has
been inconclusive (Bingham et al., 2003; Park et al., 2005),
and so far, intervention studies have tended to show either
no effect or a worsening in outcomes (Alberts et al., 2000;
Bonithon-Kopp et al., 2000). On reviewing the evidence Food
Drug Administration concluded that dietary fibre did not
protect against colon cancer (FDA, 2000).
It should be noted that the exclusion of DP o3 by the
indigestibility linked definition demonstrates a lack of
consistency, as resistant sugars such as lactulose and some
polyols share similar physiological attributes to those
suggested as the basis for characterizing carbohydrates with
DP 43 as dietary fibre, for instance prebiotic effects (Gostner
et al., 2006), and promoting calcium absorption (van den
Heuvel et al., 1999). In addition, by this approach it is not
clear how to handle carbohydrate ingredients that are
indigestible, but for which no beneficial physiological
attributes have been demonstrated.
While the addition of resistant carbohydrates to products
has no direct impact on the food-based guidelines to
consume dietary fibre as fruits, vegetables and whole grains,
there is a greater potential for confusion with respect to the
population reference intake values. For dietary fibre, these
have predominantly been derived from epidemiological
evidence linking plant-rich diets with reduced disease
incidence. However, unless supplemented foods are clearly
identified as such, then a potential conflict arises if the
consumer perceives that such preparations are directly
equivalent to the dietary fibre present in unsupplemented
foods. This could result in a situation where the consumer
selects supplemented products on the basis that they will
contribute towards the reference intake value, although in
reality this would not be a true reflection of the intention of
the dietary guidelines.
This argument does not preclude that some resistant
carbohydrate preparations cannot have a position within
diets, but it should be emphasized that such formulations are
European Journal of Clinical Nutrition
Nutritional characterization and measurement of dietary carbohydrates
K Englyst et al
researched and if shown beneficial promoted on the basis of
the usually very specific functional properties that they may
have. If resistant carbohydrate preparations were included as
dietary fibre, the population reference intake values established
by the epidemiology evidence base would become
redundant, as there would be no clear link between the food
label and the guideline.
It should also be considered that the bulking effect that
acts to self-limit the intake of foods with a naturally high
dietary fibre content, represents much less of a constraint for
some resistant starches and resistant oligosaccharides that
can be formulated within products in relatively high
amounts, and could make a considerable contribution to
some diets. This highlights the need to consider potential
detrimental effects and whether safe upper intake limits for
resistant carbohydrates are required.
Determination of substances included within the indigestibility
approach. The principle of determining carbohydrates
grouped on the basis of the physiological attribute of
indigestibility in the small intestine has already been
addressed within the earlier section on the measurement of
resistant carbohydrates. Although NSP constitutes the major
resistant carbohydrate fraction in most foods, this is typically
not determined specifically by this approach, but instead
forms part of an enzymatic–gravimetric measurement that
includes other materials in unknown amounts. The values
obtained by these enzymatic–gravimetric techniques have
previously been presented as ‘total dietary fibre’ measures,
but a number of supplementary methods have since been
proposed to determine other substances included within the
indigestibility approach. The principal enzymatic–gravimetric
techniques and the other complementing methods are
discussed in turn.
Enzymatic–gravimetric procedures. The stated aim of these
methods is to measure the sum of indigestible polysaccharides
and lignin as the weight of a residue, corrected for ash
and crude protein content, which remains after treatment
with protease and amylolytic enzymes and washing with
80% ethanol. These methods therefore do not focus on plant
cell-wall material, but seek to include RS, which may be
present in large amounts as the result of food processing or
the addition of RS preparations. Non-carbohydrate materials
are also recovered, with the given justification of measuring
lignin and other non-carbohydrate cell-wall components,
although in practice it also recovers food processing artefacts
such as Maillard reaction products, which may have adverse
physiological effects (Tuohy et al., 2006).
This approach has evolved from the early methods used
for measurement of ‘crude fibre’. The different versions of
the enzymatic–gravimetric technique can be considered as
modifications of the method proposed by Prosky and coworkers
(AOAC method 985.29; AOAC, 2005). The most
common variation is AOAC method 991.43, which uses a
different buffer system. Briefly, they utilize 400 ml beakers as

eaction vessels, with successive treatments with amylase,
protease and amyloglucosidase, each step requiring individual
pH adjustments. Four volumes of ethanol are added
and the precipitated material is transferred to a filtration
crucible where it is dried and weighed. For each sample
separate residues are collected for determination of ash and
protein.
Contrary to normal conventions of nutrient definition,
the material recovered by these methods has been presented
by some as a ‘de facto’ definition of dietary fibre (AACC,
2001). Of the various materials that may be determined, only
the intrinsic cell-wall polysaccharides are a consistent feature
of natural unrefined plant foods, and for many plant-based
products this will be the main component of the gravimetric
fibre value. However, as this methodological approach can
include some RS and other substances formed as the result of
food processing or during sample preparation for analysis,
it is not possible to identify how much, if any, of the ‘fibre’
value is plant cell wall material (Ranhotra et al., 1991;
Theander and Westerlund, 1993; Rabe, 1999). A detailed
assessment of the influence of food processing on materials
recovered by enzymatic–gravimetric approaches has been
provided elsewhere (Englyst et al., 1996).
This method cannot be considered as ‘fit for purpose’ in
meeting the requirement to consistently reflect natural
unrefined plant foods. Neither does it meet its own stated
aim of measuring indigestible polysaccharides, as the RS
recovered in the residue may have little bearing on what is
present in the food. Lignin should be excluded from further
consideration as part of a dietary fibre measure on the
grounds that (1) it is not a carbohydrate, (2) it is not present
in the human diet in significant amounts, (3) there is no
specific routine method for is analysis, (4) its inclusion has
often inappropriately been used to justify the presence of
unidentified material in the gravimetric fibre residue.
In terms of practicality for the analytical chemist, the
enzymatic–gravimetric approach is excessively cumbersome.
It requires considerable time and reagent resources and is not
well suited to large batch sizes, increasing the cost of this
analysis.
Complementary procedures for the indigestibility approach. The
intended purpose of the other stated methods associated
with this approach are to give additional information about
individual resistant carbohydrate fractions and provide
determinations of those substances that are incompletely
recovered by precipitation in 78% ethanol with the enzymatic–gravimetric
techniques. Several of these methods
determine carbohydrates as their constituent sugar components
released by hydrolysis, along the principles described
in the carbohydrate determination section.
Similar to the NSP procedure outlined earlier, the AOAC
994.13. method is primarily based on the determination of
sugars released by the acid hydrolysis of polysaccharides
isolated by precipitation in ethanol. It differs in that by this
technique starch is only partially dispersed and hydrolysed,
Nutritional characterization and measurement of dietary carbohydrates
K Englyst et al
so unlike the NSP method it does not describe a chemically
distinct grouping of carbohydrates. It also includes a
determination of Klason lignin as the material recovered in
an acid hydrolysis resistant residue. The reality is that Klason
lignin can include a considerable amount of artifact material
including Maillard reaction products formed during food
processing.
AOAC 2002.02 is a resistant starch method based on
treatment with amylolytic enzymes and precipitation of
unhydrolysed starch in 80% ethanol, which is then chemically
dispersed and determined as glucose released by
hydrolysis. As discussed in the resistant carbohydrate
determination section, the degree of starch hydrolysis is
influenced by the analytical conditions, and for this method
these have principally been designed only to determine
retrograded starch (RS3) and some RS2 in starch granules.
Therefore, this method does not consistently provide a total
RS determination, and neither does it measure the same
starch fraction recovered by the enzymatic–gravimetric
methods, making it difficult to integrate the values obtained
by these methods. Furthermore, small starch degradation
products resulting from the enzyme hydrolysis could
potentially be lost in the 80% ethanolic supernatant, and
therefore would not be included as RS.
The AOAC 995.16 method determines b-glucans and is an
example of the measurement of an individual NSP species
by selective enzymatic hydrolysis. The AOAC 999.03 and
997.08 methods determine fructans (inulin and fructooligosaccharides)
as the fructose (and small amount of
glucose) released after fructanase treatment. AOAC 997.08
uses HPLC to measure the increase in released sugars on top
of the sugars already released by hydrolysis of sucrose and
starch, leading to a high uncertainty when dealing with
small quantities of fructans. Although AOAC 999.03
attempts to address this issue by the removal of sugars by
chemical reduction prior to the fructan hydrolysis, this
approach results in an incomplete recovery of lower DP
fructooligosaccharides as their reducing end groups are also
affected by the reduction step. Along similar principles, the
AOAC 2001.02 method determines trans-galactooligosaccharides
in an aqueous extract as the galactose released
after treatment with b-galactosidase (EC 3.2.1.23), with a
separate measurement and correction for galactose from
lactose, which is also hydrolysed by this enzyme.
The stated methods for the determination of polydextrose
(AOAC 2000.11) and resistant maltodextrin (2001.03) do not
measure their component glucose parts, but instead rely
on quantification of intact oligosaccahrides by chromatography.
AOAC 2000.11 is based on an aqueous extraction
treated to hydrolyze those a 1–4 bonds of polydextrose that
are accessible to an amylolytic enzyme, as well as removing
any available starch and maltodextrins present. A fructanase
treatment is also included to prevent fructans from coeluting
with polydextrose when it is separated by high
performance anion-exchange chromatography. The AOAC
2001.03 method is actually an extension of the soluble/
S33
European Journal of Clinical Nutrition

S34
insoluble version of the enzymatic–gravimetric method
AOAC 985.29, and is intended to measure the resistant
maltodextrins that remain in the solvent filtrate from the
precipitation and washing of the water-soluble residue. This
solvent filtrate, which can be up to 500 ml, is evaporated and
then ion exchange resins are used to remove salts and
proteins from the redisolved residue, which is then dried
again and filtered before quantification by HPLC as units
with DP43. There will be crossover in the materials
measured by the AOAC 2001.03 and 2000.11 methods, and
although fructans can be removed when present, the reality
is that any resistant oligosaccharides and possibly other
substances, may co-elute and therefore inflate the values
obtained.
Taken as a whole, the enzymatic–gravimetric analysis and
the complementing AOAC methods form a disjointed
approach to the determination of resistant carbohydrates.
There is specific concern about the double counting of the
same substances by more than one of these procedures,
which severely limits the integration of values. It has been
suggested by some that the combined enzymatic-gravimetric
and resistant maltodextrin method could provide an
integrated approach to dietary fibre determination for the
indigestibility approach. This must be viewed with some
skepticism, as both these determinations would be prone to
interference due to their empirical nature, and furthermore,
there would be no primary standards available to reflect the
diversity of materials recovered by the HPLC measurement.
The other disadvantage of applying empirical methods
recovering unidentified material is that it is not possible to
indicate what material is present, or how much, if any, of it
conforms to the qualifying criteria of exhibiting beneficial
physiological properties.
Public health application of carbohydrate
measurements
The nutritional characterization of dietary carbohydrates
should acknowledge the heterogeneity in the functional
properties of carbohydrate containing foods. This ranges
from a consideration of the metabolizable energy provided,
to their varied physico-chemical characteristics in the
gastrointestinal tract and subsequent effects on physiology
and metabolism, and to the more holistic consideration of
the overall nutrient profile of the foods.
Describing these varied attributes in a consistent and
nutritionally relevant manner has proved challenging, as
chemical composition does not always adequately reflect
functionality, especially in the context of the food matrix.
The implication is that it has been difficult to assign
population reference intake values and nutrition claims
based on the commonly applied chemical divisions such
as starch and sugars. There is therefore a requirement to
incorporate additional nutritional descriptions into classification,
measurement and public health messages relating to
European Journal of Clinical Nutrition
Nutritional characterization and measurement of dietary carbohydrates
K Englyst et al
carbohydrate containing foods. Some of the challenges and
potential solutions are commented on here.
Sugars
The nutritional considerations relating to sugar containing
foods can be evaluated by their impact on dental caries,
excess energy intake, nutrient:energy ratios, and physiology
due to metabolic differences between sugars. The complexity
in the nutritional description of sugars relates to the food
groups from which they are consumed. This need to
distinguish between sugar sources was recognized by the
development of the term intrinsic sugars for those retained
in intact cellular structures (Department of Health, 1991)
and the terms free sugars, added sugars and non-milk
extrinsic sugars which are essentially synonymous with each
other. US Department of Agriculture have prepared a large
database with values for added and total sugars for a wide
range of products (Pehrsson et al., 2006).
There is no available evidence to suggest that there are any
adverse effects on health outcomes in humans from the
sugars consumed in the form of fruit, vegetables or milk. In
any case, the bulky nature of fruits and vegetables tends to
limit the absolute consumption of sugars from these sources.
In contrast, free (or added) sugars have the potential to be
consumed in large quantities and have a more direct impact
on these related health issues (van Dam and Seidell, 2007).
For this reason, guidelines have limited free sugar intake to
o10% of energy (Department of Health, 1991; WHO, 2003).
As there is no justification to have a specific limit for the
consumption of intrinsic sugars from fruit, vegetables and
milk, these are instead considered within the overall
guidance to consume 45–60% of energy from carbohydrates.
However, apart from when dealing with primary food
groups such as fruits, vegetables and milk, it can be difficult
to identify the source of sugars in foods, particularly in
products composed of multiple ingredients. Furthermore
nutrition labels only state values for total sugars, and it is
perceived as overly complex to include a division between
intrinsic and free sugars. This poses a problem for the
nutrient signposting and claims relating to sugar content. A
practical solution has been proposed that establishes a high
criteria for claim purposes based on the guidelines on free (or
added) sugars (50 g for a 2000 kcal diet), but incorporating
a small allowance (that is, 10 g) for the average consumption
of intrinsic and milk sugars consumed from manufactured
foods (FSA, 2006). This would allow the food labelling for
total sugars to be used in the nutritional signposting of
manufactured foods. This is a pragmatic approach that is
consistent with the dietary guidelines for a selective restriction
of free sugars without the reliance on an analytical
distinction between intrinsic and free (or added) sugars,
which has proved difficult for routine labelling purposes.
Starch and whole grains
Starch has been presented as a preferable source of carbohydrate
to sugars. In reality this is an oversimplification, and

similar to the situation with sugars, the food source of starch
needs to be considered when evaluating nutritional properties.
A considerable amount of starch is consumed as refined
cereal products where the germ and bran fractions have been
lost along with the majority of the associated micronutrients
and phytochemicials. This results in a lower nutrient/energy
ratio for many refined cereal products when compared with
their whole grain counterparts. Furthermore, the physicochemical
characteristics of starch are very dependent on the
biological origin and degree of processing, affecting both the
rate and extent of digestion in the small intestine.
The consequence is that in isolation, a value for the total
starch content in foods or diets is not necessarily very
informative about the functional attributes of a carbohydrate
food. Information on dietary fibre defined as ‘intrinsic plant
cell-wall polysaccharides’ will help identify whether the
starch is associated with refined or whole grain material, and
the detailed profile of the carbohydrate-release characteristics
will indicate the likely gastrointestinal and metabolic
fate of the carbohydrate food.
Glycaemic index
The GI concept has provided insight into the physiological
properties of carbohydrate containing foods, which are not
apparent from chemical composition alone. This physiological
ranking is often presented as a description of carbohydrate
quality, and it is therefore appropriate to consider
how the GI integrates with the overall strategy of characterizing
the functionality of dietary carbohydrates.
In addition to the rate of carbohydrate digestion, other
food-mediated effects on both gastrointestinal events and
postabsorptive metabolism can influence the GI. Therefore,
GI values do not represent a direct measure of carbohydrate
absorption from the small intestine, but rather reflect the
combined effect of all the properties of a food or meal that
influence the rate of influx and removal of glucose from the
circulation. However, the different mechanisms responsible
for changes in the glycaemic response to a food or meal
cannot be considered equally beneficial to health. For
instance, it would be inappropriate to promote a food as
low GI if the underlying mechanisms responsible were either
high contents of fat or fructose. In such cases, any
potentially detrimental nutritional attributes should take
precedence over the physiological GI characteristic of the
food or meal. Therefore, the GI measure should be applied
only to foods with a high carbohydrate content, and there
should be an overall consideration of the food and meal
characteristics. The in vitro carbohydrate-release profiles,
such as the measures of rapidly and slowly available glucose,
can specifically identify the low GI products that are rich in
slow-release carbohydrates, which have demonstrated health
benefits.
As the GI measurement relates to the available carbohydrate
component of foods, it is appropriate to calculate
the portion sizes of test meals based on direct determinations
Nutritional characterization and measurement of dietary carbohydrates
K Englyst et al
of available carbohydrate, thereby reducing this aspect of
GI measurement uncertainty and ensuring that resistant
carbohydrates are not mistakenly included. The accompanying
paper on GI addresses further issues relating to the
determination and application of this physiological measure
(Venn and Green, 2007).
Dietary fibre and other resistant carbohydrates
There has been considerable interest in the development and
marketing of a number of resistant carbohydrates including
polysaccharides (for example, retrograded and modified
resistant starches, modified celluloses), oligosaccharides (for
example, fructooligosaccharides, polydextrose and resistant
maltodextrins) and sugars (for example, polyols and lactulose).
As discussed earlier there has been debate about
whether some or any of these materials should be encompassed
within a dietary fibre term (that is, indigestibility
approach), or whether they should be considered separately
(that is, plant-rich diet approach). This debate is of
considerable importance, as it impacts directly on the
rationale for dietary fibre and on how different resistant
carbohydrates can be managed from a public health
perspective.
By the plant-rich diet approach, the definition very
effectively supports the existing dietary guidelines to consume
fruits, vegetables and whole grains. Likewise, this
approach is consistent with the evidence base on which
population reference intake values have been established.
This need to differentiate between dietary fibre intrinsic to
plant foods, and added preparations was also recognized by
US academy of sciences (IOM, 2001).
By the indigestibility approach, the inclusion of materials
other than ‘intrinsic plant cell wall polysaccharides’ as
dietary fibre could adversely impact on the guidelines to
consume fibre from plant foods, and it could be wrongly
interpreted as inferring that the evidence for the benefit of
the plant-rich diet can be extrapolated to these other
substances. This issue is currently very pertinent, as although
dietary guidelines are to consume at least five portions
of fruit and vegetable and three or more portions of
whole grains daily, average intakes are far lower in some
populations. As a consequence, the average dietary fibre
intakes are also less than the population reference values,
and this has been represented by some as a ‘fibre gap’ that
could be met through supplementation with other substances.
However, this would be a fundamental misinterpretation
of the evidence base for a fibre rich diet being
beneficial to health and would potentially mislead the public
in their selection of this diet.
As only intrinsic plant cell wall polysaccharides are
included within the plant-rich diet definition, other sources
of resistant carbohydrates would need to be described
separately for labelling and nutrition claims purposes. Of
course, it would only be necessary and appropriate to include
additional categories such as resistant oligosaccharides or
S35
European Journal of Clinical Nutrition

S36
resistant starch on nutrition labels if they were shown to be
of sufficient relevance to public health.
The nutrition labelling of separate categories of resistant
carbohydrates represents an opportunity for industry to
stimulate product innovation through the development
and promotion of functional ingredients based on their
specific physiological properties. This would most
appropriately take the form of health claims, for which
there is established legislation to ensure that these substances
are suitably evaluated and controlled (for example,
European Commission Regulation EC No 1924/2006). The
diversity in physiological attributes between different
functional ingredients, and the variation in the quantities
required to produce the nutritional or physiological effects,
makes it unfeasible to group all these substances under a
single set of conditions for nutrition or health claims.
Dietary recommendations and labelling
Food-based dietary guidelines have traditionally been the
most effective approach to the public communication of
nutrition. For carbohydrates, this is simply conveyed by the
promotion of fruit, vegetable and whole grain consumption.
However, the increasing predominance of manufactured
products necessitates additional strategies that address the
varied functional properties of these processed foods, thereby
allowing either beneficial or potentially adverse attributes
to be distinguished. Such functional parameters should
reflect nutritional aspects of different types of ingredients
and specific carbohydrate components, as well as any effects
of processing on the overall food characteristics. These
principles form the basis of the classification and measurement
scheme for dietary carbohydrates presented in Table 1,
which can be applied as a tool for further research into the
link between dietary carbohydrates and health.
Recommendations
1. Food-based guidelines promoting the consumption of
fruits, vegetables and whole grains are some of the most
effective public health messages. Carbohydrate classification
and measurements should support these dietary
guidelines and provide the means with which to describe
the component ingredients and functional properties of
foods, including those attributed to the food matrix.
2. For energy calculation purposes, there is a need to
describe carbohydrates in terms of their gastrointestinal
and metabolic fate, as ‘available’, which provide carbohydrate
for metabolism, and as ‘resistant’, which resist
digestion in the small intestine or are poorly absorbed/
metabolized. These categories should be further subdivided
into types describing attributes of specific nutritional and
functional relevance.
3. There should be a commitment to move away from
empirical based methods that measure unspecified materials,
European Journal of Clinical Nutrition
Nutritional characterization and measurement of dietary carbohydrates
K Englyst et al
towards rational methods providing direct and specific
measurement of different types and categories of carbohydrates
as their chemically identified components.
4. The selective restriction of free or added sugars is a useful
food-based guideline. To support this, nutrition claims
relating to the content of sugars in manufactured
products should be provided in the context of the
established maximum intake limit of free sugars (10% of
energy), but with an additional small allowance for the
intrinsic sugars provided by manufactured food groups.
This allows the content of total sugars to be used for
claims purposes and overcomes the need to distinguish
between intrinsic and free sugars on nutrition labels. This
effectively targets the restriction of free or added sugars,
without adversely influencing intakes of fruits, vegetables
or milk.
5. The GI and glycaemic load are useful nutritional terms
providing insight into a physiological parameter that is
not always apparent from chemical composition alone.
Its application should be limited to foods with high
carbohydrate contents and is most effectively interpreted
in conjunction with information on other food/meal
characteristics including detailed sugar composition and
starch digestibility profiles.
6. Dietary fibre should be considered as a public health term
supporting dietary guidelines to consume a plant-rich
diet. The definition ‘intrinsic plant cell wall polysaccharides’
provides the only consistent link with the scientific
evidence on which these guidelines are based. The
physiological characteristic of indigestibility is therefore
not considered to be an adequate basis for the definition
of dietary fibre.
7. Resistant carbohydrates other than dietary fibre should be
considered separately and by their own merits. Distinct
categories are essential for functional ingredients to be
managed effectively from a public health perspective. As
there is the potential for large amounts of added resistant
carbohydrates to be consumed, safe upper intake limits
may need to be established.
Acknowledgements
We wish to thank Professor Nils-Georg Asp, Professor John H
Cummings, Professor Timothy Key, Professor Jim Mann,
Professor HH Vorster and Dr Roger Wood 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
Dr Klaus Englyst: Director and share-holder in Englyst
Carbohydrates Ltd which is a small research-oriented

company working on dietary carbohydrates and health. The
UK Food Standards Agency is the main research partner and
sponsor. In addition, Englyst Carbohydrates provide analytical
assistance to universities and food industry worldwide,
albeit on a small scale. The complete independence of
Englyst Carbohydrates is maintained by not entering into
any consultancy agreement.
Professor Simin Liu: Member of the Scientific advisory
board for the EU Health Grain Project.
Dr Hans Englyst: Director and share-holder in Englyst
Carbohydrates Ltd which is a small research-oriented
company working on dietary carbohydrates and health.
The UK Food Standards Agency is the main research partner
and sponsor. In addition, Englyst Carbohydrates provide
analytical assistance to universities and food industry worldwide,
albeit on a small scale. The complete independence of
Englyst Carbohydrates is maintained by not entering into
any consultancy agreement.
Peer-reviewers
Professor Nils-Georg Asp: On part-time leave from university
professorship to be the Director of the Swedish
Nutrition Foundation (SNF), a nongovernmental organization
for the promotion of nutrition research and its practical
implications. SNF is supported broadly by the food sector;
the member organizations and industries are listed on the
SNF home page (www.snf.ideon.se).
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).
Professor Timothy Key: None declared.
Professor Jim Mann: None declared.
Professor HH Vorster: Member and Director of the Africa
Unit for Transdisciplinary health Research (AUTHeR), Research
grant from the South African Sugar Association
Dr Roger Wood: None declared.
References
Abrams SA, Griffin IJ, Hawthorne KM, Liang L, Gunn SK, Darlington
G et al. (2005). A combination of prebiotic short- and long-chain
inulin-type fructans enhances calcium absorption and bone
mineralization in young adolescents. Am J Clin Nutr 82, 471–476.
Alberts DS, Martinez ME, Roe DJ, Guillen-Rodriguez JM, Marshall JR,
van Leeuwen JB et al. (2000). Lack of effect of a high-fiber cereal
supplement on the recurrence of colorectal adenomas. N Engl J
Med 342, 1156–1162.
American Association of Cereal Chemists (2001). The definition of
dietary fiber. Cereal Foods World 46, 112–126.
AOAC (2005). Official Methods of Analysis 18th Ed W Horwitz. AOAC
International: Gaithersburg, MD, USA.
Bazzano LA, He J, Ogden LG, Loria CM, Whelton PK (2003). Dietary
fiber intake and reduced risk of coronary heart disease in US men
and women. The National Health and Nutrition Examination
Survey I Epidemiologic Follow-up Study. Arch Intern Med 163,
1897–1904.
Nutritional characterization and measurement of dietary carbohydrates
K Englyst et al
Bazzano LA, Serdula M, Liu S (2005). Prevention of type 2 diabetes by
diet and lifestyle modification. J Am Coll Nutr 24, 310–319.
Bingham SA, Day NE, Luben R, Ferrari P, Slimani N, Norat T et al.
(2003). Dietary fibre in food and protection against colorectal
cancer in the European Prospective Investigation into Cancer and
Nutrition (EPIC): an observational study. Lancet 361, 1496–1501.
Bonithon-Kopp C, Kronborg O, Giacosa A, Rath U, Faivre J (2000).
Calcium and fibre supplementation in prevention of colorectal
adenoma recurrence: a randomised intervention trial. European
Cancer Prevention Organisation Study Group. Lancet 356,
1300–1306.
Brand-Miller J, Hayne S, Petocz P, Colagiuri S (2003). Low-glycemic
index diets in the management of diabetes: a meta-analysis of
randomized controlled trials. Diabetes Care 26, 2466–2468.
Burkitt DP (1969). Related disease—related cause? Lancet ii,
1229–1231.
Burn J, Katheuser A, Fodde R, Coaker J, Chapman PD, Mathers JC
(1996). Intestinal tumours in the Apc 1638N mouse: aspirin not
protective and resistant starch increases small bowel tumours. Eur
J Hum Genet 4 (Suppl 1), 13.
Champ MM, Langkilde A-M, Brouns F, Kettlitz B, Bail-Collet YL
(2003). Advances in dietary fibre characterization. 2. Consumption,
chemistry, physiology and measurement of resistant starch;
implications for health and food labelling. Nutr Res Rev 16, 143–161.
Daly M (2003). Sugars, insulin sensitivity, and the postprandial state.
Am J Clin Nutr 78, 865S–872S.
Department of Health (1991). Dietary reference values for food energy
and nutrients for the United Kingdom. Report on Health and Social
41. London.
Elia M, Cummings JH (2007). Physiological aspects of energy
metabolism and gastrointestinal effects of carbohydrates. Eur J
Clin Nutr.
Ellis PR, Rayment P, Wang Q (1996). A physico-chemical perspective
of plant polysaccharides in relation to glucose absorption, insulin
secretion and the entero-insular axis. Proc Nutr Soc 55, 881–898.
Englyst HN, Cummings JH (1985). Digestion of the polysaccharides
of some cereal foods in the human small intestine. Am J Clin Nutr
42, 778–787.
Englyst HN, Cummings JH (1986). Digestion of the carbohydrates of
banana (Musa paradisiaca sapientum) in the human small
intestine. Am J Clin Nutr 44, 42–50.
Englyst HN, Cummings JH (1987). Digestion of the polysaccharides
of potato in the small intestine of man. Am J Clin Nutr 45, 423–431.
Englyst HN, Hudson GJ (1996). The classification and measurement
of dietary carbohydrates. Food Chem 57, 15–21.
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 fiber as non-starch polysaccharides with gas–liquid
chromatographic, high-performance liquid chromatographic or
spectrophotometric measurement of constituent sugars. Analyst
119, 1497–1509.
Englyst HN, Quigley ME, Englyst KN, Bravo L, Hudson GJ (1996).
Dietary fibre. Measurement by the Englyst NSP procedure.
Measurement by the AOAC procedure. Explanation of the
differences. J Assoc Pub Anal 32, 1–52.
Englyst HN, Wiggins HS, Cummings JH (1982). Determination of the
non-starch polysaccharides in plant foods by gas–liquid chromatography
of constituent sugars as alditol acetates. Analyst 107,
307–318.
Englyst KN, Englyst HN (2005). Carbohydrate bioavailability. Brit J
Nutr 94, 1–11.
Englyst KN, Englyst HN, Hudson GJ, Cole TJ, Cummings JH (1999).
Rapidly available glucose in foods. An in vitro measurement that
reflects the glycemic response. Am J Clin Nutr 69, 448–454.
Englyst KN, Vinoy S, Englyst HN, Lang V (2003). Glycaemic index of
cereal products explained by their content of rapidly and slowly
available glucose. Br J Nutr 89, 329–339.
S37
European Journal of Clinical Nutrition

S38
Nutritional characterization and measurement of dietary carbohydrates
K Englyst et al
FAO (2003). Food energy—methods of analysis and conversion
factors. Report of a technical workshop. Rome 2002. FAO Food and
Nutrition Paper 77.
FDA (2000). Docket No. 91N-0098. Letter regarding dietary supplement
health claim for fiber with respect to colorectal cancer (US
food and drug administration center for food safety and applied
nutrition office of nutritional products, labeling, and dietary
supplements October 10, 2000).
Fried SK, Rao SP (2003). Sugars, hypertriglyceridemia, and cardiovascular
disease. Am J Clin Nutr 78, 873S–880S.
FSA (2006). Rationale and discussion of the 15 g high sugars criterion
within the FSA signposting scheme. http://www.food.gov.uk/
foodlabelling/signposting/ signposttimeline/rationalesugars/.
Goodlad RA (2007). Fiber can make your gut grow. Nutrition 23,
434–435.
Gostner A, Blaut M, Schäffer V, Kozianowski G, Theis S, Klingeberg M
et al. (2006). Effect of isomalt consumption on faecal microflora
and colonic metabolism in healthy volunteers. Brit J Nutr 95,40–50.
Gross LS, Li L, Ford ES, Liu S (2004). Increased consumption of
refined carbohydrates and the epidemic of type 2 diabetes in the
United States: an ecologic assessment. Am J Clin Nutr 79, 774–779.
Gudmand-Hoyer E (1994). The clinical significance of disaccharide
maldigestion. Am J Clin Nutr 59, 735S–741S.
Institute of Medicine (2001). Dietary Reference Intakes: Proposed
Definition of Dietary Fiber. The National Acadamies Press:
Washington.
Institute of Medicine (2002). Dietary Reference Intakes for Energy,
Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino
Acids (macronutrients). The National Academies Press: Washington.
Jacobs 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.
Jenkins DJ, Kendall CW, Augustin LS, Franceschi S, Hamidi M,
Marchie A et al. (2002). Glycemic index: overview of implications
in health and disease. Am J Clin Nutr 76, S266–S273.
Liu S (2002). Intake of refined carbohydrates and whole grain foods
in relation to risk of type 2 diabetes mellitus and coronary heart
disease. J Am Coll Nutr 21, 298–306.
Liu S, Lee IM, Ajani U, Cole SR, Buring JE, Manson JE (2001). Intake
of vegetables rich in carotenoids and risk of coronary heart disease
in men: the physicians’ health study. Int J Epidemiol 30, 130–135.
Liu S, Manson J, Lee I, Cole S, Willett W, Buring J (2000b). Fruit and
vegetable intake and risk of cardiovascular disease: the women’s
health study. Am J Clin Nutr 72, 922–928.
Liu S, Manson JE, Stampfer MJ, Hu FB, Giovannucci E, Colditz GA
et al. (2000a). A prospective study of whole-grain intake and risk of type
2 diabetes mellitus in US women. Am J Public Health 90, 1409–1415.
Liu S, Willett WC, Manson JE, Hu FB, Rosner B, Colditz GA (2003).
Relation between changes in intakes of dietary fiber and grain
products and changes in weight and development of obesity
among middle-aged women. J Am Coll Nutr 78, 920–927.
Liu SM, Stampfer MJ, Hu FB, Giovannucci E, Rimm E, Manson JE
et al. (1999). Whole-grain consumption and risk of coronary heart
disease: results from the Nurses’ Health Study. Am J Clin Nutr 70,
412–419.
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.
McCance RA, Lawrence RD (1929). The Carbohydrate Content of Foods.
MRC Special Report. HMSO: London.
McCleary BV, Murphy A, Mugford DC (2000). Measurement of total
fructan in foods by enzymatic/spectrophotometric method:
collaborative study. J AOAC Int 83, 356–364.
Park Y, Hunter DJ, Spiegelman D, Bergkvist L, Berrino F, van den
Brandt PA et al. (2005). Dietary fiber intake and risk of colorectal
cancer: a pooled analysis of prospective cohort studies. JAMA 294,
2849–2857.
European Journal of Clinical Nutrition
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.
Pendlington AW, Meuree-Vanlaethem N, Brookes A (1996). The
Method Specific Certification of the Mass Fraction of Dietary Fibre in
Lyophilised Haricot Beans, Carrot, Apple, Full Fat Soya Flour and Bran
Breakfast Cereal Reference Materials. CRMs 514, 515, 516, 517 & 518.
Office for Official Publications of the European Communities:
Luxembourg.
Pietinen P, Rimm EB, Korhonen P, Hartman AM, Willett WC,
Albanes D et al. (1996). Intake of dietary fiber and risk of coronary
heart disease in a cohort of Finnish men. The Alpha-Tocopherol,
Beta-Carotene Cancer Prevention Study. Circulation 94, 2720–2727.
Quigley ME, Hudson GJ, Englyst HN (1999). Determination of
resistant short-chain carbohydrates (non-digestible oligosaccharides)
using gas-liquid chromatography. Food Chem 65, 381–390.
Rabe E (1999). Effect of processing on dietary fiber in foods. In: Cho
SS, Prosky L, Dreher, M (eds). Complex carbohydrates in foods.
Marcel Dekker, Ink: NewYork, Basel. pp 395–409.
Ranhotra GS, Gelroth JA, Eisenbraun GJ (1991). High-fiber white
flour and its use in cookie products. Cereal Chem 68, 432–433.
Rimm EB, Ascherio A, Giovannucci E, Spiegelman D, Stampfer MJ,
Willett WC (1996). Vegetable, fruit, and cereal fiber intake and risk
of coronary heart disease among men. JAMA 275, 447–451.
Rissanen TH, Voutilainen S, Virtanen JK, Venho B, Vanharanta M,
Mursu J et al. (2003). Low intake of fruits and berries and
vegetables is associated with excess mortality in men; the Kupio
Ischaemic Heart Disease Risk Factor (KIHD) Study. Am Soc Nutr Sci
133, 199–204.
Roberfroid MB (2005). Introducing inulin-type fructans. Br J Nutr 93
(Suppl 1), S13–S25.
Rumessen JJ (1992). Hydrogen and methane tests for evaluation of
resistant carbohydrates. Eur J Clin Nutr 46 (Suppl 2), S77–S90.
Sengupta S, Muir JG, Gibson PR (2006). Does butyrate protect from
colorectal cancer? J Gastroenterol Hepatol 21, 209–218.
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.
Slavin J (2003). Why whole grains are protective: biological
mechanisms. Proc Nutr Soc 62, 129–134.
Southgate DAT (1969). Determination of carbohydrates in foods. II.
Unavailable carbohydrates. J Sci Food Agric 20, 331–335.
Southgate DAT, Englyst HN (1985). Dietary fibre: chemisty, physical
properties and analysis. In: Trowell H, Burkitt D, Heaton K (eds).
Dietary fibre, fibre depleated foods and disease. Academic Press:
London.
Steffen LM, Jacobs DR, Stevens J (2003). Associations of whole-grain,
refined-grain, and fruit and vegetable consumption with risks of
all-cause mortality and incident coronary artery disease and
ischemic stroke: the Atherosclerosis Risk in Communities (ARIC)
Study. Am J Clin Nutr 78, 383–390.
Theander O, Westerlund E (1993). Determination of individual
components of dietary fiber. In: Spiller GA (ed). CRC Handbook of
Dietary Fiber in Human Nutrition. pp 77–98. CRC Press Inc: Boca
Raton, FL.
Trowell H (1972). Crude fibre, dietary fibre and atherosclerosis.
Atherosclerosis 16, 138–140.
Trowell H, Burkitt D, Heaton K (1985). Editors of Dietary Fibre, Fibredepleted
Foods and Disease. Academic Press: London.
Tuohy KM, Hinton DJ, Davies SJ, Crabbe MJ, Gibson GR, Ames JM
(2006). Metabolism of Maillard reaction products by the human
gut microbiota—implications for health. Mol Nutr Food Res 50,
847–857.
USDA/DHHS (2005). Nutrition and Your Health: Dietary Guidelines for
Americans. DHHS: Washington, DC.
van Dam RM, Rimm EB, Willett WC, Stampfer MJ, Hu FB (2002).
Dietary patterns and risk for type 2 diabetes mellitus in US men.
Annals of Internal Medicine 136, 201–209.

van Dam RM, Seidell JC (2007). Carbohydrate intake and obesity.
Eur J Clin Nutr.
van den Heuvel EG, Muijs T, Van Dokkum W, Schaafsma G (1999).
Lactulose stimulates calcium absorption in postmenopausal
women. J Bone Miner Res 14, 1211–1216.
Venn BJ, Green TJ (2007). Glycemic index and glycemic load:
measurement issues and their effect on diet–disease relationships.
Eur J Clin Nutr.
Vinjamoori DV, Byrum JR, Hayes T, Das PK (2004). Challenges
and opportunities in the analysis of raffinose oligosaccharides,
pentosans, phytate, and glucosinolates. J Anim Sci 82,
319–328.
Wacker M, Wanek P, Eder E, Hylla S, Gostner A, Scheppach W (2002).
Effect of enzyme-resistant starch on formation of 1,N(2)-propanodeoxyguanosine
adducts of trans-4-hydroxy-2-nonenal and cell
proliferation in the colonic mucosa of healthy volunteers. Cancer
Epidemiol Biomarkers Prev 11, 915–920.
Nutritional characterization and measurement of dietary carbohydrates
K Englyst et al
Wasan HS, Goodlad RA (1996). Fibre-supplemented foods may
damage your health. Lancet 348, 319–320.
WHO Technical report series 916 (2003). Diet Nutrition and the
Prevention of Chronic Disease. WHO: Geneva.
Willett W, Manson J, Liu S (2002). Glycemic index, glycemic load,
and risk of type 2 diabetes. Am J Clin Nutr 76, S274–S280.
Wolk A, Manson JE, Stampfer MJ, Colditz GA, Hu FB, Speizer FE et al.
(1999). Long-term intake of dietary fiber and decreased risk of
coronary heart disease among women. JAMA 281, 1998–2004.
Wong JM, de Souza R, Kendall CW, Emam A, Jenkins DJ (2006).
Colonic health: fermentation and short chain fatty acids. J Clin
Gastroenterol 40, 235–243.
Wood R, Englyst HN, Southgate DAT, Cummings JH (1993).
Determination of dietary fibre in foods—collaborative trials. IV.
Comparison of Englyst GLC and colorimetric measurement with
the Prosky procedure. J Assoc Pub Anal 29, 57–141.
S39
European Journal of Clinical Nutrition

REVIEW
Physiological aspects of energy metabolism and
gastrointestinal effects of carbohydrates
M Elia 1 and JH Cummings 2
1 2
Institute of Human Nutrition, University of Southampton, Southampton, UK and Department of Pathology and Neuroscience,
Ninewells Hospital and Medical School, Dundee, UK
The energy values of carbohydrates continue to be debated. This is because of the use of different energy systems, for example,
combustible, digestible, metabolizable, and so on. Furthermore, ingested macronutrients may not be fully available to tissues,
and the tissues themselves may not be able fully to oxidize substrates made available to them. Therefore, for certain
carbohydrates, the discrepancies between combustible energy (cEI), digestible energy (DE), metabolizable energy (ME) and net
metabolizable energy (NME) may be considerable. Three food energy systems are in use in food tables and for food labelling in
different world regions based on selective interpretation of the digestive physiology and metabolism of food carbohydrates. This
is clearly unsatisfactory and confusing to the consumer. While it has been suggested that an enormous amount of work would
have to be undertaken to change the current ME system into an NME system, the additional changes may not be as great as
anticipated. In experimental work, carbohydrate is high in the macronutrient hierarchy of satiation. However, studies of eating
behaviour indicate that it does not unconditionally depend on the oxidation of one nutrient, and argue against the operation of
a simple carbohydrate oxidation or storage model of feeding behaviour to the exclusion of other macronutrients. The site, rate
and extent of carbohydrate digestion in, and absorption from the gut are key to understanding the many roles of carbohydrate,
although the concept of digestibility has different meanings. Within the nutrition community, the characteristic patterns of
digestion that occur in the small (upper) vs large (lower) bowel are known to impact in contrasting ways on metabolism, while
in the discussion of the energy value of foods, digestibility is defined as the proportion of combustible energy that is absorbed
over the entire length of the gastrointestinal tract. Carbohydrates that reach the large bowel are fermented to short-chain fatty
acids. The exact amounts and types of carbohydrate that reach the caecum are unknown, but are probably between 20 and
40 g/day in countries with ‘westernized’ diets, whereas they may reach 50 g/day where traditional staples are largely cereal or
diets are high in fruit and vegetables. Non-starch polysaccharides clearly affect bowel habit and so, to a lesser extent, does
resistant starch. However, the short-chain carbohydrates, which are also found in breast milk, have little if any laxative role,
although do effect the balance of the flora. This latter property has led to the term ‘prebiotic’, which is defined as the capacity to
increase selectively the numbers of bifidobacteria and lactobacilli without growth of other genera. This now well-established
physiological property has not so far led through to clear health benefits, but current studies are focused on their potential to
prevent diarrhoeal illnesses, improve well-being and immunomodulation, particularly in atopic children and on increased
calcium absorption.
European Journal of Clinical Nutrition (2007) 61 (Suppl 1), S40–S74; doi:10.1038/sj.ejcn.1602938
Keywords: carbohydrate; energy; short-chain fatty acids; glycaemic index; food labelling; prebiotic
Introduction
European Journal of Clinical Nutrition (2007) 61 (Suppl 1), S40–S74
& 2007 Nature Publishing Group All rights reserved 0954-3007/07 $30.00
www.nature.com/ejcn
All organisms require fuels to maintain their life cycles. For
humans, carbohydrates are the major fuels, typically
accounting for 45–70% of the total energy intake and
Correspondence: Professor M Elia, Institute of Human Nutrition, University of
Southampton, Southampton General Hospital, MP 113 (F Level), Tremona Rd,
Southampton SO16 6YD, UK.
E-mail: elia@soton.ac.uk
expenditure. Despite their importance in energy metabolism,
the energy values of some carbohydrates continue to be
debated and are confused due to the existence of different
energy systems, which are not entirely consistent with each
other (for example, combustible, digestible energy (DE),
metabolizable energy (ME), net metabolizable energy
(NME) systems; general and specific Atwater systems). The
choice of energy system is of considerable importance
because it not only affects the energy values of carbohydrates,
but also those of fats, proteins and alcohol (FAO,

2003). In establishing the energy values of carbohydrates, it
is not only necessary to consider their intrinsic physicochemical
properties, but also their physiology, specifically,
their digestibility, the end products of their metabolism and
even the pathways by which they are metabolized. Since it is
desirable to use the same energy system for all macronutrients,
it is appropriate to consider at least briefly the energy
values of carbohydrates, fats and proteins as a combined
entity. Similarly, the role of carbohydrates in energy balance
can only be adequately examined by considering the broader
perspectives of energy homeostasis, which involve all
macronutrients.
The site of carbohydrate breakdown in the gut, the rate
and extent of breakdown, and the kinetics of absorption are
key to understanding many other roles that carbohydrate
plays beyond providing energy. In the large bowel, dietary
carbohydrate interacts with a rich and diverse microbiota
that produce end products unique to the body. This process
of fermentation affects bowel habit, transit time, mucosal
health, and also provides products to the portal and systemic
circulation, thus effecting metabolism both within and
beyond the gut.
In this paper, we have focused on areas where there has
been progress since the 1997 consultation, or where there is
substantial controversy. The implications for the glycaemic
response and for conditions such as diabetes, obesity and
cancer are dealt with in other papers in this expert review.
The evidence reviewed is primarily from human studies.
Energy values of carbohydrates and other
macronutrients
Figure 1, which is a modification of previous diagrams
(Warwick and Baines, 2000; Livesey, 2001; FAO, 2003), shows
the flow of energy through the body. It presents a conceptual
framework for considering energy values of foods and
individual macronutrients. The pathway begins with the
combustible energy intake that depends on the physicochemical
characteristics of ingested macronutrients, and not
on physiological processes. There are then three subsequent
consecutive steps, each of which depends on the previous
one. Each of these depends on physiological processes:
digestibility (relevant to DE), metabolizability (relevant to
ME) and relative metabolic (bioenergetic) efficiency (relevant
to NME). These are considered separately below.
Combustible energy
Combustible energy refers to the heat released during
complete combustion of foodstuffs or macronutrients in
the presence of O2, to yield CO 2 and H 2O. The heat of
combustion of amino acids/protein is reported in various
ways, for example with S being in elemental form, as SO 3,or
H 2SO 4. In the comprehensive analysis provided by Elia and
Livesey (1992), the end products are H 2O (liquid), CO 2 (gas),
Physiological aspects of energy metabolism
M Elia and JH Cummings
Gains
(Stored energy)
Ingested energy (cIE)
Digestible energy (DE)
= cIE – cFE - cGaE
Metabolisable energy (ME)
= DE – cUE - cSE
Net metabolisable energy
(NME) = ME –Hine = kME
Losses
Faecal energy (cFE)
Combustible gas (cGaE)
Urinary energy (cUE)
Surface energy (cSE)
Heat due to bio-energetic
inefficiency relative to
glucose oxidation (Hine)
Figure 1 Diagram showing the flow of energy through the body.
The abbreviations cIE, cFE, cGaE and cSE refer to combustible intake
energy (cIE), combustible faecal (cFE), gaseous (cGaE) and surface
energy (cSE), respectively. H ine is the heat released as a result of
metabolic inefficiency in performing metabolic and external work
relative to glucose. The coefficient k reflects this metabolic efficiency.
The equations shown in the diagram apply in states of nutrient
balance. In situations where excess energy is deposited, metabolizable
energy (ME) and net metabolizable energy (NME) of the diet
includes the energy that would be made available through
mobilization and oxidative metabolism of the stored energy.
N 2 (gas) and H 2SO 4.115H 2O. Combustible energy has also
been called the heat of combustion, gross energy, intake
energy and energy intake. Since some of these terms can be
confused with DE and ME intake, the term combustible
intake of energy (cIE) will be used in this paper, because it is
self-explanatory and less likely to be misinterpreted or
confused with other terms.
The combustible energy values for hexose-based carbohydrates
(kJ/g) are generally between 15.5 and 17.5 kJ/g (less
than 15% difference) (Table 1) (Livesey and Elia, 1988; Elia
and Livesey, 1992), and are more than twofold lower than
those of typical fats. Since carbohydrates are more oxidized
than fats (their carbon skeletons are linked to relatively more
oxygen and less hydrogen), they generate less heat than fat
when fully combusted with oxygen. The variability in heat
of combustion of different hexose-based carbohydrates (kJ/g)
is almost entirely determined by the number of glycosidic
bonds, the formation of which is associated with loss of one
molecule of water (water of condensation) per glycosidic
bond (Elia and Livesey, 1992). When the heat of combustion
of hexose-based carbohydrates is expressed as monosaccharide
equivalents, the values are very consistent with each
other (B15.7 kJ/g, rounded to 16 kJ/g). The value for
glycogen is only approximately 1.3% greater than that of
glucose (Elia and Livesey, 1992), the difference being due to
the heat released during the breakdown of the glycosidic
bonds (in living organisms, this is achieved through
enzymatic hydrolysis of glycosidic bonds (heat of hydrolysis).
Glucose monohydrate is unusual in being hydrated
S41
European Journal of Clinical Nutrition

S42
Table 1 The RQ (CO 2 production/O 2 consumption), combustible
energy equivalents of CO 2 (cEeqCO 2)andO 2 (cEeqO 2), and the heat
of combustion of the major macronutrients and specific carbohydrates a
RQ EeqO 2 (kJ/l) EeqCO 2 (kJ/l) Heat of
combustion (kJ/g)
Macronutrients
Carbohydrate 1.000 21.12 21.12 17.51
Fat 0.710 19.61 27.62 39.42
Protein 0.835 19.48 23.33 23.64
Alcohol 0.667 20.33 30.49 29.68
Specific carbohydrates
Starch 1.000 21.12 21.12 17.48
Glycogen 1.000 21.12 21.12 17.52
Sucrose 1.000 20.97 20.97 16.48
Maltose 1.000 20.99 20.99 16.49
Lactose 1.000 21.00 21.00 16.50
Glucose 1.000 20.84 20.84 15.56
Galactose 1.000 20.84 20.84 15.56
Fructose 1.000 20.91 20.91 15.61
Glycerol 0.857 21.16 24.69 18.03
Erythritol 0.889 20.91 23.53 17.27
Xylitol 0.909 19.92 21.91 16.96
Sorbitol 0.923 20.89 22.71 16.71
Mannitol 0.923 20.89 22.71 16.71
Maltitol 0.960 20.87 21.74 16.98
Lactitol 0.960 20.87 21.74 16.98
Isomalt 0.960 20.87 21.74 16.98
Abbreviations: EeqCO2, energy equivalents of CO2; EeqO2, energy equivalents
of O2; RQ, respiratory quotient.
a The values are based on the characteristics of the nutrients and are
independent of whether they are endogenous or exogenous or whether they
are given orally or intravenously. Based on Elia and Livesey (1992) and
unpublished data.
and having a low-energy density (14.1 kJ/g), but when
expressed as anhydrous glucose, the value becomes 15.6 kJ/g.
The heat of combustion of polyols (alcohol derivatives of
sugars), as for other carbohydrates, depends on their
composition and number of glycosidic linkages. For many
the values are B17.0 kJ/g (for example, 16.70–17.2 kJ/g for
erythritol, isomalt, lactitol, maltitol, mannitol, sorbitol,
xylitol) (Elia and Livesey, 1992; Livesey, 1992). Glycerol
(triol), which is also a polyol, has a relatively high heat of
combustion (18 kJ/g). The values for polysaccharides that
reach the colon do not differ substantially from those for
starch or glycogen (for example, 17.5 kJ/g for guar gum and
solka-floc cellulose, 17.2 kJ/g for pectin, but as low as 15.5 kJ/
g for psyllum gum and as high as 17.6 kJ/g for sugar beet
fibre) (Livesey, 1992; Livesey and Elia, 1995). The mean
values for non-starch polysaccharides (NSP) in cereals have
been reported to be 17.5 (range 16.6–18.5) kJ/g, vegetables,
16.8 (range 16.6–17.9) kJ/g, and fruits 16.5 (range 14.9–
17.3) kJ/g. Hydroxyl-propylmethyl cellulose, a chemically
modified NSP, has a particularly high heat of combustion
(22.0 kJ/g), which is due to the high heat of formation of the
substituted side chains. The values for oligosaccharides, such
as polydextrose, polyfructose and soya bean oligosaccharides
are between 16.8 and 17.0 kJ/g. It is estimated that the
European Journal of Clinical Nutrition
Physiological aspects of energy metabolism
M Elia and JH Cummings
average combustible energy of NSP in mixed conventional
diets is approximately 17.0 kJ/g, compared to 15.7 kJ/g for
carbohydrates (as monosaccharide equivalents) that are
absorbed by the small bowel (‘available carbohydrates’).
The principles of thermodynamics imply that the heat
released when carbohydrates are completely combusted in a
bomb calorimeter in the presence of oxygen is the same as
when the same quantity of carbohydrate is fully oxidized to
H2O and CO 2 by living organisms through a large series of
metabolic steps. However, ingested macronutrients may not
be made fully available to tissues, and the tissues themselves
may not be able to fully oxidize substrates made available to
them. Therefore, for certain carbohydrates, the discrepancies
between combustible energy (cEI), DE, ME and NME may be
considerable.
Digestible energy
It is obvious that substances that are not absorbed cannot be
metabolized by human tissues. For the purposes of energy
metabolism, digestibility is defined as the proportion of
combustible energy that is absorbed over the entire length of
the gastrointestinal tract. DE (DE ¼ cIE digestibility) is
estimated as the difference between the combustible energy
present in ingested food and that present in faeces
( þ combustible gases, such as H2 and CH 4 which are
excreted).
DE ¼ cIE cFE cGaE
The ‘c’ in cIE (combustible intake energy), cFE (combustible
faecal energy) and cGaE (combustible gaseous energy)
indicates that it is the heat of combustion that is being
considered, thus avoiding confusion with DE, ME and NME
values. DE is actually an ‘apparent’ energy value, because the
flux of nutrients into and out of the colon is not unidirectional.
For example, mucus is secreted into the large bowel
(estimated to be 2–3 g/day, but in certain diseases affecting
the large intestine, the amount may be considerably more).
In addition, desquamated colonic cells are shed directly into
the lumen of the colon, which may also receive desquamated
cells from the small intestine. Some of the energy from these
sources, which is not present in the diet, may be lost to
faeces, reducing the apparent digestibility of the diet.
Another problem concerns loss of energy in the form of
combustible gaseous products (flatus þ breath), which was
neglected in early human studies. The result is that DE intake
was slightly overestimated. However, this omission makes
very little overall difference to DE intake because a relatively
small amount of energy is lost in this way in humans. Whole
body calorimetry studies over 24 h in healthy subjects have
measured the excretion of H2 and CH 4 (via flatus and breath)
and found it to be small (usually o1 l/day) (Poppitt et al.,
1996; King et al., 1998), with variable and sometimes an
inverse relationship between H2 and CH 4 excretion. This
may correspond, very approximately, to about 3–5% of the

combustible energy of the fermentable carbohydrates entering
the colon (0.50–0.85 kJ/g, sometimes rounded up to
1 kJ/g of fermentable carbohydrate) and 0.3–0.5% of total
combustible carbohydrate intake when 10% of it enters the
colon as fermentable carbohydrate. Most H2 and CH 4 are
excreted in flatus, especially at high rates of net gaseous
production, but a variable proportion is absorbed and excreted
unchanged in breath (cf. definition of DE above, which
involves absorption of gases produced during fermentation,
which contributes to the overall ‘digestibility’ of energy).
It should be noted that the digestibility and DE values are
mean values obtained in subjects without disease. It is
recognized that digestibility, and therefore DE (kJ/g), vary
between individuals and are affected by both age (for
example, the fractional absorption of some macronutrients,
is less in young infants than adults) and disease (malabsorption
disorders). There may also be interaction between
nutrients. For example, studies in rats suggest that guar
gum and some pectin preparations induce substantial faecal
fat loss, although quantitatively important interactions of a
similar kind were not observed in humans (Rumpler et al.,
1998). One study in healthy volunteers showed that an
increase in dietary fibre intake from 15 to 46 g/day, through
addition of pectin to the diet, increased fatty acid excretion
from 1.5 to 2.7 g/day (Cummings et al., 1979). Another study
reported a smaller increase in fatty acid excretion when
dietary fibre intake was increased from 17 to 45 g/day
through addition of wheat bran to the diet (Cummings
et al., 1976). The possibility that quantitatively greater
interactions occur with novel fats cannot be excluded.
Dietary fibre has also been repeatedly reported to increase
N excretion, but at least some of this N originates from urea,
which passes through the colonic mucosa to reach bacteria
and which also enters the colon through the ileocaecal valve,
which allows transfer of small intestinal fluid from the small
to the large intestine. Fat and protein (non-nitrogenous
component) in biomass are generally thought to be generated
from fermentable carbohydrate. High intakes of fibre
may also increase shedding of mucosal cells, containing
some fat and protein) into the lumen of the gut.
For carbohydrates that are considered to be fully digested
and absorbed by the small intestine (digestibility ¼ 1.0; for
example, glucose, fructose, lactose, sucrose and starch), DE is
identical to cIE. For carbohydrates that are not absorbed at
all, DE and digestibility are zero (although their cIE is similar
to that of other types of carbohydrates (kJ/g)).
The DE of carbohydrate that reaches the colon (‘unavailable
carbohydrate’) is lower than its cEI, partly because some
are not fermented (faeces generally contain some of this
carbohydrate) and partly because some of the products of
fermentation are lost to faeces (for example, as bacterial
biomass). Virtually, all the short-chain fatty acids (SCFAs)
produced during fermentation are absorbed through the
colon and metabolized by human tissues (butyrate is also
actively metabolized by the colonic mucosa). Guar gum,
which is completely fermented, contributes about 60% of its
Physiological aspects of energy metabolism
M Elia and JH Cummings
cIE to SCFAs; pectin, which is about 95% fermented
contributes a little less of its energy to SCFAs; and other
types of carbohydrates, which are poorly fermented (for
example, Solka-floc) contribute very little of their energy to
SCFAs (Livesey and Elia, 1995). For traditional foods (foods
listed in the 1978 edition of McCance and Widdowson’s
Composition of Foods (Paul and Southgate, 1978), a general
fermentability value of 70% seems reasonable (Livesey and
Elia, 1995; FAO, 1998), since 30% of the NSP is excreted in
faeces over a wide range of intakes. This is in agreement with
intestinal balance studies and net production rates of SCFAs,
which are almost entirely available for metabolism by
human tissues (Livesey and Elia, 1995), with lower efficiency
than glucose (Hine is used to indicate the heat released as a
result of this metabolic inefficiency relative to glucose). A
schematic diagram (Figure 2) can help understand how
standard reference values for DE (and ME) of NSP and other
carbohydrates reaching the colon (‘fibre’) are established,
assuming that 70% are fermented there. For historical
reasons, values for the energy values of ‘fibre’ largely
developed from meta-analyses of studies using either the
AOAC (2002) or Southgate (1969) analytic procedures, which
corresponds approximately to NSP þ resistant starch (RS).
DE ¼cIE cFE cGaE
¼17:0 8:7 0:6kJ=g
¼7:7kJ=g
The value for cFE (8.7 kJ/g) is the sum of the cIE of
unfermented carbohydrate (5.1 kJ/g or 30% of the total cIE
of the carbohydrate entering the colon; Figure 2) and
products of fermentation lost to faeces (mainly bacterial
matter) (3.6 kJ/g). Another way of calculating the same result
is to use a fractional intestinal balance approach. Since 70%
of carbohydrate that reaches the colon is fermented, of
which B65% is transformed into SCFAs (B60%) plus
gaseous energy (5%), then the following equation applies:
DE ¼17:0 0:7 0:65
¼7:7kJ=g
For simplicity, the figure assumes that 5% of the energy of
fermented carbohydrate is lost as fermentation heat and
another 5% as gaseous energy, although it is more likely to
be 7 and 3%. The results also vary between subjects. For
example, calculations based on balance studies carried out in
six subjects who were studied in a metabolic unit and whole
body calorimeter indicate that 3.872.8% of fermentable
energy (NSP þ RS) was lost as gaseous energy (H2 and CH 4)
(Poppitt et al., 1996). The values for individual types of
fermented carbohydrate differ, partly because the proportion
fermented varies and partly because products of fermentation
also vary (that is, bacterial biomass plus gaseous
products, which are excreted vs SCFAs, virtually all of which
are absorbed and metabolized by human tissues) (Livesey
and Elia, 1995).
S43
European Journal of Clinical Nutrition

S44
'Fibre'
17.0 kJ/g
Physiological aspects of energy metabolism
M Elia and JH Cummings
Energy kJ/g
Similar calculation procedures to the above can be used to
establish DE values of polyols. Like NSP, polyols are a
heterogeneous group of substances (Bernier and Pascal,
1990; Japanese Ministry of Health, 1991; Livesey, 1992;
FASEB 1994; Finley and Leveille, 1996) that have different
physiological properties. For example, glycerol, like glucose,
is fully absorbed (digestibility ¼ 1.0), and therefore its DE is
the same as its cEI (18.0 kJ/g). Although other polyols, such
as isomalt, lactitol, maltitol and erythritol, have a cEI of
16.70–17.20 kJ/g, their DE values (11.0–16.61 kJ/g) are 65–
97% of their cEI. In contrast, virtually, all the lactitol in food
escapes digestion and absorption by the small bowel and
finds its way into the large bowel, where about 60% of its cIE
is converted to SCFAs, which are subsequently absorbed by
the colon (cIE, 17.0 kJ/g; DE, estimated to be about 11.1 kJ/g
(DE ¼ 0.65cIE)). With the increasing use of different types of
polyols in foods, it is becoming apparent that their digestibility,
the extent to which they are absorbed by the small bowel and
the extent to which they are subsequently metabolized by
human tissues varies so much (see below) that it might be
appropriate to use individual rather than generic energy values.
A further potential complexity is that polyols may be handled
differently by the gut (resulting in different digestibility and DE
values) when taken in foods compared to drinks. However,
food regulations normally prohibit the use of polyols in drinks.
5.1
3.6
0.6
0.6
7.1
Unfermented fibre
Bacterial matter
H2 and CH4 Fermentation heat
Hine SCFA (1.2 kJ)
Short chain fatty acids
(SCFA)
Faecal energy (8.7 kJ)
Gaseous energy (0.6 kJ)
Metabolisable = Digestible
energy (7.7 kJ)
An exception is the use of erythritol, which is hardly
metabolized once absorbed and hardly fermented when
reaching the colon. An issue for further consideration is
whether low-dose polyols should be used in drinks.
Metabolizable energy
By convention, the metabolizable dietary energy (ME) is measured
at zero nitrogen and energy balance. In these circumstances,
ME is the component of DE that produces heat during oxidative
metabolism. It does not include energy that is lost to urine
(combustible urinary energy (cUE; for example, urea, which is a
partially combusted end product of protein metabolism or
unmetabolized urinary polyols) or body surfaces (combustible
surface energy (cSE; for example, desquamated cells, hair loss,
perspiration), because this energy is not used in metabolism.
Therefore, ME can be defined mathematically as follows:
ME ¼ DE cUE cSE
Since DE ¼ cIE–cGaE, the following also applies:
ME ¼ cIE cGaE cUE cSE
H ine 'fibre'
(1.8 kJ)
Net
metabolisable
energy
(5.9 kJ)
Figure 2 The fate of ‘fibre’ ingested with conventional foods. It is assumed that 70% is fermented, so that 5.1 kJ/g (30% of the combustible
intake energy (cIE) is lost to faeces unchanged. Of the fibre that is fermented, 5% of the cIE energy is lost as gaseous products (H2 and CH4), and
another 5% as heat of fermentation, which contributes to metabolizable energy. The majority (60% of cIE) is converted to short-chain fatty acids,
almost all of which are absorbed and metabolized by human tissues, or lost to faeces (30% of cIE), mainly as bacterial biomass. Of the energy
initially present in total fibre (fermentable and non-fermentable) 51.0% is lost to faeces, 3.5% to gaseous products and the remaining 45.5%
accounts for metabolizable energy.
European Journal of Clinical Nutrition
In normal healthy subjects, loss of surface energy is small
and can be ignored. For many carbohydrates (for example,

glucose, starch, NSP) there is also negligible loss of cIE to
urine, and so cUE can also be ignored. Under these
circumstances (which do not apply to many polyols), the
following equation is valid
DE ¼ ME
For carbohydrates that are fully absorbed and fully metabolized
to CO2 and H 2O (‘available carbohydrates’, for
example, glucose, starch), the following equation also
applies:
cIE ¼ DE ¼ ME
For other types of carbohydrates, ME is less than cIE, either
because there is some loss of cIE to faeces (cFE) and/or urine
(cUE), as in the case of some polyols. Polyols, which are often
used as bulk-sweetening agents, have fewer detrimental
effects on teeth, lower glycemic indices and lower ME values
than conventional sweet sugars (for example, glucose,
fructose, sucrose). The proportion absorbed by the small
intestine varies considerably with the type of polyol (Bernier
and Pascal, 1990; Livesey, 1992; Finley and Leveille, 1996),
and tends to decrease with increasing molecular weight. As
much as 35% of the energy of polyols can be accounted for
in faeces (digestibility of energy, 65–100%; although polyols
are fermented and not recovered in faeces, some of the
products of this fermentation are recovered in faeces). In
addition, a variable proportion of absorbed polyols is lost in
urine. A detailed discussion about the extent of absorption
and metabolism of polyols by bacteria in the human colon
and by human tissues is beyond the scope of this paper.
However, the following brief points illustrate the variability
in their physiological handling. The small intestine absorbs
only about 2–5% of lactitol, which is almost completely
recovered in urine (95%), with little oxidation by human
tissues (suggested by studies involving intravenous administration
of C-labelled lactitol (Bernier and Pascal, 1990)).
The small intestine absorbs B25% of mannitol and, like
lactitol, it is almost totally excreted in urine without
metabolism by human tissues (again suggested by intravenous
administration of labelled (Nasrallah and Iber, 1969)
and unlabelled (Nasrallah and Iber, 1969) mannitol). In
contrast, glycerol is completely absorbed and metabolized by
human tissues, without excretion in urine. Sorbitol is
absorbed to the extent of B50%, of which the majority
(B85%) is metabolized by human tissues (Finley and
Leveille, 1996).
Most studies assessing ME of diets have been carried out in
subjects close to N and energy balance. The DE and ME
systems appear to overestimate DE and ME when energy
intake is low, especially when accompanied by high NSP
intake because apparent digestibility tends to be lower. Food
energy values reflect the supply of energy and not whether it
is spent. Some energy is deposited during growth and
development of obesity. The combustible energy that is
deposited in such situations corresponds to DE, which may
Physiological aspects of energy metabolism
M Elia and JH Cummings
not be made available to oxidative metabolism, if for
example it is retained during the growing process. Some of
it may not be made available even if it were mobilized
and metabolized, because incompletely combusted products
of metabolism, such as urea, are excreted in urine
(for protein, DE4ME). However, when carbohydrate is
mobilized, it is usually totally oxidized (combustible
energy ¼ DE ¼ ME). The issue is discussed further below—
‘Calculating energy balance’.
DE and ME values may be influenced by genetic factors
and disease. For example, although lactose is assigned a
digestibility coefficient of 1.0, so that cIE ¼ DE (also ¼ ME), in
subjects with lactose malabsorption (Matthews et al., 2005;
Montalto et al., 2006) (due to genetic factors that can affect
entire populations, or to damage of intestinal lactase by
gastroenteritis or enteropathy), not all the lactose is hydrolyzed
and absorbed by the small intestine. Some of it reaches
the colon, where it is fermented to produce faecal matter and
gaseous end products (both of which reduce DE and ME) and
SCFAs, which are absorbed and metabolized by human
tissues. Another example concerns fructose, which when
ingested in small quantities is fully absorbed by the small
intestine, but when ingested in large quantities, can reach
the large bowel, where it is fermented (Truswell et al.,
1988). This fermentation results in some loss of energy
to faeces (cFE) (for example, as bacterial biomass), which
again reduces its DE and ME values. Yet another example
involves loss of glucose in the urine of diabetic patients
(metabolizability of o1.0; and MEoDE). Since DE and ME
values vary between individuals and are affected by disease,
the values in general use are based on average results
obtained in ‘healthy’ subjects using doses of substrates that
are likely to be ingested. Any errors in establishing DE values
will not only affect ME values but also NME values, which is
discussed next.
Net metabolizable energy
A criticism that has been raised against the ME system is that
it fails to consider differences in the efficiency with which
substrates supply biologically useful energy. The NME system
is based on the concept of biological efficiency at the level of
ATP and takes into account two specific processes, both of
which have the effect of reducing metabolic efficiency. First,
the heat of fermentation does not yield ATP to the host.
Since this lowers the overall bioenergetic efficiency of the
fermentable substrate, some workers have suggested using
the term ‘true metabolizable energy’, which they have
defined as ME minus the heat of fermentation (Bar, 1990;
Bernier and Pascal, 1990). This approach, which is used in
animal nutrition, differs from that in human nutrition,
where the heat of fermentation is included in ME. Second,
different fuels yield net ATP (ATP gain) with different
bioenergetic efficiencies (Elia and Livesey, 1988, 1992;
Blaxter, 1989). Those with lower efficiencies (for example,
protein) will result in more heat production for the
S45
European Journal of Clinical Nutrition

S46
same useful metabolic or external physical work done than
in those with higher efficiencies (that is, more heat is
released for the same ATP gain). The NME system aims to
make the energy values of all fuels equivalent at a
biochemical level (isobioenergic). It does this by adjusting
ME values to take into account differences in the energetic
efficiency with which net ATP is generated. A theoretical
approach, which is discussed first, has been found to agree
remarkably well with an empirical approach based on
measurements obtained by indirect calorimetry, which is
discussed later.
The energy equivalent of ATP is the heat generated
during oxidation of a substrate (ME) divided by the
number of moles of ATP gained through specified
oxidative metabolic pathways. To understand this,
it is necessary to clarify four points (Elia and Livesey,
1988, 1992).
(1) Application of NME does not require an understanding
of how ATP is generated, any more that
than application of ME requires an understanding
of the complex processes of digestion and fermentation,
which are quantitatively poorly understood and
variable.
(2) ATP gain does not mean that ATP accumulates: ATP has a
small pool size with a very rapid turnover. Therefore, the
term ‘ATP gain’ refers to the ATP made available to the
body for metabolic or external work.
(3) Some pathways involve both production and utilization
of ATP. For example, during glucose oxidation, one ATP
is obligatorily used in the initial activation of glucose to
glucose 6-phosphate and another for the conversion
of fructose-6-phosphate to fructose-1,6-biphosphate.
Therefore, two extra moles of ATP are produced than
are gained. The term ‘ATP gain’ refers to the net gain of
ATP (total ATP produced minus ATP utilized obligatorily
during the oxidation of a substrate).
(4) ATP gain, which refers to the net provision of energy
in the form of ATP, has to be distinguished from
biochemical processes, including substrate cycling
(for example, fatty acid-acetyl-CoA recycling or
glucose-lactate recycling) that utilizes ATP. The two
should not be confused because they are on different
sides of the ATP balance equation (ATP gain ¼ ATP
utilized).
The NME system is based on the concept of relative
metabolic efficiency. Although uncoupling of oxidative
phosphorylation increases the amount of heat released per
ATP gained, this affects all macronutrients, so that the ratio
of metabolic efficiency between two macronutrients is
expected to change much less than that associated with a
single nutrient. This is confirmed quantitatively by theoretical
models of progressive uncoupling (Livesey, 1984, 1986).
By convention, the NME system expresses efficiency of
nutrients relative to glucose, which is assigned an efficiency
coefficient of 1.0 (kglucose ¼ 1.0)(Dutch Nutrition Council,
European Journal of Clinical Nutrition
Physiological aspects of energy metabolism
M Elia and JH Cummings
1987; Blaxter, 1989; Livesey and Elia, 1995).
k ðnutrientÞ ¼ kJðMEglucoseÞ per ATP gain=kJ ðMEnutrientÞ per ATP gain
¼ ATP gain per kJ ðMEnutrientÞ=ATP gain per kJ ðMEglucoseÞ
The following equations show how the coefficient, knutrient,
is related to its NME (NME nutrient).
The lower the value of k nutrient, the greater the metabolic
inefficiency relative to glucose and the greater the amount of
heat released for the same ATP gain (Livesey, 1984; Elia and
Livesey, 1988, 1992; Livesey and Elia, 1995).
The term Hine is the extra heat released as a result of
metabolic inefficiency relative to glucose (that is, extra heat
released for the same ATP gain). H ine is therefore a
component of total heat release.
NMEnutrient ¼ knutrient ME
¼ MEnutrient ðð1 knutrientÞ MEnutrientÞ
¼ MEnutrient Hine
Since H ine is the heat released as a result of the metabolic
inefficiency (relative to glucose) (H ine ¼ (1 k nutrient)
ME nutrient), 1 k nutrient can be regarded as the metabolic
inefficiency coefficient relative to glucose. The overall values
for k (and therefore 1 k) for carbohydrates fermented in the
large bowel takes into account not only the relative
metabolic inefficiency of oxidizing the products of fermentation
(SCFAs are absorbed and oxidized by human tissues),
but also the heat of fermentation, which is not associated
with ATP gain to the host. Both of these elevate the overall
energy equivalent of ATP gain relative to glucose, and
decrease the overall value for knutrient (or increase the
inefficiency coefficient, 1 k nutrient), with a resulting increase
in heat production (H ine).
The metabolic efficiency with which fats are oxidized is
98% that of glucose, proteins about 80%, alcohol about 90%,
SCFAs (direct oxidation) about 85–90%. Various scholars and
committees have established theoretical values of efficiency
of substrate oxidation relative to glucose from the stochiometry
of metabolic pathways (summarized by G Livesey)
with the following ranges: protein (Blaxter, 1971, 1989;
Schulz, 1975, 1978; Flatt, 1978, 1980, 1987, 1992; Livesey
and Elia, 1985; Life Sciences Research Office, 1994; Black,
2000), 0.78–0.85 (mean, 0.81); fat (Armstrong, 1969; Schulz,
1975, 1978; Flatt, 1978, 1980, 1987, 1992; Livesey and Elia,
1985; Blaxter, 1989; Black, 2000), 0.97–1.01 (mean 0.98);
fermentable carbohydrate (British Nutrition Foundation,
1990; Life Sciences Research Office, 1994; Livesey, 2002),
0.74–0.75 (mean, 0.74); and mixed short-chain organic acids
(Armstrong, 1969; Livesey and Elia, 1985; Dutch Nutrition
Council, 1987; Livesey, 1992; Life Sciences Research Office,
1994), 0.83–0.87 (mean, 0.85). In the case of oxidation of
SCFAs by human tissues, this means that about 10–15%
more heat is released for a given ATP gain than when glucose
is oxidized to yield the same ATP gain. However, in subjects
ingesting 80 g protein and only 20 g ‘fibre’, the excess heat

generated through bioenergetic inefficiency during protein
oxidation is B10-fold greater than during SCFA oxidation
(B5-fold greater than the overall Hine of fibre, which
includes the heat of fermentation). This is mainly because
more protein is oxidized than in SCFAs, and partly because
the bioenergetic inefficiency in generating ATP is greater
during protein oxidation than in SCFA oxidation. The
inefficiency associated with the metabolism of NSP as a
whole not only takes into account the bioenergetic inefficiency
of oxidizing SCFAs but also the heat of fermentation
(overall metabolic efficiency of fermentable NSP B72.5%
that of glucose). Even so, the heat released as a result of
metabolic inefficiency relative to glucose (Hine) remains
quantitatively more important with protein than with NSP
mainly because much more protein than NSP is usually
ingested. However, in subjects ingesting large amounts of
NSP and relatively little protein, as in some low-income
countries, the heat released as a result of metabolic
inefficiency associated with these two macronutrients can
be almost equivalent.
To examine the validity of the NME system, theoretically
calculated k coefficients were compared to those based on
experimental observations obtained in human studies that
involved calorimetry (for example, whole body 24 h calorimetry),
N balances, and in some cases measurement of H2
and CH 4 excretion (Livesey, 2001). In these studies, measurements
of energy expenditure were undertaken in subjects
given a control diet or a test diet, which differed from the
control diet in that available carbohydrate was exchanged
for a test substrate (for example, protein, which has a lower k
coefficient than glucose). Test diets, which had lower kdiet
coefficients than control diets, would be expected to be
associated with greater energy expenditure (more heat
production). The increased heat production (H ine) on the
test diet (after adjusting for energy and N balance) was used
to calculate k(nutrient) ((ME–H ine)/ME), assuming that ATP
gain remained unchanged under the conditions of the study,
which involved undertaking a fixed pattern of physical
activity in ambient temperatures of thermal comfort. The
mean results obtained by the theoretical and experimental
approaches are strikingly similar (Table 2). The same type of
approach has been used to establish coefficients for k for the
NME system in animals, and again the similarities between
theoretical and experimental approaches are striking (Livesey,
2001). However, as is the case for DE and ME evaluation
of foods, time-consuming studies cannot realistically be
undertaken to establish efficiency values for large numbers
of foods reported in food tables. Therefore, some of the
recent developments and applications of the human NME
system have relied heavily on the theoretical approach.
Detailed theoretical considerations about potential alternative
pathways for generating ATP gain from the same
substrate are beyond the scope of this paper (for example,
cytosolic vs mitochondrial activation of SCFAs, net lipogenesis
from glucose; Elia and Livesey, 1988; Livesey and
Elia, 1995; and oxidation of alcohol with alcohol dehydro-
Physiological aspects of energy metabolism
M Elia and JH Cummings
Table 2 A comparison of experimentally obtained and theoretical
values of metabolic efficiency of macronutrients relative to glucose
(k nutrient) a
Macronutrient
(number of studies)
Relative metabolic efficiency (k nutrient)
Experimentally
obtained
mean7s.e.m.
Theoretical
Fat (n ¼ 8) 0.9770.01 0.98
Protein (n ¼ 9) 0.7970.02 0.80
Fermentable
unavailable
carbohydrate (n ¼ 8)
0.7170.09 0.76
Alcohol (n ¼ 3) 0.9070.04 0.90
a Based on Livesey (2001).
genase or mixed function oxidase). However, alternative
pathways are considered to have small effects on the overall
energy equivalents of ATP. In addition, some processes may
also operate only in unusual circumstances. For example, 24h
whole-body indirect calorimetry shows that it is unusual
for individuals ingesting a western-type diet to achieve a
non-protein respiratory quotient above 1.0, which would
indicate net lipogenesis from carbohydrate (although persistent
overfeeding with large amounts of carbohydrate can do
this, for example, during nutritional repletion of malnourished
subjects; Pullicino and Elia, 1991). The energy
equivalent of ATP associated with low to moderate intake
of alcohol was established using the pathway involving
alcohol dehydrogenase, and its validity confirmed by
calorimetry studies (Table 2).
The NME system has been supported by many national
committees in different countries, and the FAO has recommended
that it should be considered for food labelling, for
food tables and when calculating practical food needs from
energy requirements. General values for the major macronutrients
and specific values for different types of carbohydrates
are shown in Table 3 (Livesey, 2003a), together with
combustible digestible and metabolizable conversion factors.
Variability in energy values
DE, ME and NME values for specific carbohydrates do not
necessarily apply in all circumstances. For example, in some
individuals, carbohydrates may be totally metabolized by
human tissues after complete absorption by the small
intestine, whereas in others, they are incompletely digested
and absorbed by the small intestine (for example, lactose
malabsorption, which can be elicited more readily certain
ethnic groups). In the latter situation lactose undergoes
colonic fermentation, with some converted into cFE, cGaE
and fermentation heat. The absorption and oxidation of
SCFAs derived from fermentation of lactose is associated
with higher energy equivalents of ATP (less ATP gain) than
when lactose is hydrolyzed and oxidized directly by human
S47
European Journal of Clinical Nutrition

S48
Table 3 Energy values of macronutrients and specific carbohydrates in
foods (kJ/g) a
cIE DE ME NME
Macronutrients
Fat 39.3 37.4 37.4 36.6
Protein 23.6 21.7 16.8 13.4
Available carbohydrates b
15.7 15.7 15.7 15.7
‘Fibre’/NSP 17.0 7.7 7.7 6.0
Alcohol 29.6 29.6 29.0 25.8
Specific carbohydrates
Available carbohydrates
Glucose monohydrate 14.1 14.1 14.1 14.1
Glucose 15.7 15.7 15.7 15.7
Fructose 15.7 15.7 15.7 15.2
Lactose 16.5 16.5 16.5 16.3
Sucrose 16.5 16.5 16.5 16.3
Starch 17.5 17.5 17.5 17.5
Fibre
Fermentable ‘fibre’ 17.0 11.0 11.0 8.0
Non-fermentable ‘fibre’ 17.0 0.0 0.0 0.0
Resistant starch 17.5 11.4 11.4 8.8
Non-digestible oligosaccharides
General, conventional foods
Isolated/synthetic
17.0 11.1 11.1 8.4
Fructooligosaccharides 17.0 11.1 11.1 8.4
Synthetic polydextrose
(5% glucose)
16.9 6.6 6.6 5.2
Inulin (pure) 17.5 11.4 11.4 8.8
Polyols
Erythritol 17.2 16.6 1.1 0.9
Glycerol 18.0 18.0 18.0 16.6
Isomalt 17.0 11.6 11.2 8.9
Lactitol 17.0 11.1 10.7 8.2
Maltitol 17.0 13.4 13.0 11.5
Mannitol 16.7 12.3 8.1 6.3
Polyglycitol 17.1 13.5 13.2 11.6
Sorbitol 16.7 12.0 11.7 9.7
Xylitol 17.0 14.0 13.7 12.4
Abbreviations: cIE, combustible intake of energy; DE, digestible energy; ME,
metabolizable energy; NME, net metabolizable energy; NSP, non-starch
polysaccharides.
a Based on Livesey (2003a). The values for macronutrients have also been
reported in Livesey (2001), with some trivial differences in the values for
protein and alcohol. For available carbohydrates, see also Livesey and Elia
(1988) and Elia and Livesey (1992). For fibre see text.
b Available carbohydrate (as monosaccharide equivalents) measured by direct
analysis of sugars. When carbohydrate is measured ‘by difference’, the values
increase by 1 kJ/g.
tissues. The overall result is that DE, ME and NME values for
lactose, when there is lactose malabsorption, are lower than
cIE, and lower than the usual values assigned to it, which
assume that all lactose is absorbed from the small bowel
(16.5 kJ/g for cIE, DE and ME values, and 16.3 kJ/g for the
NME). However, this problem is likely to be a minor one,
since individuals with lactose intolerance do not normally
ingest large amounts of lactose because it can produce
undesirable bloating effects and diarrhoea.
European Journal of Clinical Nutrition
Physiological aspects of energy metabolism
M Elia and JH Cummings
The standard DE, ME and NME values obviously
overestimate the actual values in patients with malabsorption
disorders to an extent that depends on the
type of malabsorption disorder and its severity. In contrast,
the standard values may underestimate the actual
values when nutrients are given intravenously and made
directly available to human tissues. This problem is
addressed in detail elsewhere (Livesey and Elia, 1985,
1988). However, the energy value for glucose, which is
usually the dominant energy source for intravenous
nutrition, remains unaltered, but this may not be so for
other types of carbohydrates that may be lost in urine to a
variable extent.
Calculating energy balance
There is interest in estimating energy balance during growth,
development of obesity and undernutrition, as well as
during their treatment. The energy balance can be calculated
from changes in body composition (Fuller et al., 1992;
Heymsfield et al., 2005), which can provide estimates of fat
and protein mass, and in some cases carbohydrate (glycogen
stores are small, but can be measured accurately and
precisely using nuclear magnetic resonance; Taylor et al.,
1992; Casey et al., 2000). Energy balance can also be
calculated from classic balance studies, which involve
measurement of energy intake and energy expenditure. In
calculating energy balance, it is important not to confuse the
energy values of endogenous and exogenous fuels and also
not to confuse different food energy systems (combustible,
ME and NME systems; Table 4).
The ME and NME values of nutrients (kJ/g) that are
mobilized from body stores and used for oxidation during
weight loss are higher than the corresponding food energy
values (Table 4), mainly because digestibility does not apply
to the former but it does to the latter. For example, the ME
values of fat in food is 37.4 kJ/g compared to the value of
39.3 kJ/g of stored fat, and for food protein 16.8 kJ/g
compared to 18.4 kJ/g of tissue protein.
Using body composition techniques, the energy balance
equations is
Energy balance ¼Energy storestime point 2 energy storestime point 1
¼energytime point 2ðcarbohydrate þ fat þ proteinÞ
energytime point 1ðcarbohydrate þ fat þproteinÞ
The energy values for all macronutrients in this equation are
the endogenous food energy values and not the exogenous
food energy values (see Table 4 for combustible energy, ME
and NME values). The body composition approach is usually
only of value in long-term studies where there are large
changes in body composition, which far outweigh measurement
precision. In some such studies, the changes in
glycogen stores are small and therefore they are either be
ignored or estimated. The errors associated with this
approach have been assessed (Elia et al., 2003).

Table 4 cE, ME and NME, and content of macronutrients in food and in
body stores
The classic energy balance technique involves measurements
of both total energy expenditure, using calorimetry
(whole-body calorimetry) or tracer techniques (the doubly
labelled water technique (Prentice, 1990) or the bicarbonate–
urea method (Elia et al., 1995; Gibney et al., 1996, 2003;
Paton et al., 1996, 2001; Tang et al., 2002; Elia, 2005) and
energy intake.
The energy balance equation can be considered in terms of
combustible energy or ME.
(i) Energy balance (combustible energy): cEstored ¼ cIE–
heat–cEnergy losses where cEnergy losses are the
combined losses through faeces urine, body surfaces
and gases (cFE þ cUE þ cSA þ cGaE). cIE can be measured
using bomb calorimetry, but can also be estimated
(Table 3).
(ii) Energy balance (ME): MEstored ¼ ME intake–heat
Since the following two equations apply,
ME stored ¼NME=kstored
ME intake ¼ME intake=kintakeðdietÞ ‘Food’ energy Body stores a
cE kJ/g ME kJ/g NME kJ/g cE kJ/g ME kJ/g NME kJ/g
Fat 39.3 37.4 36.6 39.3 39.3 38.5
CHO b
Available 15.7 15.7 15.7 15.7 15.7 15.9 c
Unavailable 17.0 7.7 5.6 — — —
Protein 23.6 16.8 13.4 23.6 18.4 o14.7 d
Alcohol 29.6 29.0 25.8 — — —
Abbreviations: cE, combustible energy; ME, metabolizable energy; NME, net
metabolizable energy.
a The values indicated assume mobilization and oxidation of ‘stored’
macronutrients. The starting points are the body stores (which are lost during
starvation), and as such represent a departure from food energy systems.
b CHO as monosaccharide equivalents.
c The value, which is higher than that of glucose, takes into account the heat of
hydrolysis of glycogen and one extra ATP gained per glucosyl residue during
oxidation of glycogen compared to oxidation of glucose.
d Protein breakdown involves both ATP-independent pathways (in which case
the value of B14.7 kJ/g applies) and ATP-dependent pathways (in which case
the value o14.7 kJ/g applies).
the energy balance equation can also be expressed in relation
to NME.
Energy balance (NME): NME/k stored ¼
ME intake/k intake(diet)–heat
In these equations, the heat released (energy expenditure)
is again measured by calorimetry in controlled laboratory
conditions or by tracer techniques in free-living conditions.
ME intake in this situation is obviously calculated from the
exogenous food energy values (Table 4). Parenthetically, it
should be noted that in short-term studies involving
external work, not all of the heat produced is lost from the
Physiological aspects of energy metabolism
M Elia and JH Cummings
body (some is transiently retained in the body, causing an
increase in body temperature).
Energy systems and food labelling
Food energy systems have evolved over time, as fundamental
physiology of macronutrient metabolism was reviewed. For
example, a value of 17 kJ/g (Atwater general factor) was
assigned to the ME for total carbohydrate including ‘fibre’ (or
unavailable carbohydrate) (this system also assigns values of
17 and 37 kJ/g of protein and fat, respectively) (Merill and
Watt, 1973). There are physiological objections to this: not
all carbohydrate that reaches the colon is fermented (some is
excreted unchanged in faeces); some is converted to
substances, such as combustible bacterial biomass in faecal
matter and combustible gases, which are lost from the body.
Therefore, more appropriate ME values for ‘fibre’/NSP have
been established (for example, ME of 7.7 kJ/g for that present
in conventional foods (see section on Digestible energy),
which can be rounded to 8 kJ/g and NME of 6 kJ/g). The same
evolution of thought occurred with polyols. Regulatory
agencies assigned a value of 17 kJ/g to polyols (as for other
carbohydrates) because none was recovered in faeces.
However, this did not agree with energy balance studies.
Basic physiological considerations (confirmed by experimental
studies) suggest that fermentation of polyols results
in some loss of energy in faeces and gaseous products (as
with ‘fibre’). In addition, absorbed polyols are frequently lost
in substantial quantities in urine. After further consultations
within the European Union, polyols were assigned a general
energy value of 10 kJ/g (European Communities, 1990).
However, because of the variability in ME values for different
polyols, European countries (for example, The Netherlands,
France) as well as non-European countries (for example, the
United States, Canada, New Zealand and Japan) have
adopted or recommended a specific value for each polyol.
An FAO consultation suggested that where one polyol
represents a substantial source of energy in a product, use
of a more specific factor may be desirable (FAO, 2003).
Three food energy systems have been widely promoted for
use in food tables and for food labelling in different world
regions: those that employ the Atwater-specific factors
(Merill and Watt, 1973); those that employ the Atwater
general factors (Merill and Watt, 1973; FAO/WHO/UNU,
1985) (17, 17, 37 kJ/g for carbohydrate, protein and fat,
respectively); and those that employ NME factors (typically
applied to polyols). The use of all these systems is a source of
confusion, especially when factors from different systems are
used in the same country. A further source of confusion is
that in some countries, energy is expressed in relation to
weight of carbohydrate, and in others in relation to weight
of monosaccharide equivalents (B11% difference in the case
of starch). There is also the issue of accuracy. A particular
problem with the Atwater general system (Merill and Watt,
1973) is that it appears to overestimate the energy values of
specific food items, to the extent of 0–30%. The bias is
S49
European Journal of Clinical Nutrition

S50
Physiological aspects of energy metabolism
M Elia and JH Cummings
especially marked for food items containing a high proportion
of ‘fibre’ and certain types of protein, both of which may
be prominent in weight-reducing diets (Brown et al., 1998). A
modification of the Atwater general system that takes into
account the presence of unavailable carbohydrate in foods
(Livesey, 1991) has the advantage of overcoming the inaccuracy
of the Atwater general system, while avoiding the complexity of
the Atwater-specific system (Merill and Watt, 1973).
Terminological and methodological issues in carbohydrate
analysis are also important, due to the varied nature of the
techniques that are in use. These are dealt with in other
papers in this consultation (for example, see Measurement of
dietary carbohydrates for food tables and labelling by Englyst
et al., 2007). All carbohydrates can be determined by direct
analysis. Carbohydrate ‘by difference’ (total carbohydrate
unavailable carbohydrate) requires assumptions about the
fate of carbohydrates in the colon that cannot always be
predicted because this will, for example in the case of
starches, be dependent on the physical form of the food,
cooking and cooling, the degree of ripeness and nature of the
starch granule. Therefore, ‘by difference methods’ should not
have a place in carbohydrate analysis. Compared to the
direct method, the indirect method (‘by difference’) has the
effect of increasing the cEI, DE, ME and NME values of
available carbohydrates by B1 kJ/g. As already suggested, the
direct analysis is preferable (FAO, 1998), although this may
pose practical problems in developing countries with poor
access to analytical facilities. Finally, it should be remembered
that food labelling applies to food items not diets, and
that some foods may acquire different properties and
different energy values before being ingested. For example,
starch in potato can become resistant if it is cooked and
allowed to cool before being ingested (Englyst and
Cummings, 1987). In contrast, RS changes into ordinary of
starch during the maturation of bananas, which occurs whilst
their skins turn from green to yellow. Although these processes
will affect DE, ME and NME values in different ways, food
labelling cannot adequately take them into account.
The idea that bioenergetic inefficiency results in heat
production is not new, since it was already in existence in
early part of the twentieth century. It is also well known that
agents such as 2,4-dinitrophenol, which uncouple oxidative
phosphorylation, produce a considerable amount of heat.
Therefore, measurement of energy expenditure (heat production)
does not necessarily reflect useful metabolic or
external work done. The idea of establishing bioenergetic
equivalence of different macronutrients has obvious attractions
to food labelling. The NME (but not the ME) system is
based on the concept that macronutrients contribute
equivalently to energy balance and energy requirements
irrespective of food composition. The FAO acknowledges
that the Atwater-specific and general systems do not take
into account Hine, which is relevant to calculations of energy
requirements, and in the annexe of its report on food energy
(FAO, 2003), it provides two tables to address this issue
(Table 3.1 provides correction factors when ME intakes are
European Journal of Clinical Nutrition
being matched with requirements and Table 3.2 expresses
food needs as the ‘food NME requirement’).
The current labelling of carbohydrates (and more broadly
foods in general) is unsatisfactory because the ME system is
used for some carbohydrates (usually the most abundant
carbohydrates, such as starch and sucrose), but the NME
system is used for polyols and polydextrose in many regions
(and in the same regions that use the ME system). It would
seem more logical to use the same energy system for all
carbohydrates, and indeed for all other nutrients. It has been
argued that since NME values are only B96% of ME values
(NMEB ¼ 0.96 ME), it makes little difference which system is
used (especially since energy intake cannot generally be
assessed with an accuracy of less than 4%). On the other
hand, values for specific foods can be affected by considerably
more (up to 30%; for example, those rich in certain
types of protein and ‘fibre’ are used for weight reduction).
These errors may fall outside the range permitted by
legislation in the European Union (Commission, 2006). In
addition, by comparison with the NME system different diets
of the same ME value can be constructed that would result in
differences of 10–15% (for example, diets rich in NSP or
protein, as for example energy-restricted diets in which
normal protein intake represents a high proportion of
energy). Therefore, it would seem sensible to use a consistent
method that is sound at all levels, so that the consumer, the
nutritionist and the researcher are not misled. In addition, it
should be remembered that physiological considerations
that make a difference of only a few percent have already
been taken into account in devising food energy systems, for
example DE of the diet is greater than ME by a few percent,
which is similar to the extent to which ME is greater than
NME. Furthermore, as indicated above, early inaccuracies in
energy values for ‘fibre’/NSP and polyols have been
corrected, even though the changes make only a few percent
difference to the total energy value of the diet. An early
example of implementing small changes involves the Atwater
energy factors that replaced the earlier Rubner factors,
for example 2.5% difference for combustible energy of
carbohydrate. Another argument is that errors associated
with measurement of dietary intake and potential errors
associated with the development of RS after cooking are
sources of greater error than the use of the ME system instead
of the NME system. This poses a problem for analysts and
those compiling food composition tables, However, it is
reasonable to suggest that food-labelling policy should aim
to establish a system that avoids bias as far as possible using a
practical user-friendly system. It has also been argued that an
enormous amount of work would have to be undertaken to
change the current ME system into an NME system. On the
other hand, if an attempt is to be made to establish either a
consistent ME or an NME system, some changes will be
necessary. If this involves conversion to a consistent NME
system, the additional changes may not be as great
as anticipated, as the following three equations (round
numbers) indicate.

ME (Atwater general system) (Merill and Watt, 1973) ¼
37F þ 17CHO total þ 17P.
ME (modified Atwater general system) (Livesey, 1991) ¼
37F þ 16CHOavailable* þ 8CHO unavailable þ 17P.
NME (Livesey, 2001) ¼
37F þ 16CHOavailable* þ 6CHO unavailable þ 13P.
In these equations, ME and NME are in kJ, the coefficients
are in kJ/g, and F, P and CHO are in g of fat, protein and
carbohydrate, which may be available or unavailable ( * ¼ as
monosaccharide equivalents).
The key issues are the extent to which the underlying
physiological considerations are sound, the extent to which
they should drive food-labelling policy and the extent to
which food energy system(s) should become consistent
within and between regional jurisdictions.
Carbohydrates, feeding behaviour and energy
balance
The above discussion regarding energy values of carbohydrates
and energy balance equations gives no indication
about physiological processes that control voluntary food
intake and energy expenditure in different environments.
Do carbohydrates play a special role in processes that affect
energy balance? Many theories of appetite regulation and
energy homeostasis have been proposed, with carbohydrates
playing a key role in several of them. None of these theories
are universally accepted. Some have evolved over time and a
few are continuing to do so. The literature is vast, and
therefore only a brief overview of certain aspects of
carbohydrates physiology will be discussed in this paper.
Since there seems to be so much confusion and conflicting
advice being given to the consumer about the types of foods
that should be used for weight control and energy homeostasis,
the origins of some concepts and the mechanisms
involved are also discussed. This provides a platform for
considering future research. It is worth emphasizing that
despite the widespread use of a large number of diets that
vary in macronutrient composition, there is continued
growth of obesity and overweight. If a diet was associated
with long-term compliance and overwhelmingly successful
reduction in weight, a decrease in the prevalence of
overweight and obesity might be expected. Several randomized
and non-randomized controlled trials have reported
weight reduction over several months when low fat (highcarbohydrate
diets) are used. A review of four separate metaanalyses
of low-fat vs control diets (or the relationship
between the fat content of the diet and weight loss) (Astrup
et al., 2002) reported greater weight loss with the low-fat
diets, and one meta-analysis reported a dose–response
relationship between reduction in percent dietary fat intake
and weight loss. And yet, these results raise a conundrum,
firstly because the meta-analyses consistently show weight
Physiological aspects of energy metabolism
M Elia and JH Cummings
loss (generally only slight to modest mean weight loss of
B2–3 kg), whereas in the general population there is
progressive weight gain; and secondly because the percent
fat intake in the United Kingdom (Henderson et al., 2003)
and the United States (Centres for Disease Control and
Prevention (CDC), 2004) appears to have decreased in recent
years, while obesity and overweight have increased. Several
questions arise from these simple observations. Would the
growth of obesity have been greater if there was no decrease
in the percent fat intake? Is the study population representative
of the general population (for example, one metaanalysis
included separate studies of patients with breast
dysplasia, breast cancer and hypercholesterolaemia (Astrup
et al., 2000))? Are confounding variables such as physical
activity adequately controlled for? Is the dietary compliance
under the study conditions of randomized controlled trials
better than non-study conditions? If the studies had been
extended over longer periods, would there have been a
decrease in dietary compliance (for example, one systematic
review included studies that lasted for only 3 weeks; Yu-Poth
et al., 1999)? Are the effects of these diets mediated by energy
density rather than macronutrient composition (low-fat
diets, especially those rich in fibre tend to have a low-energy
density)? Do the types of carbohydrates used in the diets of
studies carried out over the past 20 or more years reflect
those currently ingested (for example, carbohydrates in soft
drinks and other food items have changed over time)? This
section examines some of these issues from a physiological
perspective, focusing on appetite regulation. It is of course
recognized that there is considerable scope for interaction
between feeding behaviour and other confounding factors
that affect energy expenditure. This section does not aim to
provide a detailed summary of the effectiveness or advantages
and disadvantages of the large number of popular
weight-reducing diets (reviewed elsewhere; Freedman et al.,
2001), some of which may suit certain individuals more than
others. It primarily aims to establish some links between
physiology and clinical/public health nutrition, through
examination of feeding behaviour, which has a key role in
energy homeostasis. To do this, it is first necessary to place
the scientific issues in perspective by considering the extent
to which energy balance is regulated in man.
The extent to which energy balance is regulated
Figure 3 shows that the cumulative energy intakes of a
reference male and reference female, both of which were
established using the UK-recommended ME intakes (Department
of Health, 1991). The figure indicates that 231 GJ
(231 000 MJ) are ingested during a lifetime of 75 years by the
reference male, and 196 GJ (196 000 MJ) by the reference
female. Since the ME of stored fat is B39.4 MJ/kg and that of
fat-free tissue is B3.8 MJ/kg (of which o0.1 MJ/kg is in the
form of glycogen), it can be calculated that the total energy
content of a 65 kg man (15% fat) is 0.594 GJ. This
corresponds to 15.6% of the annual energy intake of a man
S51
European Journal of Clinical Nutrition

S52
Cumulative energy intake
(GJ)
Cumulative energy intake
(GJ)
250
200
150
100
50
0
200
150
100
50
Males
CHO
aged 20–50 years. An energy imbalance that halves the body
mass index (BMI) from 20 to 10 kg/m 2 (barely compatible
with life) and that halves body weight from 65 (15% fat;
7.25 kg fat) to 32.5 kg (1.54% fat or 0.5 kg fat) corresponds to
only 0.49 GJ. This is equivalent to only 0.21% of the total ME
intake over a lifetime (or 0.42% of the ME of carbohydrate
intake over a lifetime). It is also equivalent to 0.26% of the
intake during adult life, and 1.3% of the intake over 10 years
of adult life (for example, between 30 and 40 years of age).
The dot by the side of Figure 3 is scaled to correspond to
approximately 1% of the final cumulative intake. These
considerations suggest that any effective regulatory physiological
system (without assumptions about the type or level
of regulation or specific drivers used in regulation; Stock,
1999; Spiegelman and Flier, 2001; Liu et al., 2003; Druce and
Bloom, 2006) must achieve much better results than the
small percentages indicated above, if good health is to be
maintained. Similar calculations can be made for increments
in body weight. An energy imbalance that doubles the BMI
of a man (from 20 to 40 kg/m 2 ) and doubles body weight
(from 65 kg (15% fat) to 130 kg fat (45% fat)) corresponds to
1.98 GJ, which is less than 1% of the total energy intake over
a lifetime (less than 2% of carbohydrate intake), almost
P
Fat
0 10 20 30 40
Age
50 60 70
Females
Fat
CHO
0
0 10 20 30 40
Age
50 60 70
Figure 3 The cumulative energy intake (1 GJ ¼ 1000 MJ) of
reference male and female (based on reference nutrient intake;
Department of Health, 1991) and its distribution between carbohydrate
(CHO; 55% of total intake), fat (35% of total energy intake)
and protein (P; 15% of total energy intake). The arrow points to a
dot that corresponds to approximately 1% of the cumulative energy
intake at 75 years, which if deposited, would double adult body
weight and body mass index.
European Journal of Clinical Nutrition
Physiological aspects of energy metabolism
M Elia and JH Cummings
P
.
.
exactly 1% if this occurred during adult life, and 2.5% if it
occurred over 30 years, between 20 and 50 years of age.
Again, with a view to maintain good health, any effective
regulatory system must achieve very much better results
than these calculations indicate. The alarming increase in
the number of grossly obese individuals in high-income
countries, and to a lesser extent in low-income countries,
argues against the operation of a tightly regulated physiological
system(s) of long-term energy homeostasis to maintain
optimal health in the modern environment. Short-term
regulation of energy balance is probably less precisely
regulated, perhaps reflecting an advantage during evolutionary
development, to store excess energy when food is
available for subsequent use when it is not readily available.
Carbohydrate-specific models of feeding behaviour
Carbohydrate-specific models of feeding behaviour have
been prominent in the last 50 years. Many models established
testable hypotheses, which have been examined, and
through this examination new concepts about energy
homeostasis have emerged. Mayer’s original glucostatic
theory (Mayer, 1955), which was published in 1955,
suggested that regulation of energy balance predominantly
involved short-term ‘glucostatic’ responses that could be
modified by longer term ‘lipostatic’ responses. Russek (1963,
1976) went on to establish the hepatostatic theory in 1963
by postulating the presence of receptors in the liver. Other
studies suggested that 6–12% fall in plasma glucose concentration
predicts meal initiation in animals (Campfield
and Smith, 1990; 2003) and humans (Campfield et al., 1992;
Campfield and Smith, 2003). Simple relationships do not
necessarily prove causality, although infusions of glucose in
animals to prevent the drop in circulating glucose concentrations
were found to delay feeding behaviour. Irrespective
of the regulatory mechanisms, there would be a clear
advantage in coupling oxidative metabolism (energy expenditure)
with feeding behaviour (energy intake). Over the last
25 years, arguments have been put forward that the
oxidation (or deposition) of specific macronutrients, especially
carbohydrates, have a disproportionately large effect
on feeding behaviour. The arguments are set out below.
(1) Macronutrients have different propensities to being
oxidized and stored. Alcohol is not stored at all, and
therefore it has to be oxidized (very little is excreted
unchanged). Protein is also readily oxidized, since little
of it, if any, is accreted into the tissues of non-growing
individuals in a range of pools with varying size, rates of
turnover and ‘lability’. Only a small proportion of
ingested protein is retained during development of
muscular hypertrophy from physical training and during
development of obesity (obese individuals have more
muscle to support the extra body, the heart hypertrophies
as its cardiac output increases and organs may also
enlarge). Carbohydrate stores, which are limited

(o0.8 kg in the adult), increase after ingestion of a mixed
meal (Elia et al., 1988), but much, if not all of the
increased store, may be mobilized for oxidative purposes
before the next meal. The increased post-prandial oxidation
of carbohydrate contrasts with decreased oxidation
of fat, which is the main storage form of energy
(Elia et al., 1988). Fat too can be oxidized before the next
meal, but excess energy intake over time is predominantly
stored as fat, which has very large potential
storage capacity, rather than carbohydrate, which has
a limited storage capacity. Therefore, the oxidative
hierarchy of fuels (alcohol, protein, carbohydrate and
fat) is associated with the reverse hierarchy in storage,
with carbohydrates occupying a fairly central position.
(2) A number of workers have attempted to examine the
relative satiating efficiency of different macronutrients.
They reported that protein is more satiating than
carbohydrate, which in turn is more satiating than fat
(alcohol is enigmatic in that it may actually stimulate
appetite in some circumstances). This hierarchy was
reported to occur at a community level, elucidated
through surveys of dietary intake (de Castro and de
Castro, 1989; de Castro, 1991) (a limitation of such
studies is possible misreporting of dietary intake) and
also at the laboratory level (Weststrate, 1992), where
dietary intake is usually measured by the researchers. It
has also been reported to occur at the level of nutrient
balance in subjects ingesting an oral diet (Stubbs et al.,
1995b) as well as in those receiving nutrients intravenously
(Gil et al., 1991). The results of some of the
macronutrient balance studies undertaken in metabolic
centres do not appear to be simply due to differences in
sensory cues. For example, in a study of men who were
studied on three separate occasions so that they could
receive a high carbohydrate (low fat), medium carbohydrate
(medium fat), or low carbohydrate (high fat) diet,
differences in sensory cues (taste, smell) were minimized
by covertly manipulating the diets (Stubbs et al., 1995a).
The proportion of energy from protein was kept constant
across all diets, but the energy density decreased
progressively as proportion of energy from carbohydrate
increased. The subjects ate these covertly manipulated
diets ad libitum while in a whole body calorimeter for 7
days. Average daily balances were 2.58, 0.77 and
0.27 MJ/day on the low-, medium- and high-carbohydrate
diets, respectively. Carbohydrate appeared to be
more satiating than fat. It also appeared that oxidation
of carbohydrate and protein predicted the subsequent
day’s intake better than fat oxidation. The effects on
energy balance persisted through the entire 7-day period
of study, apparently without compensation as the study
progressed. Other studies suggested that protein, which
has even less ‘storage capacity’ than carbohydrate, is also
more satiating than fat, and probably also more satiating
than carbohydrate, especially when protein is ingested
in large quantities (41.2 MJ per meal) (Weststrate, 1992).
Physiological aspects of energy metabolism
M Elia and JH Cummings
From such studies, it appeared that macronutrients such
as carbohydrate and protein, whose balance is most
tightly regulated, exerted greater suppressive effects on
subsequent energy intake than fat, the balance of which
is not so tightly regulated.
(3) If the macronutrient hierarchy in satiation parallels the
macronutrient hierarchy in oxidation (inversely with the
hierarchy in storage), as several studies suggest, the
possibility exists that the two are causally linked. In
examining this hypothesis, it is necessary to consider at
least two issues: (i) plausible mechanisms by which such
effects might be mediated and (ii) whether the macronutrient
hierarchy on satiety are independent of the
energy density of the diet.
Potential mechanisms by which oxidation and storage are linked
to feeding behaviour. A large number of potential mechanisms
may be invoked in carbohydrate-specific models of
eating behaviour. A basic consideration is whether the sensor
is linked predominantly to storage of specific macronutrients
(glycogen in the case of carbohydrate) or oxidation.
Flatt (1987) suggested that glycogen exerts a strong
negative feedback on energy intake, giving rise to the
glycogenostatic model of appetite regulation. The model,
which was developed from observations in rodents, has its
attractions because changes in carbohydrate stores can occur
rapidly, from meal to meal, and are associated with changes
in carbohydrate oxidation. Therefore, the model could
provide a plausible physiological basis for feeding behaviour,
which typically involves ingestion of carbohydrate as the
major macronutrient. However, the model does not distinguish
between liver, which is rapidly lost during starvation,
and muscle glycogen, which appears to be depleted slowly
compared to muscle rapidly depleted during fasting in which
does not appear to be rapidly lost during starvation (rodents
muscle glycogen appears to be depleted much faster during
short-term starvation than in human). In addition, several
human studies have examined the predictions of this model,
and in some circumstances they were unable to confirm
them. For example, changes in feeding behaviour induced by
dietary manipulations that change carbohydrate stores,
without necessarily altering energy balance, or vice versa,
cannot be readily explained by the glycogenostatic theory
(van Stratum et al., 1978; Shetty et al., 1994; Stubbs et al.,
1995a, 1996; Snitker et al., 1997). Some of these studies
found that oxidation of all three macronutrients are
considerably better at predicting subsequent energy intake
than carbohydrate alone (Stubbs et al., 1995b).
The putative signal(s) and sensor(s) associated with the
glycogenostatic model have also not been identified.
Another approach is to examine the effect of blocking
oxidative pathways, either individually or in combination.
Friedman and Stricker (1976) proposed a model that was
based on the balance between oxidation and storage of fuels.
They established an integrative macronutrient model of
S53
European Journal of Clinical Nutrition

S54
feeding behaviour, which was consistent with pharmacological
inhibition of metabolic pathways. However, the model
does not specify the metabolic sensor(s), and therefore a
number of general questions arise. How are fuel oxidation
and/or storage detected? Can the oxidation of one fuel be
distinguished from that of another (a necessary consideration
in macronutrient-specific models) and how are signals
integrated to influence feeding behaviour? Another basic
issue is whether it is oxidation of fuels per se or oxidative
phosphorylation (for example, ATP (or other related substances)
that is primarily responsible for feeding behaviour.
The use of drugs that uncouple oxidative phosphorylation,
causing a major increase in oxidation of nutrients without
an increase in ATP generation, could be used as tools for
addressing this last issue. Dinitrophenol administration
in humans was found to have a linear relationship with
metabolic rate and weight loss, so that at a dose of 0.5 g/day
it increased metabolic rate by about 50% and decreased body
weight by almost 1 kg/week (Tainter et al., 1935; Harper et al.,
2001). The use of dinitrophenol to treat obesity in the United
States in the 1930s (Tainter et al., 1935; Harper et al., 2001)
made some subjects very hungry during treatment. If this
were a general effect, it would suggest that oxidation of fuels
per se, was not responsible for providing feedback on energy
intake, and that oxidative phosphorylation was more likely
to be the cause. However, the information on the effects of
dinitrophenol in humans is far from clear, since some
subjects became less hungry during treatment, so that in
reality there was no consistent overall effect on appetite. It is
difficult to assess the significance of these briefly reported
observations, partly because no measurements of dietary
intake were made, and partly because dinitrophenol can
cause a range of side effects which suppress appetite (for
example, nausea, gastrointestinal symptoms, loss of taste
(especially for salt and sweet), skin rashes and a variety of
other symptoms) (Rabinowitch and Fowler, 1934; Tainter
et al., 1935; Harper et al., 2001). In addition, uncoupling
oxidative phosphorylation would not distinguish between
carbohydrates and fat because both are uncoupled by
dinotrophenol.
The effect of blocking specific oxidative pathways on
eating behaviour has also been examined. For example,
humans experience increased hunger when given 2-deoxyglucose
(50 mg/kg body weight), which blocks glucose
oxidation by competitively inhibiting phosphohexose isomerase
(EC 5.3.1.9), an enzyme that catalyses one of the
early steps of the glycolytic pathway (Thompson and
Campbell, 1977; Welle et al., 1980; Thompson et al., 1982).
When rats were given 2-deoxyglucose, a dose-dependent
increase in food intake was observed (Friedman and Tordoff,
1986). When rats are given methyl palmoxirate, which
blocks fat oxidation by inhibiting the transport of longchain
fatty acids across the inner mitochondrial membrane
(necessary for mitochondrial fat oxidation), there was also
an increase in food intake (Friedman et al., 1990). Separate
sensory systems for detecting carbohydrate and fat oxidation
European Journal of Clinical Nutrition
Physiological aspects of energy metabolism
M Elia and JH Cummings
have been identified through investigations that have
combined surgical lesions of the nervous system and
administration of pharmacological agents that inhibit either
fat or carbohydrate oxidation (Titter and Calingasan, 1994).
Some inhibitors of glucose oxidation act both centrally in
the brain and peripherally (2-deoxyglucose), whereas others
do not penetrate the blood–brain barrier, and therefore
inhibit glucose oxidation only in peripheral tissues (for
example, 2.5-anhydro-D mannitol). Investigations using
such tools suggest that fatty acid oxidation is monitored
peripherally, whereas glucose oxidation is monitored both
peripherally and centrally. The area postrema and nucleus of
the solitary tract receive signals from both systems (the vagus
nerve relays signals from the periphery). However, lesions
that destroy the sensory but not the motor nucleus abolishes
mercaptoacetate (lipoprivic), but not the 2-deoxyglucose
(glucoprivic)-induced feeding. It is not clear to what extent
these distinct pathways, and their interactions, which have
been identified in animals (Titter and Calingasan, 1994), also
operate in humans.
Feeding behaviour and energy density. Several of the studies
that established a satiating hierarchy of macronutrients did
not control the energy density of the diet. Since highcarbohydrate/low-fat
diets tend to be less energy dense than
low-carbohydrate/high-fat diets, it is possible that energy
density rather than a macronutrient hierarchy is responsible
for many of the differential effects on satiety. If this were the
case generally, macronutrient-specific models of feeding
behaviour would decline in importance and give way to
new alternatives, but perhaps complementary models of
feeding behaviour. Few long-term physiological studies have
been undertaken to address this specific issue, but two are
noteworthy. The first is the study of van Stratum et al. (1978),
which was already published at the time when the macronutrient
hierarchy and glycogenostatic models of feeding
behaviour were being actively explored. It reported that
isoenergetically dense high-fat and high-carbohydrate diets
had similar effects on energy intake over 2 weeks in 22
Trappistine nuns. The second study of Stubbs et al. (1996),
involving men who were also studied over a period of 2
weeks, confirmed the results obtained on the Trappistine
nuns. Both of these studies covertly manipulated the diets to
conceal taste differences. A number of short-term studies
(Johnstone et al., 1996; Stubbs et al., 1996), including those
involving single preloads (Stubbs et al., 1996), found that
there may still be some subtle macronutrient effects that are
independent of energy density. Protein was still considered
by some workers to be more satiating than other macronutrients
when taken in doses of 41.2 MJ per meal
(Weststrate, 1992), although even this was challenged by a
recent short-term study with a crossover design (Raben et al.,
2003). It measured the ad libitum intake of a high-protein,
high-carbohydrate, high-fat and high-alcohol meals over a
period of 5 h in the same subjects on separate occasions. The
energy intake from the high-protein meal, which included

B1.4 MJ from protein in men, did not differ significantly
from the other meals, which had similar energy density and
‘dietary fibre’ content (although they differed in their
sensory attributes, such as taste and after taste). There was
also no difference in intake among the 10 women who
consumed B0.7 MJ of protein from the high-protein meal.
This study did not covertly manipulate the meals to conceal
taste. Nevertheless, all the above observations taken together
raise questions about the predictive value of macronutrient-specific
models of feeding behaviour, at least under
laboratory conditions, which may not necessarily apply to
free-living conditions, for reasons that will emerge later.
Inevitably, attention is focused on the weight and energy
density of foods, including ready-to-eat foods, which have
recently flooded the market in developed countries.
Before considering these issues further, it is worth briefly
reflecting on the likely limitations of physiological processes
controlling feeding behaviour that place disproportionate
importance on a single nutrient (for example, carbohydrate).
Such a system would regulate energy balance poorly because
it would be unable to adapt adequately to situations
associated with changes in the proportions of different
macronutrients in the food supply. Therefore, it is not
surprising that specific macronutrient models of feeding
behaviour were not found to be robust in some circumstances,
particularly those involving covert manipulation of
diets with the same energy density but different macronutrient
composition (see above). In addition, pharmacological
inhibition of both fat and glucose oxidation in animals were
found to elicit a massive increase in food intake, compared to
inhibition of either fat or carbohydrate oxidative pathways
(Friedman and Tordoff, 1986). This suggests that the
regulatory system is responsive to oxidation of both fat
and carbohydrate, and that oxidation of each substrate
compensates for the other, when the availability of one of
them is reduced. Therefore, eating behaviour does not
unconditionally depend on the oxidation of one nutrient,
and indeed it would be surprising if it did. Furthermore, the
satiating effect of the Eskimo diet (fat and protein) (Mowat,
1981) would argue against the operation of a simple
carbohydrate oxidation or storage model of feeding behaviour
to the exclusion of other macronutrients.
Energy density, weight of food and carbohydrates
Several reports suggest that under ad libitum feeding conditions
(mainly under laboratory conditions), people tend to
consume a fixed weight or volume of food (Duncan et al.,
1983; Lissner et al., 1987; Tremblay et al., 1991; Tremblay,
1992; Poppitt, 1995; Rolls and Bell, 1999; Bell and Rolls,
2001). This implies that when foods or diets differ in energy
density, energy intake will be affected. Indeed, it has been
suggested that this is a major mechanism for determining
energy intake at different stages of the life cycle (for example,
compared to young adults, older adults reduce energy intake
by largely decreasing the energy density of the food eaten). If
Physiological aspects of energy metabolism
M Elia and JH Cummings
satiety was mainly determined by the weight of food eaten,
then those who consume high-energy density foods would
be expected to feel satiated only after increased energy
intake, whereas those consuming low-energy density foods
would feel satiated despite lower energy intake. This
argument has been promoted by several workers (Poppitt,
1995; Poppitt and Prentice, 1996; Prentice and Poppitt, 1996;
Rolls and Bell, 1999). It is argued that in real-life energy
dense foods tend to be more palatable and less satiating than
foods with low-energy density. Typically, palatable snacks are
energy dense, and those rich in both fat and available
carbohydrates, which are rapidly digested and absorbed, are
said to be particularly palatable (Drewnowski, 1998). Interestingly,
available carbohydrate and fat also provide more
NME per unit ME than does the diet as a whole or either
protein or NSP or other carbohydrates fermented in the large
bowel. Therefore, when using the ME system, high-fat, highavailable
carbohydrate foods (or diets) are more fattening
than isoenergetic high-protein, high-fibre foods. In addition,
foods of low-energy density (often rich in NSP and water)
were less palatable in these studies.
An examination of 1032 foods revealed that the strongest
predictors of energy density are water, which is negatively
related to energy density, and fat (g per 100 g food), which is
positively related to energy density (Stubbs et al., 2000). The
carbohydrate and protein content of foods (g per 100 g food)
are both positively related to energy density, but the
association is much weaker than those obtained with fat or
water (Figure 4). This weak relationship with carbohydrate is
understandable because it is possible for foods containing a
large proportion of energy from carbohydrate to have either
a low-energy density (for example, vegetables) or highenergy
density (some sweet snacks). Diet surveys do not
generally report the energy density of the food eaten, and
this precludes extensive analyses and interpretation of data
Energy density (kJ/100g)
4000
3500
3000
2500
2000
1500
1000
500
0
0 25 50
Weight (g/100g)
75 100
Fat (r 2 =75)
Protein (r 2 =13.9)
Carbohydrate (r 2 =1.9)
Water (r 2 =66.7)
Figure 4 Relationship between energy density (outcome variable)
and % weight of dietary macronutrients and water of 1032 readyto-eat
foods (Stubbs et al., 2000). The prediction equations
are as follows: Energy density (ED) ¼ 462.6 þ (35.5 fat); ED ¼
781.4 þ (12.2 protein); ED ¼ 654.5 þ (12.5 carbohydrate); and
ED ¼ 2034–(21.2 water). The values in parentheses, derived from
regression are r 2 100.
S55
European Journal of Clinical Nutrition

S56
Physiological aspects of energy metabolism
M Elia and JH Cummings
in free-living conditions. However, below is a summary of
some observations, including some in free-living conditions,
which need to be taken into account in models of
energy homeostasis in free-living conditions. The role of
different types of carbohydrates on satiety, energy intake and
energy balance are then considered in the light of these
observations.
(1) The idea that energy intake and energy balance are
largely determined by energy density or weight of food
eaten has enormous implications for the control of
feeding behaviour and energy balance, because it implies
that normally energy intake is a secondary or passive
manifestation of the drive to eat a constant weight of
food. This would not be an ideal evolutionary strategy
for survival. The energy density of the food eaten by our
ancestors is likely to have varied from a low-energy
density (berries, fruits) to high-energy density (fatty fish,
animal meat and fat), depending on location, season and
other environmental changes, including natural disasters.
Too much reliance on weight of food eaten, rather
than on energy eaten, would bring about poor regulation
in energy homeostasis, at least in the long term.
(2) A recent epidemiological study (Ledikwe et al., 2006)
reported that US adults consuming a low-energy-dense
diet are likely to consume more food by weight, and at
the same time less energy intake (assessed by 24 h recall)
than those consuming a higher energy-dense diet.
Normal weight individuals were found to consume diets
with a lower energy density than obese individuals, but
causality is difficult to assess from the available information.
This study is of course subject to potential problems
associated with misreporting, which are briefly discussed
below.
(3) If free-living individuals primarily eat a constant weight
of food, with energy intake as a passive outcome, then
the within-subject day-to-day coefficient of variation for
weight of food ingested would be expected to be lower
than that for energy intake. This was not found to be the
case. In an analysis of 7-day-weighed food intakes in 73
subjects, the coefficients of variation were found to be
slightly higher for weight of food ingested rather than
energy intake (23 vs 21%, respectively, or 21 vs 20.5%,
when drinks were added in the analysis) (Stubbs et al.,
2000; Stubbs and Whybrow, 2004). It would be valuable
to know the extent to which coefficients of variation of
ingested energy relative to ingested weight of food
depended on the environmental conditions. Studies to
address this issue are subject to the problem of
misreporting of dietary intake. One review concluded
that there is no accurate method of measuring dietary
intake (Westerterp and Goris, 2002). Another extensive
review (Livingstone and Black, 2003) expressed dietary
energy intake per day as a fraction of energy expenditure
measured by the doubly labelled water method. The
results for 7-day-weighed food intake were variable
European Journal of Clinical Nutrition
according to the study. For example, in one study in
women aged 60 years, the values were 1.0770.08 (s.d.)
in unrestrained eaters and 0.9470.05 in restrained eaters
(Bathalon et al., 2000), whereas in another study
(Kaskoun et al., 1994) in 15 and 18-year-old males, the
value was 0.7770.023.
(4) Laboratory studies have been undertaken in which diets
with different energy densities are covertly manipulated
to maintain the same palatability (Titter and Calingasan,
1994). In such environments, studies suggest that a large
proportion of the variability in energy intake (440%)
can be explained by energy density of the diet (Prentice
and Poppitt, 1996; Rolls and Bell, 1999; Stubbs et al.,
2000) compared to only a much smaller proportion in
free-living conditions (for example, only 7% according
to one study; Stubbs and Whybrow, 2004). There are
several possible explanations for the discrepancy between
laboratory and free-living studies. (i) It is easier to
covertly manipulate diets of different energy densities or
macronutrient composition at the lower end of the
energy density range. Studies that have varied the energy
density of foods at constant palatability (Drewnowski,
1998) have used diets containing 2.5–4.4 kJ/g and
compared them with diets containing 5.6–7.0 kJ/g. They
have not extended to the higher energy density range.
To give the readers a feel for the energy density of
specific foods that can make up diets, here are some
values: chocolate cake, hamburger and chips, which can
make up a meal, fall in the range of 12–18 kJ/g;
chocolate, decreases in the range of 22–24 kJ/g; raw
vegetables and soft drinks B2 kJ/g for; the full potential
range is from 0 (water) to B39 kJ/g (pure oil or fat). This
means that the energy density (and the variety) of foods/
diets in free-living conditions are more variable than
those predetermined (lower values) in the laboratory.
There is also some evidence that specific macronutrient
effects are more likely to exert effects at the lower end of
the energy density spectrum (Westerterp-Plantenga,
2001). (ii) Learned cues, which influence eating behaviour,
are absent or dramatically reduced in laboratory
studies, because such studies often use foods that are
unfamiliar or covertly manipulated. (iii) Energy density
is fixed in some laboratory conditions, but not usually in
free-living conditions. (iv) Measurement of dietary
intake is likely to be more accurate under laboratory
than in free-living conditions. (iv) There appear to be a
longer-term compensatory mechanisms for the increase
in energy intake induced by ad libitum intake of highenergy-density
foods (Stubbs et al., 2004). A reduction in
the weight of food eaten appears to be an important
compensatory mechanism. (v) Some free-living studies
have reported an association between energy density of
the diet and increased intake. A recent study (de Castro,
2004) involving 952 subjects who provided 7-day food
diaries, reported such a relationship, but found no
relationship between energy density of the consumed

food and body weight, height or BMI. Since high-energy
density appears to be related to greater overall intake in
the short term, the author suggested that there may be
compensation over the long term with no net effect on
body size. An alternative possibility is that increased
activity may determine the energy density of the diet:
large volumes of bulky diets may be less acceptable to
individuals who require a lot of energy.
(5) Analyses of energy intake in free-living conditions using
multiple regression models suggest that there are multiple
determinants of energy intake, of which energy
density is one, and percent energy from individual
macronutrients, including carbohydrate, is another
(Stubbs and Whybrow, 2004). From this and other
works, it appears that macronutrients have effects on
energy intake that are independent of the energy density
of the food consumed. One study reported that specific
macronutrients influence intake, when foods contain
little water (high-energy density) and not when they
contain a lot of water (low-energy density) (Westerterp-
Plantenga, 2001). This could provide an explanation
why laboratory studies, which have covertly manipulated
diets at the lower end of the energy density range,
have a strong predictive effect on energy intake.
(6) Substitution of sugar for artificial sweeteners offers a way
of separating palatability from energy density of foods.
But even here the effects are not clear-cut. For example,
there is little epidemiological evidence to suggest that
intense sweeteners are causally linked to BMI. In
addition, most intervention studies with intense sweeteners
are of short duration (less than 1 or 2 days, and
often only a few hours), with the exception of a few
studies, which suggested that the intense sweetener,
aspartame, reduced energy intake in metabolic units
(Porikos et al., 1977, 1982) or free-living conditions
(Tordoff and Alleva, 1990a). Some short-term studies
using drinks (Blundell and Rogers, 1994) or gum (Tordoff
and Alleva, 1990b) suggested that intense sweeteners
might actually stimulate appetite. One study found that
aspartame-sweetened water transiently increased appetite
in lean men, but aspartame-sweetened soft drinks
suppressed appetite (Black et al., 1993; Black and
Anderson, 1994). Other studies report that sweeteners
have the same effect as water (Rogers et al., 1988; Rodin,
1990; Canty and Chan, 1991), whereas others report
suppression of overall intake when compared with soft
drinks containing sugar (Rolls, 1997). Yet, other studies
suggest that there is an increased intake when an
unpalatable food item is made more palatable by
addition of a sweetener. Methodological problems are
at least partly responsible for the variable results
obtained. In addition, the responses to sweet foods and
drinks depend on dietary restraint (Lavin et al., 1997)
and they can be conditioned. Intense sweeteners may
mimic the ingestive effects of sugars on satiety but not
the post-ingestive effects. For example, animal studies
Physiological aspects of energy metabolism
M Elia and JH Cummings
suggest that initial dislike for a bitter taste of a sucrose
octa acetate solution is reversed once the animals learn
to associate this taste with positive post-ingestive effects
(Sclafani, 1987, 2001). Finally, a 10-week study involving
overweight men and women, who consumed 28% of
their energy as sucrose (mostly as beverages), showed an
increase in body weight (1.6 kg), which was not observed
in a similar group of subjects who consumed artificial
sweeteners (Raben et al., 2002).Overall it appears that
intense sweeteners do not reliably affect appetite and
energy intake, but more longer term studies in free-living
conditions are required.
From the above, it appears that the control of appetite
operates through a redundant system that does not rely
overwhelmingly on one or two major factors, such as energy
density or on a specific macronutrient. Rather it depends on
multiple factors that can interact, compensate or override
each other, depending on environmental exposures and
their duration. Given the fundamental importance of eating
for survival, this is not surprising. The factors that influence
eating behaviours include sensory factors, diet composition
and variety of available food items, eating environment and
individual subject characteristics, such as age, habitual
dietary intake and prior social conditioning. It has been
difficult to establish the relative importance of these factors,
which are likely to differ with the environmental setting, for
example, developed or developing countries. However, it is
clear that feeding behaviour is influenced by both nutritional
and non-nutritional factors. In addition, it appears
that there is a good defence against development of underweight
(in the absence of disease) (Garrow, 1988), but less
good defence against development of overweight (Blundell
and Stubbs, 1999), especially in an environment where the
availability of abundant and varied food is coupled with
limited physical activity. This defence needs to be very tight
because deposition if only 1% of the lifetime cumulative
energy intake were deposited in the reference male or female
(equivalent to the dot by the side of Figure 3), it would
double adult body weight and BMI, and adversely affect
health.
Effect of different types of carbohydrate on feeding behaviour and
energy homeostasis
With this background, the role of different types of
carbohydrates on eating behaviour might not be expected
to have overwhelming effects on long-term energy homeostasis.
The following summary based on intervention
studies is generally consistent with this view.
Non-starch polysaccharides (dietary fibre). The majority of
studies comparing the effects of carbohydrates on energy
balance and weight loss have been undertaken with NSP in
various forms, probably because there are several potential
ways in which NSP might induce a negative energy balance.
S57
European Journal of Clinical Nutrition

S58
The following are potential mechanisms: compared to
available carbohydrates, NSP has a lower energy density
(kJ/g; see above); foods containing NSP are often made from
whole grains or whole foods such as fruit and vegetables and
are thus bulky with a low-energy density, which may
promote satiety; some types of NSP are viscous and can
delay gastric emptying, causing feelings of increased fullness
and satiety; NSP affects the intestinal absorption of other
macronutrients, such as fat, although there is little evidence
for significant losses in this way in humans; and delayed
colonic metabolism of NSP after meal ingestion may delay
the onset of hunger before the next meal (although the
possible role of SCFAs and other potential signals from the
colon appear to have been little investigated in humans).
Results of more than 50 studies have been summarized in
several reviews (Blundell and Burley, 1987; Stevens, 1988;
Burley and Blundell, 1990; Levine and Billington, 1994), but
it is difficult to establish firm overarching conclusions that
apply to all circumstances. There are several reasons for this:
the type and amount of NSP used in different studies have
varied; the intervention has sometimes involved administration
of tablets of extracted NSP and at other times diets
rich in NSP; the age and adiposity of the study population
have varied; and study designs have varied from open trials
to double-blind control trials. Nevertheless, it appears that
dietary supplementation with extracted NSP to a level that
can be tolerated has at best a modest effect in reducing body
weight over a period of several months or longer. Bulky NSPrich
diets with a low-energy density might be expected to
encourage weight loss, or prevent weight gain, by promoting
satiety, but such diets are often less palatable and less likely
to be consumed. Better modelling techniques of the available
data, including dose–response curves (for example,
studied by meta-regression), may shed further light on
possible effects of NSP on energy intake and body weight.
Starch and sugars. A few short-term preload studies have
examined the effects of different types of hexose-based
sugars on appetite and found little difference between them.
Comparisons between sugars and high glycaemic index (GI)
starches using short-term preloading protocols or 7-day
protocols (Mazlam, 2001) suggest no major differences in
satiety. However, a study comparing a high-starch diet with a
high-sucrose diet found that the ad libitum intake over 14
days was lower and weight loss greater with the high-starch
diet than the high-sucrose diet, which scored better on
palatability (Raben et al., 1997). Another study (Raben et al.,
1994) found that exchange of digestible starch for RS in a
meal reduced satiety as well as post-prandial glycaemia and
insulinaemia. The authors acknowledged that the differences
in satiety may be due to texture and palatability of the meal,
but it can be difficult to separate the sensory properties of
diets from their macronutrient composition. However, it
appears from other studies that sensory-specific satiety is
related more to the sensory characteristics of a food than to
its macronutrient composition (Johnson and Vickers, 1993).
European Journal of Clinical Nutrition
Physiological aspects of energy metabolism
M Elia and JH Cummings
As indicated in section above, energy density (see section
Energy density, weight of food and carbohydrates) is also a
potentially confounding variable. There is a lack of longterm
studies. The CARMEN multicentre trial (Saris et al.,
2000) compared the effects of ad libitum intake of low-fat
diets rich in either complex carbohydrates or simple
carbohydrates. Body weight loss in the low-fat simple
carbohydrate and low-fat complex carbohydrate groups did
not differ significantly (0.9 and 1.8 kg, respectively). There
were also no significant differences in circulating lipids.
Since the distribution of complex carbohydrate between
starch and NSP was not reported, it is not possible to assess
the effects of starch vs sugars. Therefore, it appears that there
is lack of convincing evidence from long-term studies that
starch offers advantages over sugars in weight reduction.
Polyols. There appears to be little information on the effect
of polyols relative to other carbohydrates on weight loss or
on satiety and other appetite sensations. One study found
that replacement of sucrose with increasing doses of isomalt
(12, 24 and 48 g per day), which is mostly unavailable
carbohydrate, resulted in weight loss of about 2 kg over a
period of 12 weeks (Spengler and Boehme, 1984).
Low vs high glycaemic index (GI carbohydrates, foods and
diets). Numerous reports have examined the role of different
carbohydrates, foods and diets with varying GI on
appetite and satiety (Raben et al., 1994; Anderson et al., 2002;
Kaplan and Greenwood, 2002; Anderson and Woodend,
2003; Ball et al., 2003), and weight control (Pawlak et al.,
2002; Raben, 2002; Ludwig, 2003; Warren et al., 2003). The
overall results do not appear to be consistent, although
many short-term studies (o1 day) suggest that low-GI
carbohydrates and foods promote satiety. One review
concluded that short-term feeding trials generally show an
inverse association between GI and satiety, and mediumterm
clinical trials show less weight loss on high-GI (or high
glycaemic load) diets compared to low-GI (or low glycaemic
load) diets (Pawlak et al., 2002). However, the differences in
weight between control and intervention groups were small.
Another review (Pawlak et al., 2002) (presented as part of a
debate) systematically examined 31 short-term studies (o1
day) and reported that in 15 of these, low-GI foods promoted
satiety or reduced hunger, and in the remaining 16 low-GI
foods either reduced or made no difference to satiety. Low-GI
foods reduced ad libitum food intake in seven studies, but did
not do so in eight other studies. The same systematic review
examined results from 20 longer term studies (up to 6
months). Weight loss was favoured by the low-GI diets in
four studies, by the high-GI diets in two studies and there
was no significant difference between the low and high-GI
diets in the remaining 14 studies. The average weight loss
was only 1.5 kg on the low-GI diets and 1.6 kg on the high-GI
diets. Not surprisingly, the author concluded that there was
no evidence that low-GI foods were superior to high-GI
foods for long-term weight control. The overall differences in

esults may relate not only to the duration of the studies, but
also to the energy density and sensory attributes of foods or
diets, as well as the age of subjects, which has ranged from
the young age of pre-pubertal children (Warren et al., 2003)
to that of elderly people . The sensory attributes of foods are
particularly important in free-living studies, because poorly
palatable diets are likely to be associated with poor
compliance. The obligatory ingestion of a diet or a meal
when used as a preload under short-term laboratory conditions
may well produce different results when the same diet
is promoted for longer term use under free-living conditions.
Ingestion of such a diet is not obligatory in free-living
conditions, and therefore a diet may show efficacy in
laboratory conditions but not effectiveness in free-living
conditions. In addition, a recent study (Alfenas and Mattes,
2005) examined the effect of consuming only low- or high-
GI foods ad libitum in the laboratory for 8 days in either high
(three foods per meal) or low (one food per meal) variety
conditions. It appeared that differential appetite sensations
and glycaemic responses to foods tested in isolation were not
preserved under conditions of longer term ad libitum
ingestion of mixed meals. The potential interactions
between different foods are much greater in free-living
conditions than in laboratory conditions. In summary, the
evidence suggesting that low-GI foods favour weight loss is
stronger in short-term than in long-tem studies. It would be
valuable to examine dose–response curves, especially in the
long term, partly because there are fewer long-term studies,
and partly because they are more relevant to health.
Liquid vs solid carbohydrates. Several papers and reviews
(Mattes, 1996, 2006; DiMeglio and Mattes, 2000) suggest
that supplemental energy provided as carbohydrates in fluids
is less precisely compensated than when solid foods are
manipulated. The reasons for the difference in energy
compensation are not clear but may involve the rate at
which these are ingested (liquids can be ingested several
times more rapidly than solid food items), their energy
density and sensory attributes, and gastric emptying rates,
which are generally faster for liquids than solids. Reactive
hypoglycaemia may also be implicated. One short-term
study lasting a few hours examined the effect of ingesting
the same quantity of available carbohydrate at the same rate
in the form of an apple, puree apple or juice without the cell
wall material or ‘fibre’. The juice was less satiating than the
puree, which in turn was less satiating than the apple. These
changes were associated with greater insulin responses,
which peaked at 30 min (juice4puree4apple) and a greater
subsequent rebound hypoglycaemia (also juice4puree4
apple) from similar peak glucose concentrations (measured
in venous rather than arterial blood, which could confound
interpretation) (Haber et al., 1977). There are few long-term
studies, but in 1958 Fryer (Fryer, 1958) reported the effects of
supplementing the diet of 20 college students for 2 months
with a carbohydrate-rich drink (1.8 MJ/day). Dietary compensation
for the extra liquid energy intake was only about
Physiological aspects of energy metabolism
M Elia and JH Cummings
half complete by 2 months. Other studies (observational or
epidemiological) have shown that the compensation that
does occur displaces the intake of protein and micronutrients
from the diet, the full significance of which is still
somewhat uncertain. The above intervention trials are
complemented by epidemiological studies, such as the
Nurses’ Health Study II in the United States (Schulze et al.,
2004), which reported that higher consumption of sugarsweetened
beverages is associated with greater weight gain
(4–5 kg over 4 years in those who increased consumption
from one or fewer drinks per week to one or more drinks per
day) and greater risk of type II diabetes. Conversely, weight
gain was smaller among women who decreased their intake
of sugar-sweetened drinks (0–1.5 kg over 4-year periods).
Epidemiological studies in children (for example, US Growing
Up Today Study; Berkey et al., 2004) also show a link
between sugar-added beverages and weight gain. The link
between soft drinks and childhood obesity has been
reviewed recently (James and Kerr, 2005). Finally, a longterm
intervention trial in England (The Christchurch obesity
prevention project in schools (CHOPPS); James et al., 2004)
reported that the results of a cluster randomized control trial
involving 644 children aged 7–11 years. A focused educational
program at school over a year reported the average 3day
consumption of carbonated drinks decreased by 0.6
glasses (average glass size 250 ml) in the intervention group
and increased by 0.2 glasses in the control group, with the
result that in the percentage of overweight and obese
children at the end of the year increased by 7.5% in the
control group and decreased by 0.2% in the intervention
group. In summary, there is epidemiological, physiological
evidence, as well as interventional evidence linking sugarsweetened
beverages and weight gain.
The need to expand the physiological evidence base
The early growth of the obesity epidemic in developed
countries some 30 years ago was associated with ingestion of
high-fat, low-carbohydrate diets. Many workers felt that the
macronutrient composition of the diets had a large part to
play in this, and physiological studies that varied the
carbohydrate to fat ratios (without controlling for
energy density or GI) in laboratory studies were consistent
with this notion. However, the recent trend in many
developed countries, such as the United States (Centres
for Disease Control and Prevention (CDC), 2004) and
the United Kingdom (Henderson et al., 2003), for a
decrease in the proportion of dietary energy from fat
and an increase from carbohydrate, has not prevented
the growth in obesity and overweight, which has continued
to rise unabated. It is possible that without this change
an even greater rise would have occurred. These trends
have also been associated with exposure to an increased
variety of foods and drinks containing different types of
carbohydrates, carbohydrate combinations, sweeteners
and sweet-bulking agents. Advances in technology can
S59
European Journal of Clinical Nutrition

S60
modify the structures of carbohydrates, for example, starch
to RS, and lead to the development of new types of
carbohydrates or mimetics (substances that mimic
carbohydrates). However, their functional effects and their
interactions with other dietary components need systematic
examination so that their overall effect on energy balance
can be understood.
Consumer behaviour and demands have also grown, and
they have strained the existing physiological evidence base.
For example, the above discussion indicates that macronutrients
(or particular types of macronutrients, such as sugars,
starch or NSP) can themselves influence energy density, GI
and palatability of foods, all of which have been implicated
in feeding behaviour. However, some of the effects of these
food attributes on feeding behaviour overlap with each other
to an uncertain degree (Figure 5). They probably also interact
with other factors, such as texture and viscosity, which can
also be influenced by macronutrients (for example, viscosity
can be increased by certain types of NSP, such as guar gum).
Most studies on feeding behaviour have studied only one of
these attributes, some have considered a couple, but there is
a lack of studies that have considered the effects of three or
more such attributes simultaneously. Therefore, a number of
key questions arise, which emphasize the need for integrative
research. To what extent is energy density linked with GI
and palatability? To what extent is palatability linked to GI
and energy density? How do different types of carbohydrates
influence these attributes, and can artificial modification
bring about predictable changes? There is also the issue
about the extent to which feeding behaviour is causally
Energy
density
Macronutrients
and water
Palatability
Glycaemic
index/load
Feeding behaviour
Physiological aspects of energy metabolism
M Elia and JH Cummings
Other factors
Other factors
(nutritional + non-nutritional)
Figure 5 A model showing a possible pathway between intake of
different macronutrients (and different subtypes, such as sugar,
starch, polyols and fibre, in the case of carbohydrate) and feeding
behaviour. Study of only one of the properties of foods (glycaemic
index, energy density, palatability and others) is less informative
than their simultaneous study, especially when linked to food and
macronutrient composition.
European Journal of Clinical Nutrition
linked to these food attributes, as opposed to being markers
of some other functional and causal property or properties. It
has been suggested that weight-reducing diets should avoid
high glycaemic foods, on the basis that they might cause
reactive hypoglycaemia, which in turn, it is argued, stimulates
appetite, reduces satiety and makes weight loss more
difficult (Agatston, 2003). Mention has already been made
(see section Liquid vs solid carbohydrates’) about the
hypothetical role of reactive hypoglycaemia in mediating
differences in satiety after ingestion of apple, puree and juice
(Haber et al., 1977). However, the evidence base for this
causal pathway is weak and further work at each one of these
steps is required.
Another issue concerns the energy system used in detailed
metabolic studies, for example, studies that accurately
assessed the effects of isoenergetic diets on ad libitum energy
intake. Almost invariably the ME system has been used for
this purpose (at least for the major macronutrients—fat,
protein and carbohydrate absorbed by the small bowel).
However, since this system can be considered to be flawed in
relation to energy balance, a re-evaluation of the situation
using the NME system is necessary, especially for those
studies that have used large amounts of protein, fibre or
other macronutrients that have a low bioenergetic efficiency.
Diets that are isoenergetic in relation to the ME system are
not necessarily isoenergetic in relation to the NME system
(see sections Net metabolizable energy, and Energy systems
and food labelling).
There is also a need to bridge the gap between short-term
physiological studies that are undertaken in confined
environments (laboratory studies) and longer term studies
in free-living environments. Epidemiological approaches are
useful in establishing associations, but they may have greater
difficulty in separating causes from effects. For example,
cross-sectional epidemiological studies on the role of sweeteners
on weight control may be difficult to interpret because
sweeteners may be preferentially used by those already
overweight or obese, and this may mask any associations
between sweeteners and weight loss. In addition, it should be
remembered that sweeteners are used both as replacements
for sugar and as supplements to sugar (for example, in several
sugary soft drinks) in an attempt to promote a more
prolonged and desirable after taste.
This discussion has placed much emphasis on physiological
determinants of eating behaviour. However, there have
been long-standing debates about the relative importance of
physiology and psychology on feeding behaviour, and the
boundary between them. Both are important, but their
interactions and their independent contributions to feeding
behaviour almost certainly vary with the environment.
Social conditioning, social and environmental ambiance,
and learned behaviour all play a role. Even the extent to
which compensation occurs following supplementation may
depend on whether the food is familiar to the consumer.
Future research using a multidisciplinary approach is to be
encouraged. Such an approach should also involve the

ecologist, who is pre-occupied with why animals evolved
with the behavioural programmes that control food intake.
In contrast, the physiologist is pre-occupied with more
immediate determinants of feeding behaviour. Both can
provide important insights into mechanisms of eating
behaviour and possible ways for combating over- and
undernutrition.
Requirements for dietary carbohydrate
Using the brain’s requirement of glucose as a basis for
considering carbohydrate requirements, the Institute of
Medicine has estimated the average requirement to be
100 g/day, and the Recommended Dietary Allowance to be
130 g/day day, for both men and women aged 19 years and
above (Institute of Medicine, 2005). Estimated average
requirement typically represents only 10–20% of total energy
requirements, which vary according to weight and physical
activity (FAO/WHO/UNU Expert Consultation, 2004). However,
high-fat, high-protein diets are not considered healthy
and many national and international bodies have set limits
and ranges for carbohydrate intake based mainly on supplying
energy needs for the individual that would not be
supplied by recommended amounts of fat and protein. In its
report ‘Diet and Prevention of Chronic Diseases’, WHO/FAO
recommended that carbohydrate provides 55–75% of total
energy (WHO, 2003). The upper limit is thought to allow
adequate protein and fat intake, whereas the lower limit
allows a maximum energy intake from fat of 30% and
protein of 15%.
The partial replacement of carbohydrate with monounsaturated
fats in diabetic diets has been seen to be
beneficial (Garg, 1998). This has motivated European and
UK diabetes associations to implement this advice for people
with diabetes (Diabetes and Nutrition Study Group of
European Association for the Study of Diabetes, 2000;
Connor et al., 2003). They recommend that 40–60% of total
energy should come from carbohydrate, and that carbohydrate
plus monounsaturated fat should provide 60–70% of
energy. This allows 20% of energy to come from monounsaturates.
WHO/FAO, and other bodies, also suggest that there
should be a limit on the intake of free sugars at o10% of
energy. This arises from the relation of sugar to dental caries,
an increasing problem in the developing world, and that
FAO/WHO ‘considered that studies showing no effect of free
sugars on excess weight have significant limitations’ (Nishida
et al., 2004). The 10% figure is given as a maximum, and
one would expect therefore that sugar intakes on average
would be less than this for populations as a whole.
Carbohydrate and physical performance
Carbohydrate feeding both before and during exercise can
improve performance through a variety of mechanisms
(Jeukendrup, 2004).
Physiological aspects of energy metabolism
M Elia and JH Cummings
Carbohydrates in the large bowel
Quantitative aspects
The principal carbohydrates that reach the human large
bowel are the NSP, RS, non-a-glucan oligosaccharides (shortchain
carbohydrates), and some polyols and modified
starches. In many populations of the world, there is primary
low lactase activity in the small bowel.
The amount of these carbohydrates that arrive in the
caecum is difficult to calculate because dietary intakes of
most, except NSP, are largely unknown. NSP in the diet in
European countries is between 11 and 33 g/day (Cummings
and Frolich, 1993; Bingham et al., 2003) and all of this will
reach the colon. Intakes of NSP in Mexicans are reported
as 22 and 17 g/day in rural men and women, and 18 and
16 g/day in urban men and women (Sanchez-Catillo et al.,
1997). In a West African village, intakes (corrected to 100%
energy intake) were in the range of 25–28 g/day for men and
21–24 g/day for women, and 10–13 g/day in children of
unspecified age, but sharing meals with adults (Hudson and
Englyst, 1993). Other data, reported by Cassidy et al. (1994)
give NSP intakes for Australia of 12–13 g/day, India 15–21 g/day,
Ireland 9–11 g/day, Japan 11 g/day and the United States
12–17 g/day. Higher intakes may occur in vegetarians and
populations with high intakes of fruit and vegetables such as
some Pacific Islanders.
Resistant starch intakes are much more difficult to
determine. This is because almost any handling of starchy
foods from diet collections, that is mixing with water,
heating, homogenization, freezing or cooling, will affect RS
content. Present estimates for countries with westernized
diets are in the range of 3–10 g/day (Champ et al., 2003;
Goldring, 2004). Clearly this is very diet dependent. A couple
of relatively unripe bananas will readily provide 20 g RS. A
single biscuit made with potato flour would give 10 g.
Estimates for developing countries for RS intakes are 9–10
to 30–40 g/day where starch intakes are high (Stephen et al.,
1995). The amount of RS that escapes digestion also varies
between people, partly dependent on transit time through
the small bowel (Stephen et al., 1983; Silvester et al., 1995).
Modified starches are also likely to reach the colon because
of their ether and ester bonds or increased cross-linking and
substitutions. These starches are used in small amounts by
the food industry for their functional properties and intakes
are unlikely to be more than 2–3 g/day.
Almost no data have been reported of intakes of the non-aglucan
oligosaccharides or short-chain carbohydrates. Fructooligosaccharide
and inulin intakes were estimated at
between 1 and 10 g/day in European countries by Van Loo
et al. (1995), but one must add to this the other oligosaccharides
in beans, peas, non-wheat cereals and milk. Again, a
meal containing 100 g of Jerusalem artichokes would provide
16–20 g of inulin, and babies living on breast milk will be
getting 15 g/l of oligosaccharides (Miller and McVeagh,
1999). Average oligosaccharide intakes for westernized diets
could, therefore, be in excess of 10 g/day.
S61
European Journal of Clinical Nutrition

S62
Polyols, such as lactitol, isomalt, maltitol, sorbitol,
manitol, erythritol and xylitol, are mostly used as lowcalorie
substitutes for sugar in foodstuffs and confectionary.
Daily intake of these carbohydrates is unknown but is likely
to average only 2–3 g. However, people selecting sugar-free
products including diabetics and those trying to lose weight
may easily consume 20 g/day, so variation in intakes may
well be great. Intakes are limited because of gastrointestinal
intolerance (Lee et al., 2001). Even when intakes are known,
the amount of these carbohydrates reaching the colon is
difficult to calculate because their absorption varies between
2 and 90%, whereas some, such as erythritol, are almost
completely excreted by kidney (Livesey, 2003b).
Overall, therefore, carbohydrate reaching the colon in people
on westernized diets is probably in the range of 20–40 g/day,
whereas in countries with high cereal intakes or higher intakes
of fruit and vegetables, this could reach 50 g/day.
Virtually, all carbohydrate that enters the large bowel will
be fermented by the commensal bacteria that live in the
colon at densities of up to 10 12 /g. The extensive fermentation
of NSP has been known for many years (Cummings,
1981), while recovery of oligosaccharides from faeces is
effectively nil (Cummings et al., 2001). Similar data are also
available for RS with only very resistant retrograded starches
partly surviving, the amount depending on colonic transit
time, although occasional individuals are unable to digest
some RS fractions (Cummings et al., 1996). Microcrystalline
cellulose may resist fermentation because of its highly
condensed structure, an observation that led to the belief
that cellulose was not digested in the human gut. This is
because microcrystalline cellulose was used in many early
experiments of cellulose digestion. However, cellulose
naturally present in the cell wall of food is completely
fermented unless in association with large amounts of lignin.
Other polysaccharides of the plant cell wall are also readily
fermented, even when given in purified forms such as pectin
or guar gum. This process is facilitated by the ability of these
latter substances to form gels readily accessible to the
microbiota.
Fermentation
Fermentation is an anaerobic process and, therefore, produces
unique end products. These include principally the
SCFAs acetate, propionate and butyrate (Cummings, 1995;
Cummings et al., 1995). They are rapidly absorbed. Butyrate
is the major energy source for the colonic epithelial cell, in
contrast to glutamine for the small bowel and glucose for
most other tissues. Butyrate also has differentiating properties
in the cell, arresting cell division through its ability to
regulate gene expression (Siavoshian et al., 2000). This
property provides a credible link between the dietary intake
of fermented carbohydrates, such as NSP, and protection
against colorectal cancer (Bingham et al., 2003).
Propionate is absorbed and passes to the liver where it is
taken up and metabolized aerobically. This molecule is not
European Journal of Clinical Nutrition
Physiological aspects of energy metabolism
M Elia and JH Cummings
seen as having significant regulatory properties in humans,
although it may moderate hepatic lipid metabolism (Todesco
et al., 1991; Stephen, 1993). However, in ruminant animals,
propionate is crucial to life because it is used to synthesize
glucose in the liver.
Acetate is the major SCFA produced in all types of
fermentation, the molar ratio of acetate:propionate:butyrate
is around 60:20:20. Acetate is rapidly absorbed, stimulating
sodium absorption and passes to the liver and then into the
blood from where it is available as an energy source. Fasting
blood acetate levels are about 50 mmol/l, rising 8–12 h later to
100–300 mmol/l after meals containing fermentable carbohydrate.
Acetate is rapidly cleared from the blood with a halflife
of only a few minutes and is metabolized principally by
skeletal and cardiac muscle and the brain. Acetate spares free
fatty acid oxidation in humans and its absorption does not
stimulate insulin release. Another precursor of blood acetate
is alcohol.
Fermentation also gives rise to the gases hydrogen and
carbon dioxide. Much of the hydrogen is converted to
methane by bacteria, and both hydrogen and methane are
excreted in breath and flatus. Gas production, especially if
rapid, is one of the principal complaints of people unused to
eating foods containing significant amounts of fermentable
carbohydrate. Another product of fermentation is microbial
biomass or microbial growth (Stephen and Cummings,
1980). These bacteria are excreted in faeces and this is one
of the principal mechanisms of laxation by NSP. Lactate is
also produced from fermentation, usually during rapid
breakdown of soluble carbohydrates such as oligosaccharides.
Both D- and L-lactate are produced and both are
absorbed. Some ethyl alcohol is produced during hind gut
fermentation, although it is more characteristic of fermentation
due to yeasts as in brewing and wine making.
Bowel habit
Carbohydrates in the colon have additional health benefits
beyond providing energy through SCFA absorption. The best
documented effect is that of NSP on bowel habit. Table 5 is a
compilation of faecal weight data from around 120 papers
published between 1932 and 1992 detailing 150 separate
studies (Cummings et al., 2004). It shows the average
increase in stool output expressed as grams of stool (wet
weight) per gram of ‘fibre’ fed using weighted means from
published data. Wheat bran as a source comes out as the
most effective with raw bran at 7.2 g/g more effective than
cooked bran, 4.4 (Po0.05), but has the disadvantage of
containing 3% phytate—a known inhibitor of the absorption
of divalent cations (calcium, magnesium, zinc and iron).
Fruit and vegetables are remarkably effective, 6.0 g/g, and
with wheat bran are well ahead of the rest. After fruit and
vegetables comes psyllium at 4.0 g/g and then the league
table progresses through oats 3.4, other gums and mucilages
3.1, corn 2.9, legumes (mainly soya) 1.5 and lastly pectin 1.3.
The overall differences among the various sources of NSP are

Table 5 Effect of NSP (dietary fibre) on bowel habit a
Source N b
Increase in
stool weight
(mean g/g
‘fibre’ fed)
Median Range
Raw bran 82 7.2 6.5 3–14.4
Fruit and vegetables 175 6.0 3.7 1.4–19.6
Cooked bran 338 4.4 4.9 2–12.3
Psyllium/ispaghula 119 4.0 4.3 0.9–6.6
Oats 53 3.4 4.8 1–5.5
Other gums and
mucilages
66 3.1 1.9 0.3–10.2
Corn 32 2.9 2.9 2.8–3.0
Soya and other
legumes
98 1.5 1.5 0.3–3.1
Pectin 95 1.3 1.0 0–3.6
Abbreviations: ANOVAR, analysis of variance; NSP, non-starch polysaccharides.
Difference among sources significant: ANOVAR: F ¼ 4.78; Po0.001.
a
Modified and recalculated from Cummings et al. (2001, 2004).
b
N, number of subjects involved in studies.
statistically significant. These laxative properties of NSP are
used in the prevention and treatment of constipation
(Cummings, 1994).
Are any of the changes brought about by plant cell wall
NSP unique? For most of them the answer is ‘no’, although
the physical properties of NSP in the gut, especially gel
function in the small bowel and surface effects in the large
intestine, come closest. What has put the physiological and
health effects of NSP into perspective has been the arrival on
the nutritional scene in the last 20 years of RS and prebiotic
carbohydrates (oligosaccharides).
Resistant starch shows some of the attributes of NSP in that
it provides substrate for fermentation with the production of
SCFA, but differs radically from NSP in not having physical
properties that come from the plant cell wall, and RS is more
the product of food processing rather than being an
indicator of a healthy diet. RS, through its fermentability
has mild laxative properties. Seven studies have reported
accurate RS measurements in diet and have carried out
adequate faecal collections (Table 6), from which it can be
seen that the average increase in stool weight is 1.5 g/g RS
fed. A meta-analysis including data from six of these seven
studies (Tomlin and Read, 1990; van Munster et al., 1994;
Phillips et al., 1995; Cummings et al., 1996; Silvester et al.,
1997; Heijnen et al., 1998; Hylla et al., 1998) (excluding
Tomlin and Read, 1990; because no error measurements are
given in the paper) has been carried out and the results are
given in Figure 6. This shows a highly significant overall
increase in mean daily stool weight for the group of
41.1 g/day (75.4 g s.e.m.; Po0.001), but no significant
dose-response was found using meta-regression. It was not
possible to distinguish from these data among the different
types of RS. This puts RS very much towards the bottom of
the league table of carbohydrates that affect bowel habit
(Table 5) and a minor contributor compared with NSP from
Physiological aspects of energy metabolism
M Elia and JH Cummings
Table 6 Effect of RS on bowel habit
Source N Amount
fed (g/day)
Increase in
stool weight
(mean g/g
RS fed)
Potato RS2 9 26.8 1.6
Banana RS2 8 30.0 1.7
Wheat RS3 9 17.4 2.5
Maize RS3 8 19.0 2.7
Hylon VII RS2 23 32 1.4
Hylon VII RS3 23 32 2.2
Hylon VII 12 55 0.8
Mixed potato RS2
8 40 0.9
and Hylon VII
Mixed food sources 11 39 1.8
Cornflakes 8 10 0
Hylon VII RS2 14 28 1.0
Abbreviation: RS, resistant starch.
Overall weighted mean (s.e.m.) 1.5 g/g (0.24).
N ¼ 147 diet periods in 11 studies.
For data sources see Figure 6.
whole grain cereals, fruit, vegetables, psyllium, oats and
corn. In vitro starch and RS are good sources of butyrate
(Cummings, 1995), but these findings do not lead through
to consistent changes in faecal butyrate output or molar
ratios of SCFA (Flourie et al., 1986; van Munster et al., 1994;
Phillips et al., 1995; Cummings et al., 1996; Heijnen et al.,
1998; Hylla et al., 1998).
Clinical aspects
Irritable bowel syndrome (IBS) is one of the most common
disorders seen in the gastroenterology clinic. It has two main
presenting features, namely, abdominal pain and altered
bowel habit. It is, however, a very diverse disease with no
clear aetiology. While the cause is unknown, NSP has a useful
role to play in the management of constipation-predominant
IBS. However, wheat bran is not universally beneficial
in this condition possibly because it is thought that a
significant number of IBS patients are wheat intolerant
without having the diagnostic features of coeliac disease
(Nanda et al., 1989; Francis and Whorwell, 1994; Snook and
Shepherd, 1994). Furthermore, changing people onto significantly
increased NSP intakes leads to excess gas production
and IBS patients may have a gut that is unusually
sensitive to gas (King et al., 1998; Drossman et al., 2002;
Talley and Spiller, 2002; Longstreth et al., 2006).
Colonic diverticular disease is another condition that
benefits from carbohydrate in the diet, particularly NSP. A
diverticulum is a pouch that protrudes outwards from the
wall of the bowel and is associated with hypertrophy of the
muscle layers of the large intestine, particularly the sigmoid
colon. Diverticular disease is very common in industrial
societies, the prevalence rising with age to about 30 percent
of people over the age of 65 years. Many people with
diverticula do not have symptoms, but those that do,
S63
European Journal of Clinical Nutrition

S64
complain of lower abdominal pain and changes in bowel
habit. High NSP-containing diets were introduced in the
1960s and their use revolutionized the management of this
condition (Painter et al., 1972). Wheat bran is thought to be
more effective than other sources of NSP or bulk laxatives,
although bran is not a panacea and may aggregate gas
production, feelings of abdominal distension and incomplete
emptying of the rectum (Stollman and Raskin, 2004).
The role of carbohydrate in the prevention of colorectal
cancer is dealt with elsewhere in this consultation.
Prebiotics
Physiological aspects of energy metabolism
M Elia and JH Cummings
Study name RS (dose) Difference in means and 95% CI
RS Control
Cummings et al., 1996 Banana RS 2 (30.0g) 8 12
Cummings et al., 1996 Maize RS 2 (19.0g) 8 12
Cummings et al., 1996 Potato RS 2 (26.8g) 9 12
Cummings et al., 1996 Wheat RS 3 (17.4g) 9 12
Heijnen et al., 1998 Hylon VII RS 2 (32.0g) 23 22
Heijnen et al., 1998 Hylon VII RS 3 (32.0g) 23 22
Hylla et al., 1998 Hylon VII (55g) 12 12
Silvester et al., 1997 Potato starch + Hylon VII (40g) 8 8
Phillips et al., 1995 Mixed diet (38.6g) 11 11
Van Munster et al., 1994 Hylon VII RS 2 (28g) 13 13
124 136
Definition
‘A prebiotic is a non-digestible food ingredient that beneficially
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). This use of the term ‘non-digestible’ in
this context is meant to convey the idea of carbohydrates
that are not digested and absorbed in the small intestine. As
such it conflicts with the use of the term digestion referred to
in the section on DE where ‘digestibility is defined as the
proportion of combustible energy that is absorbed over the
entire length of the gastrointestinal tract.’ (see below).
Physiology and microbiology
The main candidate prebiotics are all non-a-glucan oligosaccharides.
They are present in the normal diet, as described
above, at intakes of 5–10 g/day. Those that have been shown
-150.00 -75.00 0.00 75.00 150.00
Favours control Favours RS
Figure 6 Meta-analysis (fixed effect model) of effect of various sources of resistant starch on bowel habit (stool weight). Point estimate
þ 41.1 g/day; 95% CI: þ 30.5 to þ 51.7 g/day. The test of overall effect was highly significant (P ¼ o0.001), and the test of heterogeneity was
not significant (I 2 ¼ 0%; P ¼ 0.971).
European Journal of Clinical Nutrition
to be prebiotic are fructooligosaccharides, galactooligosaccharides
and lactulose. The ability of fructooligosaccharides,
galactooligosaccharides and inulin, when taken in
relatively small amounts in the diet of around 5–15 g/day,
to alter the composition of the flora to one dominated by
bifidobacteria and lactobacilli is now well established
(Gibson et al., 2004; Roberfroid, 2005).
Prebiotic carbohydrates are important because of the new
concept of a healthy or balanced gut flora. A healthy, or
‘balanced’ microbiota is one that is predominantly saccharolytic
and comprises significant numbers of bifidobacteria
and lactobacilli (Cummings et al., 2004). This concept is
based on a number of observations. The genera Bifidobacterium
and Lactobacillus do not contain any known pathogens,
and they are primarily carbohydrate-fermenting bacteria,
unlike other groups such as bacteroides and clostridia that
are also proteolytic and amino-acid fermenting. The products
of carbohydrate fermentation, principally SCFAs are
beneficial to host health, while those of protein breakdown
and amino acid fermentation, which include ammonia,
phenols, indoles, thiols, amines and sulphides, are not
(Cummings and Macfarlane, 1991). Furthermore, lactic
acid-producing bacteria such as bifidobacteria and lactobacilli
play a significant role in the maintenance of colonization
resistance, through a variety of mechanisms (Gibson
et al., 2005). Equally importantly, the exclusively breast-fed
neonate has a microflora dominated by bifidobacteria,
which is part of the baby’s defence against pathogenic
microorganisms, and which is an important primer for their
immune system. This microflora is nurtured by oligosaccharides
in breast milk, which can be considered to be the
original prebiotics.

As already described, almost any carbohydrate that reaches
the large bowel will provide a substrate for the commensal
microbiota, and will affect its growth and metabolic
activities. This has been shown for NSP (Stephen and
Cummings, 1980), and will occur with other substrates such
as RS, sugar alcohols and lactose. However, stimulation of
growth by these carbohydrates is a nonspecific, generalized
effect, that probably involves many of the major saccharolytic
groups, and associated cross-feeding species in the
large bowel (Macfarlane and Cummings, 1991). The selective
properties of prebiotics relate to the growth of bifidobacteria
and lactobacilli at the expense of other groups of bacteria in
the gut, such as bacteroides, clostridia, eubacteria, enterobacteria,
enterococci, and so on. In practice, studies show
that such selectivity is variable, and the extent to which
changes in the microbiota allow a substance to be called
prebiotic have not been established. There are also qualitative
aspects of the concept of selectivity. Some investigations
have shown increases in other bacterial genera such as
Roseburia, Ruminococcus and Eubacterium, with established
prebiotics like inulin (Duncan et al., 2003; Langlands et al.,
2004). Moreover, it is now recognized that many bacteria
inhabiting the large bowel have not yet been identified and
are difficult to culture routinely (Macfarlane and Macfarlane,
2004; Eckburg et al., 2005). One consequence of this is that
we do not know what the global effects of prebiotics are on
the structure of the microbiota. Furthermore, prebiotics can
only enhance the growth of bacteria that are already present
in the gut and the composition of the microbiota can be
affected by a variety of other factors, such as diet, disease,
drugs, antibiotics, age, and so on. Nevertheless, prebiotic
carbohydrates do have dramatic effects on the gut flora. Does
this lead to any health benefits?
Table 7 Effects of prebiotics on bowel habit
Physiological aspects of energy metabolism
M Elia and JH Cummings
Prebiotics should be laxative but as Table 7 and
Figure 7 show (Ito et al., 1990; Gibson et al., 1995;
Alles et al., 1996; Bouhnik et al., 1997; Castiglia-Delavaud
et al., 1998; van Dokkum et al., 1999; Gostner et al., 2005),
they are not significantly so. Meta-analysis of the studies in
Table 7 show that overall there is no significant change in
faecal weight when a wide range of sources and doses of
oligosaccharides are studied. Most of these papers do,
however, report a clear bifidogenic effect, so this alone does
not affect bowel habit. Subjects often report increased
flatulence and bloating, a sign that fermentation is occurring,
but there is no change in SCFA profile or bile acid
output. Studies in constipated subjects (Teuri and Korpela,
1998; Chen et al., 2000; Den Hond et al., 2000), report
increases in stool weight, which is surprising and should be
followed up.
Health benefits
The main potential health benefit of prebiotics should be in
strengthening gut barrier function against infection. So far,
however, few randomised controlled trial (RCT) have been
done and those that have, in traveller’s diarrhoea, IBS,
inflammatory bowel disease and pouchitis, really do not
show a clear benefit except, possibly, in well-being, although
animal studies in inflammatory bowel disease are numerous
and show reductions in inflammation. In antibioticassociated
diarrhoea, three RCT have been reported, in one
of which there was a benefit in reducing episodes of
diarrhoea in patients with Clostridium difficile-associated
symptoms treated with metronidazole and vancomycin.
Again, it is too early to draw conclusions from these studies.
(see review by Macfarlane et al., 2006).
Type Amount (g/day) N MDSW/g/day Increase in stool weight
(mean g/g ‘fibre’ fed)
Control Prebiotic
Oligomate 55 (GOS) 4.8 12 151 134 0
9.6 12 151 0
19.2 12 162 0.6
Oligofructose 15.0 8 134 154* 1.3
Inulin 15 4 92 123 2.1
Oligofructose 5 24 272 279 0
15 264 0
TOS 10 8 105 80 0
Inulin 31 9 129 204* 2.4
Inulin 15 12 129 155 1.7
Oligofructose 15 12 108 0
GOS 15 12 158 1.9
Isomalt a
30 19 99 111 0.4
Abbreviations: GOS, galacto-oligosaccharides; gram/gram increase, gram increase in stool weight per day per gram prebiotic fed; MDSW, mean daily stool weight;
N, number of subjects; TOS, transgalacto-oligosaccharide.
*Significantly different from control Po0.05.
a
Proposed as a prebiotic but not established as one.
For data sources see Figure 7.
S65
European Journal of Clinical Nutrition

S66
Study name Oloigosaccharide (OS) (g) Difference in means and 95% CI
OS Control
Ito et al 1990 GOS (10g) 12 12
Ito et al 1990 GOS (2.5g) 12 12
Ito et al 1990 GOS (5g) 12 12
Gibson et al 1990 FOS (15g) 8 8
Gibson et al 1990 Inulin (15g) 4 4
Alles et al 1996 FOS (5g) 24 24
Alles et al 1996 FOS (15g) 24 24
Bouhnik et al al 1997 TOS (10g) 8 8
Castiglia-Delavaud et al 1998 Inulin (22g) 9 9
van Dokkum et al 1999 FOS (15g) 12 12
van Dokkum et al 1999 GOS (15g) 12 12
van Dokkum et al 1999 Inulin (15g) 12 12
Gostner et al 2005 Isomalt (30g) 19 19
168 168
-200.00 -100.00 0.00 100.00 200.00
Favours control Favours OS
Figure 7 Meta-analysis (fixed effect model) of various prebiotic carbohydrates on bowel habit (stool weight). Point estimate þ 12.6 g/day; 96%
CI: 1.5 to þ 26.6 g/day. The test of overall effect was not significant (P ¼ 0.079) and nor was the test of heterogeneity (I 2 ¼ 21.6%; P ¼ 0.225).
Table 8 Effect of prebiotics on calcium absorption in humans
Subjects N Prebiotic Study design Absorption
method
Adolescents, M 14–16 years 12 FOS 15 g RCT feeding study. 9-day
periods
Adolescents, F 11–14 years 59 FOS 8 g
FOS þ inulin 8 g
Adolescents, F/M 9–13 years 100 Mixed long- and
short-chain inulin 8 g
M 20–30 years 12 Inulin 15 g, FOS 15 g,
GOS 15 g
An unexpected but potentially very important effect of
prebiotics is that on calcium absorption and bone mineral
density. Studies in animal models show clearly enhanced
absorption of calcium, magnesium and iron with galacto-
Randomized crossover
feeding study. 3-week
periods
1-year supplement to
diet
Randomized crossover
feeding study. 21-day
periods
M 9 Inulin Latin square feeding
study. 28-day periods
Post menopausal women
12 FOS 10 g RCT feeding study. 5-
50–70 years
week periods
Post menopausal women
12 TOS 20 g RCT crossover. 9-day
55–65 years
periods
8 men 7 women, 25–36 years 15 FOS 0.8–1.1 g Absorption from fortified
milk drinks
Abbreviations: FOS, fructo-oligosaccharides; TOS, transgalacto-oligosaccharides.
For data sources see Macfarlane et al. (2006).
European Journal of Clinical Nutrition
Physiological aspects of energy metabolism
M Elia and JH Cummings
Result
44 Ca, 48 Ca Fractional absorption
increased from 48717%
to 60717%
46 Ca, 42 Ca FOS—effect
FOS/inulin absorption
increased from 32710%
to 38710%
46 Ca Calcium absorption
greater. Bone mineral
density higher
44 Ca, 48 Ca No effect on calcium or
iron absorption
Balance Significant increase in
absorption. No effect on
magnesium, iron or zinc
44
Ca and balance No effect
44 48
Ca, Ca Ca absorption increased
from 2177% to 2477%
42 43 44
Ca, Ca, Ca No effect
oligosaccharides, fructooligosaccharides and inulin, and
prevention of osteopenia following gastrectomy and ovariectomy.
Table 8 summarizes the studies in humans
in which five out of eight show a benefit, most importantly,

in adolescents. Possible mechanisms are discussed elsewhere
(Macfarlane et al., 2006). How specific this effect on calcium
absorption is to this class of fermented carbohydrate remains
to be seen. Lactose has traditionally been thought to
promote calcium absorption, although this is not a consistent
finding (Zittermann et al., 2000).
Other possible health benefits of prebiotics are now being
explored in many situations, facilitated by their safety and
ease of use. A substantial literature is accumulating on
prebiotics and cancer. However, much of the published work
is in animals where the role of prebiotics looks to be
beneficial, whereas human studies are mostly concerned
with identification of early biomarkers of risk (Pool-Zobel,
2005). Prebiotics are now being added to follow-on feeds for
infants (Fanaro et al., 2005), a practice which arises from the
clear benefits to children of probiotics in preventing and
ameliorating the symptoms of acute infectious diarrhoea,
and in atopic disease. Their use to prevent necrotizing
enterocolitis shows promise in animal models (Butel et al.,
2002). Prebiotics clearly change the gut microbiota of infants
and alter large bowel function, but large clinical trials are
awaited. Another area of importance is lipid metabolism
where prebiotic studies in animals have shown reduced
blood levels of cholesterol and triglycerides and beneficial
effects on fatty liver. Clinical trials in humans have not
yielded such consistent results (Williams and Jackson, 2002;
Beylot, 2005). However, the effects on hepatic lipid metabolism
are worth further study. There is also great interest in
prebiotics in the pet food and animal feed industry
(Flickinger and Fahey, 2002), where improved control of
gastrointestinal infection is reported and enhanced growth
performance is seen in poultry especially. Other areas of
interest include prebiotics and immunomodulation and the
gut immune system, glycaemic control, behavioural effects,
especially cognitive performance, and the enhancement of
probiotic activity in synbiotics.
Prebiotics bring a new dimension to dietary carbohydrates,
which might not have been predicted. Their ability to
change the composition of the gut flora towards one that
should protect against infection and their effect on calcium
absorption are very important for public health. However,
much work needs to be done in determining intakes of
prebiotics in individuals and populations, on their mechanisms
of action in the gut and potential effect on the immune
system, inflammation and cancer.
Digestibility concepts
Within the nutrition community there is an awareness that
the digestion (mechanical, chemical and enzymic breakdown)
of food and absorption of products takes place in
characteristically different regions of the gut. Of particular
consequence has been the division of digestive processes
into those that take place in the small bowel vs the large
bowel. Small bowel digestion and absorption occurs for most
Physiological aspects of energy metabolism
M Elia and JH Cummings
nutrients, whereas the large bowel, through the versatility of
its anaerobic bacteria, deals mainly with carbohydrates such
as NSP, non-a-glucan short-chain carbohydrates (oligosaccharides),
RS and polyols.
The end products of carbohydrate breakdown vary in
different regions of the gut with sugars appearing in blood
and insulin secretion stimulated after small bowel absorption,
while from the colon the process of fermentation yields
SCFAs, which metabolically are entirely different (Remesy
et al., 1995). Other divisions within the gut are also valid,
especially between duodenum/jejunum and ileum with, for
example, iron being absorbed in the duodenum and upper
jejunum and vitamin B12 in the terminal ileum. These
patterns of digestion are significant for health and extend
across the animal kingdom depending largely on the
anatomy of the gut. For example, in ruminants, fermentation
takes place primarily in the upper gut (the stomach
and rumen).
Knowledge of the importance of this contrasting physiology
linked to different regions of the gut has led to suggested
classifications of nutrients, especially carbohydrates on the
basis of the apparent site of their digestion and absorption.
An example is fibre, originally the cell wall carbohydrates of
plants such as cellulose, hemicellulose and pectin. However,
as our understanding of carbohydrate digestion has increased,
it has become clear that while regional differences in
function occur, there is no such clarity in regional differences
in carbohydrate digestion. Most carbohydrates can
reach the colon, for example, lactose, polyols, short-chain
carbohydrates, some starches and all NSP. For some of these,
for example, polyols or starch, the amount can vary between
meals and individuals, depending on factors such as dose
and transit time (Stephen et al., 1983; Cummings and
Englyst, 1991; Silvester et al., 1995; Cummings et al., 1996;
Livesey, 2003b). To describe carbohydrates by the region of
their digestion in the gut is, therefore, not an exact science.
Moreover, the gut acts as a single organ in digestion.
Internal regulation of gut digestive processes occurs through
neuro-endocrine loops between different regions. A classic
example is the gastro-colic reflex, but much more intricate
feedback occurs between regions of the gut (Cooke and
Reddix, 1994; Nightingale et al., 1996; Camilleri, 2006).
Furthermore, as already indicated in the discussion of
energy values of food, ‘digestibility is defined as the
proportion of combustible energy that is absorbed over the
entire length of the gastrointestinal tract’. Other terms in use
include ‘true digestibility’ (small bowel), ‘fermentable’ (large
bowel) and ‘apparent digestibility’ (whole gut).
There is a strong case for looking at carbohydrate digestion
as an integrated whole gut process. For the purposes of
classifying food as carbohydrates, both their DE and
digestion and absorption should be seen as a single
integrated system. This is not to underestimate the importance
of regional differences in these processes and their
subsequent metabolic effects, but the digestion and classification
of carbohydrate should be based on their chemistry
S67
European Journal of Clinical Nutrition

S68
while considering their digestion and digestibility, a whole
gut approach should be used.
Conclusion
Being quantitatively the most important dietary energy
source for most populations, carbohydrates have a special
role to play in energy metabolism and homeostasis. The
overview provided here deals with only selected physiological
effects of energy metabolism and gastrointestinal effects
of carbohydrates and their health implications. Several
carbohydrate-specific theories of appetite regulation have
been proposed but none are universally accepted, although
high-carbohydrate, low-energy-density diets rich in fruit,
vegetables and fibre are often recommended for weight
reduction or prevention of weight gain. Appetite and
hunger, which are of fundamental to survival, appear to
have many layers of control, with one layer compensating or
dominating another in some circumstances. The recent
growth of overweight and obesity throughout the world is
related to lifestyle changes, which have placed the human in
the unusual situation (at least in evolutionary terms) of
having to defend against a combination of persistent
abundance of tasty food and reduced physical activity. There
is a need to better understand the interaction between
energy density, GI/load, palatability and other factors and
their effects on feeding behaviour. At the same time, there is
room to establish more rational national and international
dietary energy systems and to consider greater application of
the NME system, which has some advantages over the more
commonly used, ME system.
The other physiological effects of many carbohydrates
depend on the site, rate and extent of their digestion in and
absorption from the gut. The majority of mono- and
disaccharides, together with maltodextrins and most starch,
are hydrolyzed by pancreatic enzymes in the small bowel
and at the epithelial surface, and the resultant monosaccharide
mixtures absorbed transported to the liver and then
stored or metabolized. These carbohydrates are primarily
energy yielding. The non-a-glucan oligosaccharides have
other properties, in that they increase calcium absorption
and some selectively modify the composition of the large
bowel microbiota to one dominated by bifidobacteria and
lactobacilli, known as the prebiotic effect. Studies to
demonstrate proven health benefits of prebiotics are
awaited.
The prebiotic carbohydrates, along with NSP, RS and some
polyols, reach the large bowel where they are fermented. The
principal end products, SCFA, are absorbed and provide a
further energy source to the tissues. Fermentation benefits
bowel habit, although the effects can be very small, and
provides mechanisms that could be important in cancer
prevention in the colon.
In the digestion of carbohydrates, the gut acts as an
integrated organ with distinct regions of function. The
European Journal of Clinical Nutrition
Physiological aspects of energy metabolism
M Elia and JH Cummings
nature of carbohydrate, both its chemical and physical
properties, together with the physiological responses of
absorption and secretion, local and visceral neuro-endocrine
reflexes, enzyme section and microbial activity, determine
the varied responses to carbohydrate in our diet.
Acknowledgements
We thank Professor Nils-Georg Asp, Professor Arne V Astrup,
Professor Nancy Keim, Dr Geoffrey Livesey, Professor Ian
MacDonald, Dr Gabriele Riccardi, Professor A Steward
Truswell and Dr Ricardo Uauy for the valuable comments
they provided on the earlier manuscript. We also thank Dr
Klaus Englyst and Dr R James Stubbs for helpful comments.
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 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).
Professor Marinos Elia: Since undertaking this work,
Professor Elia has joined a committee to undertake a
systematic review on the role of fibre in enteral tube feeding,
with an unrestricted grant from Numico.
Peer-reviewers
Professor Nils-Georg Asp: On part-time leave from university
professorship to be the Director of the Swedish
Nutrition Foundation (SNF), a nongovernmental organization
for the promotion of nutrition research and its practical
implications. SNF is supported broadly by the food sector;
the member organizations and industries are listed on the
SNF home page (www.snf.ideon.se).
Professor Arne V Astrup: Research grants from Arla Foods,
Danish Diary Association, Danish Meat Industry, Dutch
Diary Association, Schulstad (Bakery), Unilever and Weight
Watchers; speaker for Campina/Dutch Diary Association and
Suikerstichting, Holland.
Professor Nancy Keim: None declared.
Dr Geoffrey Livesey: Director and shareholder of Independent
Nutrition Logic Ltd, which also employs him as a
consultant to work with commercial, governmental and
educational establishments and to undertake research on
commissioned works on matters regarding health and
nutrition. Received payment or other support covering a
period during 2002–2006, and has an expectation of support
for the future from several commercial entities with an
interest in the subject matter of the FAO/WHO scientific
update, even if it does not convey any benefit to him
personally, but which benefits his position or administrative

unit, such as a grant or fellowship or payment for the
purpose of financing a post or consultancy.
Professor Ian MacDonald: None declared.
Dr Gabriele Riccardi: None declared.
Professor A Steward Truswell: None declared.
Professor Ricardo Uauy: Scientific Adviser on a temporary
basis for Unilever and Wyeth; Scientific Editorial/Award
Adviser for Danone, DSM, Kelloggs, and Knowles and Bolton
on an ad hoc basis.
References
Agatston A (2003). The South Beach Diet. Headline Book Publishing:
London.
Alfenas RC, Mattes RD (2005). Influence of glycemic index/load on
glycemic response, appetite, and food intake in healthy humans.
Diabetes Care 28, 2123–2129.
Alles MS, Hautvast JG, Nagengast FM, Hartemink R, Van Laere KM,
Jansen JB (1996). Fate of fructo-oligosaccharides in the human
intestine. Br J Nutr 76, 211–221.
Anderson GH, Catherine NL, Woodend DM, Wolever TM (2002).
Inverse association between the effect of carbohydrates on blood
glucose and subsequent short-term food intake in young men. Am
J Clin Nutr 76, 1023–1030.
Anderson GH, Woodend D (2003). Effect of glycemic carbohydrates
on short-term satiety and food intake. Nutr Rev 61, S17–S26.
AOAC (2002). AOAC Official Methods of AOAC International, 17th edn.
Gaithersburg, MD, USA.
Armstrong DG (1969). Cell bioenergetics and energy metabolism. In:
Lenkeit W, Breiren D, Crassman E (eds). Handbuch Der Tierernahrung.
Parey Publishing: Hamburg.
Astrup A, Buemann B, Flint A, Raben A (2002). Low-fat diets and
energy balance: how does the evidence stand in 2002? Proc Nutr
Soc 61, 299–309.
Astrup A, Grunwald GK, Melanson EL, Saris WH, Hill JO (2000). The
role of low-fat diets in body weight control: a meta-analysis of ad
libitum dietary intervention studies. Int J Obes Relat Metab Disord
24, 1545–1552.
Ball SD, Keller KR, Moyer-Mileur LJ, Ding YW, Donaldson D, Jackson
WD (2003). Prolongation of satiety after low versus moderately high
glycemic index meals in obese adolescents. Pediatrics 111, 488–494.
Bar A (1990). Factorial calculation model for the estimation of the
physiological caloric value of polyols. In: Hosoya N (ed). Proceedings
of the International Symposium on the Caloric Evaluation of
Carbohydrates. The Japanese Association for of Dietetic and
Enriched Foods: Tokyo, pp 209–379.
Bathalon GP, Tucker KL, Hays NP, Vinken AG, Greenberg AS, McCrory
MA et al. (2000). Psychological measures of eating behavior and the
accuracy of 3 common dietary assessment methods in healthy
postmenopausal women. Am J Clin Nutr 71, 739–745.
Bell EA, Rolls BJ (2001). Energy density of foods affects energy intake
across multiple levels of fat content in lean and obese women. Am J
Clin Nutr 73, 1010–1018.
Berkey CS, Rockett HR, Field AE, Gillman MW, Colditz GA (2004). Sugaradded
beverages and adolescent weight change. Obes Res 12, 778–788.
Bernier J, Pascal G (1990). Valeur energetique des polyols (surcres
alcools). Med Nutr 26, 221–238.
Beylot M (2005). Effects of inulin-type fructans on lipid metabolism
in man and in animal models. Br J Nutr 93 (Suppl 1), S163–S168.
Bingham SA, Day NE, Luben R, Ferrari P, Slimani N, Norat T et al.
(2003). Dietary fibre in food and protection against colorectal
cancer in the European prospective investigation into cancer and
nutrition (EPIC): an observational study. Lancet 361, 1496–1501.
Black JL (2000). Modelling growth and lactation in pigs. In:
Theodorou KM, France J (eds). Feeding Systems and Feed Evaluation
Models. Wallingford, Oxon, UK, pp 363–391.
Physiological aspects of energy metabolism
M Elia and JH Cummings
Black RM, Anderson GH (1994). Sweeteners, food intake and
selection. In: Fernstrom JD, Miller GD (eds). Appetite and Body
Weight Regulation: Sugar, Fat and Macronutrient Substitutes. Boca
Raton, FL.
Black RM, Leiter LA, Anderson GH (1993). Consuming aspartame
with and without taste: differential effects on appetite and food
intake of young adult males. Physiol Behav 53, 459–466.
Blaxter KB (1989). Energy Metabolism in Animals and Man. Cambridge
University Press: Cambridge.
Blaxter KL (1971). Methods of measuring the energy metabolism
of animals and interpretation of results obtained. Fed Proc 30,
1436–1443.
Blundell JE, Burley VJ (1987). Satiation, satiety and the action of fibre
on food intake. Int J Obes 11 (Suppl 1), 9–25.
Blundell JE, Rogers PJ (1994). Sweet carbohydrate substitutes
(intense sweeteners) and control of appetite: scientific issues. In:
Fernstrom JD, Miller GD (eds). Appetite and Body Weight Regulation:
Sugar, Fat and Macronutrient Susbstitutes. Boca Raton, FL.
Blundell JE, Stubbs RJ (1999). High and low carbohydrate and fat
intakes: limits imposed by appetite and palatability and
their implications for energy balance. Eur J Clin Nutr 53 (Suppl 1),
S148–S165.
Bouhnik Y, Flourie B, D’Agay-Abensour L, Pochart P, Gramet G,
Durand M et al. (1997). Administration of transgalactooligosaccharides
increases fecal bifidobacteria and modifies
colonic fermentation metabolism in healthy humans. J Nutr 127,
444–448.
British Nutrition Foundation (1990). Energy from complex carbohydrates.
In: BNF Task Force on Complex Carbohydrates in Foods.
Chapman and Hall: London.
Brown J, Livesey G, Roe M, Faulks R, Poppitt S, Wilkinson J et al.
(1998). Metabolizable energy of high non-starch polysaccharidemaintenance
and weight-reducing diets in men: experimental
appraisal of assessment systems. J Nutr 128, 986–995.
Burley VJ, Blundell JE (1990). Action of dietary fibre on the satiety
cascade. In: Kritchevsky D, Bonfield C, Anderson JW (eds). Dietary
Fibre. Chemistry, Physiology and Health Effects. Plenum Press: New
York, NY.
Butel MJ, Waligora-Dupriet AJ, Szylit O (2002). Oligofructose and
experimental model of neonatal necrotising enterocolitis. Br J Nutr
87 (Suppl 2), S213–S219.
Camilleri M (2006). Integrated upper gastrointestinal response to
food intake. Gastroenterology 131, 640–658.
Campfield LA, Smith FJ (1990). Transient declines in blood glucose
signal meal initiation. Int J Obes 14 (Suppl 3), 15–31; discussion
31–34.
Campfield LA, Smith FJ (2003). Blood glucose dynamics and control
of meal initiation: a pattern detection and recognition theory.
Physiol Rev 83, 25–58.
Campfield LA, Smith FJ, Rosenbaum M (1992). Human hunger: is
there a role for blood glucose dynamics? Appetite 18, 244.
Canty DJ, Chan MM (1991). Effects of consumption of caloric vs
noncaloric sweet drinks on indices of hunger and food consumption
in normal adults. Am J Clin Nutr 53, 1159–1164.
Casey A, Mann R, Banister K, Fox J, Morris PG, Macdonald IA et al.
(2000). Effect of carbohydrate ingestion on glycogen resynthesis
in human liver and skeletal muscle, measured by (13)C MRS. Am
J Physiol Endocrinol Metab 278, E65–E75.
Cassidy A, Bingham SA, Cummings JH (1994). Starch intake and
colorectal cancer risk: an international comparison. Br J Cancer 69,
937–942.
Castiglia-Delavaud C, Verdier E, Besle JM, Vernet J, Boirie Y, Beaufrere
B et al. (1998). Net energy value of non-starch polysaccharide
isolates (sugarbeet fibre and commercial inulin) and their impact
on nutrient digestive utilization in healthy human subjects. Br
J Nutr 80, 343–352.
Centres for Disease Control and Prevention (CDC) (2004). Trends in
intake of energy and macronutrients—United States, 1971–2000.
MMWR Morb Mortal Wkly Rep 53, 80–82.
S69
European Journal of Clinical Nutrition

S70
Physiological aspects of energy metabolism
M Elia and JH Cummings
Champ M, Langkilde A-M, Brouns F et al. (2003). Advances in dietary
fibre characterization. 2. Consumption, chemistry, physiology and
measurement of resistant starch; implications for food health and
food labelling. Nutr Res Rev 16, 143–161.
Chen H-L, Lu Y-H, Lin JJ, Ko L-Y (2000). Effects of fructooligosaccharide
on bowel function and indicators of nutritional status
in constipated elderly men. Nutr Res 20, 1725–1733.
Commission E (2006). Directive 90/496/EEC on Nutrition Labelling of
Foodstuffs: Discussion Paper on Revision of Technical Issues. Services
for the European Commission.
Connor H, Annan F, Bunn E, Frost G, McGough N, Sarwar T et al.
(2003). The implementation of nutritional advice for people with
diabetes. Diabet Med 20, 786–807.
Cooke HJ, Reddix RA (1994). Neural regulation of intestinal
electrolyte transport. In: Johnson LR, Alpers DH, Christensen J,
Jacobson ED, Walsh JH (eds). Physiology of the Gastrointestinal Tract.
Raven Press, New York, pp 2083–2132.
Cummings JH (1981). Dietary fibre. Br Med Bull 37, 65–70.
Cummings JH (1994). Non-starch polysaccharides (dietary fibre)
including bulk laxatives in constipation. In: Kamm MA, Lennard-
Jones JE (eds). Constipation. Wrightson Biomedical Publishing Ltd:
Petersfield, UK, pp 307–314.
Cummings JH (1995). Short-chain fatty acids. In: Gibson GR,
Macfarlane GT (eds). Human Colonic Bacteria. CRC Press: Boca
Raton, FL, pp 101–130.
Cummings JH, Antoine JM, Azpiroz F, Bourdet-Sicard R, Brandtzaeg
P, Calder PC et al. (2004). PASSCLAIM—gut health and immunity.
Eur J Nutr 43 (Suppl 2), II118–II173.
Cummings JH, Beatty ER, Kingman SM, Bingham SA, Englyst HN
(1996). Digestion and physiological properties of resistant starch
in the human large bowel. Br J Nutr 75, 733–747.
Cummings JH, Englyst HN (1991). Measurement of starch fermentation
in the human large intestine. Can J Physiol Pharmacol 69,
121–129.
Cummings JH, Frolich W (1993). Dietary Fibre Intakes in Europe: A
Survey Conducted by Members of the Management Committee of COST
92. Metabolic and Physiological Aspects of Dietary Fibre in Food.
Brussels.
Cummings JH, Hill MJ, Jenkins DJ, Pearson JR, Wiggins HS (1976).
Changes in fecal composition and colonic function due to cereal
fiber. Am J Clin Nutr 29, 1468–1473.
Cummings JH, Macfarlane GT (1991). The control and consequences
of bacterial fermentation in the human colon. J Appl Bacteriol 70,
443–459.
Cummings JH, Macfarlane GT, Englyst HN (2001). Prebiotic
digestion and fermentation. Am J Clin Nutr 73, 415S–420S.
Cummings JH, Rombeau JL, Sakata T (1995). Physiological and Clinical
Aspects of Short-Chain Fatty Acids. Cambridge University Press:
Cambridge.
Cummings JH, Southgate DA, Branch WJ, Wiggins HS, Houston H,
Jenkins DJ et al. (1979). The digestion of pectin in the human gut
and its effect on calcium absorption and large bowel function. Br J
Nutr 41, 477–485.
de Castro JM (1991). Seasonal rhythms of human nutrient intake and
meal pattern. Physiol Behav 50, 243–248.
de Castro JM (2004). Dietary energy density is associated
with increased intake in free-living humans. J Nutr 134,
335–341.
de Castro JM, de Castro ES (1989). Spontaneous meal patterns of
humans: influence of the presence of other people. Am J Clin Nutr
50, 237–247.
Den Hond E, Geypens B, Ghoos Y (2000). Effect of high performance
chicory inulin on constipation. Nutr Res 20, 731–736.
Department of Health (1991). Dietary Reference Values for Food, Energy
and Nutrients for the United Kingdom. HMSO: London.
Diabetes and Nutrition Study Group of European Association for the
Study of Diabetes (2000). Recommendations for the nutritional
management of patients with diabetes mellitus. Eur J Clin Nutr 54,
353–355.
European Journal of Clinical Nutrition
DiMeglio DP, Mattes RD (2000). Liquid versus solid carbohydrate:
effects on food intake and body weight. Int J Obes Relat Metab
Disord 24, 794–800.
Drewnowski A (1998). Energy density, palatability, and satiety:
implications for weight control. Nutr Rev 56, 347–353.
Drossman DA, Camilleri M, Mayer EA, Whitehead WE (2002). AGA
technical review on irritable bowel syndrome. Gastroenterology
123, 2108–2131.
Druce M, Bloom SR (2006). The regulation of appetite. Arch Dis Child
91, 183–187.
Duncan KH, Bacon JA, Weinsier RL (1983). The effects of high and
low energy density diets on satiety, energy intake, and eating time
of obese and nonobese subjects. Am J Clin Nutr 37, 763–767.
Duncan SH, Scott KP, Ramsay AG, Harmsen HJ, Welling GW, Stewart
CS et al. (2003). Effects of alternative dietary substrates on
competition between human colonic bacteria in an anaerobic
fermentor system. Appl Environ Microbiol 69, 1136–1142.
Dutch Nutrition Council (1987). The Energy Values of Polyols. Recommendations
of the Committee on Polyols. Nutrition Council: The Hague.
Eckburg PB, Bik EM, Bernstein CN, Purdom E, Dethlefsen L, Sargent
M et al. (2005). Diversity of the human intestinal microbial flora.
Science 308, 1635–1638.
Elia M (2005). Insights into energy requirements in disease. Public
Health Nutr 8, 1037–1052.
Elia M, Folmer P, Schlatmann A, Goren A, Austin S (1988).
Carbohydrate, fat, and protein metabolism in muscle and in the
whole body after mixed meal ingestion. Metabolism 37, 542–551.
Elia M, Jones MG, Jennings G, Poppitt SD, Fuller NJ, Murgatroyd PR
et al. (1995). Estimation of energy expenditure from specific
activity of urine urea during lengthy subcutaneous NaH 14 CO3
infusion. Am J Physiol 269, E172–E182.
Elia M, Livesey G (1988). Theory and validity of indirect calorimetry
during net lipid synthesis. Am J Clin Nutr 47, 591–607.
Elia M, Livesey G (1992). Energy expenditure and fuel selection in
biological systems: the theory and practice of calculations based
on indirect calorimetry and tracer methods. World Rev Nutr Diet
70, 68–131.
Elia M, Stratton R, Stubbs J (2003). Techniques for the study of energy
balance in man. Proc Nutr Soc 62, 529–537.
Englyst HN, Cummings JH (1987). Digestion of polysaccharides
of potato in the small intestine of man. Am J Clin Nutr 45,
423–431.
Englyst KN, S Liu, Englyst HN (2007). Nutritional characterisation
and measurement of dietary carbohydrates. Eur J Clin Nutr 61
(Suppl 1), S19–S39.
European Communities (1990). Directive 90/496/EEC. Nutrition
labelling for foodstuffs. Off J Eur Commun L276, 40–44.
Fanaro S, Boehm G, Garssen J, Knol J, Mosca F, Stahl B et al. (2005).
Galacto-oligosaccharides and long-chain fructo-oligosaccharides as
prebiotics in infant formulas: a review. Acta Paediatr Suppl 94,22–26.
FAO (1998). Carbohydrates in human nutrition. Report of a Joint FAO/
WHO Expert Consultation. FAO Food and Nutrition Paper No. 66.
FAO: Rome.
FAO (2003). Food Energy—Methods of Analysis and Conversion Factors.
FAO: Rome.
FAO/WHO/UNU (1985). Energy and Protein Requirements. WHO:
Geneva.
FAO/WHO/UNU (2004). Human Energy Requirements. Food and
Nutrition Technical Series 1. FAO: Rome.
FASEB (1994). The Evaluation of the Energy of Certain Sugar Alcohols Used
as Food Ingredients. Life Science Research Office: Bethesda, MD.
Finley JW, Leveille GA (1996). Macronutrient substitutes. In:
Zeigler EE, Filer LJ (eds). Present Knowledge in Nutrition. ILSI Press:
Washington, DC, pp 581–595.
Flatt JP (1978). The biochemistry of energy expenditure. In: Bray GA
(ed). Recent Advances in Obesity Research. Newman Publications
Limited: London, pp 221–228.
Flatt JP (1980). Energetics of intermediary metabolism. In: Assessment
of Energy Metabolism in Health and Disease. Report of the First ROSS

Conference on Medical Research. Ross Laboratories: Columbus, OH,
pp 77–87.
Flatt JP (1987). The difference in the storage capacities for
carbohydrate and for fat, and its implications in the regulation
of body weight. Ann N Y Acad Sci 499, 104–123.
Flatt JP (1992). Energy costs of ATP synthesis. In: Kinney JM, Tucker
HN (eds). Energy Metabolism: Tissue Determinants and Cellular
Corollaries. Raven Press: New York, pp 319–343.
Flickinger EA, Fahey Jr GC (2002). Pet food and feed applications of
inulin, oligofructose and other oligosaccharides. Br J Nutr 87
(Suppl 2), S297–S300.
Flourie B, Florent C, Jouany JP, Thivend P, Etanchaud F, Rambaud JC
(1986). Colonic metabolism of wheat starch in healthy humans.
Effects on fecal outputs and clinical symptoms. Gastroenterology
90, 111–119.
Francis CY, Whorwell PJ (1994). Bran and irritable bowel syndrome:
time for reappraisal. Lancet 344, 39–40.
Freedman MR, King J, Kennedy E (2001). Popular diets: a scientific
review. Obes Res 9 (Suppl 1), 1S–40S.
Friedman MI, Ramirez I, Bowden CR, Tordoff MG (1990). Fuel
partitioning and food intake: role for mitochondrial fatty acid
transport. Am J Physiol 258, R216–R221.
Friedman MI, Stricker EM (1976). The physiological psychology of
hunger: a physiological perspective. Psychol Rev 83, 409–431.
Friedman MI, Tordoff MG (1986). Fatty acid oxidation and glucose
utilization interact to control food intake in rats. Am J Physiol 251,
R840–R845.
Fryer JH (1958). The effects of a late-night caloric supplement upon
body weight and food intake in man. Am J Clin Nutr 6, 354–364.
Fuller NJ, Jebb SA, Laskey MA, Coward WA, Elia M (1992). Fourcomponent
model for the assessment of body composition in
humans: comparison with alternative methods and evaluation of
the density and hydration of fat-free mass. Clin Sci 82, 687–693.
Garg A (1998). High-monounsaturated-fat diets for patients with
diabetes mellitus: a meta-analysis. Am J Clin Nutr 67, 577S–582S.
Garrow JS (1988). Obesity and Related Diseases. Churchill Livingstone:
Edinburgh.
Gibney E, Elia M, Jebb SA, Murgatroyd P, Jennings G (1996). Total
energy expenditure in patients with small cell lung cancer: results
of a validated study using the bicarbonate-urea method. Metabolism
46, 1412–1417.
Gibney ER, Murgatroyd P, Wright A, Jebb S, Elia M (2003).
Measurement of total energy expenditure in grossly obese women:
comparison of the bicarbonate-urea method with whole-body
calorimetry and free-living doubly labelled water. Int J Obes Relat
Metab Disord 27, 641–647.
Gibson GR, Beatty ER, Wang X, Cummings JH (1995). Selective
stimulation of bifidobacteria in the human colon by oligofructose
and inulin. Gastroenterology 108, 975–982.
Gibson GR, McCartney AL, Rastall RA (2005). Prebiotics and
resistance to gastrointestinal infections. Br J Nutr 93 (Suppl 1),
S31–S34.
Gibson GR, Probert HM, Van Loo J, Rastall RA, Roberfroid M (2004).
Dietary modulation of the human colonic microbiota: updating
the concept of prebiotics. Nutr Res Rev 17, 259–275.
Gibson GR, Roberfroid MB (1995). Dietary modulation of the human
colonic microbiota: introducing the concept of prebiotics. J Nutr
125, 1401–1412.
Gil KM, Skeie B, Kvetan V, Askanazi J, Friedman MI (1991). Parenteral
nutrition and oral intake: effect of glucose and fat infusions. JPEN J
Parenter Enteral Nutr 15, 426–432.
Goldring JM (2004). Resistant starch: safe intakes and legal status.
J AOAC Int 87, 733–739.
Gostner A, Schaffer V, Theis S, Menzel T, Luhrs H, Melcher R et al.
(2005). Effects of isomalt consumption on gastrointestinal and
metabolic parameters in healthy volunteers. Br J Nutr 94, 575–581.
Haber GB, Heaton KW, Murphy D, Burroughs LF (1977). Depletion
and disruption of dietary fibre. Effects on satiety, plasma-glucose,
and serum-insulin. Lancet 2, 679–682.
Physiological aspects of energy metabolism
M Elia and JH Cummings
Harper JA, Dickinson K, Brand MD (2001). Mitochondrial uncoupling
as a target for drug development for the treatment of obesity.
Obes Rev 2, 255–265.
Heijnen ML, van Amelsvoort JM, Deurenberg P, Beynen AC (1998).
Limited effect of consumption of uncooked (RS2) or retrograded
(RS3) resistant starch on putative risk factors for colon cancer in
healthy men. Am J Clin Nutr 67, 322–331.
Henderson L, Gregory J, Irving K, Swan G (2003). The National Diet &
Nutrition Survey: Adults Aged 19 to 64 Years. Energy, Protein,
Carbohydrate, Fat and Alcohol Intake. The Stationery Office:
London.
Heymsfield SB, Lohman TG, Wang W, Going SB (2005). Human Body
Composition, 2nd edn. Human Kinetics: Champaign, IL.
Hudson GJ, Englyst HN (1993). UK and some other countries. In:
Cummings JH, Frolich W (eds). Dietary Fibre Intakes in Europe.
CEC—Directorate General, Science, Research & Development:
Brussels, pp 83–87.
Hylla S, Gostner A, Dusel G, Anger H, Bartram HP, Christl SU et al.
(1998). Effects of resistant starch on the colon in healthy
volunteers: possible implications for cancer prevention. Am J Clin
Nutr 67, 136–142.
Institute of Medicine (2005). Dietary Reference Intakes for Energy,
Carbohydrate, Fiber, Fat, Fatty Acids, Protein, and Amino Acids. The
National Academies Press: Washington, DC.
Ito M, Deguchi Y, Miyamori A, Matsumoto K, Kiluchi H, Matsumoto
K et al. (1990). Effects of administration of galactooligosaccharides
on the human faecal microflora, stool weight and abdominal
sensation. Microb Ecol Health Dis 3, 285–292.
James J, Kerr D (2005). Prevention of childhood obesity by reducing
soft drinks. Int J Obes (Lond) 29 (Suppl 2), S54–S57.
James J, Thomas P, Cavan D, Kerr D (2004). Preventing childhood
obesity by reducing consumption of carbonated drinks: cluster
randomised controlled trial. BMJ 328, 1237.
Japanese Ministry of Health and Welfare (1991). Japanese Ministry
of Health and Welfare Official Notice: Evaluation of the Energy Value of
Indigestible Carbohydrates in Nutritive Foods. Docket Number Ei-shin
71. Japanese Ministry of Health and Welfare: Tokyo.
Jeukendrup AE (2004). Carbohydrate intake during exercise and
performance. Nutrition 20, 669–677.
Johnson J, Vickers Z (1993). Effects of flavor and macronutrient
composition of food servings on liking, hunger and subsequent
intake. Appetite 21, 25–39.
Johnstone AM, Stubbs RJ, Harbron CG (1996). Effect of overfeeding
macronutrients on day-to-day food intake in man. Eur J Clin Nutr
50, 418–430.
Kaplan RJ, Greenwood CE (2002). Influence of dietary carbohydrates
and glycaemic response on subjective appetite and food intake in
healthy elderly persons. Int J Food Sci Nutr 53, 305–316.
Kaskoun MC, Johnson RK, Goran MI (1994). Comparison of energy
intake by semiquantitative food-frequency questionnaire with
total energy expenditure by the doubly labeled water method in
young children. Am J Clin Nutr 60, 43–47.
King TS, Elia M, Hunter JO (1998). Abnormal colonic fermentation in
irritable bowel syndrome. Lancet 352, 1187–1189.
Langlands SJ, Hopkins MJ, Coleman N, Cummings JH (2004).
Prebiotic carbohydrates modify the mucosa associated microflora
of the human large bowel. Gut 53, 1610–1616.
Lavin JH, French SJ, Read NW (1997). The effect of sucrose- and
aspartame-sweetened drinks on energy intake, hunger and food
choice of female, moderately restrained eaters. Int J Obes Relat
Metab Disord 21, 37–42.
Ledikwe JH, Blanck HM, Kettel Khan L, Serdula MK, Seymour JD,
Tohill BC et al. (2006). Dietary energy density is associated with
energy intake and weight status in US adults. Am J Clin Nutr 83,
1362–1368.
Lee A, Livesey G, Storey D, Zumbe A (eds). (2001). International
symposium on low-digestible carbohydrates. Br J Nutr 85, S1–S64.
Levine AS, Billington CJ (1994). Appetite and Body Weight Regulation:
Sugar, Fat, and Macronutrient Substitutes. CRC Press: Boca Raton, FL.
S71
European Journal of Clinical Nutrition

S72
Physiological aspects of energy metabolism
M Elia and JH Cummings
Life Sciences Research Office (1994). The Evaluation of the Energy of
Certain Sugar Alcohols used as Food Ingredients. Life Sciences Research
Office, Federation of American Societies for Experimental Biology.
Bethesda, MD.
Lissner L, Levitsky DA, Strupp BJ, Kalkwarf HJ, Roe DA (1987).
Dietary fat and the regulation of energy intake in human subjects.
Am J Clin Nutr 46, 886–892.
Liu YJ, Araujo S, Recker RR, Deng HW (2003). Molecular and genetic
mechanisms of obesity: implications for future management. Curr
Mol Med 3, 325–340.
Livesey G (1984). The energy equivalents of ATP and the energy
values of food proteins and fats. Br J Nutr 51, 15–28.
Livesey G (1986). ATP yields from proteins, fats and carbohydrates
and mitochondrial efficiency in vivo. In: Barry EM, Blondheim SH,
Eliahou HE, Shafir E (eds). Recent Advances in Obesity Research. V.
John Libbey: London, Paris.
Livesey G (1991). Calculating the energy values of foods: towards
new empirical formulae based on diets with varied intakes of
unavailable complex carbohydrates. Eur J Clin Nutr 45, 1–12.
Livesey G (1992). The energy values of dietary fibre and sugar
alcohols for man. Nutr Res Rev 5, 61–84.
Livesey G (2001). A perspective on food energy standards for
nutrition labelling. Br J Nutr 85, 271–287.
Livesey G (2002). Thermogenesis associated with fermentable
carbohydrate in humans, validity of indirect calorimetry, and
implications of dietary thermogenesis for energy requirements,
food energy and body weight. Int J Obes Relat Metab Disord 26,
1553–1569.
Livesey G (2003a). Analytical issues in food energy and composition.
Energy in food labeling—including regulatory and trade issues.
Posited with FAO.
Livesey G (2003b). Health potential of polyols as sugar replacers,
with emphasis on low glycaemic properties. Nutr Res Rev 16,
163–191.
Livesey G, Elia M (1985). Food energy values for artificial feeds for
man. Clin Nutr 4, 99–111.
Livesey G, Elia M (1988). Estimation of energy expenditure, net
carbohydrate oxidation and net fat oxidation and synthesis by
indirect calorimetry. Evaluation of some errors with special
reference to the detailed composition of fuels. Am J Clin Nutr 47,
608–623.
Livesey G, Elia M (1995). Short chain fatty acids as an energy source
in the colon: metabolism and clinical implications. In: Cummings
JH, Rombeau JL, Sakata TT (eds). Physiological and Clinical Aspects
of Short Chain Fatty Acids. Cambridge University Press: Cambridge,
pp 1427–1482.
Livingstone MB, Black AE (2003). Markers of the validity of reported
energy intake. J Nutr 133 (Suppl 3), 895S–920S.
Longstreth GF, Thompson WG, Chey WD, Houghton LA, Mearin F,
Spiller RC (2006). Functional bowel disorders. Gastroenterology
130, 1480–1491.
Ludwig DS (2003). Dietary glycemic index and the regulation of body
weight. Lipids 38, 117–121.
Macfarlane GT, Cummings JH (1991). The colonic flora, fermentation
and large bowel digestive function. In: Phillips SF, Pemberton
JH, Shorter RG (eds). The Large Intestine: Physiology, Pathophysiology
and Disease. Raven Press: New York, pp 51–92.
Macfarlane S, Macfarlane GT (2004). Bacterial diversity in the human
gut. Adv Appl Microbiol 54, 261–289.
Macfarlane S, Macfarlane GT, Cummings JH (2006). Prebiotics in the
gastrointestinal tract. Aliment Pharmacol Ther 24, 701–714.
Mattes R (2006). Fluid calories and energy balance: the good, the bad,
and the uncertain. Physiol Behav 89, 66–70.
Mattes RD (1996). Dietary compensation by humans for supplemental
energy provided as ethanol or carbohydrate in fluids.
Physiol Behav 59, 179–187.
Matthews SB, Waud JP, Roberts AG, Campbell AK (2005). Systemic
lactose intolerance: a new perspective on an old problem. Postgrad
Med J 81, 167–173.
European Journal of Clinical Nutrition
Mayer J (1955). Regulation of energy intake and the body weight: the
glucostatic theory and the lipostatic hypothesis. Ann N Y Acad Sci
63, 15–43.
Mazlam N (2001). Effect of Fat and Carbohydrate on Energy Intake and
Macronutrient Selection in Humans. PhD thesis, University of
Aberdeen.
Merill AL, Watt BK (1973). Energy Value of Foods: Basis and Derivation.
United States Department of Agriculture: Washington, DC.
Miller JB, McVeagh P (1999). Human milk oligosaccharides: 130
reasons to breast-feed. Br J Nutr 82, 333–335.
Montalto M, Curigliano V, Santoro L, Vastola M, Cammarota G,
Manna R et al. (2006). Management and treatment of lactose
malabsorption. World J Gastroenterol 12, 187–191.
Mowat F (1981). People of the Deer. Seal Books: Toronto.
Nanda R, James R, Smith H, Dudley CR, Jewell DP (1989).
Food intolerance and the irritable bowel syndrome. Gut 30,
1099–1104.
Nasrallah SM, Iber FL (1969). Mannitol absorption and metabolism
in man. Am J Med Sci 258, 80–88.
Nightingale JM, Kamm MA, van der Sijp JR, Ghatei MA, Bloom SR,
Lennard-Jones JE (1996). Gastrointestinal hormones in short
bowel syndrome. Peptide YY may be the ‘colonic brake’ to gastric
emptying. Gut 39, 267–272.
Nishida C, Uauy R, Kumanyika S, Shetty P (2004). The joint WHO/
FAO expert consultation on diet, nutrition and the prevention of
chronic diseases: process, product and policy implications. Public
Health Nutr 7, 245–250.
Painter NS, Almeida AZ, Colebourne KW (1972). Unprocessed
bran in treatment of diverticular disease of the colon. BMJ 2,
137–140.
Paton NI, Angus B, Chaowagul W, Simpson AJ, Suputtamongkol Y,
Elia M et al. (2001). Protein and energy metabolism in chronic
bacterial infection: studies in melioidosis. Clin Sci (Lond) 100,
101–110.
Paton NI, Elia M, Jebb SA, Jennings G, Macallan DC, Griffin GE
(1996). Total energy expenditure and physical activity measured
with the bicarbonate-urea method in patients with human
immunodeficiency virus infection. Clin Sci (Lond) 91, 241–245.
Paul AA, Southgate DAT (1978). McCance and Widdowson’s The
Composition of Foods, 4th edn. HMSO: London.
Pawlak DB, Ebbeling CB, Ludwig DS (2002). Should obese patients be
counselled to follow a low-glycaemic index diet? Yes. Obes Rev 3,
235–243.
Phillips J, Muir JG, Birkett A, Lu ZX, Jones GP, O’Dea K et al. (1995).
Effect of resistant starch on fecal bulk and fermentation-dependent
events in humans. Am J Clin Nutr 62, 121–130.
Pool-Zobel BL (2005). Inulin-type fructans and reduction in colon
cancer risk: review of experimental and human data. Br J Nutr 93
(Suppl 1), S73–S90.
Poppitt SD, Prentice AM (1996). Energy density and its role in the
control of food intake: evidence from metabolic and community
studies. Appetite 26, 153–174.
Poppitt SD (1995). Energy density of diets and obesity. Int J Obes Relat
Metab Disord 19 (Suppl 5), S20–S26.
Poppitt SD, Livesey G, Faulks RM, Roe M, Prentice AM, Elia M (1996).
Circadian patterns of total 24-h hydrogen and methane excretion
in humans ingesting nonstarch polysaccharide (NSP) diets and the
implications for indirect calorimetric and D2 18O methodologies.
Eur J Clin Nutr 50, 524–534.
Porikos KP, Booth G, Van Itallie TB (1977). Effect of covert nutritive
dilution on the spontaneous food intake of obese individuals: a
pilot study. Am J Clin Nutr 30, 1638–1644.
Porikos KP, Hesser MF, van Itallie TB (1982). Caloric regulation in
normal-weight men maintained on a palatable diet of conventional
foods. Physiol Behav 29, 293–300.
Prentice AM (1990). The doubly-labelled water method for measuring
energy expenditure. Technical recommendations for use in
humans: a consensus report by the IDECG Working Group
International Atomic Energy Agency NAHRES-4.

Prentice AM, Poppitt SD (1996). Importance of energy density and
macronutrients in the regulation of energy intake. Int J Obes Relat
Metab Disord 20 (Suppl 2), S18–S23.
Pullicino E, Elia M (1991). Intravenous carbohydrate over-feeding: a
method for rapid nutritional repletion. Clin Nutr 10, 146–154.
Raben A (2002). Should obese patients be counselled to follow a lowglycaemic
index diet? No. Obes Rev 3, 245–256.
Raben A, Agerholm-Larsen L, Flint A, Holst JJ, Astrup A (2003). Meals
with similar energy densities but rich in protein, fat, carbohydrate,
or alcohol have different effects on energy expenditure and
substrate metabolism but not on appetite and energy intake. Am
J Clin Nutr 77, 91–100.
Raben A, Macdonald I, Astrup A (1997). Replacement of dietary fat by
sucrose or starch: effects on 14 d ad libitum energy intake, energy
expenditure and body weight in formerly obese and never-obese
subjects. Int J Obes Relat Metab Disord 21, 846–859.
Raben A, Tagliabue A, Christensen NJ, Madsen J, Holst JJ, Astrup A
(1994). Resistant starch: the effect on postprandial glycemia,
hormonal response, and satiety. Am J Clin Nutr 60, 544–551.
Raben A, Vasilaras TH, Moller AC, Astrup A (2002). Sucrose compared
with artificial sweeteners: different effects on ad libitum food
intake and body weight after 10 wk of supplementation in
overweight subjects. Am J Clin Nutr 76, 721–729.
Rabinowitch IM, Fowler AF (1934). Dinitrophenol. Can Med Assoc J
30, 128–133.
Remesy C, Demigne C, Morand C (1995). Metabolism of short-chain
fatty acids in the liver. In: Cummings JH, Rombeau JL, Sakata T
(eds). Physiological and Clinical Aspects of Short-Chain Fatty Acids.
Cambridge University Press: Cambridge, pp 171–190.
Roberfroid M (2005). Inulin-Type Fructans. Functional Food Ingredients.
CRC Press: Boca Raton, FL.
Rodin J (1990). Comparative effects of fructose, aspartame, glucose,
and water preloads on calorie and macronutrient intake. Am J Clin
Nutr 51, 428–435.
Rogers PJ, Carlyle JA, Hill AJ, Blundell JE (1988). Uncoupling sweet
taste and calories: comparison of the effects of glucose and three
intense sweeteners on hunger and food intake. Physiol Behav 43,
547–552.
Rolls BJ (1997). Fat and sugar substitutes and the control of food
intake. In: Rolls BJ, Anderson, GH, Steffen DG (eds). Nutritional
Implications of Macronutrient Substitutes. New York Academy of
Sciences: New York, NY, pp 180–193.
Rolls BJ, Bell EA (1999). Intake of fat and carbohydrate: role of energy
density. Eur J Clin Nutr 53 (Suppl 1), S166–S173.
Rumpler WV, Baer DJ, Rhodes DG (1998). Energy available from corn
oil is not different than that from beef tallow in high- or low-fiber
diets fed to humans. JNutr128, 2374–2382.
Russek M (1963). Participation of hepatic glucoreceptors in the
control of intake of food. Nature 197, 79–80.
Russek M (1976). Hepatic glucoreceptors in the short-term control of
feeding. Acta Physiol Pol 27, 147–156.
Sanchez-Catillo C, Grubb D, Solano ML, Franklin F, James WPT
(1997). Non-starch polysaccharide intakes in Mexican villagers
and residents of Mexico City. Br J Nutr 77, 345–357.
Saris WH, Astrup A, Prentice AM, Zunft HJ, Formiguera X,
Verboeket-van de Venne WP et al. (2000). Randomized controlled
trial of changes in dietary carbohydrate/fat ratio and simple vs
complex carbohydrates on body weight and blood lipids:
the CARMEN study. The Carbohydrate Ratio Management
in European National diets. Int J Obes Relat Metab Disord 24,
1310–1318.
Schulz AR (1975). Computer-based method for calculation of the
available energy of proteins. J Nutr 105, 200–207.
Schulz AR (1978). Simulation of energy metabolism in the simplestomached
animal. Br J Nutr 39, 235–254.
Schulze MB, Manson JE, Ludwig DS, Colditz GA, Stampfer MJ, Willett
WC et al. (2004). Sugar-sweetened beverages, weight gain, and
incidence of type 2 diabetes in young and middle-aged women.
JAMA 292, 927–934.
Physiological aspects of energy metabolism
M Elia and JH Cummings
Sclafani A (1987). Carbohydrate-induced hyperphagia and obesity in
the rat: effects of saccharide type, form, and taste. Neurosci
Biobehav Rev 11, 155–162.
Sclafani A (2001). Psychobiology of food preferences. Int J Obes Relat
Metab Disord 25 (Suppl 5), S13–S16.
Shetty PS, Prentice AM, Goldberg GR, Murgatroyd PR, McKenna AP,
Stubbs RJ et al. (1994). Alterations in fuel selection and voluntary
food intake in response to isoenergetic manipulation of glycogen
stores in humans. Am J Clin Nutr 60, 534–543.
Siavoshian S, Segain JP, Kornprobst M, Bonnet C, Cherbut C,
Galmiche JP et al. (2000). Butyrate and trichostatin A effects on
the proliferation/differentiation of human intestinal epithelial
cells: induction of cyclin D3 and p21 expression. Gut 46, 507–514.
Silvester KR, Bingham SA, Pollock JR, Cummings JH, O’Neill IK
(1997). Effect of meat and resistant starch on fecal excretion of
apparent N-nitroso compounds and ammonia from the human
large bowel. Nutr Cancer 29, 13–23.
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.
Snitker S, Larson DE, Tataranni PA, Ravussin E (1997). Ad libitum food
intake in humans after manipulation of glycogen stores. Am J Clin
Nutr 65, 941–946.
Snook J, Shepherd HA (1994). Bran supplementation in the
treatment of irritable bowel syndrome. Aliment Pharmacol Ther 8,
511–514.
Southgate DA (1969). Determination of carbohydrates in foods. II.
Unavailable carbohydrates. J Sci Food Agric 20, 331–335.
Spengler M, Boehme K (1984). Tolerance and Acceptance of Isomalt
Study 26 report no. 13850, Wupperpal, Germany (Rn Study 26, editor:
Bayer AG Pharmaceuticals Division).
Spiegelman BM, Flier JS (2001). Obesity and the regulation of energy
balance. Cell 104, 531–543.
Stephen A (1993). Proprionate—sources and effect on lipid metabolism.
In: Binder HJ, Cummings JH, Soergel KH (eds). Short-Chain
Fatty Acids. Falk Symposium. Kluwer Academic Publishers, Dordrecht,
pp 260–271.
Stephen AM, Cummings JH (1980). Mechanism of action of dietary
fibre in the human colon. Nature 284, 283–284.
Stephen AM, Haddad AC, Phillips SF (1983). Passage of carbohydrate
into the colon. Direct measurements in humans. Gastroenterology
85, 589–595.
Stephen AM, Sieber GM, Gerster YA, Morgan DR (1995). Intake of
carbohydrate and its components—international comparisons,
trends over time, and effects of changing to low-fat diets. Am J Clin
Nutr 62, 851S–867S.
Stevens J (1988). Does dietary fiber affect food intake and body
weight? J Am Diet Assoc 88, 939–942; 945.
Stock MJ (1999). Gluttony and thermogenesis revisited. Int J Obes
Relat Metab Disord 23, 1105–1117.
Stollman N, Raskin JB (2004). Diverticular disease of the colon.
Lancet 363, 631–639.
Stubbs J, Ferres S, Horgan G (2000). Energy density of foods: effects
on energy intake. Crit Rev Food Sci Nutr 40, 481–515.
Stubbs RJ, Harbron CG, Murgatroyd PR, Prentice AM (1995a). Covert
manipulation of dietary fat and energy density: effect on substrate
flux and food intake in men eating ad libitum. Am J Clin Nutr 62,
316–329.
Stubbs RJ, Harbron CG, Prentice AM (1996). Covert manipulation of
the dietary fat to carbohydrate ratio of isoenergetically dense diets:
effect on food intake in feeding men ad libitum. Int J Obes Relat
Metab Disord 20, 651–660.
Stubbs RJ, Hughes DA, Johnstone AM, Whybrow S, Horgan GW, King
N et al. (2004). Rate and extent of compensatory changes in energy
intake and expenditure in response to altered exercise and diet
composition in humans. Am J Physiol Regul Integr Comp Physiol
286, R350–R358.
Stubbs RJ, Ritz P, Coward WA, Prentice AM (1995b). Covert
manipulation of the ratio of dietary fat to carbohydrate and
S73
European Journal of Clinical Nutrition

S74
Physiological aspects of energy metabolism
M Elia and JH Cummings
energy density: effect on food intake and energy balance in freeliving
men eating ad libitum. Am J Clin Nutr 62, 330–337.
Stubbs RJ, Whybrow S (2004). Energy density, diet composition and
palatability: influences on overall food energy intake in humans.
Physiol Behav 81, 755–764.
Tainter ML, Stockton AB, Cutting WC (1935). Dinitrophenol in the
treatment of obesity. Final report. JAMA 105, 332–337.
Talley NJ, Spiller R (2002). Irritable bowel syndrome: a little
understood organic bowel disease? Lancet 360, 555–564.
Tang NL, Chung ML, Elia M, Hui E, Lum CM, Luk JK et al. (2002).
Total daily energy expenditure in wasted chronic obstructive
pulmonary disease patients. Eur J Clin Nutr 56, 282–287.
Taylor R, Price TB, Rothman DL, Shulman RG, Shulman GI (1992).
Validation of 13C NMR measurement of human skeletal muscle
glycogen by direct biochemical assay of needle biopsy samples.
Magn Reson Med 27, 13–20.
Teuri U, Korpela R (1998). Galacto-oligosaccharides relieve constipation
in elderly people. Ann Nutr Metab 42, 319–327.
Thompson DA, Campbell RG (1977). Hunger in humans induced by
2-deoxy-D-glucose: glucoprivic control of taste preference and
food intake. Science 198, 1065–1068.
Thompson DA, Welle SL, Lilavivat U, Penicaud L, Campbell RG
(1982). Opiate receptor blockade in man reduces 2-deoxy-Dglucose-induced
food intake but not hunger, thirst, and hypothermia.
Life Sci 31, 847–852.
Titter S, Calingasan NY (1994). Neural Substrates for Metabolic Control
of Feeding. CRC Press: Boca Raton, FL.
Todesco T, Rao AV, Bosello O, Jenkins DJ (1991). Propionate lowers
blood glucose and alters lipid metabolism in healthy subjects. Am J
Clin Nutr 54, 860–865.
Tomlin J, Read NW (1990). The effect of resistant starch on colon
function in humans. Br J Nutr 64, 589–595.
Tordoff MG, Alleva AM (1990a). Effect of drinking soda sweetened
with aspartame or high-fructose corn syrup on food intake and
body weight. Am J Clin Nutr 51, 963–969.
Tordoff MG, Alleva AM (1990b). Oral stimulation with aspartame
increases hunger. Physiol Behav 47, 555–559.
Tremblay A (1992). Human obesity: a defect in lipid oxidation or in
thermogenesis? Int J Obes Relat Metab Disord 16, 953–957.
Tremblay A, Lavallee N, Almeras N, Allard L, Despres JP, Bouchard C
(1991). Nutritional determinants of the increase in energy intake
associated with a high-fat diet. Am J Clin Nutr 53, 1134–1137.
Truswell AS, Seach JM, Thorburn AW (1988). Incomplete absorption
of pure fructose in healthy subjects and the facilitating effect of
glucose. Am J Clin Nutr 48, 1424–1430.
European Journal of Clinical Nutrition
van Dokkum W, Wezendonk B, Srikumar TS, van den Heuvel EG
(1999). Effect of nondigestible oligosaccharides on large-bowel
functions, blood lipid concentrations and glucose absorption in
young healthy male subjects. Eur J Clin Nutr 53, 1–7.
van Loo J, Coussement P, de Leenheer L, Hoebregs H, Smits G
(1995). On the presence of inulin and oligofructose as
natural ingredients in the western diet. Crit Rev Food Sci Nutr 35,
525–552.
van Munster IP, Tangerman A, Nagengast FM (1994). Effect of
resistant starch on colonic fermentation, bile acid metabolism,
and mucosal proliferation. Dig Dis Sci 39, 834–842.
van Stratum P, Lussenburg RN, van Wezel LA, Vergroesen AJ, Cremer
HD (1978). The effect of dietary carbohydrate: fat ratio on energy
intake by adult women. Am J Clin Nutr 31, 206–212.
Warren JM, Henry CJ, Simonite V (2003). Low glycemic index
breakfasts and reduced food intake in preadolescent children.
Pediatrics 112, e414.
Warwick PM, Baines J (2000). Point of view: energy factors for food
labelling and other purposes should be derived in a consistent
fashion for all food components. Br J Nutr 84, 897–902.
Welle SL, Thompson DA, Campbell RG, Lilavivathana U (1980).
Increased hunger and thirst during glucoprivation in humans.
Physiol Behav 25, 397–403.
Westerterp KR, Goris AH (2002). Validity of the assessment of dietary
intake: problems of misreporting. Curr Opin Clin Nutr Metab Care 5,
489–493.
Westerterp-Plantenga MS (2001). Analysis of energy density of food
in relation to energy intake regulation in human subjects. Br J Nutr
85, 351–361.
Weststrate JA (1992). Effect of Nutrients on the Regulation of Food Intake.
Unilever Vlaardingen, The Netherlands.
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.
Williams CM, Jackson KG (2002). Inulin and oligofructose: effects on
lipid metabolism from human studies. Br J Nutr 87 (Suppl 2),
S261–S264.
Yu-Poth S, Zhao G, Etherton T, Naglak M, Jonnalagadda S, Kris-
Etherton PM (1999). Effects of the national cholesterol education
program’s step I and step II dietary intervention programs on
cardiovascular disease risk factors: a meta-analysis. Am J Clin Nutr
69, 632–646.
Zittermann A, Bock P, Drummer C, Scheld K, Heer M, Stehle P (2000).
Lactose does not enhance calcium bioavailability in lactosetolerant,
healthy adults. Am J Clin Nutr 71, 931–936.

REVIEW
Carbohydrate intake and obesity
RM van Dam 1,2,3 and JC Seidell 3
1 Department of Nutrition, Harvard School of Public Health, Boston, MA, USA; 2 Channing Laboratory, Department of Medicine,
Brigham and Women’s Hospital, and Harvard Medical School, Boston, MA, USA and 3 Institute of Health Sciences, Vrije Universiteit
Amsterdam, Amsterdam, The Netherlands
The prevalence of obesity has increased rapidly worldwide and the importance of considering the role of diet in the prevention
and treatment of obesity is widely acknowledged. This paper reviews data on the effects of dietary carbohydrates on body
fatness. Does the composition of the diet as related to carbohydrates affect the likelihood of passive over-consumption and longterm
weight change? In addition, methodological limitations of both observational and experimental studies of dietary
composition and body weight are discussed. Carbohydrates are among the macronutrients that provide energy and can thus
contribute to excess energy intake and subsequent weight gain. There is no clear evidence that altering the proportion of total
carbohydrate in the diet is an important determinant of energy intake. However, there is evidence that sugar-sweetened
beverages do not induce satiety to the same extent as solid forms of carbohydrate, and that increases in sugar-sweetened soft
drink consumption are associated with weight gain. Findings from studies on the effect of the dietary glycemic index on body
weight have not been consistent. Dietary fiber is associated with a lesser degree of weight gain in observational studies.
Although it is difficult to establish with certainty that fiber rather than other dietary attributes are responsible, whole-grain
cereals, vegetables, legumes and fruits seem to be the most appropriate sources of dietary carbohydrate.
European Journal of Clinical Nutrition (2007) 61 (Suppl 1), S75–S99. doi:10.1038/sj.ejcn.1602939
Keywords: obesity; diet; carbohydrate; fiber; sugar; glycemic index
General introduction
Background
The prevalence of overweight and obesity has increased
rapidly worldwide during recent decades, acquiring epidemic
proportions in children and adults and in industrialized as
well as transitional and developing countries (Popkin and
Gordon-Larsen, 2004; Ogden et al., 2006). Excess adiposity
increases risk of type 2 diabetes, arthritis, sleep apnea,
hypertension, dyslipidemia, cardiovascular diseases, various
types of cancer and premature death (Willett et al., 1999).
Therefore, the importance of prevention and treatment of
obesity is widely acknowledged. Changes in energy storage
as body fat are affected by the balance of energy intake and
energy expenditure, making diet and physical activity
obvious targets for interventions. Effects of dietary composition
on both energy intake and energy expenditure (dietary
induced thermogenesis and resting energy metabolism) are
plausible, based on results from animal experiments and
metabolic studies in humans (Poppitt and Prentice, 1996;
Correspondence: Dr RM van Dam, Department of Nutrition, Harvard School
of Public Health, 665 Huntington Avenue, Boston, MA 02115, USA.
E-mail: rvandam@hsph.harvard.edu
European Journal of Clinical Nutrition (2007) 61 (Suppl 1), S75–S99
& 2007 Nature Publishing Group All rights reserved 0954-3007/07 $30.00
www.nature.com/ejcn
Ludwig, 2002; Bray et al., 2004a, b; Halton and Hu, 2004;
Slavin, 2005). Although substantial short-term weight loss can
be achieved by many people, successful long-term maintenance
of weight loss is much more difficult and compensatory
physiological processes appear to stimulate weight regain
(Hirsch et al., 1998). Effects on body weight found in shortterm
metabolic studies can therefore not be readily extrapolated
to long-term effects. This overview will present
evidence for the effects of dietary composition related to
carbohydrates on body fatness. Dietary factors that will be
discussed include the proportion of total carbohydrates in
the diet, free-sugars, sugar-sweetened beverages, the dietary
glycemic index (GI) and dietary fiber. Studies have been
identified through systematic and narrative reviews on these
topics supplemented with searches in MEDLINE (PubMed)
until July 2007. The emphasis is on longer-term studies, but
shorter-term studies are also discussed depending on the
availability of data for different exposures.
Methodological considerations for studies relating diet to body
weight
Methodological limitations have to be considered in the
interpretation of results on macronutrient composition in

S76
relation to weight change or the incidence of obesity. We
will discuss these methodological limitations separately for
observational and experimental studies.
Observational studies. In these studies, individuals typically
report their food intakes by interviews or questionnaires, and
the obtained estimates are related to their body mass index
(BMI) or change in body weight. There are several methodological
issues that make results from observational studies
of diet and body weight difficult to interpret. First, the
possibility of an effect of perceived body weight or changes
in body weight on dietary habits (‘reverse causation’) should
be considered. In contrast to various other health outcomes,
people tend to be highly aware of their body weight and
changes therein. In addition, many people hold strong
beliefs about the relation between the composition of the
diet and body weight and have control over two of the main
determinants of body weight: energy intake and energy
expenditure through physical activity. Cross-sectional studies
where diet and measures of obesity are assessed
simultaneously are therefore difficult to interpret: the
perception that their body weight is high or increasing
may lead persons to change their dietary habits (for example,
dieting). For this reason, prospective observational studies
relating dietary intakes to changes in body weight would be
preferable. However, because changes in energy balance are
likely to almost directly translate into changes in body
weight, it seems biologically most relevant to study changes
in dietary intakes in relation to changes in body weight over
the same period. Because the exposure is not assessed before
the outcome, such an analysis is not truly prospective and
associations may still reflect an effect of perceived changes in
body weight on changes in dietary habits over the same
period.
Second, selective underreporting of dietary intakes can be
correlated with the degree of overweight (Heitmann and
Lissner, 1995; Heerstrass et al., 1998). There are indications
that intakes of carbohydrates and fat are more subject to
underreporting than intakes of protein and this can bias
results of studies of macronutrient composition and body
weight.
Third, dietary intakes and reporting thereof can be
correlated to many other characteristics such as age, sex,
socioeconomic status and other health-related habits that
may affect energy balance (Braam et al., 1998). These
characteristics can confound the association between dietary
intakes and body weight. For example, individuals who are
able to adhere to a generally recommended diet for weight
management (for example, a low-fat diet) are also more
likely to be able to adhere to limited total energy intake.
Also, associations between (changes in) dietary intakes and
energy balance have to be interpreted in the context of
(changes in) energy expenditure. The latter is notoriously
difficult to assess and is also subject to reporting bias. These
methodological issues complicate the interpretation of data
European Journal of Clinical Nutrition
Carbohydrate intake and obesity
RM van Dam and JC Seidell
on macronutrient intakes and changes therein in relation to
adiposity in observational studies.
Experimental studies. In evidence-based medicine, a stronger
weight is generally given to long-term randomized experimental
studies than to observational studies. In the last
several decades, many disputes have been published on the
interpretation of experimental studies that have manipulated
macronutrient composition of diets and evaluated
changes in body weight. The same data can be interpreted in
different ways. Some authors have argued that an increasing
proportion of energy coming from fat leads to greater weight
gain (Bray et al., 2004b), whereas others have concluded that
the proportion of energy from fat does not substantially
influence body weight (Willett, 2002). An important issue
here is that the outcomes of experimental studies seem to
be dependent on the choice of subjects (for example,
overweight versus normal weight subjects), the duration of
the experiment (short-term trials of days or weeks versus
trials that last several years) and the choice of foods that have
been used to manipulate macronutrient composition. With
respect to the latter issue, a low-fat, high-carbohydrate diet
can be a diet consisting mainly of highly refined grains and
products with added sugar, or a diet that is close to being a
traditional vegetarian diet (that is, with plenty of whole
grains, legumes, fruits and vegetables). In addition, the effect
of macronutrients on satiety and energy expenditure may
depend on the way diets are administered. Energy intake is
the outcome of the portion size energy density frequency
of consumption, and all the three factors can be altered
experimentally in relation to macronutrient composition.
Furthermore, experimental studies in humans rely on the
degree of successful (preferably double blinded) randomization
and on the compliance of subjects with dietary
regimens. Because both issues are problematic for long-term
trials of macronutrient intakes and weight change, the
strength of the evidence provided by randomized controlled
trials can be limited. Finally, experimental studies can be
performed by changing macronutrient intake with fixed or
ad libitum energy intakes and only the latter will provide us
with insights that are directly relevant for public health.
Effects of dietary intakes in the context of weight management
in obese persons may be different from the effects on
prevention of weight gain in leaner persons. The transition
from normal weight to obesity can result in changes in the
levels of hormones such as insulin, leptin and adiponectin,
which may alter relative substrate oxidation (fat versus
carbohydrates) and appetite control (Blaak, 2004; Schwartz
and Porte, 2005). In addition, the obese state alters basal and
24-h energy requirements as well as the sensitivity to various
hormones such as insulin and leptin. One should therefore
be cautious in the extrapolation of findings from experimental
studies on macronutrient composition and weight
loss in obese persons, to the role of macronutrients in the
prevention of weight gain.

Total carbohydrate intake
Introduction
Weight gain is the result of higher energy intake than energy
expenditure. This is also known as a positive energy balance.
The total amount of energy (expressed in units of kilocalories
or kilojoules per day) ingested by food and drinks come from
four major nutrients (macronutrients).
Macronutrient kJ/g kcal/g
Fat 37 9
Alcohol 29 7
Protein 17 4
Carbohydrate 16 4
As shown in the table above, fat contains more energy per
gram than carbohydrates. Carbohydrates, however, also
provide energy and therefore contribute to the total energy
intake per day and thus potentially to a positive energy
balance. One of the most controversial questions in human
nutrition in recent decades has been whether it matters for
energy balance what the relative contribution of macronutrients
is to the total energy intake. Potentially, there
could be the differences because of variations in effects on
appetite and satiety or in effects on oxidation and energy
expenditure for different macronutrients. Specifically, it has
been suggested that a higher proportion of fat in the diet can
lead to weight gain through excess energy intake, because it
is less satiating than the same amount of energy from
carbohydrates (Bray et al., 2004b). Others have suggested
that proteins are particularly satiating (Halton and Hu,
2004). The answer to this question is important because if
energy intake in the form of one macronutrient is more
likely to lead to a positive energy balance than energy intake
from other macronutrients, this would provide the basis to
emphasize the reduction of the intake of the former
macronutrient in recommendations for prevention of weight
gain or for achieving weight loss in overweight persons.
Before discussing evidence on the relation between carbohydrate
content of the diet and body weight, we will discuss
research on energy density, because it is frequently considered
to be an important mediator of effects of dietary
composition on energy balance.
Energy density
Carbohydrates and energy density. The energy density is the
amount of energy per unit of weight of foods, meals or diets
(Prentice and Jebb, 2003; Stubbs and Whybrow, 2004).
Carbohydrate provides less energy per gram than fat and is
thus less energy dense. However, few foods only contain
macronutrients, and the fiber and particularly the water
content (or conversely, the dryness) has a major effect on the
energy density of foods (Drewnowski et al., 2004). As a result,
foods with a high energy percentage of carbohydrates can
Carbohydrate intake and obesity
RM van Dam and JC Seidell
Lettuce
Vegetable soup
Skim milk
Apple
Black beans
White fish
Yogurt
Vegetable lasagna
Roast chicken
White bread
Pretzels
Cheddar cheese
Salad dressing
Potato chips
Bacon
Butter
0 1 2 3 4 5 6 7 8
Energy Density (kcal/g)
Figure 1 Energy density of selected commonly used foods.
Reprinted from Klein et al. (2002) copyright 2002, with permission
from Elsevier (figure provided courtesy of Liane Roe).
range from a low (for example, raw vegetables and fruits) to a
high (for example, sugary candy) energy density (Drewnowski
et al., 2004). Figure 1 shows the energy density of commonly
used foods, illustrating that a dry high-carbohydrate food
such as pretzels can have similar energy density as high-fat
foods such as cheese (Klein et al., 2002). The carbohydrate
content of diets tend to have a modest inverse association
with the energy density of diets, whereas a higher fat content
is generally associated with a higher energy density of diets
(Stookey, 2001; Drewnowski et al., 2004). However, whether
a diet with a moderately high energy percentage of fat has a
high or low energy density depends to a large extent on the
amount of fruits and vegetables consumed (Ledikwe et al.,
2006). Because of their high water content, beverages
generally have a lower energy density than solid foods.
However, in the interpretation of the energy density of diets
and foods it seems appropriate to consider foods and
beverages separately given indications that energy intake
from beverages is regulated differently (Mattes, 1996; Rolls
et al., 1999).
Short-term studies of energy density. Laboratory studies testing
covertly manipulated foods for a few days or less found that
under these conditions, the weight or volume of foods is the
major determinant of satiation and satiety with persons
consuming a relatively constant weight of food regardless
of energy density (Poppitt and Prentice, 1996; Stubbs and
Whybrow, 2004). Results from intervention studies lasting
up to 2 weeks suggest that the lower satiety and satiation for
fat as compared with carbohydrate intake (per unit of
energy) can be explained by the lower energy density of
carbohydrate (Poppitt and Prentice, 1996). A randomized
cross-over study in six men using covert manipulation of a
mixed diet, tested effects of foods of three levels of energy
density (low, 373 kJ/100 g; medium, 549 kJ/100 g; high,
737 kJ/100 g), with virtually identical macronutrient composition
for 14 days each (Stubbs et al., 1998). Participants
S77
European Journal of Clinical Nutrition

S78
Figure 2 Mean (7standard error) change in body weight during
three 14-day periods in which foods of different energy densities
were provided. Reprinted by permission from Macmillan Publishers
Ltd: International Journal of Obesity, (Stubbs et al., 1998) copyright
1998.
compensated for the energy density of diets (that is,
consuming a lower weight of foods with higher energy
density), but this compensation was incomplete resulting
in statistically significant differences in changes in body
weight: 1.20 kg for low, þ 0.02 kg for medium and
þ 0.95 kg for high energy density (Figure 2). In a cross-over
trial in 13 women, a contrast in the energy density of diets
was obtained by offering participants foods that contained
35–40 energy percent as fat (intervention diet) or 20–25
energy percent as fat (control diet) (Kendall et al., 1991). A
statistically significant 2.5 kg greater weight loss was found
for the low-fat diet in the first cross-over period, but in the
second cross-over period only a non-significant 0.4 kg
difference was found. Furthermore, comparison of energy
intakes for the intervention and control diet indicated that
over time, participants compensated better for the higher
energy density of the intervention diet. These observations
suggest that it is uncertain whether the effect on weight
change can be extrapolated to long-term effects on weight.
Also, compensation would probably have been more complete
if participants would have been able to change the type
in addition to the amount of food eaten, because there is
evidence that people compensate by choosing lower energy
density foods after consumption of higher energy density
foods (Poppitt and Prentice, 1996). In addition, when foods
differ in taste, texture and appearance, instead of being
covertly manipulated, as in many of the short-term trials
(Stubbs et al., 1998), physiological consequences of foods can
be paired to these characteristics, resulting in learning effects
and more complete compensation for energy density (Stubbs
and Whybrow, 2004; Yeomans et al., 2005). On the longerterm,
people appear to acquire a greater degree of sensoryspecific
satiety for foods of a higher energy density based on
their post-ingestive effects (Stubbs and Whybrow, 2004).
Compensation for high energy density through learning
European Journal of Clinical Nutrition
Carbohydrate intake and obesity
RM van Dam and JC Seidell
effects may, however, be less effective in an environment
with a wide availability of novel unfamiliar foods (Stubbs
and Whybrow, 2004) or in combination with large portion
sizes (Ello-Martin et al., 2005). The considerations stated
above and the difference in effects on body weight that tend
to be found for long-term as compared with short-term interventions
in general warrant caution with regard to the extrapolation
of short-term effects of energy density in laboratory
studies to long-term consequences for body weight.
Longer-term studies of energy density and weight change. Few
longer-term studies have directly examined the association
between energy density and body weight. Results from crosssectional
studies of energy density of the diet and adiposity
have been inconsistent (Drewnowski et al., 2004) and
prospective observational data are sparse. The association
between energy density and weight change was examined in
a cohort of middle-aged Danish men and women (Iqbal et al.,
2006). In the overall cohort, energy density at baseline was
not substantially associated with 5-year weight gain. In
women, energy density was positively associated with 5-year
weight gain among the obese and inversely associated with
weight gain in normal-weight women, whereas no significant
interaction with baseline weight was observed among
men. In a 1-year trial, 200 overweight and obese participants
were randomized to receive low-energy density soups or
high-energy density snacks (Rolls et al., 2005). All participants
received instructions from dietitians to follow an
exchange-based energy restricted diet. Weight loss was 8.1 kg
for the control group without provided foods, 7.2 kg for the
two-soup per day group, 6.1 kg for the one-soup per day
group and 4.8 kg for the two-snack per day group. Interpretation
of these results is not straightforward, because the
greatest weight loss was achieved in the control group and
because the smaller weight loss for the snack group may have
related to characteristics other than energy density such as
detrimental effects of snacking in a non-hungry state (Rolls
et al., 2005). Further long-term studies of energy density and
body weight are needed.
Weight loss trials comparing diets of different carbohydrate
contents
Below, studies on the effects of weight-loss diets with
variable proportions of carbohydrate on body weight are
reviewed. This includes studies comparing energy-restricted
diets with low and high carbohydrate contents, very-lowcarbohydrate
diets with low-fat energy-restricted diets, highprotein
and high-carbohydrate diets, low-fat and energyrestricted
diets and low-fat and control diets. Although all
these diets may affect body weight through reductions of
energy intake, ‘energy-restricted’ refers to explicit instructions
to participants about energy restriction.
Energy-restricted diets: high versus low carbohydrate. In trials
with strictly controlled energy intakes, macronutrient

composition of the diet did not substantially affect body
weight or fat mass. Golay et al. (1996) compared the effects of
diets containing 4.2 MJ per day that contained 45% (26% fat)
or 15% (53% fat) of energy as carbohydrates in 43 obese
persons. After a 6-week hospital stay during which all foods
were provided, loss of body weight and fat mass did not
differ between the two diets. However, the important
question remains whether a specific macronutrient composition
of the diet can facilitate reduction of energy intake
under realistic ad libitum conditions. Table 1 shows the
characteristics and results of randomized trials that compared
diets that had the same explicit energy intake target,
but differed in carbohydrate content (Baron et al., 1986;
Pascale et al., 1995; Lean et al., 1997; McManus et al., 2001).
Although dietary advice included a specific target for energy
intake, the long duration of these studies and ‘free living’
conditions precluded strict control of the energy intake of
the participants. Therefore, it is of interest what carbohydrate
composition of the diet best facilitated participants to
adhere to the advice to restrict energy intakes and to lose
weight as a result. In the study by Baron et al. (1986), no
difference in weight loss between the high- and the lowcarbohydrate
diets was observed. However, the limited
3-month duration of the intervention probably reduced the
contrast in dietary composition at 12 months. The authors
reported that weight loss differed much more by weight loss
club than by macronutrient composition of the diet. Pascale
et al. (1995) compared an energy-restricted diet with the
recording of fat intake with an energy-restricted diet with
less emphasis on fat. The low-fat dietary advice resulted in a
greater weight loss in persons with type 2 diabetes, but not in
persons with only a family history of diabetes. Lean et al.
(1997) compared two diets with a 23 energy percent
difference in targets for the carbohydrate content of the
diet. No difference in weight loss between the high- and
low-carbohydrate diet was observed after 6 months. In a
subgroup of postmenopausal women who were followed for
12 months, the low-carbohydrate diet was associated with
greater weight loss than the low-fat diet. The ‘moderate-fat’
diet in the trial by McManus et al. (2001) was not particularly
low in carbohydrates, but was more liberal with regard to
unsaturated fat intake than conventional low-fat diets, and
included nuts, avocados and olive oil. The greater weight loss
and substantially lower drop-out for the ‘moderate-fat’ diet
as compared with the low-fat diet suggest that this more
liberal approach may be beneficial for long-term adherence
to diets aimed at weight loss. In summary, in trials where
participants were explicitly instructed to restrict total energy
intake, advice to consume a low-fat, high-carbohydrate
diet did not consistently lead to more or less weight
loss than advice to consume a lower-carbohydrate diet.
Very low-carbohydrate diet versus low-fat, energy-restricted
diet. Table 2 shows the characteristics and results of five
12-month weight loss trials that randomized participants
to very-low carbohydrate ‘Atkins’ type diets or low-fat diets
Carbohydrate intake and obesity
RM van Dam and JC Seidell
(Foster et al., 2003; Samaha et al., 2003; Stern et al., 2004;
Dansinger et al., 2005; McAuley et al., 2005, 2006; Gardner
et al., 2007). In four of the five studies, a substantially larger
weight loss was found after 6 months of the low-carbohydrate
diet as compared with the low-fat diet (Foster et al.,
2003; Samaha et al., 2003; McAuley et al., 2005; Gardner
et al., 2007). This agrees with findings from two other (6
month) trials (Brehm et al., 2003; Yancy et al., 2004), and
results from a recent meta-analysis that reported a pooled
3.3 kg (95% confidence intervals 1.4, 5.3) greater weight loss
for the low-carbohydrate as compared with the low-fat diet
after 6 months (Nordmann et al., 2006). However, during an
additional 6 months, regain of weight diminished the
differences between the diets, resulting in lack of substantial
differences in weight after 12 months (Table 2). This was also
found in the trial with the most intensive intervention that
continued monthly meetings until the 12-month measurements
(Stern et al., 2004). The ketogenic effect of very-low
carbohydrate diets has been suggested to facilitate weight
loss though urinary excretion of ketones or suppression of
appetite by circulating ketones. However, the amount of
energy lost through urinary excretion of ketones is minimal
(Astrup et al., 2004). Furthermore, Foster et al. (2003) did not
observe an association between urinary ketones and weight
loss. The simplicity of the diet, the restriction of the variety
of food choices and possibly a greater satiating effect of
protein seem more plausible explanations for the greater
initial weight loss on very-low carbohydrate diets (Astrup
et al., 2004). In summary, in overweight individuals in the
US and New Zealand, instructions to consume a very-low
carbohydrate diet generally led to greater weight loss during
the first 6 months than instructions to consume low-fat
diets, but due to subsequent regain of weight this may not
result in a greater long-term weight loss.
High carbohydrate versus high protein. Effects of increasing
the proportion of carbohydrates in the diet may depend on
the macronutrient that is replaced: protein or fat. The verylow-carbohydrate
diets discussed in the previous section also
had a higher protein content than the low-fat diets: a 3–7
energy percent higher protein intake was reported at 6
months (Brehm et al., 2003; Samaha et al., 2003; Yancy et al.,
2004; McAuley et al., 2005). Other trials however, have more
specifically attempted to replace carbohydrates with protein.
A Danish group compared the effects of two ad libitum
reduced-fat (30 energy percent) diets: a diet high in
carbohydrates and a diet high in protein (Skov et al., 1999).
During the first 6 months, foods were supplied through a
laboratory shop system, followed by 6 months of consultation
with a dietitian once every 2 weeks. The energy
percentage of protein in the high-protein diet (24.3%
registered in the shop at 6 months; 21.2% reported at 12
months) was substantially higher (12.5% at 6 months, 7.3%
at 12 months) than in the high-carbohydrate diet. Furthermore,
regular measurement of 24-h urinary nitrogen excretion
agreed with these differences in protein intake. After 6
S79
European Journal of Clinical Nutrition

European Journal of Clinical Nutrition
Table 1 Long-term randomized intervention studies of hypocaloric diets: high carbohydrate (low fat) versus lower carbohydrate
Reference/country Participants a
(Baron et al.,
1986) UK
(Pascale et al.,
1995) US
(Lean et al., 1997)
UK
(McManus et al.,
2001) US
M/F. High carb: n ¼ 61,
age 40, BMI 28.5; low
carb: n ¼ 59, age 40, BMI
39.5. An additional n ¼ 8
(high carb) and n ¼ 7
(low carb) were lost to
follow-up and not
included in the analysis
M/F Type 2 diabetes, age
5778, BMI 36.374.7.
Low fat: n ¼ 15; higher
fat: n ¼ 16. An additional
n ¼ 7 (low fat) and n ¼ 6
(higher fat) were lost to
follow-up and not
included in the analysis
M/F family history of
diabetes, age 4378, BMI
35.974.7. Low fat:
n ¼ 16; higher fat: n ¼ 13.
An additional n ¼ 7 (Low
fat) and n ¼ 10 (higher
fat) were lost to followup
and not included in
the analysis
F. High carb: n ¼ 42, age
51714, BMI 32.375.5.
Low carb: n ¼ 40, age
50.1714, BMI
32.875.1. An additional
n ¼ 15 (high carb) and
n ¼ 13 (low carb) lost to
follow-up at 6 months
not included in the
analysis
M/F. Low fat: n ¼ 30, age
44710, BMI 3373.
Moderate fat: n ¼ 31, age
44710, BMI 3475. In
addition n ¼ 21 (high
carb) and n ¼ 19 (low
carb) lost to follow-up at
18 months and included
in ‘last value carried
forward’ intention to
treat analysis
Duration Intervention ‘High carbohydrate’ ‘Low carbohydrate’ Compliance Results
12 months 3 months: participation
in weekly diet club
meetings and written
material. Target:
1000–1200 kcal
per day
12 months 16 weekly group sessions
and meetings at 5, 6, 8
and 10 months. Target:
1000–1500 kcal per day
(depending on baseline
weight)
6 months;
12 months
follow-up
for n ¼ 46
6 months: counseling
by dietitian and written
material. Target:
1200 kcal per day
18 months Weekly group sessions
with dietitian for whole
period. Target: 1200 (F)
or 1500 kcal per day (M)
Fat o30 g per day Carbohydrate o50 g
per day
Target: 20 en% fat.
Recording of both
amount of calories
and fat of foods used
Target: 58 en% carb,
21 en% fat, 21 en%
protein
Target: 60–65 en%
carb, 20 en% fat,
15–20 en% protein
Emphasis on low energy
intake. Recording of
calories of foods used.
Fat o30 en%
encouraged
Target: 35 en% carb, 35
en% fat, 30 en% protein
Target: 45–50 en%
carb, 35 en% fat,
15–20 en% protein.
‘Moderate fat’
Abbreviations: BMI, body mass index (kg/m 2 ); en%, energy percent; F, female; FFQ, food frequency questionnaire; M, male; s.d., standard deviation.
All trials had a parallel design.
None of the trials reported/conducted blinding of the assessors of outcomes or allocation concealment.
a Values are means7s.d.
FFQ: higher fiber intake
(18.4 vs 15.1 g per day),
bread and potato intake
for high-carb vs
low-carb diet
Diet records (12
months): low-fat 26
en% fat, higher fat 34
en% fat
Diet records (12
months): low-fat 26
en% fat, higher fat 34
en% fat
High carb: 1.6 kg
Low carb: 2.3 kg
Difference: 0.7 kg
(95% CI 1.2, 2.6)
Low fat: 5.2 kg
(s.d. 7.3)
Higher fat: 1.0 kg
(s.d. 3.9)
(P ¼ 0.06)
Low fat: 3.1 kg
(s.d. 8.9)
Higher fat: 3.2 kg
(s.d. 7.2)
Not assessed 6 months: High carb:
4.2 kg.
Low carb: 5.4 kg
(P ¼ 0.22). 12 months
(subgroup):
High carb: 3.0 kg Low
carb: 6.5 kg (Po0.05)
Attendance of sessions Low fat: þ 1.1 kg
(20% for low-fat vs 54% Moderate-fat:
for moderate fat, 2.5 kg (P ¼ 0.005)
Po0.01). FFQ at 18
months: low-fat 50 en%
carb, 35% fat; moderatefat
47% carb and 35% fat
S80
Carbohydrate intake and obesity
RM van Dam and JC Seidell

European Journal of Clinical Nutrition
Table 2 Randomized intervention studies comparing very low carbohydrate diet and low-fat energy-restricted diets: results after 6 and 12 months
Reference/
country
(Samaha
et al., 2003;
Stern et al.,
2004) US
(Foster
et al.,
2003) US
(Dansinger
et al., 2005)
US
(McAuley
et al., 2005,
2006)
New
Zealand
Participants Intervention ‘Low-carb’ diet ‘Low-fat’ diet Follow-up Compliance Weight loss (kg)
M/F. BMI X35,
83% diabetes or
metabolic
syndrome. Low
carb: n ¼ 64. Lowfat:
n ¼ 68; lost to
follow-up: n ¼ 14 at
6 months, n ¼ 6at
12 months
M/F. Obese (mean
BMI 34). Low carb:
n ¼ 33. Low-fat:
n ¼ 30; lost to
follow-up: n ¼ 21 at
6 months, n ¼ 26 at
12 months
M/F. BMI 27–42
(mean 35) with
metabolic risk
factors. Low carb:
n ¼ 40. Low-fat:
n ¼ 40; lost to
follow-up: n ¼ 19
for low carb and
n ¼ 20 for low-fat
F. Insulin resistant,
BMI427. Low carb:
n ¼ 31. Low-fat:
n ¼ 32; lost to
follow-up: n ¼ 6at
6 months, n ¼ 7at
12 months
Group counseling:
weekly sessions for
4 weeks and 11
monthly session;
written materials
One consultation
with dietitian; a
book/manual
Advice during four
group sessions in
first 2 months;
written materials
and diet book
Weekly counseling
for 16 weeks;
written materials
Carbohydrate
intake o30 g per
day
Carbohydrate
o20 g per day
for 2 weeks,
followed by a
gradual increase
Carbohydrate
o20 g per day with
gradual increase to
50 g per day
Carbohydrate
o20 g per day
for 2 weeks with
gradual increase
Fat o30 energy%
and 500 kcal per
day energy deficit
Fat B25%,
protein B15%,
carbohydrate
B60% of energy.
Energy restricted
Vegetarian ‘Ornish’
diet with 10
energy% fat
Conventional
high-fiber, low-fat,
reduced sugar
diet. No explicit
energy-restriction
6 months 24-h recall (energy
%). Low carb: C 37,
F 41, P 22 w .
Low-fat: C 51,
F 33, P 16
12 months 24-h recall
(g per day).
Carbohydrate: low
carb 120; low-fat
230
Low carb a
Low fat a
Difference
5.8 (8.6) 1.9 (4.2) P ¼ 0.002 b
5.1 (8.7) 3.1 (8.4) P ¼ 0.20 c
6 months Testing of urinary
ketone
concentrations.
Significant
difference between
groups up to 12
weeks
6.9 (6.4) 3.1 (5.5) P ¼ 0.02 d
12 months 4.3 (6.6) 2.5 (6.2) P ¼ 0.26 d
6 months Diet records
(g per day):
low carb: C 190 g,
F 81 g; low-fat: C
237 g, F 55 g
3.2 (4.9) 3.6 (6.7) P ¼ 0.76 d
12 months Diet records (g/d):
low carb: C 190 g,
F 81 g; low-fat: C
218 g, F 64 g
6 months Diet records
(energy %): low
carb: C 26, F 47,
P 24; low fat: C 45,
F 28, P 21
12 months Diet records
(energy %): low
carb: C 33, F 41,
P 21; Low fat: C 45,
F 29, P 22
2.1 (4.8) 3.3 (7.3) P ¼ 0.40 d
7.1 4.7 Po0.05 c
5.4 4.4 P40.05 c
S81
Carbohydrate intake and obesity
RM van Dam and JC Seidell

S82
Table 2 Continued
Participants Intervention ‘Low-carb’ diet ‘Low-fat’ diet Follow-up Compliance Weight loss (kg)
Reference/
country
Difference
Low fat a
Low carb a
B 5.8 B 3.1 Po0.05 d
6 months 24-h recalls (energy
%): low carb: C 30,
F 47, P 22; Low fat:
C 48, F 31, P 18
‘LEARN’ program:
55–60%
carbohydrate and
o10 energy%
saturated fat,
caloric restriction,
exercise
Carbohydrate
p20 g per day for
B2–3 months
followed by p50 g
per day
Weekly counseling
for 2 months; diet
books
F. Premenopausal
(25–50 years),
BMI 27–40.
Low carb: n ¼ 77.
Low fat n ¼ 79;
lost to follow-up:
n ¼ 21 at 6
months, n ¼ 27
at 12 months
(Gardner
et al., 2007)
US
European Journal of Clinical Nutrition
4.7 ( 6.3, 3.1) 2.2 ( 3.6, 0.8) P40.05 d
Carbohydrate intake and obesity
RM van Dam and JC Seidell
12 months 24-h recalls (energy
%): low carb: C 35,
F 44, P 21; low-fat:
C 47, F 33, P 19
Abbreviations: BMI, body mass index (kg m 2 ); C, carbohydrates; F, fat; F, female; M, male; P, protein.
All trials had a parallel design, allocation concealment was performed for all studies, and blinding of outcome assessors was not reported.
a
Values are means (s.d. or 95% CI).
b
Intention to treat: last value carried forward.
c
Random coefficient analysis.
d
Intention to treat: baseline value carried forward.
months, weight loss was 9.4 kg in the high-protein group
and 5.9 kg in the high-carbohydrate group (difference 3.5 kg,
P ¼ 0.008; Skov et al., 1999). After 12 months, weight loss was
6.2 kg for the high-protein group and 4.3 kg for the highcarbohydrate
group (difference 1.9 kg, P40.05; Due et al.,
2004). Thus, despite a remaining substantial difference in
protein intake, some weight was regained and it was unclear
whether a greater long-term weight loss could be maintained
for the high-protein protein group. The reduction in waist
circumference remained statistically significantly greater
after 12 months for the high-protein as compared with the
high-carbohydrate group (Due et al., 2004). An Australian
intervention study also compared effects of a high-protein
with a high-carbohydrate diet (McAuley et al., 2005). At 6
months, the higher protein diet was associated with a 5
energy percent higher protein intake (26 versus 21%) and a
statistically significant 2.2 kg greater weight loss (6.9 versus
4.7 kg). Regain of weight in the subsequent 6 months was
0.5 kg for the high-carbohydrate and 0.9 kg for the highprotein
group (McAuley et al., 2006). After 12 months, the
energy percent of protein was identical for the two groups,
whereas a non-statistically significant 2.3 kg lower body
weight remained for the high-protein as compared with the
high-carbohydrate group. In two intervention studies by
Brinkworth and co-workers, a high-protein diet and a highcarbohydrate
diet resulted in similar weight loss after 68
weeks, but the intervention only lasted for 16 weeks and the
urinary urea/creatinine ratio indicated no difference in
protein content after the intervention (Brinkworth et al.,
2004a, b). In summary, two trials suggest that exchanging
protein for carbohydrates can facilitate weight loss over 6
months, but more research is needed to clarify whether this
beneficial effect can be maintained after 12 or more months.
Low-fat (high carbohydrate) versus energy-restricted diet. Several
long-term intervention studies compared dietary advice
focused on reducing fat intake with dietary advice focused
on restriction of total energy intake. These trials reported
worse (Harvey-Berino, 1998), similar (Jeffery et al., 1995;
Dansinger et al., 2005) and better (Toubro and Astrup, 1997)
effects on body weight for the low-fat approach, as compared
with the energy-restricted approach. These differences in the
results probably reflect differences in the energy-restricted
program (for example, a complicated color-coded system
(Toubro and Astrup, 1997) versus a simpler method of calorie
counting (Dansinger et al., 2005)) rather than effects of
differences in macronutrient composition of the diet.
Intervention studies of carbohydrate intake not primarily aimed at
weight loss
Effect of dietary advice to consume a low-fat diet on body
weight. The Women’s Health Initiative Dietary Modification
Trial tested the effect of advice to decrease fat intake and
increase consumption of fruit, vegetables and grains on
body weight for a mean follow-up of 7.5 years in 48 835

postmenopausal US women (Howard et al., 2006). The
intervention included 18 group sessions during the first 12
months, followed by four group sessions per year for the
duration of the trial supplemented with individual sessions.
The control group only received dietary educational materials.
According to self-reported data from a food frequency
questionnaire, the percentage of energy from fat decreased
by 8.8%, the percentage of energy from carbohydrates
increased by 8.2%, the number of servings of fruits and
vegetables increased by 1.4 servings per day and fiber intake
increased by 2.2 g per day in the intervention group, whereas
no substantial changes were reported by the control group.
Body weight decreased 1.9 kg after 1 year and 0.4 kg after an
average of 7.5 years for the intervention as compared with
the control group (both P-value 0.001). Trends in body
weight for the intervention and control group were similar:
an increase in women who were not overweight, little
change in those who were moderately overweight and a
decrease in those who were obese before the study (Figure 3).
Weight loss was not an aim of this study and data on dietary
change only relied on self-reports. However, the findings do
not support a substantial long-term effect of an educational
intervention aimed at reducing the proportion of fat (or
increasing the proportion of carbohydrates) in the diet on
body weight in US women. Even the small effect on body
weight may reflect changes other than fat intake as a result of
the more intensive dietary advice in the intervention as
compared with the control group.
Carbohydrate intake and obesity
RM van Dam and JC Seidell
The results of earlier smaller randomized interventions of
at least 12 months duration that tested advice to consume
low-fat diets and were not aimed at weight loss showed a
similar lack of substantial effects on body weight (Willett,
2002). A meta-analysis of intervention studies suggested a
more substantial 3.2 kg greater weight loss as a result of an ad
libitum low-fat diet as compared with the control group.
(Astrup et al., 2004) However, interpretation of the results is
limited by the inclusion of shorter term (duration was 2–12
months) and non-randomized studies. Moreover, in most
studies, only the low-fat intervention group received intensive
dietary advice, making it unclear whether effects
were due to changes in fat intake per se or other behavioral
changes related to the greater awareness of diet. In summary,
randomized intervention studies do not consistently show
that educational efforts aimed at reducing the percentage of
energy intake from fat (or increase the percentage of energy
from carbohydrates) without additional efforts to reduce
energy intakes have important long-term effects on body
weight.
Effect of provision of reduced fat foods on body weight. In the
double-blinded multicenter ‘First Study’ of National Diet-
Heart Study, foods of variable fat contents were provided to
approximately 1000 middle-aged US men (Anonymous,
1968). The participants were provided low-fat foods resulting
in 29.7% of energy from fat (based on diet records), high-fat/
high-polyunsaturated fat foods resulting in 34.4% of energy
Figure 3 Effects of advice to consume a low-fat diet high in fruits and vegetables on body weight: differences (from baseline) in body weight by
intervention group and BMI at screening in the Women’s Health Initiative. Error bars indicate 95% confidence intervals. *Pp0.05 and 40.01
w Pp0.01 and 40.001 z Pp0.001 for the difference between the intervention and control group. Reprinted from Howard et al. (2006) copyright
(2006), American Medical Association. All rights reserved. BMI, body mass index.
S83
European Journal of Clinical Nutrition

S84
from fat or high-fat/low-polyunsaturated fat foods resulting
in 34.9% of energy from fat. Over 44 weeks, men in all
groups lost weight (2.3, 1.4 and 1.8 kg respectively), with a
0.5–0.9 kg greater weight loss for the low-fat group. In a
randomized trial of 241 Dutch non-obese individuals with
no intention to lose weight, free access was provided to B45
different foods either in reduced fat or full fat version,
covering between 30 and 40% of energy intake. The group
that consumed the reduced-fat products had on average a 7
energy percent lower fat intake and a 0.7 kg (95% confidence
interval 0.1, 1.4) decrease in body fat over 6 months as
compared with those that consumed the full-fat products
(Westerterp et al., 1996; Weststrate et al., 1998). In the
randomized CARMEN trial (Carbohydrate Ratio Management
in European National Diets), diets high in carbohydrates
were compared with a control diet that had a
macronutrient composition that was typical for each
country (Saris et al., 2000). The intervention was ad libitum
and consisted of providing foods through a laboratory shop
system in 398 men and women with a BMI between 26 and
35 kg/m 2 . Allocation concealment and blinding procedures
were not described and no intention-to-treat analysis was
conducted. According to diet records intake of fat, protein,
and carbohydrates changed significantly by 8.7, þ 2.7 and
þ 6.3 energy percent for the low-fat high-starch group
relative to the control group. Weight loss over 6 months
was 2.6 kg for the low-fat high-starch group as compared
with the control group. This weight loss was substantially
larger than for the National Diet-Heart-Study, possibly due to
the greater contrast in fat intake, the higher protein intake
for the high-starch diet relative to the control diet or to all
the CARMEN participants being overweight. Furthermore, in
the CARMEN study participants were not blinded and the
high-starch diet required a substantial change and reconsideration
of dietary habits, whereas the control diet was
similar to the usual dietary intake.
In all trials discussed above, full-fat foods were replaced by
reduced-fat versions of the foods. One could argue that the
observed small effects of this product change on body
weight may be worthwhile on a population level. However,
exchange of reduced-fat foods for full-fat versions may
reflect the effect of reducing the amount of energy per
portion without substantially changing the participants’
perception of the satiety value of the food. Thus, findings
do not necessarily apply to exchanging carbohydrates for fat
and replacement of full-calorie carbohydrate foods with
reduced-calorie versions may have similar effects.
Observational studies of carbohydrate intake and body weight
Cohort studies of fat intake and long-term weight change. As
reported in section 1.2, observational studies of dietary
intakes and body weight are prone to various biases. People
tend to hold particularly strong ideas about the effects of the
proportion of fat or carbohydrate on body weight favoring
for example commonly recommended low-fat diets or
European Journal of Clinical Nutrition
Carbohydrate intake and obesity
RM van Dam and JC Seidell
popular low-carbohydrate diets. A person’s ability to adhere
to the macronutrient composition of such a diet is likely to
be associated with a person’s ability to control total energy
intake. Therefore, an association between the proportion of
carbohydrates in the diet with lower weight or less weight
gain may well be confounded by the ability to control energy
intake. Other methodological issues include selective underreporting
of fat intake that is related to weight and residual
confounding by other dietary factors (Seidell, 1998). Results
from prospective cohort studies that study the proportion of
energy from fat in the diet (and thus indirectly the
proportion of energy from carbohydrate in the content)
have been inconsistent (Seidell, 1998). In addition to the
aforementioned methodological limitations, differences in
results may relate to differences in genetic susceptibility. In
two cross-sectional studies, the association between fat
intakes and adiposity differed by peroxisome proliferatoractivated
receptor-a Pro12Ala genotype (Memisoglu et al.,
2003; Franks et al., 2004). This variant also modified the
weight loss as result of a lifestyle intervention (Lindi et al.,
2002), but not the response to an energy restricted high or
low-fat diet (Sorensen et al., 2006).
Weight control registry. The National Weight Control Registry
is a registry of persons in the US that maintained a weight
loss of 13.6 kg or more for at least 1 year (Phelan et al., 2006).
The majority of these successful weight losers consumed a
low-fat, high-carbohydrate diet, but the average percentage
of energy from carbohydrates had decreased from 56% in
1995 to 49% in 2003, and the proportion of persons on a
low-carbohydrate diet (o90 g/d carbohydrate) increased
from 5.9 to 17.1% during that period (Phelan et al., 2006).
These results indicate that the diet associated with weight
loss at a certain point in time is to some degree dependent on
the types of diet that are generally believed to be beneficial
for weight. Still, these registry data support the findings from
intervention studies that substantial weight loss is possible
over a wide range of macronutrient compositions of the diet.
Conclusions on the carbohydrate content of the diet
Findings from the dietary intervention studies discussed
above suggest that similar weight loss can be achieved over
the course of a year with diets of substantially different
carbohydrate contents. Positive energy balance and weight
gain can be achieved with a wide range of energy percentage
of carbohydrates in the diet; thus, neither a diet with a high
nor a low energy percentage of carbohydrates will necessarily
protect a person from weight gain on the long term.
In trials in which dietary advice for weight loss was
provided in overweight persons in the US and New Zealand,
weight loss was greater after 6 months for low-carbohydrate
diets as compared with low-fat diets, although this difference
greatly attenuated after longer follow-up (Nordmann et al.,
2006; Gardner et al., 2007). Interestingly, in the European
CARMEN trial in which macronutrient intake was manipulated
by providing foods through a laboratory shop system,

weight loss was greater after 6 months with a low-fat highcarbohydrate
diet as compared with a higher fat control diet
(Saris et al., 2000). There are several possible explanations for
this difference in results. First, in the trials with dietary
advice, protein intake was substantially higher for the lowcarbohydrate
relative to the low-fat diet, whereas the
opposite was true for the CARMEN trial. The promising
hypothesis that a higher protein content of the diet may
contribute to weight management requires further research.
Second, the nature of the foods included in the low-fat diet
may have differed: in the CARMEN trial the low-fat diet
included many reduced-calorie versions of otherwise similar
full-fat foods, whereas diverse self-chosen high-carbohydrate
foods may have been included in the low-fat diets in the
trials with dietary advice. Third, the characteristics of the
interventions unrelated to physiological effects of macronutrient
composition may have played an important role. In
the trials with dietary advice, the low-carbohydrate diet was
probably the most novel, simple and restrictive diet, whereas
in the CARMEN trial the lower carbohydrate diet was a
control diet that required little change in food choice.
Long-term dietary and lifestyle interventions (X2 years)
show that consumption of a relatively high-carbohydrate
diet (B55% of energy) that includes high amounts of fiberrich
foods can be compatible with clinically relevant weight
loss (Tuomilehto et al., 2001; Knowler et al., 2002; Esposito
et al., 2004; Mayer-Davis et al., 2004). A Mediterranean-style
moderate-fat diet with a lower carbohydrate content (B50% of
energy) was also associated with substantial weight loss after
18 months (McManus et al., 2001). Taken together, the
available data do not provide strong evidence that either
increasing or decreasing the energy percentage of carbohydrate
in the diet by itself has an important effect on body weight.
Free sugars
Introduction
Free sugars are defined as added sugars plus concentrated
sugars in honey, syrups and fruit juices. Because glucose chains
in starches can be rapidly broken down in the gastrointestinal
tract, many starchy foods induce a more rapid increase of
blood glucose concentrations than many foods high in free
sugars (Foster-Powell et al., 2002). However, other characteristics
of free sugars can be relevant for energy balance. Foods
high in free sugars have been proposed to contribute to weight
gain as compared with starchy foods because of lack of dietary
fiber and high energy density (Poppitt and Prentice, 1996),
higher palatability because they are sweeter (Raben et al.,
1997), unique effects of fructose (Elliott et al., 2002) and
because these are often consumed in the form of high-caloric
liquids instead of solid foods (Mattes, 1996).
Micronutrient dilution
‘Micronutrient dilution’ refers to a reduction of the micronutrient
content of the diet as a result of the displacement of
Carbohydrate intake and obesity
RM van Dam and JC Seidell
micronutrient-rich foods by foods high in free sugars. Studies
in children, adolescents and adults have reported that a high
energy percentage of the diet as free sugars is associated with
lower intakes of various micronutrients (Alexy et al., 2003;
Charlton et al., 2005; Kranz et al., 2005). In addition, a high
free-sugar content of the diet has been associated with lower
intakes of fiber (Kranz et al., 2005) and fruit and vegetables
(Charlton et al., 2005). Regular consumption of foods high in
free sugars does not have to be associated with micronutrient
deficiencies (Ruxton, 2003). However, given that the energy
intake that is compatible with avoiding weight gain in
modern societies with little occupational physical activity is
limited, it should be considered that a high intake of energy
as free sugars will generally make it more difficult to achieve
optimal intakes of micronutrients, phytochemicals, fiber and
fruit and vegetables.
Postprandial insulin secretion and body weight
The effect of postprandial insulin secretion on energy
balance is controversial. On the one hand, it has been
postulated that a high postprandial insulin response may
lead to weight gain through various mechanisms. First, a
high postprandial insulin response may rapidly lower blood
glucose and free fatty acid concentrations, which might in
turn induce the secretion of counter-regulatory hormones.
These hormones may stimulate hunger and energy intake
and may also lower resting energy expenditure through a
proteolytic effect that reduces lean body mass over time
(Ludwig, 2002; McMillan-Price and Brand-Miller, 2006).
Second, it has been suggested that high postprandial insulin
responses may reduce fat oxidation and increase fat synthesis
and storage (McMillan-Price and Brand-Miller, 2006). However,
hyperinsulinemia has been associated with reduced
weight gain in several longitudinal studies (Valdez et al.,
1994; Hoag et al., 1995; Schwartz et al., 1995), and the
quantitative significance of the effect of higher insulin levels
on de novo fatty acid synthesis in adipose tissue in humans
has been questioned (Wolever, 2006).
On the other hand, it has been postulated that low
postprandial insulin responses may lead to reduced satiety
(Elliott et al., 2002). Lack of insulin response can lead to
reduced leptin production by adipose tissue and both leptin
and insulin can play a role in inducing satiety (de Graaf et al.,
2004; Schwartz and Porte, 2005); however, in short-term
experiments, the effect of insulin infusion on appetite and
food intake has not been consistent (Wolever, 2006). Taken
together, it has not been established whether variation in
postprandial insulin responses has substantial effects on
body weight regulation in humans.
Fructose intake, satiety and body weight
In contrast to glucose, fructose ingestion elicits little
response in blood glucose and insulin concentrations. It
has been suggested that high intake of fructose intake may
S85
European Journal of Clinical Nutrition

S86
lead to excess energy intake because of this lack of insulin
response as well as failure to suppress secretion of the
‘hunger hormone’ ghrelin (Elliott et al., 2002; Teff et al.,
2004). The effect of fructose intake on body weight has not
been examined in randomized trials. A diet supplemented
with 50–60 g fructose resulted in weight gain in 14 persons
with type 2 diabetes during 23 weeks, but this study did not
include a control group (Anderson et al., 1989).
Studies of mostly solid sugary foods
A cross-over study in 20 non-overweight women (mean BMI
23 kg/m 2 ) in which all foods were provided compared 14
days of ad libitum consumption of a high-starch/high-fiber
diet with a high-sucrose diet (Raben et al., 1997). The highstarch/high-fiber
diet resulted in a 0.7 kg reduction in body
weight, whereas the sucrose diet resulted in a non-significant
0.2 kg increase in weight (Po0.05 for difference between
diets). Possibly, this difference in weight change is related to
the participants’ higher palatability ratings for the highsucrose
diet or to the sugar-sweetened beverages in the highsucrose
diet. In the 6-month randomized CARMEN trial, a
diet higher in mono-and disaccharides was compared with a
diet higher in starch in men and women with BMI 26–35
kg/m 2 (Saris et al., 2000). The intervention was ad libitum and
consisted of providing foods through a laboratory shop
system. The diet high in mono-/disaccharides was associated
with a non-significant 0.9 kg smaller weight loss than the
high-starch diet. The same intervention in persons with the
metabolic syndrome resulted in a 4 kg greater reduction in
weight for the high-starch as compared with the high monoand
disaccharide diet (Poppitt et al., 2002). These two diets
did not only differ in the solid foods consumed, but also in
the types of beverages used: beverages high in free sugars in
the high-mono-and disaccharide diet versus artificially
sweetened beverages in the high-starch diet (Poppitt et al.,
2002). Middle-aged male office workers who were asked to
cut out sucrose from their diet and replace it with other
foods, reduced their sucrose intake from 85 to 12 g per day
according to diet records and were reported to have lost
weight after 22 weeks (Mann et al., 1970). Similar interventions
showed a non-significant tendency for greater weight
loss as compared with the control group in hypertriglyceridemic
individuals (Smith et al., 1996), whereas no difference
in weight loss was found in non-obese women (Gatenby
et al., 1997).
Sugar-sweetened beverages and body weight
Intervention studies of sugar-sweetened beverages. Several trials
have specifically examined the effects of sweetened beverages
on body weight (Table 3). Energy consumed as
liquids may induce less satiety as compared with the same
foods in a solid form because of the rapid transit of liquids
through the stomach and intestines that may lead to reduced
stimulation of satiety signals, differences in the regulation of
European Journal of Clinical Nutrition
Carbohydrate intake and obesity
RM van Dam and JC Seidell
thirst and hunger, and lower cognitively perceived energy
content (Mattes, 1996; DiMeglio and Mattes, 2000). Reported
energy intake in intervention studies of sugarsweetened
beverages suggested a lack of compensation for
the energy provided through these liquids by reduced
subsequent energy intake (Tordoff and Alleva, 1990; DiMeglio
and Mattes, 2000). Trials of sugar-sweetened beverage
consumption and body weight varied from blinded interventions
of several weeks (Tordoff and Alleva, 1990; Raben
et al., 2002) to an education program on carbonated
beverages of 1 year (James et al., 2004; Table 3). Findings in
all trials were consistent with a detrimental effect of
consumption of sugar-sweetened beverages on body weight.
Both trials in which participants were blinded with regard to
both the aim of the study and the sweeteners in the
beverages they received found a statistically significant effect
the sugar-sweetened intervention on body weight (Tordoff
and Alleva, 1990; Raben et al., 2002). Figure 4 shows the
results of one of these blinded trials conducted by Raben
et al. (2002). A 1-year dietary educational program at schools
mainly aimed at reducing use of carbonated beverages was
associated with a reduced incidence of overweight (James
et al., 2004). However, given the very limited effect on sugarsweetened
soft drink consumption, it is uncertain whether
the effect of the program on obesity are due to reduced
consumption of these beverages, to the apparently greater
reduction in consumption of artificially sweetened soft
drinks or to other behavioral changes. In a 25-week
intervention that mainly consisted of the free delivery of
non-caloric beverages, consumption of non-caloric beverages
instead of sugar-sweetened beverages was associated
with weight loss among participants with a higher baseline
BMI (Ebbeling et al., 2006). The result of this trial are also of
interest because the changed availability of beverages
resulted in a marked decrease in consumption of sugarsweetened
beverages over 25 weeks (reported 82% decrease
in the intervention group) (Ebbeling et al., 2006). Drewnowski
and Bellisle (2007) have pointed out that liquid meal
replacement shakes containing amounts of sugars that are
similar to the amount in sugar-sweetened beverages are
commonly used for weight loss treatment. However, in
contrast to sugar-sweetened soft drinks, liquid meal replacement
shakes contain substantial amounts of protein and
fiber and are used instead of meals rather than in addition to
meals (Drewnowski and Bellisle, 2007).
Cohort studies of sugar-sweetened beverages and long-term weight
change. In a real-life setting, as compared with the
described trials, additional factors may link the consumption
of sugar-sweetened beverages to excess energy intake. These
factors include huge serving sizes, free refills, massive
advertising campaigns for soft drinks and the ubiquitous
availability of soft drinks including vending machines at
schools (Bray et al., 2004a). Recently, the literature on intake
of sugar-sweetened beverages and weight gain has been
systematically reviewed and 10 prospective cohort studies

European Journal of Clinical Nutrition
Table 3 Intervention studies of sugar-sweetened soft drinks and body weight
Reference/country Participants a
(Tordoff and Alleva,
1990) US
(DiMeglio and
Mattes, 2000) US
(Raben et al., 2002)
Denmark
(James et al., 2004)
UK
(Ebbeling et al.,
2006) US
n ¼ 30. 9 F: age
28.772.7, BMI
25.471.4; 21 M:
age 22.970.8, BMI
25.170.5; 11 others
dropped out and
were not included
in the analysis
n ¼ 15 M/F: age
22.872.7, BMI
21.972.2. All
participants included
in the analysis
n ¼ 41 overweight
M/F. Sucrose group:
n ¼ 21, age
33.372.0, BMI
28.070.5. Sweetener
group: n ¼ 20, age
37.172.2, BMI 27.6.
One drop-out was
not included in the
analysis
n ¼ 644 M/F, children
age 7–11 year from
six schools. 20% were
overweight or obese
n ¼ 103 M/F, age
13–18 year. No loss
to follow-up
Trial design b
Cross-over,
participants were
blinded
Cross-over 2 4 week and
4-week washout
Randomized parallel;
participants were
blinded
Cluster randomized
(29 clusters ¼ classes)
Randomized, parallel,
allocation
concealment
Duration Intervention Control Compliance Results
3 3 weeks 1135 g per day
of supplied
HFCS-sweetened
soft drinks
1880 kJ per day of
sugar-sweetened soft
drinks (‘liquid’)
10 week Sucrose supplements
(B70% as drinks)
1 year Educational nutrition
program mainly
discouraging
carbonated beverage
use
25 weeks Home delivery of
non-caloric
beverages (four
servings per day)
and instructions
Two control
interventions: 1135 g
per day aspartamesweetened
drinks or
no specific
instructions
1880 kJ per day of
jelly beans (‘solid’)
Artificial sweetener
supplements
None Drink diaries
completed by 36%:
0.7 glass per day
reduction of mainly
diet carbonated
beverages
Instruction to
continue usual
beverage
consumption
Abbreviations: BMI, body mass index (kg/m 2 ); F, female; HFCS, high fructose corn syrup; M, male.
a Values are means7s.d.
b None of the trials reported on blinding of the assessors of outcomes, except for Ebbeling and co-workers allocation concealment was not reported.
Diet records HFCS drinks vs no
instructions:
F þ 0.97 kg,
M þ 0.72 kg. HFCS
drinks vs aspartame
drinks: F þ 0.72 kg,
M þ 0.99 kg
(P overall o0.01 for
both F and M)
Diet records Po0.05 for weight
change during liquid
period ( þ 0.5 kg),
but not significantly
different from change
in the solid period
( þ 0.3 kg)
Diet records 2.6 (95% CI 1.3–3.8,
Po0.001) kg increase
in weight and 1.6
(95% CI 0.4–2.8,
Po0.01) increase in
fat mass for the
sucrose as compared
with the artificial
sweetener group
24-h recalls: 82%
reduction of sugarsweetened
beverages
in intervention group
No substantial
decrease in mean
BMI (0.1 kg/m 2 , 95%
CI 0.1–0.3), but a
7.7% (95% CI
2.2–13.1) decrease
in prevalence of
overweight/obesity
as compared with
control
BMI: 0.14 (SE 0.21,
P40.05) for the total
group and 0.75
(SE 0.34, P ¼ 0.03)
for the upper
baseline-BMI tertile
as compared with
the control group
S87
Carbohydrate intake and obesity
RM van Dam and JC Seidell

S88
Figure 4 Mean (7standard error) changes in body weight and fat
mass during an intervention in which overweight subjects consumed
supplements containing either sucrose (K; n ¼ 21) or artificial
sweeteners (n; n ¼ 20) daily for 10 weeks. In the sucrose group,
sucrose was mostly consumed as beverages (B70%). At specific
time points for changes in body weight and fat mass, there were
significant differences between the sucrose and sweetener groups:
*Po0.05, **Po0.001 and ***Po0.0001. Reprinted from Raben et al.
(2002) copyright 2002, with permission from the American Society
for Nutrition.
were identified (Malik et al., 2006). In most of these studies,
increases in sugar-sweetened soft drink consumption were
associated with weight gain. The lack of significant association
in some of the studies may have been due to small
sample sizes, short periods of follow-up and limited variation
in soft drink consumption. In the largest study in children
(n ¼ 11 654, age 9–14 years), an increase of two or more
servings of sugar-sweetened soft drinks per day over a year
was associated with an increase in BMI in both boys (0.14
kg/m 2 , P ¼ 0.01) and girls (0.10 kg/m 2 , Po0.05; Berkey et al.,
2004). In the largest study of adults, 51 603 women were
followed for two 4-year periods (Schulze et al., 2004). Women
who increased their sugar-sweetened soft drink consumption
from less than weekly to daily had the largest weight gain
(4.7 kg for the first and 4.2 kg for the second period), whereas
women who reduced their consumption from daily to less
than weekly had the smallest weight gain (1.3 kg for the first
and 0.2 kg for the second period; P-value differences between
groups 0.001). The finding of an association between sugarsweetened
beverage consumption and weight gain has not
been limited to US populations (Bes-Rastrollo et al., 2006b).
Although these prospective studies considered confounding
in detail, remaining confounding by other dietary and
lifestyle factors cannot be excluded. However, the results of
cohort studies are consistent with the findings from intervention
studies and suggest that consumption of sugarsweetened
soft drinks is relevant for long-term weight
management. It is plausible that findings for sugarsweetened
soft drinks also apply to juices high in free sugars
European Journal of Clinical Nutrition
Carbohydrate intake and obesity
RM van Dam and JC Seidell
such as apple juice, but effects of juices on body weight have
not been tested experimentally and results from observational
studies have been mixed (Malik et al., 2006).
High-fructose corn syrup versus sucrose-sweetened beverages. It
has been suggested that the use of high-fructose corn syrup
to sweeten soft drinks has contributed to the increased
prevalence of obesity in the US (Bray et al., 2004a). The most
commonly used type of high-fructose corn syrup contains
55% fructose and 45% glucose, which is similar to sucrose.
Theoretically, the separated glucose and fructose (as in highfructose
corn syrup) may have different effects than
conjoined molecule (as in sucrose before it is broken down
in the intestine) related to slightly higher sweetness and
higher osmolarity (Bray et al., 2004a). It should be noted,
however, that the Danish trial found effects of sucrosesweetened
foods on weight gain (Raben et al., 2002).
Therefore, reduction of the consumption of both sucrosesweetened
and high-fructose corn syrup sweetened soft
drinks is warranted.
Conclusion on free sugars and sugar-sweetened beverages
Short-term experiments suggest that satiety and satiation
may be lower for carbohydrates consumed as beverages as
compared with carbohydrates consumed as solid foods.
Although long-term randomized controlled trials of sugarsweetened
beverages are lacking, evidence from short-term
blinded randomized controlled trials, medium-term nonblinded
randomized trials, and long-term prospective cohort
studies indicates that reduction of consumption of sugarsweetened
beverages is beneficial for weight management.
Some data also support that reduction of solid foods high in
free sugars can contribute to weight loss, but findings have
been less consistent than for sugar-sweetened beverages. The
high energy density and low content of micronutrients and
fiber of many foods high in free sugars such as sugary candy,
table sugar, cookies and cakes should be considered.
Dietary GI and glycemic load
Introduction
The GI of a food quantifies the strength of the blood glucose
response after consumption of a fixed amount of available
carbohydrates from that food as compared with the response
after consumption of the same amount of available carbohydrate
from a reference food (glucose or white bread) (Foster-
Powell et al., 2002). The glycemic load (GL) is the product of
the GI of a food and the amount of carbohydrate that it
contains (Foster-Powell et al., 2002). The GL was conceived
to reflect the total blood glucose-raising potential of the diet.
The GL of a diet can thus be reduced by reducing the GI of
the diet and/or by reducing the amount of carbohydrate
consumed. It is important to clearly distinguish between the
GI and the GL. First, the GL of a diet is strongly influenced by

its carbohydrate content, whereas a low-GI diet can be either
high or low in carbohydrates. Second, physiological effects
can be different for low-GI diets as compared with low-GL
diets (Wolever et al., 1995). For example, in a study that
evaluated four meals with different combinations of the GI
(low or high) and carbohydrate content (low or high), the
greatest contrast in effects on free fatty acid levels was found
between the low-GI/high-carbohydrate meal and the low-GI/
low-carbohydrate (that is, lowest GL) meal (Wolever et al.,
1995).
Mechanistic studies of the GI and GL and energy balance
It has been postulated that consumption of high-GI foods
can lead to excess energy intake through rapid rises and
rapid subsequent declines in blood glucose concentrations
and associated increases in levels of counter-regulatory
hormones (Ludwig, 2002). This is supported by the observation
that declines in blood glucose concentrations acutely
increased voluntary food intake in time-blinded volunteers
(Melanson et al., 1999). Several studies have reported
beneficial effects of low-GI meals on self-reported hunger
and short-term energy intake (Raben, 2002; Pereira et al.,
2004), but overall findings have been mixed (Raben, 2002).
In addition, a greater reduction in resting energy expenditure
has been observed in 39 overweight adults after a 10%
weight reduction as result of a high-GL hypocaloric diet than
after the same weight reduction as a result of a hypocaloric
diet with a lower GL (Pereira et al., 2004). No differences in
lean or fat mass were detected and the authors suggested that
low postprandial circulating concentrations of metabolic
substrates may have caused this difference possibly through
neuroendocrinological adaptations to conserve energy
(Pereira et al., 2004). It cannot be fully excluded that other
differences between the diets including substantial differences
in protein, fiber and micronutrients may have
contributed to the observed effects.
It is plausible that the effects on energy balance differ for
different types of low-GI foods. For example, it has been
reported that having some low-GI foods (for example,
spaghetti) for breakfast reduced the glycemic response to a
subsequent lunch as compared with high-GI white wheat
bread, whereas having other low-GI foods for breakfast (for
example, white wheat bread with added vinegar) did not
(Liljeberg et al., 1999). The authors suggested that the
‘second-meal effect’ for the former foods might have resulted
from the prolonged elevation of postprandial glucose
concentrations and avoidance of a between-meal fasting.
Furthermore, the effect of foods with a low-GI as a result
of slow absorption of starches in the intestines may be
different from that of foods with a low GI as a result of high
fructose content. For example, slowly absorbed starches may
reach the lower part of the ileum and stimulate the secretion
of gut satiety hormones such as glucagon-like peptide-1
(McMillan-Price and Brand-Miller, 2006).
Carbohydrate intake and obesity
RM van Dam and JC Seidell
Cohort studies of the GI and GL
The association between the dietary GI or GL and changes in
body weight has been examined in two cohort studies, and
lower values were associated reductions in body weights
(Spieth et al., 2000; Ma et al., 2005). However, the limited
control for other dietary factors (Spieth et al., 2000; Ma et al.,
2005) and substantial loss to follow-up (Spieth et al., 2000)
complicate the interpretation of these results.
Intervention studies of the GI and weight change
Table 4 presents data on intervention studies that examined
the dietary GI in relation to body weight and body fatness
(Slabber et al., 1994; Tsihlias et al., 2000; Bouche et al., 2002;
Wolever and Mehling, 2003; Sloth et al., 2004; Carels et al.,
2005; Raatz et al., 2005). These trials had a duration ranging
from 5 weeks (Bouche et al., 2002) to 12 months (Carels et al.,
2005), and included only dietary advice (Slabber et al., 1994;
Bouche et al., 2002; Carels et al., 2005; Raatz et al., 2005),
provision of key foods (Tsihlias et al., 2000; Wolever and
Mehling, 2003) or provision of foods that provided much of
the total energy intake (Sloth et al., 2004; Raatz et al., 2005).
The presented intervention studies tested effects of slowly
absorbed carbohydrates rather than fructose-rich low-GI
foods. Effects of the dietary GI on body weight may be
mediated by effects on satiety/satiation and energy intake.
Therefore, we did not include studies in which energy
intakes were fixed. In the study by Raatz et al. (2005) the
energy intake was fixed in the first phase of the study, but
not in the second ‘free-living’ phase. Although the dietary
advice in the Slabber et al. (1994) intervention had a specific
deficit in energy intake as a target, the characteristics of the
trial (no foods provided 12-week duration) implied that the
energy intake of the participants could still vary widely.
In one of the studies, the low-GI intervention resulted in a
statistically significant reduction in weight as compared with
the control intervention (Slabber et al., 1994). This study
consisted of a randomized parallel trial comparing a low-GI
diet (‘low insulin response’ diet) with a control diet
(‘balanced diet’), where about half of the participants
volunteered to subsequently switch the other diet resulting
in a cross-over study (Slabber et al., 1994). Although the same
trend was observed in the parallel study and the cross-over
study, differences in weight loss were only statistically
significant in the latter. The used cross-over study approach
may result in selection bias, because volunteering to switch
to the other diet may depend on the experience with the
previous diet. Another limitation of this study is that it is not
reported how the diets changed as a result of the provided
advice. Differences in diets other than the GI may have
occurred, for example the ‘low-insulin response’ advice also
included the instruction not to use any snacks. In a French
study, dietary advice to consume low-GI foods did not lead
to a lower body weight but resulted in a 0.5 kg lower fat mass
as compared with dietary advice to eat high-GI foods
(Bouche et al., 2002). Fiber intake also differed substantially
S89
European Journal of Clinical Nutrition

European Journal of Clinical Nutrition
Table 4 Intervention studies of low GI diets and body weight
Reference/country Participants a
(Slabber et al., 1994)
South Africa
(Tsihlias et al., 2000)
Canada
(Bouché et al., 2002)
France
(Wolever and Mehling,
2003) Canada
(Sloth et al., 2004)
Denmark
(Carels et al., 2005)
US
n ¼ 30 F age 3576
years, BMI 3574.
No loss to follow-up.
n ¼ 16 participants
volunteered to switch
to the other diet after
completion
M/F type 2 diabetes
High GI: n ¼ 22, age
63, BMI 28. Low GI:
n ¼ 26 age 62, BMI 28.
An additional n ¼ 7
(high GI) and n ¼ 4
(low GI) were lost to
follow up and not
included in analyses
n ¼ 11 months, age
46, BMI 28. No loss
to follow-up
M/F impaired glucose
tolerance; low GI:
n ¼ 13 age 55, BMI 30;
high GI: n ¼ 11 age 59,
BMI 29
Trial design a
Parallel design with
minimization and
cross-over study
Duration Intervention Control Compliance Results
12 week (2 12 weeks
for cross-over study):
counseling
Randomized parallel 6 months: breakfast
cereals (10–15% total
energy intake)
provided and meal
plans
Randomized crossover
study
2 5weekand5week
washout: counseling
Randomized parallel 4 months: monthly
counseling and key
foods provided
F low GI: n ¼ 23, age Randomized parallel,
29, BMI 27.6; high GI: participants blinded
n ¼ 22, age 31, BMI with regard to study
27.6; another n ¼ 6 aim
(low GI) and n ¼ 4
(high GI) lost to
follow-up not included
in primary analysis
n ¼ 40 F/M. Low GI,
age 43.4, BMI 38.0;
high GI, age 43.5,
BMI 37.2. An
additional n ¼ 13 was
lost to follow up and
not included in the
analysis
10 weeks: test foods
provided (B49% of
total energy)
Randomized parallel 20 week with weekly
group sessions and
12 months follow-up
Only carbohydrate
foods with low insulin
response,
carbohydrate foods in
separate meals, no
snacks
Low-GI cereals
provided
Foods with GI o45
recommended
Instructed X1 low-GI
food at each meal
Low-GI test foods (GI
103)
‘Balanced diet’. Both
advised intervention
and control diet
were hypocaloric
(4200–5000 kJ
per day)
High-GI cereals
provided
Foods with GI 460
recommended
Instructed X1
high-GI food at
each meal
High-GI test foods
(GI 79)
Behavioral weight loss Behavioral weight
program þ GI loss program only
education and popular
book on GI
Diet records.
Difference in GI not
reported
Diet records: low GI
reported 10 point
lower GI and
27 g per day higher
fiber intake as
compared with high
GI
Diet records: Low GI
30 point lower GI
and 12 g per day
higher fiber intake
Diet records: Low GI
decreased GI 4.3
points and increased
fiber 12 g per day.
No significant
changes for high GI
Test food diaries
(495% used),
urinary lithium
recovery from the
provided bread
(low GI 74%, high
GI 82%, P40.05),
diet records
Test of difference in
GI knowledge,
session attendance,
diet records: fivepoint
GI difference
achieved
Parallel: intervention
diet 9.3 kg, control
7.4 kg, difference
1.9 kg (95% CI 0.7,
4.6; P ¼ 0.14). Crossover:
intervention diet
7.4 kg, control
4.5 kg, difference
2.9 kg (95% CI 0.1,
5.8; P ¼ 0.04)
No statistically
significant differences
in body weight (high
GI B1 kg more weight
loss than low GI)
No statistically
significant difference
in body weight. Low
GI lost 0.52 kg body
fat, high GI 0.02
(difference 0.50 kg
Po0.05)
High GI: 0.49 kg, low
GI 0.19 kg
(difference 0.30 kg,
Po0.05)
Low GI: 1.9 (SE 0.5)
kg, high GI 1.3
(SE 0.3) kg (difference
0.6 kg, P ¼ 0.31) body
weight (intention to
treat: difference
0.4 kg, P ¼ 0.57). Fat
mass: low GI 1.0
(0.4) kg, high GI 0.4
(0.3) kg (P difference
0.20)
After 20 weeks: low
GI 7.1 kg weight loss,
controls: 8.2 kg weight
loss. Difference not
statistically significant,
also not after 12
months follow-up
S90
Carbohydrate intake and obesity
RM van Dam and JC Seidell

Table 4 Continued
Duration Intervention Control Compliance Results
Trial design a
Reference/country Participants a
Feeding phase: low GI
9.95 kg, 6.9 kg fat
mass. High GI 9.3 kg,
4.5 kg fat mass.
Additional weight
change in free-living
phase: 1.8 kg for low
GI and 1.6 kg for
high GI. No significant
differences between
groups
Diet records: no
significant difference
in GI between
groups after
free-living phase
Feeding phase:
hypocaloric (3128 kJ
per day deficit), high
GI. Free living: advised
to continue this diet
Feeding phase:
hypocaloric (3138 kJ
per dcay deficit), low
GI. Free living: advised
to continue this diet
F/M adults. Low GI, Randomized parallel 12-week feeding
n ¼ 10, BMI 37.7. High
phase (all foods
GI, n ¼ 9, BMI 34.6.
provided) þ 24-week
31% lost to follow-up
free-living phase with
in feeding phase.
counseling
During the free living
phase n ¼ 4 of the low
GI and n ¼ 1ofthe
high-GI group lost.
No intention to treat
analysis
(Raatz et al., 2005)
US
Abbreviations: BMI, body mass index (kg/m); F, female; GI, glycemic index; M, male.
Values are means (s.d.) unless reported otherwise.
a
None of the studies reported on blinding of outcome assessors or allocation concealment.
Carbohydrate intake and obesity
RM van Dam and JC Seidell
between the diets, and given the magnitude of the effect
other modest differences between the intervention diets may
also have been responsible. In a Danish study, almost half of
the foods were provided, allowing better control of differences
between the diets, and recovery of lithium from the
provided breads was used to monitor compliance (Sloth
et al., 2004). Despite the substantial differences in the GI of
the diets, no statistically significant effects on body weight
or fat mass were observed although a small beneficial effect
could not be excluded either. In the studies by Raatz et al.
(2005) (‘free living phase’ of the study) and Carels et al.
(2005) low-GI dietary advice did not have effects on body
weight as compared with the control interventions, but
achieved differences in dietary GI were small. In two
Canadian studies, provision of key low-glycemic foods
resulted in modest differences in dietary GI, but did not
result in a lower body weight over 4 or 6 months as
compared with the provision of high-GI foods (Tsihlias
et al., 2000; Wolever and Mehling, 2003). In contrast, weight
loss tended to be somewhat larger for the high-GI as
compared with the low-GI diet. A larger randomized
intervention study that evaluates substantial differences in
the dietary GI in the absence of other dietary differences
with the control group would be of interest. In addition, it
may be warranted to distinguish further between different
types of low-GI foods. However, the currently available data
provides little support for an important role of the dietary GI
in weight management.
Intervention studies of the GL and weight change
Four studies aimed to test the effects of low-GL diets on body
weight (Table 5). The effect of low-GL dietary advice on body
weight was tested in 16 obese adolescents (Ebbeling et al.,
2003). The low-GL advice resulted in a substantially larger
reduction in fat mass as compared with the conventional
control diet. Given the very small resulting differences in
dietary GI and carbohydrate intake between the intervention
and control diet, effects on body fat may also have resulted
from other beneficial characteristics of the recommended
foods (non-starchy vegetables, fruits, legumes, nuts, dairy) or
a greater acceptation of the ad libitum approach by the
adolescents (Ebbeling et al., 2003). The same low-GL advice
resulted in greater changes in GI, carbohydrate intake, and
GL in a trial in young adults, but did not result in a
statistically significant greater reduction in body fat or fat
mass (Ebbeling et al., 2005). In a substantially larger 12-week
Australian trial, four diets were compared: (A) a high-GI/
high-carbohydrate diet (highest GL), (B) a low-GI/highcarbohydrate
diet, (C) a high-GI/high-protein diet and (D)
a low-GI/high-protein diet (lowest GL) (McMillan-Price et al.,
2006). No significant differences in weight loss or decrease in
fat mass were found, although the percentage with at least
5% weight loss was greater for diets B and C than for the
other diets. Also, in a subgroup analysis in women, more fat
mass was lost for diets B and C than for the other diets.
S91
European Journal of Clinical Nutrition

European Journal of Clinical Nutrition
Table 5 Randomized intervention studies of low GL diets and body weight (all parallel design) a
Reference/
country
(Ebbeling et al.,
2003) US
(Ebbeling et al.,
2005) US
(McMillan-Price
et al., 2006) US
(Maki et al.,
2007) US
(Das et al. 2007)
US
Participants a
Obese adolescents M/F.
Low glycemic load (GL):
n ¼ 8, age 17, BMI 35, fat
mass 39 kg; high GL: n ¼ 8,
age 15, BMI 37, fat mass
49 kg (P ¼ 0.03 vs low GL);
n ¼ 2 lost to follow-up but
included in intention to
treat
F/M. Low GL: n ¼ 11, age
30, BMI 34.0. High GL:
n ¼ 12, age 27, BMI 27.2.
An additional n ¼ 6 (low
GL) and n ¼ 5 (high GL)
lost to follow up
F/M, age 18–40 BMI
X25 kg/m 2 A. High GI/
high carb: n¼ 32, B. Low
GI/high carb: n¼ 32, C.
High GI/high protein:
n ¼ 32, D. Low GI/high
protein: n ¼ 33. n ¼ 5, 2, 1
and 5 withdrew, but
included in intention to
treat
F/M, age 18–65 waist
X87 cm for women and
X90 cm for men. Low GL:
n ¼ 43, control diet:
n ¼ 43. n ¼ 10 and 7 lost to
follow-up, but all except 1
for both groups included in
intention to treat for
weight (last observation
carried forward)
F/M, age 24–42 BMI 25–
30 kg m 2 . Low GL: n ¼ 17.
High GL: n ¼ 17. n ¼ 3and
2 lost to follow-up
Duration Intervention Control Compliance Results
12 months: 12
counseling
sessions during
0–6 months and
2 during 6–12
months
Ebbeling et al.
(2003)
12 weeks:
counseling and
provision of key
foods each week
36 weeks starting
with 12–24 weeks
weigh loss phase
followed by
weight
maintenance.
Counseling at 14
visits and written
information
1 year: 6 months,
where foods were
provided followed
by 6 months with
counseling and
self-selected foods
Low to moderate GI foods,
balanced consumption of
carbohydrates with fat and
protein at each meal/
snack; ad libitum
Low glycemic load: see
Ebbeling et al.(2003)
All eating plans were
hypocaloric (B1400 kcal
for women and 1900 kcal
for men), but participants
were told to ‘eat to
appetite’ and accordingly
eat more or less than
eating plan. Diet A had the
highest GL (diet records:
129 g) and diet D the
lowest GL (59 g)
‘Ad libitum reduced GL’.
first phase: 2 weeks with
elimination of high
carbohydrate foods.
Second phase: addition of
low-GI foods
Low GL with 30% caloric
restriction. Lower GI foods
and 40 en% carb, 30 en%
fat, 30 en% protein
Reduced fat and
hypocaloric (250–
500 kcal per day
deficit)
Reduced fat and
hypocaloric (250–
500 kcal per day
deficit)
‘Portioncontrolled
low-fat
diet’. Reduction
of high-fat foods,
portion sizes, and
energy density
aiming at deficit
of 500–800 kcal
per day
High GL with
30% caloric
restriction. Higher
GI foods and 60
en% carb, 20
en% fat, 20 en%
protein and
higher GI foods
Diet records (follow-up):
low GL 10 g per
1000 kcal lower GL, 3
en% lower carbohydrate
and 3 point lower GI
than control group
Diet records. Low vs high
GL: 24 g per 1000 kcal
lower GL, 7 point lower
GI, 12 en% lower
carbohydrate
Diet records and
monitoring of 10 h
glucose and insulin
responses indicated that
substantial contrasts in GI
and GL existed between
the diets. Fiber intake
was substantially higher
for diet B than for the
other diets
FFQ (for GI and GL) and
diet records. Low GL vs
control (week 36): 55 g
lower carbohydrate, 23 g
lower sucrose/fructose, 3
point lower GI, 33%
lower GL
Provided foods for high
GL vs low GL had a 63%
higher GI 161% higher
GL. Caloric restriction
(doubly labeled water) at
3 months, 6 months and
12 months were 21%,
16%, and 17% for high
GL and 28%, 18% and
10% for low GL
Abbreviations: BMI, body mass index (kg m 2 ); en%, energy percent; F, female; FFQ, food frequency questionnaire; GI, glycemic index; GL, glycemic load; M, male. Values are means.
a Das et al. reported the blinding of outcome assessors. None of the other studies reported on blinding of outcome assessors or allocation concealment.
BMI decreased 1.8
(P ¼ 0.02) and fat mass
4.2 kg (P ¼ 0.01) for low
GL as compared with
control diet
Low GL: weight 7.8 kg,
fat mass 16.5%. High
GL: weight 6.1 kg
(P ¼ 0.89 vs low GL), fat
mass 15.7% (P ¼ 0.97).
Similar for intention to
treat analysis
Diet A: weight 3.7, fat
2.8 kg; diet B: weight
4.8, fat 4.5 kg; diet C:
weight 5.3, fat 4.3 kg;
diet D: weight 4.4, fat
3.7 kg P ¼ 0.17 for
differences in weight and
P ¼ 0.08 for differences in
fat mass
Low GL: weight 4.9, fat
2.2, fat-free mass
2.1 kg. Control: weight
2.5, fat 1.3, fat-free
mass 0.9 kg P ¼ 0.09 for
difference in weight,
P ¼ 0.33 for difference in
fat, P ¼ 0.004 for
difference in fat-free mass
Low GL: weight 10.4%
(6 months) and 7.8%
(12 months). High GL:
weight 9.1% (6
months) and 8.0% (12
months). No significant
differences between diet
groups (also not in body
fat %)
S92
Carbohydrate intake and obesity
RM van Dam and JC Seidell

Despite a substantially lower GL for diet D as compared with
diet A, according to diet records ( 54%) and 10-h incremental
area under the glucose ( 38%) and insulin ( 45%)
curve on profile days, no substantial or significant difference
in loss of fat mass was observed. Thus, although results from
this study could be seen as supportive for a low-GI /highcarbohydrate/high-fiber
diet and a high-protein diet for
weight loss, the unexpected results for the diet that
combined a low-GI and high protein intake, the sex
differences in results and the limited duration warrant a
cautious interpretation. In a US study, a reduced GL diet was
compared with a portion-controlled low-fat diet (Maki et al.,
2007). Weight loss and loss of fat mass was greater for the
low-GL diet as compared with the control diet after 12 weeks
(‘weight loss phase’), but these differences did not remain
significant after 36 weeks (‘weight maintenance phase’).
Finally, another US study in 34 participants combined a goal
of 30% caloric restriction with a low or high-GL diet (Das
et al., 2007). No significant differences in weight loss or
energy intake measured with the doubly labeled water
method were found after either the 6-month phase in which
all foods were provided or the subsequent 6-month phase
where the participants selected the foods themselves. In
summary, studies that have directly compared low- and
high-GL diets do not consistently support this hypothesis
that a low-GL diet supports weight loss.
Conclusion on GI and GL
Current evidence from studies on low-GI and low-GL diets is
too limited to warrant specific recommendation with regard
to these concepts for the prevention of obesity. Caution
with regard to the health effects of high fructose intake is
warranted, whereas other low-GI foods such as legumes that
also have other beneficial nutritional properties could be
recommended as part of a diet for weight control.
Dietary fiber
Introduction
A higher fiber content contributes to a lower energy density
of foods (Slavin, 2005). In addition, viscous fibers form a
viscous gel in contact with water and have been suggested to
reduce energy intake through increased feelings of fullness
(Pasman et al., 1997). These types of fiber may also increase
gastrointestinal transit time, slow glucose absorption and
lead to a more gradual increase in blood glucose levels (see
glycemic index section) (Slavin, 2005).
Intervention studies of ‘fiber’ supplements and body weight
Guar gum is a viscous galactomannan fiber that is
commonly used in the form of supplements in attempts to
lose weight (Pittler and Ernst, 2001). In a meta-analysis
published in 2001, the results of 11 randomized double-blind
Carbohydrate intake and obesity
RM van Dam and JC Seidell
trials of guar gum intake and body weight were pooled
(Pittler and Ernst, 2001). The results did not suggest a
difference in effect on body weight as compared with
placebo treatment (weighted mean difference 0.04 kg;
95% confidence interval 2.2 to 2.1). In a 14-month
randomized intervention study with 11 obese women,
adding 20 g of guar gum supplements per day to an ad
libitum diet after completion of a very low-calorie diet did
not reduce weight as compared with controls (Pasman et al.,
1997). Timing of the intake of fiber supplements in relation
to meals may be relevant (Pasman et al., 1997). Also, fiber
supplements may not have appreciable effects in isolation,
but rather improve adherence to hypocaloric diets. Ryttig
et al. (1989) examined the effects of a fiber supplement in
combination with a hypocaloric diet in a series of doubleblind,
randomized trials of 3- to 12-month duration in
overweight persons (Rossner et al., 1987; Rigaud et al., 1990).
The addition of the fiber supplements to a hypocaloric diet
led to a statistically significant 1–2 kg greater weight loss as
compared with a placebo in combination with the hypocaloric
diet. An effect of similar magnitude was observed for
adding viscous glucomannan fiber supplement to a hypocaloric
diet in 5-week placebo-controlled double-blind
randomized study (Birketvedt et al., 2005). In summary,
although results for ‘fiber’ supplements on body weight have
been mixed there is some evidence that these supplements
may increase adherence to hypocaloric diets and thus lead to
a small additional weight loss. The role of different types of
fiber as well as the timing of fiber intake requires further
study in this context.
Weight loss interventions with high-fiber foods
In a randomized trial in obese men and women that received
an energy-restricted diet (target 500 kcal per day deficit) for
48 weeks additional advice to increase consumption of
vegetables, fruit and whole grains was associated with a
B10 g per day reported higher fiber intake, but not with
greater weight loss (Thompson et al., 2005). In a 2-year
randomized trial, the effect of a ‘Mediterranean-style’ diet
high in fiber-rich foods was tested in 180 persons with the
metabolic syndrome (Esposito et al., 2004). The intervention
group received individualized dietary advice from a dietician
(monthly session in the 1st year, bimonthly in the second
year) and participated in group sessions with education on
reducing calories, personal goal setting and self-monitoring.
Aims included increasing consumption of fruits, vegetables,
legumes, nuts, whole grains and olive oil combined with less
than 30% of energy from fat and 50–60% from carbohydrates.
The control group did not receive individualized
advice, but only general recommendations based on the
same goals for macronutrient intakes. Weight loss after 2
years was 4.0 kg for the intervention and 1.2 kg for the
control group (difference 2.8 kg, 95% confidence interval
0.5–5.1). The 16 g per day greater increase in fiber intake as
well as other characteristics of the recommended foods or
S93
European Journal of Clinical Nutrition

S94
the educational program may have contributed to this
greater weight loss.
Intervention studies of high-fiber foods not primarily aimed at
weight loss
In the Women’s Health Initiative, a modest self-reported
increase in fruit, vegetable and fiber intake was observed in
the intervention group and this was not associated with a
substantial long-term effect on body weight (Howard et al.,
2006). Other trials that were not aimed at weight loss
generally did not find an effect of increased consumption of
fruits and vegetables on body weight (Rolls et al., 2004). For
example, in a 1-year trial in 201 men and women specifically
aimed at fruit and vegetable consumption, fruit (diet records
þ 2.8 serving per day), vegetable ( þ 1.4 serving per day),
juice ( þ 2.5 serving per day) and fiber intake ( þ 5.5 g per
day) increased substantially more in the intervention as
compared with the control group (Smith-Warner et al.,
2000). These changes were supported by increased plasma
carotenoid concentrations, but no appreciable difference in
weight change was found between the intervention
( þ 0.3 kg) and control group ( þ 0.4 kg). It has been suggested
that concomitant increases in juice consumption, a highcaloric
beverage, may have reduced possible beneficial effects
of the fruits and vegetable interventions on body weight
(Rolls et al., 2004). An alternative explanation that is
consistent with the results from prospective cohort studies
is that effects of higher consumption of fiber-rich foods on
body weight are modest and only apparent after several
years.
Cohort studies of high-fiber foods and long-term weight change
The association between fiber intake and weight change has
been examined in several large prospective cohort studies
with detailed adjustment for potential confounders. In a
study of men, a 10 g per day increase in fiber intake was
associated with a 2.25 kg lower weight gain over 8 years after
correction for measurement error in the assessment of fiber
intake (Koh-Banerjee et al., 2004). Similar results have been
obtained in women, younger adults, and a Mediterranean
population (Ludwig et al., 1999; Liu et al., 2003; Bes-Rastrollo
et al., 2006a). Higher consumption of whole grains (Liu et al.,
2003; Koh-Banerjee et al., 2004) and fruits and vegetables (He
et al., 2004; Bes-Rastrollo et al., 2006a) was also associated
with a modestly lower long-term weight gain. These findings
may reflect a dietary pattern rather than fiber intake per se,
and the possibility that confounding by unmeasured or
imprecisely measured factors contributed to these associations
cannot be excluded.
Conclusion on dietary fiber
The currently available evidence does not support a major
effect of fiber intake on weight loss, but evidence is
European Journal of Clinical Nutrition
Carbohydrate intake and obesity
RM van Dam and JC Seidell
consistent with modest long-term effects of higher consumption
of fiber-rich foods on body weight gain. An
important reason to attempt weight loss is the prevention
of morbidity associated with excess adiposity including type
2 diabetes, coronary heart disease and stroke (Willett et al.,
1999). Consumption of a diet high in fruits, vegetables and
whole grains is associated with reduced risk of these diseases
(Appel et al., 1997; de Lorgeril et al., 1999; Jacobs and
Gallaher, 2004) and should therefore be included in diets for
long-term weight management. Ad libitum intervention
studies that specifically examine the effects of whole grain
consumption on body weight appear to be lacking and this
topic warrants further research.
General discussion
Although weight gain is due to an imbalance between energy
intake and energy expenditure, this equation is deceivingly
simple in the context of obesity prevention and treatment
under realistic conditions. Complex feedback mechanisms
appear to stimulate regaining lost weight (Hirsch et al., 1998)
and potent environmental influences stimulate over-consumption
of ubiquitously available foods and a sedentary
lifestyle (Egger and Swinburn, 1997). The optimal macronutrient
composition of diet for weight management may
also differ by individual characteristics. First, the response to
diets may differ by genetic susceptibility of individuals
(Memisoglu et al., 2003), although a recent study of 26
candidate genes did not identify important modifiers of the
effects of a high or low carbohydrate weight loss diet
(Sorensen et al., 2006). Second, findings from two recent
randomized intervention studies suggest that insulin sensitivity
or glucose-induced insulin secretion can modify the
weight loss in response to low-carbohydrate or low-GL diets
(Cornier et al., 2005; Pittas et al., 2005). Larger studies are
needed to test the hypothesis that characteristics related to a
person’s glucose homeostasis may affect the optimal macronutrient
composition for weight management. Finally, the
carbohydrate content of a diet that allows long-term
adherence may to a large extent depend on individual
preferences. For prevention of weight gain, individuals
should be able to maintain a dietary pattern that prevents
excess energy intake for decades and the degree of adherence
to weight loss interventions is a strong predictor of weight
loss (Heshka et al., 2003; Dansinger et al., 2005).
This overview is limited to dietary carbohydrates in
relation to obesity. However, many other factors should be
considered in the design of weight loss interventions. Selfmonitoring
of diet and activity, structural approaches
providing foods, meal replacements or menus and recipes,
increased physical activity, and longer length of treatment
have been shown to contribute to greater weight loss (Perri
et al., 1989; Foster et al., 2005). Although, the achieved
weight loss of long-term interventions may appear small,
B5% loss of body weight as a result of dietary changes and

increased physical activity has been associated with a major
decrease in incidence of type 2 diabetes and other metabolic
disturbances (Tuomilehto et al., 2001; Knowler et al., 2002).
The high prevalence of overweight and obesity indicates that
energy imbalance is a population-wide phenomenon in
many societies across the world. In addition to interventions
aimed at individuals, rigorous changes in the social and
physical environment including the food supply that reduce
stimuli to overeat and facilitate greater physical activity will
be needed to target the epidemic of obesity.
Although more randomized longer-term trials of diet and
body weight have been conducted in recent years, many
caveats exist in the current evidence and further research
efforts to address this topic that is of pivotal importance for
public health is strongly recommended.
Conclusions and recommendations
Carbohydrates are among the macronutrients that provide
energy and can thus contribute to excess energy intake and
subsequent weight gain. If energy intake is strictly controlled,
macronutrient composition of the diet (energy
percentages of fat and carbohydrates) does not substantially
affect body weight or fat mass (Golay et al., 1996). However,
an important issue is whether among free-living individuals,
macronutrient composition of the diet increases the likelihood
of passive over-consumption. There is no clear
evidence that altering the proportion of total carbohydrate
in the diet is an important determinant of energy intake.
However, there is evidence that sugar-sweetened beverages
do not induce satiety to the same extent as solid forms of
carbohydrate and that increases in sugar sweetened soft
drink consumption are associated with weight gain. Thus,
there is a justification for the recommendation to restrict the
use of beverages high in free sugars in order to reduce the risk
of excessive weight gain and to treat obesity. Solid foods high
in free sugars tend to be energy dense, and there is some
evidence from intervention studies that reduction of solid
foods high in free sugars can contribute to weight loss. Thus,
the current recommendation to restrict free sugars to no
more than 10 percent of total energy is consistent with
appropriate diets for the prevention of obesity. A high
content of dietary fiber in whole-grain cereals, vegetables,
legumes and fruits is associated with relatively low energy
density, promotion of satiety, and in observational studies a
lesser degree of weight gain than among those with lower
intakes. Although it is difficult to establish with certainty
that dietary fiber rather than other dietary attributes are
responsible, it is considered appropriate to recommend that
whole grain cereals, vegetables, legumes and fruits are the
most appropriate sources of dietary carbohydrate. The
currently available evidence is considered to be insufficient
to recommend carbohydrate-containing foods likely to
reduce the risk of obesity or promote weight loss on the
basis of their GI.
Carbohydrate intake and obesity
RM van Dam and JC Seidell
Acknowledgements
We thank Dr Ahmad R Dorosty, Dr Cara Ebbeling, Professor
Philip James, Professor Simin Liu, Professor Jim Mann,
Professor Boyd Swinburn, Professor Carolyn Summerbell
and Dr Ricardo Uauy for their valuable comments on the
earlier manuscript.
Conflict of interest
During preparation and peer-review of this paper in 2006, the
authors and peer-reviewers declared the following interests.
Authors
Dr Rob M van Dam: None declared.
Professor Jaap C Seidell: None declared.
Peer-reviewers
Dr Ahmad R Dorosty: None declared.
Dr Cara Ebbeling: None declared.
Professor Philip James: None declared.
Professor Simin Liu: None declared.
Professor Jim Mann: None declared.
Professor Boyd Swinburn: None declared.
Professor Carolyn Summerbell: None declared.
Dr Ricardo Uauy: Scientific Adviser on a temporary basis
for Unilever and Wyeth; Scientific Editorial/Award Adviser
for Danone, DSM, Kelloggs, and Knowles and Bolton on an
ad hoc basis.
References
Alexy U, Sichert-Hellert W, Kersting M (2003). Associations between
intake of added sugars and intakes of nutrients and food groups in
the diets of German children and adolescents. Br J Nutr 90, 441–447.
Anderson JW, Story LJ, Zettwoch NC, Gustafson NJ, Jefferson BS
(1989). Metabolic effects of fructose supplementation in diabetic
individuals. Diabetes Care 12, 337–344.
Anonymous (1968). The National Diet-Heart Study Final Report.
Circulation 37, I1–I428.
Appel LJ, Moore TJ, Obarzanek E, Vollmer WM, Svetkey LP, Sacks FM
et al. (1997). A clinical trial of the effects of dietary patterns on
blood pressure. DASH Collaborative Research Group. N Engl J Med
336, 1117–1124.
Astrup A, Meinert Larsen T, Harper A (2004). Atkins and other lowcarbohydrate
diets: hoax or an effective tool for weight loss?
Lancet 364, 897–899.
Baron JA, Schori A, Crow B, Carter R, Mann JI (1986). A randomized
controlled trial of low carbohydrate and low fat/high fiber diets for
weight loss. Am J Public Health 76, 1293–1296.
Berkey CS, Rockett HR, Field AE, Gillman MW, Colditz GA (2004). Sugaradded
beverages and adolescent weight change. Obes Res 12, 778–788.
Bes-Rastrollo M, Martinez-Gonzalez MA, Sanchez-Villegas A, de la
Fuente Arrillaga C, Martinez JA (2006a). Association of fiber intake
and fruit/vegetable consumption with weight gain in a Mediterranean
population. Nutrition 22, 504–511.
Bes-Rastrollo M, Sanchez-Villegas A, Gomez-Gracia E, Martinez JA,
Pajares RM, Martinez-Gonzalez MA (2006b). Predictors of weight
gain in a Mediterranean cohort: the Seguimiento Universidad de
Navarra Study 1. Am J Clin Nutr 83, 362–370;quiz 394–395.
Birketvedt GS, Shimshi M, Erling T, Florholmen J (2005). Experiences
with three different fiber supplements in weight reduction. Med Sci
Monit 11, PI5–PI8.
S95
European Journal of Clinical Nutrition

S96
Carbohydrate intake and obesity
RM van Dam and JC Seidell
Blaak EE (2004). Basic disturbances in skeletal muscle fatty acid
metabolism in obesity and type 2 diabetes mellitus. Proc Nutr Soc
63, 323–330.
Bouche C, Rizkalla SW, Luo J, Vidal H, Veronese A, Pacher N et al.
(2002). Five-week, low-glycemic index diet decreases total fat mass
and improves plasma lipid profile in moderately overweight
nondiabetic men. Diabetes Care 25, 822–828.
Braam LA, Ocke MC, Bueno-de-Mesquita HB, Seidell JC (1998).
Determinants of obesity-related underreporting of energy intake.
Am J Epidemiol 147, 1081–1086.
Bray GA, Nielsen SJ, Popkin BM (2004a). Consumption of highfructose
corn syrup in beverages may play a role in the epidemic of
obesity. Am J Clin Nutr 79, 537–543.
Bray GA, Paeratakul S, Popkin BM (2004b). Dietary fat and obesity: a
review of animal, clinical and epidemiological studies. Physiol
Behav 83, 549–555.
Brehm BJ, Seeley RJ, Daniels SR, D’Alessio DA (2003). A randomized
trial comparing a very low carbohydrate diet and a calorierestricted
low fat diet on body weight and cardiovascular risk
factors in healthy women. J Clin Endocrinol Metab 88, 1617–1623.
Brinkworth GD, Noakes M, Keogh JB, Luscombe ND, Wittert GA,
Clifton PM (2004a). Long-term effects of a high-protein, lowcarbohydrate
diet on weight control and cardiovascular risk
markers in obese hyperinsulinemic subjects. Int J Obes Relat Metab
Disord 28, 661–670.
Brinkworth GD, Noakes M, Parker B, Foster P, Clifton PM (2004b).
Long-term effects of advice to consume a high-protein, low-fat
diet, rather than a conventional weight-loss diet, in obese adults
with type 2 diabetes: one-year follow-up of a randomised trial.
Diabetologia 47, 1677–1686.
Carels RA, Darby LA, Douglass OM, Cacciapaglia HM, Rydin S (2005).
Education on the glycemic index of foods fails to improve
treatment outcomes in a behavioral weight loss program. Eat
Behav 6, 145–150.
Charlton KE, Kolbe-Alexander TL, Nel JH (2005). Micronutrient
dilution associated with added sugar intake in elderly black South
African women. Eur J Clin Nutr 59, 1030–1042.
Cornier MA, Donahoo WT, Pereira R, Gurevich I, Westergren R,
Enerback S et al. (2005). Insulin sensitivity determines the
effectiveness of dietary macronutrient composition on weight
loss in obese women. Obes Res 13, 703–709.
Dansinger ML, Gleason JA, Griffith JL, Selker HP, Schaefer EJ (2005).
Comparison of the Atkins, Ornish, Weight Watchers, and Zone
diets for weight loss and heart disease risk reduction: a randomized
trial. JAMA 293, 43–53.
Das SK, Gilhooly CH, Golden JK, Pittas AG, Fuss PJ, Cheatham RA
et al. (2007). Long-term effects of 2 energy-restricted diets differing
in glycemic load on dietary adherence, body composition, and
metabolism in CALERIE: a 1-y randomized controlled trial. Am J
Clin Nutr 85, 1023–1030.
de Graaf C, Blom WA, Smeets PA, Stafleu A, Hendriks HF
(2004). Biomarkers of satiation and satiety. Am J Clin Nutr 79,
946–961.
de Lorgeril M, Salen P, Martin JL, Monjaud I, Delaye J, Mamelle N
(1999). Mediterranean diet, traditional risk factors, and the rate of
cardiovascular complications after myocardial infarction: final
report of the Lyon Diet Heart Study. Circulation 99, 779–785.
DiMeglio DP, Mattes RD (2000). Liquid versus solid carbohydrate:
effects on food intake and body weight. Int J Obes Relat Metab
Disord 24, 794–800.
Drewnowski A, Bellisle F (2007). Liquid calories, sugar, and body
weight. Am J Clin Nutr 85, 651–661.
Drewnowski A, Almiron-Roig E, Marmonier C, Lluch A (2004).
Dietary energy density and body weight: is there a relationship?
Nutr Rev 62, 403–413.
Due A, Toubro S, Skov AR, Astrup A (2004). Effect of normal-fat diets,
either medium or high in protein, on body weight in overweight
subjects: a randomised 1-year trial. Int J Obes Relat Metab Disord 28,
1283–1290.
European Journal of Clinical Nutrition
Ebbeling CB, Feldman HA, Osganian SK, Chomitz VR, Ellenbogen SJ,
Ludwig DS (2006). Effects of decreasing sugar-sweetened beverage
consumption on body weight in adolescents: a randomized,
controlled pilot study. Pediatrics 117, 673–680.
Ebbeling CB, Leidig MM, Sinclair KB, Hangen JP, Ludwig DS (2003). A
reduced-glycemic load diet in the treatment of adolescent obesity.
Arch Pediatr Adolesc Med 157, 773–779.
Ebbeling CB, Leidig MM, Sinclair KB, Seger-Shippee LG, Feldman HA,
Ludwig DS (2005). Effects of an ad libitum low-glycemic load diet
on cardiovascular disease risk factors in obese young adults. Am J
Clin Nutr 81, 976–982.
Egger G, Swinburn B (1997). An ‘ecological’ approach to the obesity
pandemic. BMJ 315, 477–480.
Elliott SS, Keim NL, Stern JS, Teff K, Havel PJ (2002). Fructose, weight
gain, and the insulin resistance syndrome. Am J Clin Nutr 76, 911–922.
Ello-Martin JA, Ledikwe JH, Rolls BJ (2005). The influence of food
portion size and energy density on energy intake: implications for
weight management. Am J Clin Nutr 82, 236S–241S.
Esposito K, Marfella R, Ciotola M, Di Palo C, Giugliano F, Giugliano
G et al. (2004). Effect of a mediterranean-style diet on endothelial
dysfunction and markers of vascular inflammation in the metabolic
syndrome: a randomized trial. JAMA 292, 1440–1446.
Foster GD, Makris AP, Bailer BA (2005). Behavioral treatment of
obesity. Am J Clin Nutr 82, 230S–235S.
Foster GD, Wyatt HR, Hill JO, McGuckin BG, Brill C, Mohammed BS
et al. (2003). A randomized trial of a low-carbohydrate diet for
obesity. N Engl J Med 348, 2082–2090.
Foster-Powell K, Holt SH, Brand-Miller JC (2002). International table
of glycemic index and glycemic load values: 2002. Am J Clin Nutr
76, 5–56.
Franks PW, Luan J, Browne PO, Harding AH, O’Rahilly S, Chatterjee
VK et al. (2004). Does peroxisome proliferator-activated receptor
gamma genotype (Pro12ala) modify the association of physical
activity and dietary fat with fasting insulin level? Metabolism 53,
11–16.
Gardner CD, Kiazand A, Alhassan S, Kim S, Stafford RS, Balise RR et al.
(2007). Comparison of the Atkins, Zone, Ornish, and LEARN diets
for change in weight and related risk factors among overweight
premenopausal women: the A TO Z Weight Loss Study: a
randomized trial. JAMA 297, 969–977.
Gatenby SJ, Aaron JI, Jack VA, Mela DJ (1997). Extended use of foods
modified in fat and sugar content: nutritional implications in a
free-living female population. Am J Clin Nutr 65, 1867–1873.
Golay A, Allaz AF, Morel Y, de Tonnac N, Tankova S, Reaven G (1996).
Similar weight loss with low- or high-carbohydrate diets. Am J Clin
Nutr 63, 174–178.
Halton TL, Hu FB (2004). The effects of high protein diets on
thermogenesis, satiety and weight loss: a critical review. J Am Coll
Nutr 23, 373–385.
Harvey-Berino J (1998). The efficacy of dietary fat vs total energy
restriction for weight loss. Obes Res 6, 202–207.
He K, Hu FB, Colditz GA, Manson JE, Willett WC, Liu S (2004).
Changes in intake of fruits and vegetables in relation to risk of
obesity and weight gain among middle-aged women. Int J Obes
Relat Metab Disord 28, 1569–1574.
Heerstrass DW, Ocke MC, Bueno-de-Mesquita HB, Peeters PH, Seidell
JC (1998). Underreporting of energy, protein and potassium intake
in relation to body mass index. Int J Epidemiol 27, 186–193.
Heitmann BL, Lissner L (1995). Dietary underreporting by obese
individuals—is it specific or non-specific? BMJ 311, 986–989.
Heshka S, Anderson JW, Atkinson RL, Greenway FL, Hill JO, Phinney
SD et al. (2003). Weight loss with self-help compared with a
structured commercial program: a randomized trial. JAMA 289,
1792–1798.
Hirsch J, Hudgins LC, Leibel RL, Rosenbaum M (1998). Diet
composition and energy balance in humans. Am J Clin Nutr 67,
551S–555S.
Hoag S, Marshall JA, Jones RH, Hamman RF (1995). High fasting
insulin levels associated with lower rates of weight gain in persons

with normal glucose tolerance: the San Luis Valley Diabetes Study.
Int J Obes Relat Metab Disord 19, 175–180.
Howard BV, Manson JE, Stefanick ML, Beresford SA, Frank G, Jones B
et al. (2006). Low-fat dietary pattern and weight change over 7
years: the Women’s Health Initiative Dietary Modification Trial.
JAMA 295, 39–49.
Iqbal SI, Helge JW, Heitmann BL (2006). Do energy density and
dietary fiber influence subsequent 5-year weight changes in adult
men and women? Obesity (Silver Spring) 14, 106–114.
Jacobs Jr DR, Gallaher DD (2004). Whole grain intake and
cardiovascular disease: a review. Curr Atheroscler Rep 6, 415–423.
James J, Thomas P, Cavan D, Kerr D (2004). Preventing childhood
obesity by reducing consumption of carbonated drinks: cluster
randomised controlled trial. BMJ 328, 1237.
Jeffery RW, Hellerstedt WL, French SA, Baxter JE (1995). A
randomized trial of counseling for fat restriction versus calorie
restriction in the treatment of obesity. Int J Obes Relat Metab Disord
19, 132–137.
Kendall A, Levitsky DA, Strupp BJ, Lissner L (1991). Weight loss on a
low-fat diet: consequence of the imprecision of the control of food
intake in humans. Am J Clin Nutr 53, 1124–1129.
Klein S, Wadden T, Sugerman HJ (2002). AGA technical review on
obesity. Gastroenterology 123, 882–932.
Knowler WC, Barrett-Connor E, Fowler SE, Hamman RF, Lachin JM,
Walker EA et al. (2002). Reduction in the incidence of type 2
diabetes with lifestyle intervention or metformin. N Engl J Med
346, 393–403.
Koh-Banerjee P, Franz M, Sampson L, Liu S, Jacobs Jr DR, Spiegelman
D et al. (2004). Changes in whole-grain, bran, and cereal fiber
consumption in relation to 8-y weight gain among men. Am J Clin
Nutr 80, 1237–1245.
Kranz S, Smiciklas-Wright H, Siega-Riz AM, Mitchell D (2005).
Adverse effect of high added sugar consumption on dietary intake
in American preschoolers. J Pediatr 146, 105–111.
Lean ME, Han TS, Prvan T, Richmond PR, Avenell A (1997). Weight
loss with high and low carbohydrate 1200 kcal diets in free living
women. Eur J Clin Nutr 51, 243–248.
Ledikwe JH, Blanck HM, Kettel Khan L, Serdula MK, Seymour JD,
Tohill BC et al. (2006). Dietary energy density is associated with
energy intake and weight status in US adults. Am J Clin Nutr 83,
1362–1368.
Liljeberg HG, Akerberg AK, Bjorck IM (1999). Effect of the glycemic
index and content of indigestible carbohydrates of cereal-based
breakfast meals on glucose tolerance at lunch in healthy subjects.
Am J Clin Nutr 69, 647–655.
Lindi VI, Uusitupa MI, Lindstrom J, Louheranta A, Eriksson JG, Valle
TT et al. (2002). Association of the Pro12Ala polymorphism in the
PPAR-gamma2 gene with 3-year incidence of type 2 diabetes and
body weight change in the Finnish Diabetes Prevention Study.
Diabetes 51, 2581–2586.
Liu S, Willett WC, Manson JE, Hu FB, Rosner B, Colditz G (2003).
Relation between changes in intakes of dietary fiber and grain
products and changes in weight and development of obesity
among middle-aged women. Am J Clin Nutr 78, 920–927.
Ludwig DS (2002). The glycemic index: physiological mechanisms
relating to obesity, diabetes, and cardiovascular disease. JAMA 287,
2414–2423.
Ludwig DS, Pereira MA, Kroenke CH, Hilner JE, Van Horn L, Slattery
ML et al. (1999). Dietary fiber, weight gain, and cardiovascular
disease risk factors in young adults. JAMA 282, 1539–1546.
Ma Y, Olendzki B, Chiriboga D, Hebert JR, Li Y, Li W et al. (2005).
Association between dietary carbohydrates and body weight. Am J
Epidemiol 161, 359–367.
Maki KC, Rains TM, Kaden VN, Raneri KR, Davidson MH (2007).
Effects of a reduced-glycemic-load diet on body weight, body
composition, and cardiovascular disease risk markers in overweight
and obese adults. Am J Clin Nutr 85, 724–734.
Malik VS, Schulze MB, Hu FB (2006). Intake of sugar-sweetened beverages
and weight gain: a systematic review. Am J Clin Nutr 84, 274–288.
Carbohydrate intake and obesity
RM van Dam and JC Seidell
Mann JI, Hendricks DA, Truswell AS, Manning E (1970). Effects on
serum-lipids in normal men of reducing dietary sucrose or starch
for five months. Lancet 1, 870–872.
Mattes RD (1996). Dietary compensation by humans for supplemental
energy provided as ethanol or carbohydrate in fluids.
Physiol Behav 59, 179–187.
Mayer-Davis EJ, Sparks KC, Hirst K, Costacou T, Lovejoy JC,
Regensteiner JG et al. (2004). Dietary intake in the diabetes
prevention program cohort: baseline and 1-year post randomization.
Ann Epidemiol 14, 763–772.
McAuley KA, Hopkins CM, Smith KJ, McLay RT, Williams SM, Taylor
RW et al. (2005). Comparison of high-fat and high-protein diets
with a high-carbohydrate diet in insulin-resistant obese women.
Diabetologia 48, 8–16.
McAuley KA, Smith KJ, Taylor RW, McLay RT, Williams SM, Mann JI
(2006). Long-term effects of popular dietary approaches on weight
loss and features of insulin resistance. Int J Obes (London) 30,
342–349.
McManus K, Antinoro L, Sacks F (2001). A randomized controlled
trial of a moderate-fat, low-energy diet compared with a low fat,
low-energy diet for weight loss in overweight adults. Int J Obes
Relat Metab Disord 25, 1503–1511.
McMillan-Price J, Brand-Miller J (2006). Low-glycaemic index diets
and body weight regulation. Int J Obes (London) 30 (Suppl 3),
S40–S46.
McMillan-Price J, Petocz P, Atkinson F, O’Neill K, Samman S,
Steinbeck K et al. (2006). Comparison of 4 diets of varying
glycemic load on weight loss and cardiovascular risk reduction in
overweight and obese young adults: a randomized controlled trial.
Arch Intern Med 166, 1466–1475.
Melanson KJ, Westerterp-Plantenga MS, Saris WH, Smith FJ,
Campfield LA (1999). Blood glucose patterns and appetite in
time-blinded humans: carbohydrate versus fat. Am J Physiol 277,
R337–R345.
Memisoglu A, Hu FB, Hankinson SE, Manson JE, De Vivo I, Willett
WC et al. (2003). Interaction between a peroxisome proliferatoractivated
receptor gamma gene polymorphism and dietary fat
intake in relation to body mass. Hum Mol Genet 12, 2923–2929.
Nordmann AJ, Nordmann A, Briel M, Keller U, Yancy Jr WS, Brehm BJ
et al. (2006). Effects of low-carbohydrate vs low-fat diets on weight
loss and cardiovascular risk factors: a meta-analysis of randomized
controlled trials. Arch Intern Med 166, 285–293.
Ogden CL, Carroll MD, Curtin LR, McDowell MA, Tabak CJ, Flegal
KM (2006). Prevalence of overweight and obesity in the United
States, 1999–2004. JAMA 295, 1549–1555.
Pascale RW, Wing RR, Butler BA, Mullen M, Bononi P (1995). Effects
of a behavioral weight loss program stressing calorie restriction
versus calorie plus fat restriction in obese individuals with NIDDM
or a family history of diabetes. Diabetes Care 18, 1241–1248.
Pasman WJ, Westerterp-Plantenga MS, Muls E, Vansant G, van Ree J,
Saris WH (1997). The effectiveness of long-term fibre supplementation
on weight maintenance in weight-reduced women. Int J
Obes Relat Metab Disord 21, 548–555.
Pereira MA, Swain J, Goldfine AB, Rifai N, Ludwig DS (2004). Effects
of a low-glycemic load diet on resting energy expenditure
and heart disease risk factors during weight loss. JAMA 292,
2482–2490.
Perri MG, Nezu AM, Patti ET, McCann KL (1989). Effect of length of
treatment on weight loss. J Consult Clin Psychol 57, 450–452.
Phelan S, Wyatt HR, Hill JO, Wing RR (2006). Are the eating and
exercise habits of successful weight losers changing? Obesity (Silver
Spring) 14, 710–716.
Pittas AG, Das SK, Hajduk CL, Golden J, Saltzman E, Stark PC et al.
(2005). A low-glycemic load diet facilitates greater weight loss in
overweight adults with high insulin secretion but not in overweight
adults with low insulin secretion in the CALERIE Trial.
Diabetes Care 28, 2939–2941.
Pittler MH, Ernst E (2001). Guar gum for body weight reduction:
meta-analysis of randomized trials. Am J Med 110, 724–730.
S97
European Journal of Clinical Nutrition

S98
Carbohydrate intake and obesity
RM van Dam and JC Seidell
Popkin BM, Gordon-Larsen P (2004). The nutrition transition:
worldwide obesity dynamics and their determinants. Int J Obes
Relat Metab Disord 28 (Suppl 3), S2–S9.
Poppitt SD, Prentice AM (1996). Energy density and its role in the
control of food intake: evidence from metabolic and community
studies. Appetite 26, 153–174.
Poppitt SD, Keogh GF, Prentice AM, Williams DE, Sonnemans HM,
Valk EE et al. (2002). Long-term effects of ad libitum low-fat, highcarbohydrate
diets on body weight and serum lipids in overweight
subjects with metabolic syndrome. Am J Clin Nutr 75, 11–20.
Prentice AM, Jebb SA (2003). Fast foods, energy density and obesity: a
possible mechanistic link. Obes Rev 4, 187–194.
Raatz SK, Torkelson CJ, Redmon JB, Reck KP, Kwong CA, Swanson JE
et al. (2005). Reduced glycemic index and glycemic load diets
do not increase the effects of energy restriction on weight loss
and insulin sensitivity in obese men and women. J Nutr 135,
2387–2391.
Raben A (2002). Should obese patients be counselled to follow a lowglycaemic
index diet? No. Obes Rev 3, 245–256.
Raben A, Macdonald I, Astrup A (1997). Replacement of dietary fat by
sucrose or starch: effects on 14 d ad libitum energy intake, energy
expenditure and body weight in formerly obese and never-obese
subjects. Int J Obes Relat Metab Disord 21, 846–859.
Raben A, Vasilaras TH, Moller AC, Astrup A (2002). Sucrose compared
with artificial sweeteners: different effects on ad libitum food
intake and body weight after 10 wk of supplementation in
overweight subjects. Am J Clin Nutr 76, 721–729.
Rigaud D, Ryttig KR, Angel LA, Apfelbaum M (1990). Overweight
treated with energy restriction and a dietary fibre supplement: a
6-month randomized, double-blind, placebo-controlled trial. Int J
Obes 14, 763–769.
Rolls BJ, Bell EA, Thorwart ML (1999). Water incorporated into a food
but not served with a food decreases energy intake in lean women.
Am J Clin Nutr 70, 448–455.
Rolls BJ, Ello-Martin JA, Tohill BC (2004). What can intervention
studies tell us about the relationship between fruit and vegetable
consumption and weight management? Nutr Rev 62, 1–17.
Rolls BJ, Roe LS, Beach AM, Kris-Etherton PM (2005). Provision of
foods differing in energy density affects long-term weight loss.
Obes Res 13, 1052–1060.
Rossner S, von Zweigbergk D, Ohlin A, Ryttig K (1987). Weight
reduction with dietary fibre supplements. Results of two doubleblind
randomized studies.. Acta Med Scand 222, 83–88.
Ruxton CH (2003). Dietary guidelines for sugar: the need for
evidence. Br J Nutr 90, 245–247.
Ryttig KR, Tellnes G, Haegh L, Boe E, Fagerthun H (1989). A dietary
fibre supplement and weight maintenance after weight reduction:
a randomized, double-blind, placebo-controlled long-term trial.
Int J Obes 13, 165–171.
Samaha FF, Iqbal N, Seshadri P, Chicano KL, Daily DA, McGrory J
et al. (2003). A low-carbohydrate as compared with a low-fat diet
in severe obesity. N Engl J Med 348, 2074–2081.
Saris WH, Astrup A, Prentice AM, Zunft HJ, Formiguera X, Verboeketvan
de Venne WP et al. (2000). Randomized controlled trial of
changes in dietary carbohydrate/fat ratio and simple vs complex
carbohydrates on body weight and blood lipids: the CARMEN
study. The Carbohydrate Ratio Management in European National
diets. Int J Obes Relat Metab Disord 24, 1310–1318.
Schulze MB, Manson JE, Ludwig DS, Colditz GA, Stampfer MJ, Willett
WC et al. (2004). Sugar-sweetened beverages, weight gain, and
incidence of type 2 diabetes in young and middle-aged women.
JAMA 292, 927–934.
Schwartz MW, Porte Jr D (2005). Diabetes, obesity, and the brain.
Science 307, 375–379.
Schwartz MW, Boyko EJ, Kahn SE, Ravussin E, Bogardus C (1995).
Reduced insulin secretion: an independent predictor of body
weight gain. J Clin Endocrinol Metab 80, 1571–1576.
Seidell JC (1998). Dietary fat and obesity: an epidemiologic
perspective. Am J Clin Nutr 67, 546S–550S.
European Journal of Clinical Nutrition
Skov AR, Toubro S, Ronn B, Holm L, Astrup A (1999). Randomized
trial on protein vs carbohydrate in ad libitum fat reduced diet for
the treatment of obesity. Int J Obes Relat Metab Disord 23, 528–536.
Slabber M, Barnard HC, Kuyl JM, Dannhauser A, Schall R (1994).
Effects of a low-insulin-response, energy-restricted diet on weight
loss and plasma insulin concentrations in hyperinsulinemic obese
females. Am J Clin Nutr 60, 48–53.
Slavin JL (2005). Dietary fiber and body weight. Nutrition 21,
411–418.
Sloth B, Krog-Mikkelsen I, Flint A, Tetens I, Bjorck I, Vinoy S et al.
(2004). No difference in body weight decrease between a lowglycemic-index
and a high-glycemic-index diet but reduced LDL
cholesterol after 10-wk ad libitum intake of the low-glycemic-index
diet. Am J Clin Nutr 80, 337–347.
Smith JB, Niven BE, Mann JI (1996). The effect of reduced extrinsic
sucrose intake on plasma triglyceride levels. Eur J Clin Nutr 50,
498–504.
Smith-Warner SA, Elmer PJ, Tharp TM, Fosdick L, Randall B, Gross M
et al. (2000). Increasing vegetable and fruit intake: randomized
intervention and monitoring in an at-risk population. Cancer
Epidemiol Biomarkers Prev 9, 307–317.
Sorensen TI, Boutin P, Taylor MA, Larsen LH, Verdich C, Petersen L
et al. (2006). Genetic polymorphisms and weight loss in obesity: a
randomised trial of hypo-energetic high- versus low-fat diets. PLoS
Clin Trials 1, e12.
Spieth LE, Harnish JD, Lenders CM, Raezer LB, Pereira MA, Hangen SJ
et al. (2000). A low-glycemic index diet in the treatment of
pediatric obesity. Arch Pediatr Adolesc Med 154, 947–951.
Stern L, Iqbal N, Seshadri P, Chicano KL, Daily DA, McGrory J et al.
(2004). The effects of low-carbohydrate versus conventional
weight loss diets in severely obese adults: one-year follow-up of
a randomized trial. Ann Intern Med 140, 778–785.
Stookey JD (2001). Energy density, energy intake and weight status in
a large free-living sample of Chinese adults: exploring the
underlying roles of fat, protein, carbohydrate, fiber and water
intakes. Eur J Clin Nutr 55, 349–359.
Stubbs RJ, Whybrow S (2004). Energy density, diet composition and
palatability: influences on overall food energy intake in humans.
Physiol Behav 81, 755–764.
Stubbs RJ, Johnstone AM, O’Reilly LM, Barton K, Reid C (1998). The
effect of covertly manipulating the energy density of mixed diets
on ad libitum food intake in ‘pseudo free-living’ humans. Int J Obes
Relat Metab Disord 22, 980–987.
Teff KL, Elliott SS, Tschop M, Kieffer TJ, Rader D, Heiman M et al.
(2004). Dietary fructose reduces circulating insulin and leptin,
attenuates postprandial suppression of ghrelin, and increases
triglycerides in women. J Clin Endocrinol Metab 89, 2963–2972.
Thompson WG, Rostad Holdman N, Janzow DJ, Slezak JM, Morris
KL, Zemel MB (2005). Effect of energy-reduced diets high in dairy
products and fiber on weight loss in obese adults. Obes Res 13,
1344–1353.
Tordoff MG, Alleva AM (1990). Effect of drinking soda sweetened
with aspartame or high-fructose corn syrup on food intake and
body weight. Am J Clin Nutr 51, 963–969.
Toubro S, Astrup A (1997). Randomised comparison of diets for
maintaining obese subjects’ weight after major weight loss: ad lib,
low fat, high carbohydrate diet v fixed energy intake. BMJ 314, 29–34.
Tsihlias EB, Gibbs AL, McBurney MI, Wolever TM (2000). Comparison
of high- and low-glycemic-index breakfast cereals with
monounsaturated fat in the long-term dietary management of
type 2 diabetes. Am J Clin Nutr 72, 439–449.
Tuomilehto J, Lindstrom J, Eriksson JG, Valle TT, Hamalainen H,
Ilanne-Parikka P et al. (2001). Prevention of type 2 diabetes
mellitus by changes in lifestyle among subjects with impaired
glucose tolerance. N Engl J Med 344, 1343–1350.
Valdez R, Mitchell BD, Haffner SM, Hazuda HP, Morales PA,
Monterrosa A et al. (1994). Predictors of weight change in a biethnic
population. The San Antonio Heart Study. Int J Obes Relat
Metab Disord 18, 85–91.

Westerterp KR, Verboeket-van de Venne WP, Westerterp-Plantenga
MS, Velthuis-te Wierik EJ, de Graaf C, Weststrate JA (1996). Dietary
fat and body fat: an intervention study. Int J Obes Relat Metab
Disord 20, 1022–1026.
Weststrate JA, van het Hof KH, van den Berg H, Velthuis-te-Wierik EJ, de
Graaf C, Zimmermanns NJ et al. (1998). A comparison of the effect of
free access to reduced fat products or their full fat equivalents on
food intake, body weight, blood lipids and fat-soluble antioxidants
levels and haemostasis variables. Eur J Clin Nutr 52, 389–395.
Willett WC (2002). Dietary fat plays a major role in obesity: no. Obes
Rev 3, 59–68.
Willett WC, Dietz WH, Colditz GA (1999). Guidelines for healthy
weight. N Engl J Med 341, 427–434.
Wolever TM (2006). Physiological mechanisms and observed health
impacts related to the glycaemic index: some observations. Int J
Obes (London) 30 (Suppl 3), S72–S78.
Carbohydrate intake and obesity
RM van Dam and JC Seidell
Wolever TM, Mehling C (2003). Long-term effect of varying the
source or amount of dietary carbohydrate on postprandial plasma
glucose, insulin, triacylglycerol, and free fatty acid concentrations
in subjects with impaired glucose tolerance. Am J Clin Nutr 77,
612–621.
Wolever TM, Bentum-Williams A, Jenkins DJ (1995). Physiological
modulation of plasma free fatty acid concentrations by diet.
Metabolic implications in nondiabetic subjects. Diabetes Care 18,
962–970.
Yancy Jr WS, Olsen MK, Guyton JR, Bakst RP, Westman EC (2004). A
low-carbohydrate, ketogenic diet versus a low-fat diet to treat
obesity and hyperlipidemia: a randomized, controlled trial. Ann
Intern Med 140, 769–777.
Yeomans MR, Weinberg L, James S (2005). Effects of palatability and
learned satiety on energy density influences on breakfast intake in
humans. Physiol Behav 86, 487–499.
S99
European Journal of Clinical Nutrition

REVIEW
Dietary carbohydrate: relationship to cardiovascular
disease and disorders of carbohydrate metabolism
J Mann
European Journal of Clinical Nutrition (2007) 61 (Suppl 1), S100–S111
& 2007 Nature Publishing Group All rights reserved 0954-3007/07 $30.00
www.nature.com/ejcn
Department of Human Nutrition, Edgar National Centre for Diabetes Research, University of Otago, Dunedin, New Zealand
The nature of carbohydrate is of considerable importance when recommending diets intended to reduce the risk of type II
diabetes and cardiovascular disease and in the treatment of patients who already have established diseases. Intact fruits,
vegetables, legumes and wholegrains are the most appropriate sources of carbohydrate. Most are rich in nonstarch
polysaccharides (NSPs) (dietary fibre) and other potentially cardioprotective components. Many of these foods, especially those
that are high in dietary fibre, will reduce total and low-density lipoprotein cholesterol and help to improve glycaemic control in
those with diabetes. There is no good long-term evidence of benefit when NSPs or other components of wholegrains, fruits,
vegetables and legumes are added to functional and manufactured foods. Frequent consumption of low glycaemic index foods
has been reported to confer similar benefits, but it is not clear whether such benefits are independent of the dietary fibre content
of these foods or the fact that low glycaemic index foods tend to have intact plant cell walls. Furthermore, it is uncertain whether
functional and manufactured foods with a low glycaemic index confer the same long-term benefits as low glycaemic index
plant-based foods. A wide range of carbohydrate intake is acceptable, provided the nature of carbohydrate is appropriate.
Failure to emphasize the need for carbohydrate to be derived principally from wholegrain cereals, fruits, vegetables and legumes
may result in increased lipoprotein-mediated risk of cardiovascular disease, especially in overweight and obese individuals who
are insulin resistant.
European Journal of Clinical Nutrition (2007) 61 (Suppl 1), S100–S111; doi:10.1038/sj.ejcn.1602940
Keywords: cardiovascular disease; impaired carbohydrate metabolism
Dietary carbohydrate and cardiovascular disease
Traditional dietary patterns, which are high in carbohydrate,
are associated with low rates of coronary heart disease
(CHD). This appears to be the case regardless of the
carbohydrate containing primary staple, for example rice in
most Asian countries and a range of cereals, root crops and
pulses in different parts of Africa. However, cross-cultural
comparisons provide no indication as to whether the
percentage of energy intake derived from total carbohydrate
intake, total quantity of carbohydrate, particular classes of
carbohydrate and other nutrients, which are found in
carbohydrate-containing foods or method of food preparation,
account for the cardioprotection afforded by such
traditional carbohydrate-containing diets. Furthermore, it is
possible that cardiovascular risk is reduced simply because
Correspondence: Professor J Mann, Department of Human Nutrition, Edgar
National Centre for Diabetes Research, University of Otago, Dunedin, New
Zealand.
E-mail: jim.mann@stonebow.otago.ac.nz
traditional high carbohydrate diets are low in fat, especially
saturated fat, or because they promote satiety and thus
protect against overweight and obesity. Indeed, it is
conceivable that high carbohydrate diets simply act as a
marker for some other protective factor. Prospective epidemiological
studies and a range of experimental approaches
examining the effects of carbohydrates on cardiovascular risk
factors have attempted to clarify the role of total carbohydrate
and carbohydrate classes (sugars, oligosaccharides
and polysaccharides) and subgroups (for example starch,
nonstarch polysaccharide (NSP)) in CHD and stroke.
Prospective epidemiological studies
Several limitations are common to all prospective studies
examining the relationship between foods and nutrients and
disease risk. However, there are issues that are of particular
relevance when considering the role of carbohydrates. The
lack of consistency in the methods used for the measurement
of different classes and subgroups of carbohydrate,

especially the NSP, complicates comparisons between the
results of studies and their extrapolation into nutritional
recommendations. The term ‘dietary fibre’ is often incorrectly
regarded as being synonymous with NSP. However, in
this paper, the terms will be used as they occur in the
relevant literature, dietary fibre having been measured and
reported in most of the epidemiological studies and much of
the experimental research. In some countries, total carbohydrate
is measured ‘by difference’, after water fat, protein
and ash have all been measured. This is a less reliable
measure than that derived from adding the various classes of
carbohydrate measured individually (see accompanying
papers on ‘definition’ and ‘measurement’).
Of equal, or arguably greater, importance is the extent to
which sources of dietary carbohydrate have changed over
time or differ throughout the world. In some developing
countries, a high proportion of total sugars may be derived
from intact fruits and vegetables. The same may have been
true in the past for societies that are now relatively affluent.
However, nowadays, in most countries, sucrose, high
fructose corn syrup and other free sugars contribute a
substantial proportion of total sugars. Foods rich in free
sugars are usually energy dense and are often poor sources of
essential micronutrients (Baghurst et al., 1991), and are thus
likely to be associated with different physiological effects
and clinical consequences when compared with fruits and
vegetables. Thus, knowledge of total sugar content provides
little information relating to physiology. Similarly, until
relatively recently, dietary fibre was likely to be derived
principally from minimally processed cereals, vegetables,
fruits, legumes and ‘wholegrain’ cereal foods likely to
contain at least a reasonable proportion of intact or lightly
processed grains. Now, dietary fibre is often an added
ingredient and the definition of ‘wholegrain’ permits foods
to be described as such if they contain the constituents of the
grain without requiring the structure to be intact. Again,
the physiological effects of dietary fibre, when derived from
intact fruits, vegetables and grains, may differ from that
added to manufactured products. Thus, it may not be
appropriate to extrapolate the findings of epidemiological
studies involving the consumption of conventional foods to
carbohydrate-containing manufactured and functional
foods, which provide a substantial proportion of total energy
in most affluent and many developing counties.
Early prospective studies reporting the cardioprotective
effect of total carbohydrate intakes and intake of dietary fibre
were published in the 1970s and early 1980s (Morris et al.,
1977; Garcia-Palmieri et al., 1980; McGee et al., 1984). They
were relatively underpowered and confounding of results by
a range of factors certainly could not be excluded. During the
past 20 years, results have been published from a substantial
number of prospective studies involving cohorts of sufficient
size to enable examination of potentially confounding
factors. There has been particular emphasis on the cardioprotective
role of cereal grains and dietary fibre. Most studies
define wholegrain as either intact or milled grain with bran,
Cardiovascular disease and diabetes
J Mann
germ and endosperm in the same proportion as the unmilled
grain. Wholegrain foods have generally been arbitrarily
defined as those foods with more than 25% wholegrain or
bran content by weight (Flight and Clifton, 2006). The
studies have been systematically reviewed (for example Liu,
2002; Truswell, 2002; Hu, 2003; Slavin, 2003; Flight and
Clifton, 2006) and meta-analysed (Anderson, 2002, 2003).
Despite the use of different instruments for assessing dietary
intake, the results of the studies show a remarkably
consistent trend. Meta-analysis, involving four of the largest
published studies, suggested a 28% reduction in risk of CHD
when comparing individuals in the highest and lowest
quintiles of intake of wholegrains (relative risk 0.72, 95%
confidence intervals: 0.48, 0.94) (Anderson 2003). The size of
the most recent cohort studies has enabled analyses to
examine the extent to which these findings might be
confounded by other cardioprotective factors. While residual
confounding cannot be excluded with absolute certainty,
such analyses provide reasonable evidence that wholegrains
protect against CHD regardless of the degree of adiposity and
other measured variables associated with healthy lifestyles
and eating habits. However, it is acknowledged that people
who consistently eat wholegrain breads do indeed have
different lifestyles from those who do not, and it may not be
possible in epidemiological studies to measure all attributes
of lifestyle.
The possibility that an association is truly causal is
strengthened by the existence of a dose response effect.
Such an effect has been demonstrated in the Iowa Women’s
Health Study. After adjustment for cardiovascular risk
factors, relative risks for cardiovascular disease were 1.0,
0.96, 0.71, 0.64 and 0.70 in ascending quintiles of wholegrain
intake, P for trend ¼ 0.02 (Jacobs et al., 1998). However,
the epidemiological studies do not provide an indication of
whether all grains are equal in this respect. There is no clear
evidence as to which constituent of the grain provides
cardioprotection or as to whether the structure of the grain
needs to be wholly or partially intact.
Relatively few groups have examined the extent to which
wholegrain intakes influence the risk of stroke. Reporting on
the largest of the prospective studies, the Nurses’ Health
Study, Liu et al. (2000a) observed risk reductions of
magnitude similar to those observed for CHD (relative risk
0.69, 95% confidence intervals: 0.50, 0.98 when comparing
highest relative to lowest quintile of intake of wholegrains).
Similar risk reductions have been observed in three other
somewhat smaller cohorts.
Wholegrains differ in a number of respects from highly
refined cereals, and several of the constituents may explain
their cardioprotective effects (Table 1). Dietary fibre has been
studied more than the other constituents. Several groups of
researchers have pooled the results from a number of studies.
For example, Pereira et al. (2004) reported a relative risk of
0.90 (95% confidence intervals: 0.77, 1.07 that is not
statistically significant) for total CHD events for each
10 g/day increase in cereal fibre. When considering CHD
S101
European Journal of Clinical Nutrition

S102
Cardiovascular disease and diabetes
J Mann
Table 1 Constituents of wholegrains which may confer cardioprotection
deaths, the relative risk, 0.75 (95% confidence intervals:
0.63, 0.91) was statistically significant, the association being
independent of a number of dietary factors and other
cardiovascular risk factors.
However, other analyses suggest that while cereal fibre is
inversely associated with CHD rates, the relationship is not
as strong as the inverse relationship between CHD and
wholegrain bread (Anderson et al., 2000). Furthermore, the
effect of dietary fibre may be largely explained by fibre
derived from wholewheat, rye or pumpernickel breads
(Mozaffarian et al., 2003). A single study suggested risk
reduction among those with the highest intake of added
bran compared with those who consumed no added bran
(Jensen et al., 2004). The suggestion that reduced cardiovascular
risk principally results from consumption of the
wholegrain rather than dietary fibre is supported by findings
from the Iowa Women’s Health Study. CHD rates were
compared in women consuming similar amounts of cereal
fibre from either predominantly refined grain sources or
predominantly wholegrains. After adjustments, all cause
mortality was significantly lower, and CHD appreciably
(though not statistically significantly) reduced among the
latter group (Jacobs et al., 2000). The absence of a universally
adopted definition of ‘wholegrain’, and the even more
inconsistent use of the terms ‘wholemeal’ and ‘wholewheat’
preclude more definitive conclusions and complicate extrapolation
into guidelines.
Fruits and vegetables are important sources of carbohydrate
and also contain a range of potentially cardioprotective
components, including dietary fibre, folate,
potassium, flavonoids and antioxidant vitamins. Pooled
analyses of the Nurses’ Health Study and the Health
Professionals’ Follow-up Study suggested risk reductions of
20 and 30% for CHD and ischaemic stroke, respectively,
when comparing extreme quintiles of fruit and vegetable
intake (Joshipura et al., 2001). Similar findings were reported
in a recent meta-analysis (He et al., 2006). A dose–response
effect was reported with lowest risk for CHD being associated
with eight servings or more per day and for stroke, four or
more servings daily. It is noteworthy that in this, the largest
study reported to date, individual items relatively low in
total carbohydrate appear to have the most striking protective
effect—green leafy vegetables in the case of CHD and
cruciferous vegetables, citrus fruits and juices and vitamin C-rich
Antioxidant Lipid-lowering Enzyme modulator
Dietary fibre (nonstarch polysaccharide) O
Unsaturated cis: n–3 and n–6 O
Folate O
Vitamin E O
Selenium O
Flavenoids O
Phytooestrogens O
European Journal of Clinical Nutrition
fruits and vegetables in the case of stroke. In the Physicians’
Health Study, Liu et al. (2001) reported an inverse association
between carotenoid rich vegetables and CHD risk, and
Finnish data suggested that flavonoid-rich foods (for example
berries, onions, apples) were especially beneficial in this
regard (Knekt et al., 1994). Legume consumption four or
more times per week compared with less than once per week
has also been associated with cardiovascular risk reduction
(Bazzano et al., 2001).
During the 1960s, Yudkin (1964) and co-workers reported
strong associations between high sugar intakes and CHD.
The data were cross-sectional and flawed (Keys, 1971;
Truswell, 1987). The Iowa Women’s Health Study found no
relationship between intake of sweets and desserts and CHD
in over 30 000 women followed for 9 years (Jacobs et al.,
1998). The glycaemic index (GI; and more recently the
glycaemic load, GL) has provided a means of determining
the extent to which carbohydrate-containing foods may
determine cardiovascular risk through their potential to raise
blood glucose. Liu et al. (2000b) reported from the Nurses’
Health Study that women who consumed diets with a highglycaemic
load (that is high in rapidly digested starches)
were at increased risk of CHD compared with those with a
lower consumption; a twofold increase in risk over a 10-year
follow-up was observed when comparing those in the
highest and lowest quintiles of intake. The effect appeared
to be independent of total energy intake and other
cardiovascular risk factors.
The potential of dietary patterns, rather than the effect of
individual foods and nutrients on human health, have been
examined in two of the largest cohort studies—the Nurses’
Health Study and the Health Professionals’ Study (Hu et al.,
2000; Fung et al., 2001b). Factor analysis was used to
examine the association between CHD and two major
dietary patterns identified: ‘western’ and ‘prudent’. The
prudent pattern was characterized by higher intakes of
fruits, vegetables, legumes, fish, poultry and wholegrains
and the western pattern, by higher intakes of red and
processed meats, sweets and desserts, French fries
and refined grains. After adjustment for cardiovascular risk
factors, the prudent diet score was associated with relative
risks of 1.0, 0.87, 0.79, 0.75 and 0.70 from the lowest to the
highest quintiles. Conversely, the relative risks across
increasing quintiles of the western pattern score were 1.0,

1.21, 1.35, 1.40 and 1.64. The patterns were also related to
biochemical markers of CHD. Positive correlations were
noted between the western pattern score and plasminogen
activator antigen, fasting insulin, C-peptide, leptin, C-reactive
protein and homocysteine after adjustment for confounders.
A significant inverse correlation was observed between
plasma folate and the western dietary pattern score. Inverse
correlations were observed between the prudent pattern
score and fasting insulin and homocysteine, and a positive
correlation was observed with folate (Fung et al., 2001a).
In summary, prospective epidemiological studies provide
strong evidence for a protective role of wholegrain cereals,
fruits and vegetables and dietary patterns characterized by
relatively high intakes of such foods. While a high
consumption of dietary fibre derived from cereals is also
associated with a reduced risk of cardiovascular disease, it is
not clear whether the cardioprotection can entirely be
attributed to the polysaccharides per se. The effect seems to
stem largely from dietary fibre in dark breads and may
therefore be at least partially attributable to other constituents
of wholegrains. It is not possible to determine the extent
to which other potentially protective constituents of wholegrains
are relevant. Epidemiological studies do not permit
separating the effects of different wholegrains. Even the best
prospective studies cannot conclusively eliminate the possibility
of residual confounding, so recommendations regarding
the intake of carbohydrates in relation to cardiovascular
disease also depend on the intervention studies described
below.
Carbohydrates and cardiovascular risk factors
Lipids and lipoproteins
The effect of carbohydrates on lipids and lipoproteins has
dominated discussions regarding the amounts and classes of
carbohydrate likely to reduce cardiovascular risk. While
there is no doubt that increasing total carbohydrate at the
expense of fat, especially saturated and trans fatty acids in
the Western diet, will result in reduction of total and lowdensity
lipoprotein (LDL) cholesterol, concern has been
expressed that other predictors of lipoprotein-mediated
cardiovascular risk may be adversely affected by substantial
increases in total carbohydrate. The potential of
high carbohydrates to increase fasting triglycerides was first
demonstrated in the 1960s (Ahrens et al., 1961) but later
considered to be a transient phenomenon, the hypertriglyceridaemia
diminishing with prolonged exposure to a high
carbohydrate intake (Antonis and Bersohn, 1961; Stone and
Connor, 1963). However, more recent and more sophisticated
studies suggest that, at least in certain circumstances, a
low-fat, high carbohydrate diet may be associated with
slightly elevated triglyceride levels for as long as 2 years
(Retzlaff et al., 1995). Furthermore, such a diet may result in
the appearance of small dense LDL particles that are
particularly atherogenic and in an adverse ratio of total to
Cardiovascular disease and diabetes
J Mann
high-density lipoprotein (HDL) cholesterol considered to be
a more specific marker of CHD than total LDL cholesterol
(Kinosian et al., 1995).
The effect of carbohydrate as replacement energy for fatty
acids has been examined in a fairly recent meta-analysis
involving 60 mostly short-term controlled trials (Mensink
et al., 2003). The results appeared to offer a strong message:
regardless of the source of the fat displaced, an increase in
total carbohydrate was associated with an increased ratio of
total to HDL cholesterol. However, although these studies
paid careful attention to ensure that the comparisons were
carried out under isoenergetic conditions and food intake
was thoroughly controlled, there was less consistency with
regard to the source and class of carbohydrate that was used
as fat replacement. A fairly substantial body of evidence
suggests that the type of carbohydrate influences lipid and
lipoprotein levels. The effect of sugars on triglyceride levels
differs from that of polysaccharides. Compared with starches
and NSPs, sugars, especially sucrose and fructose, may be
associated with appreciable increases in triglyceride levels,
especially when the sugars are consumed in the context of a
diet high in total carbohydrate or when dietary fat comprises
principally of saturated fatty acids (Truswell, 1994). The
effect appears to be more marked in insulin resistant
individuals with abdominal obesity (Fried and Rao, 2003).
Unfortunately, few long-term data are available. However,
the CARMEN study suggested that after 6 months, a very
modest weight loss could mitigate the hypertriglyceridaemic
effect of a moderate increase in sucrose (Saris et al., 2000).
This has also been observed for diets high in total
carbohydrate and starches, which might otherwise have
been associated with hypertriglyceridaemia when compared
with diets lower in total carbohydrates (Kasim-Karakas et al.,
2000).
Dietary fibre has a potentially important effect on lipids
and lipoproteins when consumed in plant foods or as
supplements Viscous subgroups, including pectins, b-glucans,
glucomannans, guar and psyllium, which are generally water
soluble, all lower total and LDL cholesterol between 5 and
10 g/day, lowering LDL by about 5% (Truswell, 2002). The
mechanism appears to be by reducing ileal bile acid
absorption (Morgan et al., 1993). These soluble forms of
dietary fibre appear to have a negligible effect on triglyceride
or HDL levels. Insoluble dietary fibre subgroups are derived
largely from cereal sources and have little or no effect on
lipids and lipoproteins (Jenkins et al., 2000). A Cochrane
Review has been undertaken of randomized controlled trials
(RCTs) investigating the effects of wholegrain cereals on
CHD and cardiovascular risk factors. Only 10 studies were
identified that met the strict criteria required for such a
review. None had clinical end points. They compared
wholegrain foods or diets high in wholegrain foods with
other foods or diets with lower levels or no wholegrains.
Eight of the ten studies evaluated wholegrain oats and
clearly demonstrated the potential of such foods to reduce
total LDL cholesterol. Levels were about 0.2 mmol/l lower
S103
European Journal of Clinical Nutrition

S104
on the diet rich in oats compared with the control diet.
No changes were observed in triglyceride or HDL cholesterol
levels. Lack of appropriate studies precluded conclusions
on wholegrains other than oats (Brunner et al., 2005).
A Cochrane Review has also been undertaken to explore
the extent to which low glycaemic index diets might
influence CHD and risk factors. The average reduction in
total cholesterol when comparing low and high GI diets was
0.17 mmol/l, P ¼ 0.03 (95% CI 0.32 to 0.02). No convincing
differences were apparent in LDL cholesterol, HDL
cholesterol or triglyceride levels (Kelly et al., 2004).
In studies of free living individuals in whom fruits,
vegetables and legumes rich in the viscous forms of NSPs
replaced some of the relatively high fat foods typically
consumed in a western diet, total and LDL cholesterol fell as
expected, and the ratio of total (or LDL) cholesterol to HDL
cholesterol improved with no change reported in triglyceride
despite an appreciable increase in total carbohydrate (Turley
et al., 1998). Total carbohydrate provided 59 and 43% total
energy on the high and low carbohydrate diets. While the
aim of these studies was not to achieve weight loss, small
reductions in body weight did occur on the high carbohydrate
diets, perhaps as a result of the enhanced satiety
associated with the more bulky, less energy dense foods.
This, together with the effects of dietary fibre, are likely to
have contributed to the less atherogenic lipid profile
observed on the high carbohydrate diet. Such findings
provide a clear indication of the need to consider the nature
of dietary carbohydrate when carbohydrate rich foods are
recommended as replacement for dietary saturated fatty
acids.
Type II diabetes, impaired carbohydrate metabolism and insulin
resistance
See later section, which deals more fully with this topic.
Inflammatory markers
There has been considerable recent interest in the role of
inflammation as a determinant of cardiovascular risk. Much
of the research relating to the role of nutritional determinants
has centred around different dietary fats (Basu et al.,
2006). Jenkins et al. (2003) reported reduced C-reactive
protein levels in hyperlipidaemic patients consuming a high
carbohydrate diet rich in viscous fibre-containing foods.
However, the diets were also high in nuts (almonds), plant
sterols and soy proteins, and it is therefore impossible to
disentangle separate effects. A very recent study (Kasim-
Karakas et al., 2006) found that when carbohydrate replaced
a substantial proportion of dietary fat under eucaloric
conditions, the levels of several inflammatory markers
increased along with an increase in triglyceride. However,
when the participants (post-menopausal women) consumed
the 15% fat diet ad libitum under free living conditions, they
European Journal of Clinical Nutrition
Cardiovascular disease and diabetes
J Mann
lost weight and triglyceride and the levels of inflammatory
markers decreased.
Hypertension
Several ‘high carbohydrate’ dietary approaches have been
shown to be associated with reduced levels of blood pressure.
However, these have invariably involved many dietary
changes. For example, the Dietary Approaches to Stop
Hypertension approaches have involved an increase in
intakes of fruits and vegetables, low fat dairy products and
a reduction in sodium. Thus, it has been impossible to
disentangle individual effects (Appel et al., 1997).
RCT involving clinical end points
A single RCT has examined the effect of altering dietary
carbohydrate intakes on clinical end points. Burr et al. (1989)
randomized 2033 men who had survived a myocardial
infarction to receive one of eight possible combinations of
dietary advice. Those who were advised to increase cereal
fibre had a similar rate of CHD recurrences and deaths as
those who did not receive this advice when considering a
10-year follow-up period. However, the study was underpowered
and the degree of compliance is uncertain (Ness
et al., 2002). Other RCTs, which involved an increase in
dietary carbohydrate, from fruits, vegetables, legumes and
wholegrain, as part of a multifactorial dietary approach to
primary and secondary prevention of CHD, demonstrated
reduction in clinical events among those receiving intensive
dietary advice compared with control groups (Hjermann
et al., 1981; de Lorgeril et al., 1994). The Women’s Health
Initiative Randomised Control for Dietary Modification Trial
involved the intervention group being recommended a high
carbohydrate low fat diet. Failure to demonstrate significant
differences in CHD rates in control and intervention groups
contributed little to the understanding of the role of
carbohydrates since long-term compliance with the intervention
diet was poor (Howard et al., 2006).
Dietary carbohydrate and cardiovascular disease:
recommendations
The joint WHO/FAO Expert Consultation on Diet, Nutrition
and the Prevention of Chronic Disease (WHO Technical
Report Series 916, 2003) considered that the evidence
suggesting a cardio-protective effect of fruits and vegetables
was ‘convincing’ and that relating to wholegrain cereals and
NSP was ‘probable’. The evidence that high intakes of total
carbohydrate might increase the risk of cardiovascular
diseases was considered to be ‘insufficient’ (Table 10, p88
TRS 916). Difficulties with regard to the definition of
wholegrains and inability to exclude the possibility of

esidual confounding provide the justification for regarding
the protective effect being graded as ‘probable’ rather than
‘convincing’. The Expert Consultation recommended that
total carbohydrate should provide 55–75% total energy and
that free sugars should provide less than 10%. Recommended
intake of fruits and vegetables was 400 or more g per
day, excluding tubers (that is potatoes, cassava). Precise
amounts of NSPs or dietary fibre were not recommended.
However, it was considered that appropriate intakes of fruits,
vegetables, legumes and regular consumption of wholegrain
cereals would provide in excess of 20 g/day of NSP and over
25 g of total dietary fibre. There was also no recommendation
regarding glycaemic index. These recommendations
would appear to be generally compatible with this update of
the relevant literature with the following caveats:
Many western countries have average intakes of carbohydrate,
which are below 55% total energy. There is no
good evidence that total carbohydrate intake for populations
or individuals must reach the lower recommended
intake in order to achieve a cardioprotective dietary
pattern provided guidelines relating to dietary fat and
other nutrients are met. Several national guidelines
suggest a lower limit of 50% total energy.
A wide range of carbohydrate intakes is compatible with
cardioprotection. However, it is important to be prescriptive
with regard to the nature of carbohydrate, especially
when total carbohydrate intakes are at the upper end of
the recommended range. Failure to emphasize the need
for carbohydrate to be derived principally from wholegrain
cereals, fruits, vegetables and legumes may result in
increased lipoprotein-mediated risk of CHD associated
with an increase in the ratio of total (or LDL) cholesterol
to HDL cholesterol and an increase in triglyceride. This
may apply particularly to overweight and obese individuals
who are insulin resistant (see separate section on
diabetes and insulin resistance).
While fruits and legumes rich in viscous (soluble) forms of
NSP and dietary fibre are associated with reduced LDL
cholesterol, insoluble forms derived from wholegrain
cereals also appear to confer cardioprotection possibly in
association with other constituents of wholegrains. There
is no convincing evidence that fibre supplements and
manufactured and functional foods containing them
reduce cardiovascular risk.
Dietary carbohydrate and diabetes, impaired
carbohydrate metabolism and the metabolic
syndrome
The escalating rates of type II diabetes worldwide and the
realization that insulin resistance and its associated metabolic
abnormalities contribute increasingly to the global epidemic
of cardiovascular disease (Mann, 2000) have resulted in
renewed interest in the nutritional determinants of these
Cardiovascular disease and diabetes
J Mann
metabolic derangements. In the 1960s, Yudkin suggested that
sugar (sucrose) was a major determinant of type II diabetes
(then known as ‘maturity onset’ and later as ‘non-insulin
dependent’ diabetes). The conclusions were based largely on
highly selected, within country and cross country, comparisons
(Yudkin, 1964; Nuttall and Gannon, 1981) later
acknowledged to be totally inappropriate analyses (Nuttall
et al., 1985). At about the same time, Trowell, then working in
Uganda, suggested that the low rates of diabetes in rural Africa
were likely to result from a protective effect of large intakes of
dietary fibre found in the minimally processed or unprocessed
carbohydrate, which provided a high proportion of total
dietary energy in such societies (Trowell, 1975). Although it
was not possible to know the extent to which these
observations implied true causation or protection or simply
represented confounding, they stimulated much subsequent
epidemiological and experimental research.
Prospective epidemiological studies
A number of limitations apply to all prospective studies that
have examined the association between foods and nutrients
and the risk of various diseases. The issues that are
potentially relevant to dietary carbohydrate have been
considered in the previous section on cardiovascular disease,
and they apply equally when relating various carbohydrates
to risk of developing diabetes. Recent prospective studies
provide some insight.
In the Nurses’ Health Study, there was no association
between total intake of grains and type II diabetes. However,
wholegrain consumption appeared to be protective. When
comparing the highest and lowest quintiles of intake, age
and energy adjusted relative risk was 0.62 (95% CI: 0.53,
0.71). Further adjustment for other risk factors did not
appreciably alter this risk estimate (Liu et al., 2000). A
virtually identical risk reduction was observed among men
participating in the Health Professionals Study (Fung et al.,
2002). A relative risk of 0.63 was reported in association with
three or more servings per day of wholegrains. In the Iowa
Women’s Health Study, postmenopausal women in the
upper quintile of wholegrain consumption (more than 33
servings per week) were 20% less likely to develop type II
diabetes than those in the lowest quintile (fewer than 13
servings per week) (Meyer et al., 2000). Wholegrains have
also been shown to be protective against type II diabetes in
African-Americans (van Dam et al., 2006), Finns (Montonen
et al., 2003) and Iranians (Esmaillzadeh et al., 2005). Cereal
fibre, as distinct from fibre derived from other sources,
appears to be associated with a protective dose–response
effect that is present after controlling for a range of
potentially confounding factors. In the Nurses’ Health Study,
multivariate relative risks for quintiles 1–5 were 1.0, 0.85,
0.87, 0.82 and 0.64, respectively (Salmeron et al., 1997a). The
role of wholegrains and dietary fibre in diabetes has been
reviewed in detail in Venn and Mann (2004).
S105
European Journal of Clinical Nutrition

S106
In the Nurses’ Health Study and the Health Professionals
Follow-up Study, a statistically significant relative risk of
about 1.4 was observed when comparing diabetes rates in the
highest and lowest quintiles of dietary glycaemic load, after
adjusting for potentially confounding factors (Salmeron
et al., 1997a, b). A high dietary glycaemic index was also
found to be associated with an increased risk of type II
diabetes in the Melbourne Collaboration Cohort Study
(Hodge et al., 2004).
Rapidly digested starches and sugars, especially glucose,
contribute to the glycaemic load of the diet, and thus such
observations provide some evidence for a promotive role
of these carbohydrates in type II diabetes. However,
no association was found between sucrose intake and risk
of diabetes during a 6-year follow-up of the Nurses’ Health
Study (Colditz et al., 1992). Meyer et al. (2000), reporting the
findings from the Iowa Women’s Health Study, found no
relationship between type II diabetes and either glycaemic
index or glycaemic load. However, dietary glucose and
fructose (but not sucrose) were associated with increased risk.
Women who developed gestational diabetes in the Nurses’
Health Study had lower intakes of dietary fibre and higher
dietary glycaemic load than those who did not (Zhang et al.,
2006).
No prospective studies have examined the relationship
between dietary carbohydrate and insulin resistance,
considered to be the underlying abnormality in most
cases of type II diabetes. However, two large cross-sectional
studies, using validated food frequency questionnaires
to assess nutrient intake and either the frequently
sampled intravenous glucose tolerance test or homoeostasis
model assessment for insulin resistance, found that
intake of dietary fibre was inversely associated with the
probability of having insulin resistance (Lau et al., 2005;
Liese et al., 2005). In the Insulin Resistance Atheroscelerosis
Study (Liese et al., 2005), it was possible to
demonstrate that fibre was associated with increased
insulin sensitivity even after adjustment for body mass
index. Interestingly, neither study found any relationship
between insulin sensitivity and glycaemic index or
glycaemic load.
While diets rich in wholegrains and dietary fibre may
protect against diabetes and pre-diabetic states by virtue of
their potential to promote satiety and weight loss in those
who are overweight or obese (excess adiposity being the
major determinant of these metabolic abnormalities),
the epidemiological data do provide evidence for a protective
effect independent of that on fat mass. The data do not
permit distinction among different grains nor the extent to
which structure of the grain or its constituents explain the
protective effect. Several studies suggest that diets with a
relatively low glycaemic index or glycaemic load are
protective against disorders of carbohydrate metabolism,
but the data are insufficient to establish with certainty the
extent to which this is independent of other attributes of
carbohydrates.
European Journal of Clinical Nutrition
Cardiovascular disease and diabetes
J Mann
Experimental studies
Most studies that have compared the effects of carbohydrates
and fats and the effects of different carbohydrates have been
undertaken in people with established diabetes; however,
some information is available concerning people with
insulin resistance and the metabolic syndrome. Insulin
resistance and its associated dyslipidaemia (high triglyceride,
low HDL) have been shown to be more marked on a high
carbohydrate diet when compared with a diet rich in
monounsaturated fatty acids. However, while such studies
may distinguish between sugars and polysaccharides, they
often do not distinguish between starches, many of which
are rapidly digested and absorbed, and NSPs (Mann, 2001).
A study by Pereira et al. (2002) clearly demonstrates the
importance of doing so. In a controlled 6-week study,
overweight hyperinsulinaemic adults consumed diets providing
55% energy from carbohydrate and 30% from fat.
Carbohydrates were derived from predominantly wholegrain
or refined grain cereals, with dietary fibre content of the
wholegrain diet 28 g compared with 17 g on the refined grain
diet. Dietary fibre was predominantly from cereal sources.
Total carbohydrate and fat and fat sources were virtually
identical on the two diets. Insulin sensitivity measured by a
euglycaemic hyperinsulinaemic clamp was appreciably improved
on the wholegrain compared with the refined grain
diet. Fasting insulins and area under the 2-h insulin curve
were lower, despite body weight not being significantly
different on the two diets. Rye bread (Juntunen et al., 2003)
and purified cereal fibre (Weickert et al., 2006) have been
shown to improve insulin sensitivity in overweight and
obese women.
Analysis of a subgroup of individuals within the CARMEN
Study, who were diagnosed with the metabolic syndrome,
showed that the reduction in body weight and improvement
in metabolic indices seen when simple carbohydrate (sugars)
were replaced with complex ones were more striking in
people with this disorder than in the general population
(Poppitt et al., 2002). Studies by McAuley et al. (2005; 2006)
have compared different nutritional approaches in overweight
insulin-resistant women. Although a low carbohydrate
high fat diet resulted in initial weight loss and
improvement in the metabolic derangements associated
with insulin resistance, the improvement was not sustained.
A diet relatively high in protein and unsaturated fat sources
and with moderate amounts of fibre-rich carbohydrate, and
one low in total fat and relatively high in fibre-rich
carbohydrate produced sustained weight loss and improvement
in a range of metabolic measurements over a period of
a year. Thus, while weight loss in those who are overweight
or obese is the cornerstone of management aimed at
improving insulin sensitivity, it appears that this can be
achieved by diets relatively high or relatively low in
carbohydrates, provided the carbohydrate sources are
wholegrain and rich in dietary fibre. Other studies that have
compared the effects of diets of varying amounts of

carbohydrate have generated broadly similar results (Nordmann
et al., 2006).
The fact that fructose, compared with glucose, is preferentially
metabolized to lipid in the liver and that fructose
consumption induces insulin resistance, impaired glucose
tolerance, hyperinsulinaemia, hypertriglyceridaemia and
hypertension in animal models has led to the suggestion
that fructose, sucrose or high fructose corn syrup may have
uniquely untoward effects compared with other carbohydrates
in humans when fed in the context of energy balance.
However, while these sugars may contribute to increasing
the risk of overweight and obesity and thus to the
clinical and metabolic disorders (Mann, 2004), the evidence
regarding fructose remains limited and conflicting, and no
conclusions can be drawn (Elliott et al., 2002; Daly, 2003).
A large number of studies have been undertaken to
determine the extent to which total quantity and nature of
dietary carbohydrate influence glycaemic control, insulin
resistance and cardiovascular risk factors in people with type
II diabetes. While such studies apply principally to diabetes
management, it seems reasonable to assume that the data
also have some relevance to disease prevention.
A meta-analysis of nine studies has been undertaken to
compare the effects of high carbohydrate and high monounsaturated
fatty acid diets. The overall results provide
evidence of higher fasting triglycerides and VLDL cholesterol,
slightly lower HDL cholesterol, slightly higher glucose,
but not glycated haemoglobin on the high carbohydrate
diets, typically providing 55–60% total energy from carbohydrate
(Garg, 1998). There was considerable variation in the
results of the individual studies, partly due to the fact that
they were all relatively underpowered and also because the
severity of diabetes may influence response to dietary
carbohydrate. However, equally important is the fact that
the nature of dietary carbohydrate was not clearly specified.
It appears that a range of ‘complex carbohydrates’ provided a
substantial proportion of total carbohydrate on the high
carbohydrate diets. It seems that there may have been a lack
of appreciation regarding the different effects of starches and
NSPs and indeed the different effects of NSPs from different
sources. This is surprising, given that in the late 1980s, it had
been clearly shown that high carbohydrate diets were only
associated with beneficial effects in terms of glycaemic
control and blood lipids when carbohydrate was derived
principally from foods rich in soluble forms of dietary fibre,
notably pulses, legumes and intact fruits and vegetables
(Rivellese et al., 1980; Simpson et al., 1981; Riccardi et al.,
1984; Mann, 2001). Higher intakes of cereal (insoluble) fibre
were associated with less marked or no benefit in glucose or
lipid levels (Simpson et al., 1979; 1982; Riccardi et al., 1984).
The beneficial effects of soluble forms of dietary fibre derived
from fruits and vegetables appear to have been rediscovered
more recently in randomized crossover studies in type II
diabetes (Chandalia et al., 2000). Dietary fibre has also been
shown to be of benefit in type I diabetes (Giacco et al., 2000).
Cross-sectional epidemiological data, based on the
Cardiovascular disease and diabetes
J Mann
EURODIAB Complications Study, which included over
2000 patients in 31 European centres, showed an inverse
association between dietary fibre intake of HbA1c and LDL
cholesterol (in men only) and a positive association with
HDL cholesterol in men and women (Toeller et al., 1999). An
RCT (parallel design) in type I patients continued for 6
months, confirmed the potential for around 40 g/day dietary
fibre (half of the soluble type from legumes, fruits and
vegetables) to improve glycaemic control (Giacco et al.,
2000). Some studies have shown improved glycaemic control
when soluble fibres have been fed as supplements to patients
with diabetes. However, evidence-based nutritional guidelines
of the Nutrition Study Group of the European
Association for the Study of Diabetes offer no recommendations
regarding supplements of functional foods providing
increased dietary fibre. Long-term clinical trials are deemed
necessary before recommendations can be considered (Mann
et al., 2004).
In the 1980s, several RCT with crossover designs demonstrated
no adverse effects on glycaemic control, lipids and
lipoproteins when diets containing small amounts of sucrose
(usually around 50 g) were compared under eucaloric conditions
with virtually sucrose-free diets in type I and type II
diabetes (Slama et al., 1984; Peterson et al., 1986; Mann, 1987).
While there is thus clear evidence for the acceptability of
moderate intakes of sucrose for most people with diabetes,
there are few data from which to derive acceptable upper
limits. For those who are overweight or obese or are markedly
insulin resistant, fairly severe restriction may be appropriate.
For all others, moderate intakes of free sugars (up to 50 g/day)
are acceptable, provided that free sugars do not exceed 10%
total energy. There are suggestions that sugar in beverages
increases body weight to a greater extent than the same
amount of sugar in the solid form (diMeglio and Mattes, 2000).
A number of randomized trials have examined the extent
to which diets with a low glycaemic index can improve
glycaemic control and cardiovascular risk factors. Several
meta-analyses have been published (Brand-Miller et al., 2003;
Opperman et al., 2004) and the results are broadly comparable.
For example, Opperman et al. reported a reduction in
HbA1c of 0.27 (95% confidence intervals: 0.5 to 0.03)
when comparing high and low glycaemic index diets. This
issue was examined in a Cochrane Systematic Review, which
suggested some evidence of an effect of low GI diets
compared with high GI diets after 12 weeks. Pooled analysis
of studies with a parallel design demonstrated mean HbA1c
levels 0.45% less than the high GI diets (95% confidence
intervals: 0.82 to 0.09, P ¼ 0.02). It is of importance to
note that the low GI foods utilized in most of the studies
were plant based (for example peas, lentils, beans, pasta,
barley, parboiled rice, oats and cereals) as were the high
GI foods (potatoes, wheatmeal and white bread and high GI
breakfast cereals). The studies did not generally include
many of the now widely available manufactured and
functional foods, which are not necessarily predominantly
plant based.
S107
European Journal of Clinical Nutrition

S108
RCT with clinical end points
Two RCTs involving people with impaired glucose tolerance
consuming ‘western’ diets have examined the extent to
which progression of impaired glucose tolerance to type II
diabetes can be reduced by intensive lifestyle modification
(Tuomilehto et al., 2001; Knowler et al., 2002). In both the
Finnish and US intervention studies, advice regarding
frequent consumption of wholegrain products, vegetables
and fruits has been a pivotal component of the advice to
achieve weight loss. Other components included recommendations
to use monounsaturated oils, soft margarines and
low fat meat and milk products, as well as to increase
physical activity of moderate intensity. These interventions
achieved a nearly 60% reduction of progression from
impaired glucose tolerance to type II diabetes. While it is
impossible to accurately disentangle the separate contributions
of the different components of the intervention
package, and weight loss was clearly the major determinant
of risk reduction, increased intake of dietary fibre appeared
to be independently associated with reduced risk of progression
(Lindstrom et al., 2003). A similar study has been
conducted in India (Ramachandran et al., 2006).
Dietary carbohydrate and diabetes:
recommendations
The Joint WHO/FAO Expert Consultation on Diet, Nutrition
and the Prevention of Chronic Disease (WHO Technical
Report Series 916, 2003) describes the protective effect of NSP
as ‘probable’. However, an explanatory note indicates that
the level of evidence is graded as ‘probable’ rather than
‘convincing’ because of the apparent discrepancy between
the experimental studies, in which it appears that soluble
forms of NSP exert benefit in terms of glycaemic control,
lipids and lipoproteins, and the prospective cohort studies,
which suggest that cereal-derived insoluble forms are
protective. The evidence regarding low glycaemic index
foods as protective is graded as ‘possible’. The evidence table
(Table 9, p 77 TRS 916) leads to the disease-specific
recommendation regarding intakes of NSP, which quite
appropriately disregards a clear distinction between soluble
and insoluble forms and suggests that adequate intakes of
NSP or dietary fibre can be achieved through regular
consumption of wholegrain cereals, legumes, fruits
and vegetables. There is no clear evidence on which to
base precise quantities although a minimum intake of 20 g
of NSP is recommended. These recommendations appear
to be compatible with this update of the relevant literature
with two caveats:
There appears to be a reasonable body of evidence to
suggest a protective effect of wholegrains.
The apparently different effects of soluble and insoluble
forms of NSP or dietary fibre may result from difficulty in
accurately distinguishing between the two forms.
European Journal of Clinical Nutrition
Cardiovascular disease and diabetes
J Mann
Similar approaches to those suggested for reducing disease
risk have been recommended for the management of
diabetes. The European evidence-based nutritional approaches
(Mann et al., 2004) to the treatment of diabetes indicate that
a wide range of intakes of total carbohydrate (45–60% total
energy) is acceptable, provided a high proportion of
carbohydrate energy is derived from vegetables, legumes,
fruits and wholegrain cereals. This should provide the
recommended intake of total dietary fibre (more than
40 g/day or 20 g/1000 kcal/day), about half of which should
be soluble. Cereal-based foods should whenever possible be
wholegrain and high in fibre. Daily consumption of at least
five servings of fibre-rich vegetables or fruits and at least four
servings of legumes per week help to provide the minimum
requirements for fibre. While a choice anywhere within this
range is acceptable for most people, it is acknowledged that
some, especially those with marked hypertriglyceridaemia
and/or severe insulin resistance, may respond more appropriately
to an intake at the lower end of the recommended
range. Use of carbohydrate-containing foods with a low
glycaemic index is encouraged. While a moderate amount of
sugar is acceptable in the context of energy balance, those
who are overweight should substantially restrict sugar-containing
energy-dense foods. Sugar-containing beverages,
while not energy dense, promote overweight and obesity
and are best avoided. Finally, it is important to emphasize
that the beneficial effects of certain carbohydrate-containing
foods have been based principally on the consumption of
conventional foods. Evidence for the benefit of manufactured
and functional foods containing added dietary fibre or
other components of wholegrains, fruits, vegetables and
legumes is limited. This applies also to foods with a low
glycaemic index. Some foods now marketed as having a low
glycaemic index are high in fat and sugar and therefore may
not be suitable for people with diabetes, especially those
who are overweight. The paper describing the European
recommendations (Mann et al., 2004) provides a detailed list
of references justifying the conclusions summarized here.
Conclusions
Nutritional attributes that are likely to protect against
diabetes and cardiovascular diseases are remarkably similar
as are the principles of nutritional management for those
who already have established disease. In summary, they are
as follows:
A wide range of intakes of carbohydrate containing foods
is acceptable.
Nature of carbohydrate rather than quantity is principally
what matters.
Intact fruits, vegetables, legumes and wholegrains are
excellent sources of carbohydrate likely to be rich in NSP
or dietary fibre and other potentially cardioprotective
components.

There is no good evidence of benefit when NSP or other
components of wholegrains, fruits, vegetables and
legumes are added to functional and manufactured foods.
Low glycaemic index foods may confer benefits in terms
of lowering total cholesterol and improving glycaemic
control in people with diabetes. However, it is not clear
whether these benefits are independent of the effects of
dietary fibre or the fact that low glycaemic index foods
tend to have intact plant cell walls. Furthermore, it is
uncertain whether functional and manufactured foods
with a low glycaemic index confer the same long-term
benefits as low glycaemic index predominantly plantbased
foods.
A low dietary glycaemic load may reduce the risk of type II
diabetes and cardiovascular disease. The same caveats as
those described above for the therapeutic role of low GI
foods apply.
Limitations to the use of the glycaemic index and
glycaemic load concepts in the clinical setting are further
described in the accompanying paper in this series.
Acknowledgements
I thank Professor Kjeld Hermansen, Dr Andrew Neil, Dr
Gabriele Riccardi, Dr Angela Rivellese, Dr Monika Toeller,
Professor A Stewart Truswell, Dr Rob M van Dam and Dr HH
Vorster for the valuable comments they provided on the
earlier paper.
Conflict of interest
During the preparation and peer-review of this paper in
2006, the author and peer-reviewers declared the following
interests.
Author
Professor Jim Mann: none declared.
Peer-reviewers
Professor Kjeld Hermansen: none declared.
Dr Andrew Neil: none declared.
Dr Gabriele Riccardi: none declared.
Dr Angela Rivellese: none declared.
Dr Monika Toeller: none declared.
Professor A Stewart Truswell: none declared.
Dr Rob M van Dam: none declared.
Dr HH Vorster: member and director of the Africa Unit for
Transdisciplinary health Research (AUTHeR), Research grant
from the South African Sugar association.
References
Ahrens E, Hirsch J, Oette K, Farquhar J, Stein Y (1961). Carbohydrateinduced
and fat-induced lipemia. Trans Assoc Am Physicians 74,
134–146.
Cardiovascular disease and diabetes
J Mann
Anderson JW (2002). Whole-grains intake and risk for coronary heart
disease. In: Marquart L, Slavin JL, Fulcher RD (eds). Whole-grain
Foods in Health and Disease. American Association of Cereal Chemists.
St Paul: Minnesota, USA. pp 187–200.
Anderson JW (2003). Whole grains protect against atherosclerotic
cardiovascular disease. Proc Nutr Soc 62, 135–142.
Anderson JW, Hanna TJ, Peng X et al. (2000). Whole grain foods and
heart disease risk. J Am Coll Nutr 19, 291S–299S.
Antonis A, Bersohn I (1961). The influence of diet on serum-triglycerides
in South African White and Bantu prisoners. Lancet 7, 39.
Appel LJ, Moore TJ, Obarzanek E et al. (1997). For the DASH
Collaborative Research Group. A clinical trial of the effects of
dietary patterns on blood pressure. N Engl J Med 336, 1117–1124.
Baghurst KI, Record SJ, Baghurst PA (1991). The relationship between
refined sugar and micronutrient intakes. Proc Nutr Soc Australia
16, 33.
Basu A, Devaraj S, Jialal I (2006). Dietary factors that promote or
retard inflammation. Arterioscler Thromb Vasc Biol 26, 995–1001.
Bazzano LA, He J, Ogden LG, Loria C, Vupputuri S, Myers L et al.
(2001). Legume consumption and risk of coronary heart disease in
US men and women: NHANES 1 Epidemiologic Follow-up Study.
Arch Intern Med 161, 2573–2578.
Brand-Miller J, Hayne S, Petocz P et al. (2003). Low-glycemic index
diets in the management of diabetes: a meta-analysis of randomized
controlled trials. Diabetes Care 26, 2261–2267.
Brunner EJ, Thorogood M, Rees K, Hewitt G (2005). Dietary advice
for reducing cardiovascular risk. Cochrane Database of Systematic
Reviews 4, Art. No: CD002128.DOl: 10.1002/.
Burr ML, Fehily AM, Gilbert JF, Rogers S, Holliday RM, Sweetnam PM
et al. (1989). Effects of changes in fat, fish and fibre intakes on
death and myocardial reinfarction: diet and reinfarction trial
(DART). Lancet ii, 757–761.
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.
Colditz GA, Manson JF, Stampfer MJ, Rosner B, Willett WC, Speizer
FF (1992). Diet and risk of clinical diabetes in women. Am J Clin
Nutr 55, 1018–1023.
Daly M (2003). Sugars, insulin sensitivity, and the postprandial state.
Am J Clin Nutr 78, 865S–872S.
de Lorgeril M, Renaud S, Mamelle N, Salen P, Martin J, Monjaud I
et al. (1994). Mediterranean alpha-linolenic acid-rich diet in
secondary prevention of coronary heart disease. Lancet 343,
1454–1459.
DiMeglio DP, Mattes RD (2000). Liquid versus solid carbohydrate:
effects on food intake and body weight. Int J Obes Relat Metab
Disord 24, 794–800.
Elliott SS, Keim NS, Stern JS, Teff K, Havel PJ (2002). Fructose,
weight gain, and the insulin resistance syndrome. Am J Clin Nutr
76, 911–922.
Esmaillzadeh A, Mirmiran P, Azizi F (2005). Whole-grain consumption
and the metabolic syndrome: a favorable association in
Tehranian adults. Eur J Clin Nutr 59, 353–362.
Flight I, Clifton P (2006). Cereal grains and legumes in the
prevention of coronary heart disease and stroke: a review of the
literature. Eur J Clin Nutr 60, 1145–1159.
Fried SK, Rao SP (2003). Sugars, hypertriglyceridemia, and cardiovascular
disease. Am J Clin Nutr 78 (Suppl 8), 73S–80S.
Fung TT, Hu FB, Pereira MA et al. (2002). Whole-grain intake and the
risk of type 2 diabetes: a prospective study in men. Am J Clin Nutr
76, 535–540.
Fung TT, Rimm EB, Spiegelman D, Rifai N, Tofler GH, Willett WC
et al. (2001a). Association between dietary patterns and plasma
biomarkers of obesity and cardiovascular disease risk. Am J Clin
Nutr 73, 61–67.
Fung TT, Willett WC, Stampfer MJ, Manon JE, Fu FB (2001b). Dietary
patterns and risk of coronary heart disease in women. Arch Intern
Med 161, 1857–1862.
S109
European Journal of Clinical Nutrition

S110
Cardiovascular disease and diabetes
J Mann
Garcia-Palmieri MR, Sorlie P, Tillotson J, Costas Jr R, Cordero E,
Rodriguez M (1980). Relationship of dietary intake to subsequent
coronary heart disease incidence: The Puerto Rico Heart Health
Program. Am J Clin Nutr 33, 1818–1827.
Garg A (1998). High-monounsaturated-fat diets for patients
with diabetes mellitus: a meta-analysis. Am J Clin Nutr 67,
577S–582S.
Giacco R, Parillo M, Rivellese AA et al. (2000). Long-term dietary
treatment with increased amounts of 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.
He FJ, Nawson CA, Macgregor GA (2006). Fruit and vegetable
consumption and stroke: metaanalysis of cohort studies. Lancet
367, 320–326.
Hjermann I, Byer K, Holme I et al. (1981). Effect of diet and smoking
intervention on the incidence of coronary heart disease. Lancet ii,
1303–1310.
Hodge AM, English DR, O’Dea K et al. (2004). Glycaemic Index and
dietary fiber and the Risk of type 2 diabetes. Diabetes Care 27,
2701–2706.
Howard BV, van Horn JH, Manson JE et al. (2006) (reprinted) Low-fat
dietary pattern and risk of cardiovascular disease: the women’s
health initiative randomized controlled dietary modification trial.
JAMA 295, 655–666.
Hu FB (2003). Plant-based foods and prevention of cardiovascular
disease: an overview. Am J Clin Nutr 78 (Suppl 3), 544S–551S.
(review).
Hu FB, Rimm EB, Stampfer MJ, Ascherio A, Spiegelman D,
Willett WC (2000). Prospective study of major dietary patterns
and risk of coronary heart disease in men. Am J Clin Nutr 72,
912–921.
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.
Jacobs Jr DR, Pereira MA, Meyer KA, Kushi LH (2000). Fiber from
whole grains, but not refined grains, is inversely associated with
all-cause mortality in older women: the Iowa women’s health
study. J Am Coll Nutr 19, 326S–330S.
Jenkins DJ, Kendall CW, Marchie A, Faulkner DA, Wong JM, de Souza
R et al. (2003). Effects of a dietary portfolio of cholesterol-lowering
foods vs lovastatin on serum lipids and C-reactive protein. JAMA
290, 502–510.
Jenkins DJA, Kendall CWC, Axelsen M, Augustin LSA, Vuksan V
(2000). Viscous and nonviscous fibres, nonabsorbable and low
glycaemic indeed carbohydrates, blood lipids and coronary heart
disease. Curr Opin Lipidol 11, 49–56.
Jensen MK, Koh-Banerjee P, Hu FB, Franz M, Sampson L, Gronbaek 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.
Joshipura KJ, Hu FB, Manson JE et al. (2001). The effect of fruit and
vegetable intake on risk for coronary heart disease. Ann Intern Med
134, 1106–1114.
Juntunen KS, Laaksonen DE, Poutanen KS, Niskanen LK, Mykkanen
HM (2003). High-fiber rye bread and insulin secretion and
sensitivity in healthy postmenopausal women. Am J Clin Nutr
77, 385–391.
Kasim-Karakas SE, Almario RU, Mueller WM, Peerson J (2000).
Changes in plasma lipoproteins during low-fat, high-carbohydrate
diets: effects of energy intake. Am J Clin Nutr 71, 1439–1447.
Kasim-Karakas SE, Tsodikov A, Singh U, Jialal I (2006). Responses of
inflammatory markers to a low-fat, high carbohydrate diet: effects
of energy intake. Am J Clin Nutr 83, 774–779.
Kelly S, Frost G, Whittaker V et al. (2004). Low glycaemic index diets
for coronary heart disease. Cochrane Database Syst Rev 18,
CD004467.
European Journal of Clinical Nutrition
Keys A (1971). Sucrose in the diet and coronary heart disease.
Atherosclerosis 14, 193–202.
Kinosian B, Glick H, Preiss L, Puder KL (1995). Cholesterol and
coronary heart disease: predicting risks in men by changes in
levels and ratios. J Invest Med 43, 443–450.
Knekt P, Reunanen A, Jarvinen R, Seppanen R, Heliovaara M, Aromaa
A (1994). Antioxidant vitamin intake and coronary mortality in a
longitudinal population study. Am J Epidemiol 139, 1180–1189.
Knowler WC, Barrett-Connor E, Fowler SE, et al. , For the Diabetes
Prevention Program Research Group (2002). Reduction in the
incidence of type 2 diabetes with lifestyle intervention or
metformin. N Engl J Med 346, 393–403.
Lau C, Faerch K, Glumer C, Tetens I, Pedersen O, Carstensen B
et al. (2005). Dietary glycemic index, glycemic load, fiber, simple
sugars, and insulin resistance: the Inter 99 study. Diabetes Care 28,
1397–1403.
Liese AD, Schulz M, Fang F, Wolever TM, D’Agostino RB, Sparks KC
et al. (2005). Dietary glycemic index and glycemic load, carbohydrate
and fiber intake, and measures of insulin sensitivity,
secretion, and adiposity in the Insulin Resistance Atherosclerosis
Study. Diabetes Care 28, 2832–2838.
Lindstrom J, Louheranta A, Mannelin M, Rastas M, Salminen V,
Eriksson J et al. (2003). The Finnish Diabetes Prevention Study
(DPS): lifestyle intervention and 3-year results on diet and physical
activity. Diabetes Care 26, 3230–3236.
Liu S (2002). Intake of refined carbohydrates and whole grain foods
in relation to risk of type 2 diabetes mellitus and coronary heart
disease. J Am Coll Nutr 21, 298–306.
Liu S, Lee IM, Ajani U, Cole SR, Buring JE, Manson JE (2001).
Intake of vegetables rich in carotenoids and risk of coronary heart
disease in men: the Physicians’ Health Study. Int J Epidemiol 30,
130–135.
Liu S, Manson JE, Stampfer MJ et al. (2000). A prospective study of
whole-grain intake and risk of type 2 diabetes mellitus in US
women. Am J Public Health 90, 1409–1415.
Liu S, Manson JE, Stampfer MJ, Rexrode KM, Hu FB, Rimm EG
et al. (2000a). Whole grain consumption and risk of ischemic
stroke in women: a prospective study. JAMA 284, 1534–1540.
Liu S, Willett WC, Stampfer MJ, Hu FB, Franz M, Sampson L
et al. (2000b). A prospective study of dietary glycaemic load,
carbohydrate intake, and risk of coronary heart disease in US
women. Am J Clin Nutr 71, 1455–1461.
Mann J (2001). Dietary fibre and diabetes revised. Eur J Clin Nutr 55,
919–921.
Mann J (2004). Free sugars and human health: sufficient evidence for
action? Lancet 363, 1068–1070.
Mann JI (1987). Simple sugars and diabetes. Diabet Med 4, 135–139.
Mann JI (2000). Stemming the tide of the diabetes epidemic. Lancet
356, 1454–1455 (editorial).
Mann JI, De Leeuw I, Hermansen K, Karamanos B, Karlström B,
Katsilambros N et al. (2004). Evidence-based nutritional
approaches to the treatment and prevention of diabetes mellitus.
Nutr Metab Cardiovasc Dis 14, 373–394.
McAuley KA, Hopkins CM, Smith KJ, McLay RT, Williams SM, Taylor
RW et al. (2005). Comparisons of high-fat and high-protein diets
with a high-carbohydrate diet in insulin-resistant obese women.
Diabetologia 48, 8–16.
McAuley KA, Smith KJ, Taylor RW, McLay RT, Williams SM, Mann JI
(2006). Long-term effects of popular dietary approaches on weight
loss and features of insulin resistance. Int J Obes 30, 342–349.
McGee DL, Reed DM, Yano K, Kagan A, Tillotson J (1984). Ten-year
incidence of coronary heart disease in the Honolulu Heart
Program: relationship to nutrient intake. Am J Epidemiol 119,
667–676.
Mensink RP, Zock PL, Kester ADM, Katan MB (2003). Effects of
dietary fatty acids and carbohydrates on the ratio of serum total to
HDL cholesterol and on serum lipids and apolipoproteins: a metaanalysis
of 60 controlled trials. Am J Clin Nutr 77, 1146–1155.

Meyer KA, Kushi LH, Jacobs Jr DR, Slavin J, Sellers TA, Folsom AR
(2000). Carbohydrates, dietary fiber, and incident type 2 diabetes
in older women. Am J Clin Nutr 71, 921–930.
Montonen J, Knekt P, Jarvinen R, Aromaa A, Reunanen A (2003).
Whole-grain and fiber intake and the incidence of type 2 diabetes.
Am J Clin Nutr 77, 622–629.
Morgan LM, Tredger JA, Shavila Y, Travis JS, Wright J (1993). The
effect of non-starch polysaccharide supplementation on circulating
bile acids, hormone and metabolite levels following a fat meal
in human subjects. Br J Nutr 70, 491–501.
Morris JN, Marr JW, Clayton DG (1977). Diet and heart: a postscript.
BMJ 2, 1307–1314.
Mozaffarian D, Kumanyika SK, Lemaitre RN, Olson JL, Burke GL,
Siscovick DS (2003). Cereal, fruit and vegetable fiber intake and
the risk of cardiovascular disease in elderly individuals. JAMA 289,
1659–1666.
Ness AR, Hughes J, Elwood PC, Whiteley E, Smith GD, Burr ML
(2002). The long-term effect of dietary advice in men with
coronary disease: follow-up of the Diet and Reinfarction trial
(DART). Eur J Clin Nutr 56, 512–518.
Nordmann AJ, Nordmann A, Briel M, Keller U, Yancy Jr WS, Brehm BJ
et al. (2006). Effects of low-carbohydrate vs low-fat diets on weight
loss and cardiovascular risk factors: a meta-analysis of randomized
controlled trials. Arch Intern Med 166, 285–293.
Nuttall FO, Gannon MC, Wald JL, Ahmed M (1985). Plasma glucose
and insulin profiles in normal subjects ingesting diets of varying
carbohydrate, fat, and protein content. J Am Coll Nutr 4, 437–450.
Nuttall FO, Gannon MC (1981). Sucrose and disease. Diabetes Care
4L, 305–310.
Opperman AM, Venter CS, Oosthuizen W et al. (2004). Meta-analysis
of the health effects of using the glycaemic index in mealplanning.
Br J Nutr 92, 367–381.
Pereira MA, Jacobs Jr DR, Pins JJ, Raatz SK, Gross MD, Slavin JL et al.
(2002). Effect of whole grains on insulin sensitivity in overweight
hyperinsulinaemic adults. Am J Clin Nutr 75, 848–855.
Pereira MA, O’Reilly E, Augustsson K, Fraser GE, Goldbourt U,
Heitmann BL et al. (2004). Dietary fiber and risk of coronary heart
disease: a pooled analysis of cohort studies. Arch Intern Med 164,
370–376.
Peterson DB, Lambert J, Gerring S et al. (1986). Sucrose in the diet
of diabetic patients-just another carbohydrate? Diabetologia 29,
216–220.
Poppitt SD, Keogh GS, Prentice AM, Williams DE, Sonnemans HM,
Valk EE et al. (2002). Long-term effects of ad libitum low-fat,
high-carbohydrate diets on body weight and serum lipids in
overweight subjects with metabolic syndrome. Am J Clin Nutr 75,
11–20.
Ramachandran A, Snehalatha C, Mary S, Mukesh B, Bhaskar AD,
Vijay V (2006). The Indian Diabetes Prevention Programme shows
that lifestyle modification and metformin prevent type 2 diabetes
in Asian Indian subjects with impaired glucose tolerance (IDPP—1).
Diabetologia 49, 289–297.
Retzlaff BM, Walden CE, Dowdy AA, McCann BS, Anderson KV,
Knopp RH (1995). Changes in plasma triacylglycerol concentrations
among free-living hyperlipidemic men adopting different
carbohydrate intakes over two years: The Dietary Alternatives
Study. Am J Clin Nutr 62, 988–995.
Riccardi G, Rivellese A, Pacioni D, Genovese S, Mastranzo P, Mancini
M (1984). Separate influence of dietary carbohydrate and fibre on
the metabolic control in diabetes. Diabetologia 26, 116–121.
Rivellese A, Riccardi G, Giacco A et al. (1980). Effect of dietary fibre
on glucose control and serum lipoproteins in diabetic patients.
Lancet 2, 447–450.
Cardiovascular disease and diabetes
J Mann
Salmeron J, Ascherio A, Rimm EB et al. (1997b). Dietary
fiber, glycemic load, and risk of NIDDM in men. Diabetes Care
20, 545–550.
Salmeron J, Manson JE, Stampfer MJ et al. (1997a). Dietary fiber,
glycemic load, and risk of non-insulin-dependent diabetes
mellitus in women. JAMA 277, 472–477.
Saris WH, Astrup A, Prentice AM et al. (2000). Randomized controlled
trial of changes in dietary carbohydrate/fat ratio and simple vs
complex carbohydrates on body weight and blood lipids: the
CARMEN study. The Carbohydrate Ratio Management in European
National diets. Int J Obes Relat Metab Disord 24, 1310–1318.
Simpson HCR, Carter RD, Lousley S, Mann JL (1982). Digestible
carbohydrate—an independent effect on diabetic control in type
11 (non-insulin dependent) diabetic patients? Diabetologia 23,
235–239.
Simpson HCR, Simpson RW, Lousley S et al. (1981). A high
cabohydrate leguminous fibre diet improves all aspect of diabetic
control. Lancet 1, 1–5.
Slama G, Haardt MJ, Jean-Joseph P et al. (1984). Sucrose taken during
mixed meal has no additional hyperglycaemic action over
isocaloric amounts of starch in well-controlled diabetes. Lancet
2, 122–125.
Slavin J (2003). Why whole grains are protective: biological
mechanisms. Proc Nutr Soc 62, 129–134.
Stone D, Connor W (1963). The prolonged effects of a lowcholesterol,
high-carbohydrate diet upon serum lipids in diabetic
patients. Diabetes 12, 127–132.
Toeller M, Buyken AE, Heitkamp G, de Pergola G, Giorgino F, Fuller
JH (1999). Fiber intake, serum cholesterol levels, and cardiovascular
disease in European individuals with type 1 diabetes.
EURODIAB IDDM Complications Study Group. Diabetes Care 22
(Suppl 2), B21–B28.
Trowell H (1975). Dietary-fiber hypothesis of the etiology of diabetes
mellitus. Diabetes 24, 762–765.
Truswell AS (1987). Sugar and health: a review. Food Technology in
Australia 39, 134–140.
Truswell AS (1994). Food carbohydrates and plasma lipids—an
update. Am J Clin Nutr 59, 710S–718S.
Truswell AS (2002). Cereal grains and coronary heart disease.
Eur J Clin Nutr 56, 1–14.
Tuomilehto J, Lindström J, Eriksson JG et al. (2001). For the Finnish
Diabetes Prevention Study Group. Prevention of type 2 diabetes
mellitus by changes in lifestyle among subjects with impaired
glucose tolerance. N Engl J Med 334, 1343–1350.
Turley ML, Skeaff CM, Mann JI, Cox B (1998). The effect of a low-fat,
high-carbohydrate diet on serum high density lipoprotein cholesterol
and triglyceride. Eur J Clin Nutr 52, 728–732.
Van Dam RM, Hu FB, Rosenberg L, Krishnan S, Palmer JR (2006).
Dietary calcium and magnesium, major food sources, and risk of
type 2 diabetes in US black women. Diabetes Care 29, 2238–2243.
Venn BJ, Mann JI (2004). Cereal grains, legumes and diabetes.
Eur J Clin Nutr 58, 1443–1461.
Weickert MO, Mohlig M, Schofl C et al. (2006). Cereal fiber improves
whole-body insulin sensitivity in overweight and obese women.
Diabetes Care 29, 775–780.
WHO Technical Report Series 916 (2003). Diet, nutrition and the
prevention of chronic diseases. Report of a Joint WHO/FAO Expert
Consultation (Geneva).
Yudkin J (1964). Dietary fat and dietary sugar in relation to ischaemic
heart disease and diabetes. Lancet 4, 4–5.
Zhang C, Liu S, Solomon CG et al. (2006). Dietary fiber intake,
dietary glycemic load and the risk for gestational diabetes mellitus.
Diabetes Care 29, 2223–2230.
S111
European Journal of Clinical Nutrition

REVIEW
Carbohydrates and cancer: an overview of the
epidemiological evidence
TJ Key and EA Spencer
Cancer Research UK Epidemiology Unit, University of Oxford, Oxford, UK
Objective: To assess the epidemiological evidence on dietary carbohydrates and the risk of developing cancer.
Method: Review of published studies, concentrating on recent systematic reviews, meta-analyses and large prospective studies.
Conclusions: Carbohydrates have not been intensively investigated in epidemiological studies of diet and cancer. There is a
moderately large amount of data on the possible association between dietary fibre and the risk for colorectal cancer; the results
of studies have varied and no firm conclusion can be drawn, but the available data suggest that high intakes of dietary fibre
possibly reduce the risk for colorectal cancer. There are also limited data which suggest that high intakes of sucrose might
increase the risk for colorectal cancer and that high intakes of lactose might increase the risk for ovarian cancer. For other
components of carbohydrates and other types of cancer, the available data are too sparse to draw even tentative conclusions.
Further research is needed on the possible associations of carbohydrates with cancer risk.
European Journal of Clinical Nutrition (2007) 61 (Suppl 1), S112–S121; doi:10.1038/sj.ejcn.1602941
Keywords: carbohydrates; sugar; dietary fibre; cancer; review
Introduction
European Journal of Clinical Nutrition (2007) 61 (Suppl 1), S112–S121
& 2007 Nature Publishing Group All rights reserved 0954-3007/07 $30.00
www.nature.com/ejcn
General
The causes of cancer are incompletely understood, and
cancers in different organs in the body can have different
causes. Reviews have provided estimates that dietary factors
may account for approximately 30% of cancers in industrialized
countries and approximately 20% in developing
countries, but it has proved difficult to clearly establish the
effects of dietary factors on cancer risk (WHO, 2003a,b; Key
et al., 2004). The WHO/FAO Expert Consultation in 2002
categorized putative nutritional risk factors according to the
strength of the evidence available (WHO, 2003a). The only
diet-related factors for which there is convincing evidence of
an effect on cancer risk are overweight and obesity, alcohol,
aflatoxin and Chinese-style salted fish; there is also evidence
that fruit and vegetables probably reduce the risk for some
cancers, and that high intakes of preserved meat, saltpreserved
foods and salt and very hot drinks and food
probably increase the risk for some types of cancer (WHO,
2003a).
Correspondence: Professor TJ Key, Cancer Research UK Epidemiology Unit,
University of Oxford, Richard Doll Building, Roosevelt Drive, Oxford OX3 7LF,
UK.
E-mail: Tim.Key@ceu.ox.ac.uk
The body of evidence relating to diet and cancer risk in
humans derives mostly from observational epidemiological
studies conducted during the last 30 years, together with a
few randomized controlled trials. Most attention has been
given to a small number of prominent hypotheses in relation
to certain food groups, in particular meat, fruit and
vegetables; certain macronutrients, especially fat and saturated
fat; and some of the micronutrients. With the
exception of dietary fibre, relatively little attention has been
given to the possibility that consumption of carbohydrates
may affect cancer risk, although some attention has been
given to sucrose and recently several studies have investigated
the possible associations of glycaemic index and
glycaemic load with cancer risk.
Approach and methods of this overview
Our aim in this overview is to summarize the current
evidence and understanding of the possible role of carbohydrates
in determining the risk for various cancers. This is not
a systematic review of all relevant literature; rather, we have
discussed previous assessments by expert panels (World
Cancer Research Fund, 1997; Department of Health UK,
1998; WHO, 2003a), meta-analyses and systematic reviews,
and data from recent publications (1995 onwards), especially

from prospective studies which have not been considered in
previous reviews. Publications were identified by searching
on PubMed using the terms carbohydrate, sucrose, lactose,
galactose, glycaemic index, glycaemic load, dietary fibre and
individual cancer types, together with searching the reference
lists of each paper identified. It is noted that the World
Cancer Research Fund is currently updating their previous
systematic review of diet and cancer; this will include all
studies relevant to the possible associations between carbohydrate
consumption and cancer risk, and is scheduled to be
published during 2007.
For each cancer site we summarize the findings available in
relation to sucrose, glycaemic index and glycaemic load and
dietary fibre. Some studies have reported on total carbohydrate;
we describe these findings briefly for stomach
cancer, but for other cancers have followed the conclusion
of the World Cancer Research Fund report (1997), that such
results are very difficult to interpret due to the differences
between different types of carbohydrate, especially starches
and sugars. The estimation and interpretation of glycaemic
index and glycaemic load is discussed fully in the paper by
Venn and Green (2007); we note here simply that glycaemic
index is the blood glucose response to a test food compared
to a standard reference food, that the glycaemic load is
usually calculated by multiplying the glycaemic index of a
food by the content of available carbohydrate, and that the
interpretation of these estimates requires caution because of
variations due to factors such as food preparation and total
meal composition.
We have restricted our review to studies with cancer as
the endpoint, excluding studies of intermediate endpoints
such as precancerous lesions, with the single exception
of colorectal cancer, where we discuss briefly the results
of randomized trials and large prospective studies with colorectal
adenoma as the endpoint. We have not discussed
possible mechanisms in any detail, but where the epidemiological
evidence suggests that there might be an association
of carbohydrates with cancer risk we have mentioned briefly
potential mechanisms which have been proposed.
Limitations of the research
The study of carbohydrates in relation to cancer risk suffers
several limitations. As with all epidemiological studies of
nutrition, it is difficult to obtain valid measures of long-term
dietary intake. Dietary assessment in general is not very
accurate, and for carbohydrates there are some particular
problems; for example, substantial amounts of sucrose can
be incorporated in many manufactured foods but this can be
hard to quantify in epidemiological studies; most food
composition tables do not distinguish the sugars naturally
present in foods such as fruits from sugars added during food
manufacturing; the definition of fibre varies between food
tables; the amount of fibre in foods such as bread varies
several fold according to the method of milling and sieving
of the grain and flour, and it is difficult to identify the exact
Carbohydrates and cancer
TJ Key and EA Spencer
types of breads chosen and the quantity consumed using
dietary questionnaires. Another problem concerns the interrelations
of different dietary components; for energyproviding
nutrients such as most carbohydrates (fibre is an
exception because it provides little energy), the intake of
carbohydrate is usually inversely correlated with the intake
of some of the other macronutrients, such as fat, therefore it
can be difficult to determine which macronutrient is
genuinely associated with cancer risk.
Review of the role of carbohydrates in the
aetiology of the major cancers
Cancers of the oral cavity, pharynx and oesophagus
The dietary factors known to be important in the aetiology
of these cancers of the upper gastro-intestinal tract are
alcohol, which can cause all these cancers, and obesity,
which increases the risk for adenocarcinoma of the oesophagus
(WHO, 2003a; Crew and Neugut, 2004). Few studies
have examined the risk for these cancers in relation to the
consumption of any type of carbohydrate, although it has
been suggested that dietary fibre might reduce the risk for
adenocarcinoma of the oesophagus (Pera et al., 2005).
Stomach cancer
Diet, together with infection by the bacterium Helicobacter
pylori, is thought to be important in the aetiology of
stomach cancer, but the exact nature of this relationship
has not been established (Crew and Neugut, 2006). Risk is
probably decreased by a high intake of fruit and vegetables
and increased by a high intake of salt-preserved foods and
salt (Key et al., 2004). Some studies have suggested that the
risk for stomach cancer might be increased by diets high in
starch or total carbohydrates (Muñoz et al., 2001; De Stefani
et al., 2004; Lissowska et al., 2004), but a recent large
prospective study found no association (Larsson et al., 2006a)
and the positive association observed in some studies may
reflect the association of this cancer with low socioeconomic
status. A few studies have examined the association
between dietary fibre and risk for stomach cancer; one
small case–control study reported an inverse association
between dietary fibre and stomach cancer (De Stefani et al.,
1999), but others observed no association (Botterweck et al.,
2000; Muñoz et al., 2001; Nomura et al., 2003; Lissowska
et al., 2004).
Whereas overall stomach cancer rates have fallen in most
western countries so that these countries now have lower
rates for this cancer than do many developing countries, one
sub-type of stomach cancer, adenocarcinoma of the gastric
cardia, is increasing and is more common among western
than developing countries (Crew and Neugut, 2006). The risk
for this sub-type of stomach cancer is positively associated
with obesity (Hampel et al., 2005), and some evidence has
led to the suggestion that a high intake of cereal fibre might
S113
European Journal of Clinical Nutrition

S114
Carbohydrates and cancer
TJ Key and EA Spencer
reduce the risk for this type of stomach cancer (Terry et al.,
2001).
Overall, the data are insufficient to draw any conclusions
as to whether the consumption of carbohydrates has any
effect on the risk for stomach cancer.
Colorectal cancer
The aetiology of colorectal cancer is poorly understood, but
rates are generally much higher in western countries than
in developing countries, and much evidence suggests that
dietary factors are likely to be important in the development
of this disease (World Cancer Research Fund, 1997; Potter,
1999; Key et al., 2004). However, despite much research on
diet and colorectal cancer, it is still unclear which dietary
factors are important. In relation to carbohydrates, the most
prominent hypothesis is that high intakes of dietary fibre
may reduce risk; there has also been some research into the
possible roles of sugar and, more recently, of glycaemic index
and glycaemic load.
Sucrose. The possibility that a high consumption of sucrose
might increase the risk for colorectal cancer has been
investigated in several epidemiological studies, but has not
been studied extensively. In a review by Burley (1997), 18
case–control and three prospective studies of sugar and
colorectal cancer were considered; 7 of these 21 studies
reported risk estimates consistent with a positive association
between sugar consumption and the risk for colorectal
cancer, but the author concluded that, due to the limited
number of well-conducted studies, there was insufficient
information to assess the degree of risk associated with sugar
consumption. In the World Cancer Research Fund review
(World Cancer Research Fund, 1997), 12 case–control studies
and one cohort study were assessed (the smaller number of
studies included in this review than in the Burley (1997)
review is largely explained by the stricter application of
criteria as to which studies were relevant among studies
which reported on a limited number of foods); 8 of the 12
case–control studies, and the single cohort study, reported
that diets comparatively high in refined sucrose or sucrosecontaining
foods were associated with increased risk of
colorectal cancer. The reviewers noted that high sugar
intakes are associated with other dietary factors that may
affect the risk for colorectal cancer, and concluded that diets
high in extrinsic (refined) sugars possibly increase the risk of
Table 1 Prospective studies of sucrose intake and colorectal cancer risk
colorectal cancer, and that the evidence was strongest for
sucrose.
Table 1 summarizes the findings of prospective studies
which have reported on sucrose intake and the risk for
colorectal cancer; three of these four studies were published
after the systematic reviews discussed above. For each study
we have given the relative risk for people with the highest
level of sucrose consumption compared to people with the
lowest level of consumption within the same study. Three of
the four relative risks reported were above unity, but none
was statistically significant (Terry et al., 2003 did not cite a
relative risk but stated that there was no association with
sucrose).
Overall, therefore, the available data are compatible with
the hypothesis that high intakes of sugar may increase the
risk for colorectal cancer. However, the results from prospective
studies are inconsistent, few studies have addressed
this hypothesis as their primary, pre-specified objective and
there may be reporting bias in that investigators may have
selectively reported findings where there was a positive
association of sucrose intake with the risk for colorectal
cancer. Possible mechanisms by which high intakes of
sucrose might increase the risk for colorectal cancer include
increasing mouth to anus transit time and increasing the
faecal concentration of secondary bile acids (Bostick et al.,
1994).
Glycaemic index and glycaemic load. Several case–control
studies have suggested that there might be a positive
association between diets with a high glycaemic index or
glycaemic load and the risk for colorectal cancer (Slattery
et al., 1997; Franceschi et al., 2001; Levi et al., 2002).
Subsequently, four prospective studies have investigated
the associations of estimates of glycaemic index and
glycaemic load with the risk for colorectal cancer (Table 2).
For glycaemic index, all four relative risks for high versus low
index were above unity, but none was statistically significant.
For glycaemic load, four out of five relative risks
reported were greater than unity, of which one was
statistically significant. The available data are too few to
draw any conclusions on whether there is any association
between either of these variables and the risk for colorectal
cancer.
Fibre. Burkitt proposed in 1969 that a high intake of dietary
fibre reduces the risk for colorectal cancer (Burkitt, 1969).
First author and year Country Cases Sex Comparison Relative risk (95% CI) in highest category
Bostick et al., 1994 USA 212 Women Highest vs lowest fifth 1.45 (0.88–2.39)
Terry et al., 2003 Canada 616 Women Highest vs lowest fifth No association
Higginbotham et al., 2004a USA 174 Women Highest vs lowest fifth 1.51 (0.90–2.54)
Michaud et al., 2005 USA 683 Men Highest vs lowest fifth 1.30 (0.99–1.69)
Michaud et al., 2005 USA 1096 Women Highest vs lowest fifth 0.89 (0.72–1.11)
European Journal of Clinical Nutrition

Table 3 Prospective studies of dietary fibre and risk for colorectal cancer
First author and year Region Cases Sex Comparison Relative risk in highest category
Bingham et al., 2005 Europe a
1721 Both Highest vs lowest fifth 0.79 (0.63–0.99)
Park et al., 2005 North America and Europe b
8081 Both Highest vs lowest fifth 0.94 (0.86–1.03)
Otani et al., 2006 Japan 567 Men Highest vs lowest fifth 0.92 (0.67–1.3)
Otani et al., 2006 Japan 340 Women Highest vs lowest fifth 1.4 (0.95–2.2)
a Data from 10 European countries.
b Pooled analysis of data from 13 studies in six countries.
This hypothesis has subsequently been investigated in
numerous observational studies. In a meta-analysis of 13
case–control studies, Howe et al. (1992) reported that people
with a high intake of dietary fibre had a risk for colorectal
cancer about 50% lower than that of people with a relatively
low intake of dietary fibre (relative risk in highest versus
lowest fifth of consumption was 0.53). However, the results
from prospective studies have been much closer to null
(Table 3). In the Pooling Project of Prospective Studies of Diet
and Cancer, 8081 colorectal cancer cases were identified
among men and women from 13 cohort studies (Park et al.,
2005). For comparison of the highest versus lowest studyand
sex-specific fifth of dietary fibre intake, a significant
inverse association was found in the age-adjusted model
(pooled relative risk 0.84; 95% confidence interval [CI],
0.77–0.92). However, the association was attenuated and no
longer statistically significant after adjusting for other dietary
and non-dietary risk factors (pooled multivariate RR 0.94;
95% CI, 0.86–1.03). In the European Prospective Investigation
into Cancer and Nutrition (EPIC), a large prospective
study, which was not included in the Pooling Project, the
association of dietary fibre with colorectal cancer was
investigated among 1721 cases from nine European countries.
The relative risk for people in the highest versus the
lowest fifth of dietary fibre intake, adjusted for dietary and
non-dietary covariates, was 0.79 (0.63–0.99; Bingham et al.,
2005). Thus the results from EPIC show a significantly lower
risk for colorectal cancer associated with high-fibre intake,
whereas the main results from the Pooling Project show no
significant association; however, further analyses in the
report from the Pooling Project show a significantly
increased risk for colorectal cancer among participants with
a very low intake of fibre (dietary fibre intake of o10 g/day
Carbohydrates and cancer
TJ Key and EA Spencer
Table 2 Prospective studies of glycaemic index, glycaemic load and colorectal cancer risk
First author and year Country Cases Sex Comparison Glycaemic index,
relative risk in highest category
Glycaemic load,
relative risk in highest category
Terry et al., 2003 Canada 616 Women Highest vs lowest fifth Not reported 1.05 (0.73–1.53)
Higginbotham et al., 2004a USA 174 Women Highest vs lowest fifth 1.71 (0.98–2.98) 2.85 (1.40–5.80)
Michaud et al., 2005 USA 683 Men Highest vs lowest fifth 1.14 (0.88–1.48) 1.32 (0.98–1.79)
Michaud et al., 2005 USA 1096 Women Highest vs lowest fifth 1.08 (0.87–1.34) 0.89 (0.71–1.11)
McCarl et al., 2006 USA 954 Women Highest vs lowest fifth 1.08 (0.88–1.32) 1.09 (0.88–1.35)
versus intake of X30 g/day, relative risk 1.18 (1.05–1.31)).
Both the Pooling Project and EPIC reported results according
to the source of fibre, categorized as cereals, fruit or
vegetables, but neither study showed clear differences in
the association with risk according to fibre source (Bingham
et al., 2005; Park et al., 2005). A further recent prospective
study in Japan, which was not included in the Pooling
Project, reported no significant association of dietary fibre
with colorectal cancer among 907 incident cases (Otani et al.,
2006).
In the Women’s Health Initiative (Beresford et al., 2006), a
randomized controlled trial designed to evaluate the effects
of a low-fat eating pattern, the intervention aimed to reduce
total fat and to increase consumption of vegetables, fruits
and grains, and the intake of dietary fibre in the intervention
group increased from 15.4 to 17.9 g/day. The hazard ratio for
colorectal cancer during 8 years of follow-up was 1.08 (0.90–
1.29), thus the small increase in dietary fibre intake did not
reduce the risk of colorectal cancer. In three randomized
controlled trials with colorectal adenoma as the endpoint,
the recurrence of adenomas was not reduced by a fibre
supplement, a high-fibre cereal supplement or by a highfibre
low-fat diet (Alberts et al., 2000; Bonithon-Kopp et al.,
2000; Schatzkin et al., 2000), although a recent pooled
analysis of the latter two trials suggested that there was a
significant reduction in adenoma recurrence among men
but not among women (Jacobs et al., 2006), and observational
prospective studies have suggested that high intakes of
dietary fibre may reduce the risk for adenomas (Peters et al.,
2003; Robertson et al., 2005). However, although some
colorectal adenomas progress to cancer, many do not,
therefore factors which affect adenomas may not have the
same effect on colorectal cancer.
S115
European Journal of Clinical Nutrition

S116
In summary, despite a substantial amount of research, the
relationship of dietary fibre intake with the development of
colorectal cancer is not well understood, and we conclude
that high intakes of dietary fibre possibly reduce the risk for
colorectal cancer.
Several mechanisms which might explain a protective
effect of fibre have been proposed; dietary fibre might reduce
the risk for colorectal cancer by increasing the speed of
transit of food material through the large intestine, by
fermentation to short-chain fatty acids which may promote
cell differentiation, induce apoptosis and/or inhibit the
production of secondary bile acids by reducing luminal pH,
or by other mechanisms (Hague et al., 1995; Nagengast et al.,
1995; Potter, 1999).
Some studies have investigated whether consumption of
whole grains rather than refined grains is associated with a
relatively low risk for colorectal cancer. In a review and metaanalysis
in 1998, Jacobs et al. (1998) concluded that the
evidence from case–control studies supported the hypothesis
that a high intake of whole grains reduces the risk for
colorectal cancer. However, the Pooling Project of Prospective
Studies of Diet and Cancer showed no significant
association between whole grain intake and the risk for
colorectal cancer (Park et al., 2005).
Cancer of the liver and biliary tract
The major causes of liver cancer (hepatocellular carcinoma)
include hepatitis viruses, cirrhosis, liver flukes and aflatoxins
(WHO, 2003b; Bosch et al., 2005). Carbohydrates are not
thought to be directly relevant to the aetiology of this
cancer, although under poor storage conditions some staple
foods such as maize and rice, which are rich sources of
carbohydrates, can become contaminated with aflatoxins
produced by aspergillus.
For cancer of the biliary tract, Zatonski et al. (1997)
reported in a case–control study that a high intake of
carbohydrates was associated with a substantial increase in
risk; however, this might reflect dietary changes due to longstanding
symptoms of gallbladder disease. There are few
other data on the possible role of carbohydrates in the
aetiology of this cancer, and no conclusions can be drawn
(Pandey, 2003).
Cancer of the pancreas
The only well-established risk factor for cancer of the
pancreas is smoking (Lowenfels and Maisonneuve, 2005).
Chronic pancreatitis and type-II diabetes are associated with
pancreatic cancer (Michaud, 2004; Huxley et al., 2005) but
some uncertainties remain as to whether these represent
largely causal associations or whether these non-malignant
conditions are sometimes due to the presence of early
pancreatic cancer. The role of diet in the aetiology of this
cancer is not well understood. Obesity is weakly positively
associated with risk (Berrington de Gonzalez et al., 2003), but
European Journal of Clinical Nutrition
Carbohydrates and cancer
TJ Key and EA Spencer
no other dietary risk factors have been confirmed. Based on
the association with diabetes, it has been proposed that the
risk for cancer of the pancreas may be increased by a diet
with a high glycaemic load and by impaired glucose
tolerance, insulin resistance and hyperinsulinaemia. Three
prospective studies have examined the associations of sugar,
sugar-sweetened soft drinks, glycaemic index and glycaemic
load with the risk for pancreatic cancer: sucrose itself was not
associated with risk and the results for glycaemic index and
load were inconsistent (Michaud et al., 2002; Johnson et al.,
2005; Silvera et al., 2005a), whereas there was some evidence
that risk increased with a high intake of sugar-sweetened soft
drinks in women (Schernhammer et al., 2005). Case–control
studies have suggested that high intakes of dietary fibre
might reduce risk (World Cancer Research Fund, 1997), but
few data are available from prospective studies (Stolzenberg-
Solomon et al., 2002). There are insufficient data to draw any
conclusions.
Lung cancer
Most cases of lung cancer are caused by smoking. The
possibility that diet might also have an effect was raised by
studies in the 1970s (Key et al., 2004). Subsequent research
has focused on fruit, vegetables and related micronutrients,
with little attention given to carbohydrates. No dietary
effects on lung cancer have been established, and the weak
associations between some dietary factors and risk observed
in many studies might be due to residual confounding by
smoking.
Breast cancer
The established risk factors for breast cancer are hormonal
and reproductive factors (Key et al., 2003). Risk among
postmenopausal women is increased by obesity (van den
Brandt et al., 2000), probably because obese women have
relatively high oestrogen production due to synthesis in the
adipose tissue from androgen precursors (Endogenous
Hormones and Breast Cancer Collaborative Group, 2003).
In relation to carbohydrates, it is possible that high intakes
of sucrose or high glycaemic load might increase risk by
leading to obesity (see van Dam and Seidell, 2007) which
causes an increase in endogenous oestrogen levels, while it is
possible that high intakes of fibre might reduce risk, perhaps
by interrupting the entero-hepatic circulation of oestrogens
and therefore reducing oestrogen levels in the breasts (Key
et al., 2003).
Sucrose. Few studies have investigated the association
between sucrose consumption and breast cancer risk. In
previous reviews it was concluded that the little evidence
available was inconsistent and therefore that no judgement
was possible (World Cancer Research Fund, 1997) or that
although there was some suggestion of a small increase in
risk of breast cancer in association with a high consumption

of sucrose-containing foods, such as cakes and biscuits, there
was insufficient evidence to conclude whether sugar has
any association with breast cancer risk (Burley, 1998). Some
subsequent case–control studies have also observed an
increased risk for breast cancer among women with a
relatively high consumption of sugar (Romieu et al., 2004)
or sweet foods (Potischman et al., 2002). Subsequent
prospective studies have mainly examined glycaemic index
and glycaemic load (see below) rather than sucrose, and the
few prospective data published on sucrose have not
suggested that there is an association with breast cancer risk
(Nielsen et al., 2005; Silvera et al., 2005b).
Glycaemic index and glycaemic load. Seven prospective
studies have investigated the associations of estimates of
glycaemic index and glycaemic load with the risk for breast
cancer (Table 4). For glycaemic index, the relative risks for a
high index versus a low index ranged from 0.88. to 1.15,
with a median of 1.03, and one of the relative risk was
significantly greater than unity. For glycaemic load the
relative risks for a high load versus a low load ranged from
0.87 to 1.19, with a median of 1.02 and none was statistically
Table 4 Prospective studies of glycaemic index, glycaemic load and breast cancer risk
significant. The available data therefore suggest that there is
no association between either of these variables and the risk
for breast cancer.
Fibre. In previous systematic reviews, the World Cancer
Research Fund (1997) concluded that, based largely on the
findings from case–control studies, dietary fibre possibly
decreases the risk of breast cancer, whereas the Department
of Health UK report concluded simply that the evidence
was inconsistent (Department of Health UK, 1998). In six
subsequent prospective studies, the relative risks for high
versus low consumption of total dietary fibre ranged from
0.58. to 1.08, with a median of 0.92 (Table 5), and one of
these relative risks was significantly below unity. Analyses
according to the food source of fibre, categorized as cereals,
fruit and vegetables, did not show any significant associations
with risk (Table 5). Overall, the data do not support the
hypothesis that a high intake of dietary fibre might reduce
breast cancer risk.
In a prospective study, there was no clear association
between the consumption of whole and refined grain and
breast cancer risk (Nicodemus et al., 2001).
First author and year Country Cases Menopausal status Comparison Glycaemic index, relative
risk in highest category
Glycaemic load, relative
risk in highest category
Cho et al., 2003 USA 714 Premenopausal Highest vs lowest fifth 1.05 (0.83–1.33) 1.06 (0.78–1.45)
Jonas et al., 2003 USA 1442 Postmenopausal Highest vs lowest fifth 1.03 (0.87–1.22) 0.90 (0.76–1.08)
Higginbotham et al., 2004b USA 946 Any Highest vs lowest fifth 1.03 (0.84–1.28) 1.01 (0.76–1.35)
Holmes et al., 2004 USA 854 Premenopausal Highest vs lowest fifth 1.02 (0.82–1.28) 0.87 (0.70–1.12)
Holmes et al., 2004 USA 2924 Postmenopausal Highest vs lowest fifth 1.15 (1.02–1.30) 1.03 (0.90–1.16)
Nielsen et al., 2005 Denmark 634 Postmenopausal Per 10 units GI, per 100 units GL 0.94 (0.80–1.10) 1.04 (0.90–1.19)
Silvera et al., 2005b Canada 1461 Any Top fifth 0.88 (0.63–1.22) 0.95 (0.79–1.14)
Giles et al., 2006 Australia 324 Postmenopausal 1 s.d. 0.98 (0.88–1.10) 1.19 (0.93–1.52)
Table 5 Prospective studies of dietary fibre and breast cancer risk
First author
and year
Verhoeven
et al., 1997
Terry et al.,
2002
Cho et al.,
2003
Holmes
et al., 2004
Holmes
et al., 2004
Mattisson
et al., 2004
Giles et al.,
2006
Country N
cases
Menopausal
status
Carbohydrates and cancer
TJ Key and EA Spencer
Comparison Relative risk in highest category
Total fibre Cereal fibre Fruit fibre Vegetable fibre
Netherlands 650 Postmenopausal Highest
vs lowest fifth
0.83 (0.56–1.24)
Canada 2536 Any Highest
vs lowest fifth
0.92 (0.78–1.09) 0.90 (0.78–1.04) 1.07 (0.92–1.25) 0.90 (0.75–1.08)
USA 714 Premenopausal Highest
vs lowest fifth
0.88 (0.67–1.14) 0.91 (0.69–1.21) 1.13 (0.88–1.46) 0.97 (0.75–1.24)
USA 854 Premenopausal Highest
vs lowest fifth
0.99 (0.75–1.29) 0.99 (0.78–1.25) 0.86 (0.67–1.10) 0.95 (0.72–1.25)
USA 2924 Postmenopausal Highest
vs lowest fifth
0.96 (0.83–1.10) 1.08 (0.96–1.22) 0.92 (0.81–1.04) 0.94 (0.82–1.08)
Sweden 342 Postmenopausal Highest
vs lowest fifth
0.58 (0.40–0.84)
Australia 324 Postmenopausal 1 s.d. 1.08 (0.92–1.26) 1.08 (0.95–1.23) 1.00 (0.88–1.13) 1.07 (0.95–1.20)
S117
European Journal of Clinical Nutrition

S118
Cancer of the endometrium
The development of endometrial cancer is strongly related to
hormonal factors: both parity and use of the oral contraceptive
pill reduce risk, whereas obesity and oestrogen-only
hormone replacement therapy increase risk (Amant et al.,
2005). There are few data on dietary carbohydrates and
endometrial cancer risk; available results do not suggest an
association of risk with total carbohydrates or with dietary
fibre, but have suggested a possible association with high
glycaemic index or glycaemic load (Jain et al., 2000; Augustin
et al., 2003; Folsom et al., 2003; Silvera et al., 2005c). Since
obesity increases risk by about threefold (Bergstrom et al.,
2001; International Agency for Research on Cancer, 2002), it
is possible that any associations with the glycaemic index or
load of the diet may be explained by the effect of obesity.
One prospective study has reported that high consumption
of whole grain is associated with a reduced risk for
endometrial cancer among women who had never used
hormone replacement therapy (Kasum et al., 2001).
Cancer of the cervix
The principal cause of cancer of the cervix is persistent
infection with oncogenic strains of the human papilloma
virus (zur Hausen, 2002). The possibility that dietary factors
might influence risk has been investigated in a number of
studies, but these have focused largely on carotenoids and
folic acid, with little attention given to carbohydrates
(Garcia-Closas et al., 2005).
Cancer of the ovary
The development of ovarian cancer is strongly related to
hormonal factors: both parity and use of the oral contraceptive
pill reduce risk (Banks et al., 1997). Cramer et al.
(1989) proposed that a high consumption of lactose might
increase the risk for ovarian cancer, perhaps due to adverse
effects of galactose on the ovaries. A number of studies have
subsequently addressed this hypothesis. In a meta-analysis,
Larsson et al. (2006b) concluded that the results of case–
control studies were heterogeneous and did not suggest an
association of lactose intake with the risk for ovarian cancer.
However, they reported that in the three prospective studies
examined there was evidence of a significant positive
association between lactose consumption and ovarian
cancer risk. A pooled analysis of 12 prospective studies,
which included the three studies reviewed by Larsson et al.
(2006b), reported that the risk for ovarian cancer was higher
among people with the highest lactose intake than among
those with the lowest lactose intake (relative risk 1.19 (1.01–
1.40)), although the test for a linear trend was not
statistically significant (Genkinger et al., 2006). Overall, we
conclude that a high intake of lactose is possibly associated
with a small increase in the risk for ovarian cancer. As
proposed by Cramer et al. (1989), one possible mechanism
European Journal of Clinical Nutrition
Carbohydrates and cancer
TJ Key and EA Spencer
for such an association is a direct adverse effect of galactose
on the ovaries.
For other carbohydrates, there are few data in relation to
the risk for ovarian cancer (Schulz et al., 2004).
Prostate cancer
The causes of prostate cancer are poorly understood, and
despite considerable research on the association of dietary
factors with risk, no foods or nutrients have been clearly
established as risk factors for this disease (Chan et al., 2005).
Very little of this research has been focused on carbohydrates,
and the data available have not suggested that
there is an association between carbohydrate intake and
prostate cancer risk (Chan et al., 2005).
Bladder cancer
The principal known risk factor for bladder cancer is tobacco
(WHO, 2003b). Some studies have investigated the possible
role of dietary factors, but the results have been inconclusive
and little attention has been given to carbohydrates (Zeegers
et al., 2004).
Kidney cancer
The aetiology of kidney cancer is poorly understood. Obesity
increases risk, but the mechanism for this is unknown (Wolk
et al., 1996). Studies have suggested possible associations
with some dietary factors, but have focused mainly on the
possible increased risks associated with the consumption of
meat, eggs and dairy products and the possible reduction in
risk associated with high intakes of fruit and vegetables. The
results have been inconclusive and little attention has been
given to carbohydrates (Wolk et al., 1996; Lindblad et al.,
1997; Handa and Kreiger, 2002; Nicodemus et al., 2004;
Rashidkhani et al., 2005).
Other cancers
For other cancers, not discussed above, little or no research
has been conducted on whether dietary carbohydrates may
have any direct impact on the development of disease.
Obesity: associations with carbohydrate intake and cancer risk
Obesity is an established risk factor for cancers of the
oesophagus (adenocarcinoma) colorectum, breast, endometrium
and kidney (International Agency for Research on
Cancer, 2002; Key et al., 2004). The role of carbohydrates,
especially sucrose and fibre, in the aetiology of obesity is
discussed in the paper by van Dam and Seidell (2007) in this
supplement. Where dietary carbohydrates contribute to
obesity then they almost certainly also contribute to
increasing the risk for these particular cancer sites, but this
link may not be clearly seen in observational studies if the
magnitude of both these associations is moderate or small.

Acrylamide
Acrylamide is a chemical which has been shown to be
carcinogenic in some tests on laboratory animals, and can
be formed from carbohydrates during the production or
cooking of foods at high temperatures. The possible effects of
dietary acrylamide on cancer risk in humans have been
reviewed by the Joint FAO/WHO Expert Committee on Food
Additives (http://www.who.int/foodsafety/chem/chemicals/
acrylamide/en/). This committee concluded in 2005 that,
based on comparisons with the intake shown to cause
mammary tumours in rats, acrylamide intakes in humans are
generally low but that there is a possible human health
concern for high consumers. Epidemiological data available
are limited but have not suggested any association of
estimated dietary acrylamide intake with cancer risk in
humans (Mucci and Adami, 2005; Pelucchi et al., 2006;
Wilson et al., 2006).
Conclusions and recommendations
Carbohydrates as a group have not been intensively
investigated in epidemiological studies of diet and cancer.
There is a moderately large amount of data on the possible
association between dietary fibre and the risk for colorectal
cancer; the results of studies have varied and no firm
conclusion can be drawn, but the available data suggest
that high intakes of dietary fibre possibly reduce the risk
for colorectal cancer. There are also limited data which
suggest that high intakes of sucrose might increase the
risk for colorectal cancer and that high intakes of lactose
might increase the risk for ovarian cancer. For other
components of carbohydrates and other types of cancer,
the available data are too sparse to draw any conclusions.
Further research is needed on the possible associations
of carbohydrates with cancer risk. The systematic review
being conducted by the World Cancer Research Fund
will provide a detailed update in 2007. Further information
could also be provided by extending the Pooling Project
of Prospective Studies of Diet and Cancer to examine
specific classes of carbohydrates in relation to a range of
cancer sites.
Acknowledgements
We thank Dr Arthur Schatzkin, Professor Stephanie Smith-
Warner, Professor Carolyn Summerbell, Dr Anne Tjonneland,
Professor PA van den Brandt and Dr Martin Wiseman for their
constructive comments. The authors are supported by Cancer
Research UK.
Conflict of interest
During the preparation and peer-review of this paper in
2006, the authors and peer-reviewers declared the following
interests.
Carbohydrates and cancer
TJ Key and EA Spencer
Authors
Professor Timothy J Key: None declared.
Dr Elizabeth A Spencer: None declared.
Peer-reviewers
Dr Arthur Schatzkin: None declared.
Professor Stephanie Smith-Warner: None declared.
Professor Carolyn Summerbell: None declared.
Dr Anne Tjonneland: None declared.
Professor PA van den Brandt: None declared.
Dr Martin Wiseman: None declared.
References
Alberts DS, Martinez ME, Roe DJ, Guillen-Rodriguez JM, Marshall JR,
van Leeuwen JB et al. (2000). Lack of effect of a high-fiber cereal
supplement on the recurrence of colorectal adenomas. Phoenix
Colon Cancer Prevention Physicians’ Network. N Engl J Med 342,
1156–1162.
Amant F, Moerman P, Neven P, Timmerman D, Van Limbergen E,
Vergote I (2005). Endometrial cancer. Lancet 366, 491–505.
Augustin LS, Gallus S, Bosetti C, Levi F, Negri E, Franceschi S et al.
(2003). Glycemic index and glycemic load in endometrial cancer.
Int J Cancer 105, 404–407.
Banks E, Beral V, Reeves G (1997). The epidemiology of epithelial
ovarian cancer: a review. Int J Gynecol Cancer 7, 425–438.
Beresford SA, Johnson KC, Ritenbaugh C, Lasser NL, Snetselaar LG,
Black HR et al. (2006). Low-fat dietary pattern and risk of colorectal
cancer: the women’s health initiative randomized controlled
dietary modification trial. JAMA 295, 643–654.
Bergstrom A, Pisani P, Tenet V, Wolk A, Adami HO (2001).
Overweight as an avoidable cause of cancer in Europe. Int J Cancer
91, 421–430.
Berrington de Gonzalez A, Sweetland S, Spencer E (2003). A metaanalysis
of obesity and the risk of pancreatic cancer. Br J Cancer 89,
519–523.
Bingham SA, Norat T, Moskal A, Ferrari P, Slimani N, Clavel-
Chapelon F et al. (2005). Is the association with fiber from foods
in colorectal cancer confounded by folate intake? Cancer Epidemiol
Biomarkers Prev 14, 1552–1556.
Bonithon-Kopp C, Kronborg O, Giacosa A, Rath U, Faivre J (2000).
Calcium and fibre supplementation in prevention of
colorectal adenoma recurrence: a randomised intervention trial.
European cancer prevention organisation study group. Lancet 356,
1300–1306.
Bosch FX, Ribes J, Cleries R, Diaz M (2005). Epidemiology of
hepatocellular carcinoma. Clin Liver Dis 9, 191–211.
Bostick RM, Potter JD, Kushi LH, Sellers TA, Steinmetz KA, McKenzie
DR et al. (1994). Sugar, meat, and fat intake, and non-dietary risk
factors for colon cancer incidence in Iowa women (United States).
Cancer Causes Control 5, 38–52.
Botterweck AA, van den Brandt PA, Goldbohm RA (2000). Vitamins,
carotenoids, dietary fiber, and the risk of gastric carcinoma: results
from a prospective study after 6.3 years of follow-up. Cancer 88,
737–748.
Burkitt DP (1969). Related disease—related cause? Lancet 2,
1229–1231.
Burley VJ (1997). Sugar consumption and cancers of the digestive
tract. Eur J Cancer Prev 6, 422–434.
Burley VJ (1998). Sugar consumption and human cancer in sites
other than the digestive tract. Eur J Cancer Prev 7, 253–277.
Chan JM, Gann PH, Giovannucci EL (2005). Role of diet in prostate
cancer development and progression. J Clin Oncol 23, 8152–8160.
Cho E, Spiegelman D, Hunter DJ, Chen WY, Colditz GA, Willett WC
(2003). Premenopausal dietary carbohydrate, glycemic index,
S119
European Journal of Clinical Nutrition

S120
Carbohydrates and cancer
TJ Key and EA Spencer
glycemic load, and fiber in relation to risk of breast cancer. Cancer
Epidemiol Biomarkers Prev 11, 1153–1158.
Cramer DW, Harlow BL, Willett WC, Welch WR, Bell DA, Scully RE
et al. (1989). Galactose consumption and metabolism in relation
to the risk of ovarian cancer. Lancet 8, 66–71.
Crew KD, Neugut AI (2004). Epidemiology of upper gastrointestinal
malignancies. Semin Oncol 31, 450–464.
Crew KD, Neugut AI (2006). Epidemiology of gastric cancer. World J
Gastroenterol 12, 354–362.
De Stefani E, Boffetta P, Deneo-Pellegrini H, Mendilaharsu M,
Carzoglio JC, Ronco A (1999). Carbohydrates and risk of stomach
cancer in Uruguay. Int J Cancer 82, 618–621.
De Stefani E, Correa P, Boffetta P, Deneo-Pellegrini H, Ronco AL,
Mendilaharsu M (2004). Dietary patterns and risk of gastric cancer:
a case–control study in Uruguay. Gastric Cancer 7, 211–220.
Department of Health UK (1998). Nutritional aspects of the development
of cancer (Report of the Working Group on Diet and Cancer of the
Committee on Medical Aspects of Food and Nutrition Policy). The
Stationery Office: London.
Endogenous Hormones and Breast Cancer Collaborative Group
(2003). Body mass index, serum sex hormones, and breast cancer
risk in postmenopausal women. J Natl Cancer Inst 95, 1218–1226.
Folsom AR, Demissie Z, Harnack L (2003). Iowa women’s health
study: glycemic index, glycemic load, and incidence of endometrial
cancer: the Iowa women’s health study. Nutr Cancer 46, 119–124.
Franceschi S, Dal Maso L, Augustin L, Negri E, Parpinel M, Boyle P
et al. (2001). Dietary glycemic load and colorectal cancer risk.
Ann Oncol 12, 173–178.
Garcia-Closas R, Castellsague X, Bosch X, Gonzalez CA (2005). The
role of diet and nutrition in cervical carcinogenesis: a review of
recent evidence. Int J Cancer 117, 629–637.
Genkinger JM, Hunter DJ, Spiegelman D, Anderson KE, Arslan A,
Beeson WL et al. (2006). Dairy products and ovarian cancer: a
pooled analysis of 12 cohort studies. Cancer Epidemiol Biomarkers
Prev 15, 364–372.
Giles GG, Simpson JA, English DR, Hodge AM, Gertig DM, Macinnis
RJ et al. (2006). Dietary carbohydrate, fibre, glycaemic index,
glycaemic load and the risk of postmenopausal breast cancer. Int J
Cancer 118, 1843–1847.
Hague A, Elder DJ, Hicks DJ, Paraskeva C (1995). Apoptosis in
colorectal tumour cells: induction by the short chain fatty acids
butyrate, propionate and acetate and by the bile salt deoxycholate.
Int J Cancer 60, 400–406.
Hampel H, Abraham NS, El-Serag HB (2005). Meta-analysis: obesity
and the risk for gastroesophageal reflux disease and its complications.
Ann Intern Med 143, 199–211.
Handa K, Kreiger N (2002). Diet patterns and the risk of renal cell
carcinoma. Public Health Nutr 5, 757–767.
Higginbotham S, Zhang ZF, Lee IM, Cook NR, Buring JE, Liu S
(2004b). Dietary glycemic load and breast cancer risk in the
Women’s Health Study. Cancer Epidemiol Biomarkers Prev 13, 65–70.
Higginbotham S, Zhang ZF, Lee IM, Cook NR, Giovannucci E, Buring
JE et al. (2004a). Dietary glycemic load and risk of colorectal cancer
in the women’s health study. J Natl Cancer Inst 96, 229–233.
Holmes MD, Liu S, Hankinson SE, Colditz GA, Hunter DJ, Willett WC
(2004). Dietary carbohydrates, fiber, and breast cancer risk. Am J
Epidemiol 159, 732–739.
Howe GR, Benito E, Castelleto R, Cornee J, Esteve J, Gallagher RP
et al. (1992). Dietary intake of fiber and decreased risk of cancers of
the colon and rectum: evidence from the combined analysis of 13
case-control studies. J Natl Cancer Inst 84, 1887–1896.
http://www.who.int/foodsafety/chem/chemicals/acrylamide/en/.
Huxley R, Ansary-Moghaddam A, Berrington de Gonzalez A, Barzi F,
Woodward M (2005). Type-II diabetes and pancreatic cancer: a
meta-analysis of 36 studies. Br J Cancer 92, 2076–2083.
International Agency for Research on Cancer (2002). Overweight and
lack of exercise linked to increased cancer risk. In: IARC Handbooks
of Cancer Prevention, volume 6 World Health Organisation:
Geneva.
European Journal of Clinical Nutrition
Jacobs Jr DR, Marquart L, Slavin J, Kushi LH (1998). Whole-grain
intake and cancer: an expanded review and meta-analysis.
Nutr Cancer 30, 85–96.
Jacobs ET, Lanza E, Alberts DS, Hsu CH, Jiang R, Schatzkin A et al.
(2006). Fiber, sex, and colorectal adenoma: results of a pooled
analysis. Am J Clin Nutr 83b, 343–349.
Jain MG, Rohan TE, Howe GR, Miller AB (2000). A cohort study of
nutritional factors and endometrial cancer. Eur J Epidemiol 16,
899–905.
Jonas CR, McCullough ML, Teras LR, Walker-Thurmond KA, Thun MJ
et al. (2003). Dietary glycemic index, glycemic load, and risk
of incident breast cancer in postmenopausal women. Cancer
Epidemiol Biomarkers Prev 12, 573–577.
Johnson KJ, Anderson KE, Harnack L, Hong CP, Folsom AR (2005).
No association between dietary glycemic index or load and
pancreatic cancer incidence in postmenopausal women. Cancer
Epidemiol Biomarkers Prev 14, 1574–1575.
Kasum CM, Nicodemus K, Harnack LJ, Jacobs Jr DR, Folsom AR, Iowa
Women’s Health Study (2001). Whole grain intake and incident
endometrial cancer: the Iowa Women’s Health Study. Nutr Cancer
39, 180–186.
Key TJ, Allen NE, Spencer EA, Travis RC (2003). Nutrition and breast
cancer. Breast 12, 412–416.
Key TJ, Schatzkin A, Willett WC, Allen NE, Spencer EA, Travis RC
(2004). Diet, nutrition and the prevention of cancer. Public Health
Nutr 7 (1A), 187–200.
Larsson SC, Bergkvist L, Wolk A (2006a). Glycemic load, glycemic
index and carbohydrate intake in relation to risk of stomach
cancer: a prospective study. Int J Cancer 118, 3167–3169.
Larsson SC, Orsini N, Wolk A (2006b). Milk, milk products and
lactose intake and ovarian cancer risk: a meta-analysis of
epidemiological studies. Int J Cancer 118, 431–441.
Levi F, Pasche C, Lucchini F, Bosetti C, La Vecchia C (2002).
Glycaemic index, breast and colorectal cancer. Ann Oncol 13,
1688–1689.
Lindblad P, Wolk A, Bergstrom R, Adami HO (1997). Diet and risk of
renal cell cancer: a population-based case–control study. Cancer
Epidemiol Biomarkers Prev 6, 215–223.
Lissowska J, Gail MH, Pee D, Groves FD, Sobin LH, Nasierowska-
Guttmejer A et al. (2004). Diet and stomach cancer risk in Warsaw,
Poland. Nutr Cancer 48, 149–159.
Lowenfels AB, Maisonneuve P (2005). Risk factors for pancreatic
cancer. J Cell Biochem 95, 649–656.
Mattisson I, Wirfalt E, Johansson U, Gullberg B, Olsson H, Berglund
G (2004). Intakes of plant foods, fibre and fat and risk of breast
cancer–a prospective study in the Malmo diet and cancer cohort.
Br J Cancer 90, 122–127.
McCarl M, Harnack L, Limburg PJ, Anderson KE, Folsom AR (2006).
Incidence of colorectal cancer in relation to glycemic index and
load in a cohort of women. Cancer Epidemiol Biomarkers Prev 15,
892–896.
Michaud DS (2004). Epidemiology of pancreatic cancer. Minerva Chir
59, 99–111.
Michaud DS, Fuchs CS, Liu S, Willett WC, Colditz GA, Giovannucci E
(2005). Dietary glycemic load, carbohydrate, sugar, and colorectal
cancer risk in men and women. Cancer Epidemiol Biomarkers Prev
14, 138–147.
Michaud DS, Liu S, Giovannucci E, Willett WC, Colditz GA,
Fuchs CS (2002). Dietary sugar, glycemic load, and
pancreatic cancer risk in a prospective study. J Natl Cancer Inst
94, 1293–1300.
Mucci LA, Adami HO (2005). The role of epidemiology in understanding
the relationship between dietary acrylamide and cancer
risk in humans. Adv Exp Med Biol 561, 39–47.
Muñoz N, Plummer M, Vivas J, Moreno V, De Sanjose S, Lopez G et al.
(2001). A case–control study of gastric cancer in Venezuela. Int J
Cancer 93, 417–423.
Nagengast FM, Grubben MJ, van Munster IP (1995). Role of bile acids
in colorectal carcinogenesis. Eur J Cancer 31A, 1067–1070.

Nicodemus KK, Jacobs Jr DR, Folsom AR (2001). Whole and refined
grain intake and risk of incident postmenopausal breast cancer
(United States). Cancer Causes Control 12, 917–925.
Nicodemus KK, Sweeney C, Folsom AR (2004). Evaluation of dietary,
medical and lifestyle risk factors for incident kidney cancer in
postmenopausal women. Int J Cancer 108, 115–121.
Nielsen TG, Olsen A, Christensen J, Overvad K, Tjonneland A (2005).
Dietary carbohydrate intake is not associated with the breast
cancer incidence rate ratio in postmenopausal Danish women.
J Nutr 135, 124–128.
Nomura AM, Hankin JH, Kolonel LN, Wilkens LR, Goodman MT,
Stemmermann GN (2003). Case–control study of diet and other
risk factors for gastric cancer in Hawaii (United States). Cancer
Causes Control 14, 547–558.
Otani T, Iwasaki M, Ishihara J, Sasazuki S, Inoue M, Tsugane S et al.
(2006). Dietary fiber intake and subsequent risk of colorectal
cancer: the Japan public health center-based prospective study. Int
J Cancer 119, 1475–1480.
Pandey M (2003). Risk factors for gallbladder cancer: a reappraisal.
Eur J Cancer Prev 12, 15–24.
Park Y, Hunter DJ, Spiegelman D, Bergkvist L, Berrino F, van den
Brandt PA et al. (2005). Dietary fiber intake and risk of colorectal
cancer: a pooled analysis of prospective cohort studies. JAMA 294,
2849–2857.
Pelucchi C, Galeone C, Levi F, Negri E, Franceschi S, Talamini R et al.
(2006). Dietary acrylamide and human cancer. Int J Cancer 118,
467–471.
Pera M, Manterola C, Vidal O, Grande L (2005). Epidemiology of
esophageal adenocarcinoma. J Surg Oncol 92, 151–159.
Peters U, Sinha R, Chatterjee N, Subar AF, Ziegler RG, Kulldorff M
et al. (2003). Dietary fibre and colorectal adenoma in a colorectal
cancer early detection programme. Lancet 361, 1491–1495.
Potischman N, Coates RJ, Swanson CA, Carroll RJ, Daling JR, Brogan
DR et al. (2002). Increased risk of early-stage breast cancer related
to consumption of sweet foods among women less than age 45 in
the United States. Cancer Causes Control 13, 937–946.
Potter JD (1999). Colorectal cancer: molecules and populations.
J Natl Cancer Inst 91, 916–932.
Rashidkhani B, Akesson A, Lindblad P, Wolk A (2005). Major dietary
patterns and risk of renal cell carcinoma in a prospective cohort of
Swedish women. JNutr135, 1757–1762.
Robertson DJ, Sandler RS, Haile R, Tosteson TD, Greenberg ER, Grau
M et al. (2005). Fat, fiber, meat and the risk of colorectal
adenomas. Am J Gastroenterol 100, 2789–2795.
Romieu I, Lazcano-Ponce E, Sanchez-Zamorano LM, Willett W,
Hernandez-Avila M (2004). Carbohydrates and the risk of breast
cancer among Mexican women. Cancer Epidemiol Biomarkers Prev
13, 1283–1289.
Schatzkin A, Lanza E, Corle D, Lance P, Iber F, Caan B et al. (2000).
Lack of effect of a low-fat, high-fiber diet on the recurrence of
colorectal adenomas. Polyp Prevention Trial Study Group. N Engl J
Med 342, 1149–1155.
Schernhammer ES, Hu FB, Giovannucci E, Michaud DS, Colditz GA,
Stampfer MJ et al. (2005). Sugar-sweetened soft drink consumption
and risk of pancreatic cancer in two prospective cohorts. Cancer
Epidemiol Biomarkers Prev 14, 2098–2105.
Schulz M, Lahmann PH, Riboli E, Boeing H (2004). Dietary
determinants of epithelial ovarian cancer: a review of the
epidemiologic literature. Nutr Cancer 50, 120–140.
Silvera SA, Jain M, Howe GR, Miller AB, Rohan TE (2005b). Dietary
carbohydrates and breast cancer risk: a prospective study of the
Carbohydrates and cancer
TJ Key and EA Spencer
roles of overall glycemic index and glycemic load. Int J Cancer 114,
653–658.
Silvera SA, Rohan TE, Jain M, Terry PD, Howe GR, Miller AB (2005a).
Glycemic index, glycemic load, and pancreatic cancer risk
(Canada). Cancer Causes Control 16, 431–436.
Silvera SA, Rohan TE, Jain M, Terry PD, Howe GR, Miller AB (2005c).
Glycaemic index, glycaemic load and risk of endometrial
cancer: a prospective cohort study. Public Health Nutr 8,
912–919.
Slattery ML, Benson J, Berry TD, Duncan D, Edwards SL, Caan BJ,
Potter JD (1997). Dietary sugar and colon cancer. Cancer Epidemiol
Biomarkers Prev 6, 677–685.
Stolzenberg-Solomon RZ, Pietinen P, Taylor PR, Virtamo J, Albanes D
(2002). Prospective study of diet and pancreatic cancer in male
smokers. Am J Epidemiol 155, 783–792.
Terry P, Jain M, Miller AB, Howe GR, Rohan TE (2002). No association
among total dietary fiber, fiber fractions, and risk of breast cancer.
Cancer Epidemiol Biomarkers Prev 11, 1507–1508.
Terry P, Lagergren J, Ye W, Wolk A, Nyren O (2001). Inverse
association between intake of cereal fiber and risk of gastric cardia
cancer. Gastroenterology 120, 387–391.
Terry PD, Jain M, Miller AB, Howe GR, Rohan TE (2003). Glycemic
load, carbohydrate intake, and risk of colorectal cancer in
women: a prospective cohort study. J Natl Cancer Inst 95,
914–916.
van Dam RM, Seidell JC (2007). Carbohydrate intake and obesity. Eur
J Clin Nutr 61 (Suppl 1), S75–S99.
van den Brandt PA, Spiegelman D, Yaun SS, Adami HO, Beeson L,
Folsom AR et al. (2000). Pooled analysis of prospective cohort
studies on height, weight, and breast cancer risk. Am J Epidemiol
152, 514–527.
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.
Verhoeven DT, Assen N, Goldbohm RA, Dorant E, van ‘t Veer P,
Sturmans F et al. (1997). Vitamins C and E, retinol, beta-carotene
and dietary fibre in relation to breast cancer risk: a prospective
cohort study. Br J Cancer 75, 149–155.
Wilson KM, Rimm EB, Thompson KM, Mucci LA (2006). Dietary
acrylamide and cancer risk in humans: a review. J Verbr Lebensm 1,
19–27.
Wolk A, Lindblad P, Adami HO (1996). Nutrition and renal cell
cancer. Cancer Causes Control 7, 5–18.
World Cancer Research Fund (1997). Food, Nutrition, and the
Prevention of Cancer: A Global Perspective. American Institute for
Cancer Research: Washington DC.
World Health Organization (2003a). Diet, Nutrition and the Prevention
of Chronic Diseases. WHO Technical Report Series 916. WHO:
Geneva.
World Health Organization (2003b). World Cancer Report. BW Stewart
and P Kleihues (eds) IARC Press: Lyon.
Zatonski WA, Lowenfels AB, Boyle P, Maisonneuve P, Bueno de
Mesquita HB, Ghadirian P et al. (1997). Epidemiologic aspects of
gallbladder cancer: a case-control study of the SEARCH program of
the international agency for research on cancer. J Natl Cancer Inst
89, 1132–1138.
Zeegers MP, Kellen E, Buntinx F, van den Brandt PA (2004). The
association between smoking, beverage consumption, diet
and bladder cancer: a systematic literature review. World J Urol
21, 392–401.
zur Hausen H (2002). Papillomaviruses and cancer: from basic studies
to clinical application. Nat Rev Cancer 2, 342–350.
S121
European Journal of Clinical Nutrition

REVIEW
Glycemic index and glycemic load: measurement
issues and their effect on diet–disease relationships
BJ Venn and TJ Green
Department of Human Nutrition, University of Otago, Dunedin, New Zealand
Glycemic index (GI) describes the blood glucose response after consumption of a carbohydrate containing test food relative to a
carbohydrate containing reference food, typically glucose or white bread. GI was originally designed for people with diabetes as
a guide to food selection, advice being given to select foods with a low GI. The amount of food consumed is a major
determinant of postprandial hyperglycemia, and the concept of glycemic load (GL) takes account of the GI of a food and the
amount eaten. More recent recommendations regarding the potential of low GI and GL diets to reduce the risk of chronic
diseases and to treat conditions other than diabetes, should be interpreted in the light of the individual variation in blood
glucose levels and other methodological issues relating to measurement of GI and GL. Several factors explain the large inter- and
intra-individual variation in glycemic response to foods. More reliable measurements of GI and GL of individual foods than are
currently available can be obtained by studying, under standard conditions, a larger number of subjects than has typically been
the case in the past. Meta-analyses suggest that foods with a low GI or GL may confer benefit in terms of glycemic control in
diabetes and lipid management. However, low GI and GL foods can be energy dense and contain substantial amounts of sugars
or undesirable fats that contribute to a diminished glycemic response. Therefore, functionality in terms of a low glycemic
response alone does not necessarily justify a health claim. Most studies, which have demonstrated health benefits of low GI or GL
involved naturally occurring and minimally processed carbohydrate containing cereals, vegetables and fruit. These foods have
qualities other than their immediate impact on postprandial glycemia as a basis to recommend their consumption. When the GI
or GL concepts are used to guide food choice, this should be done in the context of other nutritional indicators and when values
have been reliably measured in a large group of individuals.
European Journal of Clinical Nutrition (2007) 61 (Suppl 1), S122–S131; doi:10.1038/sj.ejcn.1602942
Keywords: glycemic index; glycemic load; methodology; diet–disease relationships
Introduction
European Journal of Clinical Nutrition (2007) 61 (Suppl 1), S122–S131
& 2007 Nature Publishing Group All rights reserved 0954-3007/07 $30.00
www.nature.com/ejcn
The glycemic index (GI) concept was introduced by Jenkins
et al. (1981) in the early 1980s as a ranking system for
carbohydrates based on their immediate impact on blood
glucose levels. GI was originally designed for people with
diabetes as a guide to food selection, advice being given to
select foods with a low GI (Jenkins et al., 1983). Lower GI
foods were considered to confer benefit as a result of the
relatively low glycemic response following ingestion compared
with high GI foods. The GI concept has been extended
to also take into account the effect of the total amount of
carbohydrate consumed. Thus glycemic load (GL), a product
of GI and quantity of carbohydrate eaten provides an
Correspondence: Dr B Venn, PO Box 56, Department of Human Nutrition,
University of Otago, Dunedin, New Zealand.
E-mail: bernard.venn@stonebow.otago.ac.nz
indication of glucose available for energy or storage following
a carbohydrate containing meal. Although GI is usually
tested on individual foods, there are methods described
whereby the GI and GL of meals and habitual diets can be
estimated (Wolever and Jenkins, 1986; Salmeron et al.,
1997a). In addition to a role in the treatment of diabetes,
low GI and GL diets have more recently been widely
recommended for the prevention of chronic diseases including
diabetes, obesity, cancer and heart disease and in the
treatment of cardiovascular risk factors, especially dyslipidaemia
(Jenkins et al., 2002).
The usefulness of GI and GL has been questioned on
several counts: failure to consider the insulin response
(Coulston et al., 1984), the high intra- and inter-subject
variation in glucose response to a food (Pi-Sunyer, 2002), and
a loss of discriminating power when foods are combined in a
mixed meal (Flint et al., 2004). Furthermore, foods with a
high sugar (sucrose) content and those containing both

carbohydrate and fat may have a low GI, but may not be
regarded as particularly appropriate choices because of their
energy density and nature of dietary fat (Freeman, 2005).
This review considers the reliability of the measurement and
the practical application of GI. Its value in relating dietary
attributes to chronic diseases is considered in other papers in
this series.
Definition and measurement
GI is defined as the blood glucose response measured as area
under the curve (AUC) in response to a test food consumed
by an individual under standard conditions expressed as a
percentage of the AUC following consumption of a reference
food consumed by the same person on a different day (FAO/
WHO, 1998). The test food and reference food (usually 50 g
glucose) must contain the same amount of available
carbohydrate (Figure 1). It is important to standardize GI
testing conditions, and the procedure for the measurement
of GI is described in detail in the 1998 FAO/WHO report on
carbohydrates in human nutrition (FAO, 1998). Hundreds of
foods have been tested for GI with the aim of ranking foods
within and between food categories. A GI classification
system is in common use in which foods are categorized as
having low (o55), medium (55–69) or high GI (470) (Brand-
Miller et al., 2003a).
Glucose, a monosaccharide, induces a large glycemic
response and is often used as the reference food and assigned
a GI of 100. Some polysaccharides, such as those present in
instant potato for example, may also result in large glycemic
Blood glucose (mmol/L)
10
8
6
4
2
0
Blood glucose (mmol/L)
8
7
6
5
4
3
2
1
0
0 30 60 90 120
Time (min)
responses when consumed in an amount containing 50 g
available carbohydrate because of rapid and near complete
digestion and absorption in the small intestine. From the
International Tables, the GI for instant potato, determined as
the mean of six studies, was 85 (Foster-Powell et al., 2002).
Sucrose, a disaccharide of glucose and fructose, has a
somewhat lower GI of 68, resulting from the fructose
component which has an exceptionally low GI of 19. Adding
protein or fat to a carbohydrate containing food can also
lower overall GI (Miller et al., 2006). Resistant starch and
dietary fibre are largely undigested and not absorbed in the
small intestine and therefore contribute little to postprandial
glycemia. However, a lowering of glycemic response has
been found when purified extracts of fibre, particularly of
the type that forms a viscous gel in water such as guar gum,
are added to a test food in sufficient quantity (Jenkins et al.,
1976; Doi et al., 1979; Wolever et al., 1991; Tappy et al.,
1996). GI cannot be predicted from the fibre content of a
carbohydrate containing food or from the terms wholemeal
and wholegrain for which there are no universally accepted
definitions. For example, from the International Tables, the
mean GI of wholemeal bread from 13 studies is 71, while
that of white wheat bread (mean of six studies) is 70 (Foster-
Powell et al., 2002). Whole grains, when largely intact, have
been found to lower GI (Jenkins et al., 1986; 1988; Liljeberg
et al., 1992; Granfeldt et al., 1994; 1995), but wholegrain
products contain a variable proportion of intact grains.
GI does not take into account the amount of carbohydrate
consumed, an important determinant of glycemic response.
For example, watermelon has a high GI (Foster-Powell et al.,
2002) and may not be considered a good food selection as
0 30 60 90 120
Time (min)
Blood glucose (mmol/L)
Glycemic index and glycemic load
BJ Venn and TJ Green
10
8
6
4
2
0
0 30 60 90 120
Time (min)
Figure 1 Example of an individual’s data used to estimate glycemic index (GI). Area under the curve (AUC) refers to the area included between
the baseline and incremental blood glucose points when connected by straight lines. The area under each incremental glucose curve is
calculated using the trapezoid rule (note: only areas above the baseline are used). GI ¼ AUCFood/mean (AUCReference) 100.
S123
European Journal of Clinical Nutrition

S124
Blood glucose iAUC (mmol/L·min)
350
300
250
200
150
100
50
0
0 12.5 25 37.5 50 62.5 75
Available carbohydrate (g)
Figure 2 Blood glucose area under the curve (AUC) responses to
increasing amounts of glucose and granola bar tested in 20 people.
part of a low GI diet. However, watermelon only contains 5 g
of carbohydrate per 100 g, thus it would have a minimal
glycemic effect. GL takes into account how much carbohydrate
a serving of a food contains and may be determined by
indirect and direct methods.
The indirect method involves multiplying the GI of a food
by the amount of available carbohydrate in the portion of
food consumed. This method implies that GL is directly
proportional to the amount of the particular food eaten. This
is perhaps counterintuitive, because the blood glucose AUC
does not increase in direct proportion to the amount
consumed. For example, eating six times the amount of
bread results in an approximately threefold increase in AUC
(Brand-Miller et al., 2003c). In other words, as the amount of
food increased, the rate of increase in AUC declines, an effect
shown in Figure 2 (Venn et al., 2006). Therefore, it is implicit
in the calculation of GL that the AUC for both the test and
the reference foods are attenuated to the same degree with
increasing amounts consumed.
Glycemic equivalence is a method of directly determining
GL. For each subject an AUC for glucose is calculated for a
range of doses of the reference food measured on different
days. A standard curve is constructed for each subject with
increasing amounts of the reference on the x axis with its
corresponding AUC for blood glucose on the y axis (Venn
et al., 2006). The AUC in response to a food consumed at any
portion size, typically a usual serving, is compared to that
individual’s glucose standard curve as depicted in Figure 3
(Venn et al., 2006). Using this technique, glycemic equivalence
is the amount of glucose that would theoretically
produce the same blood glucose AUC as that particular
portion size of food consumed. Major drawbacks of the direct
method are increased time and cost required to determine
the GL of a food. The reference must be tested at several
doses in each subject and the GL of a food cannot be
estimated from currently available GI values. Data from our
laboratory support the premise that GL is linearly related to
the amount of food consumed that is, GL calculated using
European Journal of Clinical Nutrition
Glycemic index and glycemic load
BJ Venn and TJ Green
Blood glucose iAUC (mmol/L·min)
175
150
125
100
75
50
25
0
iAUC food
0 12.5 25 37.5 50 62.5 75
Glycaemic load (g)
GL direct measure
Figure 3 Example of an individual’s standard glucose curve
generated using glucose doses of 12.5, 25, 50 and 75 g. A test food
is consumed and the resulting area under the curve (AUC) used to
impute the glycemic load. AUC refers to the area included between
the baseline and incremental blood glucose points when connected
by straight lines. The area under each incremental glucose curve is
calculated using the trapezoid rule (Note: only areas above the
baseline are used.)
Table 1 Examples of GL arranged by classification taken from the
international tables (Foster-Powell et al., 2002)
Food GI Serving
size (g)
Available
carbohydrate (g)
Watermelon 72 120 6 4
Ice cream (high fat) 37 50 9 4
Mashed potato 74 150 20 15
Macaroni 47 180 48 23
Parboiled rice 64 150 36 23
Chocolate bar 65 60 40 26
Porridge 58 250 22 13
Corn flakes 81 30 26 21
Abbreviations: GI, glycaemic index; GL, glycaemic load.
GI available carbohydrate agrees well with GL measured
directly, at least when food is consumed over a range of usual
intakes (Venn et al., 2006). A GL classification system is used
in which foods are categorized as having low (p10), medium
(410–o20) or high GL (X20).
The relationship between GI and GL is not straightforward;
for example, a high GI food can have a low GL if eaten
in small quantities. Conversely, a low GI food can have a
high GL dependent upon the portion size eaten. This effect is
demonstrated in Table 1, in which various foods from the
International Tables have been selected (Foster-Powell et al.,
2002). A ‘serving size’ of watermelon, a high GI food, has the
same GL as a serving size of high fat ice cream, a low GI food.
Mashed potato and macaroni may be contrasted with the
lower GI food (macaroni) having a higher GL per serving.
GL

Foods having very different nutrient profiles can have
similar GIs and GLs per serving, such as parboiled rice and
a chocolate bar. GI and GL can also be positively related to
each other, for example comparing porridge and corn flakes,
in which the higher GI food (corn flakes) predicts a higher
GL per serving. Although a food is assigned a fixed GI value,
any food could have a low, medium or high GL because GL is
dependent upon the amount eaten.
The glycemic load of a diet can be calculated by summing
the glycemic loads for all foods consumed in the diet. A low
GL diet could be achieved by choosing small servings of
foods relatively high in carbohydrate having a low GI.
Alternatively, a low GL diet could comprise foods having a
high fat, high protein, low carbohydrate content. The
heterogeneity of foods that could be used to construct a
low GL diet indicates that food selection should not be made
on GL alone. Knowledge of other qualities of the food, for
example fat content, type of fat, energy density, fibre
content and appropriate serving size should be taken into
consideration.
GI and GL labelling
Voluntary GI labelling of foods by food manufacturers occurs
in several countries. Products may need to meet nutritional
compositional requirements to be eligible for GI testing and
labelling, such as a limit on the type or amount of fat
contained in the food. However, compositional requirements
are not standardized either within a country, where
more than one laboratory may provide a GI-testing service,
or among countries around the world. Standardized eligibility
criteria would give consumers, health professionals
and regulators more confidence in the suitability of a food to
display its GI. It could be argued that GL should be labeled
because GL more closely reflects the glycaemic impact
associated with consuming an amount of the food.
Factors affecting the measurement
Postprandial glucose concentrations are dependent upon
several factors. In people with impaired glucose tolerance
and diabetes the glycemic response measured as blood
glucose AUC is increased compared with healthy individuals.
However, GI is the AUC in response to a test food relative to
that of a reference food and given that each person acts as
his/her own control the GI of a food should not differ in
those with and without abnormalities of glucose metabolism.
GI testing has been carried out, and values published in
international tables, using normoglycaemic individuals as
well as those with impaired glucose tolerance (Foster-Powell
et al., 2002). Despite broadly comparable results Brouns et al.
(2005) have recommended using people with normal glucose
tolerance for the determination of GI because variability in
glycemic response is greater in people with impaired glucose
tolerance or diabetes.
Glycemic index and glycemic load
BJ Venn and TJ Green
The use of a test food referenced to a standard could be
used as an argument that GI is a property of food, rather than
a characteristic of the individual consuming the food.
However, postprandial glycemia may be influenced by the
extent to which individuals chew food prior to swallowing
(Read et al., 1986; Suzuki et al., 2005) as well as the expected
biological variation in rates and extent of digestion and
absorption. These variables may not apply equally to test and
reference foods; a reference food commonly used is a glucose
beverage. The observed intra- and inter-individual differences
in GI and GL which are apparent even when measured
under standardized conditions may be further exaggerated
by differing physical and chemical nature of apparently
similar food products (Wolever, 1990). For example, gelatinization,
the process of rendering starches water soluble;
retrogradation, a realignment of starch molecules during
cooling and storage; starch type; and dietary fibre, are all
factors with potential glycemic-modifying effects. Some of
these factors are affected by cooking times and methods, and
the temperature of the food consumed, potentially providing
a source of variability both in GI measurement and in
day-to-day variability of glycemic responses to the same
food.
Reliability of GI values for individual foods
The 1998 Joint FAO/WHO Expert Consultation on carbohydrates
suggested that for the determination of the GI of a
food, six subjects would be required; although the basis for
this number was not given (FAO, 1998). More recently, it has
been recommended that a sample of 10 should be used, on
the grounds that it allows for a ‘reasonable degree of power
and precision for most purposes’, although it was acknowledged
that more people would be necessary if greater
precision was required (Brouns et al., 2005). However, there
are strong indications that using 10 people is insufficient to
obtain reliable estimates of GI, particularly if GI levels are
high, because variance increases with the mean. Large
variation in glycemic response between- and within-people
makes it difficult to show differences among foods. For
example, Henry et al. (2005) tested eight varieties of potato
in groups of 10 people and reported mean7s.e.m. GIs
ranging from 5673 to94716. Despite a wide range of GIs, it
was not possible to demonstrate statistically significant
differences among the potato varieties. Because of the large
variation there is the potential to miss-classify foods into
categories of low, medium, or high GI.
In an inter-laboratory study, seven laboratories tested the
GIs of centrally provided foods, each using 8–12 participants
(Wolever et al., 2003). A range in mean GI values among
laboratories was obtained for each of the test foods; potato
65.2744.6–98.5720.6; bread (locally sourced) 64.2715.4–
78.9726.1; rice 54.8724.1–85.0728.6; spaghetti 36.4735.8
–69.9718.8; and barley 23.2724.6–47.1749.7. Rice would
have been classified as low GI (54.8724.1) by one laboratory,
S125
European Journal of Clinical Nutrition

S126
medium GI (62.6725.0, 63.378.1, 68.4748.0) by three
laboratories, and high GI (76.9712.9, 85.0728.6,
87.0775.9) by the other three laboratories. It appeared that
results were more consistent when GI was calculated using
capillary blood obtained by finger prick rather than venous
blood. A much better estimate was obtained when data from
the laboratories using capillary blood were combined. Using
a pooled sample of 47 participants, mean GI for rice was 69
with narrower confidence intervals (95% CI: 63, 76). Thus, in
addition to indicating the most appropriate method for
blood sampling, the data demonstrate the enhanced reliability
of a measurement when studying large numbers of
individuals. Most of the variation in GI differences between
laboratories was attributed to random within-person variability
(Wolever et al., 2003). There is no ready explanation for
this random day-to-day variability in glycemic response that
occurs even to repeat challenges of the same food under
standardized conditions.
Within-person variability can be reduced to some extent
by increasing the number of replicates for each subject. The
current recommendation is that the reference food should be
tested two or three times in each subject (Brouns et al., 2005).
However, within-person variability is also present for the test
food. Using data obtained from our laboratory in which a
test food and a reference food were tested three times and
four times, respectively, in 20 people, we have calculated
sample sizes necessary to be confident of a difference of 10
GI units between foods. The sample size is dependent on the
level of GI. For a difference of 10 units, between 30 and 40
for instance, it was estimated that 25 people would be
required if the food was tested once and the reference food
three times, or 19 people if the test and reference foods were
both tested twice. These estimates are comparable with the
sample size estimates shown by Brouns et al. (2005). The
same likelihood of detecting differences of 10 GI units
between foods having GIs toward the upper end of the scale
(70 and 80), would require a sample size of 114 people
testing the food once and the reference food three times, or
86 people if the test and reference foods are both tested
twice. Increasing the sample and/or repeating the test food
would appreciably increase the cost of testing, perhaps a
necessary expense, if more precision in GI measurement is to
be achieved.
Mixed meals
The ranking of meals by GI has been found to reflect the
ranking of the major carbohydrate component in the meal.
For example, baked potato was found to have a higher GI
than rice (Jenkins et al., 1984). When these foods were
incorporated into meals, there was a tendency for the
postprandial glycemic ranking of the meals to be maintained
according to the GI ranking of the food that provided the
major carbohydrate source (Wolever and Jenkins, 1986).
However, there is debate as to whether summing the
European Journal of Clinical Nutrition
Glycemic index and glycemic load
BJ Venn and TJ Green
individual GIs of foods in a meal can be used to reliably
calculate the GI of the meal. Flint et al. (2004) used GI values
taken from the International Table of GI (Foster-Powell et al.,
2002) to predict the GI of 13 simple breakfast meals, each
providing 50 g available carbohydrate. There was no association
between the GI calculated from the Tables and the
measured GI. It was pointed out that the energy content of
the meals had not been standardized (Brand-Miller and
Wolever, 2005); however, it might be argued that GI testing
is standardized to an available amount of carbohydrate, not
to energy content. In another study, a closer relationship was
reported between calculated GI and glycemic responses to
various breakfast meals, but the agreement was not entirely
consistent (Wolever et al., 2006). Two of the meals, a bagel
with cream cheese and orange juice; and a meal of rye bread,
margarine, cereal, milk, sugar and orange juice each
contained approximately 69 g available carbohydrate. The
mean7s.e. glycemic responses to the meals, measured as
AUC, were similar (148714 and 143713 mmol/l min, respectively).
However, using published values the calculated
GIs of the meals were predicted to be 67 and 51, respectively.
Sugiyama et al. (2003) found that the ingestion of milk with
rice resulted in a significantly lower GI than when rice was
eaten alone. When cheese was added to potato, a dramatic
lowering of the potato’s mean7s.e.m. GI from 9378 to
3975 was found (Henry et al., 2006). Thus while it is clear
that combining foods does influence GI and that the
addition of protein and fat to a carbohydrate containing
meal can appreciably reduce the glycemic response (Collier
and O’Dea, 1983; Nuttall et al., 1984) there is insufficient
information to accurately predict the effect of different
combinations of foods. Aggregating the GIs of individual
components of a meal does not reliably predict the observed
GI of the meal as a whole.
Published glycemic index and glycemic load values
Care must be taken when using published GI and GL values.
Variability of GI and GL among apparently similar foods has
led to recommendations that some foods, for example rice,
should be tested in the geographical region in which they are
consumed. This may be necessary to account for differences
in variety and cooking conditions (Foster-Powell et al., 2002).
Testing in specific populations may also be important if GI is
not solely a property of food. Mettler et al. (2007) found that
the training state of athletes affected GI of the same food. It
was suggested that Flint et al. (2004) who reported that
combining the GI of single foods did not predict the GI of
mixed meals, chose incorrect GI values from published tables
(Wolever et al., 2006). This problem is likely to occur when
multiple entries for the same food are presented. For
example, in the International Tables baked Russet Burbank
potatoes eaten without added fat (that is butter) are listed as
having GI values of 56, 78, 94 and 111 (Foster-Powell et al.,
2002). Multiple entries for the same foods also complicate

esearch activities and food selection for individuals trying to
follow a low GI diet. For example, boiled carrots have mean
GI values listed in the International Tables of 92, 49 and 32
(Foster-Powell et al., 2002). If GI were a major criterion in
food selection, a value of 92 might have discouraged people
from eating boiled carrots. On the other hand, low GI values
of 32 and 49 would have suggested boiled carrots as a highly
suitable choice. It has been suggested that reliability of the
estimates might have contributed to the differences in GI
(Foster-Powell et al., 2002), an argument in favour of
studying large numbers of individuals under the standardized
conditions described above.
Dietary instruments
Questions have been raised regarding the appropriateness of
the dietary instruments used when examining the relationship
between GI and GL and various diseases in observational
studies (Pi-Sunyer, 2002). Food frequency
questionnaires used in several studies were not designed
specifically to obtain information on GI and GL. Individual
foods were not assigned GI or GL values, rather they were
collapsed into categories. Most of the published studies have
not described how foods were grouped, but an Australian
study gave some insight into the process (Hodge et al., 2004).
The food frequency questionnaire used in the Australian
study had a category of ‘cereal foods, cakes and biscuits’.
Within that category were 17 items, two of which were
‘muesli’ and ‘other breakfast cereals’. A GI value of 46% was
assigned to muesli and 62% to ‘other breakfast cereals.’ The
use of a single value to describe the GI of breakfast cereals
seemed inappropriate given that the GIs of Australian cereals
range from 30 (All-Bran) to 85 (Rice Bubbles) (Foster-Powell
et al., 2002). Pi-Sunyer has drawn attention to the even
broader groupings that were used in the Health Professionals’
Follow-up Study and the Nurses’ Health Study.
Categories used were—all whole grains; all refined grains;
all cold breakfast cereals; all fruit; and all fruit juices (Pi-
Sunyer, 2002). Such observations do raise some concerns
about the degree of confidence that can be placed on the
findings of these cohort studies with respect to dietary GI or
GL and disease outcome. On the other hand one might argue
that misclassification leads to an underestimate of the true
association between GI and GL and disease (see paper by
Mann in this series). The use of dietary questionnaires
specifically designed to gather information on GI and GL,
such as those currently being developed, may be expected to
generate data which can be interpreted with greater
confidence (Flood et al., 2006; Neuhouser et al., 2006).
Overall comment on reliability
Biological variation, differing chemical and physical structure
of apparently similar foods and method of food
Glycemic index and glycemic load
BJ Venn and TJ Green
preparation and consumption may all contribute to the
marked inter- and intra-individual variation observation in
the glycemic response to foods. GI and GL testing on larger
numbers of individuals than previously undertaken or
increasing the number of replicates carried out on an
individual will improve the reliability and precision of GI
estimates. Specifying origin and other details of product (for
example, variety of fruit or vegetable) will further enhance
confidence in the measurement. However, the usefulness of
the index will always be limited to some extent by the
variation between and within individuals. It is common
practice to place foods into broad categories of GI. Classification
works well when there is a large separation in GI values
between foods, but there are still uncertainties into which
category of GI many foods belong because of the variability
around group mean GI values. More certainty in the relative
ranking of foods by GI would be attained if larger group sizes
were used to estimate GI. The degree to which limitations of
currently available data influence present use of the concepts
of GI and GL in understanding cause or influencing
management of disease is considered further in the section
on recommendations.
Glycemic index and glycemic load and human
health
The GI/GL concept has been widely advocated as a means of
identifying foods that might protect against chronic diseases
or be useful in disease management. Potential protection
against diabetes and cardiovascular disease is considered in
the paper by Mann (2007). Several other health issues are
discussed here.
Long-term glycemic control in diabetes
The GI concept was used in the management of diabetes
before being used in other clinical situations. Glycated
haemoglobin (HbA1C) is measured in people with diabetes
to assess overall glycemic control over a period of approximately
2–3 months prior to the measurement being made
(Goldstein et al., 2004). Fructosamine, another glycated
protein, is also occasionally used as a measure of glucose
control over the preceding 2–3 weeks. Several studies have
been carried out in people with diabetes to examine the
effect on HbA1C or fructosamine of diets differing principally
with respect to GI. Data from these studies form the basis of
two meta-analyses. There was a modest reduction in HbA 1C
in people consuming low GI diets, estimated to be 0.33%
units (95% CI: 0.07, 0.59) in one meta-analysis (Brand-Miller
et al., 2003b), and 0.27% units (95% CI: 0.03, 0.5) in the
other (Opperman et al., 2004). Fructosamine concentrations
were also lower in favour of low GI dietary interventions. In
one meta-analysis, the estimated difference between low and
high GI dietary periods was 0.19 mmol/l (95% CI: 0.06, 0.32)
S127
European Journal of Clinical Nutrition

S128
(Brand-Miller et al., 2003b), and in the other 0.1 mmol/l
(95% CI: 0.00, 0.20) (Opperman et al., 2004). Although these
reductions in HbA1C or fructosamine are small it is important
to note that these effects are in addition to other dietary
changes or pharmacological treatments used in diabetes
management. Whether changes in glycated proteins of this
magnitude affect long-term health outcomes is untested.
Trials using drugs such as acarbose, which lower postprandial
hyperglycaemia, suggest that acarbose may be effective in
reducing cardiovascular complications in people with type 2
diabetes mellitus (Hanefeld et al., 2004). However, reductions
in HbA1C of 0.6–0.8% were achieved in these trials (Hanefeld
et al., 2004; van de Laar et al., 2005). Nevertheless, any
dietary strategy that resulted in improved glycaemic control
would be welcome and given the difference in the acute
effect that low and high GI foods have on postprandial
hyperglycaemia, the proposition that changing foods in the
diet from high to low GI might improve markers of glycemic
control is entirely plausible. However, some caveats may be
appropriate. In many of the studies included in the metaanalyses
described above, the low GI foods tended to have
low energy density and a high fibre content, such as whole
fruit, oats, whole grain, pulses and pasta (Frost et al., 1998;
Heilbronn et al., 2002). Thus, modest changes in glycaemic
control were achieved under study conditions that required
people to be compliant with relatively major changes in
dietary habits. The findings may not necessarily apply to the
many low GI functional and convenience foods currently
available, which may be relatively high in sugars and energy
dense.
Glycemic index and glycemic load and blood lipids
Relationships between dietary GI and blood lipid fractions
have been assessed in several prospective observational
studies. A reasonably consistent finding has been an inverse
association between fasting HDL cholesterol concentrations
and dietary GI (Liu et al., 2001; Amano et al., 2004; Slyper
et al., 2005), although one study found no association
(Murakami et al., 2006). Ma et al. (2006) found inverse
associations between dietary GI and GL in a cross-sectional
analysis, but the associations were lost during follow-up. An
inverse association between GI and HDL-cholesterol concentration
has also been found in a nationally representative
sample of US adults (Ford and Liu, 2001).
Findings from intervention trials differed from those of
observational studies. Kelly et al. (2004) conducted a metaanalysis
of intervention trials that had examined the effect of
low GI diets on coronary heart disease risk factors. Results
from that analysis showed limited and weak evidence of an
inverse relationship between GI and total cholesterol, with
no effect of dietary GI on LDL and HDL cholesterol,
triglycerides, fasting glucose and fasting insulin. Opperman
et al. (2004) conducted a meta-analysis of 14 randomized
controlled trials relating to the effects on blood lipids of
European Journal of Clinical Nutrition
Glycemic index and glycemic load
BJ Venn and TJ Green
altering the GI of test diets. There was a difference in total
and LDL-cholesterol concentrations of 0.33 (95% CI: 0.18,
0.47) mmol/l and 0.15 (95% CI: 0.00, 0.31) mmol/l, favoring
the low GI diets, but no difference in HDL cholesterol
concentrations between people consuming low and high GI
diets.
Thus, the reasonably consistent finding in observational
studies of an inverse association between dietary GI and HDL
cholesterol concentrations is not confirmed by meta-analyses
of randomized controlled trials in which people
consumed diets that had been designed specifically to
achieve differences in GI (Kelly et al., 2004; Opperman
et al., 2004). On the other hand, the meta-analyses show
differences in total and LDL cholesterol not found in the
observational data. There is no obvious explanation for this
inconsistency.
Glycemic index and glycemic load and insulin
response
The GI of a food is affected not only by the rate of absorption
of carbohydrate, but also by the rate of glucose removal from
the plasma. When comparing two breakfast cereals with
different GI values (131733 and 54.577.2), the rate of
glucose removal was a major determinant of postprandial
hyperglycaemia (Schenk et al., 2003). It was found that the
lower GI breakfast cereal had induced hyperinsulinaemia
earlier than the higher GI cereal, resulting in an earlier
increase in more rapid removal of glucose from circulation. It
has been known for some time that insulin response cannot
be predicted based solely on the glycemic response to a food.
Collier and O’Dea (1983) found marked differences in the
glycemic response to potato with or without added butter,
but a very similar insulin response. The effect of GI on
insulin response may also depend upon insulin sensitivity.
Dietary GI has not been shown to have a marked effect on
insulin sensitivity whereas dietary fibre has (McAuley and
Mann, 2006).
Glycemic index and satiety
An important justification for the claim of an overall health
benefit of low GI foods is that low GI foods may aid weight
control because they promote satiety (Brand-Miller et al.,
2002). Ideally, weight loss studies comparing low and high
GI diets would need to assess differences between diets based
on ad libitum intake to show that the apparently greater
satiating effect of low GI foods led to a reduced energy
intake. Holt and colleagues have carried out the most
comprehensive study investigating the relationship between
GI and satiety, reporting the same work in several articles
(Holt et al., 1995; Holt et al., 1996; Holt et al., 1997). Isoenergetic
(1000 kJ) servings of 38 foods were tested for satiety
rating and glucose and insulin response. The food with the

highest satiety score was boiled potato. When comparing a
high GI food (potato) with a low GI food (white pasta) on an
iso-energetic, equi-carbohydrate (49 g) basis, the high GI
food had the highest satiety rating. The opposite was true
when comparing oranges (lower GI) and white bread (higher
GI), where the lower GI food had the higher satiety rating.
Porridge and natural muesli had similar glycemic and
insulinemic scores, but porridge had a greater satiety index
than muesli (Po0.001). These results suggest that there is
little or no relationship between GI and satiety, at least when
comparing food portions of equal energy content. Rather,
energy density appeared to be inversely related to satiety,
presumably because of the high bulk required to obtain a
serving containing 1000 kJ when low energy-dense foods
were tested. When iso-energetic, iso-volumetric carbohydrate-containing
beverages were tested, high GI beverages
resulted in lower energy intakes during a subsequent meal,
while low GI beverages were found not to suppress appetite
and food intake in the short-term (Anderson et al., 2002). A
review of the effect of glycemic carbohydrates on short-term
satiety has been published (Anderson and Woodend, 2003).
One conclusion was that high GI carbohydrates suppress
short-term (1 h) food intake more effectively than low GI
carbohydrates, whereas low GI carbohydrates appeared to be
more effective over longer periods (6 h).
How dietary GI and GL affects satiety and food intake over
a number of years is not entirely clear. The effectiveness of
dietary GI and GL on weight loss or maintenance is covered
by van Damm in this series. The results of several observational
studies have shown little difference in body mass
index (BMI) across categories of GI and GL (Salmeron et al.,
1997a, b; Hodge et al., 2004; Schulze et al., 2004). Murakami
et al. (2006) found a positive association between GI and
body mass index in Japanese female farmers, but no
association between GL and body mass index. On the other
hand, Ford and Liu reported inverse associations between GI
and GL and body mass index in a nationally representative
sample of US adults (Ford and Liu, 2001). These contradictory
findings might suggest that dietary GI and GL is not
a major determinant of dietary energy intake over the longterm.
A plausible reason is that GI appears not to be related
to energy density. Potatoes and lentils for example represent
foods with widely differing GIs but comparable energy
densities of around 3–4 kJ/g. On the other hand, cakes,
cookies and fresh oranges have similar GIs in the low to
medium range, but energy densities some 10-fold different
(Holt et al., 1996).
Recommendations
The FAO/WHO Report on Carbohydrates in Human Nutrition
suggests that the concept of GI provides a useful means
of selecting the most appropriate carbohydrate containing
foods for the maintenance of health and the treatment of
several disease states (FAO, 1998). Since the publication of
Glycemic index and glycemic load
BJ Venn and TJ Green
that report some of the limitations of the GI and GL concepts
have become increasingly apparent. With regard to measurement
there is clearly a need to study a larger number of
subjects under standard conditions to obtain more precise
estimates of the GI and GL of individual foods. The
introduction of instruments for assessing dietary intake in
epidemiological studies that have been designed to include
more direct measures of GI and GL will enhance the
confidence in findings from such studies. Despite these
reservations it does appear that distinguishing between foods
with appreciable differences in the indices may produce
some benefit in terms of glycemic control in diabetes and
lipid management. However, caution should be exercised in
food choice based solely on GI or GL because low GI and GL
foods may be energy dense and contain substantial amounts
of sugars or undesirable fatty acids that contribute to the
diminished glycemic response but not necessarily to good
health outcomes. This may apply especially to some of the
manufactured products that have been GI and GL tested and
are available in many countries. Given that most of the
studies which have demonstrated a health benefit of low GI
and GL involved the use of naturally occurring and
minimally processed foods it would seem to be appropriate
for such products to be further tested for their health benefits
directly, rather than on the basis of their functionality (that
is, a low glycemic response). Although some data suggest
that the low GI effect is not explained by the dietary fibre
content of the foods it remains conceivable that food
structure or composition explain some of the health
benefits. GI may be a useful indicator to guide food choice
if for example bread with a high GI is replaced on a slice-forslice
basis with a lower GI bread, thereby achieving a lower
GL. However, the complexity of the relationship between GI
and GL is probably not well understood whereby GI and the
amount of a food eaten are both important determinants of
the postprandial glycemic response. For the present it would
seem appropriate that when GI or GL are used to guide food
choice, it should only be done in the context of other
nutritional indicators and when values have been measured
in a large group of individuals.
Acknowledgements
We wish to thank Dr Jennie Brand-Miller, Professor Gary
Frost, Professor Philip James, Professor Simin Liu, Professor
Jim Mann, Dr Gabriele Riccardi, Dr M Robertson and
Professor HH Vorster for their valuable comments.
Conflict of interest
During the preparation and peer-review of this paper in
2006, the authors and peer-reviewers declared the following
interests.
Authors
Dr Bernard J Venn: None declared.
S129
European Journal of Clinical Nutrition

S130
Dr Tim Green: Affiliated with GI Otago, a commercial
glycemic index testing service.
Peer-reviewers
Dr Jennie Brand-Miller: Publishing books in the popular
press: ‘The New Glucose Revolution Series’; Director of a
University-based service for GI testing; Director of a not-forprofit
food-labelling programme based on the GI.
Professor Gary Frost: None declared.
Professor Philip James: None declared.
Professor Simin Liu: None declared.
Professor Jim Mann: None declared.
Dr Gabriele Riccardi: None declared.
Dr M Robertson: Research Grant from National Chemical
and Starch.
Professor HH Vorster: Member and Director of the Africa
Unit for Transdisciplinary health Research (AUTHeR), Research
grant from the South African Sugar Association.
References
Glycemic index and glycemic load
BJ Venn and TJ Green
Amano Y, Kawakubo K, Lee JS, Tang AC, Sugiyama M, Mori K (2004).
Correlation between dietary glycemic index and cardiovascular
disease risk factors among Japanese women. Eur J Clin Nutr 58,
1472–1478.
Anderson GH, Catherine NL, Woodend DM, Wolever TM (2002).
Inverse association between the effect of carbohydrates on blood
glucose and subsequent short-term food intake in young men. Am
J Clin Nutr 76, 1023–1030.
Anderson GH, Woodend D (2003). Effect of glycemic carbohydrates
on short-term satiety and food intake. Nutr Rev 61, S17–S26.
Brand-Miller J, Hayne S, Petocz P, Colagiuri S (2003b). Low-glycemic
index diets in the management of diabetes: a meta-analysis of
randomized controlled trials. Diabetes Care 26, 2261–2267.
Brand-Miller JC, Holt SH, Pawlak DB, McMillan J (2002). Glycemic
index and obesity. Am J Clin Nutr 76, 281S–285S.
Brand-Miller JC, Thomas M, Swan V, Ahmad ZI, Petocz P, Colagiuri S
(2003c). Physiological validation of the concept of glycemic load
in lean young adults. J Nutr 133, 2728–2732.
Brand-Miller JC, Wolever TM (2005). The use of glycaemic index
tables to predict glycaemic index of breakfast meals. Br J Nutr 94,
133–134.
Brand-Miller J, Wolever TM, Foster-Powell K, Colagiuri S (2003a). The
New Glucose Revolution. Marlowe & Company: New York.
Brouns F, Bjorck I, Frayn KN, Gibbs AL, Lang V, Slama G et al. (2005).
Glycaemic index methodology. Nutr Res Rev 18, 145–171.
Collier G, O’Dea K (1983). The effect of coingestion of fat on the
glucose, insulin, and gastric inhibitory polypeptide responses to
carbohydrate and protein. Am J Clin Nutr 37, 941–944.
Coulston AM, Hollenbeck CB, Reaven GM (1984). Utility of studies
measuring glucose and insulin responses to various carbohydratecontaining
foods. Am J Clin Nutr 39, 163–167.
Doi K, Matsuura M, Kawara A, Baba S (1979). Treatment of diabetes
with glucomannan (konjac mannan). Lancet 1, 987–988.
FAO (1998). Carbohydrates in human nutrition. Report of a Joint
FAO/WHO Expert Consultation. FAO Food Nutr Pap 66,
1–140.
Flint A, Moller BK, Raben A, Pedersen D, Tetens I, Holst JJ et al.
(2004). The use of glycaemic index tables to predict glycaemic
index of composite breakfast meals. Br J Nutr 91, 979–989.
Flood A, Subar AF, Hull SG, Zimmerman TP, Jenkins DJ, Schatzkin A
(2006). Methodology for adding glycemic load values to the
European Journal of Clinical Nutrition
National Cancer Institute Diet History Questionnaire database.
J Am Diet Assoc 106, 393–402.
Ford ES, Liu S (2001). Glycemic index and serum high-density
lipoprotein cholesterol concentration among US adults. Arch
Intern Med 161, 572–576.
Foster-Powell K, Holt SH, Brand-Miller JC (2002). International table
of glycemic index and glycemic load values: 2002. Am J Clin Nutr
76, 5–56.
Freeman J (2005). The glycemic index debate: does the type of
carbohydrate really matter? Diabet Forecast 58, 11.
Frost G, Leeds A, Trew G, Margara R, Dornhorst A (1998). Insulin
sensitivity in women at risk of coronary heart disease and the
effect of a low glycemic diet. Metabolism 47, 1245–1251.
Goldstein DE, Little RR, Lorenz RA, Malone JI, Nathan DM, Peterson
CM (2004). Tests of glycemia in diabetes. Diabetes care 27
(Suppl 1), S91–S93.
Granfeldt Y, Liljeberg H, Drews A, Newman R, Bjorck I (1994).
Glucose and insulin responses to barley products: influence of
food structure and amylose-amylopectin ratio. Am J Clin Nutr 59,
1075–1082.
Granfeldt Y, Hagander B, Bjorck I (1995). Metabolic responses to
starch in oat and wheat products. On the importance of food
structure, incomplete gelatinization or presence of viscous dietary
fibre. Eur J Clin Nutr 49, 189–199.
Hanefeld M, Cagatay M, Petrowitsch T, Neuser D, Petzinna D, Rupp
M (2004). Acarbose reduces the risk for myocardial infarction in
type 2 diabetic patients: meta-analysis of seven long-term studies.
Eur Heart J 25, 10–16.
Heilbronn LK, Noakes M, Clifton PM (2002). The effect of high- and
low-glycemic index energy restricted diets on plasma lipid and
glucose profiles in type 2 diabetic subjects with varying glycemic
control. J Am Coll Nutr 21, 120–127.
Henry CJ, Lightowler HJ, Kendall FL, Storey M (2006). The impact of
the addition of toppings/fillings on the glycaemic response
to commonly consumed carbohydrate foods. Eur J Clin Nutr 60,
763–769.
Henry CJ, Lightowler HJ, Strik CM, Storey M (2005). Glycaemic index
values for commercially available potatoes in Great Britain. Br J
Nutr 94, 917–921.
Hodge AM, English DR, O’Dea K, Giles GG (2004). Glycemic index
and dietary fiber and the risk of type 2 diabetes. Diabetes Care 27,
2701–2706.
Holt SH, Miller JC, Petocz P (1997). An insulin index of foods: the
insulin demand generated by 1000-kJ portions of common foods.
Am J Clin Nutr 66, 1264–1276.
Holt SH, Miller JC, Petocz P, Farmakalidis E (1995). A satiety index of
common foods. Eur J Clin Nutr 49, 675–690.
Holt SH, Brand Miller JC, Petocz P (1996). Inter-relationships among
postprandial satiety, glucose and insulin responses and changes in
subsequent food intake. Eur J Clin Nutr 50, 788–797.
Jenkins DJ, Goff DV, Leeds AR, Alberti KG, Wolever TM, Gassull MA
et al. (1976). Unabsorbable carbohydrates and diabetes: decreased
post-prandial hyperglycaemia. Lancet 2, 172–174.
Jenkins DJ, Kendall CW, Augustin LS, Franceschi S, Hamidi M,
Marchie A et al. (2002). Glycemic index: overview of implications
in health and disease. Am J Clin Nutr 76, 266S–273S.
Jenkins DJ, Wolever TM, Buckley G, Lam KY, Giudici S, Kalmusky J
et al. (1988). Low-glycemic index starchy foods in the diabetic diet.
Am J Clin Nutr 48, 248–254.
Jenkins DJ, Wolever TM, Jenkins AL, Giordano C, Giudici S,
Thompson LU et al. (1986). Low-glycemic response to traditionally
processed wheat and rye products: bulgur and pumpernickel
bread. Am J Clin Nutr 43, 516–520.
Jenkins DJ, Wolever TM, Jenkins AL, Josse RG, Wong GS (1984).
The glycaemic response to carbohydrate foods. Lancet 2,
388–391.
Jenkins DJ, Wolever TM, Jenkins AL, Thorne MJ, Lee R, Kalmusky J
et al. (1983). The glycaemic index of foods tested in diabetic

patients: a new basis for carbohydrate exchange favouring the use
of legumes. Diabetologia 24, 257–264.
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.
Kelly S, Frost G, Whittaker V, Summerbell C (2004). Low-glycaemic
index diets for coronary heart disease. Cochrane Database Syst Rev
CD004467.
Liljeberg H, Granfeldt Y, Bjorck I (1992). Metabolic responses to
starch in bread containing intact kernels versus milled flour. Eur
J Clin Nutr 46, 561–575.
Liu S, Manson JE, Stampfer MJ, Holmes MD, Hu FB, Hankinson SE
et al. (2001). Dietary glycemic load assessed by food-frequency
questionnaire in relation to plasma high-density lipoprotein
cholesterol and fasting plasma triacylglycerols in postmenopausal
women. Am J Clin Nutr 73, 560–566.
Ma Y, Li Y, Chiriboga DE, Olendzki BC, Hebert JR, Li W et al. (2006).
Association between carbohydrate intake and serum lipids. JAm
Coll Nutr 25, 155–163.
Mann J (2007). Dietary carbohydrate: relationship to cardiovascular
disease and disorders of carbohydrate metabolism. Eur J Clin Nutr
61(Suppl 1), S100–S111.
McAuley K, Mann J (2006). Thematic review series: patient-oriented
research. Nutritional determinants of insulin resistance. J Lipid Res
47, 1668–1676.
Mettler S, Lamprecht-Rusca F, Stoffel-Kurt N, Wenk C, Colombani PC
(2007). The influence of the subjects’ training state on the
glycemic index. Eur J Clin Nutr 61, 19–24.
Miller CK, Gabbay RA, Dillon J, Apgar J, Miller D (2006). The effect of
three snack bars on glycemic response in healthy adults. J Am Diet
Assoc 106, 745–748.
Murakami K, Sasaki S, Takahashi Y, Okubo H, Hosoi Y, Horiguchi H
et al. (2006). Dietary glycemic index and load in relation to
metabolic risk factors in Japanese female farmers with traditional
dietary habits. Am J Clin Nutr 83, 1161–1169.
Neuhouser ML, Tinker LF, Thomson C, Caan B, Horn LV, Snetselaar L
et al. (2006). Development of a glycemic index database for food
frequency questionnaires used in epidemiologic studies. J Nutr
136, 1604–1609.
Nuttall FQ, Mooradian AD, Gannon MC, Billington C, Krezowski P
(1984). Effect of protein ingestion on the glucose and insulin
response to a standardized oral glucose load. Diabetes Care 7, 465–470.
Opperman AM, Venter CS, Oosthuizen W, Thompson RL, Vorster HH
(2004). Meta-analysis of the health effects of using the glycaemic
index in meal-planning. Br J Nutr 92, 367–381.
Pi-Sunyer FX (2002). Glycemic index and disease. Am J Clin Nutr 76,
290S–298S.
Read NW, Welch IM, Austen CJ, Barnish C, Bartlett CE, Baxter AJ
et al. (1986). Swallowing food without chewing; a simple way to
reduce postprandial glycaemia. Br J Nutr 55, 43–47.
Salmeron J, Ascherio A, Rimm EB, Colditz GA, Spiegelman D, Jenkins
DJ et al. (1997a). Dietary fiber, glycemic load, and risk of NIDDM
in men. Diabetes Care 20, 545–550.
Glycemic index and glycemic load
BJ Venn and TJ Green
Salmeron J, Manson JE, Stampfer MJ, Colditz GA, Wing AL, Willett
WC (1997b). Dietary fiber, glycemic load, and risk of non-insulindependent
diabetes mellitus in women. JAMA 277, 472–477.
Schenk S, Davidson CJ, Zderic TW, Byerley LO, Coyle EF (2003).
Different glycemic indexes of breakfast cereals are not due to
glucose entry into blood but to glucose removal by tissue. Am
J Clin Nutr 78, 742–748.
Schulze MB, Liu S, Rimm EB, Manson JE, Willett WC, Hu FB (2004).
Glycemic index, glycemic load, and dietary fiber intake and
incidence of type 2 diabetes in younger and middle-aged women.
Am J Clin Nutr 80, 348–356.
Slyper A, Jurva J, Pleuss J, Hoffmann R, Gutterman D (2005).
Influence of glycemic load on HDL cholesterol in youth. Am J Clin
Nutr 81, 376–379.
Sugiyama M, Tang AC, Wakaki Y, Koyama W (2003). Glycemic
index of single and mixed meal foods among common Japanese
foods with white rice as a reference food. Eur J Clin Nutr 57,
743–752.
Suzuki H, Fukushima M, Okamoto S, Takahashi O, Shimbo T, Kurose
T et al. (2005). Effects of thorough mastication on postprandial
plasma glucose concentrations in nonobese Japanese subjects.
Metabolism 54, 1593–1599.
Tappy L, Gugolz E, Wursch P (1996). Effects of breakfast cereals
containing various amounts of beta-glucan fibers on plasma
glucose and insulin responses in NIDDM subjects. Diabetes Care
19, 831–834.
van de Laar FA, Lucassen PL, Akkermans RP, van de Lisdonk EH,
Rutten GE, van Weel C (2005). Alpha-glucosidase inhibitors for
patients with type 2 diabetes: results from a Cochrane systematic
review and meta-analysis. Diabetes Care 28, 154–163.
Venn BJ, Wallace AJ, Monro JA, Perry T, Brown R, Frampton C et al.
(2006). The glycemic load estimated from the glycemic index does
not differ greatly from that measured using a standard curve in
healthy volunteers. J Nutr 136, 1377–1381.
Wolever TM, Jenkins DJ (1986). The use of the glycemic index in
predicting the blood glucose response to mixed meals. Am J Clin
Nutr 43, 167–172.
Wolever TM, Vorster HH, Bjorck I, Brand-Miller J, Brighenti F, Mann
JI et al. (2003). Determination of the glycaemic index of foods:
interlaboratory study. Eur J Clin Nutr 57, 475–482.
Wolever TM (1990). The glycemic index. World Rev Nutr Diet 62,
120–185.
Wolever TM, Vuksan V, Eshuis H, Spadafora P, Peterson RD, Chao ES
et al. (1991). Effect of method of administration of psyllium on
glycemic response and carbohydrate digestibility. J Am Coll Nutr
10, 364–371.
Wolever TM, Yang M, Zeng XY, Atkinson F, Brand-Miller JC
(2006). Food glycemic index, as given in glycemic index
tables, is a significant determinant of glycemic responses
elicited by composite breakfast meals. Am J Clin Nutr 83,
1306–1312.
S131
European Journal of Clinical Nutrition

European Journal of Clinical Nutrition (2007) 61 (Suppl 1), S132–S137
& 2007 Nature Publishing Group All rights reserved 0954-3007/07 $30.00
www.nature.com/ejcn
FAO/WHO Scientific Update on carbohydrates in
human nutrition: conclusions
J Mann, JH Cummings, HN Englyst, T Key, S Liu, G Riccardi, C Summerbell, R Uauy, RM van Dam,
B Venn, HH Vorster and M Wiseman
European Journal of Clinical Nutrition (2007) 61 (Suppl 1), S132–S137; doi:10.1038/sj.ejcn.1602943
Keywords: carbohydrates; human nutrition; chronic diseases; FAO; WHO; scientific update
The Scientific Update involved consideration of a number of
key issues that have arisen since the Joint FAO/WHO Expert
Consultation on Carbohydrates in Human Nutrition was
held in 1997 (FAO, 1998) or where new data may have
altered conclusions drawn some 10 years ago. The Scientific
Update enabled some firm conclusions to be drawn and
identified a number of areas where more research is required
to enable definitive recommendations. The review papers
prepared as part of this Scientific Update applied the criteria
used by the 2002 WHO/FAO Expert Consultation on Diet,
Nutrition and the Prevention of Chronic Diseases to describe
strength of evidence for drawing the conclusions of the
scientific review (WHO, 2003). The following are agreedupon
summaries of both the review papers, and discussions
from the authors’ meeting (Geneva, 17–18 July 2006) on
each issue area.
The experts who participated at the authors meeting were:
John H Cummings, Hans Englyst Timothy Key, Simin Liu,
Jim Mann (Chairman), Rob van Dam, Bernard Venn, Carolyn
Summerbell (Rapporteur), Gabriele Riccardi, Ricardo Uauy,
HH Vorster (Rapporteur) and Martin Wiseman. The FAO
Secretariat members were Kraisid Tontisirin and Frank
Martinez Nocito and the WHO Secretariat members were
Chizuru Nishida and Denise Costa Coitinho.
Terminology and classification
Terminology and classification of carbohydrates remain
difficult issues. The Scientific Update endorsed the primary
classification, recommended by the 1997 Expert Consultation,
based on chemical form, while acknowledging that
classification of carbohydrates based on chemistry should
also have dimensions of physical effects (food matrix),
functional/physiological effects and health outcomes
(Cummings and Stephen, 2007). The chemical classification
Correspondence: Professor J Mann, Department of Human Nutrition, Edgar
National Centre for Diabetes Research, University of Otago, New Zealand.
E-mail: jim.mann@stonebow.otago.ac.nz
provides a practical basis for measurement and labelling, but
does not allow a simple translation into nutritional effects.
Each chemical class of carbohydrate has overlapping physiological
properties and health effects. As a result several
terms have been used to describe their functional properties.
For example, many terms exist to describe sugars in the diet.
While it is straightforward from an analytical point of view
to determine the total sugar content of food and their
division into monosaccharides and disaccharides, it is
acknowledged that the nature of the food matrix may
influence the nutritional effects of foods-containing sugars.
The term free sugars, defined as ‘all monosaccharides and
disaccharides added to foods by the manufacturer, cook or
consumer, plus sugars naturally present in honey, syrups and
fruit juices’ (WHO, 2003), describes sugars which may have
different physiological consequences from sugars incorporated
within intact plant cell walls. This concept, which has
been recommended by the 2002 WHO/FAO Expert Consultation
(WHO, 2003), is theoretically useful to nutritionists
but there is currently no standardized approach to their
determination.
The terms ‘prebiotic’, ‘resistant starch’ and ‘dietary fibre’
are used to indicate some of the physiological properties and
health effects of oligosaccharides and polysaccharides. The
concepts of ‘prebiotic’ and ‘resistant starch’ are quite clearly
defined. However, the current existence of several definitions
of dietary fibre, reflecting different physiological and
botanical properties or health effects of a range of naturally
occurring and synthetic carbohydrates and their associated
substances has resulted in confusion. The Scientific Update
considered that the term ‘dietary fibre’ should be reserved for
the cell wall polysaccharides of vegetables, fruits and wholegrains,
the health benefits of which have been clearly
established, rather than synthetic, isolated or purified
oligosaccharides and polysaccharides with diverse, and in
some cases unique, physiological effects. Thus, the Scientific
Update agreed to define ‘dietary fibre’ as ‘intrinsic plant cell
wall polysaccharides’.
It is acknowledged that certain fibre rich whole foods (for
example, pulses) can affect glycaemic control in diabetes,

and lipid levels, to a greater extent than cereal-based fibrecontaining
foods. All forms of dietary fibre influence bowel
habit, and other aspects of large bowel function. However,
the analytical distinction between soluble and insoluble
forms of dietary fibre may not always be useful since the
separation of soluble and insoluble fractions is pH dependent
and the link with specific physiological properties is
inconsistent. Further research is needed to determine the
exact properties of different non-starch polysaccharides
and their food sources, to explain their metabolic and
physiological effects. The whole-grain concept, along with
fruit and vegetables is central to the healthy diet message,
but the term whole-grain requires much clearer definition.
In particular, the extent to which health properties are
influenced by milling as compared with consumption of the
intact grain should be established.
Characterization and measurement
Carbohydrate determinations should describe chemical
composition accurately, and provide information of nutritional
relevance (Englyst et al., 2007). The traditional
calculation of carbohydrate ‘by difference’ does not conform
to either of these criteria as it combines the analytical
uncertainties of the other macronutrient measures and any
unidentified material present, and a single value for
carbohydrate cannot reflect the range of carbohydrate
components or their diverse nutritional properties.
The first principle of chemical analysis is to ensure that the
fraction of interest is completely extracted from the food
matrix in its native form (for example, sugars), or dispersed
to such an extent that it can be hydrolysed (for example,
total starch and non-starch polysaccharide). Once appropriately
isolated, oligosaccharides and polysaccharides may
be subjected to enzymatic (adds specificity) or acidic (when
appropriate enzyme is unavailable) hydrolysis to release their
constituent sugars. Detection is then by gas chromatography
or high-performance liquid chromatography, which can be
used to measure specific monosaccharides, disaccharides and
small oligosaccharides. More rapid colorimetric assays can be
used to measure sugars with reducing groups, and individual
sugar species can be determined with enzyme-linked assays.
There is more diversity of opinion regarding measurement
of dietary components defined on the basis of functionality
rather than chemical composition. However, the scientific
update did recognize the nutritional importance of considering
both chemical and food matrix aspects of carbohydrate-containing
foods. Fundamental to any approach,
where categorization is based on physiological or nutritional
properties, is that they should be measured as chemically
identified components. Such nutritional categories can
reflect physiological fate (e.g. resistant starch), specific
functionalities (e.g. prebiotics) or botanical origin (e.g. free
sugars, dietary fibre).
Scientific Update on carbohydrate in human diet
J Mann et al
In view of the considerable interest in the health effects of
dietary fibre, much discussion centred around the various
approaches to its measurement. Indigestibility in the small
intestine was not considered to be a satisfactory basis for the
definition of dietary fibre. It was agreed that the term
‘dietary fibre’ be limited to polysaccharides that are intrinsic
to the plant cell wall, and the methods for measuring dietary
fibre are, therefore, those which can reliably quantify the
component polysaccharides. Direct chemical measurement
was favoured over empirical gravimetric methods for this
purpose.
Physiology
Carbohydrates are the principal energy source in the diets of
most people and have a special role to play in energy
metabolism and homoeostasis. Despite this the energy
values of some carbohydrates continue to be debated. This
is because of the use of different energy systems such as
combustible, digestible and metabolizable. Furthermore,
ingested macronutrients may not be fully available to tissues
and the tissues themselves may not be fully able to oxidize
substrates made available to them. Therefore, for certain
carbohydrates the discrepancies between combustible energy,
digestible energy, metabolizable energy (ME) and net
metabolizable energy (NME) may be considerable.
In defining the role that carbohydrate plays in metabolism
it is important to realize that the site, rate and extent of
carbohydrate digestion in and absorption from the gut is key
to understanding the many roles that this group of
chemically related compounds and their metabolic products
play in the body. However, even the concept of digestibility
has different meanings. Within the nutrition community it
is an accepted convention that digestion occurs in the small
(upper) bowel, while fermentation occurs in the large (lower)
bowel. Although both processes result in the breakdown of
food and absorption of energy yielding substrates, the term
digestibility is reserved for events occurring in the upper gut.
However, in the discussion of the energy value of foods
digestibility is defined as the proportion of combustible
energy that is absorbed over the entire length of the
gastrointestinal tract. For the sake of determining energy
values of carbohydrate that are of practical use, and of
ascribing health benefits to individual classes of them, some
coherence needs to be brought to the use of the terms
digestible, digestion and digestibility and to integrating the
whole of the gut process into the equation of energy balance.
Three food energy systems are in use in food tables and for
food labelling in different countries and regions in the world
based on selective interpretation of the digestive physiology
and metabolism of food carbohydrates. These are the
systems which employ the Atwater specific factors; the
Atwater general factors (17, 17, 37 kJ/g for carbohydrate,
protein and fat, respectively); NME factors (typically applied
to polyols) (Elia and Cummings, 2007). This is clearly
S133
European Journal of Clinical Nutrition

S134
unsatisfactory and confusing to the consumer. Food labelling
policy should aim to establish a system that avoids bias as far
as possible using a practical user-friendly system. While it
has been suggested that an enormous amount of work would
have to be undertaken to change the current ME system into
an NME system the additional changes may not be as great as
anticipated. The key issues are the extent to which the
underlying physiological considerations are sound, the
extent to which they should drive food-labelling policy and
the extent to which food energy system(s) should become
consistent within and between regional jurisdictions.
The role of carbohydrate as a regulator of appetite and
energy expenditure is a subject for intensive study, propelled
by the rise in obesity in many countries. Carbohydrate is
generally more satiating than fat, but less satiating than
protein. However, studies of eating behaviour indicate that
appetite regulation does not unconditionally depend on the
oxidation of one nutrient and argues against the operation
of a simple carbohydrate oxidation or storage model of
feeding behaviour to the exclusion of other macronutrients.
The control of appetite operates through a system with many
redundancies that does not rely overwhelmingly on one or
two major factors, such as energy density or on a specific
macronutrient. Rather it depends on multiple factors that
can interact, compensate or override each other, depending
on environmental exposures and their duration. These
include sensory factors, diet composition and variety of
available food items, eating environment and individual
subject characteristics, such as age, habitual dietary intake
and prior social conditioning. It has been difficult to
establish the relative importance of these factors, which are
likely to differ with the environmental setting. In the light of
this evidence, the role of different types of carbohydrates on
eating behaviour might not be expected to have large effects
on long-term energy homoeostasis and intervention studies
are generally consistent with this view.
Carbohydrates that reach the large bowel enter a very
different type of metabolism, determined by the anaerobic
microbiota of this organ and in so doing exert an important
influence on its function. However, uncertainty remains
regarding the exact amounts and types of carbohydrate that
reach the caecum and are available for fermentation, largely
because of the difficulties of studying this area of the gut and
of variations in food processing, stage of maturity at which
plant foods are eaten, post harvest changes, day to day
fluctuations in food intake and individual differences in gut
function. Current estimates of the amount of non-starch
polysaccharide þ resistant starch þ non-a-glucan oligosaccharides
þ polyols þ lactose that reach the large bowel are between
20 and 40 g/day in countries with ‘westernised’ diets, while
they may reach 50 g/day where traditional staples are largely
cereal or diets are high in fruit and/or vegetables.
Since the 1997 FAO/WHO Expert Consultation (FAO, 1998),
there has been progress in understanding the role of these
‘unavailable’, but fermented, carbohydrates in the large bowel.
Non-starch polysaccharide clearly affect bowel habit and so
European Journal of Clinical Nutrition
Scientific Update on carbohydrate in human diet
J Mann et al
does resistant starch, but to a lesser extent. However, the non-aglucan
oligosaccharides or non-digestible oligosaccharides have
little laxative role, if any, although they do affect the
composition of the flora. This latter property has led to the
invention of the term ‘prebiotic’, which designates the selective
effect of these carbohydrates on the flora, in particular their
capacity to increase numbers of Bifidobacteria and Lactobacilli
without growth of other genera. This now well-established
physiological property has not so far led through to clear health
benefits. There is some evidence that use of prebiotic supplements
increases absorption of calcium, but this has not been
consistently observed. Prebiotics are incorporated into some
brands of infant formulae, and a range of food for adults.
Overweight and obesity
The prevalence of overweight and obesity has increased
rapidly worldwide during recent decades reaching epidemic
proportions in children and adults in both industrialized and
developing countries, in particular those which are going
through rapid economic transition. Carbohydrates are
among the macronutrients that provide energy and can
thus contribute to weight gain when consumed in excess of
energy requirements. If energy intake is strictly controlled,
macronutrient composition of the diet (energy percentages
of fat and carbohydrates) does not substantially affect body
weight or fat mass. However, an important issue is whether,
among free-living individuals, the macronutrient composition
of the diet affects the likelihood of passive overconsumption.
There is no clear evidence that altering the
proportion of total carbohydrate in the diet is an important
determinant of energy intake. However, there is evidence
that sugar-sweetened beverages do not induce satiety to the
same extent as solid forms of carbohydrate and that increases
in consumption of sugar-sweetened soft drinks are associated
with weight gain (van Dam and Seidell, 2007). Thus, there is
justification for the recommendation to restrict the consumption
of beverages high in free sugars to reduce the risk
of excessive weight gain. Solid foods high in free sugars tend
to be energy dense and there is some evidence from
intervention studies that reduction of solid foods high in
free sugars can contribute to weight loss. Thus, the outcomes
of the Scientific Update support the population nutrient
intake goals on free sugars (that is, o10% of total energy)
that were recommended by the 2002 WHO/FAO Expert
Consultation (WHO, 2003). A high content of dietary fibre
in whole-grain, vegetables, legumes and fruits is associated
with relatively low energy density, promotion of satiety and
in observational studies a lesser degree of weight gain than
among those with lower intakes. Although it is difficult
to establish with certainty that dietary fibre rather than
other dietary attributes are responsible, it is considered
appropriate to recommend that whole-grain cereals, vegetables,
legumes and fruits are the most appropriate sources of
dietary carbohydrate. The available evidence is considered

insufficient for the use of glycaemic index (GI) of carbohydrate-containing
foods to predict the likelihood of their
ability to reduce the risk of obesity in normal weight
individuals or promote weight loss in those who are overweight
or obese.
Glycaemic index and glycaemic load
The 1997 FAO/WHO Expert Consultation suggested that the
concept of GI might provide a useful means of helping to
select the most appropriate carbohydrate-containing foods
for the maintenance of health and the treatment of several
diseases (FAO, 1998). It is acknowledged that choice of
carbohydrate-containing foods should not be based solely on
GI since low-GI foods may be energy dense and contain
substantial amounts of sugars, fat or undesirable fatty acids
that contribute to the diminished glycaemic response but
not necessarily to good health outcomes. The inter-individual
variation in glycaemic responses to foods is a further
limitation of the GI concept and underlines the need to
study a larger number of subjects under standard conditions
than have been investigated previously to obtain more
precise estimates of the GI of individual foods. In addition
to the inter-individual variation, responses in a single
individual are not necessarily consistent. Despite these
reservations it does appear that distinguishing between
foods with appreciable differences in the indices may
produce some benefit in terms of glycaemic control in
diabetes and lipid management. The benefits demonstrated
in randomized-controlled trials are smaller than those
conferred by other dietary and lifestyle changes especially
those which facilitate weight loss in the overweight and
obese (Venn and Green, 2007). Although some data
suggest that the dietary fibre content of the foods does not
explain the low-GI effect it remains conceivable that food
structure or composition explain some of the health
benefits.
The limitation of the GI concept may apply especially to
some of the manufactured products that have been GI tested
and are available in many countries. Given that most of the
studies which have demonstrated a health benefit of low GI
involved the use of naturally occurring and minimally
processed foods, it would seem to be appropriate for
manufactured foods to be further examined rather than to
assume health benefits only on the basis of their functionality
(that is, a low-glycaemic response).
GI is perhaps most appropriately used to guide food
choices when considering similar carbohydrate-containing
foods, for example bread with a low GI may be preferable to a
higher GI bread, with a resultant lower glycaemic load (GL).
Ideally, GI should be measured in a relatively larger group of
individuals, but even then should be interpreted in the
knowledge of the inter- and intra-individual variation.
Furthermore, GI and GL should always be considered in
the context of other nutritional indicators.
Scientific Update on carbohydrate in human diet
J Mann et al
Carbohydrates in the aetiology of diabetes and
cardiovascular disease
The Scientific Update also considered the relationship
between dietary carbohydrate and cardiovascular disease,
disorders of carbohydrate metabolism and cancer. A wide
range of intakes of carbohydrate-containing foods is acceptable
in the context of dietary patterns, which are protective
against cardiovascular disease, diabetes and prediabetic
states. The nature of carbohydrate eaten is important.
Whole-grains, legumes, vegetables and intact fruits are the
most appropriate sources of carbohydrate. There is impressive
evidence that they are associated with a reduced risk of
cardiovascular disease, although inconsistencies with regard
to the definition of whole-grain explain why the cardiovascular
effect of fruits and vegetables was described by the 2002
WHO/FAO Expert Consultation (WHO, 2003) as ‘convincing’
whereas the protective effect of whole-grains was
graded as ‘probable’. These carbohydrate-containing foods
are rich sources of dietary fibre (defined as non-starch
polysaccharide in the 2002 WHO/FAO Expert Consultation),
which protects against type II diabetes, and other cardioprotective
components. However, there is no good evidence of
protection against cardiovascular disease and diabetes when
various oligosaccharides or polysaccharides or other isolated
components of whole-grains, fruits, vegetables and legumes
are added to functional and manufactured foods. This
provides further justification for defining dietary fibre as
‘intrinsic plant cell wall polysaccharide’ as developed
through this Scientific Update. Both the 1997 FAO/WHO
Expert Consultation and the 2002 WHO/FAO Expert Consultation
recommended that total carbohydrate should
provide 55–75% total energy and that intake of fruits and
vegetables (excluding tubers, for example, potatoes and
cassava) should be 400 g/day or more. Precise amounts of
dietary fibre as non-starch polysaccharide were not recommended
by the 2002 WHO/FAO Expert Consultation (WHO,
2003). It was considered that the recommended intakes of
fruit, vegetables, legumes and regular consumption of
whole-grain cereals would provide adequate intakes of total
dietary fibre. These recommendations of the 2002 WHO/
FAO Expert Consultation are compatible with the outcomes
of the Scientific Update, although some caveats are
suggested. Given that a wide range of carbohydrate intakes
is compatible with cardioprotection and that many western
countries have average intakes that are below 55% of total
energy intakes, the conclusion of the Scientific Update was
that a lower limit of around 50% total energy was acceptable.
Several national guidelines suggest a lower limit of 50%. It is
more important to be prescriptive with regard to the nature
of dietary carbohydrate, especially when total carbohydrate
intakes are at the upper end of the recommended range.
Failure to emphasize the need for carbohydrates to be
derived principally from whole-grain cereals, fruits, vegetables
and legumes may result in increased lipoproteinmediated
risk of coronary heart disease, especially with an
S135
European Journal of Clinical Nutrition

S136
increase in the ratio of total low-density lipoprotein
cholesterol to high-density lipoprotein cholesterol and an
increase in triglycerides. This may apply particularly to
overweight and obese individuals who are insulin-resistant.
A low-dietary GL may reduce the risk of type II diabetes and
cardiovascular disease but similar caveats to those described
for GI apply when interpreting such observations.
Carbohydrates in the treatment of diabetes and
cardiovascular risk factors
Similar issues to those described above relating to dietary
carbohydrate and disease prevention apply to the management
of people with diabetes and to those with risk factors
for cardiovascular disease, especially to those with abnormal
lipid profiles. While legumes, pulses and certain fruits,
vegetables and cereal grains appear to confer particular
benefit in terms of reducing glycaemia and lowering lowdensity
lipoprotein cholesterol, difficulties of measurement
and the possibility that other constituents of plant foods
may be involved preclude the definitive conclusion that the
content of ‘soluble’ dietary fibre is the determinant of this
effect (Mann, 2007). Low-GI foods may confer benefits in
terms of improving glycaemic control in people with
diabetes. However, it is not clear whether these benefits are
fully independent of the effects of dietary fibre or the fact
that foods with intact plant cell walls tend to have low GI.
Furthermore, it is uncertain whether functional and manufactured
foods with a low GI confer the same long-term
benefits as low GI predominantly plant-based foods.
Cancer
The review relating to cancer concluded that the most
frequently observed association with carbohydrates was that
between low intakes of dietary fibre and increased risk for
colorectal cancer. However, it was noted that the data were
not entirely consistent. The possible links between high
intakes of sucrose and colorectal cancer and high intakes of
lactose and ovarian cancer were considered to be rather more
tenuous. The link between obesity and several cancers was
considered to be convincing (Key and Spencer, 2007), thus,
nutritional determinants of obesity may be regarded as
causally related to all obesity-related cancers. Key and Spencer
(2007) noted that the World Cancer Research Fund Report
(WCRF) on Diet and Cancer to be published in November
2007 was likely to generate more definitive data on nutritional
determinants of cancer than had previously been available.
Discussion
The Scientific Update has enabled us to draw a number of
important conclusions. The importance of improving the
definition of dietary fibre was discussed and it was agreed
European Journal of Clinical Nutrition
Scientific Update on carbohydrate in human diet
J Mann et al
that the definition should be based on well-established
health benefits and the ability to fulfill regulatory requirements.
We, therefore, proposed that dietary fibre should be
defined as intrinsic plant cell wall polysaccharides. Thus the
method(s), which enable the measuring of the component
polysaccharides would be appropriate for determining dietary
fibre as defined above. The three food energy systems
currently used in food tables and for labelling continue to
create confusion for the consumer. There is an urgent need
for an internationally agreed readily understood system, and
further consideration of the net metabolizable energy
approach is suggested. Review of the recent literature
endorses the recommendations of the 2002 WHO/FAO
Expert Consultation concerning the restriction of beverages
high in free sugars and the limitation of total intake of free
sugars to reduce the risk of overweight and obesity. The
positive messages of that consultation are also endorsed,
notably the potential of whole-grains, legumes, vegetables
and intact fruits to protect against diabetes and cardiovascular
disease. Many of these foods, high in dietary fibre, help
to improve glycaemic control in people with diabetes and to
reduce cardiovascular risk factors. However, while foods with
a low GI may also confer benefit in some of these contexts,
the Scientific Update suggested caution regarding the use of
the GI as the sole determinant of the quality of carbohydrate-containing
foods. Given the association between
obesity and cancer at several sites and the possibility that
dietary fibre may reduce the risk of colorectal cancer, the
dietary messages aimed at reducing the risk of obesity,
diabetes and cardiovascular disease have the potential to also
reduce cancer risk.
Finally, the need to review the current recommended
range for dietary carbohydrate intake (55–75% total energy)
was identified. There appeared to be insufficient justification
for the recommended lower limit, therefore a possible
revision to 50% was suggested. A wide range of intakes, as
a proportion of total energy intake, is compatible with low
risk of chronic diseases although excess intake of any
macronutrient is likely to lead to obesity. The nature of
dietary carbohydrate appears to be a more important
determinant of health outcomes than the proportion of
total energy derived from carbohydrate intake. It is hoped
that these papers will stimulate discussion among the
scientific community and relevant health professions and
inform the formal Expert Consultation on Carbohydrate to
be convened in the foreseeable future.
Conflict of interest
During the preparation and peer-review of this paper in 2006,
the authors and peer-reviewers declared the following interests.
Experts
Professor John 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 Hans Englyst: Director and share-holder of Englyst
Carbohydrates Ltd—a small research-oriented company working
on dietary carbohydrates and health within the Medical
Research Council. The UK Food Standards Agency is the main
research partner and sponsor. In addition, Englyst Carbohydrates
provide analytical assistance and reagents
to universities and food industry worldwide, albeit on a small
scale. The complete independence of Englyst Carbohydrates is
maintained by not entering into any consultancy agreement.
Professor Timothy J Key: None declared.
Professor Simin Liu: None declared.
Professor Jim Mann: None declared.
Dr Gabriele Riccardi: None declared.
Professor Carolyn Summerbell: None declared.
Professor Ricardo Uauy: Scientific Adviser on a temporary
basis for Unilever and Wyeth; Scientific Editorial/Award
Adviser for Danone, DSM, Kelloggs, and Knowles and Bolton
on a ad hoc basis.
Dr Rob M van Dam: None declared.
Dr Bernard Venn: None declared.
Dr HH Vorster: Member and Director of the Africa Unit for
Transdisciplinary health Research (AUTHeR), Research grant
from the South African Sugar Association.
Dr Martin Wiseman: None declared.
FAO/WHO Secretariat members
Dr Denise Costa Coitinho: None declared.
Mr Frank Martinez Nocito: None declared.
Scientific Update on carbohydrate in human diet
J Mann et al
Dr Chizuru Nishida: None declared.
Dr Kraisid Tontisirin: None declared.
References
Cummings JH, Stephen AM (2007). Carbohydrate terminology and
classification. Eur J Clin Nutr 61 (Suppl 1), S5–S18.
Elia M, Cummings JH (2007). Physiological aspects of energy
metabolism and gastrointestinal effects of carbohydrates. Eur
J Clin Nutr 61 (Suppl 1), S40–S74.
Englyst K, Liu S, Englyst H (2007). Nutritional characterisation
of dietary carbohydrates providing defined measurements
for labeling and research. Eur J Clin Nutr 61 (Suppl 1),
S19–S39.
FAO (1998). Carbohydrates in Human Nutrition. Report of a Joint FAO/
WHO Expert Consultation (FAO Food and Nutrition Paper 66)
Food and Agriculture Organization: Rome.
Key TJ, Spencer EA (2007). Carbohydrates and cancer: an overview
of the epidemiological evidence. Eur J Clin Nutr 61 (Suppl 1),
S112–S121.
Mann J (2007). Dietary carbohydrate: relationship to cardiovascular
disease and disorders of carbohydrate metabolism. Eur J Clin Nutr
61 (Suppl 1), S100–S111.
van Dam RM, Seidell JC (2007). Carbohydrate intake and obesity. Eur
J Clin Nutr 61 (Suppl 1), S75–S99.
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.
WHO (2003). Diet, Nutrition and the Prevention of Chronic
Diseases. Report of a Joint WHO/FAO Expert Consultation
(WHO Technical Report Series 916) World Health Organization:
Geneva.
S137
European Journal of Clinical Nutrition

Annex 1.3: Council Directive 90/496/EEC on Nutrition
Labelling of Foodstuffs.

Annex 1.4: US‐FDA GRAS Notification No. GRN 00184
(Isomaltulose).

FDA/CFSAN/OFAS: Agency Response Letter: GRAS Notice No. GRN 000184
FDA Home Page | CFSAN Home | Search/Subject Index | Q & A | Help
CFSAN/Office of Food Additive Safety
March 20, 2006
Agency Response Letter
GRAS Notice No. GRN 000184
William A. Olson, Ph.D.
Center for Regulatory Services, Inc.
5200 Wolf Run Shoals Road
Woodbridge, VA 22192-5755
Dear Dr. Olson:
Re: GRAS Notice No. GRN 000184
The Food and Drug Administration (FDA) is responding to the notice, dated October 31,
2005, that you submitted on behalf of S‹DZUCKER AG Mannheim/Ochsenfurt
(S‹DZUCKER) in accordance with the agency's proposed regulation, proposed 21 CFR
170.36 (62 FR 18938; April 17, 1997; Substances Generally Recognized as Safe (GRAS);
the GRAS proposal). FDA received the notice on November 1, 2005, filed it on November
4, 2005, and designated it as GRAS Notice No. GRN 000184.
The subject of the notice is isomaltulose. The notice informs FDA of the view of
S‹DZUCKER that isomaltulose is GRAS, through scientific procedures, for use as a
nutritive sweetener in a variety of foods as described in Table 1.
Table 1
Conditions of use proposed by S‹DZUCKER
Food Category Use levels (percent)]
Baked goods and baking mixes (21 CFR 170.3(n)(1)) 10-25
Beverages (21 CFR 170.3(n)(2), 21 CFR 170.3(n)(3)) 1-10
Cereal-based products (21 CFR 170.3(n)(4))
cereals 20-35;
bars 5-20
Confectionery and frostings (21 CFR 170.3(n)(9)) 15-99
Chewing gum (21 CFR 170.3(n)(6)) 5-35
Frozen dairy desserts and mixes (21 CFR 170.3(n)(20)) 30
Fruit and water ices (21 CFR 170.3(n)(21)) 15
Gelatins, desserts, puddings, etc. (21 CFR 170.3(n)(22)) 15-30
http://www.cfsan.fda.gov/~rdb/opa-g184.html
Seite 1 von 3
19.02.2008

FDA/CFSAN/OFAS: Agency Response Letter: GRAS Notice No. GRN 000184
Jams, jellies and spreads (21 CFR 170.3(n)(28)) 25-40
Nuts and peanut spreads (21 CFR 170.3(n)(32)) 45
Milk products (21 CFR 170.3(n)(31)) 3-20
Processed fruit and fruit juices or vegetable juices
(21 CFR 170.3(n)(35)), (21 CFR 170.3(n)(36))
The subject of GRN 000184 is 6-O-·-D-glucopyranosyl-D-fructofuranose, monohydrate
(CAS Reg No. 13718-94-0; molecular formula C 12 H 22 O 11 , referred to as isomaltulose.
Isomaltulose is a reducing disaccharide consisting of one glucose and one fructose moiety
linked by an ·-1,6- glycosidic bond, and is a water soluble, white or colorless crystalline
powder.
S‹DZUCKER describes the method of manufacture and provides product specifications for
isomaltulose. Isomaltulose is manufactured from food-grade sucrose. An aqueous sucrose
solution is applied to a column with an immobilized enzyme preparation consisting of nonviable
cells of Protaminobacter rubrum (strain designated by the Dutch culture collection
(Centraalbureau voor Schimmelcultures (CBS)) as CBS 574.77). (1) The enzyme sucrose-6glucosylmutase
(EC 5.4.99.11) converts the ·-1,2 glycosidic bond of sucrose into the ·-1,6
bond of isomaltulose. The resultant isomaltulose is crystallized, dried, then purified by
filtration and ion exchange. Specifications include an assay content of at least 98 % 6-O-·-
D-glucopyranosyl- D-fructofuranose and limits on lead (less than 0.1 milligrams per
kilogram). S‹DZUCKER intends isomaltulose to be used as a nutritive sweetener that
would totally or partially replace sucrose or other highly digestible carbohydrates.
Isomaltulose provides a moderate sweetness, bulk, and texture to foods. S‹DZUCKER
considers that the greater cost and lower solubility in water of isomaltulose compared with
sucrose would limit those foods in which it would replace sucrose.
The notice includes a published review summarizing published in vivo and in vitro studies
demonstrating that isomaltulose is completely hydrolyzed and absorbed in the small
intestine as glucose and fructose. The notifer concludes that the safety of isomaltulose is,
therefore, equivalent to that of sucrose, which, like isomaltulose, is a disaccharide composed
of glucose and fructose. The review discusses biological data, toxicological studies,
metabolic studies, and studies on gastrointestinal tolerance, and concludes that the use of
isomaltulose would not be a health concern.
S‹DZUCKER used the reported per capita refined sugar consumption in the United States
and assumed a five-to-ten percent market share replacement of isomaltulose for sucrose to
estimate daily intake of isomaltulose at approximately 3 to 6 grams (g)/person per day.
S‹DZUCKER notes that isomaltulose has been found at low concentrations in honey and
cane sugar juice. Dietary intake of isomaltulose from consumption of honey would likely be
1-10
Snack foods (21 CFR 170.3(n)(37)) 10-25
Sugar substitutes (21 CFR 170.3(n)(42)) 2 - >99
Sweet sauces, toppings, syrups (21 CFR 170.3(n)(43)) 15-30
Other categories:
nutritive formula
energy-reduced foods
meal replacements/slimming foods
http://www.cfsan.fda.gov/~rdb/opa-g184.html
†
5-20
5-40
5-20
Seite 2 von 3
19.02.2008

FDA/CFSAN/OFAS: Agency Response Letter: GRAS Notice No. GRN 000184
less than one gram. S‹DZUCKER notes that isomaltulose has been in use as a food
ingredient in Japan since 1985 and has recently been authorized as a novel food or food
ingredient in Europe. (2)
Based on the information provided by S‹DZUCKER, as well as other information available
to FDA, the agency has no questions at this time regarding S‹DZUCKER's conclusion that
isomaltulose is GRAS under the intended conditions of use. The agency has not, however,
made its own determination regarding the GRAS status of the subject use of isomaltulose.
As always, it is the continuing responsibility of S‹DZUCKER to ensure that food
ingredients that the firm markets are safe, and are otherwise in compliance with all
applicable legal and regulatory requirements.
In accordance with proposed 21 CFR 170.36(f), a copy of the text of this letter, as well as a
copy of the information in your notice that conforms to the information in proposed 21 CFR
170.36(c)(1), is available for public review and copying on the homepage of the Office of
Food Additive Safety (on the Internet at http://www.cfsan.fda.gov/~lrd/foodadd.html).
Sincerely,
Laura M. Tarantino, Ph.D.
Director
Office of Food Additive Safety
Center for Food Safety and Applied Nutrition
(1) Non-viable cells of P. rubrum are immobilized by entrapment in beads of calcium
alginate gel, formed from calcium chloride (21 CFR 184.1193) and sodium alginate (21
CFR 184.1724). S‹DZUCKER notes that the immobilization system is consistent with 21
CFR 173.357.
(2) (European Commission decision of 25 July 2005 authorizing the placing on the market of
isomaltulose as a novel food or novel food ingredient under Regulation (EC) No 258/97 of
the European Parliament and of the Council).
Food Ingredients and Packaging † | † Summary of all GRAS Notices
CFSAN Home | CFSAN Search/Subject†Index | CFSAN Disclaimers†&†Privacy†Policy | CFSAN Accessibility/Help
FDA Home Page | Search FDA Site | FDA A-Z Index | Contact FDA
http://www.cfsan.fda.gov/~rdb/opa-g184.html
FDA/Center for Food Safety & Applied Nutrition
Hypertext updated by rxm April 24, 2006
Seite 3 von 3
19.02.2008

Annex 1.5: US Health Claims Regulation on Dietary
Noncariogenic Carbohydrate Sweeteners and Dental
Caries, Fed. Reg. Vol. 73, No. 102, May 27, 2008, 30299‐
30301.

jlentini on PROD1PC65 with RULES
Federal Register / Vol. 73, No. 102 / Tuesday, May 27, 2008 / Rules and Regulations
manure and litter from the infected
premises must be moved to the
composting site at the same time;
(5) Following the composting process,
the composted manure or litter remains
undisturbed for an additional 15 days
before movement;
(6) After this 15-day period, all of the
composted manure or litter from the
infected site is removed at the same
time;
(7) The resulting compost must be
transported either in a previously
unused container or in a container that
has been cleaned and disinfected, since
last being used, in accordance with part
71 of this chapter;
(8) The vehicle in which the resulting
compost is to be transported has been
cleaned and disinfected, since last being
used, in accordance with part 71 of this
chapter; and
(9) Copies of the permit
accompanying the compost derived
from the manure and the litter are
submitted so that a copy is received by
the State animal health official and the
veterinarian in charge for the State of
destination within 72 hours of arrival of
the compost at the destination listed on
the permit.
■ 7. Section 82.8 is amended as follows:
■ a. In paragraph (a)(2), by removing the
citation ‘‘7 CFR part 59’’ and adding the
citation ‘‘9 CFR part 590’’ in its place.
■ b. By revising paragraph (a)(3) to read
as set forth below.
§ 82.8 Interstate movement of eggs, other
than hatching eggs, from a quarantined
area.
(a) * * *
(3) The establishment that processes
the eggs, other than hatching eggs, for
sale establishes procedures adequate to
ensure that the eggs are free of END,
including:
(i) The establishment separates
processing and layer facilities, the
incoming and outgoing eggs at the
establishment, and any flocks that may
reside at the establishment;
(ii) The establishment implements
controls to ensure that trucks, shipping
companies, or other visitors do not
expose the processing plant to END;
(iii) Equipment used in the
establishment is cleaned and disinfected
in accordance with part 71 of this
chapter at intervals determined by the
Administrator to ensure that the
equipment cannot transmit END to the
eggs, other than hatching eggs, being
processed; and
(iv) The eggs are packed either in
previously unused flats or cases, or in
used plastic flats that were cleaned or
disinfected since last being used, in
accordance with part 71 of this chapter;
* * * * *
■ 8. Section 82.9 is amended as follows:
■ a. In paragraph (b), by removing the
word ‘‘and’’ at the end of the paragraph.
■ b. By redesignating paragraph (c) as
paragraph (d).
■ c. By adding a new paragraph (c) to
read as set forth below.
§ 82.9 Interstate movement of hatching
eggs from a quarantined area.
* * * * *
(c) The hatching eggs have been kept
in accordance with the sanitation
practices specified in § 147.22 and
§ 147.25 of the National Poultry
Improvement Plan; and
* * * * *
■ 9. Section 82.14 is amended as
follows:
■ a. In paragraph (c)(2), in the
introductory text, by revising the second
sentence to read as set forth below.
■ b. In paragraph (e)(2), by removing the
first sentence and by adding two new
sentences in its place to read as set forth
below.
■ c. By adding a new paragraph (i) to
read as set forth below.
§ 82.14 Removal of quarantine.
* * * * *
(c) * * *
(2) * * * The birds and poultry must
be composted according to the following
instructions or according to another
procedure approved by the
Administrator as being adequate to
prevent the dissemination of END:
* * * * *
(e) * * *
(2) Composting. If the manure and
litter is composted, the manure and
litter must be composted in the
quarantined area. The manure and litter
must be composted according to the
following method, or according to
another procedure approved by the
Administrator as being adequate to
prevent the dissemination of END: Place
the manure and litter in rows 3 to 5 feet
high and 5 to 10 feet at the base. * * *
* * * * *
(i) After the other conditions of this
section are fulfilled, an area will not be
released from quarantine until followup
surveillance over a period of time
determined by the Administrator
indicates END is not present in the
quarantined area.
* * * * *
VerDate Aug2005 16:07 May 23, 2008 Jkt 214001 PO 00000 Frm 00029 Fmt 4700 Sfmt 4700 E:\FR\FM\27MYR1.SGM 27MYR1
30299
Done in Washington, DC, this 20th day of
May 2008.
Kevin Shea,
Acting Administrator, Animal and Plant
Health Inspection Service.
[FR Doc. E8–11741 Filed 5–23–08; 8:45 am]
BILLING CODE 3410–34–P
DEPARTMENT OF HEALTH AND
HUMAN SERVICES
Food and Drug Administration
21 CFR Part 101
[Docket No. FDA–2006–P–0404] (Formerly
Docket No. 2006P–0487)
Food Labeling: Health Claims; Dietary
Noncariogenic Carbohydrate
Sweeteners and Dental Caries
AGENCY: Food and Drug Administration,
HHS.
ACTION: Final rule.
SUMMARY: The Food and Drug
Administration (FDA) is adopting as a
final rule, without change, the
provisions of the interim final rule that
amended the regulation authorizing a
health claim on noncariogenic
carbohydrate sweeteners and dental
caries, i.e., tooth decay, to include
isomaltulose as a substance eligible for
the health claim. FDA is taking this
action to complete the rulemaking
initiated with the interim final rule.
DATES: This rule is effective May 27,
2008.
FOR FURTHER INFORMATION CONTACT:
Jillonne Kevala, Center for Food Safety
and Applied Nutrition (HFS–830), Food
and Drug Administration, 5100 Paint
Branch Pkwy., College Park, MD 20740–
3835, 301–436–1450.
SUPPLEMENTARY INFORMATION:
I. Background
In the Federal Register of September
17, 2007 (72 FR 52783), FDA published
an interim final rule to amend the
regulation in part 101 (21 CFR part 101)
that authorizes a health claim on the
relationship between noncariogenic
carbohydrate sweeteners and dental
caries (§ 101.80) to include the
noncariogenic sugar isomaltulose.
Under section 403(r)(3)(B)(i) and section
403(r)(7) of the Federal Food, Drug, and
Cosmetic Act (the act) (21 U.S.C.
343(r)(3)(B)(i) and 343(r)(7)), FDA
issued this interim final rule in response
to a petition filed under section
403(r)(4) of the act. Section
403(r)(3)(B)(i) of the act states that the
Secretary of Health and Human Services
(and, by delegation, FDA) shall issue a

jlentini on PROD1PC65 with RULES
30300 Federal Register / Vol. 73, No. 102 / Tuesday, May 27, 2008 / Rules and Regulations
regulation authorizing a health claim if
he or she ‘‘determines, based on the
totality of publicly available scientific
evidence (including evidence from welldesigned
studies conducted in a manner
which is consistent with generally
recognized scientific procedures and
principles), that there is significant
scientific agreement, among experts
qualified by scientific training and
experience to evaluate such claims, that
the claim is supported by such
evidence’’ (see also § 101.14(c)). Section
403(r)(4) of the act sets out the
procedures that FDA is to follow upon
receiving a health claim petition.
Section 403(r)(7) of the act permits FDA
to make a proposed regulation issued
under section 403(r) effective upon
publication pending consideration of
public comment and publication of a
final regulation if the agency determines
that such action is necessary for public
health reasons.
On August 31, 2006, Cargill, Inc.
(petitioner), submitted a health claim
petition to FDA requesting that the
agency amend the ‘‘dietary
noncariogenic carbohydrate sweeteners
and dental caries’’ claim at § 101.80 to
authorize a noncariogenic dental health
claim for isomaltulose. FDA filed the
petition for comprehensive review in
accordance with section 403(r)(4) of the
act on December 8, 2006. The petitioner
requested that FDA grant an interim
final rule by which foods containing
isomaltulose could bear the health claim
prior to publication of the final rule.
FDA and the petitioner mutually agreed
to extend the deadline for the agency’s
decision on the petition to September 5,
2007.
As part of its review of the scientific
literature on isomaltulose and dental
caries, FDA considered the scientific
evidence presented in the petition as
well as information previously
considered by the agency on the
etiology of dental caries and the effects
of slowly fermentable carbohydrates.
The agency summarized this evidence
in the interim final rule (72 FR 52783
at 52784 to 52786). Based on the
available evidence, FDA concluded that
isomaltulose, like other noncariogenic
carbohydrate sweeteners listed in
§ 101.80(c)(2)(ii), does not promote
dental caries. Consequently, FDA
amended § 101.80(c)(2)(ii) to broaden
the health claim to include isomaltulose
as an additional substance eligible for
the health claim.
II. Summary of Comments and the
Agency’s Response
FDA solicited comments on the
interim final rule. The comment period
closed on December 3, 2007. The agency
received four letters of response, three
from consumers and one from a
manufacturer. The manufacturer
supported the interim rule. Two of the
consumers’ comments addressed issues
that are outside the scope of this
rulemaking and will not be addressed
here. The remaining comment suggested
that there had been insufficient testing
to demonstrate the safety of
isomaltulose, but did not provide any
information or analysis to support
revision of the agency’s conclusion.
Given the absence of contrary
evidence on the agency’s decisions
announced in the interim final rule,
FDA is adopting as a final rule, without
change, the interim final rule that
amended § 101.80 to include
isomaltulose as a substance eligible for
the dental caries health claim.
III. Analysis of Impacts
FDA has examined the impacts of the
final rule under Executive Order 12866,
the Regulatory Flexibility Act (5 U.S.C.
601–612), and the Unfunded Mandates
Reform Act of 1995 (Public Law 104–4).
Executive Order 12866 directs agencies
to assess all costs and benefits of
available regulatory alternatives and,
when regulation is necessary, to select
regulatory approaches that maximize
net benefits (including potential
economic, environmental, public health
and safety, and other advantages;
distributive impacts; and equity). The
agency believes that this final rule is not
a significant regulatory action under the
Executive order.
The Regulatory Flexibility Act
requires agencies to analyze regulatory
options that would minimize any
significant impact of a rule on small
entities. Because this final rule allows
new voluntary behavior and imposes no
additional restrictions on current
practices, the agency certifies that the
final rule will not have a significant
economic impact on a substantial
number of small entities.
Section 202(a) of the Unfunded
Mandates Reform Act of 1995 requires
that agencies prepare a written
statement which includes an assessment
of anticipated costs and benefits before
proposing ‘‘any rule that includes any
Federal mandate that may result in the
expenditure by State, local, and tribal
governments, in the aggregate, or by the
private sector, of $100,000,000 or more
(adjusted annually for inflation) in any
one year.’’ The current threshold after
adjustment for inflation is $127,000,000,
using the most current (2006) Implicit
Price Deflator for the Gross Domestic
Product. FDA does not expect this final
rule to result in any one-year
VerDate Aug2005 16:07 May 23, 2008 Jkt 214001 PO 00000 Frm 00030 Fmt 4700 Sfmt 4700 E:\FR\FM\27MYR1.SGM 27MYR1
expenditure that would meet or exceed
this amount.
FDA received no comments relevant
to economic impact. The costs and
benefits of available regulatory
alternatives analyzed in the interim
final rule (72 FR 52783 at 52787 to
52788) are adopted without change in
this final rule. By now affirming that
interim final rule, FDA has not imposed
any new requirements. Therefore, there
are no additional costs and benefits
associated with this final rule.
IV. Environmental Impact
The agency has determined under 21
CFR 25.32(p) that this action is of a type
that does not individually or
cumulatively have a significant effect on
the human environment. Therefore,
neither an environmental assessment
nor an environmental impact statement
is required.
V. Paperwork Reduction Act
FDA concludes that the labeling
provisions of this final rule are not
subject to review by the Office of
Management and Budget because they
do not constitute a ‘‘collection of
information’’ under the Paperwork
Reduction Act of 1995 (44 U.S.C. 3501–
3520). Rather, the food labeling health
claim on the association between
consumption of isomaltulose and the
nonpromotion of dental caries is a
‘‘public disclosure of information
originally supplied by the Federal
Government to the recipient for the
purpose of disclosure to the public’’ (5
CFR 1320.3(c)(2)).
VI. Federalism
FDA has analyzed this final rule in
accordance with the principles set forth
in Executive Order 13132. FDA has
determined that the rule will have a
preemptive effect on State law. Section
4(a) of the Executive order requires
agencies to ‘‘construe * * * a Federal
statute to preempt State law only where
the statute contains an express
preemption provision or there is some
other clear evidence that the Congress
intended preemption of State law, or
where the exercise of State authority
conflicts with the exercise of Federal
authority under the Federal statute.’’
Section 403A of the act (21 U.S.C. 343–
1) is an express preemption provision.
Section 403A(a)(5) of the act provides
that:
* * * no State or political subdivision of
a State may directly or indirectly establish
under any authority or continue in effect as
to any food in interstate commerce—* * *(5)
any requirement respecting any claim of the
type described in section 403(r)(1) made in
the label or labeling of food that is not

jlentini on PROD1PC65 with RULES
Federal Register / Vol. 73, No. 102 / Tuesday, May 27, 2008 / Rules and Regulations
identical to the requirement of section 403(r)
* * *
On September 17, 2007, FDA
published an interim final rule which
imposed requirements under section
403(r) of the act. This final rule affirms
the September 17, 2007, amendment to
the existing food labeling regulations to
add isomaltulose to the authorized
health claim for noncariogenic
carbohydrate sweeteners and dental
caries. Although this rule has a
preemptive effect in that it precludes
States from issuing any health claim
labeling requirements for isomaltulose
and the nonpromotion of dental caries
that are not identical to those required
by this final rule, this preemptive effect
is consistent with what Congress set
forth in section 403A of the act. Section
403A(a)(5) of the act displaces both
State legislative requirements and State
common law duties. Riegel v.
Medtronic, 128 S. Ct. 999 (2008).
FDA believes that the preemptive
effect of this final rule is consistent with
Executive Order 13132. Section 4(e) of
the Executive order provides that ‘‘when
an agency proposes to act through
adjudication or rulemaking to preempt
State law, the agency shall provide all
affected State and local officials notice
and an opportunity for appropriate
participation in the proceedings.’’ On
August 1, 2007, FDA’s Division of
Federal and State Relations provided
notice via fax and e-mail transmission to
State health commissioners, State
agriculture commissioners, food
program directors, and drug program
directors, as well as FDA field
personnel, of FDA’s intent to amend the
health claim regulation authorizing
health claims for noncariogenic
carbohydrate sweeteners and dental
caries (§ 101.80). FDA received no
comments from any States in response
to this notice.
In addition, the agency sought input
from all stakeholders through
publication of the interim final rule in
the Federal Register on September 17,
2007 (72 FR 52783). FDA received no
comments from any States on the
interim final rule.
In conclusion, the agency believes
that it has complied with all of the
applicable requirements of Executive
Order 13132 and has determined that
the preemptive effects of this rule are
consistent with the Executive order.
List of Subjects in 21 CFR Part 101
Food labeling, Nutrition, Reporting
and Recordkeeping requirements.
■ Therefore, under the Federal Food,
Drug, and Cosmetic Act and under
authority delegated to the Commissioner
of Food and Drugs, 21 CFR part 101 is
amended as follows:
PART 101—FOOD LABELING
■ Accordingly, the interim final rule
amending § 101.80 that was published
in the Federal Register of September 17,
2007 (72 FR 52783), is adopted as a final
rule without change.
Dated: May 19, 2008.
Jeffrey Shuren,
Associate Commissioner for Policy and
Planning.
[FR Doc. E8–11802 Filed 5–23–08; 8:45 am]
BILLING CODE 4160–01–S
DEPARTMENT OF THE TREASURY
Internal Revenue Service
26 CFR Part 1
[TD 9400]
RIN 1545–BG97
Treatment of Property Used To Acquire
Parent Stock in Certain Triangular
Reorganizations Involving Foreign
Corporations
AGENCY: Internal Revenue Service (IRS),
Treasury.
ACTION: Final and temporary
regulations.
SUMMARY: This document contains final
and temporary regulations under section
367(b) of the Internal Revenue Code
(Code). The final regulations revise an
existing final regulation and add a crossreference.
The temporary regulations
implement the rules described in Notice
2006–85 and Notice 2007–48. The
regulations affect corporations engaged
in certain triangular reorganizations
involving one or more foreign
corporations. The text of the temporary
regulations serves as the text of the
proposed regulations (REG–136020–07)
set forth in the notice of proposed
rulemaking on this subject published in
the Proposed Rules section in this issue
of the Federal Register.
DATES: Effective Date: These regulations
are effective May 27, 2008.
Applicability Dates: For dates of
applicability, see § 1.367(a)–
3T(b)(2)(i)(C) and 1.367(b)–14T(e).
FOR FURTHER INFORMATION CONTACT:
Daniel McCall, (202) 622–3860 (not a
toll-free number).
SUPPLEMENTARY INFORMATION:
Background
On September 22, 2006, the IRS and
Treasury Department issued Notice
VerDate Aug2005 16:07 May 23, 2008 Jkt 214001 PO 00000 Frm 00031 Fmt 4700 Sfmt 4700 E:\FR\FM\27MYR1.SGM 27MYR1
30301
2006–85 (2006–41 IRB 677), which
announced that regulations would be
issued under section 367(b) to address
certain triangular reorganizations under
section 368(a) involving one or more
foreign corporations. On May 31, 2007,
the IRS and Treasury Department issued
Notice 2007–48 (2007–25 IRB 1428),
which amplified Notice 2006–85 and
announced that additional regulations
would be issued under section 367(b).
Each notice describes transactions the
IRS and Treasury Department believe
raise significant policy concerns.
Notice 2006–85 describes triangular
reorganizations in which a subsidiary
(S) purchases stock of its parent
corporation (P) from P in exchange for
property, and then exchanges the P
stock for the stock or assets of a target
corporation (T), but only if P or S (or
both) is foreign. Notice 2006–85
announced that regulations to be issued
under section 367(b) would make
adjustments that would have the effect
of a distribution of property from S to
P under section 301 (deemed
distribution). Notice 2006–85 further
announced that regulations would
address similar transactions where S
acquires the P stock from a related party
that purchased the P stock in a related
transaction.
Notice 2007–48 describes transactions
in which S purchases all or a portion of
the P stock exchanged in the
reorganization from a person other than
P (such as from public shareholders on
the open market). Notice 2007–48
announced that regulations to be issued
under section 367(b) would also make
adjustments that would have the effect
of a distribution of property from S to
P (under section 301) followed by a
deemed contribution of such property
by P to S. Notice 2007–48 further
announced that the regulations would
take into account the earnings and
profits of other corporations, as
appropriate, if a principal purpose of
creating, organizing, or funding S is to
avoid the adjustments to be made by the
regulations.
These temporary regulations set forth
the regulations described in Notices
2006–85 and 2007–48. The existing final
regulations under § 1.367(b)–13 are
revised to conform the definitions of the
terms P, S, and T in those regulations to
the definitions of such terms in these
temporary regulations. The existing
final regulations under § 1.367(b)–2 are
revised to clarify that the definition of
earnings and profits in § 1.367(b)–2(l)(8)
applies only for purposes of §§ 1.367(b)–
7 and 1.367(b)–9.

Annex 1.6: “PalatinoseTM (isomaltulose) – a new
innovative carbohydrate”

PALATINOSEô - a new innovative carbohydrate
What is PALATINOSEô?
Palatinoseô is a new innovative carbohydrate. While providing the glucose-based energy to the body in
a prolonged, slow and low glycemic way, it is fully available and is toothfriendly. It further supports the
body in burning fat. This ingredient profile enables the development of products for a healthy lifestyle.
More detailed information is given below.
PALATINOSEô (generic name: isomaltulose) was approved for use as a novel food/novel food
ingredient in the EU in 2005 1 . In the U.S. Palatinoseô is GRAS (GRAS notification # 184 accepted by
FDA in March 2006). The food status was confirmed by many other countries worldwide.
PALATINOSEô ñ nutritional/physiological properties
Isomaltuloseís nutritional and physiological properties differ distinctly from those of sugar, as a result of this
-1,6-linkage:
The enzymes in the small intestine split isomaltulose down about 4 to 5 times more slowly than sucrose,
as demonstrated by enzyme kinetic studies. This means that a smaller amount of glucose and fructose
from isomaltulose is absorbed into the blood per time unit, compared to the more rapid rate at which the
monosaccharides from sucrose or cooked starches are absorbed.
Furthermore, this absorption does not only take place in the upper parts of the small intestine (as is the
case for quickly absorbed sugars), but along the entire small intestine. This means that isomaltulose still
supplies glucose as energy for the body at a time when sucrose has already been digested and would
no longer supply energy. The overall digestion in the small intestine is virtually complete, as
demonstrated by recovery studies.
Complete absorption in the small intestine means that isomaltulose provides the same amount of
calories as all digestible carbohydrates (sugars and starches 4 kcal/g) and is equally well tolerated (as
no significant amount reaches the large intestine no gastrointestinal distress occurs).
The slow but complete hydrolysis and absorption leads to a low blood glucose response (Glycemic Index
= 32) and a low insulinemic response (Insulinemic
Index = 30). This provides a tool for food
manufacturers to prepare low glycemic food
products (e.g. beverages etc) that help the
consumers in their food choice to achieve a lower
insulin day-profile which is desirable as explained in
the following: The hormone insulin has many
functions in the body, e.g. it down-regulates high
blood glucose levels by ìopening the doorî for
cellular glucose uptake. At the same time insulin
promotes fat storage and inhibits fat burning. High
levels of insulin over a long period of time are
discussed to contribute to obesity and the
development of diabetes.
Consistent with the lower insulin response after
intake of isomaltulose, it was demonstrated that fat
oxidation was increased (measurements of the
respiratory quotient). This may have an effect on
body fat.
0 30 60 90 120
-1
BENEO GmbH68165 MannheimGermanywww.BENEO-Group.comGeneral Management: Hildegard BauerDr. Matthias MoserYves Servotte
Court of Registration: County Court MannheimNo. HRB 701800VAT Identification Number DE 253 691 060Bank Details: Deutsche Bank AG Mannheim
Acc. 40651200 Bank Code 670 700 10SWIFT: DEUT DE SMIBAN DE94 6707 0010 0040 6512 00
Plasma glucose change (mmol/L)
5
4
3
2
1
0
-2
Time (min)
Sucrose
Palatinoseô
Blood glucose curves of Palatinoseô or sucrose (50g)
determined in 10 healthy volunteers at the Sydney Universityís
Glycaemic Index Research Service (2002), (mean±SD).
1 Commission Decision 2005/581/EC of 25 July 2005 authorising the placing on the market of isomaltulose as a novel food or novel
food ingredient under Regulation (EC) No 258/97 of the European Parliament and of the Council (notified under document number
C(2005) 2776). OJ L 199, 29.7.2005, p. 90ñ91.

Isomaltulose is tooth friendly, unlike starches or other sugars. Isomaltulose is not used by the oral micro
flora as a nutrient source. Therefore no significant amount of harmful acids that lead to cariogenic
lesions are produced from isomaltulose, as demonstrated in vivo by pH-telemetry tests.
PALATINOSEô as an ingredient in food production
Isomaltulose is derived from sugar (sucrose) by enzymatic conversion of the 1,2-linkage of glucose and
fructose into the more stable -1,6-bond. A commercial process was developed by S‹DZUCKER.
Isomaltulose is a carbohydrate providing 4 kcal/g. It has a natural sugar-like taste with a mild sweetness that
is about half that of sucrose. 2 It is used to replace sucrose in recipes but it is used for other nutritional
reasons as well (e.g. in sports drinks for longer energy supply). Because of its low hygroscopicity and good
flowability it is used in powder applications (e.g. instant powder drink applications), its high stability under
acidic conditions is important in liquid applications. The main area of application is in drinks. Other categories
of use are, e.g. cereals and cereal products, milk-based products, confectionery/baked goods etc.
Products with Palatinoseô contribute to a healthy lifestyle.
Contact:
Anke Sentko
Vice President Regulatory Affairs & Nutrition Communication
BENEO Group
ORAFTI-PALATINIT-REMY
Phone +49 621 421 144
Fax +49 621 421 165
Mobil +49160905820008
anke.sentko@beneo-group.com
2 Being half as sweet as sugar does not result in higher uses compared to sucrose but in less sweet products. Double the amount does
not lead to the same sweetness of sugar! To use the simple model of a candy of 3 gram: Doubling the amount of isomaltulose means
doubling the size and weight of the candy and it still would not be as sweet as a sucrose candy.
BENEO GmbH68165 MannheimGermanywww.BENEO-Group.comGeneral Management: Hildegard BauerDr. Matthias MoserYves Servotte
Court of Registration: County Court MannheimNo. HRB 701800VAT Identification Number DE 253 691 060Bank Details: Deutsche Bank AG Mannheim
Acc. 40651200 Bank Code 670 700 10SWIFT: DEUT DE SMIBAN DE94 6707 0010 0040 6512 00

Annex 1.7: “Dossier for the Scientific Substantiation of
Claims related to PalatinoseTM and its Nutritional
Physiological Properties”

CONFIDENTIAL
Regulatory Affairs &
Nutrition Communication – AJ/Se
28 May 2008
Page 1 of 16
Dossier for the Scientific Substantiation of Claims related to Palatinose
(isomaltulose) and its Nutritional / Physiological Properties
Table of Content
Summary
1. Isomaltulose – What it is!
2. Its Physiology – Digestion, Absorption & Metabolism of Isomaltulose
3. Studies on the Effect of Isomaltulose on Blood Glucose and Insulin Levels and
its Prolonged Energy Supply
4. Discussion: General Characteristics of the Gycemic and Insulinemic Properties
of Isomaltulose as compared to Sucrose
5. Conclusions
Summary
Isomaltulose is a disaccharide and – like sucrose – consists of glucose and fructose. But
unlike sucrose, isomaltulose has low glycemic and low insulinemic properties. It is
characterised by an α-1,6 glucosidic linkage between the glucose and fructose moieties
which is more stable than the α-1,2 glucosidic linkage in sucrose.
Isomaltulose is almost completely hydrolysed in the small intestine into its components
glucose and fructose which are absorbed and metabolised in the same pathway as if
derived from sucrose. Therefore, isomaltulose provides the same amount of calories as
other fully available carbohydrates do (4 kcal/g in food labelling). The difference to sucrose
is that isomaltulose is hydrolysed at a much lower rate which is reflected in its glycemic and
insulinemic properties. The rise in blood glucose concentrations after isomaltulose intake is
slower and remains at a lower level while it lasts for a longer period of time compared to
sucrose. The insulin response corresponds, which results in a lower insulin demand. As no
significant amounts of isomaltulose reach the large intestine, isomaltulose is tolerated as
well as sucrose.
Reflecting these characteristics in digestion and absorption, the uniqueness of isomaltulose
when compared to sucrose and other available carbohydrates (e.g. maltodextrin) becomes
obvious: Isomaltulose provides glucose, the fuel or energy for the body, in a more balanced
and prolonged way.
WHO recommends following a high-carbohydrate-based but low glycemic diet for a healthy
lifestyle. Isomaltulose can contribute significantly to reach such a healthy lifestyle.

CONFIDENTIAL
Regulatory Affairs &
Nutrition Communication – AJ/Se
28 May 2008
Page 2 of 16
1. Isomaltulose - What it is!
Isomaltulose is a carbohydrate which occurs in minor amounts naturally in honey and sugar
cane juice (Siddiqui and Furgala 1967; Eggleston and Grisham 2003). In 1957,
SÜDZUCKER discovered and described isomaltulose for the first time (Weidenhagen and
Lorenz 1957). This was the basis for the commercial, large scale production of
isomaltulose, which takes place in Offstein, Germany. An enzyme (non-GMO) transfers the
α-1,2 linkage in sucrose to an α-1,6 linkage. This glucose-fructose combination is
isomaltulose (6-O-α-D-glucopyranosyl-D-fructofuranose) (Figure 1).
HO
HO
Sucrose
OH
O
OH
1 O
OH
1
O
2 HO
OH
[glucose] [fructose]
OH
6
Enzyme
HO
HO
PalatinoseTM PalatinoseTM OH
O
OH
[glucose]
1
6
O
(isomaltulose)
OH
OH
OH
O
OH
2
1
[fructose]
Figure 1: Enzymatic rearrangement of sucrose to isomaltulose (Palatinose).
Isomaltulose is a food / food ingredient (like e.g. sugar, starch or maltodextrin). It has been
used as such in Japan and other Asian countries since 1985. Its food status was confirmed
in the European Union as well after undergoing the novel food approval process.
SÜDZUCKER / PALATINIT’s Isomaltulose was approved as a novel food / novel food
ingredient for use in food in the EU by Commission Decision 2005/251/EC of 25 th July
2005. And it received GRAS status by FDA acknowledgement in March 2006 (FDA 2006).
Compared to sucrose, isomaltulose has different key physiological properties, though:
• It is hydrolysed by the same enzyme complex as is necessary for sucrose hydrolysis
in the small intestine, but hydrolysis takes place at a much lower rate.
• It is fully digested (as sucrose) yet slowly (and thus is not a low digestible but a SLOW
digestible carbohydrate).
• Its characteristic stable bond (α-1,6) is well known from starch, rather than from
sugars.
• Isomaltulose provides the same amount of energy yet in a more constant flow and
within a lower insulin profile. Isomaltulose has a low effect on blood glucose and
insulin levels.
• Extreme blood glucose fluctuations and in particular a drastic fall below fasting
glucose concentrations into the so-called relative hypoglycaemia, physiological

CONFIDENTIAL
General or usual name: Isomaltulose
Trade Name: Palatinose
Chemical name: 6-O-α-D-glucopyranosyl-D-fructose
Chemical classification: Carbohydrate (Disaccharide)
CAS Reg. No. 13718-94-0
Total molecular formula: C12H22O11 x H2O
Molecular weight: 360.32 (monohydrate)
Figure 2: Chemical description of isomaltulose.
Regulatory Affairs &
Nutrition Communication – AJ/Se
28 May 2008
Page 3 of 16
reactions typical for highly digestible sugars and starches (sucrose, glucose or
dextrose, maltodextrin etc.), can be avoided with isomaltulose.
• Isomaltulose provides energy in form of blood glucose continuously over a longer
period of time,
Isomaltulose’s names and chemical description are summarised in Figure 2.
In the following, the physiology of isomaltulose is described in more detail.
2. Its Physiology - Digestion, Absorption & Metabolism of Isomaltulose
Due to the more stable bonding between glucose and fructose, isomaltulose is more
resistant to acid hydrolysis and enzymatic splitting by oral bacteria or by digestive enzymes
in the small intestine than sucrose is. This means, isomaltulose is hydrolysed slowly to
glucose and fructose and then absorbed, virtually completely, in the small intestine.
Isomaltulose does not lead to gastrointestinal discomfort as no significant amount reaches
the large intestine. Isomaltulose provides the same amount of calories as other full
digestible carbohydrates do.
These basic characteristics of isomaltulose have been demonstrated in a number of
studies:
Isomaltulose is slowly hydrolysed into glucose and fructose
Sucrose and isomaltulose are hydrolysed by the same sucrase / isomaltase enzyme
complex located in the brush border of the small intestine. However, this occurs at different
sites within this enzyme complex. While sucrose is hydrolysed at the sucrase site (for α-1,2
bonds), isomaltulose is hydrolysed by the isomaltase site (for α-1,6 bonds) of the complex
(Figure 3). (Heinz 1987; Heymann and Heinz 1987; Günther and Heymann 1998; Dahlqvist
et al 1963). Heinz (1987) demonstrated with human enzymes that the rate of hydrolysis
slowed down by a factor of 5 with isomaltulose as a substrate compared to sucrose.

CONFIDENTIAL
Regulatory Affairs &
Nutrition Communication – AJ/Se
28 May 2008
Page 4 of 16
Various in vitro studies with material from different species (rat, pig, human) addressed the
rate of hydrolysis of isomaltulose versus sucrose or maltose and showed a rate of
hydrolysis of typically less than 25 % that of sucrose (Table 1).
Table 1: Rate of hydrolysis of isomaltulose in vitro
Species Rate of hydrolysis of isomaltulose (%)
Reference
Versus sucrose Versus maltose
Pig 10 – 20 2 – 5 Dahlqvist, 1961
Pig 5.6 4.7 Heinz, 1987
Rat 11.8 4.4 Yamada et al., 1985
Rat 11.4 2.4 Tsuji et al., 1986
Rat 1.8 3.7 Heinz, 1987
Human 44.7 11 Grupp & Siebert, 1978
Human 12.7 9 Ziesenitz, 1986a
Human 26* 8* Ziesenitz, 1986b
Human 16 18 Heinz, 1987
*5 sugars added in a mixed preparation; in all other studies sugars added individually.
Isomaltulose is virtually completely absorbed in the small intestine
The absorption of isomaltulose was observed using various study designs and species.
Recovery of radio labelled isomaltulose was compared to labelled sucrose in a rat study
that showed a similar metabolism of both sugars (Macdonald and Daniel 1983).
Absorption in the small intestine was examined in a study using pigs (60-70 kg bw) with a
re-entrant fistula at the end of the small intestine. With this fistula it was possible to collect
the chyme that arrived at the end of the small intestine. This chyme was analysed for its
content of sugars. The pigs received 20% sucrose or isomaltulose as part of their diet. It
was demonstrated that there was hardly any passage of isomaltulose at the terminal end of
the ileum. Recovery of ingested isomaltulose was less than 3%. Both sugars were almost
completely digested and absorbed in the small intestine. The amounts of end products
(short chain fatty acids) of bacterial fermentation were similar for both sucrose and
isomaltulose, indicating that fermentation in the large intestine was not a significant factor.
Moreover, the flow of wet ileal chyme was identical in the sucrose and isomaltulose group,
demonstrating the absence of non-absorbed and thus osmotically active nutrients (van
Weerden et al 1983).
An ileostomy study performed at the University of Würzburg, Germany, examined digestion
and absorption of isomaltulose in humans in vivo (Gostner et al 2006). The 10 ileostomy
patients consumed 50g isomaltulose in form of a drink (500ml) or a combination of a drink

CONFIDENTIAL
Regulatory Affairs &
Nutrition Communication – AJ/Se
28 May 2008
Page 5 of 16
(250ml) and biscuits (140g) for breakfast after overnight fasting, followed by hourly
collection of the ileostoma bags over an 8-hour period. On average, 96% (drinks) or 99%
(drinks + biscuits) of the 50g isomaltulose intake was digested (hydrolysed), and 94%
(drinks) or 96% (drinks + biscuits) was absorbed in the small intestine. There was no
significant effect of the food type. This study confirms that isomaltulose is a fully digestible
carbohydrate.
Its gastrointestinal tolerance is similar to sucrose
While low-digestible carbohydrates are known to have the potential to cause
gastrointestinal discomfort as they are osmotically active and fermented in the large
intestine to short chain fatty acids and gases, isomaltulose shows a slow but virtually
complete absorption and is therefore tolerated well, similar to sucrose.
In conclusion, the totality of in vitro and in vivo data shows that isomaltulose is a
well tolerated carbohydrate that is absorbed slowly but virtually completely in form
of glucose and fructose.

CONFIDENTIAL
Regulatory Affairs &
Nutrition Communication – AJ/Se
28 May 2008
Page 6 of 16
3. Studies on the Effect of Isomaltulose on Blood Glucose and Insulin Levels and its
Prolonged Energy Supply
The slower hydrolysis of isomaltulose and subsequent slower absorption of glucose and
fructose compared to sucrose are reflected in its characteristic blood glucose response with
a slow, low and sustained rise in blood glucose levels and corresponding low insulin levels.
Blood glucose and insulin response studies
The effect of isomaltulose on blood glucose and insulin levels was studied in healthy and
diabetic humans (Macdonald & Daniel 1983; Kawai et al 1985 and 1989, Sydney
University’s Glycaemic Index Research Service (SUGiRS), 2002). In all the studies similar
study designs were used for blood glucose and insulin measurements e.g. intake of the test
and control substances dissolved in water and after overnight fasting.
The most recent study at Sydney University’s Glycaemic Research Service (SUGiRS,
2002) – one of the leading research institutes in this scientific area - was performed to
determine the Glycemic Index (GI) and the Insulinemic Index (II) of isomaltulose according
to internationally recognised standard methodology 1 . In this study, 10 healthy volunteers
(Caucasian, 18-24 years old, BMI of 19-24 kg/m 2 ) received 50 g isomaltulose or sucrose,
maltodextrine, or glucose (control) dissolved in 250 ml water after overnight fasting. Blood
glucose and insulin concentrations were determined over a period of 120 minutes. The
blood glucose and insulin curves for isomaltulose and sucrose are shown in Figures 3 and
4.
Blood glucose concentration after sucrose intake increased by a maximum of almost 4
mmol/l (76 mg/dl) within 30 minutes, while they returned to baseline after about 60 minutes
followed by a phase below baseline (relative hypoglycaemia) for the most of the second
hour of the measurement. In contrast, blood glucose concentrations after isomaltulose
increased by less than a third of that of sucrose (1,2 mmol/l or 22 mg/dl) and remained
above baseline for most of the duration of the two hour experiment. The insulin curves
corresponded. The overall blood glucose and insulin responses were lower for isomaltulose
than for sucrose as reflected by the corresponding GI and II values shown in Table 2.
1 The Glycemic Index (GI) is defined by FAO/WHO (1998) as the incremental area under the blood glucose
response curve (IAUC) of a 50g carbohydrate portion of a test food expressed as a percentage of the
response to the same amount of carbohydrate from a standard food (typically glucose or white bread) taken
by the same subject. The Insulinemic Index (II) can be determined for the insulin curve, respectively.

CONFIDENTIAL
Plasma glucose change (mmol/L)
5
4
3
2
1
0
-1
-2
Regulatory Affairs &
Nutrition Communication – AJ/Se
28 May 2008
Page 7 of 16
0 30 60 90 120
Time (min)
Sucrose
Palatinose
Figure 3: Blood glucose curves of isomaltulose (Palatinose or sucrose (50g) determined in 10
healthy volunteers at the Sydney University’s Glycaemic Index Research Service (2002),
(mean±SD).
Plasma insulin change (pmol/L)
350
300
250
200
150
100
50
0
-50
0 30 60 90 120
Time (min)
Sucrose
Palatinose
Figure 4: Plasma insulin curves of isomaltulose (Palatinose) or sucrose (50g) determined in 10
healthy volunteers at the Sydney University’s Glycaemic Index Research Service (2002),
(mean±SD).

CONFIDENTIAL
Regulatory Affairs &
Nutrition Communication – AJ/Se
28 May 2008
Page 8 of 16
Table 2: GI and II values (based on glucose as reference) for isomaltulose, sucrose and maltodextrin
determined at the Sydney University’s Glycaemic Index Research Service (2002).
Carbohydrate GI II
Isomaltulose 32 30
Sucrose 68 64
Maltodextrin 86 79
Glucose 100 100
Similar findings were made in previous studies. In the study of Macdonald and Daniel
(1983), ten healthy volunteers (age 18-35), who were adapted to isomaltulose by 8 test
meals, were given varying doses of up to 1 g/kg body weight of isomaltulose or sucrose
dissolved in water after overnight fasting. Blood glucose samples were taken before and at
certain intervals for a period of 90 minutes after intake for the determination of glucose,
fructose and insulin concentrations.
Blood glucose and insulin levels increased more slowly and reached lower peaks after
isomaltulose than after sucrose intake. While the blood glucose level typically returned to
baseline within 60 minutes after sucrose intake, it had not returned to baseline by 80
minutes after isomaltulose intake at the higher concentrations. Also plasma insulin and
fructose levels were only about half with isomaltulose compared with sucrose.
Livesey (2004) reanalysed the blood glucose and insulin curves of this study and calculated
GI values between 34 and 49 and II values between 17 and 31 from them.
Kawai et al (1985, 1989) compared changes in blood glucose and insulin levels in response
to the consumption of isomaltulose or sucrose (50g in 150 ml water, after overnight fasting)
in healthy (n=8 resp. n=10) and non-insulin dependent type 2 diabetic volunteers (n=10)
over 120 minutes (healthy) or 180 minutes (diabetics), respectively. In healthy subjects,
plasma glucose after isomaltulose intake gradually increased and maintained plateau until
120 minutes after intake. After sucrose intake, as opposed to isomaltulose, blood glucose
levels showed a significantly higher increase with a distinct peak at 30 minutes and
returned to the initial baseline level within 120 minutes. The corresponding insulin curves
followed these patterns, respectively. The overall blood glucose and insulin responses were
lower after isomaltulose than after sucrose.
Livesey (2004) calculated a GI of 32 (both times) and II values of 22 and 26, respectively,
for these blood glucose and insulin curves of isomaltulose in healthy volunteers.
In diabetics, the blood glucose and insulin levels following isomaltulose and sucrose
increased more gradually than in healthy subjects. However, as with the healthy, the blood
glucose rise after isomaltulose occurred more slowly and was lower than after sucrose.

CONFIDENTIAL
Regulatory Affairs &
Nutrition Communication – AJ/Se
28 May 2008
Page 9 of 16
The data demonstrate that isomaltulose has a lower glycemic response than sucrose in
healthy subjects as well as in diabetics.
Similar results were reported by Liao et al (2001) after intake of 75 g isomaltulose or
sucrose in an aqueous solution in diabetic and non-diabetic volunteers (n=10 in each
group). Blood glucose, insulin and lipid levels were measured over 180 minutes and
significantly lower peak values and areas under the curves of blood glucose and insulin
after isomaltulose than after sucrose were reported in both groups.
In all studies mentioned above isomaltulose was consumed as an aqueous solution,
accordingly these studies reflect the intake in drinks/beverages. In the more complex food
matrixes of solid food, other factors that have an influence on blood glucose and insulin
levels may play a role. Examples of influencing factors are given in Table 3.
Table 3: Factors also influencing the glycemic effect of a food
Factors influencing GI
(Examples)
Decreasing effects Increasing effects
Nature of starch High amylose part High amylopectin part
Nature of mono-, di- and
oligosaccharides
Isomaltulose, fructose,
oligofructose
Dietary fibres Guar, ß-glucan, inulin
Glucose, sucrose
Method of processing Parboiling, cold extrusion Conv. cooking, HTHS 2
extrusion, flaking, puffing
Particle size Large particles Small particles
Ripeness and food storage Unripeness, cooling Ripeness
Alpha-amylase inhibitors Lectins, phytates
Interactions with other nutrients Fat / protein (high) Fat / protein (low)
2 HTHS: High Temperature High Shear

CONFIDENTIAL
Regulatory Affairs &
Nutrition Communication – AJ/Se
28 May 2008
Page 10 of 16
4. Discussion: General Characteristics of the Glycemic and Insulinemic Properties of
Isomaltulose as compared to Sucrose
After more than 20 years of research, in the past years there has been increasing
awareness and attention within the nutrition and health community as well as on a
consumer level towards the role of low glycemic carbohydrate foods within a healthy diet. A
steadily growing body of research shows that the carbohydrate-based diet recommended
for the general population should be based on mainly low-glycemic carbohydrates, as
parameters associated with the development of typical Western chronic diseases like
obesity, diabetes, coronary heart disease and possibly some types of cancer are influenced
beneficially, as compared to a carbohydrate-based high-glycemic diet. In other words, the
low-fat high-carbohydrate diet as advocated by WHO as well as health organisations in
Western countries could be further improved by switching from high glycemic to low
glycemic food choices. In this respect, WHO recommended in their expert consultation
paper on carbohydrates in 1998 to also consider the glycemic effect of food carbohydrates
when choosing between foods of similar composition within food groups. Since then, a
number of publications reviewed the scientific literature of the last 20 years of research in
this field in detail and strengthened the “good evidence” mentioned above (e.g. Augustin et
al 2002; Ludwig 2002; Jenkins 2002; Leeds 2002; Willett et al 2002; Brand-Miller et al
2003; Oppermann et al 2004; Berg et al 2005).
While high glycemic carbohydrates are characterised by a fast release and higher blood
glucose levels resulting in greater insulin demand, low glycemic carbohydrates cause only
a slower and overall low increase in blood glucose and corresponding insulin levels.
Further influencing factors are summarized earlier in this paper (Table 3).
Typical blood glucose and insulin changes in response to 50 g intake of isomaltulose and
sucrose are shown in Figures 5 and 6. These curves are based on a retrospective analysis
by Livesey (2004) taking the above mentioned studies (see under point 3) into account.

CONFIDENTIAL
Rise in blood glucose (mmol/L)
3,00
2,50
2,00
1,50
1,00
0,50
0,00
0 30 60 90 120
-0,50
Time (min)
Regulatory Affairs &
Nutrition Communication – AJ/Se
28 May 2008
Page 11 of 16
Sucrose
Palatinose
Figure 5: Retrospective analysis (Livesey 2004) - Typical blood glucose curves in response to isomaltulose
(Palatinose) and sucrose, as mean values of blood glucose and insulin measurements after intake of 50 g
isomaltulose or sucrose in healthy adults.
Rise in blood insulin (pmol/L)
120
100
80
60
40
20
0
0 30 60 90 120
Time (min)
Sucrose
Palatinose
Figure 6: Retrospective analysis (Livesey 2004) - Typical insulin curves in response to isomaltulose
(Palatinose) and sucrose, as mean values of blood glucose and insulin measurements after intake of 50 g
isomaltulose or sucrose in healthy adults.

CONFIDENTIAL
Regulatory Affairs &
Nutrition Communication – AJ/Se
28 May 2008
Page 12 of 16
In this retrospective analysis, Livesey also calculated the corresponding GI and II values for
the individual studies (Table 6). He obtained GI values between 32 and 49 with a mean GI
of 37 and II values between 17 and 31 with a mean II of 25.
Table 4: Average GI and II values of isomaltulose (intake as drink solution) calculated by Livesey
(2004) in a retrospective data analysis of published blood glucose measurements.
Study Dose [g] Type of subjects GI II
SUGiRS 50 Healthy 32 27
Macdonald and Daniel
1983
0,25 g/kg b.w. (~17,5) Healthy 45 17
0,5 g/kg b.w. (~35) Healthy 49 28
0,75 g/kg b.w. (~52,5) Healthy 37 31
1,0 g/kg b.w. (~70) Healthy 34 25
Kawai et al 1985 50 Healthy 32 22
Kawai et al 1989 50 Healthy 32 26
GI 37
(32 – 49)
II 25
(17 – 31)
Sucrose is known to cause a fast and high increase in blood glucose and insulin levels
followed by a fast decrease which might include a drop down below baseline as an
overreaction (hypoglycaemic reaction). This is a common pattern for medium to high
glycemic carbohydrates. The physiological characteristics of isomaltulose differ significantly
from those of sucrose.
Isomaltulose is low glycemic
Blood glucose measurements with both healthy and diabetic volunteers showed that – as a
consequence of the lower rate of hydrolysis and subsequent slower absorption from the
small intestine – isomaltulose exerts an overall low blood glucose rise over a longer period
of time. The maximum rise in blood glucose concentration (peak) after isomaltulose was
typically less than half that of sucrose. And also the total blood glucose response to the
consumption of isomaltulose, as reflected by its GI, was significantly lower (about 32% of
glucose) than sucrose (about 68% of glucose), (Sydney University’s Glycaemic Research
Service (SUGiRS) 2002).
For the purpose of comparison, foods or individual carbohydrates are classified – based on
their GI – into high glycemic (GI of 70 or more), medium glycemic (GI of 56-69), and low
glycemic (GI of 55 or less), (Brand-Miller et al 1999, Livesey 2003). A further class of very
low glycemic (GI of 40 or less) was defined as well (Livesey 2003). According to this
classification, isomaltulose is low or even very low glycemic and has the potential to reduce
the glycemic properties of a food.

CONFIDENTIAL
Regulatory Affairs &
Nutrition Communication – AJ/Se
28 May 2008
Page 13 of 16
Isomaltulose is low insulinemic
Blood glucose concentrations are tightly regulated by a hormonal regulatory system in
order to facilitate constant glucose supply. A rise in blood glucose concentrations following
a meal (hyperglycemia) stimulates the release of the hormone insulin (and inhibits the
release of the antagonist glucagons and others). Insulin promotes the uptake of glucose
(and other nutrients) by the muscles and adipose tissue bringing blood glucose levels back
to normal.
The low rise in blood glucose concentrations after isomaltulose intake is associated with a
low insulin demand, respectively. The insulin response to isomaltulose ingestion, according
to the II value, was 30% of that of glucose, while the insulin response to sucrose was 64%
of that of glucose (Sydney University’s Glycaemic Research Service (SUGiRS) 2002).
Isomaltulose is not associated with a pronounced relative hypoglycemia
Typical blood glucose curves of sucrose show a pronounced phase of relative
hypoglycemia. Sucrose is rapidly digested and absorbed from the small intestine. The
corresponding rapid and high rise in blood glucose concentrations exerts a strong
stimulation on insulin release. While nutrient absorption from sucrose rapidly declines
shortly after, the effects of high insulin and low glucagon levels still remain causing blood
glucose levels to drop rapidly and below baseline into a state of relative hypoglycemia.
Isomaltulose, unlike sucrose, is not associated with a pronounced hypoglycaemic phase.
Isomaltulose provides glucose over a longer period of time (prolonged energy
supply)
After sucrose ingestion, blood glucose levels returned to near baseline levels (or even
below baseline) after 1 hour in most of the studies with healthy volunteers. In contrast,
blood glucose levels after isomaltulose remained elevated at a lower level but for a longer
period of time, typically lasting for most of the entire measurement of 90 minutes
(Macdonald and Daniel 1983) or 120 minutes (Kawai et al 1985, 1989; SUGiRS 2002). This
means, the phase during which glucose is supplied to the cells is longer after isomaltulose
than after sucrose. Glucose delivery time to the cells is prolonged by isomaltulose.
Glucose is ‘fuel’ for the body, it is the energy supply to the cells, e.g. in the brain and
muscles. An extended glucose supply means an extended energy supply in the form of
glucose to the cells.

CONFIDENTIAL
Regulatory Affairs &
Nutrition Communication – AJ/Se
28 May 2008
Page 14 of 16
5. Conclusion
Isomaltulose is a ‘slow release carbohydrate’. Due to its digestion properties, isomaltulose
is a well tolerated ‘available carbohydrate’. It is a glucose provider for the body and an
energy provider as well, supplying glucose at a lower level (avoiding pronounced
hyperglycaemic and hypoglycaemic phases) for a prolonged time compared to sucrose.
Isomaltulose is low glycemic and low insulinemic whilst providing the same total calories as
sucrose.
Having these physiological properties in mind, isomaltulose is a carbohydrate for
people following a healthy and active lifestyle. Isomaltulose is a useful tool to bring a
carbohydrate-based low-glycemic diet into practice.

CONFIDENTIAL
References
Regulatory Affairs &
Nutrition Communication – AJ/Se
28 May 2008
Page 15 of 16
Augustin, L.S., Franceschi, S., Jenkins, D.J.A., Kendall, C.W.C. , La Vecchia, C. (2002) Glycemic index in
chronic disease - a review. Eur J Clin Nutr 56, 1049-1071.
Brand-Miller, J., Wolever, T.M.S., Colagiuri, S., Foster-Powell, K. (1999) The Glucose Revolution. Marlow
and Company, New York.
Brand-Miller, J., Wolever, T.M.S., Colagiuri, S., Foster-Powell, K. (1999) The Glucose Revolution. Marlow
and Company, New York.
Brand-Miller, J., Hayne, S., Petocz, P., Colagiuri, S. (2003) Low glycemic index diets in the management of
diabetes: a meta-analysis of randomised controlled trials. Diabetes Care 26, 2261-2267.
Dahlqvist, A. (1961) Hydrolysis of Palatinose (isomaltulose) by pig intestinal glycosidases. Acta Chem Scan
15, 808-816.
Dahlquist, A., Auricchio, S., Semenza, G., Prader,A. (1963) Human intestinal disaccharidases and heriditary
disaccaride intolerance. The hydrolysis of sucrose, isomaltulose, Palatinose (isomaltulose) and a
1,6-α-oligosaccharide (isomalto-oligosaccharide) preparation. J Clin Invest, 42, 556-562.
Eggleston, G., Grisham, M. (2003) Oligosacchrides in cane and their formation on cane deterioration. In:
ACS Symposium Series 849,Chapter 16, 211-232.
EU Commission Decision 2005/581/EC of 25 th July 2005 authorising the placing on the market of
isomaltulose as a novel food or novel food ingredient under Regulation (EC) No 258/97of the European
Parliament and of the Council. Official Journal of the European Union of 29.7.2005, L199/90 – 91.
FAO/WHO (1998) Carbohydrates in human nutrition. In FAO Nutrition Paper No.66. Rome, FAO.
FDA (2006) – US GRAS Notification No 0184 on isomaltulose.
Gostner, A., Holub, S., Theis, S., Volk, A., Kozianowski, G., Scheppach, W. (2006) Intestinal digestibility of
isomaltulose (Palatinose) in patients with ileostoma [Intestinale Verdaulichkeit von Palatinose bei
Patienten mit Ileostoma]. Z. Gastroenterol 44, 1073-1094. [Abstract P 18; poster presented at the 34 th
Congress of the German Society of Gastroenterology (Gesellschaft für Gastroenterologie), 26.-
28.10.2006, Rosenheim].
Grupp, U., Siebert, G. (1978) Metabolism of hydrogenated isomaltulose, an equimolar mixture of alpha-Dglucopyranosido-1,6-sorbitol
and alpha-D-glucopyranosido-1,6-mannitol. Res Exp Med (Berl.) 173, 261-
278.
Günther, S., Heymann, H. (1998) Di- and oligosaccharide substrate specificities and subsite binding
energies of pig intestinal glucoamylase-maltase. Arch Biochem Biophys 354, 111-116.
Heinz, F. (1987) Enzymatische Spaltung von Zuckeraustauschstoffen durch isolierte Enzyme und
Enzymkomplexe der Dünndarmmukosa. Medizinische Hochschule Hannover, Zentrum Biochemie,
Universität Hannover, Forschungsvorhaben 6539.
Heymann, H., Heinz,F. (1987) Kinetic studies on glucoamylase/maltase and sucrase/isomaltase complex of
human, pig and rat intestinal mucosa. 4 th European Carbohydrate Symposium, Darmstadt, (12.-
17.07.87).
Jenkins, D.J.A., Kendall, C.W.C., Augustin, L.S.A., Franceschi, S., Hamidi, M., Marchie, A., Jenkins, A.L.,
Axelsen, M. (2002) Glycemic index: overview of implications in health and disease. Am J Clin Nutr 76,
266S-273S.

CONFIDENTIAL
Regulatory Affairs &
Nutrition Communication – AJ/Se
28 May 2008
Page 16 of 16
Kawai, K., Okuda, Y., Yamashita, K. (1985) Changes in blood glucose and insulin after an oral Isomaltulose
administration in normal subjects. Endocrinol Jpn 32, 933-936.
Kawai, K., Yoshikawa, H., Muryama, Y., Okuda, Y., Yamashita, K. (1989) Usefulness of Isomaltulose as a
caloric sweetener for diabetic patients. Horm Metabol Res 21, 338-340.
Leeds, A.R. (2002) Glycemic index and heart disease. Am J Clin Nutr 76 (1) 286S-189S.
Liao, Z.H., Li, Y.B., Yao, B., Fan, H.D., Hu, G.L., Weng, J.P. (2001). The effects of isomaltulose on blood
glucose and lipids for diabetic subjects. Diabetes 50 Suppl., 1530-P, A366.
Lina, B.A.R., Jonker, D., Kozianowski, G. (2002) Isomaltulose (Palatinose): a review of biological and
toxicological studies. Food Chem Tox 40, 1375-1381.
Livesey, G. (2003) Health potential of polyols as sugar replacers, with emphasis on low glyaemic properties.
Nutr Res Rev 16, 163-191.
Livesey, G. (2004) Retrospective analysis on the glycemic and insulinemic response curves to isomaltulose
in healthy humans. Expert opinion within personal communication.
Ludwig (2002) The gylcemic index – physiological mechanisms relating to obesity, diabetes, and
cardiovascular disease. JAMA 287, 2414-2423.
Macdonald, M., Daniel, J.W. (1983) The bio-availability of isomaltulose in man and rat. Nutr Rep Int 28,
1083-1090.
Oppermann, A.M., Venter, C.S., Oosthuizen, W., Thompson, R.L., Vorster, H.H. (2004) Meta-analysis of the
health effects of using the glycaemic index in meal-planning. Br J Nutr 92, 367-81.
Siddiqui, I.R., Furgala, B. (1967) Isolation and characterization of oligosaccharides from honey. Part I:
Disaccharides. J Agric Res 6, 139-145.
Sydney University’s Glycaemic Research Service (SUGiRS) (2002) Glycaemic Index Report – Isomaltulose.
Unpublished. See also: Sydney University: GI Database at www.glycemicindex.com
Tsuji, Y., Yamada,K., Hosoya, N., Moriuchi, S. (1986) Digestion and absorption of sugars and sugar
substitutes in rat small intestine. J Nutr Sci Vitaminol 32, 93-100.
Van Weerden, E:J:, Huisman, J., Van Leeuwen, P. (1983) Digestion processes of isomaltulose and
saccharose in the small and large intestine of the pig. ILOB-Report No. 520.
Weidenhagen, R., Lorenz, S. (1957) Isomaltulose TM (6-α-Glucopyranosido-fructofuranose), ein neues
bakterielles Umwandlungsprodukt der Saccharose. Zeitschrift für die Zuckerindustrie 7 (11), 533-534.
Willet, W., Manson, J., Liu, S. (2002) Glycemic index, glycemic load, and risk of type 2 diabetes. Am J Clin
Nutr 76 (suppl) 274S-280S.
Yamada, K., Shinohara, H. Hosoya, N. (1985) Hydrolysis of 1-0-alpha-D-glucopyranosyl-D-fructofuranose
(Trehalulose) by rat intestinal sucrase-isomaltase complex. Nutr Rep Int 32, 1211-1220.
Ziesenitz, S.C. (1986a) Zur Verwertung des Zuckeraustauschstoffs Palatinit® im Stoffwechsel. Beiträge zu
Infusionstherapie und Klinische Ernährung 16, 120-132.
Ziesenitz, S.C. (1986b) Stufenweises Prüfschema für Zuckeraustauschstoffe – Vorprüfung mittels
Enzymen. 3. Carbohydrasen aus Jejunalmucosa des Menschen. Z Ernährungswiss 25, 253-258.

Annex 1.8: Report from Imfeld (University of Zürich): “Pilot
Study on the Dental Care Properties of Isomaltulose
(PalatinoseTM)”.

Annex 2.1: “The Determination of a Caloric Value for Inulin
and Oligofructose”, Cantox Inc., 1999.

Annex 2.2: Roberfroid, M.B., 1999; “Caloric Value of Inulin
and Oligofructose”.

Annex 3.1: Scientific Evaluation by FASEB, 1994;
“Evaluation of the Energy of Certain Sugar Alcohols used as
Food Ingredients”

Annex 3.2: “Letter of no objection” from USA for the
caloric value of 2.0 kcal/g for isomalt.

Annex 4.1: Codex Alimentarius CAC/GL 23‐1997
“Guidelines for Use of Nutrition and Health Claims”, Table
of Conditions for Nutrient Content.

1 Nutrition and Health Claims (CAC/GL 23-1997)
1. SCOPE
GUIDELINES FOR USE OF NUTRITION AND HEALTH CLAIMS
Adopted in 1997. Revised in 2004. Amended in 2001, 2008 and 2009.
CAC/GL 23-1997
Nutrition claims should be consistent with national nutrition policy and support that policy. Only nutrition claims
that support national nutrition policy should be allowed.
Health claims should be consistent with national health policy, including nutrition policy, and support such
policies where applicable. Health claims should be supported by a sound and sufficient body of scientific
evidence to substantiate the claim, provide truthful and non-misleading information to aid consumers in
choosing healthful diets and be supported by specific consumer education. The impact of health claims on
consumers’ eating behaviours and dietary patterns should be monitored, in general, by competent authorities.
Claims of the type described in section 3.4 of the Codex General Guidelines on Claims are prohibited.
1.1 These guidelines relate to the use of nutrition and health claims in food labelling and, where required by the
authorities having jurisdiction, in advertising 1 .
1.2 These guidelines apply to all foods for which nutrition and health claims are made without prejudice to specific
provisions under Codex standards or Guidelines relating to Foods for Special Dietary Uses and Foods for
Special Medical Purposes.
1.3 These guidelines are intended to supplement the Codex General Guidelines on Claims and do not supersede
any prohibitions contained therein.
1.4 Nutrition and health claims shall not be permitted for foods for infants and young children except where
specifically provided for in relevant Codex standards or national legislation.
2. DEFINITIONS
2.1 Nutrition claim means any representation which states, suggests or implies that a food has particular
nutritional properties including but not limited to the energy value and to the content of protein, fat and
carbohydrates, as well as the content of vitamins and minerals. The following do not constitute nutrition claims:
(a) the mention of substances in the list of ingredients;
(b) the mention of nutrients as a mandatory part of nutrition labelling;
(c) quantitative or qualitative declaration of certain nutrients or ingredients on the label if required by national
legislation.
2.1.1 Nutrient content claim is a nutrition claim that describes the level of a nutrient contained in a food.
(Examples: “source of calcium”; “high in fibre and low in fat”.)
2.1.2 Nutrient comparative claim is a claim that compares the nutrient levels and/or energy value of two or more
foods.
(Examples: “reduced”; “less than”; “fewer”; “increased”; “more than”.)
2. 2 Health claim means any representation that states, suggests, or implies that a relationship exists between a
food or a constituent of that food and health. Health claims include the following:
2.2.1 Nutrient function claims – a nutrition claim that describes the physiological role of the nutrient in growth,
development and normal functions of the body.
1 Advertising means any commercial communication to the public, by any means other than labelling, in order to promote directly or indirectly, the sale or
intake of a food through the use of nutrition and health claims in relation to the food and its ingredients.

2 Nutrition and Health Claims (CAC/GL 23-1997)
Example:
“Nutrient A (naming a physiological role of nutrient A in the body in the maintenance of health and
promotion of normal growth and development). Food X is a source of/ high in nutrient A.”
2.2.2 Other function claims – These claims concern specific beneficial effects of the consumption of foods or their
constituents, in the context of the total diet on normal functions or biological activities of the body. Such claims
relate to a positive contribution to health or to the improvement of a function or to modifying or preserving
health.
Examples:
“Substance A (naming the effect of substance A on improving or modifying a physiological function or
biological activity associated with health). Food Y contains x grams of substance A.”
2.2.3 Reduction of disease risk claims – Claims relating the consumption of a food or food constituent, in the
context of the total diet, to the reduced risk of developing a disease or health-related condition.
Risk reduction means significantly altering a major risk factor(s) for a disease or health-related condition.
Diseases have multiple risk factors and altering one of these risk factors may or may not have a beneficial
effect. The presentation of risk reduction claims must ensure, for example, by use of appropriate language and
reference to other risk factors, that consumers do not interpret them as prevention claims.
Examples:
“A healthful diet low in nutrient or substance A may reduce the risk of disease D.
Food X is low in nutrient or substance A.”
“A healthful diet rich in nutrient or substance A may reduce the risk of disease D.
Food X is high in nutrient or substance A.”
3. NUTRITION LABELLING
Any food for which a nutrition or health claim is made should be labelled with a nutrient declaration in accordance
with Section 3 of the Codex Guidelines on Nutrition Labelling.
4. NUTRITION CLAIMS
4.1 The only nutrition claims permitted shall be those relating to energy, protein, carbohydrate, and fat and
components thereof, fibre, sodium and vitamins and minerals for which Nutrient Reference Values (NRVs)
have been laid down in the Codex Guidelines for Nutrition Labelling.
5. NUTRIENT CONTENT CLAIMS
5.1 When a nutrient content claim that is listed in the Table to these Guidelines or a synonymous claim is made,
the conditions specified in the Table for that claim should apply.
5.2 Where a food is by its nature low in or free of the nutrient that is the subject of the claim, the term describing
the level of the nutrient should not immediately precede the name of the food but should be in the form “a low
(naming the nutrient) food” or “a (naming the nutrient)-free food”.
6. COMPARATIVE CLAIMS
Comparative claims should be permitted subject to the following conditions and based on the food as sold,
taking into account further preparation required for consumption according to the instructions for use on the
label:
6.1 The foods being compared should be different versions of the same food or similar foods. The foods being
compared should be clearly identified.
6.2 A statement of the amount of difference in the energy value or nutrient content should be given. The following
information should appear in close proximity to the comparative claim:
6.2.1 The amount of difference related to the same quantity, expressed as a percentage, fraction, or an absolute
amount. Full details of the comparison should be given.
6.2.2 The identity of the food(s) to which the food is being compared. The food(s) should be described in such a
manner that it (they) can be readily identified by consumers.

3 Nutrition and Health Claims (CAC/GL 23-1997)
6.3 The comparison should be based on a relative difference of at least 25% in the energy value or nutrient
content, except for micronutrients where a 10% difference in the NRV would be acceptable, between the
compared foods and a minimum absolute difference in the energy value or nutrient content equivalent to the
figure defined as “low” or as a “source” in the Table to these Guidelines.
6.4 The use of the word “light” should follow the same criteria as for “reduced” and include an indication of the
characteristics which make the food “light”.
7. HEALTH CLAIMS
7.1 Health claims should be permitted provided that all of the following conditions are met:
7.1.1 Health claims must be based on current relevant scientific substantiation and the level of proof must be
sufficient to substantiate the type of claimed effect and the relationship to health as recognized by generally
accepted scientific review of the data and the scientific substantiation should be reviewed as new knowledge
becomes available. 2 The health claim must consist of two parts:
1) Information on the physiological role of the nutrient or on an accepted diet-health relationship; followed by
2) Information on the composition of the product relevant to the physiological role of the nutrient or the
accepted diet-health relationship unless the relationship is based on a whole food or foods whereby the
research does not link to specific constituents of the food.
7.1.2 Any health claim must be accepted by or be acceptable to the competent authorities of the country where the
product is sold.
7.1.3 The claimed benefit should arise from the consumption of a reasonable quantity of the food or food constituent
in the context of a healthy diet.
7.1.4 If the claimed benefit is attributed to a constituent in the food, for which a Nutrient Reference value is
established, the food in question should be:
(i) a source of or high in the constituent in the case where increased consumption is recommended; or,
(ii) low in, reduced in, or free of the constituent in the case where reduced consumption is recommended.
Where applicable, the conditions for nutrient content claims and comparative claims will be used to
determine the levels for “high”, “low”, “reduced”, and “free”.
7.1.5 Only those essential nutrients for which a Nutrient Reference Value (NRV) has been established in the Codex
Guidelines on Nutrition Labelling or those nutrients which are mentioned in officially recognized dietary
guidelines of the national authority having jurisdiction, should be the subject of a nutrient function claim.
7.2 Health claims should have a clear regulatory framework for qualifying and/or disqualifying conditions for
eligibility to use the specific claim, including the ability of competent national authorities to prohibit claims
made for foods that contain nutrients or constituents in amounts that increase the risk of disease or an
adverse health-related condition. The health claim should not be made if it encourages or condones excessive
consumption of any food or disparages good dietary practice.
7.3 If the claimed effect is attributed to a constituent of the food, there must be a validated method to quantify the
food constituent that forms the basis of the claim.
7.4 The following information should appear on the label or labelling of the food bearing health claims:
7.4.1 A statement of the quantity of any nutrient or other constituent of the food that is the subject of the claim.
7.4.2 The target group, if appropriate.
7.4.3 How to use the food to obtain the claimed benefit and other lifestyle factors or other dietary sources, where
appropriate.
7.4.4 If appropriate, advice to vulnerable groups on how to use the food and to groups, if any, who need to avoid the
food.
7.4.5 Maximum safe intake of the food or constituent where necessary.
2 Reference to the Scientific Criteria for Health Related Claims being developed by the Codex Committee on Nutrition and Foods for Special Dietary Uses
should be inserted here.

4 Nutrition and Health Claims (CAC/GL 23-1997)
7.4.6 How the food or food constituent fits within the context of the total diet.
7.4.7 A statement on the importance of maintaining a healthy diet.
8. CLAIMS RELATED TO DIETARY GUIDELINES OR HEALTHY DIETS
Claims that relate to dietary guidelines or “healthy diets” should be permitted subject to the following
conditions:
8.1 Only claims related to the pattern of eating contained in dietary guidelines officially recognized by the
appropriate national authority.
8.2 Flexibility in the wording of claims is acceptable, provided the claims remain faithful to the pattern of eating
outlined in the dietary guidelines.
8.3 Claims related to a “healthy diet” or any synonymous term are considered to be claims about the pattern of
eating contained in dietary guidelines and should be consistent with the guidelines.
8.4 Foods which are described as part of a healthy diet, healthy balance, etc., should not be based on selective
consideration of one or more aspects of the food. They should satisfy certain minimum criteria for other major
nutrients related to dietary guidelines.
8.5 Foods should not be described as “healthy” or be represented in a manner that implies that a food in and of
itself will impart health.
8.6 Foods may be described as part of a “healthy diet” provided that the label carries a statement relating the food
to the pattern of eating described in the dietary guidelines.
Table of conditions for nutrient contents
COMPONENT CLAIM CONDITIONS (not more than)
Energy
Fat
Saturated Fat
Cholesterol
Low
40 kcal (170 kJ) per 100 g (solids)
or
20 kcal (80 kJ) per 100 ml (liquids)
Free 4 kcal per 100 ml (liquids)
Low
3 g per 100 g (solids)
1.5 g per 100 ml (liquids)
Free 0.5 g per 100 g (solids) or 100 ml (liquids)
Low 3
Free
Low 3
Free
Sugars Free
1.5 g per 100 g (solids)
0.75 g per 100 ml (liquids)
and 10% of energy
0.1 g per 100 g (solids)
0.1 g per 100 ml (liquids)
0.02 g per 100 g (solids)
0.01 g per 100 ml (liquids)
0.005 g per 100 g (solids)
0.005 g per 100 ml (solids)
and, for both claims, less than:1.5 g saturated fat per 100 g (solids)
0.75 g saturated fat per 100 ml (liquids)
and 10% of energy of saturated fat
0.5 g per 100 g (solids)
0.5 g per 100 ml (liquids)
3
In the case of the claim “low in saturated fat”, trans fatty acids should be taken into account where applicable. This provision consequentially applies to
foods claimed to be “low in cholesterol” and “cholesterol free”.

5 Nutrition and Health Claims (CAC/GL 23-1997)
COMPONENT CLAIM CONDITIONS (not less than)
Sodium
Protein
Vitamins and Minerals
Dietary Fibre
Low 0.12 g per 100 g
Very Low 0.04 g per 100 g
Free 0.005 g per 100 g
Source
10% of NRV per 100 g (solids)
5% of NRV per 100 ml (liquids)
or 5% of NRV per 100 kcal (12% of NRV per 1 MJ)
or 10% of NRV per serving
High 2 times the values for “source”
Source
15% of NRV per 100 g (solids)
7.5% of NRV per100 ml (liquids)
or 5% of NRV per 100 kcal (12% of NRV per 1 MJ)
or 15% of NRV per serving
High 2 times the value for “source”
Source
High
3 g per 100 g 4 or 1.5 g per 100 kcal
or 10 % of daily reference value per serving 5
6 g per 100 g 4 or 3 g per 100 kcal
or 20 % of daily reference value per serving 5
4 Conditions for nutrient content claims for dietary fibre in liquid foods to be determined at national level.
5 Serving size and daily reference value to be determined at national level.

Annex 3.1: Scientific Evaluation by FASEB, 1994;
“Evaluation of the Energy of Certain Sugar Alcohols used as
Food Ingredients”

Annex 3.2: “Letter of no objection” from USA for the
caloric value of 2.0 kcal/g for isomalt.

Annex 4.1: Codex Alimentarius CAC/GL 23‐1997
“Guidelines for Use of Nutrition and Health Claims”, Table
of Conditions for Nutrient Content.

1 Nutrition and Health Claims (CAC/GL 23-1997)
1. SCOPE
GUIDELINES FOR USE OF NUTRITION AND HEALTH CLAIMS
Adopted in 1997. Revised in 2004. Amended in 2001, 2008 and 2009.
CAC/GL 23-1997
Nutrition claims should be consistent with national nutrition policy and support that policy. Only nutrition claims
that support national nutrition policy should be allowed.
Health claims should be consistent with national health policy, including nutrition policy, and support such
policies where applicable. Health claims should be supported by a sound and sufficient body of scientific
evidence to substantiate the claim, provide truthful and non-misleading information to aid consumers in
choosing healthful diets and be supported by specific consumer education. The impact of health claims on
consumers’ eating behaviours and dietary patterns should be monitored, in general, by competent authorities.
Claims of the type described in section 3.4 of the Codex General Guidelines on Claims are prohibited.
1.1 These guidelines relate to the use of nutrition and health claims in food labelling and, where required by the
authorities having jurisdiction, in advertising 1 .
1.2 These guidelines apply to all foods for which nutrition and health claims are made without prejudice to specific
provisions under Codex standards or Guidelines relating to Foods for Special Dietary Uses and Foods for
Special Medical Purposes.
1.3 These guidelines are intended to supplement the Codex General Guidelines on Claims and do not supersede
any prohibitions contained therein.
1.4 Nutrition and health claims shall not be permitted for foods for infants and young children except where
specifically provided for in relevant Codex standards or national legislation.
2. DEFINITIONS
2.1 Nutrition claim means any representation which states, suggests or implies that a food has particular
nutritional properties including but not limited to the energy value and to the content of protein, fat and
carbohydrates, as well as the content of vitamins and minerals. The following do not constitute nutrition claims:
(a) the mention of substances in the list of ingredients;
(b) the mention of nutrients as a mandatory part of nutrition labelling;
(c) quantitative or qualitative declaration of certain nutrients or ingredients on the label if required by national
legislation.
2.1.1 Nutrient content claim is a nutrition claim that describes the level of a nutrient contained in a food.
(Examples: “source of calcium”; “high in fibre and low in fat”.)
2.1.2 Nutrient comparative claim is a claim that compares the nutrient levels and/or energy value of two or more
foods.
(Examples: “reduced”; “less than”; “fewer”; “increased”; “more than”.)
2. 2 Health claim means any representation that states, suggests, or implies that a relationship exists between a
food or a constituent of that food and health. Health claims include the following:
2.2.1 Nutrient function claims – a nutrition claim that describes the physiological role of the nutrient in growth,
development and normal functions of the body.
1 Advertising means any commercial communication to the public, by any means other than labelling, in order to promote directly or indirectly, the sale or
intake of a food through the use of nutrition and health claims in relation to the food and its ingredients.

2 Nutrition and Health Claims (CAC/GL 23-1997)
Example:
“Nutrient A (naming a physiological role of nutrient A in the body in the maintenance of health and
promotion of normal growth and development). Food X is a source of/ high in nutrient A.”
2.2.2 Other function claims – These claims concern specific beneficial effects of the consumption of foods or their
constituents, in the context of the total diet on normal functions or biological activities of the body. Such claims
relate to a positive contribution to health or to the improvement of a function or to modifying or preserving
health.
Examples:
“Substance A (naming the effect of substance A on improving or modifying a physiological function or
biological activity associated with health). Food Y contains x grams of substance A.”
2.2.3 Reduction of disease risk claims – Claims relating the consumption of a food or food constituent, in the
context of the total diet, to the reduced risk of developing a disease or health-related condition.
Risk reduction means significantly altering a major risk factor(s) for a disease or health-related condition.
Diseases have multiple risk factors and altering one of these risk factors may or may not have a beneficial
effect. The presentation of risk reduction claims must ensure, for example, by use of appropriate language and
reference to other risk factors, that consumers do not interpret them as prevention claims.
Examples:
“A healthful diet low in nutrient or substance A may reduce the risk of disease D.
Food X is low in nutrient or substance A.”
“A healthful diet rich in nutrient or substance A may reduce the risk of disease D.
Food X is high in nutrient or substance A.”
3. NUTRITION LABELLING
Any food for which a nutrition or health claim is made should be labelled with a nutrient declaration in accordance
with Section 3 of the Codex Guidelines on Nutrition Labelling.
4. NUTRITION CLAIMS
4.1 The only nutrition claims permitted shall be those relating to energy, protein, carbohydrate, and fat and
components thereof, fibre, sodium and vitamins and minerals for which Nutrient Reference Values (NRVs)
have been laid down in the Codex Guidelines for Nutrition Labelling.
5. NUTRIENT CONTENT CLAIMS
5.1 When a nutrient content claim that is listed in the Table to these Guidelines or a synonymous claim is made,
the conditions specified in the Table for that claim should apply.
5.2 Where a food is by its nature low in or free of the nutrient that is the subject of the claim, the term describing
the level of the nutrient should not immediately precede the name of the food but should be in the form “a low
(naming the nutrient) food” or “a (naming the nutrient)-free food”.
6. COMPARATIVE CLAIMS
Comparative claims should be permitted subject to the following conditions and based on the food as sold,
taking into account further preparation required for consumption according to the instructions for use on the
label:
6.1 The foods being compared should be different versions of the same food or similar foods. The foods being
compared should be clearly identified.
6.2 A statement of the amount of difference in the energy value or nutrient content should be given. The following
information should appear in close proximity to the comparative claim:
6.2.1 The amount of difference related to the same quantity, expressed as a percentage, fraction, or an absolute
amount. Full details of the comparison should be given.
6.2.2 The identity of the food(s) to which the food is being compared. The food(s) should be described in such a
manner that it (they) can be readily identified by consumers.

3 Nutrition and Health Claims (CAC/GL 23-1997)
6.3 The comparison should be based on a relative difference of at least 25% in the energy value or nutrient
content, except for micronutrients where a 10% difference in the NRV would be acceptable, between the
compared foods and a minimum absolute difference in the energy value or nutrient content equivalent to the
figure defined as “low” or as a “source” in the Table to these Guidelines.
6.4 The use of the word “light” should follow the same criteria as for “reduced” and include an indication of the
characteristics which make the food “light”.
7. HEALTH CLAIMS
7.1 Health claims should be permitted provided that all of the following conditions are met:
7.1.1 Health claims must be based on current relevant scientific substantiation and the level of proof must be
sufficient to substantiate the type of claimed effect and the relationship to health as recognized by generally
accepted scientific review of the data and the scientific substantiation should be reviewed as new knowledge
becomes available. 2 The health claim must consist of two parts:
1) Information on the physiological role of the nutrient or on an accepted diet-health relationship; followed by
2) Information on the composition of the product relevant to the physiological role of the nutrient or the
accepted diet-health relationship unless the relationship is based on a whole food or foods whereby the
research does not link to specific constituents of the food.
7.1.2 Any health claim must be accepted by or be acceptable to the competent authorities of the country where the
product is sold.
7.1.3 The claimed benefit should arise from the consumption of a reasonable quantity of the food or food constituent
in the context of a healthy diet.
7.1.4 If the claimed benefit is attributed to a constituent in the food, for which a Nutrient Reference value is
established, the food in question should be:
(i) a source of or high in the constituent in the case where increased consumption is recommended; or,
(ii) low in, reduced in, or free of the constituent in the case where reduced consumption is recommended.
Where applicable, the conditions for nutrient content claims and comparative claims will be used to
determine the levels for “high”, “low”, “reduced”, and “free”.
7.1.5 Only those essential nutrients for which a Nutrient Reference Value (NRV) has been established in the Codex
Guidelines on Nutrition Labelling or those nutrients which are mentioned in officially recognized dietary
guidelines of the national authority having jurisdiction, should be the subject of a nutrient function claim.
7.2 Health claims should have a clear regulatory framework for qualifying and/or disqualifying conditions for
eligibility to use the specific claim, including the ability of competent national authorities to prohibit claims
made for foods that contain nutrients or constituents in amounts that increase the risk of disease or an
adverse health-related condition. The health claim should not be made if it encourages or condones excessive
consumption of any food or disparages good dietary practice.
7.3 If the claimed effect is attributed to a constituent of the food, there must be a validated method to quantify the
food constituent that forms the basis of the claim.
7.4 The following information should appear on the label or labelling of the food bearing health claims:
7.4.1 A statement of the quantity of any nutrient or other constituent of the food that is the subject of the claim.
7.4.2 The target group, if appropriate.
7.4.3 How to use the food to obtain the claimed benefit and other lifestyle factors or other dietary sources, where
appropriate.
7.4.4 If appropriate, advice to vulnerable groups on how to use the food and to groups, if any, who need to avoid the
food.
7.4.5 Maximum safe intake of the food or constituent where necessary.
2
Reference to the Scientific Criteria for Health Related Claims being developed by the Codex Committee on Nutrition and Foods for Special Dietary Uses
should be inserted here.

4 Nutrition and Health Claims (CAC/GL 23-1997)
7.4.6 How the food or food constituent fits within the context of the total diet.
7.4.7 A statement on the importance of maintaining a healthy diet.
8. CLAIMS RELATED TO DIETARY GUIDELINES OR HEALTHY DIETS
Claims that relate to dietary guidelines or “healthy diets” should be permitted subject to the following
conditions:
8.1 Only claims related to the pattern of eating contained in dietary guidelines officially recognized by the
appropriate national authority.
8.2 Flexibility in the wording of claims is acceptable, provided the claims remain faithful to the pattern of eating
outlined in the dietary guidelines.
8.3 Claims related to a “healthy diet” or any synonymous term are considered to be claims about the pattern of
eating contained in dietary guidelines and should be consistent with the guidelines.
8.4 Foods which are described as part of a healthy diet, healthy balance, etc., should not be based on selective
consideration of one or more aspects of the food. They should satisfy certain minimum criteria for other major
nutrients related to dietary guidelines.
8.5 Foods should not be described as “healthy” or be represented in a manner that implies that a food in and of
itself will impart health.
8.6 Foods may be described as part of a “healthy diet” provided that the label carries a statement relating the food
to the pattern of eating described in the dietary guidelines.
Table of conditions for nutrient contents
COMPONENT CLAIM CONDITIONS (not more than)
Energy
Fat
Saturated Fat
Cholesterol
Low
40 kcal (170 kJ) per 100 g (solids)
or
20 kcal (80 kJ) per 100 ml (liquids)
Free 4 kcal per 100 ml (liquids)
Low
3 g per 100 g (solids)
1.5 g per 100 ml (liquids)
Free 0.5 g per 100 g (solids) or 100 ml (liquids)
Low 3
Free
Low 3
Free
Sugars Free
1.5 g per 100 g (solids)
0.75 g per 100 ml (liquids)
and 10% of energy
0.1 g per 100 g (solids)
0.1 g per 100 ml (liquids)
0.02 g per 100 g (solids)
0.01 g per 100 ml (liquids)
0.005 g per 100 g (solids)
0.005 g per 100 ml (solids)
and, for both claims, less than:1.5 g saturated fat per 100 g (solids)
0.75 g saturated fat per 100 ml (liquids)
and 10% of energy of saturated fat
0.5 g per 100 g (solids)
0.5 g per 100 ml (liquids)
3
In the case of the claim “low in saturated fat”, trans fatty acids should be taken into account where applicable. This provision consequentially applies to
foods claimed to be “low in cholesterol” and “cholesterol free”.

5 Nutrition and Health Claims (CAC/GL 23-1997)
COMPONENT CLAIM CONDITIONS (not less than)
Sodium
Protein
Vitamins and Minerals
Dietary Fibre
Low 0.12 g per 100 g
Very Low 0.04 g per 100 g
Free 0.005 g per 100 g
Source
10% of NRV per 100 g (solids)
5% of NRV per 100 ml (liquids)
or 5% of NRV per 100 kcal (12% of NRV per 1 MJ)
or 10% of NRV per serving
High 2 times the values for “source”
Source
15% of NRV per 100 g (solids)
7.5% of NRV per100 ml (liquids)
or 5% of NRV per 100 kcal (12% of NRV per 1 MJ)
or 15% of NRV per serving
High 2 times the value for “source”
Source
High
3 g per 100 g 4 or 1.5 g per 100 kcal
or 10 % of daily reference value per serving 5
6 g per 100 g 4 or 3 g per 100 kcal
or 20 % of daily reference value per serving 5
4 Conditions for nutrient content claims for dietary fibre in liquid foods to be determined at national level.
5 Serving size and daily reference value to be determined at national level.