Professional Documents
Culture Documents
UNIB20014
SEMESTER 2, 2016
Topic 4: Food Biotechnology & Food Issues Ford and Salisbury (lectures 15 - 18)
Topic One: Food Health and Nutrition - and , Steenkamp, n and Stevens
Topic Two: Food Safety - Tribe
Topic Three: Food Economy and Politics - Freebairn
Topic Four: Food Biotechnology and Food Issues - d and Salisbury
Forums
There will be a total of FIVE Forums. Forums are presented by invited speakers.
Forums will be recorded, however recordings will only be released in SWAT vac.You are
strongly encouraged to attend all Forums. Forums will be examined in the final exam.
Assessment?
1 Mid-Semester Test worth 2 % on 29th August 2016 to be held during the lecture time slot in
Wilson Hall or University Hall
One 1000-word essay worth 1 % due nd 2016 by 10 pm
Two hours Exam worth 60% during the examination period. You will be permitted to bring
into the final exam one A4 page of notes. You can have notes on both sides.
The Mid-Semester test will be on Topic One: Health and Nutrition and Topic Two: Food Safety.
You will be examined on lectures 1 to 10, tutorials 1 to 4 and tutorial readings required for
tutorials 1 to 4.
The test will comprise of 20 multiple choice questions and four short-answer questions.
Past tests will be provided on LMS.
You are allowed to see the turn-it-in report and you are allowed to submit to turn-it-in multiple
times before the due date. Please be aware of our late penalties and plagiarism policies.
Final Exam (60%)
The final exam will be of two hours duration plus an extra 15 minutes for reading. This subject is
a Melbourne University Breadth subject which means that students are from a wide range of
backgrounds with varying skills and interests. For examination purposes students will be
expected to understand the important principles but not necessarily have mastered in depth the
detailed and complex underlying disciplinary knowledge.
The final exam will consist of 3 parts and will cover lectures, tutorials, forums and essential
readings. Part A is worth 30%, Part B is worth 50%, Part C is worth 20%
Plagiarism
The University takes plagiarism very seriously. It is recommended you visit the university
Academic Honest and Plagiarism website on http://academichonesty.unimelb.edu.au/index.html
Extensions
If you dont think you can hand in your assessment on time, you can apply directly to your
lecturer or tutor for an extension (make sure you speak to them before the due date). In general
extensions are permitted for unforeseen factors affecting your ability to complete a task on time.
The lecturer or tutor may advise you to also apply for special consideration depending on the
circumstances. Special consideration is available when illness or other causes:
significantly affected your university studies at any time during the semester;
prevented you from preparing for all or part of a component of assessment;
affected you during the performance of a component of assessment.
See the Universitys submission guidelines for Special Consideration. You can follow links from
this page to print out the HCAP or Statutory Declaration form.
You can submit your special consideration application through your Student Portal (Admin Tab
Paragraphs
Introduction: introductory paragraph should include a statement something like: In this
essay it will be shown . and then clearly state your core question/hypothesis.
Major point in defense of your argument
Secondary point in defense of your argument
Third and/or tangential point in defense of your argument
Conclusion: concluding paragraph should include a statement something like; In this
essay it has been shown., and then clearly state how you have answered the core
question/hypothesis.
Finally, spell check, then edit, then read your essay aloud, then re-edit.
Sources of Information
Recognize that you have many sources of information available to you. Lectures and tutorials will
provide you with some general knowledge and some specific information. You are expected to
conduct research to find the specific information you need, using the library and electronic sources
of journals (use the internet with discretion) Library websites are a good place to start. The
Language and Learning Skills Unit has a number of web- links to information that will help you
with hints on general aspects of writing and research.
For more on Writing reports and essay go to http://unilearning.uow.edu.au/main.html For Rules
and advice from the University Library go to http://www.lib.unimelb.edu.au/elib/techniq.html
For Search on-line go to http://www.lib.unimelb.edu.au/find.html
For Research skills go to http://www.lib.unimelb.edu.au/sg/tools.html
Referencing
Use recent information from a range of sources to show the breadth of your research Always use
credible publications (e.g. journal papers, books, DAFTA approved publications) to strengthen
your academic argument
Keep web-based material/ references to a minimum. (Not allowed to use Wikipedia as a reference
source because it is a blog site).
The source of photos, figures and graphs must be included in your reference list.
Choose a reference style and be consistent Harvard and APA styles are recommended
Always acknowledge the source of your information. This is done by including an in-text citation,
which gives the basic publication details (see Table 1). Secondly, you must include this citation in
your reference list (see Table 2).
Use of Quotes
Direct quotes should be used sparingly (5% of an assignment) if at all. It is best to summarize
and/or paraphrase information obtained from references, with proper in-text citation and
referencing. Failure to acknowledge direct quotes is plagiarism even if a reference is cited. Short
quotes (less than 30 words) should be set within the text, italicized and enclosed by single
quotation marks. Large quotes should be indented, italicized and placed in single quotation
marks. In-text citations with direct quotes should always indicate the page number(s) from which
the quote was taken.
Personal communication is a form of citation that should be avoided. The abbreviation pers.
comm. can be used in the in-text citation. The entry in the reference list must include the full
name, year, position title if known, exact date (day and month), and contact address (mail, email,
phone or fax) of the person providing the information. It is usually unacceptable to have more than
a few of these references.
Table 2: References
Hard copy books One Elder, BJ 1995, The magic of Australia, Beaut Books,
author Sydney.
Hard copy books - Two Yeric, J & Todd, J 1989, Public opinion: the visible politics,
authors Peacock Publishers, Chicago.
Hard copy books - Three Yeric, J, Todd, J & Muller, P 1999, Political perspectives,
authors and 2nd edition 2nd edn, Peacock Publishers, Chicago.
Edited book Danaher, P (ed.) 1998, Beyond the Ferris wheel, CQU Press,
Rockhampton.
Chapter in edited book Byrne, J 1995, Disabilities in tertiary education, in L
Rowan & J McNamee (eds), Voices of a margin, CQU
Press, Rockhampton.
Journal article Peterson, J & Schmidt, A 1999, Widening the horizons for
One or more authors secondary schools, Journal of Secondary Education, vol. 3,
no.8, pp.89-106
Conference proceedings Callaghan, S & Martin, B 2004, Igniting concern about
from a website refugee injustice, paper presented at the Education and
Social Action Conference, 6-8 December 2004,
http://www.uow.edu.au/arts/sts/bmartin/pubs/04esa.html
Lecturers
Professor Mohan Singh is subject coordinator. Professor Singh teaches in Molecular
Biology, Breeding and Biotechnology, Advanced Plant Breeding, Bioinformatics project
in Genome Science. His research interests are: Molecular biology of plant reproduction;
Transcriptome of flowering plant male gametes; Transcriptional regulation of male
gamete expressed genes; Molecular determinants of stem cells in legume meristems;
Genetic engineering of male sterility in crop plants; Genetic modification of Grass pollen
allergens for developing allergy vaccines; DNA repair genes in plants; and Expression
systems for Biopharmaceuticals.
Seona Candy is a research fellow with the Victorian Eco-innovation Lab at the University
of Melbourne with a particular interest in the integration of food and urban systems to
improve the resilience of cities. Her recent research work used scenario modelling to link
land and resource use with food consumption at a national level on an ARC Linkage
project, and at a city level with the Foodprint Melbourne Project. Dr Candy has a
background in engineering, renewable energy and international development. Her
PhD research focused on the use of appropriate technology to improve food security and
nutrition in remote mountain regions of Nepal. She has also worked on food security and
energy related projects with communities in Australia, Cambodia and East Timor.
Julia Steenkamp is a dietitian and food scientist. Julia has worked as a clinical dietitian
in the hospital setting as well as in private practice. During this time Julia managed
patients with obesity, diabetes, eating disorders, allergies and food intolerance. Julia has
also worked as a nutritionist for one of the largest food companies in the world. Over the
last six years, Julia's interest has been on the methods of securing safe and sufficient food
for the world and she recently completed a Masters in the area of food security and
sustainable rural development.
Dianne Stevens is a Lecturer in Nutrition and Dietetics and Community Domain Leader
in the Masters of Science in Dietetics for the College of Health & Biomedicine. She has
worked in clinical and community dietetics in Australia. Her particular interest is in
maternal, newborn and child health and nutrition in resource poor settings. Dianne has
gained considerable experience in managing programs in nutrition, food security and
health in international settings both in development and in emergencies. She has
experience in working with non-government organizations, the United Nations and with
donors.
Associate Professor David Tribe, is a senior lecturer in Department of Microbiology
and Immunology, University of Melbourne. Dr Tribe has expertise in genetics,
biochemistry, biotechnology, microbiology and food science.
Professor John Freebairn holds the Ritchie Chair in the Department of Economics at the
University of Melbourne. He received his tertiary education in agricultural economics
from the University of New England and the University of California, Davis. He has held
appointments with the NSW Department of Agriculture, ANU, La Trobe University, the
Business Council of Australia and Monash University before joining Melbourne in 1997.
John is an applied micro economist and policy analyst with current research interests in
taxation reform and environmental economics.
Dr John Ford is a marine and fisheries ecologist with a research interest on the coastal
habitat restoration and the sustainability of Australias fisheries. He combines expertise in
experimental and observational field ecology with a passion for our sustainable use of
marine resources. Dr Ford has spent the last eight years at The University of Melbourne
researching fish ecology, conducting over 500 scientific dives and has become expert on
the underwater world around Melbourne and the Victorian coastline. John is known
locally and internationally as an expert on seafood sustainability and has worked as an
assessor for the Marine Stewardship Council and the Sustainable Australia Seafood
Assessment Program. Current research projects include working with commercial fishers
to address seagrass decline in Corner Inlet, and restoring the lost shellfish reefs of Port
Phillip Bay. Dr Ford is also a regular presenter on RRR radios Sunday marine show
Radio Marinara and is a contributor to the goodfishbadfish.com.au website.
S U STA I N A B I L I T Y
CAN WE FEED
T
THE WORLD
S
AND SUSTAIN
E
F
THE PLANET? A five-step global plan
F
environmental damage
By Jonathan A. Foley
R
IGHT NOW ABOUT ONE BILLION PEOPLE SUFFER FROM CHRONIC HUNGER. THE WORLDS
farmers grow enough food to feed them, but it is not properly distributed and,
even if it were, many cannot afford it, because prices are escalating. But anoth-
er challenge looms. By 2050 the worlds population will increase by two billion
or three billion, which will likely double the demand for food, according to sev-
eral studies. Demand will also rise because many more people will have higher
incomes, which means they will eat more, especially meat. Increasing use of
cropland for biofuels will put additional demands on our farms. So even if we solve todays
problems of poverty and accessa daunting taskwe will also have to produce twice as much
to guarantee adequate supply worldwide. And thats not all.
84
By clearing tropical forests, farming tains, tundra, ice, cities, parks and other feeding crops to animals reduces the
marginal lands, and intensifying industri- unsuitable growing areas. The few re- worlds potential food supply. Grain-fed
al farming in sensitive landscapes and maining frontiers are mainly in tropical cattle operations typically use at least 100
watersheds, humankind has made agri- forests and savannas, which are vital to kilograms of grain to make one kilogram
culture the planets dominant environ- the stability of the globe, especially as of edible, boneless beef protein [see The
mental threat. Agriculture already con- stores of carbon and biodiversity. Expand- Greenhouse Hamburger, on page 108].
sumes a large percentage of the earths ing into those areas is not a good idea, yet Chicken and pork are more efficient, and
land surface and is destroying habitat, five million to 10 million hectares of crop- grass-fed beef converts nonfood material
using up freshwater, polluting rivers and land have been created in each of the past into protein. Overall, grain-fed meat pro-
oceans, and emitting greenhouse gases 20 years, with a significant portion of that duction systems are a drain on the global
more extensively than almost any other land conversion happening in the tropics. food supply.
human activity. To guarantee the globes These additions enlarged the net area of Another deterrent to growing more
long-term health, we must dramatically cultivated land by only 3 percent, howev- food is damage to the environment, which
reduce agricultures adverse impacts. er, because of farmland losses caused by is already extensive. Only our use of ener-
The worlds food system faces three urban development and other forces, par- gy, with its profound impacts on climate
incredible, interwoven challenges. It must ticularly in temperate zones. and ocean acidification, rivals the sheer
T guarantee that all seven billion people Improving yield also sounds enticing. magnitude of agricultures environmental
H alive today are adequately fed; it must footprint. Our research team has estimat-
double food production in the next 40 ed that by 2010 agriculture had already
E
years; and it must become truly environ- TO GUARANTEE cleared or radically transformed 70 per-
S
mentally sustainableall at the same time. THE GLOBES cent of the worlds prehistoric grasslands,
Could these simultaneous goals pos- 50 percent of the savannas, 45 percent of
C
sibly be met? An international team of
LONG-TERM HEALTH, the temperate deciduous forests and 25
I experts, which I coordinated, has settled WE MUST percent of the tropical forests. Since the
E on five steps that, if pursued together, DRAMATICALLY last ice age, nothing has disrupted ecosys-
could increase by more than 100 percent tems more. Agricultures physical foot-
N
the food available for human consump- REDUCE print is nearly 60 times that of the worlds
C
tion globally while substantially reducing AGRICULTURES pavements and buildings.
E greenhouse gas emissions, biodiversity Freshwater is another casualty. Hu-
losses, water use and water pollution.
ADVERSE IMPACTS. mans use an astounding 4,000 cubic kilo-
Tackling the triple challenge will be one of meters of water per year, mostly with-
the most important tests humanity has Yet our research team found that average drawn from rivers and aquifers. Irrigation
O
ever faced. It is fair to say that our response global crop yield increased by just 20 per- accounts for 70 percent of the draw. If we
F will determine the fate of our civilization. cent from 1985 to 2005far less than had count only consumptive water usewater
been reported. Cereals yield has been ris- that is used and not returned to the water-
F BUMPING UP AGAINST BARRIERS ing at less than 2 percent a year since 2000 shedirrigation climbs to 80 or 90 per-
O AT FIRST BLUSH, the way to feed more peo- and yields of pulses (beans, lentils) and cent of the total. As a result, many large
ple seems clear: grow more food, by root crops by less than 1 percentrates rivers, such as the Colorado, have dimin-
O
expanding farmland and improving yield that are nowhere near enough to double ished flows, some have dried up altogeth-
D (crops harvested per hectare). Unfortu- food production by midcentury. er, and many places have rapidly declining
nately, the world is running into signifi- Feeding more people would be easier water tables, including regions of the U.S.
cant barriers on both counts. if all the food we grew went into human and India.
Society already uses about 37 percent hands. But only 60 percent of the worlds Water is not only disappearing, it is be-
of the earths land surface, not counting crops are meant for people: mostly grains, ing contaminated. Fertilizers, herbicides
Greenland or Antarctica, for farms or pas- followed by pulses, oil plants, vegetables and pesticides are being spread at incredi-
tures. Agriculture is by far the biggest and fruits. Another 35 percent is used for ble levels and are found in nearly every
human use of land on the planet; nothing animal feed, and the final 5 percent goes ecosystem. The flows of phosphorus and
else comes close. And most of that 37 per- to biofuels and other industrial products. nitrogen through the environment have
cent covers the best farmland. Much of the Meat is the biggest issue here. Even with more than doubled since 1960, causing
remainder is covered by deserts, moun- the most efficient meat and dairy systems, widespread water pollution and enor-
IN BRIEF
The world must solve three food prob- Five solutions, pursued together, can ciency of water and fertilizer use world- how well each one delivers nutrition
lems simultaneously: end hunger, dou- achieve these goals: stop agriculture wide, reduce per capita meat con- and food security and limits environ-
ble food production by 2050, and do from consuming more tropical land, sumption, and reduce waste in food mental and social costs would help the
both while drastically reducing agricul- boost the productivity of farms that production and distribution. public choose products that push agri-
tures damage to the environment. have the lowest yields, raise the effi- A system for certifying foods based on culture in a more sustainable direction.
86
L AY O F T H E L A N D
Type of None
S
Agriculture
C
The world could grow much more food if the productivity of the poorest E
farms were raised toward the maximum possible, given climate and soil
Yield of Maize Farmland conditions. For example, the yield for maize (shown) could rise signifi- N
cantly across parts of Mexico, West Africa and eastern Europe if seeds, C
Far from maximum Close to maximum irrigation, fertilizer and markets were improved.
E
mous hypoxic dead zones at the mouths expand or the mounting environmental Slowing deforestation would dramati-
O
of many of the worlds major rivers. Ironi- harm it imposes. Previous approaches to cally reduce environmental damage while
cally, fertilizer runoff from farmlandin solving food issues were often at odds with imposing only minor constraints on glob- F
the name of growing more foodcompro- environmental imperatives. We could al food production. The resulting dip in
mises another crucial source of nutrition: boost food production by clearing more farm capacity could be offset by reducing F
coastal fishing grounds. Fertilizer certain- land or using more water and chemicals the displacement of more productive O
ly has been a key ingredient of the green but only at a cost to forests, streams and croplands by urbanization, degradation
O
revolution that has helped feed the world, wetlands. Or we could restore ecosystems and abandonment.
but when nearly half the fertilizer we ap- by taking farmland out of cultivation but D
Many proposals have been made to
ply runs off rather than nourishes crops, only by reducing food production. This reduce deforestation. One of the most
we clearly can do better. either-or approach is no longer accept- promising has been the Reducing Emis-
Farming also accounts for 10 to 12 per- able. We need truly integrated solutions. sions from Deforestation and Degrada-
cent of the warming effects of greenhouse After months of research and deliber- tion (REDD) mechanism. Under REDD,
gases released by human activitya con- ationbased on analysis of newly gener- rich nations pay tropical nations to pro-
COURTESY OF JAMES GERBER Institute on the Environment, University of Minnesota (maps)
tribution equal to that of all the road vehi- ated global agricultural and environmen- tect their rain forests, in exchange for car-
cles on the planet. Most of the direct emis- tal dataour international team settled bon credits. Other mechanisms include
sions from farming come from methane on a five-point plan that deals with food developing certification standards for ag-
produced by animals and rice paddies and environmental challenges together. ricultural products so that supply chains
and from nitrous oxide released by over- can be assured that crops were not grown
fertilized soils. Add in the effects of trop- Stop expanding agricultures foot- on lands created by deforestation. Also,
ical deforestation and other land clear- print. Our first recommendation is to better biofuel policyone that relies on
ing, and agricultures share of global slow and ultimately stop the expansion of nonfood crops such as switchgrass in-
emissions rises to 24 percent of the total. agriculture, particularly into tropical for- stead of food cropscould make vital
ests and savannas. The demise of these farmland newly available.
FIVE SOLUTIONS ecosystems has far-reaching impacts on
MODERN AGRICULTURE has been an incredi- the environment, especially through lost Close the worlds yield gaps. To double
bly positive force in the world, but we can biodiversity and increased carbon dioxide global food production without expanding
no longer ignore its dwindling ability to emissions (from clearing land). agricultures footprint, we must signifi-
87
Reduce food waste. A final, obvious but house gas emissions, biodiversity losses, ing approaches. H
often neglected recommendation is to re- water use and water pollution. Imagine the possibilities: sustainable E
duce waste in the food system. Roughly It is important to emphasize that all citrus and coffee from the tropics, con-
30 percent of the food produced on the five points (and perhaps more) must be nected to sustainable cereals from the S
planet is discarded, lost, spoiled or con- pursued simultaneously. No single strate- temperate zone, supplemented by locally
C
sumed by pests. gy is sufficient to solve all our problems. grown greens and root vegetables, all
In rich countries, much of the waste Think silver buckshot, not a silver bullet. grown under transparent, performance- I
takes place at the consumer end of the We have tremendous successes from the based standards. Use your smartphone E
system, in restaurants and trash cans. green revolution and industrial-scale ag- and the latest sustainable food app, and N
Simple changes in our daily consumption riculture to build on, along with innova- you will learn where your food came
C
patternsreducing oversize portions, the tions in organic farming and local food from, who grew it, how it was grown, and
food thrown in the garbage, and the num- systems. Lets take the best ideas and in- how it ranks against various social, nutri- E
ber of takeout and restaurant meals corporate them into a new approacha tional and environmental criteria. And
could significantly trim losses, as well as sustainable food system that focuses on when you find food that works, you can
our expanding waistlines. In poorer coun- nutritional, social and environmental per- tweet about it to your social network of
O
tries, the losses are similar in size but formance, to bring responsible food pro- farmers and foodies.
occur at the producer end, in the form of duction to scale. The principles and practices of our dif- F
failed crops, stockpiles ruined by pests or We can configure this next-generation ferent agricultural systemsfrom large-
spoilage, or food that is never delivered system as a network of local agricultural scale commercial to local and organic F
because of bad infrastructure and mar- systems that are sensitive to nearby cli- provide the foundation for grappling with O
kets. Improved storage, refrigeration and mate, water resources, ecosystems and the worlds food security and environmen-
O
distribution systems can cut waste appre- culture and that are connected through tal needs. Feeding nine billion people in a
ciably. Moreover, better market tools can efficient means of global trade and trans- D
truly sustainable way will be one of the
connect people who have crops to those port. Such a system could be resilient greatest challenges our civilization has
who need them, such as cell-phone sys- and pay farmers a living wage. had to confront. It will require the imagi-
tems in Africa that link suppliers, traders One device that would help foster this nation, determination and hard work of
and purchasers. new food system would be the equivalent countless people from all over the world.
Although completely eliminating waste of the Leadership in Energy and Envi- There is no time to lose.
from farm to fork is not realistic, even ronmental Design program now in place
small steps would be extremely benefi- for constructing new commercial build- Jonathan A. Foley is executive director of the
cial. Targeted effortsespecially reduc- ings sustainably. This LEED program California Academy of Sciences in San Francisco.
ing waste of the most resource-intensive awards increasingly higher levels of cer- From 2008 to 2014, he led the Institute on the
foods such as meat and dairycould tification based on points that are accu- Environment at the University of Minnesota.
make a big difference.
MOVING TOWARD M O R E TO E X P L O R E
A NETWORKED FOOD SYSTEM Global Consequences of Land Use. Jonathan A. Foley et al. in Science, Vol. 309, pages 570574; July 22, 2005.
IN PRINCIPLE, our five-point strategy can Enough: Why the Worlds Poorest Starve in an Age of Plenty. Roger Thurow and Scott Kilman. Public Affairs, 2010.
address many food security and environ- Food Security: The Challenge of Feeding 9 Billion People. H. Charles J. Godfray et al. in Science, Vol. 327, pages 812
818; February 12, 2010.
mental challenges. If they were done
Solutions for a Cultivated Planet. Jonathan A. Foley et al. in Nature, Vol. 478, pages 337342; October 20, 2011.
together, the steps could increase the
worlds food availability by 100 to 180 per- s c i e n t i f i c a m e r i c a n . c o m /m a g a z i n e /s a
89
ABSTRACT ward carbon and calories. However, nitrogenor protein for that
Food sustainability and food security are increasingly in the spotlight matteris an important but often disregarded macronutrient
and increasingly intertwined. According to some projections we will with a pivotal role in biodiversity loss, climate change, and
need to nearly double food production in the next 4 decades. This article human health risks.
argues that protein production and consumption are pivotal to sustain- Therefore, the current article argues that food security, food
ability, because anthropogenic contributions to the nitrogen cycle are sustainability, and nutrition are increasingly intertwined and
100200% compared with a contribution of 12% to the carbon cycle should be addressed in an integrated way, by identifying trade-
by mineral fuel combustion, with biodiversity as the main casualty. offs and setting priorities. Subsequently, it is argued that a re-
Because 1 kg animal protein requires w6 kg plant protein, its large- duction in the consumption of animal products is a prominent
scale production by means of factory farming is a major driver of option toward more sustainable food production. The role of
Am J Clin Nutr 2014;100(suppl):483S9S. Printed in USA. ! 2014 American Society for Nutrition 483S
484S AIKING
which should be understood as a challenge to preserve the After the 2008 price increases, Lawrence Haddad, one of the
adaptability and resilience of the natural (biotic and abiotic) authors of the Foresight report (38), warned in an interview, The
systems that form the basis of economic and social development. last 3 to 4 years have seen alarming spikes in hunger. The price
rises in 20078 were actually quite modest in a historical context
but it led to 100 million more people going hungry. Bigger price
FOOD SUSTAINABILITY rises could wipe out the development gains of the last 20 years
Many definitions of sustainability are known and are generally and promote violent conflict and migration (39). In 2012, after
based on its 3 pillars of ecology, economy, and society. The same droughts that threatened US corn harvests, similar warnings
holds with respect to food sustainability (14). However, distinct were voiced by the FAO (40).
definitions of food sustainability may, or may not, address issues In fact, the effects of the seemingly inevitable price increases
such as human health, equity, and animal welfare, indicating are not restricted to developing countries (41). Browning et al
that the issue is fraught with ethical considerations. In fact, it is (42) summarized the impacts in the United Kingdom as follows:
often unclear whether food sustainability definitions refer to There are over 4 million people in the UK currently living in
production or consumption. The gap between production and food poverty and The lowest income households tend to be the
consumption can be 30%, which may be in the form of waste (15) hardest hit by changes in food affordability. The relative af-
or other uses of food such as pet food. As an illustration of the fordability of foodmeasured by the share of total consumer
latter, De Silva and Turchini (16) conservatively estimated that spending that goes on food and nonalcoholic drink for house-
2.48 million tons of wild-caught fish is used directly by the cat hold suppliesis therefore a key indicator of household food
food industry, raising the ethical point that the use of a limited security.
biological resource to feed pets competes with human food
from feed crops into animal protein for human consumption may would benefit health, environment, and animal welfare. Similar
be economically feasible, but it is inherently resource-inefficient, recommendations could be phrased for the United States.
which makes intensive livestock production responsible for Of course, a successful diet transition requires local differ-
a disproportionate share of environmental pressure (6971). On ences to be addressed. For example, the global population has
average, 6 kg of plant protein is required to yield 1 kg of meat been projected to increase by 50% between 2000 and 2050, and
protein (17, 72). Consequently, only 15% of protein and energy both meat and dairy demand have been projected to double (70).
provided by feed crops will be consumed by humans indirectly Most of the projected population increase will take place in
and 85% of these crops are wasted (and, incidentally, 85% of Africa and India, although meat demand will increase primarily
fertilizer inputs to grow them). In 2000, for example, 942 and in China and South America and dairy demand in India. Such
617 million tons of grain were used for food and feed, re- geographic distinctions should be taken into careful consider-
spectively (36). Of the latter, .500 million tons constitute ation.
a tremendous loss of resources. Maybe even more important,
this huge amount of potential food is turned into massive
emissions of reactive nitrogen (eg, ammonia from manure), STAKEHOLDERS AND POLICY
which pollutes the terrestrial and aquatic environment and re- Reducing consumption in general or that of livestock products
sults in biodiversity loss. in particular is easier said than done. In essence, there is no
Both resource depletion and pollution by livestock production problem owner. Although lip service is paid to less is more,
were treated in a comprehensive FAO report (70). Moreover, the bottom line is that to reduce consumption (or to be told to do
intensive livestock production was shown to play a crucial role in so) is in sharp conflict with the desires of the average consumer,
all 3 of the planetary boundaries that have already been industrialist, and politician.
overstepped by humanity (ie, biodiversity loss, nitrogen cycle As indicated above, some countries have integrated policies
Seona Candy
Research Fellow: Sustainable Food Systems, University of Melbourne
Jennifer Sheridan
Researcher in sustainable food systems, University of Melbourne
Rachel Carey
Research Fellow, University of Melbourne
You know that feeling when you open the fridge and are met with something on the nose.
We all know what food waste looks and smells like.
But food waste stinks in more ways than one. It is expensive, costing the average household
over A$2,200 a year, and it undermines the resilience and sustainability of our food supply.
A new report from our Foodprint Melbourne Project has estimated the amount of food that is
wasted in feeding Melbourne. We found that feeding Melbourne generates more than 900,000
tonnes of edible food waste every year, or over 200 kg per person.
http://theconversation.com/melbourne-wastes-200-kg-of-food-per-person-a-year-its-tim... 8/07/2016
Melbourne wastes 200 kg of food per person a year: it's time to get serious Page 2 of 3
Undermining sustainability
Growing this wasted food uses 180 gigalitres of water each year, or 113 litres per person per
day. This is equivalent to running your shower for an extra 10 minutes a day.
This wasted food also uses around 3.6 million hectares of land around 41 ha per person, or
more than 20 times the area of the Melbourne Cricket Ground.
And this wasted food is responsible for around 2.5 million tonnes of greenhouse gas
emissions, 60% of which is generated by food waste rotting in landfill, and the rest in
producing the wasted food.
This uneaten food is not only a source of unnecessary greenhouse gas emissions. It
represents a waste of natural resources that are in increasingly limited supply.
Australia is a water-scarce region that is likely to become drier due to climate change, while
only 6% of Australias land is suitable for growing crops.
With the associated waste of natural resources, high levels of food waste add to the challenge
of producing sufficient food to feed a growing population.
There are many ways to reduce food waste at home. These include making meal plans,
sharing leftover food with friends or neighbours, checking the fridge before going shopping
and storing food correctly.
The Cloud-Freezer app can help you to keep track of whats in your freezer and fridge. Worm
farms, bokashi bins and other forms of composting are also great ways to divert food waste
from landfill.
While we can all take steps to reduce food waste at home, we need to look at the bigger
picture. Our research shows that more than 60% of food waste is generated before food
reaches your fridge or freezer.
Strict standards defining the shape, size and colour of fresh fruit and vegetables in
supermarkets can mean that a significant proportion of a crop never leaves the farm.
Low prices for second-grade produce can make it financially unviable for farmers to pick, pack
and ship imperfect produce. Pressure to keep supermarket shelves full for appearances sake,
losses during food processing and storage problems also lead to food being wasted.
http://theconversation.com/melbourne-wastes-200-kg-of-food-per-person-a-year-its-tim... 8/07/2016
Melbourne wastes 200 kg of food per person a year: it's time to get serious Page 3 of 3
Initiatives that aim to make more imperfect fruit and vegetables available, such as
Woolworths Odd Bunch campaign, go some way to reducing this problem, but more needs to
be done.
Our research estimates that if food waste was halved across the food supply chain,
Melbourne could save 1.8 million hectares of land, 90 million litres of water and avoid 1.3
million tonnes of greenhouse gases each year.
In recognition of the significant challenge that food waste represents to sustainable food
systems, the new Sustainable Development Goals set a target to halve the global food waste
per person that is generated by retailers and consumers by 2030.
The United States government has also set a national target to reduce food waste by 50% by
2030. It has established a cross-sector partnership of stakeholders across the food system to
tackle the problem.
The UK government has been an early mover in taking action to tackle food waste. In 2007, it
launched the WRAP Love Food Hate Waste program aimed at reducing food waste. An
evaluation in 2012 showed that avoidable waste of food and drink (that could have been
eaten) had fallen by 21% in five years following the launch of the program.
Most of this reduction has been in household food waste. The WRAP program is now working
with the food industry to reduce waste in other sectors. The successful UK Love Food Hate
Waste program aimed at reducing household food waste has been taken up by state
governments in Victoria and New South Wales.
Australia is developing a national food waste strategy the Food Waste 2025 Strategy and
stakeholders from across the food supply chain meet this month to discuss how to reduce
food waste.
Australia should follow suit in setting a target to halve food waste across the food supply chain
to put Australias food system on a more sustainable footing.
*Correction: This figure has been updated. It previously incorrectly stated that Melbournes
food waste is enough to feed 2,000 people per year.
Food
Waste
Farming
Greenhouse gases
Melbourne
Carbon emissions
food waste
Greenhouse Gas
ugly food
eating
Tweet163
Share334
Get newsletter
http://theconversation.com/melbourne-wastes-200-kg-of-food-per-person-a-year-its-tim... 8/07/2016
Running on empty: Australia's risky approach to oil supplies
Vlado Vivoda
Research Fellow, Griffith Asia Institute, Griffith University
Australia only keeps an estimated 23 days' worth of fuel in the country. Bidgee/Wikimedia Commons
It might sound unlikely, but Australias fuel gauge is worryingly low. Were one of the worlds
top energy exporters, but our stocks of liquid fuels such as the oil on which almost the whole
transport sector depends are far from healthy.
Australias worsening liquid fuel security problem was detailed this week in the latest
instalment of a report prepared for the National Roads and Motorists' Association.
A year after the issue was first raised, there has been no action to address the growing risks,
both logistical and financial, of keeping so little fuel in the tank.
http://theconversation.com/running-on-empty-australias-risky-approach-to-oil-supplies-... 8/07/2016
Running on empty: Australia's risky approach to oil supplies
Lack of refinement
According to the report, refinery closures coupled with growing demand for petroleum have
increased Australias dependency on imported refined petroleum products. As a result, oil-
dependent sectors such as transport, mining and agriculture are vulnerable to supply-chain
interruptions or future oil-price spikes.
The situation is made worse by the fact that Australia is consistently the only International
Energy Agency (IEA) member state that fails to maintain the mandated stockpile of 90 days'
worth of net oil imports. Australias stockpiles stood at 57 days as of November 2013. The
NRMA report estimates Australias in-country stockpile, which excludes shipments en route to
Australia, at only 23 days.
With growing demand for liquid fuels and a continued hands-off approach from government,
by 2030 Australia is on track to find itself with no refining capacity, less than 20 days' worth of
liquid fuel stocks, and entirely at the mercy of the international oil market.
The reports author, retired Air Vice-Marshall John Blackburn, has called on the government to
step in. This years National Energy Security Assessment and Energy White Paper both offer
opportunities for public debate. The report also suggests making fuel stocks a matter of
national security risk analysis.
Relying on imported oil isnt just risky its expensive too. In 2012, Australias net oil imports
were 561,000 barrels per day. At an average of US$109.08 a barrel for that year, the bill was
US$22.3 billion, or 1.7% of Australias gross domestic product.
At the same time, there are fears of a slowdown in Australias mineral export revenues as
Chinas economy loses pace.
Four decades later the picture is very different, although people dont necessarily know it.
Blinded by positive headlines about energy exports, only 1% of Australians consider the
energy crisis, petrol prices and fossil-fuel supplies to be our most pressing issue.
http://theconversation.com/running-on-empty-australias-risky-approach-to-oil-supplies-... 8/07/2016
Running on empty: Australia's risky approach to oil supplies
Perhaps Australias good fortune during the 1970s has helped to entrench a free-market
approach to the oil market, while other nations that were more badly burned by OPECs price-
gouging have taken stronger steps towards oil independence.
The Hawke governments free-market attitude continued through the Howard years, as
evidenced in Australias first ever Energy White Paper. Released in 2004, it openly
acknowledged Australias oil deficit. But instead of raising the issue as a significant energy
security threat, it sought to reassure by pointing to surpluses in gas, uranium and coal. The
policy implication was clear: as long as Australia remained a net energy exporter, its liquid fuel
balance was nothing to worry about.
None of this changed with Labors return to power. The 2012 Energy White Paper also struck
a positive tone, based on Australias position as an energy producer and net exporter, which
implied overall energy security. And now the Abbott government, in preparing this years
Energy White Paper, seems to consider the international oil market so resilient that the supply
of crude oil and refined petroleum to Australia is guaranteed. However, it is worth noting that
its recent issues paper includes a reference to asking whether Australia needs to increase its
stockpiles to meet international obligations.
Real impacts
Energy security issues may impact ordinary Australians directly through high consumer
prices, fuel shortages and the environmental impacts associated with energy consumption.
Given the importance of energy to the Australian economy, it is vital that the government
reevaluates its approach to energy policy and engages in a systematic risk analysis regarding
both Australias growing liquid fuel import dependency and potential threats to demand for its
energy exports.
While Australia remains faithful to the basic principles of free-market economics, if they are
wrong the consequences could be dire. A bipartisan hands-off approach and general public
complacency regarding Australias energy future may not serve the nations best interest in
the long term.
In terms of practical solutions to our growing liquid fuels import dependency, meeting the
IEAs mandated stockpile levels would be a first step in the right direction. The government
should also consider measures aimed at curbing the demand for liquid fuels in Australia.
These might include higher fuel-efficiency standards, improving public transport infrastructure,
promoting carpooling, or introducing congestion charges in the CBDs of major cities.
Energy security
Transport
OPEC
Exports
Petroleum
Oil refining
International Energy Agency
Tweet32
Share88
Get newsletter
http://theconversation.com/running-on-empty-australias-risky-approach-to-oil-supplies-... 8/07/2016
ANALYTIC ESSAYS
Published online ahead of print March 17, 2016 AJPH Sabat et al. Peer Reviewed Analytic Essays e1
ANALYTIC ESSAYS
e2 Analytic Essays Peer Reviewed Sabat et al. AJPH Published online ahead of print March 17, 2016
ANALYTIC ESSAYS
spread of infectious disease,4 re- now on the forefront of a new physical and social dimensions assessed, resulting in a compre-
duced food yield in some regions, expansion to incorporate envi- of the food system and its un- hensive assessment of each food
and increased threat to food ronmental considerations (Figure B, intended and undesired conse- item or dietary pattern in terms of
security.11,39 available as a supplement to the quences on the biosphere, the environmental and public
online version of this article at including public health impacts. health impacts.
http://www.ajph.org). The nu- Although this is relevant to all As depicted in Figure 2, there
trition of individuals and com- types of food systems, Western- is a direct feedback between
THREE DIMENSIONS munities can only be maintained ized industrial food systems natural resource use and envi-
OF NUTRITION within an environmentally sus- centrally controlled by govern- ronmental pollution as the latter
SCIENCES tainable context, which is cur- ments or multinational con- inuences the former and vice
Currently, human nutrition rently under serious threat.6 glomerates are the selected focus, versa. The types and quantities of
covers 2 main aspects: individuals Environmental nutrition seeks to given their growing dominance resources used inuence the type
and communities. Although address the sustainability of food and inuence on other food and amount of pollution created,
public health and commu- systems by integrating the envi- systems. which in turn compromises the
nity are not necessarily in- ronmental sciences with the nu- Through the ENM, the inputs quality of natural resources. The
terchangeable, the use of tritional sciences, addressing and outputs related to the whole required output of the food sys-
community here purposefully a range of issues from production life cycle are considered and tem (i.e., food) serves and in turn
allows differentiation of the practices to societal demands on quantied for individual food inuences societal demands and
existing and proposed aspects of a biospheric scale (see the box on items, which can then be used to composes the dietary patterns of
public health nutrition. At its core, page e4). The existing multidis- assess and compare whole diets. diverse populations. Diet-related
the discipline of nutrition is, and ciplinary nature of public health The life cycle of food includes the health issuesincluding obe-
will remain, a biological science nutrition lends itself to in- production, processing, trans- sity, chronic disease, and
centered on biochemistry, physi- corporating further disciplines. portation, storage, retail, and malnourishmentare a major
ology, and medicine. It is con- Given the intrinsic interrelation- consumption and disposal prac- consequence of the food system.
cerned with the interaction of ships between the environment tices employed. For each food Food consumption patterns or
food and other sources of nutri- and nutrition,26 the current public item, the ENM can produce consumer preferences can inu-
ents and with the physiological, health nutrition agenda must ex- a comparable list of inputs such as ence the extent of environmental
metabolic, and genomic systems pand to embrace environmental water, energy, land, and chem- degradation (as outlined in the
of the human body. By extension, nutrition as a new frontier. icals (including fertilizer and section The Evolution of Food
it addresses the effect of these as- pesticide), and of outputs, in- Systems). Policy can also sig-
pects on health and disease48 cluding GHG emissions, manure, nicantly inuence consumer
(see the box on page e4). Public and waste throughout the entire demandsfor example, by sub-
health nutrition emerged in life cycle. One of the aims of the sidizing the production of certain
recognition of the importance THE ENVIRONMENTAL ENM is to calculate a compre- food types more than others,
of community welfare, including NUTRITION MODEL hensive range of environmental which in turn articially lowers
social, economic, political, and Incorporating environmental outputs and impacts; therefore, retail price and increases con-
human aspects. Thus, nutrition nutrition into public health will for certain food items, additional sumption. In addition, policy and
extended its disciplinary scope to necessitate a comprehensive un- aspects must be considered. For governance are crucial factors for
incorporate such concerns. This derstanding of food systems and example, for livestock products, food sovereigntyfor example,
research expansion has had a tre- their impact on the biosphere. In respiration and manure can be food systems can become largely
mendous impact on public policy, an effort to clarify the interaction factored into the life cycle as- controlled by a small number
both nationally and internationally. between current food systems, sessment. Another important as- of corporations rather than by
Specically, the public health the environment, and public pect to take into account is the everyone involved, including
nutrition agenda incorporates health, we propose an environ- missed opportunity for carbon consumers.49
a range of social and multidisci- mental nutrition model (ENM). sequestration in relation to land The environmental impacts of
plinary sciencesincluding ep- This model, portrayed in use and crop type.42 An assess- the food system can in turn in-
idemiology, anthropology, and Figure 2, is intended to provide ment of the biodiversity impacts uence nutritional outcomes.
sociologyto address nutri- a useful didactic tool to explain, should also be included.25 For example, GHGs are an out-
tional deciencies, disease epi- understand, and ultimately con- Societal demands would be put of the food system that
demics, disease prevention, and tribute to the necessary modi- considered in the ENM, in- contribute to climate change,
food security (see the box on cations and changes to the cluding consumer preferences, which in turn affects food pro-
page e4). current food system to achieve policy, marketing, and social duction through reduced yields
As public health nutrition sustainability. The ENM is dynamics. The nutritional ade- and nutritional content.11 These
emerged from the science of a conceptual framework that quacy of each food item or di- impacts subsequently increase
nutrition for individuals, we are encompasses the multifaceted etary pattern would also be malnutrition and food insecurity.
Published online ahead of print March 17, 2016 AJPH Sabat et al. Peer Reviewed Analytic Essays e3
ANALYTIC ESSAYS
Scientic disciplines covered by each dimension Biological sciences Social and multidisciplinary sciences Environmental sciences
d Biochemical d Epidemiology d Physical
d Economics d Geography
d Sociology
Scope of Dimensions
Public policy
a
Environmental nutrition also covers disciplines covered by human and community nutrition.
Coupled with the global impacts largely controlled by 4 corpora- people per cropland hectare by biodiversity also has implications
of climate change which the food tions. The environmental out- shifting crop calories to direct for public health, including an
system signicantly contributes puts of concern include solid, human consumption.53 exacerbation of existing health
to,9 malnutrition and food in- liquid, and air pollution (from inequities through exposure
security in turn raise issues related animal manure, urine, blood, and to environmental hazards or
to environmental justice and hair; antibiotic residues; fossil fuel ENVIRONMENTAL through loss of livelihoods.49,55
intra- and intergenerational dis- combustion; nitrogen and phos- NUTRITION AGENDA Hence, biodiversity impacts must
tributive justice.10,16 phorus), and GHGs (primarily Expanding public health to also be a strong consideration in
As a specic example, to fully methane, nitrous oxide, and include environmental nutrition food production.24
demonstrate the complex range carbon dioxide). Other outputs will require dialogue and research To successfully feed a growing
of inputs and outputs of the food include nutrients for human in 4 overarching areas of the world population while pro-
system, we provide a case study of consumption and contributions global food system: food pro- tecting and remediating the en-
beef. In the United States, the to chronic disease52 and in- duction, food consumption, vironment, food consumption
production of 1 kilogram of beef fectious disease.5 Equity and nutrition policy, and the in- patterns must be considered and
requires the input of 10 kilograms food security issues are relevant, tegrated assessment of these areas. altered.14,56 The potential level
of grain feed,50 more than 8000 as producing livestock for So far, food production has of control that consumers have
liters of water, almost 8000 human consumption is a very attracted the most attention, over their purchases, particularly
kilojoules of energy, 150 grams of resource-inefcient process, with much of which is focused on in developed nations, implies
fertilizer, 7 grams of pesticides,38 about an 89% loss of biomass reducing the inputs required a level of elasticity in demand and
and 21 square meters of land.51 through the life cycle.50 Com- through technological means, hence a signicant potential to
Inputs on the demand side in- pared with producing 1 kilogram rather than replacing less efcient modify the environmental im-
clude marketing, cultural norms, of protein from kidney beans, the food crops with more efcient pacts incurred within the food
technology, and such policies as production of 1 kilogram of ones.13 Hence, it is essential to system and to alter personal
government subsidies, either as protein from beef requires ap- identify resource-intensive health outcomes. The level of
direct payments to farmers or proximately 18 times more land, commodities and those that consumer control depends on
through public nutritional assis- 10 times more water, 9 times could replace them, and also to many factors, including the
tance programs, both of which more fuel, 12 times more fertil- assess the relative health impli- amount of money they have
increase availability to con- izer, and 10 times more pesti- cations. Although biodiversity available and their access to al-
sumers. Governance issues can cide.51 In addition, the United loss ranks among the major ternative foods. The key ques-
also be an important consider- States uses 67% of total calorie drivers of ecosystem change, it tions in this area are how, where,
ation; in the United States, for production for livestock feed and remains a relatively neglected and when can the sustainability
example, beef processing is could feed almost 3 times as many area of investigation.54 Loss of and nutritional quality of dietary
e4 Analytic Essays Peer Reviewed Sabat et al. AJPH Published online ahead of print March 17, 2016
ANALYTIC ESSAYS
choices be simultaneously opti- have been explored to some sustainability of food systems of food delivery, which can be
mized? So far, there is a lack extent, have been shown to re- should be evaluated in terms of simple or complex. For example,
a comprehensive analysis to ad- duce dietary GHG emissions.57 whether they require techno- the life cycle of a potato involves
dress these issues.13 However, there remains much to logical change for the producer or growing, harvesting, washing,
The main area requiring at- explore, particularly in relation to behavioral change for the con- packing, transporting to a dis-
tention within policy research how societal demands can be sumer, and the type and level of tributor and then to a retailer,
relates to the integration of policy adjusted to embrace healthy and policy intervention required to purchasing at a store, transporting
options to address problems from sustainable foods, and how implement such measures. to a home, cooking, and dis-
a holistic perspective. For ex- measures such as a restructuring To conduct comprehensive posing of any waste. This is
ample, there is a need to assess of food subsidies could support assessments of food systems, the a fairly straightforward route
how food production can be such change.13 development of an analytical tool compared with that of a tomato
adjusted on a large scale to reduce Given the interrelationships, it is required. Such a tool would grown to produce ketchup. The
environmental impacts and pro- is essential to integrate food take into account the health, process is more complex, and
vide health-promoting foods production, consumption, and nutritional, and environmental even more so if the ketchup
while also accounting for the other relevant aspects in a way aspects of food choices and allow becomes one of several food
required changes in food con- that allows a comprehensive for the exploration of additional items in a ready mealthe route
sumption on the demand side. analysis of food systems. It is factors such as policy scenarios, becomes staggering.
Essentially, policies need to be important to identify foods that monetary costs for the consumer, Currently, the focus of life
developed in a way that ensures are both sustainable and healthy and equity and cultural consid- cycle assessments is on the GHGs
that all of the stages from food and to develop a food-labeling erations.12,58 One of the key from dietary patterns. This must
production to postconsumption scheme to communicate such components within such a tool is be expanded to incorporate
are coherent and cooperative information to consumers. De- an environmental life cycle as- a broad range of environmental
with a global-health-promoting, veloping a set of sustainability and sessment of foods across the entire aspects, including water use,
environmental-conservancy social indicators is likely to be food system. The growing body energy use, chemical use (pesti-
agenda. This could involve important in assessing the scale of research in the life cycle as- cide and fertilizer), land use, and
a mixture of regulatory and scal and speed of the transition. sessment of food offers a useful waste products (including solid,
measures.13 The latter, which Measures that increase the insight into the current system gaseous, and efuent) from each
aspect.12 The focus on GHGs has
partly arisen through a lack of
sufcient information on other
environmental aspectsfor ex-
Natural World ample, there is currently very
Human Societies
little analysis of the impacts of
global food wastage from an
Resources Demands environmental perspective.36
Sun Food System Consumer Hence, research efforts are also
Water preferences
Land Life Cycle Policy
required to develop the data in-
Energy Agriculture Processing Technology puts that would facilitate the in-
Chemicals Livestock Packaging Marketing
Fisheries Transportation Social dynamic clusion of such environmental
Storage Cultural norms
Retail
impacts. In addition, regionally
Waste specic analyses must be de-
veloped to increase accuracy and
applicability.12 Most life cycle
Pollution Food
Solid and fluid waste assessments currently focus on
Consumption Patterns of
Biological contaminants
Greenhouse gas emissions
Diverse Populations food production, so there is
a need to expand the analysis to
Undernourished Overfed include other stages of the food
Climate Malnutrition Obesity system.59
Change Low Regional Food Deficiency Chronic
Food Yields and/or Insecurity diseases diseases
Reduced
Nutritional Content
CONCLUSIONS
Paradoxically, although the
FIGURE 2The Environmental Nutrition Model, Including an Example of the Relationships Between
current food system is reliant on
Climate Change and Food Security
the natural world, it is also a major
Published online ahead of print March 17, 2016 AJPH Sabat et al. Peer Reviewed Analytic Essays e5
ANALYTIC ESSAYS
contributor to its degradation. 3. Foley JA, Ramankutty N, Brauman 18. Scarborough P, Appleby P, Mizdrak 33. Church N. Why our food is so de-
KA, et al. Solutions for a cultivated planet. A, et al. Dietary greenhouse gas emissions pendent on oil. Energy Bulletin, 2005.
Food production, processing, Available at: http://www.energybulletin.
Nature. 2011;478(7369):337342. of meat-eaters, sh-eaters, vegetarians and
transportation, storage, and vegans in the UK. Clim Change. 2014;125 net/node/5045. Accessed December 1,
4. McMichael AJ. Globalization, climate
packaging, coupled with an change, and human health. N Engl J Med.
(2):179192. 2014.
ever-increasing population and 2013;368(14):13351343. 19. Hallstro m E, Carlsson-Kanyama A, 34. Gunders D. Wasted: how America is
Bo rjesson P. Environmental impact of losing up to 40 percent of its food from
consumer demand, is using nat- 5. Akhtar AZ, Greger M, Ferdowsian H,
dietary change: a systematic review. J farm to fork to landll. Natural Resources
ural resources and producing Frank E. Health professionals roles in
Clean Prod. 2015;91:111. Defense Council issue paper, August
animal agriculture, climate change, and
waste at unsustainable rates. In 20. Soret S, Mejia A, Batech M, Jaceldo-
2012. Available at: https://www.nrdc.
human health. Am J Prev Med. 2009;36(2): org/food/les/wasted-food-ip.pdf.
turn, this results in serious im- 182187. Siegl K, Harwatt H, Sabat J. Climate
Accessed December 1, 2014.
plications for public health, in- change mitigation and health effects of
6. Rockstro m J, Steffen W, Noone K, 35. Venkat K. The climate change and
varied dietary patterns in real-life settings
cluding climate change, food et al. A safe operating space for humanity.
throughout North America. Am J Clin economic impacts of food waste in the
insecurity, and changing food Nature. 2009;461(7263):472475. United States. Int J Food System Dynamics.
Nutr. 2014;100(suppl 1):490S495S.
consumption patterns. The 7. Steffen W, Richardson K, Rockstro m J, 2011;2(4):431446.
21. Cannon G. The rise and fall of dietetics
et al. Planetary boundaries: guiding hu- 36. Food Wastage Footprint: Impacts on
food system must operate within man development on a changing planet.
and of nutrition science, 4000 BCE2000
CE. Public Health Nutr. 2005;8(6A):701705. Natural ResourcesSummary Report.
environmental constraints to Science. 2015;347(6223):736746. Rome, Italy: Food and Agriculture Or-
avoid disastrous consequences for 22. Scully M. Dominion: The Power of Man,
8. Smith JB, Schneider SH, Oppenheimer ganization of the United Nations; 2013.
the Suffering of Animals, and the Call to
the biosphere. Given the intrinsic M, et al. Assessing dangerous climate Mercy. New York, NY: St. Martins 37. Gustavsson J, Cederberg C, Sonesson
relationships between the envi- change through an update of the In- Grifn; 2003. U, van Otterdijk R, Meybeck A. Global
tergovernmental Panel on Climate Food Losses and Food Waste. Rome, Italy:
ronmental sciences and the nu- Change (IPCC) reasons for concern. 23. Tilman D. The greening of the green Food and Agriculture Organization of the
tritional sciences, it is imperative Proc Natl Acad Sci USA. 2009;106(11): revolution. Nature. 1998;396(6708): United Nations; 2011.
41334137. 211212.
that public health research and 38. Marlow HJ, Harwatt H, Soret S,
24. Tscharntke T, Clough Y, Wanger TC,
practice begin a concerted focus 9. Vermeulen SJ, Campbell BM, Ingram
et al. Global food security, biodiversity con-
Sabat J. Comparing the water, energy,
JSI. Climate change and food systems. Annu pesticide and fertilizer usage for the pro-
on the new discipline of envi- servation and the future of agricultural in- duction of foods consumed by different
Rev Environ Resour. 2012;37:195222.
ronmental nutrition, which seeks tensication. Biol Conserv. 2012;151(1):5359. dietary types in California. Public Health
10. Loos J, Abson DJ, Chappell MJ, et al.
to comprehensively address Putting meaning back into sustainable
25. Burlingame B, Dernini S, eds. Sus- Nutr. 2015;18(13):24252432.
tainable Diets and Biodiversity: Directions and
the sustainability of food intensication. Front Ecol Environ. 2014; 39. McMichael AJ, Powles JW, Butler
Solutions for Policy, Research and Action. CD, Uauy R. Food, livestock production,
systems. 12(6):356361.
Rome, Italy: Food and Agriculture Or- energy, climate change, and health. Lancet.
11. Porter J, Xie L, Challinor A, et al. Food ganization of the United Nations; 2012. 2007;370(9594):12531263.
CONTRIBUTORS security and food production systems. In:
26. Sutton MA, Bleeker A, Howard CM, 40. Steinfeld HGP. Livestocks Long
J. Sabat obtained funding, conceptual- Field CB, Barros VR, Dokken DJ, et al.,
et al. Our Nutrient World: The Challenge to Shadow: Environmental Issues and Options.
ized and designed the research, and edited eds. Climate Change 2014: Impacts, Adap-
Produce More Food and Energy With Less Rome, Italy: Food and Agriculture Or-
the article. H. Harwatt conceptualized the tation, and Vulnerability. Part A: Global and
Pollution. Edinburgh, UK: Centre for ganization of the United Nations; 2006.
research and wrote and edited the article. Sectoral Aspects. Contribution of Working
Ecology and Hydrology; 2013.
S. Soret conceptualized the research and Group II to the Fifth Assessment Report of the 41. Garnett T. Where are the best op-
edited the article. Intergovernmental Panel on Climate Change. 27. Ripple WJ, Newsome TM, Wolf C, portunities for reducing greenhouse gas
Cambridge, UK: Cambridge University et al. Collapse of the worlds largest her- emissions in the food system (including the
ACKNOWLEDGMENTS Press; 2014:485533. bivores. Sci Adv. 2015;1(4). Available food chain)? Food Policy. 2011;36(suppl 1):
This research was funded by the McLean at: http://advances.sciencemag.org/ S23S32.
12. Heller MC, Keoleian GA, Willett content/1/4/e1400103.full-text.pdf
Fund for Nutrition Research. WC. Toward a life cycle-based, diet-level 42. Schmidinger K, Stehfest E. Including
The concepts described in this essay +html. Accessed May 1, 2015.
framework for food environmental im- CO2 implications of land occupation in
were presented at the 2nd World Congress 28. Antimicrobial Resistance: Global Report LCAsmethod and example for livestock
pact and nutritional quality assessment.
of Public Health Nutrition; September on Surveillance. Paris, France: World products. Int J Life Cycle Assess. 2012;17(8):
Environ Sci Technol. 2013;47(22):
2325, 2010; Porto, Portugal. Health Organization; June 2014. 962972.
1263212647.
Note. The funder was not involved in
13. Garnett T. Food sustainability: prob- 29. Aktar MW, Sengupta D, Chowdhury 43. Stehfest E, Bouwman L, van Vuuren
the content or production of this essay.
lems, perspectives and solutions. Proc Nutr A. Impact of pesticides use in agriculture: DP, den Elzen MGJ, Eickhout B, Kabat P.
their benets and hazards. Interdiscip Climate benets of changing diet. Clim
HUMAN PARTICIPANT Soc. 2013;72(1):2939.
Toxicol. 2009;2(1):112. Change. 2009;95(12):83102.
PROTECTION 14. Sabat J, Soret S. Sustainability of
Institutional review board approval was 30. Eshel G, Martin PA. Geophysics and 44. Food and Agriculture Organization.
plant-based diets: back to the future. Am J
not needed for this article because no nutritional science: toward a novel, uni- Food outlook. Global market analysis.
Clin Nutr. 2014;100(suppl 1):476S482S.
human participants were involved. ed paradigm. Am J Clin Nutr. 2009;89(5): November 2012. Available at: http://
15. Uauy R. Dening and addressing the 1710S1716S. www.fao.org/docrep/016/al993e/
nutritional needs of populations. Public al993e00.pdf. Accessed December 1, 2014.
REFERENCES 31. Vitousek PM, Naylor R, Crews T,
Health Nutr. 2005;8(6A):773780.
1. Tilman D, Clark M. Global diets link et al. Nutrient imbalances in agricultural 45. Horrigan L, Lawrence RS, Walker P.
environmental sustainability and human 16. McAlpine CA, Seabrook LM, Ryan development. Science. 2009;324(5934): How sustainable agriculture can address
health. Nature. 2014;515(7528):518522. JG, et al. Transformational change: cre- 15191520. the environmental and human health
ating a safe operating space for humanity. harms of industrial agriculture. Environ
2. Food and Agriculture Organization of 32. Pew Commission on Industrial Farm
Ecol Soc. 2015;20(1). Available at: http:// Health Perspect. 2002;110(5):445456.
the United Nations. How to feed the Animal Production. Putting meat on
www.ecologyandsociety.org/vol20/iss1/
world in 2050. Rome, October 1213, the table: industrial farm animal pro- 46. Gerber P, Steinfeld H, Henderson B,
art56. Accessed May 1, 2015.
2009. Available at: www.fao.org/wsfs/ duction in America. 2008. Available at: et al. Tackling Climate Change Through
forum2050/wsfs-background- 17. Buttriss J, Riley H. Sustainable diets: http://www.livablefutureblog.com/ LivestockA Global Assessment of Emissions
documents/wsfs-expert-papers/en. harnessing the nutrition agenda. Food pdf/Putting_Meat_on_Table_FULL. and Mitigation Opportunities. Rome, Italy:
Accessed December 1, 2014. Chem. 2013;140(3):402407. pdf. Accessed December 22, 2010. Food and Agriculture Organization; 2013.
e6 Analytic Essays Peer Reviewed Sabat et al. AJPH Published online ahead of print March 17, 2016
REVIEW
CURRENT
OPINION What an anticardiovascular diet should be in 2015
David R. Jacobs Jr a and Linda C. Tapsell b
Purpose of review
Given scientific and public debate about optimal diet to prevent cardiovascular disease, and interest in diet
and other chronic diseases, we propose that following a few simple dietary principles would reduce
chronic disease incidence.
Recent findings
Nutrition research has been criticized for focusing on individual nutrients and foods, treated like drug
therapy. With a few important exceptions, clinical trials of supplemental nutrients have not shown benefit.
Although highly specific nutrition information is elusive, diet patterns have provided consistent answers,
important for public health. Observational cohort studies have found that some dietary patterns are
reported with high reliability over long periods and predict future cardiovascular and other inflammatory-
related diseases. Two randomized clinical trials confirmed this finding. There are many common features
of Mediterranean and prudent diets, particularly the plant-centered aspect, coupled with variety of foods
eaten. A dietary pattern characterized by high fruit, vegetable, legume, whole grain, nut, berry, seed, and
fish intakes, and possibly by intakes of dairy, coffee, tea, chocolate, and alcohol (not in excess), but low
meat and detrimentally processed foods is associated with reduced incidence of cardiovascular disease
and rates of noncardiovascular, noncancer chronic inflammatory-related mortality.
Summary
A plant-centered diet may be broadly recommended.
Keywords
cardiovascular disease, chronic disease, diet patterns, plant-centered diet
0957-9672 Copyright ! 2015 Wolters Kluwer Health, Inc. All rights reserved. www.co-lipidology.com 271
[29]. That coronary heart disease reduction also studying middle-aged male health professionals,
follows serum cholesterol reduction even when total Hu et al. [33] performed principal components
mortality is not reduced [30] strengthens the prop- analysis and found two patterns: prudent, with
osition that cholesterol in the serum plays a role high weightings on green leafy vegetables, dark-
in atherosclerotic disease, because it suggests that yellow vegetables, cruciferous vegetables, other
cholesterol reduction is specific to coronary heart vegetables, legumes, fruit, tomatoes, fish, garlic,
disease, whatever else might eventuate from serum poultry, and whole grains; and Western, with high
cholesterol lowering. Yet, this efficient logic is not weightings on red meat, processed meat, refined
necessarily correct. Saturated fat is not the only grains, sweets and desserts, french fries, high-fat
component of saturated fat-containing food; other dairy products, eggs, high-sugar drinks, snacks,
food components might be important. Many bio- condiments, margarine, potatoes, and butter. The
logical changes occur with cholesterol changes, prudent diet was inversely related to incident cor-
particularly considering the pleiotropic nature of onary heart disease, whereas the Western diet was
statins, and these also could be important. Direct adversely related. This list of foods closely resembles
observation of the association between saturated fat the Dietary Approaches to Stop Hypertension diet,
and coronary heart disease did not find the associ- in which vegetables, fruit, fruit juice, whole grains,
ation of saturated fat intake with coronary heart nuts, seeds, legumes, and dairy products other than
which as expected by the syllogism [31]. Such syllo- whole milk were increased, and meat, poultry, fish,
gisms are standard in science, for linking and inter- refined grain, whole milk, sweets, snack foods, fats,
preting the chains of evidence, but they should be oils, and salad dressings were decreased [38].
used cautiously because they are not always right. The Prevencion con Dieta Mediterranea (PRE-
In contrast, direct observation of certain dietary DIMED) study [9 ] instituted a 14-point Mediterra-
&&
patterns with various cardiovascular and other nean diet compliance score. Foods favored are olive
chronic diseases has consistently shown reduced oil, vegetables, fruit (including natural juices), wine,
risk [3235,36 ], although meta-analysis of specific legumes, fish or shellfish, nuts (including peanuts),
&&
foods and nutrients was found to be confusing [37]. chicken, turkey, or rabbit meat instead of veal, pork,
This difference in findings possibly relates to the hamburger, or sausage, and sofrito (a sauce made
stability of diet patterns over time and reliability of with tomato and onion, leek, or garlic, simmered
reporting. Jacobs and Orlich [19] pointed out that in with olive oil). Foods not favored are red meat,
most cases, diet patterns derived from food fre- hamburger, or meat products (ham, sausage, and
quency questionnaires had high correlation within so on), butter, margarine, or cream, sweet/carbo-
person over periods of years; two studies reported nated beverages, and commercial sweets or pastries
tracking correlation over 0.5 over 18 and 20 years. (not homemade), such as cakes, cookies, biscuits, or
Correlations over time of most foods and nutrients custard. Participants randomized to Mediterranean
are much lower, on the order of 0.2 or less, although diet groups had about 2 out of 14 points higher score
a few well-defined foods that are habitually con- than the randomized control group, starting soon
sumed by some people, such as coffee, tea, or alco- after baseline and throughout the study (Table S1
hol, have high-tracking correlations. In addition, and Figure S3 in [9 ]). Higher baseline scores pre-
&&
diet patterns integrate over foods eaten and over dicted cardiovascular disease outcomes [39].
long periods, whereas individual foods and Mursu et al. [36 ] studied various mortality out-
&&
nutrients do not capture the sum of complex nutri- comes in the Iowa Womens Health Study (IWHS)
tion effects. according to two diet scores, the A Priori Diet Qual-
ity Score (APDQS) [4045] and the Alternative
Healthy Eating Index 2010 (AHEI) [46].
MEDITERRANEAN AND PRUDENT DIET The APDQS originated in a case"control study
PATTERNS THAT PROTECT AGAINST to differentiate myocardial infarction from healthy
CHRONIC DISEASE controls according to combinations of foods [40]. It
There are many ways of describing diet patterns that was applied in several studies subsequently, show-
attempt to capture key characteristics, and to a ing relationships with demographics [41], several
certain extent, culture, such as Mediterranean (char- biomarkers [4245], cognitive function [47], and
acterizing foods from that region), prudent (imply- fitness [48]. It has been reformulated for different
ing wise, judicious food choices), and Western studies, depending on the questionnaire data avail-
(reflecting certain aspects of food industrialization able. Approximately 40 food groups, covering many
in Western culture). The central focus of diet pattern aspects of diet, are rated as favorable, adverse or
research has been Mediterranean or prudent diets neutral and placed in quantiles according to intake.
with various definitions [32]. In an early article For example, in the IWHS [36 ], 34 food groups
&&
were rated favorable (n 17), adverse (n 10), or for the AHEI-2010. We regard a difference of 36.9
neutral (n 7) relative to future chronic disease, versus 30.2 as of considerable clinical importance in
based on the literature and expert judgment, and women initially average age 62 years, followed into
then placed in quartiles (or a large nonconsumer old age. Furthermore, the women were resurveyed
group and tertiles among consumers). Mursu et al. in 2004 (n 15 076) at average age 79 years, and
[36 ] stated: The positively rated food groups were findings were similar or numerically stronger for
&&
beans and legumes, beer, coffee, fish, fruit, green diet score prediction of all outcomes.
vegetables, low-fat dairy, liquor, oil, other veg- One of the goals of Mursu et al. [36 ] was to
&&
etables, poultry, seeds, and nuts, soy products, tea, identify information in each score beyond what was
tomato, whole grains, and wine. The negatively found in the other score. Each score did add some
rated food groups were butter, fried foods, fried prediction when the other was in the model, but it
potatoes, red meat, liver, processed meat, salty was difficult to sort out what that information was.
snacks, soft drinks, sweets, and whole-fat dairy. More to the point for this review is the commonality
The remaining neutrally rated food groups were between these scores and of these scores with other
chocolate, diet soft drinks, eggs, fruit juice, margar- prudent and Mediterranean diet patterns.
ine, potatoes, and refined grains.
Groups were given scores of 0 for lowest quartile
of consumption or nonconsumers to 3 for highest CONCLUSION
consumption grouping. The APDQS was formed by The many diet pattern analyses cited here, although
adding these values in a positive direction for bene- consistent across many settings and having substan-
ficially rated food groups, in a negative direction for tial commonality across patterns studied, are obser-
adversely rated food groups, and 0 for neutrally vational and therefore considered insufficient in
rated food groups, for a theoretical maximal score themselves to make causal statements about diet
of 81. Thus, 1 point in the APQDS corresponds to a protecting against future chronic disease. However,
difference of one higher intake category in benefi- coupled with the randomized findings of the Lyon
cially rated food groups or one lower intake category Diet-Heart Study [8] and PREDIMED [9 ], we assert
&&
in adversely rated food groups. Mean $ standard that certain diet patterns can prevent or delay a
deviation of the APDQS was 38.4 $ 8.2 in the substantial amount of currently existing chronic
29 634 women aged 5569 years and free of self- disease, particularly cardiovascular and noncardio-
reported heart disease, diabetes, and cancer at base- vascular, noncancer inflammatory-related diseases.
line in 1986 and followed for 22 years. Such diet patterns are plant-centered, at the same
The AHEI-2010 [46] was formulated as an time minimizing intake of nutritionally poor plant
improvement to an earlier similar score [49] that foods such as sugar, refined grains, and highly proc-
provided dietary guidance related to the US Dietary essed and very salty foods. Intake of plant seeds,
Guidelines. AHEI-2010 contains 11 components, including whole grains, nuts, berries, coffee, and
each of which receives 010 points (theoretical chocolate, is promising [50 ]. The recommended
&&
maximum 110). Components included vegetables, prudent/Mediterranean diet patterns are highly var-
whole fruit, whole grains, nuts and legumes, red and ied, as can be gleaned from the small increments to
processed meat, long-chain (n-3) fatty acids eicosa- the APDQS from 34 widely varying food groups.
pentaenoic acid docosahexaeneoic acid, polyun- Apparently, there are many ways to achieve a pro-
saturated fatty acids, and moderate alcohol intake tective diet pattern.
(all rated favorably) and trans fatty acids, sugar- The prudent/Mediterranean diet pattern
sweetened beverages and fruit juices, and sodium solution appears to be robust, based on its consist-
(all rated adversely). In the IWHS, the mean AHEI- ency across many cohort studies, the Lyon Diet-
2010 was 40.0 $ 10.4, with correlation of 0.65 Heart Study and PREDIMED. However, some caveats
between the AHEI-2010 and the APDQS. are in order. It seems to us that vegetarian diet
Both the APDQS and AHEI-2010 predicted total, patterns work because of the presence of nutrition-
cardiovascular, cancer, and noncardiovascular, non- ally rich plant foods, not because of the absence of
cancer inflammatory-related mortality, but not meat. Thus, prudent/Mediterranean diet patterns
other external causes of mortality. Prediction was are not necessarily vegetarian, although much of the
numerically strongest for inflammatory-related meat in the diet is rated poorly in all patterns.
mortality and weakest for cancer mortality, but However, meat is less sustainable than other foods,
statistically significant for all. The risk factor given the energy cost of growing meat products of
adjusted total mortality rates were 36.9, 34.6, all types, and sustainability is a major issue for the
32.5, and 30.2 per 100 women followed for 22 years future [51]. Another caveat concerns fish, which
across the increasing APDQS quartiles, and similarly is highly recommended across all diet patterns. A
0957-9672 Copyright ! 2015 Wolters Kluwer Health, Inc. All rights reserved. www.co-lipidology.com 273
19. Jacobs DR Jr, Orlich MJ. Diet pattern and longevity: do simple rules suffice? A 38. Jacobs DR Jr, Gross MD, Steffen L, et al. The effects of dietary patterns on
commentary. Am J Clin Nutr 2014; 100 (Suppl 1):313S319S. urinary albumin excretion: results of the Dietary Approaches to Stop Hyper-
20. Fardet A, Rock E. Toward a new philosophy of preventive nutrition: from a tension (DASH) Trial. Am J Kidney Dis 2009; 53:638646.
reductionist to a holistic paradigm to improve nutritional recommendations. 39. Schroder H, Salas-Salvado J, Martnez-Gonzalez MA, et al. Baseline adher-
Adv Nutr 2014; 5:430446. ence to the Mediterranean diet and major cardiovascular events: Prevencion
21. Mozaffarian D, Ludwig DS. Dietary guidelines in the 21st century: a time for con Dieta Mediterranea trial. JAMA Intern Med 2014; 174:16901692.
food. JAMA 2010; 304:681682. 40. Lockheart MS, Steffen LM, Rebnord HM, et al. Dietary patterns, food groups
22. Keys A, Anderson JT, Grande F. Serum cholesterol response to changes in and myocardial infarction: a case-control study. Br J Nutr 2007; 98:380387.
the diet, IV: particular saturated fatty acids in the diet. Metabolism 1965; 41. Sijtsma FP, Meyer KA, Steffen LM, et al. Longitudinal trends in diet and effects
14:776784. of sex, race, and education on dietary quality score change: the Coronary
23. Mensink RP, Katan MB. Effect of dietary trans fatty acids on high-density and Artery Risk Development in Young Adults study. Am J Clin Nutr 2012;
low-density lipoprotein cholesterol levels in healthy subjects. N Engl J Med 95:580586.
1990; 323:439445. 42. Jacobs DR Jr, Sluik D, Rokling-Andersen MH, et al. Association of 1-y changes
24. Hegsted DM. A look back at lessons learned and not learned. J Nutr 1994; in diet pattern with cardiovascular disease risk factors and adipokines: results
124 (Suppl):1867S1870S. from the 1-y randomized Oslo Diet and Exercise Study. Am J Clin Nutr 2009;
25. Sabine J. Cholesterol. New York: Marcel Dekker, Inc; 1977. 89:509517.
26. Wissler RW. New insights into the pathogenesis of atherosclerosis as 43. Meyer KA, Sijtsma FP, Nettleton JA, et al. Dietary patterns are associated with
revealed by PDAY. Pathobiological Determinants of Atherosclerosis in Youth. plasma F2-isoprostanes in an observational cohort study of adults. Free Radic
Atherosclerosis 1994; 108 (Suppl):S3S20. Biol Med 2013; 57:201209.
27. Jacobs D, Blackburn H, Higgins M, et al. Report of the conference on low 44. Sijtsma FP, Meyer KA, Steffen LM, et al. Diet quality and markers of endothelial
blood cholesterol: Mortality Associations. Circulation 1992; 86:10461060. function: the CARDIA study. Nutr Metab Cardiovasc Dis 2014; 24:632638.
28. Prospective Studies Collaboration. Lewington S, Whitlock G, et al. Blood 45. Nettleton JA, Schulze MB, Jiang R, et al. A priori-defined dietary patterns and
cholesterol and vascular mortality by age, sex, and blood pressure: a meta- markers of cardiovascular disease risk in the Multi-Ethnic Study of Athero-
analysis of individual data from 61 prospective studies with 55 000 vascular sclerosis (MESA). Am J Clin Nutr 2008; 88:185194.
deaths. Lancet 2007; 370:18291839. 46. Chiuve SE, Fung TT, Rimm EB, et al. Alternative dietary indices both strongly
29. Ebrahim S, Taylor FC, Brindle P. Statins for the primary prevention of predict risk of chronic disease. J Nutr 2012; 142:10091018.
cardiovascular disease. BMJ 2014; 348:g280. 47. Zhu N, Jacobs DR, Meyer KA, et al. Cognitive function in a middle aged cohort
30. Gould AL, Rossouw JE, Santanello NC, et al. Cholesterol reduction yields is related to higher quality dietary pattern 5 and 25 years earlier: The CARDIA
clinical benefit. A new look at old data. Circulation 1995; 91:22742282. Study. J Nutr Health Aging 2015; 19:3338.
31. Hu FB, Stampfer MJ, Manson JE, et al. Dietary fat intake and the risk of 48. Shikany JM, Jacobs DR Jr, Lewis CE, et al. Associations between food
coronary heart disease in women. N Engl J Med 1997; 337:14911499. groups, dietary patterns, and cardiorespiratory fitness in the Coronary Artery
32. Trichopoulou A, Martnez-Gonzalez MA, Tong TY, et al. Definitions and Risk Development in Young Adults study. Am J Clin Nutr 2013; 98:1402
potential health benefits of the Mediterranean diet: views from experts around 1409.
the world. BMC Med 2014; 12:112. 49. McCullough ML, Feskanich D, Stampfer MJ, et al. Diet quality and major
33. Hu FB, Rimm EB, Stampfer MJ, et al. Prospective study of major dietary chronic disease risk in men and women: moving toward improved dietary
patterns and risk of coronary heart disease in men. Am J Clin Nutr 2000; guidance. Am J Clin Nutr 2002; 76:12611271.
72:912921. 50. Ros E, Hu FB. Consumption of plant seeds and cardiovascular health:
34. Stradling C, Hamid M, Taheri S, Thomas GN. A review of dietary influences on && epidemiological and clinical trial evidence. Circulation 2013; 128:553565.
cardiovascular health: part 2 " dietary patterns. Cardiovasc Hematol Disord This review and thought piece is a very clever amalgamation of a series of thought
Drug Targets 2014; 14:5063. threads that combines and expands the important work on whole cereal grains and
35. Martinez-Gonzalez MA, Bes-Rastrollo M. Dietary patterns, Mediterranean diet, nuts.
and cardiovascular disease. Curr Opin Lipidol 2014; 25:2026. 51. Sabate J, Soret S. Sustainability of plant-based diets: back to the future. Am J
36. Mursu J, Steffen LM, Meyer KA, et al. Diet quality indexes and mortality in Clin Nutr 2014; 100 (Suppl 1):476S482S.
&& postmenopausal women: the Iowa Womens Health Study. Am J Clin Nutr 52. Nstbakken OJ, Hove HT, Duinker A, et al. Contaminant levels in Norwegian
2013; 98:444453. farmed Atlantic salmon (Salmo salar) in the 13-year period from 1999 to
Among many articles reporting disease outcomes according to diet patterns in 2011. Environ Int 2015; 74:274280.
prospective observational studies, this one stands out because it compares and 53. Ruzzin J, Bethune C, Goksyr A, et al. Comment on Contaminant levels in
contrasts results for two highly predictive diet pattern formulations. Although the Norwegian farmed Atlantic salmon (Salmo salar) in the 13-year period from
authors focused on differences between the diet patterns, in the sense of the 1999 to 2011 by Nstbakken et al. Environ Int 2015. [Epub ahead of print].
current review the commonality of the two diet patterns is worthy of note. doi: 10.1016/j.envint.2015.01.003.
37. Berciano S, Ordovas JM. Nutrition and cardiovascular health. Rev Esp Cardiol 54. Latham LS, Hensen ZK, Minor DS. Chocolate: guilty pleasure or healthy
(Engl Ed) 2014; 67:738747. supplement? J Clin Hypertens (Greenwich) 2014; 16:101106.
0957-9672 Copyright ! 2015 Wolters Kluwer Health, Inc. All rights reserved. www.co-lipidology.com 275
Am J Clin Nutr 2009;89(suppl):1699S703S. Printed in USA. ! 2009 American Society for Nutrition 1699S
1700S MARLOW ET AL
irrigation systems (25). These effects may have greater signifi- Waste generation
cance during times of seasonal or extended drought (21). In addition to the previously identified pollution problems,
wastes generated by intensified animal production often result in
Energy consumption significant water, soil, and air pollution (40). In the United States,
Increased use of fossil fuels and concurrent technologic 7 billion livestock generate 130 times more waste than produced
advances have allowed humans to increase the productivity of by 300 million humans (41). These wastes, most of which go
natural systems by manipulating the environment (17, 23). The untreated, contain high concentrations of nitrogen, phosphorous,
energy intensiveness of agricultural production varies with type and potassium compounds and traces of metals and antibiotics;
of crop produced, amount of chemical inputs, and geographic these represent a serious public health problem according to the
location (18). With market globalization and convenient trans- World Health Organization and US Department of Agriculture
portation choices, food has become available during seasons (4, 23, 40, 41). Concentrated livestock operations and livestock
when they were typically absent, and the increased energy re- waste also produce gases. Some, such as ammonia, have a more
quirements are largely borne by consumers and driven by market local effect and are generally regarded as nuisance odors (40).
demand. In the United States, fossil fuel consumption doubled Others such as carbon dioxide, methane, and nitrous oxide exert
during a 20-y span while the caloric return per calorie of input a global effect and have been implicated in climate change (4,
on most crops diminished (15). Cheap sources of fossil fuels will 26, 40, 42).
continue to allow for massive energy inputs to agricultural
systems, but, as prices increase and supplies dwindle, this Land degradation
practice is likely to change (14). Conservation and optimization
of energy use will certainly be in the future of agriculture. Livestock production exacts a significant toll on natural
Adventist vegetarian diet conserves the equivalent of 54% of the 16. Brown LR. Human food production as a process in the biosphere. Sci
average weekly per capita indoor water consumption. This can Am 1970;223:16170.
17. Cleveland CJ. The direct and indirect use of fossil fuels and electricity in
be compared with a savings of 35%, estimated by the American USA agriculture, 19101990. Agric Ecosyst Environ 1995;55:11121.
Water Works Association, by installing more-efficient water 18. Pimentel D, Pimentel M. US food production threatened by rapid
fixtures and regularly checking for leaks. From this comparison population growth. Washington, DC: Prepared for Carrying Capacity
it is apparent that a plant-based diet provides a significant water Network, 1997.
19. Baroni L, Cenci L, Tettamanti M, Berati M. Evaluating the environ-
conservation benefit. A similar ecologic cost effectiveness can mental impact of various dietary patterns combined with different food
be determined for each of the other inputs in the study. production systems. Eur J Clin Nutr 2007;61:27986.
Considering the surmounting ecologic pressures that a bur- 20. Tanji KK, Enos CA. Global water resources and agricultural use. In:
geoning human civilization exerts on our planet, there is a need to Tanji KK, Yaron B, eds. Management of water use in agriculture. Berlin,
Germany: Springer-Verlag, 1994:324.
make hard decisions. Among these hard decisions, many societies,
21. Zilberman D, Dunarm A, MacDougal N, Khanna M, Brown C, Castillo
and governments in particular, will have to reconsider the increasing F. Individual and institutional responses to the drought: the case of
demand for an animal-based diet. Many governments, including California agriculture. J Contemp Water Res Educ 2002;121:1723.
both the European Union and the US government, may need to 22. Pimentel D, Berger B, Filiberto D, et al. Water resources: agricultural
reassess agricultural subsidies (59, 60) and divert some of the and environmental issues. Bioscience 2004;54:90918.
23. Pimentel D. Environmental and economic benefits of sustainable agri-
funding to support additional research, development, and applica- culture. In: Kihn J, Gowdy J, Hinterberger F, van der Straaten J, eds.
tion of sustainable methods of food production. Outreach programs Sustainability in question. Cheltenham, United Kingdom: Edward Elgar
may be necessary to educate and inform people about the health Publishing Ltd, 1999:15370.
and environmental benefits of a vegetarian diet. (Other articles in 24. Lemly AD, Kingsford RT, Thompson JR. Irrigated agriculture and
wildlife conservation: conflict on a global scale. Environ Manage 2000;
this supplement to the Journal include references 6187.) 25:485512.
Am J Clin Nutr doi: 10.3945/ajcn.113.071522. Printed in USA. ! 2014 American Society for Nutrition 1S of 7S
FIGURE 2. Graphic summary of the various types of the food production: ratio of energy required to food energy delivery. Reprinted with permission from reference 6.
SUSTAINABILITY OF PLANT-BASED DIETS 3S of 7S
EFFICIENCY: ANIMAL COMPARED WITH PLANT FOODS industrialized agricultural practices, causes significant emissions
Raising animals for human food is an intrinsically inefficient of greenhouse gases (GHGs). These may occur directly from
process. As we move up in the trophic chain there is a progressive carbon dioxide emissions from fossil fuel use on the farm or in the
loss of energy. Grass-fed livestock subsists, but this is not the supply chain, nitrous oxide emissions resulting from fertilizer
main source of meat for human consumption in developed na- application, and methane emissions from animals or indirectly as
tions. Modern husbandry (animal farms) is based on intensive a result of land use change. In addition to agricultural production
feeding of grain crops to animals (5). This grain could be a source process, the transport, processing, packaging, marketing, sales,
of food for humans. The same standards apply to the production purchasing, and cooking of food also contribute to GHG emis-
of other animal products such as eggs and dairy. Several authors sions. Berners-Lee et al (11) computed the GHG emissions from
have computed the efficiency ratios of animal compared with 61 food categories at point of purchase. With one exception, each
plant foods for human consumption. The amount of grain needed food with GHG emissions .10 kg CO2 equivalents/kg of product
to produce the same amount of meat varies from a ratio of 2.3 for is a meat or dairy food.
chicken to 13 for beef (Table 1). Pimentel and Pimentel (8) Approximately half of all food-related GHG emissions are
established that, on average, 11 times greater fossil energy is generated during farming. Farm-stage emissions include nitrous
required to produce animal protein than plant protein for human oxide and methane from livestock and carbon dioxide from ag-
consumption. However, the energy-to-protein efficiency ratio riculturally induced change in land use, especially deforestation.
varies greatly by type of meat. More specifically, it is only 4 Nitrous oxide from pasture land and arable land used to grow
times greater for chicken protein compared with grain protein food crops and methane from the digestive processes of ruminant
but 40 times greater for beef protein compared with grain pro- animals (eg, cows and sheep) account for 80% of all agricultural
tein. We have previously reported that the ratio for water used in GHG emissions (12). The emissions per unit of livestock product
the production of soy protein compared with the same quantity vary by animal type and are much higher in ruminant animals
of animal protein is from 4 to 26 and showed that the ratio such as in cattle, sheep, and dairy farming than in pig and poultry
between soy protein and the different types of animal proteins farming (13).
varies from 6 to 20 for fossil fuel usage (9). The land required to Beyond contributing disproportionately to GHG emissions
raise the feed to produce animal protein is 617 times greater than to the atmosphere, industrial husbandry operations damage the
for soy protein (9).Thus, the conversion of plant foods to foods of environment by chemical runoff to water and land and animal waste.
animal origin is an intrinsically inefficient process (w10:1). This chemical pollution may cause acidification, algal blooms, and
The ratio of energy inputs to protein delivery is also qualitatively dead zones in lakes and coastal areas; soil quality degradation;
different for animal compared with plant foods. As the concen- habitat change; and biodiversity loss (5, 12, 14). We previously
tration of protein increases in plant foods, so does the efficiency. It computed 3 environmental degradation ratios of meat compared
does not change or may even decrease in animal protein sources with soy protein production and reported that the emissions of
(Figure 3) (10). High-protein plant foods such as soy beans and acidifying substances, pesticides, and metals are respectively 7, 6,
other legumes have greater protein delivery energy efficiency and 100 times greater for meat protein compared with soy (9).
than cereals, which have a lower protein concentration. Therefore, Thus, the intensive production of meat is considerably more taxing
less energy is needed to produce the same amount of protein from to the environment than nutritionally equivalent plant protein foods.
soy than from corn. However, very similar amounts of energy are
used to produce equivalent amounts of protein from different
COMPARING ENVIRONMENTAL INDICATORS OF
sources of animal protein. In animal foods, the degree of protein
PLANT-BASED AND MEAT-BASED DIETS
concentration seems to decrease the efficiency ratio of energy
inputs compared with protein outputs. Compared with plant foods, meat and dairy products are clearly
responsible for a hefty share of the natural resource utilization
and environmental burden of food production. However, looking
ENVIRONMENTAL INDICATORS: ANIMAL COMPARED at dietary patterns, instead of single foods, is a more integrated
WITH PLANT FOODS and realistic approach in the assessment of the environmental
Research has been conducted on individual food items to determine impacts of producing foods for human consumption. We pre-
the impact of production and consumption on the environment. viously compared the environmental impacts of producing the
The production of food for human consumption, particularly by foods differentially consumed by vegetarians and nonvegetarians
from California and showed that the agricultural inputs required
for producing the nonvegetarian diet were 2.9 times more water,
TABLE 1
2.5 times more primary energy, 13 times more fertilizer, and 1.4
Ratio of different inputs to animal outputs in US husbandry practices1
times more pesticides than for the vegetarian diet (15).
Grain fed:meat produced Fossil fuel energy:protein energy Earlier studies reported several-fold increases in GHG emis-
kg:kg kcal:kcal sions and/or in environmental degradation scores for conventional
Beef 13:1 40:1 meat diets compared with vegetarian diets (16, 17). However,
Eggs 11:1 39:1 these notable differences between dietary patterns also encom-
Pork 5.9:1 14:1 passed factors other than the diet composition, such as agricultural
Milk 14:1 production methods, food processing, and transportation. Such
Turkeys 3.8:1 10:1 factors, albeit important contributors to the overall environmental
Chickens 2.3:1 4:1 footprint, are confounders in the assessment of the strict impact of
1
Adapted from reference 8. the composition of the diet.
4S of 7S SABATE AND SORET
FIGURE 3. Protein delivery efficiency in terms of energy use as a function of food protein content. Squares, plant-based foods; circles, animal-based foods.
The correlation coefficient r is shown in linear interpolations. Reprinted with permission from reference 10. GH, greenhouse; Vegs, vegetables.
Most of the research on the environmental impact of dietary GHG emissions for a vegan diet compared with the mean US
patterns has focused on the quantification of GHG emissions. diet on the basis of national food disappearance data. They cal-
Berners-Lee et al (11) contrasted 6 meatless daily meal scenarios culated a per capita annual decrease of 1.5 tons of CO2 equiv-
with the average UK diet. Under the different dietary scenarios, alents, which is equivalent to 33% fewer emissions from the
meat was replaced with either dairy products or plant-based vegan food pattern than from the average US diet. This repre-
alternatives. Compared with the benchmark UK diet, reductions sents an 8% reduction in the per capita total GHG emissions in
in GHG emissions of 22% for vegetarian and 26% for vegan the United States, which is similar to the reduction found for the
scenarios were estimated (Figure 4). With the use of data from Finnish consumer (18).
Finland, Risku-Norja et al (18) estimated that a vegan diet would Modeled reductions in GHG emissions ranging from 19% to
reduce 48% of the agricultural and 34% of the overall food 30% were estimated for several dietary scenarios involving the
system GHG emissions compared with the average Finnish diet. partial replacement of animal products, meat and dairy, in the
Reductions in agricultural and overall food system GHG emis- average UK diet (20, 21). Furthermore, Stehfest et al (22) esti-
sions of only 33% and 23%, respectively, result when dairy is mated reductions in GHG emissions of 17% for CO2, 24% for
removed and beef and mutton are replaced in the diet with pork CH4, and 21% N2O at the global level from a complete switch to
and poultry. Eshel and Martin (14, 19) estimated the decrease in meatless diets, in which all protein is derived from plants.
FIGURE 4. Reduction in greenhouse gas emissions of different prototype vegetarian and vegan diets compared with an average UK diet. Adapted by using
data from reference 11. CO2e, CO2 equivalents; p, person.
SUSTAINABILITY OF PLANT-BASED DIETS 5S of 7S
The aforementioned studies provide useful estimates of en- individual quantities and separate processes, the boundaries are
vironmental impacts, but researchers relied on prototype vege- tightly coupled. Humankind does not have the luxury of con-
tarian diets for the comparative estimates, an approach that has centrating its efforts on any one of them in isolation from the
inherent limitations. Prototype vegetarian diets may reflect the others (27). If one boundary is transgressed, then other bound-
preferences and potential bias of the researcher, and they do not aries are also under serious risk.
capture the variability and complexity of the foods consumed on The world population is growing exponentially and is expected
a daily basis by free-living individuals. Thus, impact assessments to reach 9 billion by the year 2050 (25). The parallel increase in
based on ideal or designed diets may produce only theoretical wealth in large segments of the population of transitional and
results and are not grounded in reality. We recently conducted developing countries has resulted in a markedly increased de-
an assessment of the carbon footprint of dietary patterns using mand for foods of animal origin, particularly meat and dairy
a novel approach (23), and our methodology was based not on foods. Since 1963, there has been a 62% increase worldwide in
researcher-designed, but on subject-defined food consumption meat consumption, but a much greater increase of w300% has
for a large and geographically representative study population occurred in developing nations. China has experienced a 9-fold
across North America. Furthermore, our study population exhibits increase in per capita meat consumption of meat since the 1980s
a great dietary variability with respect to the intake of animal and (Figure 5) (28). In addition to wealth, other factors driving the
plant foods, and we determined the emissions with global warming worldwide consumption of meat include urbanization, trade
potential (CO2, CH4, and N2O) of 210 foods included in a vali- liberalization, transnational food corporations, retailing growth,
dated food-frequency questionnaire by implementing a life-cycle food industry marketing, and consumer attitudes and behaviors
impact assessment. We estimated the GHG emissions for each of (29).
the .73,000 participants in the Adventist Health Study 2 according The projected increased size of the worlds population and the
to self-reported dietary preference. A reduction in emissions of increase in the appetite for meat are pushing our food systems to
31% and 22% was estimated for vegetarians and semivegetarians, unsustainable levels. Reports from authoritative agencies have
respectively, compared with nonvegetarians (23). These esti- accumulated in the past few years that document the negative
mates should be regarded as conservative because the average environmental impacts of industrial meat production (12, 3032).
meat consumption of 64 g/d for nonvegetarian Adventists con- Industrial meat production is intrinsically resource-inefficient and
trasts with the much higher 220 g/d in a typical Western diet. requires high inputs of nonrenewable natural resources and growth-
promoting antibiotics. It damages the environment through gaseous
emissions and chemical runoff into water and soil. Meat and dairy
THE CASE FOR PLANT-BASED DIETS AT THE GLOBAL production accounts for .80% of all GHG emissions from the
LEVEL food sector (30) and for 24% of global GHG emissions (12). Thus,
Is food security in the context of the current and projected the livestock industry is a major contributor to anthropogenic
dramatic increase in global demand for animal products com- GHG emissions liable for global climate change, one of the
patible with environmental sustainability? The answer to this planetary boundaries already transgressed by humanity and directly
question must be considered against the backdrop of the worlds threatening its sustainability.
demographic explosion and the increase in wealth among large Animal waste has become a public health problem and en-
segments of the population in transitional and developing na- vironmental hazard. Annually, 7 billion livestock in the US meat
tions (24, 25). These trends result in a dramatic increase in the industry generate 1.4 billion tons of wasteor w5 tons of waste
global demand for foods of animal origin, particularly meats and for every US citizen (5). These wastes, most of which go un-
dairy foods. Industrial livestock production is intrinsically re- treated, not only contain high concentrations of nitrogen, phos-
source-inefficient and highly taxing on the environment, ren- phorous, and potassium compounds and traces of metals but also
dering the current food system environmentally and societally antibiotics, and are the source of .100 zoonotic pathogens that
unsustainable. We argue in this section that for a sustainable may contaminate food and water supplies (33), thus representing
future, a drastic reduction in the consumption of meat and dairy a direct threat to human health.
foods by large segments of the worlds population is unavoidable. Food security and food sustainability are on a collision course.
Thus, plant-based diets at the global level are imperative (26). The past half-century has seen a marked growth in food pro-
The overall size and economic activity of humankind are exceeding duction, allowing for a dramatic decrease in the proportion of the
the biocapacity of the world. Until recently, it was assumed that worlds people who are hungry despite a doubling of the total
the worlds living and physical resources were inexhaustible. How-
ever, this is a false assumption attributable to a rapidly growing
reliance on nonrenewable natural resources, ie, fossil fuels. Many
human activities, including industrial agriculture, have reached
a level that could damage the systems that keep Earth in a desirable
state of ecological balance. The outcome could be irreversible
and, in some cases, lead to abrupt environmental change. Rockstrom
et al (3) identified 9 Earth-system self-regulatory processes and
associated thresholds that, if crossed, could generate unaccept-
able environmental change. Their analysis suggests that 3 of the
boundariesclimate change, rate of biodiversity loss, and inter-
ference with the nitrogen cyclehave already been trespassed. FIGURE 5. Trends in per capita meat consumption. Adapted by using
Although the planetary boundaries are described in terms of data from reference 28.
6S of 7S SABATE AND SORET
population. Nevertheless, 1 in 7 persons today consume diets dietary shift is complex and implicates behavioral and policy
insufficient in protein and energy, and 2 in 7 persons have some challenges at many levels.
micronutrient deficiency. With a larger-than-ever world population The adequacy of meatless diets has been a recurrent theme
that has a greater appetite for meat and with the threat of lower in the nutrition literature. On the basis of a higher concentration
food yields because of substantial climate change (34), the world of essential nutrients in animal products, meat and dairy were
food system is facing a set of intersecting challenges. These considered essential in large proportions for adequate nutrition in
challenges require radical changes not only in the way food is the daily diet, and consumption of plant-based diets was con-
produced, processed, stored, distributed, and accessed but also in sidered inadequate. This nutritional paradigm has changed in the
the types of food consumed. past few decades (38) as data now support that most plant-based
Previously, the main solution to food shortages was to bring diets are healthier than meat-based diets and yield greater lon-
more land into agriculture and to exploit fish stocks. These are not gevity and lower chronic diseases among those who consume
sustainable options. Neither is increasing the intensification of vegetarian diets (3942). Furthermore, there is growing evidence
existing underperforming agricultural landscapes. Closing the linking meat consumption, in particular red meat and processed
food yield gap, while considering sustainability goals, will re- meat, with detrimental health outcomes (4345). From a strict
quire new approaches, including the adaptation of conventional health perspective, there is no need to consume meat.
agriculture to more low-input and precision practices (3537). For millennia, large segments of the worlds population thrived
Waste and losses along the supply chain are estimated to involve on diets with little or no meat. In the past century, however, the
more than one-third of the food produced in the world (36). concept of eating meat as the paramount source of protein has
Reducing food waste could substantially improve food security become deeply engrained in the psyche and culture of Western
and simultaneously decrease environmental degradation. Improv- countries and now pervades many other cultures and nations.
ing food yields, reducing agricultures environmental impacts, Undertaking a drastic downshift in meat consumption will face
and reducing waste are necessary but are not sufficient strategies serious obstacles and opposition at many levels: the consumers
to ensure global food security. More food can be delivered by taste preferences; some culinary traditions; established social norms;
realigning agricultural and dietary preferences. economic forces, such as the livestock industry; and current national
Shifting diets from animal-based to plant-based at the global and international food policies. Several proposals have been
level is of paramount importance in achieving food security and advanced to accomplish the transition from animal protein to
sustainability goals. Decreasing consumption of meat and other plant-based protein consumption. Some of them include con-
animal products will free up large amounts of food that could be sumer education focused on the environmental and health merits
consumed directly by humansfor example, soy and grains. Foley of plant-based diets, the promotion of food guidelines based on
et al (36) recently estimated the potential to increase the global health and sustainability criteria, developing attractive and cul-
food supply by shifting 16 major crops from the current mix of turally acceptable plant-based meat-alternative foods, and re-
uses (eg, human consumption, livestock feed, and biofuels) to aligning current fiscal policy (food subsidies and taxation) with
human food consumption only: they estimated a 28% increase in efficiency and environmental criteria (4650).
food availability, or the equivalent of a 49% increase in dietary
energy for human consumption. These staggering figures, re-
sulting from the implementation of a single strategy, have pro- CONCLUSIONS
found implications for both food security and environmental Plant-based diets in comparison to meat-based diets are more
sustainability. By 2050, the worlds population is expected to sustainable because they use substantially less natural resources
plateau at 9 billion, a 28% increase from the current 7 billion and are less taxing on the environment. The worlds demographic
(25). Shifting the portion of current crops directed to livestock explosion and the increase in the appetite for animal foods render
feed and nonfood uses to human consumption could fill the the food system unsustainable. Food security and food sustain-
future food gap and exceed dietary energy needs. In addition, the ability are on a collision course. Changing course (to avoid the
proposed dietary shift would result in major environmental collision) will require extreme downward shifts in meat and
benefits because it does not require an overall increase in plant dairy consumption by large segments of the world population.
crops at the global level, and it would eliminate the environ- Although other approaches should be pursued, they are insufficient
mental impacts related to livestock. As presented above, in- to make the global food system sustainable, and therefore the
dustrial livestock production contributes disproportionally to dietary shift is an inevitable strategy.
numerous forms of environmental degradation. For example, Throughout history, forced either by necessity or by choice,
livestock-related GHG emissions are .80% of those generated large segments of the worlds population have thrived on plant-
by the food sector and 24% of global GHG emissions. based diets. In the past, meatless diets have been advocated on
the basis of religious, ethical, or philosophical values, not science.
It is only in the past 150 y that empirical evidence has yielded
ISSUES AND CHALLENGES dietary recommendations.
The proposal to drastically reduce meat consumption at the global Will going back to plant-based diets be the best way forward
level is ground-shaking. Some have even branded it a revolu- for a sustainable future? Agricultural and nutritional policies
tionary approach and have argued that wholesale dietary shifts that lead to the adoption of plant-based diets at the global level
may not be realistic (36). However, the proposed transition does not will simultaneously optimize the food supply, health, environ-
need to be an all or nothing process because even only incremental mental, and social justice outcomes for the worlds population.
steps could be extremely helpful in solving food availability and Implementing such policies is not free of political challenges
sustainability challenges. We are fully aware that such a drastic but is perhaps the most rational, scientific, and moral path for
SUSTAINABILITY OF PLANT-BASED DIETS 7S of 7S
a sustainable future of the human race and other living creatures in real-life settings throughout North America. Am J Clin Nutr 2014;
of the biosphere that we share. 100(suppl):490S5S.
24. Credit Suisse Research Institute. Global Wealth Report 2013. October
The authors responsibilities were as followsJS: wrote the first draft of 2013. Available from: https://publications.credit-suisse.com/tasks/render/
the manuscript; and SS: contributed to the content and editing of the man- file/?fileID=BCDB1364-A105-0560-1332EC9100FF5C83 (cited 25
uscript. The authors had no conflicts of interest to declare. November 2013).
25. US Census Bureau. Available from: http://www.census.gov/population/
international/data/idb/worldpoptotal.php (cited 5 October 2013).
26. Saxena AM. The vegetarian imperative. Baltimore, MD: The Johns
REFERENCES Hopkins University Press, 2011.
27. Aiking H. Protein production: planet, profit, plus people? Am J Clin
1. Burlingame B, Dernini S, eds. Sustainable diets and biodiversity: di- Nutr 2014;100(suppl):483S9S.
rections and solutions for policy, research and action. Proceedings of 28. FAO. Food balance sheets. 2009. Available from: http://faostat3.fao.
the International Scientific Symposium. Biodiversity and sustainable
org/home/index.html#DOWNLOAD (cited 5 October 2013).
diets united against hunger; 35 Nov 2010; Rome, Italy. Rome, Italy:
29. Kearney J. Food consumption trends and drivers. Phil Trans R Soc B
Food and Agriculture Organization, 2012.
2010;365:2793807.
2. Fanzo J, Cogill B, Mattei F. Technical brief: metrics of sustainable diets
30. McMichael AJ, Powles JW, Butler CD, Uauy R. Food, livestock pro-
and food systems. Maccarese, Italy: Bioversity International, 2012:18.
duction, energy, climate change, and health. Lancet 2007;370:125363.
3. Rockstrom J, Steffen W, Noone K, Persson A, Chapin FS III, Lambin
31. Solomon S, Qin D, Manning M, Chen Z, Marquis M, Averyt KB, Tignor
EF, Lenton TM, Scheffer M, Folke C, Schellnhuber HJ, et al. A safe
M, Miller HL, eds. Climate Change 2007: The Physical Science Basis.
operating space for humanity. Nature 2009;461:4725.
Contribution of Working Group I to the Fourth Assessment Report of the
4. Altieri MA. Linking ecologists and traditional farmers in the search for
Intergovernmental Panel on Climate Change. New York, NY and Cam-
sustainable agriculture. Front Ecol Environ 2004;2:3542.
bridge, United Kingdom: Cambridge University Press, 2007 [996 pp].
5. Horrigan L, Lawrence RS, Walker P. How sustainable agriculture can
32. Pew Commission on Industrial Farm Animal Production. Putting meat
address the environmental and human health harms of industrial agri-
culture. Environ Health Perspect 2002;110:44556. on the table: industrial farm animal production in America. Available
6. Steinhart CE, Steinhart JS. Energy: sources, use and role in human from: http://www.pewtrusts.org/our_work_report_detail.aspx?id=38442
affairs. North Scituate, MA: Duxbury Press, 1974. (cited 18 February 2013).
7. Center for Sustainable Systems, University of Michigan. U.S. Food 33. Akhtar AZ, Greger M, Ferdowsian H, Frank E. Health professionals
System Factsheet. Publication CSS01-06. Available from: http://css. roles in animal agriculture, climate change, and human health. Am J
snre.umich.edu/css_doc/CSS01-06.pdf (cited 23 April 2013). Prev Med 2009;36:1827.
8. Pimentel D, Pimentel M. Sustainability of meat-based and plant-based 34. Schmidhuber J, Tubiello FN. Global food security under climate
diets and the environment. Am J Clin Nutr 2003;78:660S3S. change. Proc Natl Acad Sci USA 2007;104:197038.
9. Reijnders L, Soret S. Quantification of the environmental impact of 35. Mueller ND, Gerber JS, Johnston M, Ray DK, Foley JA. Closing the yield
different dietary protein choices. Am J Clin Nutr 2003;78:664S8S. gap through nutrient and water management. Nature 2012;490:2547.
10. Gonzalez AD, Frostell B, Carlsson-Kanyama A. Protein efficiency per 36. Foley JA, Ramankutty N, Brauman KA, Cassidy ES, Gerber JS,
unit energy and per unit greenhouse gas emissions: potential contribution Johnston M, Mueller ND, OConnell C, Ray DK, West PC, et al. So-
of diet choices to climate change mitigation. Food Policy 2011;36:56270. lutions for a cultivated planet. Nature 2011;478:33742.
11. Berners-Lee M, Hoolohan C, Cammack H, Hewitt CN. The relative 37. Godfray HCJ, Beddington JR, Crute IR, Haddad L, Lawrence D, Muir
greenhouse gas impacts of realistic dietary choices. Energy Policy JF, Pretty J, Robinson S, Thomas SM, Toulmin C. Food security: the
2012;43:18490. challenge of feeding 9 billion people. Science 2010;327:8128.
12. Steinfeld HGP. Livestocks long shadow: environmental issues and 38. Sabate J. The contribution of vegetarian diets to health and disease:
options. Rome, Italy: Food and Agriculture Organization of the United a paradigm shift? Am J Clin Nutr 2003;78(suppl):502S7S.
Nations, 2006. 39. Rizzo NS, Sabate J, Jaceldo-Siegl K, Fraser GE. Vegetarian dietary
13. Friel S, Danogour AD, Garnett T, Lock K, Chalabi Z, Roberts I, Butler patterns are associated with a lower risk of metabolic syndrome: the
A, Butler CD, Waage J, McMichael AJ, et al. Public health benefits of Adventist Health Study 2. Diabetes Care 2011;34:12257.
strategies to reduce greenhouse-gas emissions: food and agriculture. 40. Orlich MJ, Singh PN, Sabate J, Jaceldo-Siegl K, Fan J, Knutsen S,
Lancet 2009;374:201625. Beeson WL, Gary E, Fraser GE. Vegetarian dietary patterns and mortality
14. Eshel G, Martin PA. Geophysics and nutritional science: toward in Adventist Health Study 2. JAMA Intern Med 2013;173(13):12308.
a novel, unified paradigm. Am J Clin Nutr 2009;89:1710S6S. 41. Tonstad S, Stewart K, Oda K, Batech M, Herring RP, Fraser GE.
15. Marlow HJ, Hayes WK, Soret S, Carter RL, Schwab ER, Sabate J. Diet Vegetarian diets and incidence of diabetes in the Adventist Health
and the environment: does what you eat matter? Am J Clin Nutr 2009; Study-2. Nutr Metab Cardiovasc Dis 2013;23:2929.
89(suppl):1966703S. 42. Tantamango-Bartley Y, Jaceldo-Siegl K, Fan J, Fraser G. Vegetarian
16. Carlsson-Kanyama A, Gonzalez AD. Potential contributions of food diets and the incidence of cancer in a low-risk population. Cancer
consumption patterns to climate change. Am J Clin Nutr 2009;89 Epidemiol Biomarkers Prev 2013;22:28694.
(suppl):1704S9S. 43. American Institute for Cancer Research; World Cancer Research Fund.
17. Baroni L, Cenci L, Tettamanti M, Berati M. Evaluating the environ- Food, nutrition, physical activity and the prevention of cancer: a global
mental impact of various dietary patterns combined with different food perspective. Washington, DC: AICR, 2007.
production systems. Eur J Clin Nutr 2007;61:27986. 44. Sinha R, Cross A, Graubard B, Leitzmann M, Schatzkin A. Meat intake
18. Risku-Norja H, Kurppa S, Helenius J. Dietary choices and greenhouse and mortality. Arch Intern Med 2009;169:56271.
gas emissions assessment of impact of vegetarian and organic options 45. Mozaffarian D, Hao T, Rimm EB, Willett WC, Hu FB. Changes in diet
at national scale. Progr Ind Ecol 2009;6:34054. and lifestyle and long-term weight gain in women and men. N Engl J
19. Eshel G, Martin PA. Diet, energy and global warming. Earth Interact Med 2011;364:2392404.
2006;10:117. 46. deBoer J, Aiking H. On the merits of plant-based proteins for global
20. Macdiarmid JI, Kyle J, Gorgan GW, Loe J, Fyfe C, Johnstone A, McNeill G. food security: marrying macro and micro perspectives. Ecol Econ
Sustainable diets for the future: can we contribute to reducing greenhouse 2011;70:125965.
gas emissions by eating a healthy diet? Am J Clin Nutr 2012;96:6329. 47. Vinnari M, Tapio P. Sustainability of diets: from concepts to gover-
21. Scarborough P, Allender S, Clarke D, Wickramasinghe K, Rayner M. nance. Ecol Econ 2012;74:4654.
Modelling the health impact of environmentally sustainable dietary 48. Clonan A, Holdsworth M. The challenges of eating a healthy and
scenarios in the UK. Eur J Clin Nutr 2012;66:7105. sustainable diet. Am J Clin Nutr 2012;96:45960.
22. Stehfest E, Bouwman L, van Vuuren DP. Climate benefits of changing 49. Goodland R. Environmental sustainability in agriculture: diet matters.
diet. Clim Change 2009;96:83102. Ecol Econ 1997;23:189200.
23. Soret S, Mejia A, Batech M, Jaceldo-Siegl K, Harwatt H, Sabate J. 50. Lang T, Barling D. Nutrition and sustainability: an emerging food
Climate change mitigation and health effects of varied dietary patterns policy discourse. Proc Nutr Soc 2013;72:112.
Dietary Fatty Acids
BRETT WHITE, MD, University of Southern California, Keck School of Medicine, Los Angeles, California
Fatty acids can be divided into four general categories: saturated, monounsaturated, polyun-
saturated, and trans fats. Saturated fatty acids and trans fats are associated with an increased
risk of coronary heart disease. Monounsaturated fatty acids and polyunsaturated fatty acids
are associated with a decreased risk of coronary heart disease, although these associations are
not uniformly supported in the literature. Omega-3 fatty acids, which are a type of polyun-
saturated fatty acid, have been studied as potential therapy for a variety of medical conditions
because of their suspected anti-inflammatory properties. Omega-3 fatty acids have also been
shown to provide some benefit to patients with cystic fibrosis, and may have a protective effect
against dementia. Physicians should counsel patients about the importance of avoiding hydro-
genated oils and foods containing trans fats because of their association with coronary heart
disease in observational studies. (Am Fam Physician. 2009;80(4):345-350, 372. Copyright
2009 American Academy of Family Physicians.)
F
See related editorial atty acids are long-chain hydro- hydrogen atoms is missing because of a double
T
on page 330. carbons that can be separated into bond between two carbon atoms, it is called
Patient information: four categories: saturated, mono- an unsaturated fatty acid. A fatty acid with
T
A handout on trans fats, unsaturated, polyunsaturated, and a single double bond is monounsaturated,
written by the authors of trans fats. More than 20 types of fatty acids whereas a fatty acid with more than one dou-
this article, is provided on
page 372. are found in foods; some of these are listed in ble bond is polyunsaturated (Figure 1). The
Table 1. Sources of fatty acids include fruits, carbon-carbon double bond found in mono-
vegetable oils, seeds, nuts, animal fats, and unsaturated or polyunsaturated fatty acids
fish oils. Essential fatty acids, such as omega-3 can exist in the cis or trans configuration.
fatty acids, serve important cellular functions. When the two hydrogen atoms are on oppo-
They are a necessary part of the human diet site sides of the double bond, the configura-
because the body has no biochemical pathway tion is called trans. When the hydrogen atoms
to produce these molecules on its own. are on the same side of the double bond, the
configuration is called cis (Figure 2).
Biochemical Structure
In saturated fatty acids, the carbon chain has Saturated Fats
the maximum number of hydrogen atoms Studies have shown that consuming satu-
attached to every carbon atom. If a pair of rated fatty acids has a detrimental effect on
Number of Essential
Name carbon atoms Type of fatty acid fatty acid Common sources
Downloaded from the American Family Physician Web site at www.aafp.org/afp. Copyright 2009 American Academy of Family Physicians. For the private, noncommer-
August 15,
cial 2009
use of oneVolume 80, Number 4 site. All other rights reserved.
www.aafp.org/afp American requests. 345
Family Physician
V
individual user of the Web Contact copyrights@aafp.org for copyright questions and/or permission
SORT: KEY RECOMMENDATIONS FOR PRACTICE
Evidence
Clinical recommendation rating References
Increased intake of saturated fats and trans B 4-7, 9, 40, 41 serum lipids by increasing low-density lipo-
fats may increase the risk of coronary heart
disease.
protein (LDL) cholesterol levels.1 There is
Increased intake of monounsaturated and B 4, 7-13
some evidence that short-chain fatty acids
polyunsaturated fats may decrease the risk (fewer than 10 carbon atoms) are less likely
of coronary heart disease. to affect serum cholesterol levels, whereas
Regular omega-3 fatty acid supplementation B 20, 21 longer-chain fatty acids (12, 14, or 16 carbon
may provide some benefits for persons with
atoms) are more likely to raise LDL levels.2
cystic fibrosis and dementia.
One exception to this is stearic acid (18 car-
There is no clear effect of omega-3 fatty B 18
acid consumption on cardiovascular bon atoms), which does not appear to raise
events in persons with, or at high risk of, serum cholesterol levels.3 Increased con-
cardiovascular disease. sumption of saturated fatty acids has also
been associated with an increased risk of
A = consistent, good-quality patient-oriented evidence; B = inconsistent or limited-
quality patient-oriented evidence; C = consensus, disease-oriented evidence, usual coronary heart disease (CHD; Table 2).4-13
practice, expert opinion, or case series. For information about the SORT evidence
rating system, go to http://www.aafp.org/afpsort.xml. Monounsaturated Fats
Several large observational studies have
found an association between an increased
Saturated fatty acid (stearic acid) intake of monounsaturated fatty acids and a decreased
risk of CHD (Table 2).4-13 One large study failed to
H H H H H H H H H H H H H H H H H
| | | | | | |
O
| | | | | | | | | | find a similar association, although it was limited to a
HCCCCCCCCCCCCCCCCCC=O
| | | | | | | | | | | | | | | | | H trial of beta-carotene and alpha-tocopherol in persons
H H H H H H H H H H H H H H H H H who smoke.9 Evidence from controlled clinical stud-
ies has shown that monounsaturated fatty acids favor-
Monounsaturated fatty acid (oleic acid) ably affect a number of risk factors for CHD, including
H H H H H H H H H H H H lowering total and LDL cholesterol levels, protecting
| | | | | | | | | | | | H H H against thrombogenesis, reducing LDL susceptibility
HCCCCCCCCC=C C C | | | H H
| | | | | | | |
| | C C C
| |
to oxidation, and producing a more favorable glycemic
| | C C
H H H H H H H H H H C = O
H H
| |
| profile.14
H H
O
H
H
Polyunsaturated Fats
Polyunsaturated fatty acid Polyunsaturated fatty acids, which include omega-3 fatty
(linolenic acidan omega-3 fatty acid)
acids, have been studied extensively for their effect on
H H H H H H H several diseases (Table 3).15-25 Omega-3 fatty acids have
H H| | | | | | | | H H
| | H H H
|
C= CCC=CC C= C C | | | H H H
been shown to provide some benefit to patients with
C
H C| | | | | C
| | C C C
| |
| cystic fibrosis, and may have a protective effect against
H H H H
H H
| | C C
C = O
H H H
| |
| dementia.20,21 Omega-3 fatty acids are thought to be ben-
H H
O
C=C C=C
|
346 American Family Physician www.aafp.org/afp Volume 80, Number 4 V August 15, 2009
Dietary Fatty Acids
Table 2. Fatty Acids and Coronary Heart Disease
Years of
Study Number of patients Study type follow-up Findings
Ireland-Boston Diet- 1,001 men Prospective 20 Patients who died of CHD had a higher intake of
Heart Study 6 cohort saturated fatty acids and cholesterol
Nurses Health Study 7 80,082 women Prospective 14 Positive association between percent of energy
cohort intake from saturated fatty acids and increased
risk of CHD
Nurses Health Study 7 80,082 women Prospective 14 Association between increased energy intake from
cohort monounsaturated fatty acids and a relative risk
reduction in coronary disease
Alpha-Tocopherol, 21,930 men who Prospective 6.1 No association between monounsaturated fatty
Beta-Carotene Cancer smoke cohort acids and the risk of CHD
Prevention Study 9
Los Angeles Veteran 424 men RCT 8 Increased consumption of polyunsaturated fatty
Hospital 11 acids associated with reduction of CHD
Oslo Diet-Heart Study 12 206 men with history RCT 5 Increased energy from polyunsaturated fatty
of myocardial acids associated with significant reduction in
infarction cholesterol and coronary events
CHD = coronary heart disease; NA = not applicable; RCT = randomized controlled trial.
Information from references 4 through 13.
association between increased omega-3 fatty acid intake risk of cardiovascular disease.18 The authors concluded
and reduced cancer incidence.31 that increased consumption of omega-3 fatty acids did
Various studies 4-13 and a report from the Institute of not significantly alter total mortality or combined car-
Medicine32 support the increased consumption of polyun- diovascular events in persons with cardiovascular dis-
saturated fatty acid for the prevention of CHD (Table 2 4-13). ease, in persons at high risk of cardiovascular disease,
However, a Cochrane review of 48 randomized controlled or in the general population.18 Excessive consumption of
trials (RCTs) and 41 cohort analyses suggests that fur- polyunsaturated fats is not recommended because of the
ther high-quality trials are needed to confirm the protec- increased risk of excessive weight gain and the increased
tive effect of omega-3 fatty acids in persons at increased risk of gallstone formation in some persons.33
August 15, 2009 V Volume 80, Number 4 www.aafp.org/afp American Family Physician 347
Table 3. Cochrane Reviews of Omega-3 Fatty Acids
Asthma Little evidence to recommend persons with asthma modify intake of omega-3 fatty acids15
Cancer cachexia Insufficient data16
Claudication Omega-3 has limited benefit in persons with intermittent claudication, but no evidence of consistent
improved clinical outcomes17
Coronary heart disease No clear effect of omega-3 fatty acid consumption on cardiovascular events in persons with, or at high
risk of, cardiovascular disease18
Crohn disease Insufficient data to recommend routine use of omega-3 supplements for maintenance of remission19
Cystic fibrosis Regular omega-3 supplements may provide some benefits for persons with cystic fibrosis 20
Dementia May be a protective effect of omega-3 fatty acids against dementia 21
Diabetes (type 2) Omega-3 supplementation lowers triglyceride and very-low-density lipoprotein levels in patients with
diabetes, but may raise low-density lipoprotein levels; no effect on glycemic control or fasting insulin 22
Kidney transplant recipients Insufficient evidence to recommend fish oil therapy 23
Schizophrenia Inconclusive results; use of omega-3 remains experimental 24
Ulcerative colitis No evidence to support omega-3 fatty acids in maintenance of remission 25
Butter 1 tablespoon 11 7 35 0 7 30 10
Cake (pound cake) 1 slice (80 g) 16 3.5 18 4.5 8 0 0
Candy bar 1 (40 g) 10 4 20 3 7 <5 1
Cookies (cream filled) 3 (30 g) 6 1 5 2 3 0 0
Doughnut 1 18 4.5 23 5 9.5 25 8
French fried potatoes Medium size 27 7 35 8 15 0 0
(fast food) (147 g)
Margarine, stick 1 tablespoon 11 2 10 3 5 0 0
Margarine, tub 1 tablespoon 7 1 5 0.5 1.5 0 0
Mayonnaise (soybean 1 tablespoon 11 1.5 8 0 1.5 5 2
oil)||
Milk, skim 1 cup 0 0 0 0 0 5 2
Milk, whole 1 cup 7 4.5 23 0 4.5 35 12
Potato chips Small bag 11 2 10 3 5 0 0
(42.5 g)
Shortening 1 tablespoon 13 3.5 18 4 7.5 0 0
*Nutrient values rounded based on the U.S. Food and Drug Administration (FDA) nutrition labeling regulations.
From the FDA Table of Trans Values, January 20, 1995.
1995 U.S. Department of Agriculture (USDA) composition data.
From the 2002 USDA National Nutrient Database for Standard Reference, Release 15.
||Prerelease values derived from the 2003 USDA National Nutrient Database for Standard Reference, Release 16.
Adapted from the U.S. Food and Drug Administration. Revealing trans fats. FDA Consumer Magazine. SeptemberOctober 2003. Publication no.
FDA051329C. http://www.fda.gov/FDAC/features/2003/503_fats.html. Accessed March 11, 2009.
Trans Fats remain stable at room temperature. These fats can then
Most trans fats are manufactured through a process be incorporated into certain food products (e.g., cook-
called hydrogenation, which is the artificial addition of ies, chips) to increase shelf life (Table 434). Until recently,
hydrogen atoms to unsaturated oils. Hydrogenation con- trans fats have been the predominant fat used in most
verts liquid vegetable oils to solid or semi-solid fats that types of commercial baked goods.
348 American Family Physician www.aafp.org/afp Volume 80, Number 4 V August 15, 2009
Dietary Fatty Acids
August 15, 2009 V Volume 80, Number 4 www.aafp.org/afp American Family Physician 349
Annals of Internal Medicine Perspective
Cholesterol Control Beyond the Clinic: New York Citys Trans Fat
Restriction
Sonia Y. Angell, MD, MPH; Lynn Dee Silver, MD, MPH; Gail P. Goldstein, MPH; Christine M. Johnson, MBA; Deborah R. Deitcher, MPH;
Thomas R. Frieden, MD, MPH; and Mary T. Bassett, MD, MPH
Decades after key modifiable risk factors were identified, cardiovas- vember 2008, the restriction was in full effect in all New York City
cular disease remains the leading cause of preventable death, and restaurants and estimated restaurant use of artificial trans fat for
only one quarter of persons with high cholesterol levels have at- frying, baking, or cooking or in spreads had decreased from 50%
tained recommended levels of control. Cholesterol control efforts to less than 2%. Preliminary analyses suggest that replacement of
have focused on consumer education and medical treatment. A artificial trans fat has resulted in products with more healthful fatty
powerful, complementary approach is to change the makeup of acid profiles. For example, in major restaurant chains, total satu-
food, a route the New York City Department of Health and Mental rated fat plus trans fat in French fries decreased by more than
Hygiene took when it restricted artificial trans fata contributor to 50%. At 2 years, dozens of national chains had removed artificial
coronary heart diseasein restaurants. The Department first under- trans fat, and 13 jurisdictions, including California, had adopted
took a voluntary campaign, but this effort did not decrease the similar laws. Public health efforts that change food content to make
proportion of restaurants that used artificial trans fat. In December default choices healthier enable consumers to more successfully
2006, the Board of Health required that artificial trans fat be meet dietary recommendations and reduce their cardiovascular risk.
phased out of restaurant food. To support implementation, the Ann Intern Med. 2009;151:129-134. www.annals.org
Department provided technical assistance to restaurants. By No- For author affiliations, see end of text.
hydrogenation of vegetable oils and found mostly in pro- the city, 15 000 suppliers and supermarkets, and hundreds
cessed foods especially baked and fried goods and of thousands of consumers. Mailings urged restaurants to
spreads. Partial hydrogenation was patented in 1903. Eight remove artificial trans fat from food, suppliers to promote
years later, Crisco (J.M. Smucker, Orrville, Ohio) became 0-grams trans fat products, and patrons to inquire about
the first such product widely marketed in the United States oils used. Press coverage was extensive (2729). More than
(17). These chemically modified oils, like several food in- 7800 restaurant operators received training on the issue
gredients in the 20th century, entered the food supply during the mandatory food protection course.
without a full evaluation of human health effects and were Nearly a year later, inspectors surveyed 1021 restau-
generally recognized as safe on the basis of experience rants in April and May 2006 by following the same meth-
(18). Industrial advantages, such as extended product shelf odology as the previous year. Despite the educational cam-
life and fry oil stability, led to increased use in the 1950s. paign, prevalence of artificial trans fat use remained
In the 1970s, margarine made from partially hydrogenated virtually unchanged at 51% in places where it could be
oil was promoted as a healthier alternative to butter (19). assessed. Given survey findings of a wide range of restau-
However, by the 1990s, studies began to identify trans fat rants that cooked without artificial trans fat, successful re-
as a health hazard (9). formulation of packaged foods, and the Danish experience,
In 2003, the U.S. Food and Drug Administration the Department concluded that food choice and quality
(FDA) required that all packaged goods display informa- would be unaffected by replacing artificial trans fat and
tion on trans fat content by 2006 (20). The regulation proposed using its regulatory authority to reduce artificial
allows products that contain less than 0.5 grams of trans fat trans fat use (26).
per serving to be labeled zero grams trans fat. This
prompted a widespread industry response of reformulating
products to declare zero grams trans fat on front-of- CRAFTING REGULATION
package labeling (2123), although such products may still Department clinical staff, food safety experts, and law-
contain trans fat in small amounts, which can add up. yers considered a range of regulatory options, from
The nutrition facts labeling allowed attentive consum- disclosure-only requirements to content restrictions. Label-
ers to avoid trans fat in packaged foods; however, ing trans fat would benefit only consumers who were aware
restaurantswhich provide one third of daily caloric in- of and would act on the dangers of trans fat and would be
take (24) generally do not disclose nutritional informa- difficult to enforce. A restriction, however, would benefit
tion, leaving consumers no practical way to avoid trans fat. all New Yorkers, regardless of language, literacy, level of
At the time of the FDA action, only Denmark limited health awareness, or age. Protecting children, who con-
trans fat in its entire food supply (25). sume French fries and other quick-service foods that are
Although local health departments do not commonly common sources of artificial trans fat (30), was particularly
regulate packaged goods, inspecting, licensing, and regulat- important. Because artificial trans fat is both harmful and
ing restaurants and retail food outlets are core local public fully replaceable, allowing continued use, even with disclo-
health functions (26). Building on its food safety infra- sure, could not be justified. Similarly, restricting trans fat
structure, the Department trained restaurant inspectors in in cooking oils and spreads onlyalthough easier to
May 2005 to assess the presence of artificial trans fat in achievewould be insufficient because baked goods are
products used for frying, baking, or cooking or in spreads the largest dietary source of trans fat (16). Given that an
by means of food label inspection. Because inspection visits estimated one third of calories comes from away-from-
are scheduled each year on the basis of original restaurant home sources, removing most trans fat from restaurants
licensing date and compliance records, a survey fielded could be expected to reduce trans fatassociated deaths
during regularly scheduled inspections approximates a ran- proportionately. The Department drafted a broad proposal
dom sample of these establishments. In May and June to restrict artificial trans fat use in fry oils; spreads; and all
2005, inspectors assessed the use of trans fat in oils, short- other ingredients and products, including baked goods.
enings, and spreads used for frying, baking, or cooking or For the regulation to be clear and enforceable, the
in spreads in 529 restaurants. Of the 478 restaurants that Department adopted the FDA threshold used for labeling
used fats for any of these purposes and for which complete of up to 0.5 grams of trans fat per serving for products
data were collected, trans fat presence could not be deter- containing any trans fat (20). The regulation would apply
mined in one third. Where it could be assessed, 50% of to all licensed food establishments, including restaurants,
restaurants used artificial trans fat to prepare food. school cafeterias, caterers, senior centers, and street-food
vendors. It intentionally did not affect products that con-
tained only natural trans fat.
EDUCATIONAL CAMPAIGN The Department designed enforcement to rely on ex-
In August 2005, the Department launched an educa- isting infrastructure. Inspectors check product labels,
tional campaign to reduce restaurant artificial trans fat use. which restaurants now must maintain, during their regular
Information was sent to the 30 000 licensed food outlets in annual inspections. The rule included a 6-month phase-
130 21 July 2009 Annals of Internal Medicine Volume 151 Number 2 www.annals.org
Cholesterol Control Beyond the Clinic Perspective
Table 1. Resources Created to Help Restaurants Comply With the Artificial Trans Fat Restriction
* Includes lists of replacement products classified by saturated fat content; general price lists for 0-grams trans fat products; information on how to comply with the citys
regulation; pages on frying, baking, and purchasing prepared products with 0-grams trans fat; and a baking troubleshooter guide for working with 0-grams trans fat products.
The site is offered in English, Spanish, and Chinese.
out period for artificial trans fat in fry oils and spreads and RESTAURANT RESPONSE
18 months for all other items, to give restaurants time to By June 2008, on the basis of inspection reports, 99%
identify alternatives and reformulate (31). Fines for viola- of all restaurants had successfully adapted to the fry oil and
tions range from $200 to $2000. spread restriction. The second phase, which covered all
other products, became effective on 1 July 2008. Nearly 6
FROM CONCEPT TO ACTION
months later, adherence was 92%, with most violations
The regulation was proposed as an amendment to the due to missing documentation. When November 2008 ad-
citys Health Code, approvable by the Board of Health. herence figures are adjusted to provide approximate com-
Nearly 2300 public comments were received, with sup- parison with preregulation survey denominators, an esti-
porters outnumbering opponents 31 to 1. Local, regional, mated 98% of restaurants were not using artificial trans fat
and national organizations and professional medical asso- in oils, shortenings, and spreads, compared with 50% in
ciations made statements of support. 2005 (Figure).
Opposition came primarily from industry associations, As implementation proceeded, the Department fol-
including the National Restaurant Association and its state lowed the use of saturated fat in replacement oils. A num-
affiliate and the Grocery Manufacturers Association (32). ber of chains announced early that replacement of trans
The restaurant industry, which often opposes regulation on fat containing fry oils would also reduce saturated fat by
principle, alleged that the elimination of artificial trans fat 20% to 35% (35) (Colon R. Personal communication, 14
would result in more expensive and less flavorful food and June 2007). Leading fast food chains decreased the satu-
also claimed that the supply of alternative oils was insuffi-
cient and the regulatory timeline unreasonable. Some
media derided the proposal as nanny state meddling (33, Figure. New York City restriction of artificial trans fat:
34). reduction in use in frying, baking, or cooking or in spreads.
In response to public comment, the Department mod-
ified the rule to allow a longer phase-out period for baked
Containing Oils, Shortenings, or Spreads, %
Table 2. Trans and Saturated Fat Content of French Fries Sold in Major U.S. Fast Food Chains*
McDonalds
Small 2.5 3.5 6.0 1.5 0.0 1.5 !1.0 !3.5 !4.5
Medium 4.0 5.0 9.0 2.5 0.0 2.5 !1.5 !5.0 !6.5
Large 6.0 8.0 14.0 3.5 0.0 3.5 !2.5 !8.0 !10.5
Arbys
Small 4.0 3.0 7.0 4.0 0.0 4.0 0.0 !3.0 !3.0
Medium 4.0 4.0 8.0 4.0 0.0 4.0 0.0 !4.0 !4.0
Large 7.0 5.0 13.0 7.0 1.0 8.0 0.0 !4.0 !4.0
White Castle
Regular, 4.12 oz 3.0 5.0 8.0 3.0 0.0 3.0 0.0 !5.0 !5.0
Sack, 8.6 oz 6.0 11.0 17.0 6.0 0.0 6.0 0.0 !11.0 !11.0
Mean grams 4.0 4.0 7.9 3.5 0.1 3.6 !0.4 !3.9 !4.3
Mean gram change, % !10.5 !97.9 !54.2
* Analysis restricted to French fries to illustrate the effect of the first phase of the trans fat regulation, which covered fry oils and spreads. The second phase covered all other
products and went into effect on 1 July 2008.
Values are based on nutrition information published online at each chains Web site. Access dates: Wendys, 1 December 2006 (before) and 1 April 2008 (after);
McDonalds, 25 January 2007 (before) and 22 May 2008 (after); Arbys, 6 February 2007 (before) and 23 May 2008 (after); and White Castle, 20 February 2007 (before)
and 25 August 2008 (after).
rated fat in their French fries by 10.5% and the total trans this complicated topic, harness political and community
fat plus saturated fat by 54%, on average (Table 2). Recent support for a regulatory approach, and navigate industry
experience in Canada (21) also suggests that saturated fat opposition while addressing legitimate industry concerns.
content can remain the same or even be reduced when Terms with which the public is unfamiliar, such as partially
trans fat is reduced. In New York City, the evaluation of hydrogenated vegetable oil and trans fatty acids, were aban-
replacement products and changes in fatty acid composi- doned in favor of the more accessible term artificial trans
tion are under way, and surveillance of cardiovascular dis- fat. Although the voluntary initiative did not result in mea-
ease is ongoing. However, as with other population inter- surable changes, it generated public and industry awareness
ventions, isolating the specific effect of this regulation on and demonstrated that a voluntary effort was insufficient to
low-density lipoprotein cholesterol levels and cardiovascu- eliminate the health risk. The restaurant industry was sup-
lar health is challenging. portive of the voluntary initiative; however, it organized in
Two years after New York Citys action, 12 U.S. local opposition when the regulation was proposed (32) even
governments and 1 state had adopted similar measures. though the Health Code amendment was clearly within
Following the approval of the Health Code amendment, the legal purview of the Board of Health and went unchal-
more than 50 restaurant chains announced or reiterated lenged in court. To address replacement product availabil-
their intention to discontinue use of artificial trans fat ity and cost concerns and to reduce the risk for price goug-
containing products nationally, and most have already ing at the retail level, the Department identified food
done so (37). Dozens of hotel groups, food manufacturers, suppliers and manufacturers and tracked availability and
grocery retailers, and other companiesfrom Disney and price of these products. Restaurants that switched to newer
Marriott to American Airlines and Kraft Foods have also fry oils with 0 grams of trans fat reported that the change
announced elimination of artificial trans fat. was cost neutral (38). Supply concerns proved unfounded
because major oil producers expanded production of oils
with 0 grams of trans fat (22, 39, 40) and ingredients for
REFLECTIONS ON THE EXPERIENCE baked goods with reduced levels of saturated fat (41). Al-
To develop and implement the trans fat regulation, though evaluation continues, no negative economic effects
Department staff needed to communicate effectively about have been observed to date.
132 21 July 2009 Annals of Internal Medicine Volume 151 Number 2 www.annals.org
Cholesterol Control Beyond the Clinic Perspective
PUBLIC HEALTH ACTION CHANGES THE DEFAULT Current author addresses are available at www.annals.org.
Reducing risk factors for cardiovascular disease is dif-
ficult for patients and physicians. Artificial trans fata
recognized contributor to coronary heart disease riskis References
1. Carroll MD, Lacher DA, Sorlie PD, Cleeman JI, Gordon DJ, Wolz M, et al.
widespread in our food supply and, at the time of the
Trends in serum lipids and lipoproteins of adults, 1960-2002. JAMA. 2005;294:
Departments action, was virtually impossible for the aver- 1773-81. [PMID: 16219880]
age consumer to avoid. An effective way to reduce this risk 2. U.S. Department of Health and Human Services. Nutrition and overweight.
was to change the food supply through regulation. None- In: Healthy People 2010 Midcourse Review. Washington, DC: U.S. Department
of Health and Human Services; 2005:1-25.
theless, proposing regulation as a tool to protect health
3. Panel on Macronutrients. Letter report on dietary reference intakes for trans
prompts industry opposition and complaints about govern- fatty acids. In: Institute of Medicine. Dietary Reference Intakes for Energy, Car-
ment intrusion. This opposition can often delay important bohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids. Wash-
policy actions and lead to preventable disease. For example, ington, DC: Institute of Medicine; 2002.
4. Hyre AD, Muntner P, Menke A, Raggi P, He J. Trends in ATP-III-defined
it took decades to achieve a national prohibition of lead
high blood cholesterol prevalence, awareness, treatment and control among U.S.
paint, despite unequivocal evidence for harm. New York adults. Ann Epidemiol. 2007;17:548-55. [PMID: 17395483]
Citys Board of Health banned the product in 1960, 18 5. Mann D, Reynolds K, Smith D, Muntner P. Trends in statin use and
years before federal action. low-density lipoprotein cholesterol levels among U.S. adults: impact of the 2001
National Cholesterol Education Program guidelines. Ann Pharmacother. 2008;
Throughout the Health Code amendment process, de-
42:1208-15. [PMID: 18648016]
tractors likened the city to a nanny state (33, 34), as they 6. Brunner EJ, Rees K, Ward K, Burke M, Thorogood M. Dietary advice for
had with such previously controversial measures as seat belt reducing cardiovascular risk. Cochrane Database Syst Rev. 2007:CD002128.
requirements, water fluoridation, and tobacco control. Just [PMID: 17943768]
as the debates surrounding those historic public health in- 7. Beran MS, Fowles JB, Kind EA, Craft CE. State of the art reviews: patient
and physician communication about weight management: can we close the gap?
terventions calmed after they were implemented, trans fat American Journal of Lifestyle Medicine 2008;2:75-83.
restriction is now a largely unnoticed part of New York 8. Mozaffarian D, Katan MB, Ascherio A, Stampfer MJ, Willett WC. Trans
City life. By the time the second phase went into effect, fatty acids and cardiovascular disease. N Engl J Med. 2006;354:1601-13. [PMID:
press coverage was light-heartednoting that popular 16611951]
9. Ascherio A, Katan MB, Zock PL, Stampfer MJ, Willett WC. Trans fatty
foods, from cannoli to knishes, were still available (42). acids and coronary heart disease. N Engl J Med. 1999;340:1994-8. [PMID:
Seventy-nine percent of trans fat in the diet is artifi- 10379026]
cially produced (16). It is present only because we fail to 10. Lopez-Garcia E, Schulze MB, Meigs JB, Manson JE, Rifai N, Stampfer MJ,
prevent industry from using it, even as we recommend that et al. Consumption of trans fatty acids is related to plasma biomarkers of inflam-
mation and endothelial dysfunction. J Nutr. 2005;135:562-6. [PMID:
consumers avoid it. Removing trans fat from the food sup- 15735094]
ply will improve the lipid profiles of millions of persons 11. Willett WC, Skerrett PJ, Giovannucci EL, Callahan M. Eat, Drink, and Be
without requiring complex behavioral efforts and may re- Healthy: The Harvard Medical School Guide to Healthy Eating. New York:
duce the need for medication. Although regulatory ap- Simon and Schuster; 2001.
12. Lichtenstein AH, Appel LJ, Brands M, Carnethon M, Daniels S, Franch
proaches are not appropriate for all health risks, our society HA, et al; American Heart Association Nutrition Committee. Diet and lifestyle
could improve health if available food choices more closely recommendations revision 2006: a scientific statement from the American Heart
resembled those recommended for heart health (43). Association Nutrition Committee. Circulation. 2006;114:82-96. [PMID:
16785338]
From the New York City Department of Health and Mental Hygiene, 13. U.S. Department of Health and Human Services, U.S. Department of
New York, New York. Agriculture. Dietary Guidelines for Americans. Washington, DC: U.S. Depart-
ment of Health and Human Services; 2005.
Acknowledgment: The authors thank Elliott Marcus, MSSW, and Rob- 14. American Medical Association House of Delegates. Reducing Trans Fat
Content of Americans Diet. Chicago: American Med Assoc; 2006.
ert Edman, BSc, of the Departments Bureau of Food Safety and Com-
15. World Health Organization, Food and Agriculture Organization of the
munity Sanitation, as well as all the Bureaus directors and sanitarians for
United Nations. Joint WHO/FAO Expert Consultation on diet, nutrition and
implementation of the work described here. They also thank Wilfredo the prevention of chronic diseases. Section Report of a joint WHO/FAO expert
Lopez, JD; Thomas Merrill, JD; Martha Robinson, JD; and Anna Caf- consultation. Geneva, Switzerland: World Health Organization; 2003.
farelli, MHS; the members of the New York City Board of Health; 16. Revealing trans fats. FDA Consum. 2003;37:20-6. [PMID: 14666902]
Walter Willett, MD, MPH, DrPH; Colin McCord, MD; and the other 17. Allen RR. Hydrogenation. In: Hui YH, ed. Baileys Industrial Oil Fat Prod-
staff and interns of the Department whose work was critical to the suc- ucts, Edible Oil and Fat Products: Processing Technology. 5th ed. New York: J
cess of this initiative. Wiley; 1996:213-6.
18. U.S. Food and Drug Administration. History of the GRAS list and SCOGS
Grant Support: By the City of New York. The American Heart As- Review. Silver Spring, MD: U.S. Food and Drug Administration; 2006. Accessed at
sociation provided an educational grant to support the Trans Fat www.fda.gov/Food/FoodIngredientsPackaging/GenerallyRecognizedasSafeGRAS
Help Center. /GRASSubstancesSCOGSDatabase/ucm084142.htm on 2 June 2009.
19. Chrysam MM. Margarine and spreads. In: Hui YH, ed. Baileys Industrial
Oil and Fat Products, Edible Oil and Fat Products: Products and Application
Potential Financial Conflicts of Interest: None disclosed.
Technology. 5th ed. New York: J Wiley; 1996:65-114.
20. U.S. Food and Drug Administration. Food Labeling: Trans Fatty Acids in
Requests for Single Reprints: Lynn Dee Silver, MD, MPH, New York Nutrition Labeling, Nutrient Content Claims, and Health Claims and Possible
City Health Department, 2 Lafayette Street, 20th Floor, CN-46, New Footnote or Disclosure Statements. Federal Register. Washington, DC: U.S. De-
York, NY 10007; e-mail, lsilver@health.nyc.gov. partment of Health and Human Services; 11 July 2003:41434-506.
www.annals.org 21 July 2009 Annals of Internal Medicine Volume 151 Number 2 133
Nutrition & Dietetics 2012; 69: 256259 DOI: 10.1111/j.1747-0080.2012.01600.x
INSIGHT
Bill SHRAPNEL
Shrapnel Nutrition Consulting Pty Ltd, Sydney, New South Wales, Australia
Abstract
Aim: The aim of this paper is to critically assess recent calls for increased regulation to lower the level of trans fats
in the Australian diet.
Methods: Key milestones in the elucidation of the effects on trans fats on health were identified and reviewed.
Trends in intakes of trans fats in Australia and factors affecting those trends were described and compared with
those in Denmark, which has regulated to lower population intake of trans fats.
Results: The scientific evidence demonstrating adverse effects of trans fats on human health is consistent and
strong. Australian health authorities were quick to identify the potential risk of trans fats and communicate it to
health professionals and the food industry. The response from the margarine industry resulted in large falls in the
trans fat content of the Australian diet in the mid-1990s. A second wave of trans fat reduction across many foods
categories has occurred subsequently. Total intake of trans fats in Australia is now low, half the upper limit
recommended by the World Health Organization and lower than the intake in Denmark. Trans fats of industrial origin
comprise just one-eighth of 1% of dietary energy. These falls in trans fat intake mirror the large falls that occurred
in Denmark prior to regulation.
Conclusions: The case for increased regulation to lower intake of trans fats in Australia cannot be sustained. The
trans fat issue stands as a good example of self-regulation through collaboration between Australian health
agencies, the food industry and the government.
Introduction There are two sources of trans fats in the human diet,
industrial and ruminant, and both are the result of a process
In recent years, there have been calls from Australian called partial hydrogenation.7 This process can be applied
politicians and former politicians for a ban on trans-fatty industrially to harden a vegetable oil, that is, raise the
acids (trans fats),1,2 echoing similar calls overseas that have melting point of an unsaturated oil by generating increased
resulted in changes to food regulations in Denmark3 and the proportions of saturated and trans fats. Until the mid-1990s,
USA.4,5 Is there a sound case for regulatory intervention to partial hydrogenation was used by the margarine industry in
drive down the level of trans fats in the Australian diet or Australia to generate a hard fraction, which when mixed
does current progress obviate the need for regulation? with liquid vegetable oils yielded a solid product suitable
Unsaturated fatty acids comprise approximately half of for spreading. Hardened fats were also used in the baking
the fatty acids in the typical adult Australian diet6 and their industry where fats with comparable physical properties to
carboncarbon double bonds are generally in the cis con- butter were required. Trans fats are also generated in the
figuration. This conformation ensures that unsaturated fats gastrointestinal tracts of ruminant animals, such as cattle,
have lower melting points than fats dominated by saturated sheep and goats.8 Once ingested, unsaturated fats from grass
fatty acids. In a small proportion of unsaturated fatty acids, and seeds are retained in the rumen of the animal where they
the carboncarbon double bonds are in the trans configura- undergo partial hydrogenation, increasing the proportion of
tion and the physical properties of these trans fats are similar saturated and trans fats. Hence, the meat and milk of rumi-
to those of saturated fats.7 nant animals contain trans fats, which comprise about 35%
of total fatty acids in these foods.
B. Shrapnel, MHP, APD, Director
In the popular media, it has become common to refer to
Correspondence: B. Shrapnel, Shrapnel Nutrition Consulting Pty industrially generated trans fats as artificial and, less com-
Ltd, 10 Blackwood Close, Beecroft, Sydney, NSW 2119, Australia. monly, for ruminant trans fats to be described as natural.
Email: shrapnelnc@bigpond.com This nomenclature is misleading as many 18:1 trans isomers
Industry links: Bill Shrapnel consults to the food industry and current are common to both industrial and ruminant trans fats.7
clients include Goodman Fielder. For example, elaidic acid, with the trans bond in the 9 posi-
Accepted September 2011 tion, which is prevalent in industrially hardened fats and
sometimes singled out as a focus of concern, is also present There have been three estimations of the content of trans
in ruminant fats. Vaccenic acid, with the trans bond in the 11 fats in the Australian diet. In 1994, Noakes and Nestel
position, which is the most common ruminant trans isomer, estimated that trans fats contributed 2.02.5% of the dietary
is also present in industrially hardened fats.7 Industrial and energy of Australian adults.16 Two estimates of trans fats in
ruminant trans fats largely represent different combinations the US diet in the early 1990s indicated that trans fat intake
of the same molecules. in that country was twice17 or three times18 higher than that
in Australia. The high intakes of trans fats in the USA at the
Methods time could be attributed to the widespread use of partially
hydrogenated soybean oil as a frying medium. In contrast,
In relation to the health effects of trans fats, a computerised
the major frying fats used in Australia were tallow (beef and
Medline search was conducted of English language papers
lamb fat), which is low in trans fats, and palm oil, which is
published after 1985, using the following search terms: trans
virtually free of trans fats. Noakes and Nestel estimated that
fatty acids, serum lipids, coronary heart disease and cardio-
approximately 60% of trans fats in the Australian diet were of
vascular disease. Key milestones in the elucidation of the
industrial origin, with domestic margarines being by far the
effects on trans fats on health were identified and briefly
major source.16 FSANZ estimated the trans fat content of
reviewed. Data on trans fat levels in the Australian diet
the Australian and New Zealand diets in 2007 and 2009 in
were accessed through the Food Standards Australia New
response to a request from The Australia and New Zealand
Zealand (FSANZ) website and through manual literature
Food Regulation Ministerial Council.19,20 In the more recent
searches. Information about the chronology of changes to
study,20 mean trans fat intake in Australia was estimated to be
the criteria for the Tick programme was gathered from the
0.5% of dietary energy, half the upper limit recommended by
Heart Foundation via personal communications. Details
the World Health Organization.21 The fall in trans fat intake
about the extent and timing of industry initiatives to lower
in Australia over the preceding 15 years appears to have
the trans-fatty acid content of products were gathered from
been significant and, based on the differences between the
food companies through personal communications. Data on
two FSANZ studies, was still underway in the late 2000s.
trans fat levels in the Danish diet were gathered via comput-
Falls in intakes of both industrial and ruminant trans fats
erised literature searches using the names of known Danish
have occurred since the early 1990s, the fall in ruminant
researchers into trans fats.
trans fats most likely being due to lower use of tallow and
increased use of palm oil as a frying medium and in baked
Results and Discussion
goods. The fall in intake of industrial trans fats has been
Investigations of the health effects of trans fats in the 1980s significantly greater as the use of partially hydrogenated fats
found that their effect on total blood cholesterol was almost has been phased down or out in many product categories.
neutral, similar to that of monounsaturated fats, and there Industrial trans fats now comprise just one-eighth of 1% of
appeared to be little cause for concern.9 As understanding of dietary energy. Ruminant sources contribute 73% of total
blood lipid-related coronary heart disease risk became more trans fat in the current Australian diet.20
sophisticated, the focus shifted from total blood cholesterol Two factors appear to have been instrumental in the
to the different lipoprotein fractions, in particular the detri- appreciable falls in the trans fat content of the Australian
mental LDL-cholesterol (LDL-C) and the protective HDL- dietscientific leadership and the response from the food
cholesterol (HDL-C). When the health effects of trans fats industry. Scientific leadership on the issue of trans fats has
were assessed through this prism, their unique double- come from two quarters. In 1992, just two years after the
negative effect on blood lipids was revealedtrans fats adverse effects of trans fats on serum lipoproteins were
raised the LDL-C concentration and lowered HDL-C.10 In first observed,10 the National Health and Medical Research
their 2003 meta-analysis, Mensink and colleagues found that Council released a report recommending that trans fats be
trans fats raised the total cholesterol to HDL-C ratio signifi- considered as equivalent to saturated fats in relation to their
cantly more than saturated fat, suggesting trans fats con- effect on blood cholesterol.22 Two years later, the National
ferred higher coronary heart disease risk.11 Adverse effects on Heart Foundation of Australia argued a similar case in a
this ratio have been observed with both industrial and rumi- review of dietary fats and cholesterol and forecast likely
nant trans fats, as well as conjugated linoleic acida poly- changes to relevant criteria for the Foundations Tick
unsaturated trans-fatty acid found predominantly in dairy programme.23
fat.12 The possibility that dietary trans fats may pose greater After extensive consultation with the margarine industry,
coronary heart disease risk than saturated fat was supported the Heart Foundation introduced a new Tick programme
by the findings of prospective cohort studies.13,14 The mag- criterion for table margarines in early 1996, the previous
nitude of coronary heart disease risk associated with dietary limit on saturated fats being replaced by a limit on the sum
trans fats in these studies suggests that additional mecha- of saturated and trans fats. In that year, one of Australias
nisms may also be at play, for example, adverse effects on leading margarine manufacturers removed trans fats from all
inflammation, insulin resistance and blood vessel function.15 its table spreads as part of a global initiative by the parent
Nutrition scientists now generally agree that dietary trans fats company24 and the other leading margarine company largely
pose a greater coronary heart disease risk than saturated fat followed in 1997, resulting in a considerable fall in the
and other macronutrients. amount of trans fat present in the Australian food supply.
Insight into the effects of this change on individual intakes of ing regulation and the other adopting non-regulatory mea-
trans fats was gleaned from a serendipitous study conducted sures. It can be concluded that regulation has played a very
at the Commonwealth Scientific and Industrial Organisation minor role in the Danish efforts to lower trans fat intakes and
in Adelaide.25 Adipose tissue biopsies of fatty acids were would have a negligible effect if enacted in Australia. Given
collected from two groups of subjects between 1995 and the successful strategies employed by the scientific commu-
1997, which coincided with the first margarine companys nity, the food industry and government to lower trans fats in
removal of trans fats from its products. In both groups of the Australian diet to their current low level, the case for
subjects, adipose tissue levels of trans fats fell by 23% during regulatory intervention in relation to trans fats is lacking.
the study, suggesting a substantial fall in dietary intake of In summary, although there is consistent evidence that
trans fats. In 2005, the Heart Foundation again amended dietary trans fats are associated with adverse effects on blood
its Tick programme criteria for margarines, stipulating that lipids and increased risk for coronary heart disease, the case
trans fats should comprise 1% or less of total fatty acids. This for increased regulation in this area is weak. Australian
encouraged smaller players in the margarine industry to health authorities were quick to identify the potential risk
adopt low trans fat formulations. There are now 37 catego- associated with trans fats and communicate it to health-care
ries of supermarket foods in the Heart Foundations Tick professionals and the food industry. The timely response
programme that include criteria for trans fats (personal from major margarine companies followed by a wider indus-
communication). try response has seen levels of industrial trans fats in the
Although the major fall in the trans fat content of the Australian diet fall to just one-eighth of 1% of dietary energy.
Australian diet is likely to have occurred during the 1990s, Total trans fat intakes in Australia are already lower than
there have been many initiatives undertaken in various in Denmark, which has regulated to lower the trans fat
sectors of the food industry during the last decade that have content of its food supply. Rather than there being a good
served to drive trans fat intake down to its current low level. case for increased regulation, the trans fat issue stands as
At least two multinational companies operating in Australia a good example of self-regulation though collaboration
have participated in global company programmes to lower between Australian health agencies, the food industry and
trans fats in their product ranges;26,27 the leading chain of the government.
quick service restaurants has adopted a frying oil low in both
saturated and trans fats;28 the leading supermarket chain in
Australia has included trans fats in the nutrition information
Acknowledgements
panel of its house brand products thereby highlighting Preparation of this paper was partly supported by Goodman
problem foods in its portfolio (personal communication); Fielder.
and a major donut manufacturer that had been the focus of
criticism for the high trans fat content of its frying oil has
adopted a low trans fat formulation.29 The Commonwealth References
Government played a role in facilitating the removal of 1 ABC television, Stateline (NSW). Carrs Crusade. 28 August
industrial trans fats from products available through quick 2009.
service restaurants, convening a roundtable in March 2007 2 Kaye J. The Greens. Greens call for NSW govt support for bill
and reporting good progress in September of that year30 and to terminate trans fats. Media release, 27 July 2008.
more recently.31 3 Danish Veterinary and Food Administration. Executive Order
Advocates of increased regulation of trans fats in Australia2 No. 160 pursuant to section 13, section 55(2) and section 78(3)
have highlighted the situation in Denmark, which intro- of Act No. 471 (the Danish Food Act). 11 March 2003.
duced a regulation imposing a limit on the content of trans 4 Board of Health, New York City. The regulation to phase out
artificial trans fats in New York City food service establishments.
fats in food ingredients in 2003. Although there has been a
Section 81.08, New York City Health Code 2007.
remarkable (85%) fall in the amount of trans fats in the 5 State of California. Health and Safety Code. Part 7, Division
Danish diet since the late 1970s, it should be noted that by 104, Chapter 12.6, Trans fats 2008.
far, the majority of this fall occurred before the year 2000, 6 Australian Bureau of Statistics. National Nutrition Survey.
that is, before the introduction of regulation.32 The factors Nutrient Intakes and Physical Measurements, Australia, 1995.
underpinning the Danish trend are very similar to those at 1998 Commonwealth of Australia. ABS Catalogue No. 4805.0.
play in Australiascientific leadership from the Danish 7 Expert Panel on Trans Fatty Acids and Coronary Heart Disease
Nutrition Council33 and a rapid response from industry, Risk, Kris-Etherton PM, ed. Physicochemical properties, intake,
including the removal of trans fats from table margarines, and metabolism. Am J Clin Nutr 1995; 62: 659S69S.
which corresponded with the period of the steepest fall in 8 Sommerfeld M. Trans unsaturated fatty acids in natural pro-
ducts and processed foods. Prog Lipid Res 1983; 22: 22133.
the trans fat content of the Danish diet between 1990 and
9 British Nutrition Foundation. Trans Fatty Acids. London: British
1995.32 The current trans fat content of the Danish diet is at
Nutrition Foundation, 1987.
least 0.7% of dietary energy,32,34 which is slightly higher than 10 Mensink RP, Katan MB. Effect of dietary trans fatty acids on
that in the Australian diet.19,20 By far, the majority of trans fats high-density and low-density lipoprotein cholesterol levels in
in the Danish diet is ruminant in origin, as is the case in the healthy subjects. N Engl J Med 1990; 323: 43945.
Australian diet. Thus, Denmark and Australia have arrived at 11 Mensink RP, Zock PL, Kester AD, Katan MB. Effects of dietary
near identical positions in relation to trans fats, one employ- fatty acids and carbohydrates on the ratio of serum total to HDL
ggWg$
U d ^g [g ^
^ggd W[ l Wg q W
v6AWvW^dcg
gW[ggz EWcd#
gdgfd}W/&&/d
[HdcWc^gW}g^acgW
WgqWv^gWg^^g
&& 9V
\p
g _d &=^
)
")
K
l
i
iR
`
1
&2&( "
z
%8!88 !%*l%e*8*8e!!%* !hU%! !%!%*%!! *h*
d
9 -Wk 1
E ;a
"2&
9V
E
Q- )
22
"=^
&=&"2(
w
1
Qu
a
8M@;eH'e>,/@e
& &
Ay p&
)
// m 4 ;1e-
Z" @!aeGe7`KD<eH(e?-@e
H .Re
AD
0
u
(_ 0
v
t ^ ^ saTh il:
<:i (bKW
)9K@=eH(e[eN2Ae%ePT0]e X
T@
L ([
F&z %
' I
O
982 /NTO)6 k
[eK1Ae,&ePT0^e % %
$F D
e
w
L
Z# / f 3) X 3P
; P-
e J L8[ 7`LB=e [eO1Ae X#3e
0 k
d46 ? I
= Ho00
j-(w
{
%
Ux Ux
R S $
%|
; je .l$"
?=
@@ Y_3 #_cc
Image description. Better Health Channel logo End of image description.
Fatisimportantformanybodyprocesses.Youneedtoeatsomefatinyourdiet.Fatprotectsyour
organs,keepsyouwarmandhelpsyourbodyabsorbandmovenutrientsaround.Italsohelps
hormoneproduction.However,somefatsarebetterthanothersandhavingtoomuchofanytype
isnotagoodidea.
Dietaryfatsareclassifiedbytheirstructure.Differenttypesoffatsreactdifferentlyinsidethebody.
Saturatedfats(foundmostlyinanimalproducts)increasebloodcholesterol,whichisariskfactor
incoronaryheartdisease.Mono-unsaturatedandpolyunsaturatedfatstendtolowerblood
cholesterol.
LDLisconsideredthebadcholesterolbecauseitcontributestothenarrowingofthearteries,
whichcanleadtoheartdiseaseandstroke.HDLcholesterolisconsideredtobethegood
cholesterolbecauseitactuallycarriescholesterolfromthebloodbacktotheliver,whereitis
brokendown,reducingtheriskofcardiovasculardisease.
Groups of fats
Eachgroupoffatsbehavesdifferentlyinsidethebody.Dietaryfatcanbeclassifiedintofour
groups.Theseare:
saturated
mono-unsaturated
polyunsaturated
transfats.
Mono-unsaturatedandpolyunsaturatedfatsbothtendtolowerbloodcholesterolwhentheyreplace
saturatedfatsinthediet.Polyunsaturatedfatshaveaslightlygreaterimpactthanmono-
unsaturatedfats.
Replacesaturatedfatsinyourdietwitheithermono-unsaturatedorpolyunsaturatedfatswhenever
possible.Forexample,replacebutterinsomecookeddisheswitholiveoilormargarine.
Transfattyacidstendtobehavelikesaturatedfatsinthebody,astheyraiseLDLlevelsand
increasetheriskofheartdisease.Unlikesaturatedfats,theytendtolowerHDLcholesterol,soare
potentiallyevenmoredamaging.
Fatsandoils Page1of4
Itisthetransfatsthatareproducedduringfoodmanufacturingthatyoushouldbemostconcerned
about,notthetransfatspresentnaturallyincertainfoods.Lookformargarinesthathavelessthan
onepercenttransfatsonthelabelorchoosefoodswiththeHeartFoundationTick.Limithow
muchtakeawayfoodandpackagedsnackfoodsyoueat.
Saturated fatssourcesincludefattycutsofmeat,full-fatmilk,cheese,butter,cream,
mostcommerciallybakedproductssuchasbiscuitsandpastries,mostdeep-friedfast
foods,coconutandpalmoil.
Mono-unsaturated fatssourcesincludemargarinespreadssuchascanolaoroliveoil
basedchoices,oilssuchasolive,canolaandpeanutoils,avocado,andnutssuchas
peanuts,hazelnuts,cashewsandalmonds.
Polyunsaturated fatssourcesincludefish,seafood,polyunsaturatedmargarines,
vegetableoilssuchassafflower,sunflower,cornorsoyoils,nutssuchaswalnutsand
Brazilnuts,andseeds.
Omega-3 fatsarefoundinbothplantandmarinefoods,althoughitistheomega-3fats
frommarinesourcesthathavethestrongestevidenceforhealthbenefits(including
reducingtheriskofheartdisease).Plantfoodsourcesincludecanolaandsoyoils,and
canola-basedmargarines.Marinesourcesincludefish,especiallyoilyfishsuchasAtlantic
salmon,mackerel,Southernbluefintuna,trevallyandsardines.
Omega-6fatsarefoundprimarilyinnuts,seedsandplantoils,suchascorn,soyand
safflower.
lowertriglyceridelevels,whichareimportantriskfactorsincoronaryheartdisease
improvebloodvesselelasticity
keeptheheartrhythmbeatingnormally
thintheblood,whichmakesitlessstickyandlesslikelytoclot
reduceinflammationandsupporttheimmunesystem
reducebloodpressure
mayplayaroleinpreventingandtreatingdepression
contributetothenormaldevelopmentofthefetalbrain.
Plant sterols
Plantsterolsarepresentinallplants.Intakesof23gofplantsterolsperdayhavebeenshownto
reducebloodcholesterollevelsbyanaverageof10percent.Thisisbecausetheyblockthebodys
abilitytoabsorbcholesterol,whichleadstoareducedlevelofcholesterolintheblood.
However,itishardtoeatthisamountofplantsterolsfromnaturalsources,sotherearenowplant
sterol-enrichedmargarinesanddairyproductsonthemarket.Eating11.5tablespoonsofsterol-
enrichedmargarineeachdaycanhelptolowerbloodcholesterollevels.
Fatsandoils Page2of4
Carryingtoomuchbodyfatisariskfactorinmanydiseases,includingcoronaryheartdisease,type
2diabetesandmanycancers.
Cholesterol in food
Peoplewithhighbloodcholesterolorwhoareatriskofheartdiseaseshouldalsotrytolimittheir
intakeofcholesterol-richfoods.However,whilecholesterolinfoodcanraisebloodcholesterol
levels,theeffectissmallcomparedtotheeffectsofsaturatedfat.
Dietarycholesterolisonlyfoundinanimalproductssuchas:
full-fatdairyproducts
fattymeats
eggyolks
shellfish
offalforexample,liver,kidneyandbrains.
Studieshaveshownthatoliveoilconsumptionmayhaveaprotectiveroleonbreast,colon,lung,
ovarianandskincancerdevelopment.Compoundsspecifictooliveoil,knownasphenolics,seemto
possessfreeradical-scavengingpropertiesandsomaybeabletoreduceoxidativedamagetoDNA.
Anumberofstudieshavealsoshownthatoliveoilmayhaveadditionalbeneficialeffectsonblood
pressure,obesity,rheumatoidarthritisandimmunefunction.
Wemustremember,though,thattheMediterraneandietcontainsmuchmorethanoliveoil.Its
possiblethatthelowrateofcoronaryheartdiseaseinthesecountriesrelatestoahighintakeof
vegetables,legumes,fruitsandcereals,whicharerichinantioxidants.
Usemargarinespreadsinsteadofbutterordairyblends.
Usesaladdressingsandmayonnaisemadefromoilssuchascanola,sunflower,soyand
oliveoils.
Useloworreduced-fatmilkandyoghurtoraddedcalciumsoydrinks.
Trytolimitcheesetotwiceaweek.
Havefish(freshorcanned)atleasttwiceaweeklimitfishthatishighinmercurytoonce
aweek.Veryyoungchildrenandwomenwhoarepregnant(orintendingtobecome
pregnantwithinthenextsixmonths)areadvisedtoeatfishthatislowinmercury
(shellfish,salmonandcannedtuna).
Selectleanmeat(meattrimmedoffatandchickenwithoutskin).Trytolimitfattymeats,
includingsausagesanddelicatessenmeatssuchassalami.
Snackonplain,unsaltednutsandfreshfruit.
Incorporatedriedpeas(forexample,splitpeas),beans(forexample,haricotbeans,kidney
beans,threebeanmix)orlentilsintotwomealsaweek.
Makevegetablesandgrain-basedfoods,suchasbreakfastcereals,bread,pasta,noodles
andrice,themajorpartofeachmeal.
Trytolimittakeawaytoonceaweekorless.
Trytolimitsnackfoodssuchaspotatocrispsandcorncrispstoonceaweekorless.
Trytolimitcakes,pastriesandchocolateorcreamybiscuitstoonceaweekorless.
Peoplewithhighcholesterolshouldtrytolimitcholesterol-richfoodssuchaseggyolks,
andoffallikeliver,kidneyandbrains.
Fatsandoils Page3of4
EBIOM-00561; No of Pages 10
EBioMedicine xxx (2016) xxxxxx
EBioMedicine
Research Paper
a r t i c l e i n f o a b s t r a c t
Article history: Nutrition plays a signicant role in the increasing prevalence of metabolic and brain disorders. Here we employ
Received 1 February 2016 systems nutrigenomics to scrutinize the genomic bases of nutrienthost interaction underlying disease predispo-
Received in revised form 5 April 2016 sition or therapeutic potential. We conducted transcriptome and epigenome sequencing of hypothalamus
Accepted 7 April 2016
(metabolic control) and hippocampus (cognitive processing) from a rodent model of fructose consumption,
Available online xxxx
and identied signicant reprogramming of DNA methylation, transcript abundance, alternative splicing, and
Keywords:
gene networks governing cell metabolism, cell communication, inammation, and neuronal signaling. These sig-
Systems nutrigenomics nals converged with genetic causal risks of metabolic, neurological, and psychiatric disorders revealed in humans.
Fructose Gene network modeling uncovered the extracellular matrix genes Bgn and Fmod as main orchestrators of the ef-
Omega-3 fatty acid fects of fructose, as validated using two knockout mouse models. We further demonstrate that an omega-3 fatty
DHA acid, DHA, reverses the genomic and network perturbations elicited by fructose, providing molecular support for
Epigenome nutritional interventions to counteract diet-induced metabolic and brain disorders. Our integrative approach
Transcriptome complementing rodent and human studies supports the applicability of nutrigenomics principles to predict dis-
Brain networks
ease susceptibility and to guide personalized medicine.
Metabolic diseases
2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
Brain disorders
Extracellular matrix (http://creativecommons.org/licenses/by-nc-nd/4.0/).
http://dx.doi.org/10.1016/j.ebiom.2016.04.008
2352-3964/ 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Please cite this article as: Meng, Q., et al., Systems Nutrigenomics Reveals Brain Gene Networks Linking Metabolic and Brain Disorders,
EBioMedicine (2016), http://dx.doi.org/10.1016/j.ebiom.2016.04.008
2 Q. Meng et al. / EBioMedicine xxx (2016) xxxxxx
in the utilization of nutrient-based strategies for the prevention and molecular signals from our rodent models with human genome-wide
treatment of common complex disorders. association studies (GWAS) of metabolic and brain disorders.
As fundamental aspects of gene regulation, disruptions in
epigenomic reprogramming, transcript abundance, alternative splicing, 2.2. Rat model of fructose consumption and DHA supplementation
and genegene interactions are increasingly recognized as core aspects
of wide-ranging pathogenesis (Chen et al., 2008, Zhang et al., 2013, Male SpragueDawley rats (Charles River Laboratories, Inc., MA,
Rhinn et al., 2013, Narayanan et al., 2014, Makinen et al., 2014, Yang USA) of 2 months old weighing 200-220 g were randomly assigned to
et al., 2009). Systems nutrigenomics is emerging as a powerful ap- 15% fructose treatment (n = 8, 15% w/v fructose in the drinking
proach to reveal the hidden aspects of pathogenesis under dietary mod- water), 15% fructose plus an omega-3 fatty acid diet rich in DHA (n =
ulation (Zhao et al., 2015, De Graaf et al., 2009, Panagiotou and Nielsen, 8; 0.5% of axseed oil supplying ALA and 1.2% of DHA capsule oil, Nordic
2009). Here we apply systems nutrigenomics to unveil the multidimen- Naturals, Inc., CA, USA), or a control group (n = 8, without fructose in
sional molecular interactions driven by fructose and DHA that regulate drinking water or DHA supplement) for six weeks. Sample size was cho-
pathogenesis and recovery, and to provide proof-of-principle on the po- sen to yield N 80% statistical power to detect 30% between-group differ-
tential of systems nutrigenomics to guide personalized medicine. The ence with 10% within-group difference in a phenotype using two-sided
comparative account of nutrigenomics signals between fructose and Student's t-test. The rats were singly housed in polyacrylic cages with
DHA is crucial to understand how select diets impact the molecular free access to water and their respective diets, and maintained under
substrates governing the balance between normal brain function and standard housing conditions (room temperature 2224 C) with 12 h
disease, and holds potential for guiding effective preventative and light/dark cycle. Daily food and drink intake were monitored. The
therapeutic strategies to mitigate common human diseases. fructose intake level is approximately equivalent to long-term
daily consumption of 130 g sugar in 12 l soda drinks in a 60 kg
2. Materials and Methods human. The total fat content in the control and DHA diets was 10 g
per 100 g of diet. The rats were then examined for changes in MetD-
We describe essential methods in the main text and detailed exper- related phenotypes (serum levels of insulin, glucose, and triglycerides,
imental procedures are available in the Supplemental Materials. and insulin resistance index (fasting glucose [mg/dl] fasting insulin
[ng/ml] / 16.31)). Rats were trained in the Barnes maze test for 5 days
to learn the task prior to diet treatment and then tested for memory
2.1. Overall Study Design
retention in the Barnes Maze after 6 weeks of treatment as previously
described (Agrawal and Gomez-Pinilla, 2012). Mice were sacriced,
As depicted in the analysis ow in Fig. 1, we focus our study on two
and hypothalamus and hippocampus were dissected out, ash frozen,
key regions in the rodent brain that are important for the regulation of
and stored at 70 C for transcriptome and DNA methylome sequenc-
metabolism (hypothalamus) and cognition (hippocampus), and there-
ing experiments.
fore can play a major role in fructose-induced metabolic and brain
dysregulation as well as DHA-mediated recovery. We analyzed the tran-
2.3. RNA Sequencing (RNA-Seq) and Data Analyses
scriptome and DNA methylome using next generation sequencing,
followed by investigation of the regulatory relationship between the
RNA-Seq libraries were prepared for 24 RNA samples (n = 4 per
methylome and transcriptome. Network approaches were then applied
treatment group per brain region) and sequenced in paired-end mode
to model genegene interactions and to predict essential perturbation
by HiSeq 2000 (Illumina Inc., CA, USA) as detailed in Supplemental Ma-
or regulatory points. Knockout mouse models were subsequently used
terials. The short reads data were analyzed for different transcription
to validate the predicted regulatory genes with regard to their ability
between treatment and control groups using the Tuxedo tool package
to modulate metabolic and behavior phenotypes. To infer translatability
(Trapnell et al., 2012) and false discovery rate (FDR) was estimated
to human pathophysiology, we assessed the intersection of the
using the q-value approach. One hypothalamus sample from the
fructose group failed standard quality control and was removed from
analysis. Genes and transcripts showing differential expression or alter-
native splicing at p b 0.01 in each brain region were dened as a gene
signature for further integrative analyses. The reliability of tran-
scriptome signals at p b 0.01 from RNA-Se1 was conrmed using quan-
titative real-time PCR (qPCR) for 12 selected genes (Supplemental
Materials). Relative gene expression was represented by delta Ct =
Ctgene CtGapdh. The RNA-Seq data was deposited to Gene Expression
Ominbus (GEO) under accession numbers GSE59918 (control and
fructose groups) and GSE 64815 (DHA).
Please cite this article as: Meng, Q., et al., Systems Nutrigenomics Reveals Brain Gene Networks Linking Metabolic and Brain Disorders,
EBioMedicine (2016), http://dx.doi.org/10.1016/j.ebiom.2016.04.008
Q. Meng et al. / EBioMedicine xxx (2016) xxxxxx 3
trans-regulatory effects of DNA methylome on transcriptome were fur- group, or between KO mice (Bgn KO or Fmod KO) and wild-type mice.
ther examined (Supplemental Materials). RRBS data was deposited to Two-way ANOVA with Holm-Sidak post hoc analysis for multiple com-
GEO under accession numbers GSE59893 (control and fructose groups) parisons was performed for the learning curves. For RNA-Seq, RRBS,
and GSE64816 (DHA). KDA, and SSEA analyses, the statistical tests and multiple testing correc-
tion methods were described in detail under the respective method
2.5. Assessing Overlap Between Fructose Signatures and Candidate Causal sections. All statistical and bioinformatics analyses were performed in
Genes From Human GWAS of MetDs and Brain Disorders R (http://www.R-project.org/).
We tested the consistency between the genes identied from fruc- 2.9. Study Approval
tose treatment in our study and those from human GWAS related to
MetDs and brain functions in two ways. First, we cross-checked genes All experiments were performed in accordance with the National In-
identied in our study with top GWAS candidate genes reported in stitutes of Health Guide for the Care and Use of Laboratory Animals and
the GWAS catalog (http://www.genome.gov/gwastudies/) (Hindorff were approved by the Chancellor's Animal Research Committee of the
et al., 2009) for direct overlap. Second, we used a SNP Set Enrichment University of California, Los Angeles.
Analysis (SSEA) (Makinen et al., 2014), which goes beyond the top
GWAS hits and tests the overall enrichment of the genes affected in
our fructose animal model for SNPs that demonstrated disease associa- 3. Results
tion in humans using KolmogorovSmirnov (KS) test and Fisher's exact
test. Bonferroni-corrected p b 0.05 by either KS test or Fisher's exact test 3.1. Induction of Metabolic and Behavioral Abnormalities by Fructose and
was considered signicant. The detailed methods and GWAS datasets Normalization by DHA
involved are described in Supplemental Materials.
Compared to the control group (n = 8), adult rats consuming fruc-
2.6. Network Modeling and Identication of Key Drivers (KD) of Fructose tose (n = 8) had MetD characteristics including signicantly increased
and DHA Signatures blood glucose, triglycerides, insulin, and insulin resistance index
(Fig. 2a). The animals on fructose also displayed impaired memory as
We obtained both proteinprotein interaction (PPI) networks from demonstrated by prolonged latency in the Barnes Maze test (Fig. 2a).
the HPRD database (http://www.hprd.org) (Keshava Prasad et al., On the other hand, fructose-treated animals with DHA supplementation
2009) and tissue-specic Bayesian networks (BNs) (Zhu et al., 2004, (DHA + fructose) showed signicantly improved metabolic parameters
2008) for hypothalamus and hippocampus constructed based on genet- including lower serum triglycerides, insulin, insulin resistance index,
ic and transcriptomic data from multiple large-scale mouse studies as well as improved memory as reected by the reduced latency
(Supplemental Materials). We used the fructose or DHA signature time in the Barnes Maze test (Fig. 2a). The measurements in the
genes identied from the RNA-Seq analysis as seeds to extract the top DHA + fructose group were not signicantly different from the control
most connected subnetworks as described previously (Yang et al., group with no fructose treatment for triglycerides, insulin, and latency
2009). The network components including nodes and edges as well as time, indicating that DHA supplementation counteracted the fructose
topological structures were visualized using Cytoscape (Smoot et al., effects. Of note, total caloric intake between groups was similar between
2011). A Key Driver Analysis (KDA; detailed in Supplemental Materials) groups.
(Yang et al., 2010, Zhang et al., 2013) was used to identify potential key
regulators for the fructose and DHA signatures based on the topology of 3.2. Transcriptomic Changes in Hypothalamus and Hippocampus Induced
BNs and PPI networks. Briey, the neighboring subnetwork of each gene by Fructose and Normalization by DHA
in a given network was rst extracted and then compared with the
tissue-matched fructose or DHA signature to assess the enrichment of Using RNA-Seq, we identied 581 and 146 differentially expressed
genes in the latter using Fisher's exact test. Network genes that reach genes, 352 and 79 differentially expressed transcripts, and 99 and 53
Bonferroni-adjusted p b 0.05 were reported as KDs. KDs identied genes showing alternative exon usage in hypothalamus and hippocam-
from multiple networks were ranked according to a composite score pus, respectively, at p b 0.01 (Dataset S1). These results support large-
that favors KDs that are highly signicant in each KDA and are consis- scale alterations in transcriptional activities affecting both differential
tent across network models used for KDA (Supplemental Materials). expression and alternative splicing, and dene 734 unique hypothalam-
ic genes (312 passed FDR b 5%) and 206 hippocampal genes (56 passed
2.7. KD Validation Using Bgn and Fmod Knockout (KO) Models FDR b 5%) as fructose gene signatures. The reliability of the differen-
tially signals at p b 0.01 was supported by consistent qPCR experiments
Bgn KO mice (Xu et al., 1998) and Fmod KO mice (Svensson et al., on 12 selected genes (Table S1). Among the signatures were transcrip-
1999) were given as gifts from National Institute of Dental and Cranio- tion factors (e.g., Aff3, Junb, Zbtb16, Tcf7l2), epigenetic regulators
facial Research and re-derived at the University of California, Los (e.g., Trmu, Suv420h2, Rgs9bp), and alternative splicing regulators
Angeles. Mice were housed in standard polyethylene cages (34 mice (e.g., Bicc1, Prpf31, Rbpms) (Dataset S1), which may partially explain
per cage) in an environmentally controlled room (2224 C) with a the large-scale transcriptional alterations observed.
12 h light/dark cycle and free access to food and water. Homozygous The fructose signatures showed strong tissue-specicity, with 664
male KO mice (Bgn: n = 16, Fmod: n = 11) of 9-week age were tested genes unique to the hypothalamus, 136 unique to the hippocampus,
against male wild type (WT) controls (n = 8) for phenotypic changes in and 70 overlapping between tissues. In particular, only in hypothalamus
metabolic (glucose tolerance, plasma lipid panel) and learning and we observed upregulation of Slc2a5 (encoding the major fructose trans-
memory traits (Barnes maze) as described in the Supplemental porter GLUT5) and downregulation of Slc2a4 (encoding the glucose
Materials. transporter protein GLUT4). The signatures also differ by direction of
change between the two brain compartments: 68% of hypothalamic sig-
2.8. Statistics natures showed down-regulation compared to 42% in hippocampus.
Among the overlapping genes between the two brain regions (Dataset
For metabolic and behavior phenotypes, two-sided Student's t-test S1), only 6 showed consistent upregulation (Atp5f1, Pnpla1, Pomc,
was used to test statistical differences in between fructose-treated Sepw1, Slc39a12, and Ubc) and 5 showed consistent downregulation
mice and control mice, or between the fructose and fructose + DHA (Alox15, Dstn, Rpl30, Slc39a4, and Slc5a7). The consistent genes are
Please cite this article as: Meng, Q., et al., Systems Nutrigenomics Reveals Brain Gene Networks Linking Metabolic and Brain Disorders,
EBioMedicine (2016), http://dx.doi.org/10.1016/j.ebiom.2016.04.008
4 Q. Meng et al. / EBioMedicine xxx (2016) xxxxxx
Please cite this article as: Meng, Q., et al., Systems Nutrigenomics Reveals Brain Gene Networks Linking Metabolic and Brain Disorders,
EBioMedicine (2016), http://dx.doi.org/10.1016/j.ebiom.2016.04.008
Q. Meng et al. / EBioMedicine xxx (2016) xxxxxx 5
involved in mitochondria function, energy balance, antioxidation, localized with DMLs within 50 kb distance (Table S3), suggesting cis-
inammation, and cognition. regulation of gene expression by the local DMLs. Notably, some of the
Compared to the fructose group, the DHA + fructose group showed genes with co-localizing DMLs are transcription factors (Aff3, Junb, and
differential expression of 651 and 462 genes, 449 and 277 transcripts, Zbtb16) or epigenetic factors (Parp9) (Dataset S3), which could in turn
and alternative exon usage for 155 and 81 genes in hypothalamus and trans-regulate downstream target genes. Additionally, microRNAs adja-
hippocampus at p b 0.01, respectively (Dataset S1), which together de- cent (within 20 kb or 50 kb) to the DMLs had the potential to regulate
ne DHA gene signatures of 910 unique genes (382 passed FDR b 5%) 36%45% and 17%45% of hypothalamic and hippocampal signatures
in hypothalamus and 569 unique genes (246 passed FDR b 5%) in hippo- which were the predicted target genes of the corresponding microRNAs
campus. Strikingly, these DHA signatures signicantly overlapped with (Table S3; details of the microRNAs in Dataset S3). These results suggest
the fructose signatures (Table 1), with DHA reversing the expression of that fructose-induced DNA methylation changes could contribute
the common signatures perturbed by fructose in both brain regions to transcriptional alteration through both cis and trans-regulatory
(Fig. 2b; Dataset S1). mechanisms.
3.3. Functional Categorization of Fructose and DHA Transcriptomic 3.6. Perturbation of Gene Subnetworks by Fructose and Normalization by
Signatures DHA
We found enrichment of both tissue-specic and shared biological To explore the genegene relations among the fructose signature
pathways between brain regions and between treatment groups (select genes and to identify key perturbation points of fructose, we employed
pathways shown in Fig. 2d; full details in Dataset S2). Fructose and DHA two types of networks - proteinprotein interaction (PPI) networks
signatures shared similar over-represented pathways, although the di- (Keshava Prasad et al., 2009) that capture a majority of the well-
rection of expression changes of the overlapping genes was predomi- known interactions between proteins, and data-driven Bayesian
nantly opposite (exemplied in Fig. 2d insert; details in Dataset S1). networks (BNs) (Zhu et al., 2004, 2008) that elucidate genegene regu-
The pathways unique to hypothalamus included those highly relevant latory relationships (detailed in Materials and Methods). Using these
to cardiometabolic, inammatory, and neuronal functions. The pathway networks and a network topology-based key driver analysis (KDA),
affected in the hippocampus was neurological processes. Those affect- we identied multiple candidate key driver (KD) genes whose network
ed in both regions were related to extracellular matrix (ECM), adhesion, neighboring genes highly overlapped with the fructose signature genes
complement, and insulin-like growth factor (IGF) signaling. (Dataset S4). As shown in Fig. 3, the top 5 KDs for each brain region form
central hubs connecting many fructose signature genes in coherent
3.4. Induction of DNA Methylomic Changes by Fructose and Normalization gene subnetworks, which contained genes involved in diverse process-
by DHA es such as the ECM (e.g., Col6a2, Lama4, Thbs1, Fmod, Bgn), small mole-
cule transporters (e.g., Slc6a13, Slc22a2, Slc22a6, Slc13a3), and forkhead
To identify epigenomic changes which may play a role in the transcription factors (e.g., Foxc2, Foxd2). Between the subnetworks of
transcriptomic alternation associated with fructose and DHA, we exam- the two brain compartments, Fmod, Bgn, and Foxc2 were consistent
ined the DNA methylome using RRBS that measures millions of poten- KDs revealed from our data-driven approach (Dataset S4). Remarkably,
tial DNA methylation sites at single base resolution. We found that these KDs and subnetworks were largely reversed by DHA supplemen-
fructose consumption induced a large-scale switch of the methylation tation (Fig. 3).
patterns in both brain regions: 734 and 810 of differentially methylated
loci (DMLs) showed hypomethylation and hypermethylation in hypo- 3.7. Validation of Bgn and Fmod as Key Regulators of MetDs and Behavioral
thalamus, respectively; 972 and 900 DMLs showed hypomethylation Aberrations in KO Mouse Models
and hypermethylation in hippocampus at a FDR b 5%, respectively
(Fig. 2c). DHA largely reversed the fructose-induced methylation To evaluate our network KD prediction, we used Bgn and Fmod KO
changes in these two brain regions (Fig. 2c; Table 1). mice (Xu et al., 1998; Svensson et al., 1999) to test the effect of
perturbing these KDs on phenotypes related to MetDs and cognitive
3.5. Relationship Between Differential Methylation and Differential function. Compared to wild type (WT) controls, we observed signi-
Expression cantly elevated triglyceride, total cholesterol, HDL cholesterol, and un-
esteried cholesterol in the plasma of both Bgn and Fmod KO mice, sig-
To explore the potential relationship between DMLs and gene ex- nicant increases in LDL cholesterol in the Bgn KO, and elevated free
pression, we mapped the fructose DMLs in each brain region to adjacent fatty acids in the Fmod KO (Fig. 4a). We also found decreased levels of
genes within 10 kb distance. We found 50% of DMLs to be located in the plasma glucose, insulin, as well as insulin resistance index in Fmod KO
intergenic regions, 34% in the gene body (including 5-UTR, intron, (Fig. 4bd). The intraperitoneal glucose tolerance test of Bgn KO
coding sequence, and 3-UTR), and 16% in the upstream or downstream demonstrated signicantly lower plasma glucose at 15 min after
regions of the genes, with a paucity of DMLs in the promoter and tran- glucose injection and a similar trend in the Fmod KO (Fig. 4e).
scription start site (Fig. S1; Table S2). Assessing the methylome- To assess the potential inuence of Bgn and Fmod for spatial learning
transcriptome relationship revealed co-localization of DMLs and gene and memory, we tested the animals in the Barnes Maze. Because Fmod
signatures: 74 out of the 734 (10.1%) hypothalamic signature genes and Bgn KOs have been previously shown to have impaired tendon
and 18 out of the 206 (8.7%) hippocampal signature genes co- and joints which may affect latency time (Ameye et al., 2002,
Fig. 2. Changes in metabolic and behavior phenotypes, transcriptome, DNA methylome, and biological pathways in response to fructose treatment and DHA supplementation.
(a) Metabolic and behavior phenotypes. From left to right: blood glucose, serum triglycerides, serum insulin, insulin resistance index, and latency time in the memory retention probe
in the Barnes Maze test. Fructose group was compared with the control group, and the fructose + DHA group was compared to the fructose only group. p b 0.01 and #p b 0.05 by 2-
sided Student's t-test. Error bars in the plots are standard errors. N = 8/group. (b) Heatmap of gene expression changes in hypothalamus and hippocampus. Blue to red colors indicate
low to high expression values. (c) Heatmap of DNA methylation changes in hypothalamus and hippocampus. Blue to red colors indicate low to high methylation levels. (d) Select
biological pathways affected by fructose and DHA in hypothalamus and hippocampus. Bars are the log10 enrichment p values of the pathways. The insert plot on the right shows the
opposite direction of changes in genes in the focal adhesion pathway in the hypothalamus between fructose and DHA: fructose mostly inhibits whereas DHA reverses the expression
levels. In the insert plot Y-axis indicates log10 fold changes. The fructose + DHA group is labeled as F + D in panels bd. (For interpretation of the references to color in this gure
legend, the reader is referred to the web version of this article.)
Please cite this article as: Meng, Q., et al., Systems Nutrigenomics Reveals Brain Gene Networks Linking Metabolic and Brain Disorders,
EBioMedicine (2016), http://dx.doi.org/10.1016/j.ebiom.2016.04.008
6 Q. Meng et al. / EBioMedicine xxx (2016) xxxxxx
Table 1
Overlap between fructose and DHA signatures. Numbers of signicant genes or methylation loci are shown. Enrichment p value was calculated using 2-sided Fisher's exact test.
Data Tissue Fructose signature DHA signature Overlap Background Fold enrichment p value
Chakravarti, 2002; Jepsen et al., 2002), we analyzed the Barnes Maze fructose and DHA actions was substantiated by the signicant alter-
data by number of mistakes to eliminate the potential confound origi- ations in the methylome of hypothalamus and hippocampus. cis-Regu-
nated by differences in mobility. In addition, number of mistakes has lation of transcriptome by DNA methylation was evidenced by co-
been demonstrated as a more accurate measure of cognitive abilities localization of a subset of signature genes with local DMLs; trans-
in mice (O'leary and Brown, 2013). During the learning phase, there regulation mechanism was supported by the co-localization of DMLs
was a signicant effect of animal type (F2,35 = 10.14; p = 0.0003 by with a number of transcriptional regulators (e.g., Aff3, Junb, Zbtb16,
2-way ANOVA) and day (F3,105 = 40.44; p b 0.0001 by 2-way ANOVA) and Parp9) and microRNAs (e.g., rno-miR-421, rno-miR-143) whose
on the number of mistakes made in the Barnes Maze (Fig. 4f). Post predicted target genes were enriched among the signature genes. An in-
hoc analysis revealed that both Bgn KO and Fmod KO made fewer creasing body of evidence indicates that neurological and psychiatric
mistakes compared with WT during the learning and the memory disorders have an epigenetic component (Jakovcevski and Akbarian,
phases. Representative track plot analysis show a difference in strat- 2012, Tsankova et al., 2007). In particular, disruption in cell metabolism
egy to nd the escape hole (Fig. 4g, inset): the WT animals use a trial seems to be a driving force for epigenetic changes associated with cog-
and error approach while the Bgn and Fmod KO animals follow a nitive function (Tyagi et al., 2015). The current results showing that
more direct path. DHA can counteract the broad transcriptomic and methylomic alter-
ations induced by fructose imply that diet has the ability to regulate
3.8. Relevance of Fructose Signatures and KDs to Human Diseases the susceptibility to disease by modulating the multifaceted transcrip-
tional machinery.
To assess the relevance of our nutrigenomic signals from the rodent Our studies also revealed broad impact of fructose and DHA on fun-
models to human pathophysiology, we compared the fructose signature damental cell processes encompassing metabolism, inammation, cell
genes as well as the KDs to human GWAS of MetDs and brain related communication, and neuronal signaling, as well as critical differences
diseases or traits. We found numerous overlapping genes between fruc- between the two brain regions. For instance, the fructose and glucose
tose signature genes and top GWAS hits in the GWAS catalog for a broad transporter genes Slc2a5 and Slc2a4, and a broader category of pathways
range of brain disorders and MetDs-related diseases or traits (Dataset including insulin signaling, cardiomyopathy, transforming growth
S4). Importantly, KDs such as Fmod, Col1a2, Col18a1, and Dcn were factor (TGF) beta signaling, mitogen-activated protein kinases (MAPK)
GWAS candidate genes associating with obesity, cardiovascular disease, signaling, platelet-derived growth factor (PDGF) signaling, toll like
metabolic syndrome, and brain morphology and cognition, respectively. receptor (TLR) signaling, brain-derived neurotropic factor (BDNF) sig-
When using full sets of GWAS results we had accessed for plasma lipids, naling, neural cell adhesion molecule (NCAM) signaling, and axon guid-
T2D, systolic and diastolic blood pressure traits, cognitive traits, and bi- ance were perturbed only in the hypothalamus. In fact, most of the
polar disorder, we found that human orthologs of the fructose signature hypothalamic pathways have been recently implicated in leptin resis-
genes derived from both hypothalamus and the hippocampus were sig- tance, hypothalamic deregulation and peripheral metabolism impair-
nicantly enriched for genetic polymorphisms showing stronger associ- ment (Milanski et al., 2012, Yan et al., 2014). The prominent response
ations with these diseases or traits (Table 2). of the hypothalamus to diets is consistent with its role as the master
metabolic sensor and regulator, and the current results suggest that
4. Discussion these actions intermingle with inammatory and cognitive signals. In
contrast, the effects of fructose on the hippocampus were prominent
In this systems nutrigenomics study, we show that fructose re- in pathways engaged in cognitive function and neurological processes.
models fundamental aspects of gene regulation such as differential For instance, hippocampal signature genes Pdc, Chrnb4, and Gch1 were
DNA methylation, differential gene expression, alternative splicing, previously implicated in schizophrenia (Sullivan et al., 2008), addictive
and implications for microRNA alterations, all of which have the poten- behavior (Picciotto and Kenny, 2013), attention and vigilance (Yasuda
tial to impact pathogenesis. In addition, we found that fructose alters et al., 2014), and early-onset Parkinson's disease (Cederfjall et al.,
the organization of genes in brain region-specic networks interrelating 2013). The biological pathways affected in both brain regions were
cell metabolism, immune function, inammation, and cell communica- related to ECM, focal adhesion, complement cascade, smooth muscle
tion via key regulators such as the extracellular matrix genes Bgn and contraction, and IGF signaling, which have been implicated in both
Fmod. Remarkably, we found that DHA supplementation was capable metabolic and brain disorders (Soleman et al., 2013, Frischknecht and
to reverse the genomic, epigenomic, network, and phenotypic perturba- Gundelnger, 2012, Chen et al., 2011, Phieler et al., 2013, Skaper, 2007).
tions induced by fructose via the same key drivers. A highlight of the study was the identication of the extracellular
Our ndings indicate that fructose and DHA induce transcriptomic matrix genes Bgn and Fmod as key intermediates of the fructose effects
reprogramming by engaging core transcription factors, splicing factors, on the gene network and phenotypic alterations, and these ndings
and epigenetic modications. For instance, fructose altered classic were corroborated in two knockout mouse models. Small leucine-rich
transcription factors such as Tcf7l2 (the most signicant and robust proteoglycans Fmod and Bgn consistently emerged as KDs in both
signal for T2D from human GWAS) (Voight et al., 2010) and Zbtb16. Ad- brain regions for both dietary components, and the expression levels
ditionally, N20% of the fructose and DHA signature genes only showed of both genes were also signicantly altered by fructose and DHA.
changes at isoform or alternative splicing levels but not overall expres- Fmod and Bgn are important structural components of the ECM. The
sion, and alternative splicing factors such as Bicc1, Prpf31, and Rbpms role of the ECM in the regulation of neurite outgrowth, and neural
were among the signature genes, supporting the importance of RNA functionality and plasticity (Dityatev et al., 2010) as well as metabolic
splicing in the pathogenesis of human diseases highlighted recently diseases (Soleman et al., 2013, Frischknecht and Gundelnger, 2012,
(Xiong et al., 2015). The involvement of epigenetic mechanisms in Chan et al., 2014) is becoming increasingly recognized. Our study
Please cite this article as: Meng, Q., et al., Systems Nutrigenomics Reveals Brain Gene Networks Linking Metabolic and Brain Disorders,
EBioMedicine (2016), http://dx.doi.org/10.1016/j.ebiom.2016.04.008
Q. Meng et al. / EBioMedicine xxx (2016) xxxxxx 7
Fig. 3. Gene subnetworks and top network key drivers (KDs) of fructose and DHA. (a) KDs and gene subnetwork in hypothalamus. (b) KDs and gene subnetwork in hippocampus. Larger
nodes depict KDs; grey nodes are network genes in the neighborhood of KDs that are not affected by fructose or DHA; top and bottom halves of each node, if colored, denotes genes affected
by fructose (top) and DHA (bottom), respectively; red and blue colors denotes increased and decreased expression, respectively. Direction of change for the fructose group is determined
by comparison with the control group; direction of change for the DHA + fructose group is determined by comparison with the fructose group. (For interpretation of the references to color
in this gure legend, the reader is referred to the web version of this article.)
provides unique nutrigenomic-based evidence that proteoglycans are diseases, and various metabolic diseases. As human GWAS implies a
likely main orchestrators of important metabolic and cognitive process- causal role of the candidate genes based on genetic evidence, the over-
es in the brain. lapping signatures or network KDs in response to fructose likely play
To assess the translatability of our rodent ndings to human patho- causal roles in the development of MetDs or brain disorders. The same
physiology, we found signicant intersections between the fructose sig- genes also likely mediate the benecial actions of DHA.
natures or the network KDs identied in our rodent models and human The strength of our study lies in the use of powerful systems
GWAS hits for neurological and neurodegenerative diseases, psychiatric nutrigenomic approach to comprehensively characterize phenotypic
Please cite this article as: Meng, Q., et al., Systems Nutrigenomics Reveals Brain Gene Networks Linking Metabolic and Brain Disorders,
EBioMedicine (2016), http://dx.doi.org/10.1016/j.ebiom.2016.04.008
8 Q. Meng et al. / EBioMedicine xxx (2016) xxxxxx
Fig. 4. Phenotypic validation of Bgn and Fmod using knockout (KO) mice. (a) Lipid traits including triglyceride (TG), total cholesterol, HDL cholesterol, un-esteried cholesterol (UC), LDL
cholesterol and free fatty acids (FFA). (b) Glucose. (c) Insulin. (d) Insulin resistance (IR) measured by insulin sensitivity index. (e) Intraperitoneal glucose tolerance test (IPGTT).
(f) Mistakes made during the Barnes Maze test in four days of the learning phase. (g) Track plots and mistakes made during Barnes Maze test for spatial memory. Two-sided Student's
t-test was used to test statistical difference between knockout mice (Bgn KO or Fmod KO) and wild-type (WT) mice for analyses in (a)(g). Two-way ANOVA with Holm-Sidak post
hoc analysis for multiple comparisons was performed for the learning curves in panel (f) with *WT vs. Bgn KO; #Bgn KO vs. Fmod KO; and WT vs. Fmod KO. Error bars in the plots are
standard errors. p b 0.05. Sample size n = 816.
and multi-tissue transcriptomic and methylomic proles of fructose and evidence for the potential causal role of the identied genes and
DHA, followed by network modeling, in vivo validation, and human data pathways in human diseases.
integration. The highly integrative nature not only enables a compre- We note the following limitations of the current study. First, the Bgn
hensive and in-depth exploration of the molecular mechanisms mediat- and Fmod KO animals used are not brain-specic. It is possible that the
ing the actions of nutrition on health and disease, but also pinpoints phenotypic changes in peripheral metabolism observed in these models
critical regulators of the biological pathways and provides compelling are the results of perturbations of these genes outside of the brain.
Future tissue-specic studies are warranted to further investigate the
action sites of these ECM genes. Second, the behavioral changes ob-
Table 2 served in our systemic treatment studies of fructose and DHA as well
Enrichment of human GWAS signals in fructose signatures and KDs by SNP set enrichment
as in the genetically modied animal models of Bgn and Fmod could
analysis (SSEA).
be inuenced by numerous variables that are either not investigated
Tissue GWAS GWAS p-Value p-Value or difcult to tease out. Third, we only focused on two brain regions in
disease/trait studya (KS test)b (Fisher's the current study and future investigations of additional tissues and
exact testb
cell types are warranted.
Hypothalamus HDL cholesterol GLGC 4.35E05 1.25E12 The staggering increase in the prevalence of MetDs, its negative im-
LDL cholesterol GLGC 1.05E03 1.90E05
pact on brain function and pathogenesis of brain disorders, and the role
Total cholesterol GLGC 8.51E06 3.65E14
Triglycerides GLGC 5.88E04 3.09E09 played by fructose consumption as a dietary perturbation in these pro-
Diastolic blood pressure ICBP 2.40E05 7.55E05 cesses are becoming major concerns for public health and basic and
Systolic blood pressure ICBP 3.06E03 1.16E02 clinical science. Our study provides molecular evidence supporting the
Type 2 diabetes DIAGRAM+ 2.07E02 3.39E02
ability of fructose to disrupt critical genes and fundamental physiologi-
Cognitive function FHS 1.78E02 3.18E03
Bipolar disorder WTCCC 3.06E02 4.04E02
cal processes such as insulin, IGF, TLR, TGF beta, BDNF, MAPK, NCAM,
and PDGF signaling pathways and neurological processes. The majority
Hippocampus HDL cholesterol GLGC 6.48E05 2.02E07
of these pathways have been implicated in both MetDs and brain disor-
LDL cholesterol GLGC 2.98E05 2.16E04
Total cholesterol GLGC 2.37E05 4.19E13 ders. In addition, genes affected by fructose signicantly overlapped
Triglycerides GLGC 1.34E02 6.55E06 with human GWAS genes for brain disorders (e.g., Alzheimer's disease,
Diastolic blood pressure ICBP 4.52E01 4.59E03 attention decient hyperactive disorder, depression, addiction,
Cognitive function FHS 9.43E03 4.28E02
Parkinson's disease) and MetDs-related diseases (e.g., blood pressure,
Schizophrenia CATIE 2.26E02 3.02E01
cardiovascular disease, lipids, obesity, metabolic syndrome, and T2D).
a
GLGC: Global Lipids Genetics Consortium; ICBP: The International Consortium for Our results point to the capacity of fructose to reprogram molecular
Blood Pressure; DIAGRAM +: The Diabetes Genetics Replication And Meta-analysis
Consortium; FHS: Framingham Heart Study; WTCCC: Wellcome Trust Case Control
substrates underlying pathology. The remarkable capacity of DHA sup-
Consortium; CATIE: Clinical Antipsychotic Trials for Intervention Effectiveness. plementation to restore genomic, epigenomic, pathway, network, and
b
Bolded p values are those passed Bonferroni-corrected p b 0.05. phenotypic traits vulnerable to the effects of fructose provides multi-
Please cite this article as: Meng, Q., et al., Systems Nutrigenomics Reveals Brain Gene Networks Linking Metabolic and Brain Disorders,
EBioMedicine (2016), http://dx.doi.org/10.1016/j.ebiom.2016.04.008
Q. Meng et al. / EBioMedicine xxx (2016) xxxxxx 9
level support for the potential of omega-3 fatty acids as a nutritional Ameye, L., Aria, D., Jepsen, K., Oldberg, A., Xu, T., Young, M.F., 2002. Abnormal collagen
brils in tendons of biglycan/bromodulin-decient mice lead to gait impairment,
remedy for metabolic and brain dysregulation. This points to an effec- ectopic ossication, and osteoarthritis. FASEB J. 16, 673680.
tive means to guide personalized medicine by matching the disease- Bomm, T.R., Forny-Germano, L., Sathler, L.B., Brito-Moreira, J., Houzel, J.C., Decker, H.,
promoting molecular signatures with an agent capable of reversing Silverman, M.A., Kazi, H., Melo, H.M., Mcclean, P.L., Holscher, C., Arnold, S.E., Talbot,
K., Klein, W.L., Munoz, D.P., Ferreira, S.T., De Felice, F.G., 2012. An anti-diabetes
the patterns. agent protects the mouse brain from defective insulin signaling caused by
In summary, we show that DHA reverses biological pathways and Alzheimer's disease-associated Abeta oligomers. J. Clin. Invest. 122, 13391353.
gene networks perturbed by fructose via transcriptional regulators Bremer, A.A., Stanhope, K.L., Graham, J.L., Cummings, B.P., Ampah, S.B., Saville, B.R., Havel,
P.J., 2014. Fish oil supplementation ameliorates fructose-induced hypertriglyc-
(such as transcription factors, epigenetic regulators, splicing factors) eridemia and insulin resistance in adult male rhesus macaques. J. Nutr. 144, 511.
and essential network regulators (such as Bgn and Fmod) that may con- Cederfjall, E., Nilsson, N., Sahin, G., Chu, Y., Nikitidou, E., Bjorklund, T., Kordower, J.H., Kirik,
trol the balance between health and disease. Our study reveals critical D., 2013. Continuous DOPA synthesis from a single AAV: dosing and efcacy in
models of Parkinson's disease. Sci. Rep. 3, 2157.
information supporting an action of ECM in orchestrating brain function
Chakravarti, S., 2002. Functions of lumican and bromodulin: lessons from knockout
through gene network organizations. Our experimental validation high- mice. Glycoconj. J. 19, 287293.
lights the predictive power of our systems genomics approaches in Chan, K.H., Huang, Y.T., Meng, Q., Wu, C., Reiner, A., Sobel, E.M., Tinker, L., Lusis, A.J., Yang,
identifying vulnerable disease genes and mechanisms. In addition, our X., Liu, S., 2014. Shared molecular pathways and gene networks for cardiovascular
disease and type 2 diabetes mellitus in women across diverse ethnicities. Circ.
paradigm based on the integration between rodent studies and Cardiovasc. Genet. 7, 911919.
human GWAS serves as a precedent to apply nutrigenomics principles Chassaing, B., Koren, O., Goodrich, J.K., Poole, A.C., Srinivasan, S., Ley, R.E., Gewirtz, A.T.,
to predict disease susceptibility and to guide personalized medicine. 2015. Dietary emulsiers impact the mouse gut microbiota promoting colitis and
metabolic syndrome. Nature 519, 9296.
Supplementary data to this article can be found online at http://dx. Chen, Y., Zhu, J., Lum, P.Y., Yang, X., Pinto, S., Macneil, D.J., Zhang, C., Lamb, J., Edwards, S.,
doi.org/10.1016/j.ebiom.2016.04.008. Sieberts, S.K., Leonardson, A., Castellini, L.W., Wang, S., Champy, M.F., Zhang, B.,
Emilsson, V., Doss, S., Ghazalpour, A., Horvath, S., Drake, T.A., Lusis, A.J., Schadt, E.E.,
2008. Variations in DNA elucidate molecular networks that cause disease. Nature
Role of Funding Sources 452, 429435.
Chen, D.Y., Stern, S.A., Garcia-Osta, A., Saunier-Rebori, B., Pollonini, G., Bambah-Mukku, D.,
Blitzer, R.D., Alberini, C.M., 2011. A critical role for IGF-II in memory consolidation and
XY is supported by NIH grant R01DK104363, American Heart
enhancement. Nature 469, 491497.
Association Scientist Development Grant 13SDG17290032, Leducq De Caterina, R., 2011. n-3 fatty acids in cardiovascular disease. N. Engl. J. Med. 364,
Foundation, Hellman Fellows Award, UCLA Faculty Research Grant, 24392450.
and UCLA CTSI Grant UL1TR000124. FGP is supported by NIH grants De Graaf, A.A., Freidig, A.P., De Roos, B., Jamshidi, N., Heinemann, M., Rullmann, J.A., Hall,
K.D., Adiels, M., Van Ommen, B., 2009. Nutritional systems biology modeling: from
R01DK104363, R01NS050465 and The Letten Foundation. MFY was molecular mechanisms to physiology. PLoS Comput. Biol. 5, e1000554.
supported in part by the Intramural Program of the NIDCR, NIH. XX is Dityatev, A., Schachner, M., Sonderegger, P., 2010. The dual role of the extracellular matrix
supported by NIH grants R01HG006264 and U01HG007013, and NSF in synaptic plasticity and homeostasis. Nat. Rev. Neurosci. 11, 735746.
Farooqui, A.A., Farooqui, T., Panza, F., Frisardi, V., 2012. Metabolic syndrome as a risk factor
grant 1262134. GW was supported by NSFC (91019016), NBRPC for neurological disorders. Cell. Mol. Life Sci. 69, 741762.
(2012CB316503), and the China Scholarship Council. None of the Frischknecht, R., Gundelnger, E.D., 2012. The brain's extracellular matrix and its role in
funding sources had any role in the design and conduct of the study, synaptic plasticity. Adv. Exp. Med. Biol. 970, 153171.
Guo, W., Fiziev, P., Yan, W., Cokus, S., Sun, X., Zhang, M.Q., Chen, P.Y., Pellegrini, M., 2013.
the collection, analysis, or interpretation of the data, or the preparation, BS-Seeker2: a versatile aligning pipeline for bisulte sequencing data. BMC Genomics
review, or approval of the manuscript. 14, 774.
Hindorff, L.A., Sethupathy, P., Junkins, H.A., Ramos, E.M., Mehta, J.P., Collins, F.S., Manolio,
T.A., 2009. Potential etiologic and functional implications of genome-wide association
Conicts of Interests loci for human diseases and traits. Proc. Natl. Acad. Sci. U. S. A. 106, 93629367.
Jakovcevski, M., Akbarian, S., 2012. Epigenetic mechanisms in neurological disease. Nat.
The authors declare no conict of interests. Med. 18, 11941204.
Jepsen, K.J., Wu, F., Peragallo, J.H., Paul, J., Roberts, L., Ezura, Y., Oldberg, A., Birk, D.E.,
Chakravarti, S., 2002. A syndrome of joint laxity and impaired tendon integrity in
Author Contributions lumican- and bromodulin-decient mice. J. Biol. Chem. 277, 3553235540.
Keshava Prasad, T.S., Goel, R., Kandasamy, K., Keerthikumar, S., Kumar, S., Mathivanan, S.,
Telikicherla, D., Raju, R., Shafreen, B., Venugopal, A., Balakrishnan, L., Marimuthu, A.,
QM contributed to study design, project coordination, experiments, Banerjee, S., Somanathan, D.S., Sebastian, A., Rani, S., Ray, S., Harrys Kishore, C.J., Kanth,
data analyses, writing and editing of the manuscript; XY and FGP de- S., Ahmed, M., Kashyap, M.K., Mohmood, R., Ramachandra, Y.L., Krishna, V., Rahiman,
B.A., Mohan, S., Ranganathan, P., Ramabadran, S., Chaerkady, R., Pandey, A., 2009.
signed and directed the study, and contributed to data interpretation,
Human protein reference database2009 update. Nucleic Acids Res. 37, D767D772.
writing and editing of the manuscript; EN contributed to experiments, Lowette, K., Roosen, L., Tack, J., Vanden Berghe, P., 2015. Effects of high-fructose diets on
data analysis, and writing of the manuscript; ZY, YZ, RA, YZhuang, ET, central appetite signaling and cognitive function. Front. Nutr. 2, 5.
AM, QZ, JL, MM, LO, WG, JZ, BZ, MP, TMK, XX, and MFY contributed to Lustig, R.H., Schmidt, L.A., Brindis, C.D., 2012. Public health: the toxic truth about sugar.
Nature 482, 2729.
experiments and/or data analyses, and reviewed and edited the Lyssiotis, C.A., Cantley, L.C., 2013. Metabolic syndrome: F stands for fructose and fat. Na-
manuscript. ture 502, 181182.
Makinen, V.P., Civelek, M., Meng, Q., Zhang, B., Zhu, J., Levian, C., Huan, T., Segre, A.V.,
Ghosh, S., Vivar, J., Nikpay, M., Stewart, A.F., Nelson, C.P., Willenborg, C., Erdmann, J.,
Acknowledgements Blakenberg, S., O'donnell, C.J., Marz, W., Laaksonen, R., Epstein, S.E., Kathiresan, S.,
Shah, S.H., Hazen, S.L., Reilly, M.P., Lusis, A.J., Samani, N.J., Schunkert, H.,
Quertermous, T., Mcpherson, R., Yang, X., Assimes, T.L., 2014. Integrative genomics re-
We thank Dr. Oldberg Ake from Experimental Medical Sciences, Uni- veals novel molecular pathways and gene networks for coronary artery disease. PLoS
versity of Lund, Lund, Sweden for kindly providing the Fmod knockout Genet. 10, e1004502.
mouse model for this study. Milanski, M., Arruda, A.P., Coope, A., Ignacio-Souza, L.M., Nunez, C.E., Roman, E.A.,
Romanatto, T., Pascoal, L.B., Caricilli, A.M., Torsoni, M.A., Prada, P.O., Saad, M.J.,
Velloso, L.A., 2012. Inhibition of hypothalamic inammation reverses diet-induced in-
References sulin resistance in the liver. Diabetes 61, 14551462.
Narayanan, M., Huynh, J.L., Wang, K., Yang, X., Yoo, S., Mcelwee, J., Zhang, B., Zhang, C.,
Agrawal, R., Gomez-Pinilla, F., 2012. Metabolic syndrome in the brain: deciency in Lamb, J.R., Xie, T., Suver, C., Molony, C., Melquist, S., Johnson, A.D., Fan, G., Stone,
omega-3 fatty acid exacerbates dysfunctions in insulin receptor signalling and D.J., Schadt, E.E., Casaccia, P., Emilsson, V., Zhu, J., 2014. Common dysregulation net-
cognition. J. Physiol. 590, 24852499. work in the human prefrontal cortex underlies two neurodegenerative diseases.
Agrawal, R., Noble, E., Vergnes, L., Ying, Z., Reue, K., Gomez-Pinilla, F., 2015. Dietary Mol. Syst. Biol. 10, 743.
fructose aggravates the pathobiology of traumatic brain injury by inuencing energy Newcomer, J.W., 2007. Metabolic syndrome and mental illness. Am. J. Manag. Care 13,
homeostasis and plasticity. J. Cereb. Blood Flow Metab. (Online before print). S170S177.
Akalin, A., Kormaksson, M., Li, S., Garrett-Bakelman, F.E., Figueroa, M.E., Melnick, A., O'leary, T.P., Brown, R.E., 2013. Optimization of apparatus design and behavioral measures
Mason, C.E., 2012. methylKit: a comprehensive R package for the analysis of for the assessment of visuo-spatial learning and memory of mice on the Barnes maze.
genome-wide DNA methylation proles. Genome Biol. 13, R87. Learn. Mem. 20, 8596.
Please cite this article as: Meng, Q., et al., Systems Nutrigenomics Reveals Brain Gene Networks Linking Metabolic and Brain Disorders,
EBioMedicine (2016), http://dx.doi.org/10.1016/j.ebiom.2016.04.008
10 Q. Meng et al. / EBioMedicine xxx (2016) xxxxxx
Panagiotou, G., Nielsen, J., 2009. Nutritional systems biology: denitions and approaches. Marre, M., Meitinger, T., Midthjell, K., Morken, M.A., Narisu, N., Nilsson, P., Owen,
Annu. Rev. Nutr. 29, 329339. K.R., Payne, F., Perry, J.R., Petersen, A.K., Platou, C., Proenca, C., Prokopenko, I.,
Phieler, J., Garcia-Martin, R., Lambris, J.D., Chavakis, T., 2013. The role of the complement Rathmann, W., Rayner, N.W., Robertson, N.R., Rocheleau, G., Roden, M., Sampson,
system in metabolic organs and metabolic diseases. Semin. Immunol. 25, 4753. M.J., Saxena, R., Shields, B.M., Shrader, P., Sigurdsson, G., Sparso, T., Strassburger, K.,
Picciotto, M.R., Kenny, P.J., 2013. Molecular mechanisms underlying behaviors related to Stringham, H.M., Sun, Q., Swift, A.J., Thorand, B., et al., 2010. Twelve type 2 diabetes
nicotine addiction. Cold Spring Harb. Perspect. Med. 3, a012112. susceptibility loci identied through large-scale association analysis. Nat. Genet. 42,
Rhinn, H., Fujita, R., Qiang, L., Cheng, R., Lee, J.H., Abeliovich, A., 2013. Integrative genomics 579589.
identies APOE epsilon4 effectors in Alzheimer's disease. Nature 500, 4550. Xiong, H.Y., Alipanahi, B., Lee, L.J., Bretschneider, H., Merico, D., Yuen, R.K., Hua, Y.,
Skaper, S.D., 2007. The brain as a target for inammatory processes and neuroprotective Gueroussov, S., Najafabadi, H.S., Hughes, T.R., Morris, Q., Barash, Y., Krainer, A.R.,
strategies. Ann. N. Y. Acad. Sci. 1122, 2334. Jojic, N., Scherer, S.W., Blencowe, B.J., Frey, B.J., 2015. RNA splicing. The human splic-
Smoot, M.E., Ono, K., Ruscheinski, J., Wang, P.L., Ideker, T., 2011. Cytoscape 2.8: new fea- ing code reveals new insights into the genetic determinants of disease. Science 347,
tures for data integration and network visualization. Bioinformatics 27, 431432. 1254806.
Soleman, S., Filippov, M.A., Dityatev, A., Fawcett, J.W., 2013. Targeting the neural extracel- Xu, T., Bianco, P., Fisher, L.W., Longenecker, G., Smith, E., Goldstein, S., Bonadio, J., Boskey,
lular matrix in neurological disorders. Neuroscience 253, 194213. A., Heegaard, A.M., Sommer, B., Satomura, K., Dominguez, P., Zhao, C., Kulkarni, A.B.,
Steffen, B.T., Steffen, L.M., Zhou, X., Ouyang, P., Weir, N.L., Tsai, M.Y., 2015. n-3 Fatty acids Robey, P.G., Young, M.F., 1998. Targeted disruption of the biglycan gene leads to an
attenuate the risk of diabetes associated with elevated serum nonesteried fatty osteoporosis-like phenotype in mice. Nat. Genet. 20, 7882.
acids: the multi-ethnic study of atherosclerosis. Diabetes Care 38, 575580. Yan, J., Zhang, H., Yin, Y., Li, J., Tang, Y., Purkayastha, S., Li, L., Cai, D., 2014. Obesity- and
Suez, J., Korem, T., Zeevi, D., Zilberman-Schapira, G., Thaiss, C.A., Maza, O., Israeli, D., aging-induced excess of central transforming growth factor-beta potentiates diabetic
Zmora, N., Gilad, S., Weinberger, A., Kuperman, Y., Harmelin, A., Kolodkin-Gal, I., development via an RNA stress response. Nat. Med. 20, 10011008.
Shapiro, H., Halpern, Z., Segal, E., Elinav, E., 2014. Articial sweeteners induce glucose Yang, X., Deignan, J.L., Qi, H., Zhu, J., Qian, S., Zhong, J., Torosyan, G., Majid, S., Falkard, B.,
intolerance by altering the gut microbiota. Nature 514, 181186. Kleinhanz, R.R., Karlsson, J., Castellani, L.W., Mumick, S., Wang, K., Xie, T., Coon, M.,
Sullivan, P.F., Lin, D., Tzeng, J.Y., Van Den Oord, E., Perkins, D., Stroup, T.S., Wagner, M., Lee, Zhang, C., Estrada-Smith, D., Farber, C.R., Wang, S.S., Van Nas, A., Ghazalpour, A.,
S., Wright, F.A., Zou, F., Liu, W., Downing, A.M., Lieberman, J., Close, S.L., 2008. Zhang, B., Macneil, D.J., Lamb, J.R., Dipple, K.M., Reitman, M.L., Mehrabian, M., Lum,
Genomewide association for schizophrenia in the CATIE study: results of stage 1. P.Y., Schadt, E.E., Lusis, A.J., Drake, T.A., 2009. Validation of candidate causal genes
Mol. Psychiatry 13, 570584. for obesity that affect shared metabolic pathways and networks. Nat. Genet. 41,
Svensson, L., Aszodi, A., Reinholt, F.P., Fassler, R., Heinegard, D., Oldberg, A., 1999. 415423.
Fibromodulin-null mice have abnormal collagen brils, tissue organization, and Yang, X., Zhang, B., Molony, C., Chudin, E., Hao, K., Zhu, J., Gaedigk, A., Suver, C., Zhong, H.,
altered lumican deposition in tendon. J. Biol. Chem. 274, 96369647. Leeder, J.S., Guengerich, F.P., Strom, S.C., Schuetz, E., Rushmore, T.H., Ulrich, R.G.,
Trapnell, C., Roberts, A., Goff, L., Pertea, G., Kim, D., Kelley, D.R., Pimentel, H., Salzberg, S.L., Slatter, J.G., Schadt, E.E., Kasarskis, A., Lum, P.Y., 2010. Systematic genetic and genomic
Rinn, J.L., Pachter, L., 2012. Differential gene and transcript expression analysis of analysis of cytochrome P450 enzyme activities in human liver. Genome Res. 20,
RNA-seq experiments with TopHat and cufinks. Nat. Protoc. 7, 562578. 10201036.
Tsankova, N., Renthal, W., Kumar, A., Nestler, E.J., 2007. Epigenetic regulation in psychiat- Yasuda, Y., Hashimoto, R., Ohi, K., Yamamori, H., Fujimoto, M., Umeda-Yano, S., Fujino, H.,
ric disorders. Nat. Rev. Neurosci. 8, 355367. Fukunaga, M., Horiguchi, M., Takeda, M., Ichinose, H., 2014. A functional polymor-
Tyagi, E., Zhuang, Y., Agrawal, R., Ying, Z., Gomez-Pinilla, F., 2015. Interactive actions of phism of the GTP cyclohydrolase 1 gene predicts attention performance. Neurosci.
Bdnf methylation and cell metabolism for building neural resilience under the inu- Lett. 566, 4649.
ence of diet. Neurobiol. Dis. 73, 307318. Zhang, B., Gaiteri, C., Bodea, L.G., Wang, Z., Mcelwee, J., Podtelezhnikov, A.A., Zhang, C., Xie,
Virtanen, J.K., Mursu, J., Voutilainen, S., Uusitupa, M., Tuomainen, T.P., 2014. Serum T., Tran, L., Dobrin, R., Fluder, E., Clurman, B., Melquist, S., Narayanan, M., Suver, C.,
omega-3 polyunsaturated fatty acids and risk of incident type 2 diabetes in men: Shah, H., Mahajan, M., Gillis, T., Mysore, J., Macdonald, M.E., Lamb, J.R., Bennett,
the Kuopio ischemic heart disease risk factor study. Diabetes Care 37, 189196. D.A., Molony, C., Stone, D.J., Gudnason, V., Myers, A.J., Schadt, E.E., Neumann, H.,
Voight, B.F., Scott, L.J., Steinthorsdottir, V., Morris, A.P., Dina, C., Welch, R.P., Zeggini, E., Zhu, J., Emilsson, V., 2013. Integrated systems approach identies genetic nodes
Huth, C., Aulchenko, Y.S., Thorleifsson, G., Mcculloch, L.J., Ferreira, T., Grallert, H., and networks in late-onset Alzheimer's disease. Cell 153, 707720.
Amin, N., Wu, G., Willer, C.J., Raychaudhuri, S., Mccarroll, S.A., Langenberg, C., Zhao, Y., Barrere-Cain, R.E., Yang, X., 2015. Nutritional systems biology of type 2 diabetes.
Hofmann, O.M., Dupuis, J., Qi, L., Segre, A.V., Van Hoek, M., Navarro, P., Ardlie, K., Genes Nutr. 10, 481.
Balkau, B., Benediktsson, R., Bennett, A.J., Blagieva, R., Boerwinkle, E., Bonnycastle, Zhu, J., Lum, P.Y., Lamb, J., Guhathakurta, D., Edwards, S.W., Thieringer, R., Berger, J.P., Wu,
L.L., Bengtsson Bostrom, K., Bravenboer, B., Bumpstead, S., Burtt, N.P., Charpentier, M.S., Thompson, J., Sachs, A.B., Schadt, E.E., 2004. An integrative genomics approach
G., Chines, P.S., Cornelis, M., Couper, D.J., Crawford, G., Doney, A.S., Elliott, K.S., to the reconstruction of gene networks in segregating populations. Cytogenet. Ge-
Elliott, A.L., Erdos, M.R., Fox, C.S., Franklin, C.S., Ganser, M., Gieger, C., Grarup, N., nome Res. 105, 363374.
Green, T., Grifn, S., Groves, C.J., Guiducci, C., Hadjadj, S., Hassanali, N., Herder, C., Zhu, J., Zhang, B., Smith, E.N., Drees, B., Brem, R.B., Kruglyak, L., Bumgarner, R.E., Schadt,
Isomaa, B., Jackson, A.U., Johnson, P.R., Jorgensen, T., Kao, W.H., Klopp, N., Kong, A., E.E., 2008. Integrating large-scale functional genomic data to dissect the complexity
Kraft, P., Kuusisto, J., Lauritzen, T., Li, M., Lieverse, A., Lindgren, C.M., Lyssenko, V., of yeast regulatory networks. Nat. Genet. 40, 854861.
Please cite this article as: Meng, Q., et al., Systems Nutrigenomics Reveals Brain Gene Networks Linking Metabolic and Brain Disorders,
EBioMedicine (2016), http://dx.doi.org/10.1016/j.ebiom.2016.04.008
Am J Physiol Endocrinol Metab 299: E685E694, 2010.
First published September 7, 2010; doi:10.1152/ajpendo.00283.2010. Review
AS THE GLOBAL INCIDENCE OF OBESITY continues to increase, the lower glycemic index compared with glucose (14, 60). An
search to identify the dietary components that contribute to this additional characteristic of fructose that might have been well
phenomenon is intense and ongoing. A controversial topic to suited for diabetic patients is that fructose does not stimulate
be sure, fructose has garnered considerable attention in the last insulin secretion or require insulin for the initial steps of its
decade or so as a possible contributor to this worldwide rise in hepatic metabolism. Following glycolysis to pyruvate, fructose
overweight and obesity rates (131). In addition to contributing carbon is a substrate for pyruvate dehydrogenase (PDH), an
to the caloric overconsumption and the associated energy enzyme metabolically regulated by insulin. As insulin secre-
imbalance that are currently rampant in developed countries, tion is not stimulated by fructose, substantial amounts of
the unique metabolism of fructose may hold intriguing insight fructose are anaerobically metabolized to lactate (1, 132).
into why fructose has been implicated in the etiology of a However, fructose has been cast in a much more pejorative
number of metabolic diseases (22, 126). It is, of course, light in recent years.
important to note that the human diet rarely (if ever) encounters The contribution of fructose (and particularly HFCS) to the
fructose as a single nutrient. In most cases, the dietary exposure increased incidences of obesity and the MetS is under consid-
to fructose comes through the coingestion of glucose via erable scrutiny and is hotly debated. Critical evaluations of the
sucrose (glucose and fructose) or industrial blends of fructose literature have emerging as being both for (35) and against (42)
and glucose (high-fructose corn syrup, HFCS) in ratios very dietary restriction. It is important to note that, in many of the
similar to sucrose (50% fructose, 50% glucose). There is a studies outlining the biological pathways of fructose, high
large amount of interest in the metabolic syndrome (MetS), the levels of pure fructose (60% of diet) have been utilized. The
collection of obesity-related risk factors associated with insulin outcomes observed are not necessarily applicable to the
resistance (32), in experimental and clinical obesity research. amount of fructose currently encountered in the human diet,
Initially, fructose drew great interest as a potentially beneficial particularly that fructose is typically consumed concomitantly
sweetener for patients with diabetes mellitus due to its much with glucose. Nonetheless, fructose consumption in human
subjects has been linked to each of the MetS identifying
Address for reprint requests and other correspondence: K. Adeli, The
features, which include dyslipidemia (69, 126), visceral adi-
Hospital for Sick Children, Rm 3652, 555 University Ave., Toronto, ON, M5G posity (126), insulin resistance (126), and high blood pressure
1X8 Canada (e-mail: khosrow.adeli@sickkids.ca). (for an extensive review, see Ref. 131). Although calorically
http://www.ajpendo.org 0193-1849/10 Copyright 2010 the American Physiological Society E685
Review
E686 METABOLIC EFFECTS OF FRUCTOSE
identical to glucose, fructose metabolism (hepatic metabolism signaling pathways may also contribute to the deleterious
in particular) differs from that of glucose in several important effects of fructose (Fig. 1). The goals of this review are to
ways. The rate of hepatic uptake of fructose from portal highlight emerging concepts in fructose metabolism, link these
circulation is greater than the rate of glucose uptake, and concepts to current knowledge of MetS, and identify the truly
because fructose metabolism bypasses phosphofructokinase, exciting areas with substantial research potential.
fructose metabolism is not under the regulatory control of
insulin (140). An underinvestigated aspect of fructose is the Fructose Takes Aim at the Liver
possibility that glucose, when consumed along with fructose,
may facilitate fructose uptake and vice versa, as suggested by The liver is an essential organ for the maintenance of lipid,
plasma glucose excursions measured during the consumption glucose, and hormonal homeostasis. As such, the liver is at the
of glucose-fructose-, sucrose-, and HFCS-sweetened beverages crossroads of metabolic health and disease. Since the identifi-
(124). As lipids are strongly linked to the development of cation of sterol regulatory element-binding proteins (SREBPs)
obesity and insulin resistance (2, 21, 33), the prevailing hy- by Browns and Goldsteins groups as transcription factors that
potheses concerning the mechanisms by which fructose pro- modulate lipid homeostasis (59, 145), there has been a consid-
motes MetS focus on the lipogenic nature of the sugar. The erable focus on gene expression in the regulation of intracel-
induction of lipogenesis would increase deposition of triglyc- lular carbohydrate and lipid concentrations. Fructose may ac-
eride (TG) in adipose tissue (obesity) and ectopic tissues such tivate SREBP-1c independently of insulin, which activates
as liver (hepatic steatosis) and muscle, eventually resulting in genes involved in de novo lipogenesis (DNL) (38, 81, 93).
impaired insulin signaling and dyslipidemia (125). Fructose When glucose is replaced with fructose, carbohydrate regula-
feeding has therefore been historically utilized as a model for tory element-binding protein (ChREBP) activity and nuclear
studying various aspects of hepatic dyslipidemia and insulin SREBP-1 are increased following two weeks of dietary treat-
resistance (6, 13). The results of a recent study suggest that ment (66). Prolonged consumption of fructose, but not glucose,
fructose consumption may specifically promote lipid deposi- also increased hepatic fractional DNL in humans when mea-
tion in visceral adipose tissue, particularly in men, whereas sured during energy-balanced feeding (126). The increased rate
glucose consumption appears to favor lipid deposition in sub- of fructose-induced DNL generates fatty acids that can then be
cutaneous adipose tissue (126). Interestingly, 24-hour plasma incorporated into hepatic TG or other lipid species. Increased
TG profiles are increased to a greater extent following the hepatic lipid levels are associated with increased very-low-
consumption of fructose- and HFCS-sweetened beverages density lipoprotein (VLDL) synthesis and secretion, specifi-
compared with glucose consumption (124). This suggests that cally that of VLDL1 (4). Apolipoprotein B-100 (apoB) is
the coingestion of fructose with glucose may also elicit an essential for the intracellular assembly of TG into VLDL, and
unfavorable TG profile similar to that of fructose alone. Al- apoB degradation is reduced when hepatic lipid is increased
though lipogenesis is certainly at the core of the biochemical (100), which leads to the accumulation of apoB in the hepatic
changes induced by fructose, emerging evidence suggests that endoplasmic reticulum (ER) and causes ER stress (130). Re-
the simultaneous stimulation of alternate physiological and cent experimental evidence has demonstrated that ER stress
stimulated states (94). This important study identified differ- The effects of fructose on the central nervous system may be
ential effects of PGC-1# ASO on the liver in regular-chow-fed more far reaching than its ability to elicit insulin resistance.
rats and high-fructose-fed rats. In regular-chow-fed rats, Recent studies have shown a more direct effect of fructose on
PGC-1# ASO induced slight but statistically significant reduc- the brain and in the general function of neurons. The fructose
tions in the sensitivity of hepatic glucose production to the transporter GLUT5 showed increased mRNA expression in
suppressive effects of insulin (94). The reduced insulin sensi- brains of fructose-fed rats, accompanied by increased protein
tivity was due to decreased mitochondrial fatty acid oxidation levels of GLUT5 in the hippocampus (120), implying that
and the accumulation of hepatic diacylglycerol (DAG), which neuronal fructose uptake can be increased with fructose feed-
in turn inhibited insulin signaling via activation of protein ing. Lindqvist et al. (77) have reported increased ghrelin as
kinase C$ (114). Metabolism in high-fat diet-fed rats was well as leptin levels in fructose-fed rats. Increased circulating
unaffected when treated with the same PGC-1# ASO (94). leptin may relate to development of leptin resistance upon
This evidence suggests that fructose may modulate many of its high-fructose feeding (118). This would imply that despite
metabolic effects, including stimulation of DNL, via activation increases in leptin levels there is a decreased ability to respond
of specific metabolic targets by PGC-1#. to leptin, leading to decreased satiety. This observation implies
PGC-1" and -# have been reported to coactivate both the potential for increased food intake with fructose feeding. It
PPAR" and nuclear respiratory factor (NRF)-1 (115). PPAR" should be noted, however, that the changes in ghrelin and
plays a key role in the transcriptional control of genes encoding leptin are not excusive to the one sugar, as both glucose and
mitochondrial fatty acid oxidation enzymes such as long-chain sucrose overconsumption elicited similar increases in both
acyl coenzyme-A dehydrogenase and medium-chain acyl co- leptin and ghrelin. In addition to these observations, central
enzyme-A dehydrogenase. Fructose administration (10% wt/ administration of fructose can cause increased appetite as well
vol) in the drinking water of rats reduced the transactivating as significant changes in neuronal function. Hyperphysiologi-
and transrepressing activity of PPAR" (109). The mRNA cal doses of fructose administered directly to the brain have
levels of PPAR" target genes, such as CPT I (51) and PPAR" been shown to induce increased feeding, whereas glucose
itself (45), were decreased as a result of the reduced transcrip- causes a decrease in food consumption (89). This was found to
tional activity of PPAR". Decreased PPAR" activity compro- be associated with differential induction of malonyl-CoA be-
mises fatty acid oxidation in mitochondria, a catabolic process tween the two sugars. Malonyl-CoA is known to cause an
directly controlled by CPT I (51). The accumulation of FFA in anorectic effect, which leads to the loss of appetite. Glucose
liver may then induce oxidative stress (109). The same study administration in the brain activates the glycolytic cycle, which
also observed an increased activity of proinflammatory tran- leads to increased neuronal ATP levels; this is associated with
scription factor nuclear factor-%B (NF-%B). These changes a fall in AMP levels. Since AMP is an activator of AMP kinase
were not observed in rats given glucose in their drinking water. (AMPK), there is a decrease in AMPK phosphorylation/activ-
The leptin-signal transducer and activator of transcription-3 ity. AMPK activity, in turn, catalyzes the phosphorylation/
pathway may be involved in this regulation (109). activation of acetyl-CoA carboxylase (ACC), glucose blocks
this activation and this leads to increased malonyl-CoA and
Extrahepatic Effects of Fructose Feeding decreased food intake. Fructose administered centrally de-
creases ATP levels and leads to a subsequent decrease in
In addition to well-characterized hepatic effects of fructose hypothalamic malonyl-CoA levels (27, 143). This is thought to
feeding (such as steatosis, increased lipoprotein secretion, lead to a decrease in satiety. In contrast to the above obser-
hepatic insulin resistance), fructose has been identified to alter vations, fructose given as a pre-meal load showed no dif-
biological pathways in several extrahepatic tissues, including ference in food intake compared with either glucose or
the central nervous system, adipose tissue, and the gut. These sucrose pre-meal loads (5), and all three sugars, when given
new areas of research have expanded the physiological role of to rats in drinking water, showed the same effects on food
high-fructose diets and now introduce a whole body approach consumption (77).
to investigating the metabolic fate of fructose. While direct central administration of fructose may not be a
typical physiological condition, as dietary fructose is rapidly
Central Effects of Fructose taken up and metabolized by the liver, the signaling changes
induced by fructose are intriguing and may point to larger
The role of insulin resistance in neuronal dysfunction has systemic signaling changes that could lead to alterations in
been well documented, and the observation that fructose can satiety factors. For example, fructose feeding has a well-
induce these changes is also now well documented. Several documented effect of dramatically increasing hepatic lactate
studies have linked insulin resistance and type 2 diabetes to production (1, 132). As neurons utilize lactate as well as
deficits in hippocampal declarative memory (28, 86, 128), and glucose as energy substrates, increased lactate and reduced
Stranahan et al. (129) have shown that mice on hypercaloric insulin responses seem likely signals from the periphery to the
diets have impaired memory, reduced dendritic spine density, brain conveying metabolic status following fructose intake.
and impaired long-term potentiation. When fructose-specific
neuronal effects are evaluated, fructose-fed rats and hamsters Evidence for Fructose-Induced Leptin Resistance
show hippocampal insulin resistance associated with impaired
memory retention and decreased long-term depression forma- Leptin is a well-characterized adipose-derived hormone with
tion in hippocampal neurons (87, 110). These effects are likely central and peripheral effects on food intake and dietary sub-
the result of systemic insulin resistance, as insulin is known to strate handling (for review, see Ref. 20). Leptin receptor
play a key role in neuronal synaptic plasticity (96). knockout mice (db/db) and mice that secrete mutant leptin
microbiota and increased LPS concentration (23, 25). This concentrations (24). Intestinal inflammation, absorption of LPS
hypothesis is strengthened by the findings that nonabsorbable into circulation, and gastrointestinal integrity therefore repre-
antibiotic treatment of mice consuming fructose prevented sents potential contributors to the metabolic consequences of
fructose-induced increases in LPS, hepatic lipids, and hepatic fructose consumption.
TG (15). Changes in gut microbiota are increasingly being The gut actively secretes hormones that maintain gut integ-
linked to obesity and the metabolic syndrome (for review, see rity, modulate gastric emptying, and promote insulin secretion
Refs. 26 and 74). As this is a rapidly expanding area of (34). Glucagon-like peptide-1 (GLP-1) is secreted in response
research, future investigation into the role that fructose may to nutrient ingestion (133, 134, 139) and can abolish the
have in microbiota health and species colonization should postprandial rise in TG following infusion in human subjects
provide intriguing results. Although the traditional concept of (85). Consumption of a high-fructose beverage was associated
fructose-induced metabolic dysregulation has focused on lipo- with increased TG and prolonged GLP-1 release during a
genesis and insulin regulation, these data demonstrate an al- 24-hour period (132). However, others have observed a de-
ternate method by which fructose induces insulin resistance, creased GLP-1 response in human subjects following fructose
through increased bacterial induction of an inflammatory re- ingestion compared with an equal amount of glucose (65).
sponse accompanied by steatosis. Recent evidence from our laboratory demonstrates a role for
Finally, although there is a dramatic increase in hepatic TG GLP-1 in the dyslipidemia observed following fructose feeding
following fructose feeding, it is now well established that TG of hamsters (58). Pharmaceutical inhibition of dipeptidyl pep-
per se are not responsible for the deleterious effects of hepatic tidase-4 (the enzyme responsible for GLP-1 breakdown) atten-
steatosis but that the accumulation of lipid species such as uated the fructose-induced rise in plasma TG, predominantly in
DAG (114) and ceramide (55, 113) impair insulin signaling the VLDL fraction (58). As the postprandial response to GLP-1
and may induce inflammation. Although much of these data is blunted in obesity (105) and type 2 diabetes (79), the
were derived from high-fat feeding studies, Vila et al. (138) evidence from fructose feeding suggests that increased GLP-1
have recently demonstrated that ceramides are significantly concentrations improve the metabolic response to a deleterious
increased in the livers of fructose-fed rats. As has already been diet. Although these data are intriguing, very few studies have
discussed, fructose is a potent lipogenic stimulus in the liver. focused on fructose and gut hormone secretion. This makes it
The study of lipid intermediates is an emerging area of re- difficult to draw a conclusion regarding the particular effects of
search, and more directed studies are required to unravel a fructose, but it also represents an opportunity for further
possible role for fructose. investigation into potential mechanisms of action.
Very recent evidence suggests that a single bolus of fructose
Role of the Gut in Meditating the Fructose Response initiates a positive regulatory control loop involving stimula-
tion of leptin secretion and intestinal carbohydrate transport
Increased lipoprotein secretion associated with intestinal (112). During feeding, the stomach releases leptin (11), which
insulin resistance has been identified as a factor strongly enters the intestinal lumen (49, 50). Sakar et al. (112) demon-
associated with aspects of MetS, including central adiposity strated that oral fructose increased gastric juice leptin concen-
(18, 19), insulin resistance (8, 9), and elevated fasting TG (29, trations 15 minutes after gavage without altering circulating
64). In fact, the gut may be a prime candidate for the conver- leptin concentrations. When leptin was given orally at a dose
gence of many of the emerging factors involved in fructose- similar to the leptin concentrations in gastric juice following
related disturbances. High-fructose feeding increased chylomi- fructose ingestion, increased GLUT2 and GLUT5 transport
cron production and secretion (52) and induced an insulin- activities and hepatic expression of SREBP-1c, ACC1, and
resistant state characterized by impaired IRS-1 and Akt FAS and activation of ACC1 in the liver were observed.
phosphorylation in the hamster small intestine (39). The ad- Although these results have yet to be fully explained, they
ministration of the PPAR! agonist rosiglitazone, an insulin highlight a diverse and integrated regulatory system that is
sensitizer, reduced the effect of fructose feeding in hamsters activated during the briefest of exposures to fructose. Taken
(73). Interestingly, PPAR! agonist treatment may also increase together, the role of fructose in the gut offers new and exciting
body weight and body fat distribution from visceral to subcu- opportunities for future research and therapeutic target devel-
taneous adipose tissue depots (142). Further study on meta- opment.
bolic consequences of PPAR! agonist treatment in MetS is
required. As has been previously mentioned, infusion of TNF" Concluding Remarks
increased the release of intestinal chylomicrons (104). Chylo-
microns are known to promote the intestinal absorption of LPS This review outlines several emerging molecular mecha-
(46); therefore, the increase in postprandial chylomicron levels nisms and biochemical pathways that are stimulated following
in insulin resistance may expose LPS to the portal circulation fructose exposure. High quantities of each food item would
(including mesenteric adipose tissue) and induce inflammation need to be consumed to achieve the dietary exposures repre-
and insulin resistance. This inflammatory environment may sented in many of the rodent studies. However, as this review
propagate the secretion of TNF", as has been seen following article clearly demonstrates, dietary exposure to fructose can
fructose feeding (122, 136), thus leading to enhanced chylo- potentially have substantial and profound metabolic conse-
micron production and increased exposure to LPS, establishing quences that possibly predispose individuals to chronic condi-
a vicious cycle. Conversely, treatment of fructose-fed rats with tions such as type 2 diabetes and cardiovascular disease. A
the probiotic oligofructose, an oligosaccharide that has been combination of exposure to fructose as a component of com-
suggested to promote intestinal integrity and inflammatory mon sweeteners and increased caloric intake may, in fact,
health (48, 72), blunted hepatic steatosis and circulating TG culminate in increased consumption of fructose. It is also
EB MAN
IE LI
ONN mong
BY B
o d p re ssure. A
blo
as high
n a d ia n adults h e ou t o
f ten.
fi ve C a it s n in
One out
of ties, t a t t ack
,
eir eigh o f hear
w h o r each th r r is k ts
those yo u . It acco
un
boost s ementia
e adly. It d d h
nsion is d a se , a n too ucm
H yper te r e, k id ney dise r a ilm ent. And
ilu th e
hear t fa an any o
stroke, c to r visits th
or e d o
for far m sure is
m a jo r culprit. ig h b lo o d pr e s
salt is a t of h nsion
u t 3 0 p er c en b e ll , a hyper te
s, ab o a mp od
analyse N orm C t o f B lo
In our in ta k e , says a n d p re siden
igh salt ry
due to h f C alga
e U n iv er sit y o
at th organs
ex p e r t da . d other
C a n a s , a n
P r e s s ur
e idne y atch:
e th e hear t, k e s s u r e . T he c
d a mag pr e
lt may n blood efore w
A nd s a d it s ef fec t o e d to foods b
o n d d
n d b ey eat is a
a b o ve a f th e salt we
cen t o
Photo: joanna wnuk/fotolia.com.
8 0 per .
the salt
buy the
m.
a n e e d s to cut o n p. 3
.
w h y C a nad C o n ti n u e d
d
h ow a n
Heres
CO V E R S T O RY
H ypertension is the leading risk factor for death in the world, says a new report
from Statistics Canada.1 Thats because high blood pressure boosts your risk
of dying of a heart attack or stroke more than smoking, high cholesterol, obesity,
Thats becauseunless you live in a
society where people eat very little salt
blood pressure rises as you age. In the
U.S. Atherosclerosis Risk in Communities
or any other risk factor does. study, which followed more than 15,000
Americans aged 45 to 64, average sys-
Cutting salt would make a huge difference. The beauty of reducing salt is that it tolic blood pressure (the upper number)
jumped five points in five years.6
would reduce cardiovascular disease and reduce the number of people who are
Blood pressures drift upward as
required to take blood-pressure-lowering drugs, says hypertension authority Norm people get older and theyre exposed to
Campbell. long-term excess sodium, explains U.S.
physician Stephen Havas, a former vice
Heres why youand, more importantly, the food industryshould hold the salt. president at the American Medical As-
sociation. Thats why almost all adults
are going to get blood pressures that put
them at higher risk for heart disease and
1 2
stroke.
Less salt means lower blood Nearly everyone gets high
3
pressure and less disease. blood pressure. Risk rises before your blood
Its no surprise that cutting salt lowers Why worry about salt if you havent been pressure is high.
blood pressure. That has been shown diagnosed with high blood pressure?
Doctors consider prescribing drugs when
in studies that compare higher- versus Odds are, you will be.
your blood pressure is highthat is, its
lower-salt diets in both adults and chil- Over time, 90 per cent of people in this
at least 140 over 90 (see How High Is Too
dren.2,3 country develop hypertension, says Norm
High?). But its a threat to your blood
And a recent meta-analysis of 13 Campbell, a professor of medicine, physi-
vessels before it crosses that line.
studies found not just lower blood pres- ology, and pharmacology at the University
People dont realize that blood pres-
sures, but a lower risk of heart attacks, of Calgary.
sure higher than 120 over 80 is associated
strokes, and other cardiovascular events
with increased risk, says Havas.
among people who cut their salt
intake.4 HOW HIGH IS TOO HIGH? Between normal and hyperten-
sion you have a huge number of heart
For example, in the U.S. Trials of
disease and stroke deaths attributable
Hypertension Prevention studies, some
SYSTOLIC
DIASTOLIC
5 Drugs havent solved the When doctors told them to not just take and maintain the weight loss over time,
the drugs but to also eat a lower-salt diet
problem. (1,050 milligrams a day), their blood pres-
though we should try because some
people will do it, notes Havas. We cant
So what if you get high blood pressure? sures were dramatically lower (by an aver- get people to exercise regularly and main-
Cant you just take a drug to lower it? age of 22 over 9 points) than when they tain it over time, though its worth trying
You dont want to wait until your blood were on a high-salt diet (5,750 mg a day).8 because some people will do it.
pressure crosses that magic threshold of The authors conclusion: High dietary And its not easy to get people to eat
140 over 90 because by that point youve salt ingestion is an important cause of more vegetables and fruit. If the average
already done a fair amount of damage to resistant hypertension. person eats 3 or 4 servings a day and we
your heart, vascular system, kidneys, and
need to get to 8 or 9 or more, its going to
8
dition that doesnt make the patient feel 90 per cent of Canadians eventually get
anything, explains Kaplan. If people high blood pressure, and a principal cause Salts harm goes beyond
with, say, rheumatoid arthritis dont take is exposure to excess sodium. blood pressure.
their medication, they hurt. So theyll So most people over time dont do Salt appears to damage the heart and
take that medication. But people with well with high salt loads, says Havas. blood vessels above and beyond its im-
hypertension dont experience anything Whats more, there is no predictor or test pact on blood pressure.
obvious. So they may stop. of salt sensitivity, he adds. So one has to With most animals, if you give them a
Doctors may also share some of the assume that almost all of us are sensitive high intake of salt, their blood pressure will
blame. Some physicians will put pa- to long-term sodium exposure. go up, they develop cardiovascular disease,
tients on anti-hypertensive medication, Thats why expert panels recommend and they die prematurely, explains the
and then say, Okay, Im done, without no more than 1,500 mg of sodium a University of Calgarys Norm Campbell.
monitoring to see if the patients blood day if youre middle-aged or older, are But if you genetically breed animals so
pressure is still elevated, says Kaplan. of African descent, or already have high you can feed them salt and they have no
But eating less salt would make blood blood pressurea sizeable chunk of the increase in blood pressure, they still develop
pressure drugs more effective in those population. Everyone else should shoot cardiovascular disease prematurely. That
who need them. With lower sodium for 2,300 mg of sodium a day. suggests that salt has direct toxic effects on
+ =
O Stiff arteries. Stiff arteries are often an order of Kung Pao chicken.
an early sign of heart disease. When
researchers put overweight or obese ALTERNATIVES: Add steamed or
people with normal blood pressure on a sauted vegetables to any take-
diet with a typical sodium intake (about out Chinese dish with a Szech-
3,500 mg a day), their arteries were stiffer
uan, Hunan, garlic, or other
than when they ate a lower-salt diet
(1,150 mg a day).12 strong-flavoured sauce.
These findings suggest additional
cardioprotective effects of salt reduc- Packaged Grains
tion beyond blood pressure reduction,
ADD: 1 package (250 g) Uncle
conclude the researchers.
+ =
Bens Bistro Express Wholegrain
O Kidney disease. High blood pressure Brown Rice.
damages the kidneys. But salt may make
it worse. In some studies, people who ALTERNATIVES: Add whole wheat
consume more salt excrete more protein couscous, orzo, or bulgur to any
in their urine.13 Thats a sign that their seasoned grains. Add whole-grain
kidneys are under stress. pasta to any seasoned pasta dish.
The presence of protein in the filter-
ing surfaces of the kidneys is associated
Frozen Dishes
Photos: Namita Davis (packages), Stephen Schmidt (finished dishes).
+ =
disease, a reduction in sodium can ADD: 250 g asparagus and 250 g
reduce protein in the urine. broccoli florets.
O Osteoporosis. High-salt diets increase ALTERNATIVES: Add fresh (or
calcium losses in urine. When theres
frozen) vegetables to any dish.
excess salt in the kidneys tubules, it
draws out the calcium, explains Kaplan. (Most packaged foods have too
But only a few studies have looked at few veggies.)
salts impact on bone.14 Stay tuned.
Indian Dishes
9 Small cuts in salt can save
lives and dollars. ADD: 1 can of unsalted canned
The average Canadian consumes roughly
3,500 milligrams of sodium a day. What
would be saved by cutting that intake to
1,700 mg? (Thats a typical target intake
+ chick peas (garbanzos).
)
(mg
tions. And go to sodium101.ca for more advice.
garys Norm Campbell. That would mean saving about 11,500
es
ium
ori
cardiovascular events per year. Food
Sod
Ca l
However, more Canadians are being treated for high blood Extreme Pita Market Fresh Veggie Pita (regular) 280 530
pressure now than when the initial estimates were made. In
Swiss Chalet Classic Chicken Breast & veg. medley 360 560
1999, we started a huge effort to improve hypertension control
Caseys Salmon Nioise Salad with dressing 760 760
and treatment in Canada, explains Campbell. Well rerun the
analyses when the new national control rates are released. Milestones Fire-Grilled Wild Coho Salmon & rice & veg. 740 780
But theres a downside to treating more people for high blood Boston Pizza Delicious Alternatives Pollo Pomodoro Linguini 510 820
pressure. Right now 48 per cent of women and 42 per cent of Boston Pizza Delicious Alternatives Spicy Garlic Chicken Pizza 650 860
men over 60 are on an anti-hypertensive drug, explains Camp- Starbucks Italian Cheese, Tomato & Pesto Sandwich 470 900
bell. Thats a success in the sense that those people are protect- Kelseys Veggie Burger 690 950
ed from the ravages of hypertension, but thats a huge portion of Mr. Greek Spanakopita 410 970
the population taking pharmacotherapy. Kelseys Broccoli Cheddar Soup 190 1,020
Cutting salt would trim the cost of that treatment, because Manchu Wok Green Bean Chicken with Mixed
lower blood pressure would curb the number of people who need Vegetables & rice 660 1,070
drugs.15 We would save between $430 and $540 million per Starbucks Reduced-Fat Chicken Bruschetta Sandwich 430 1,090
year purely related to a reduced need for hypertension treatment Kelseys Classic Sirloin Burger 780 1,140
and physician visits, says Campbell. Milestones Grilled Chicken Caesar Salad with dressing 580 1,160
Smaller cuts in sodium (say, to 2,300 mg a day rather than to
Mr. Greek Greek Salad with dressing 600 1,160
1,700 mg) would have less impact on heart attacks, strokes, and
Boston Pizza Vegetarian Pizza (individual) 680 1,160
heart failure.16 On the other hand, we only examined the effect
The Keg Grilled Top Sirloin (8 oz.) & buttered baked potato 1,400 1,230
of salt on blood pressure, not its direct effects on cardiovascular
disease and other health issues, says Campbell. So our num- Milestones Spinach & Feta Omelette 410 1,300
bers are an underestimate. Swiss Chalet Chalet Chicken Soup 170 1,380
Swiss Chalet Chicken Pot Pie 550 1,420
1 9 saltinstitute.org/Issues-in-focus/Food-salt-health/How-the-body-handles-salt.
Statistics Canada, Blood pressure in Canadian adults. Cat. no. 82-003-XPE, Health
10 JAMA 288: 1882, 2002.
Reports, Vol. 21, no. 1, Mar. 2010.
2 J. Human Hypertens. 16: 761, 2002. 11 Circulation 87: 476, 1993.
3 Hypertension 48: 861, 2006. 12 Am. J. Clin. Nutr. 89: 485, 2009.
4 BMJ 339: b4567, 2009. 13 J. Intern. Med. 256: 324, 2004.
5 BMJ 334: 885, 2007. 14 J. Bone Miner. Res. 23: 1477, 2008.
6 Circulation 106: 703, 2002. 15 Can. J. Cardiol. 23: 437, 2007.
7 J. Clin. Hypertens. 12: 203, 2010. 16 Can. J. Cardiol. 24: 497, 2008.
8 Hypertension 54: 475, 2009.
Review article
Medical Progress
T
he fact that salt (sodium chloride) is essential for life has From the Medical College of Wisconsin,
been recognized for millennia. Historically, the exchange value of salt played Milwaukee (T.A.K., A.W.C.); and Ochsner
Medical Center, New Orleans (E.D.F.).
an important role in establishing trade routes, securing alliances, and provok- Address reprint requests to Dr. Kotchen
ing revolutions. Homer referred to salt as a divine substance, and Plato described it at the Department of Medicine, Medical
as especially dear to the gods. Salt has been associated with sexual potency, fertil- College of Wisconsin, 8701 Watertown
Plank Rd., Milwaukee, WI 53226, or at
ity, and immortality. tkotchen@mcw.edu.
In sodium-deficient states, salt consumption is driven by salt appetite an
N Engl J Med 2013;368:1229-37.
innate and motivated behavioral response that drives a human or animal to seek DOI: 10.1056/NEJMra1212606
and ingest salt-containing foods and fluids.1 However, under usual circumstances, Copyright 2013 Massachusetts Medical Society.
the ambient salt diet is in excess of physiological need, and in humans, it has been
difficult to distinguish innate salt appetite and salt need from salt preference.2 The
hunger for salt is also influenced by taste, culture, social custom, the widespread
availability of salt, and habit independent of the need for salt.3 Despite its historical
value and physiological importance, high salt consumption has been recognized as
detrimental to health. In this article, we provide an overview of the current under-
standing of the relation of salt consumption to hypertension and cardiovascular
disease.
* The numbers in the sodium intake and blood pressure columns are average reductions. An amount of 17 mmol of sodium is equivalent to
400 mg of sodium or 1.0 g of sodium chloride.
The reduction in the blood-pressure value was significant.
ship between salt consumption and blood pres- not a binary, trait, depending on the methods
sure. Although meta-analyses are potentially used for assessment and the definition of salt
subject to criticism owing to variations in the sensitivity, approximately 30 to 50% of persons
inclusion and exclusion criteria of the trials and with hypertension and a smaller percentage of
other variations among the study protocols, sev- persons with normal blood pressure are thought
eral meta-analyses of randomized clinical trials to have salt-sensitive blood pressure.16 Pheno-
have consistently shown that persons with hy- types associated with salt-sensitive blood pres-
pertension have a greater response to a reduced sure include low-renin hypertension, older age,
salt intake than do persons with normal blood African American ethnicity, obesity, and the met-
pressure10-13 (Table 1). In a meta-analysis of 10 abolic syndrome.17,18
controlled trials involving a total of 966 children Blood-pressure responses to salt may be mod-
(median age, 13 years; range, 8 to 16), a 42% ified by other components of the diet. Low dietary
reduction in salt intake was associated with intakes of potassium and calcium potentiate the
small but significant reductions of both systolic salt sensitivity of blood pressure.19 Conversely,
pressure (1.17 mm Hg; 95% confidence interval high dietary intakes of potassium and calcium
[CI], 1.78 to 0.56) and diastolic pressure attenuate the development of salt-induced hyper-
(1.29 mm Hg; 95% CI, 1.94 to 0.65).14 tension in several animal models. In genetic ex-
Trials involving abrupt and severe salt restric-perimental models of hypertension, blood-pressure
tion have shown significant increases in plasma responses to salt are modulated by the protein,
renin activity, serum aldosterone, and plasma carbohydrate, and fat composition of the diet. In
levels of noradrenaline and adrenaline, total addition, the full expression of salt-sensitive hy-
cholesterol, and triglycerides.12,15 The implicationpertension depends on the concomitant intake of
is that these neural and hormonal responses may sodium and chloride, rather than sodium with
have adverse cardiovascular consequences. Stud- some other anion.20 However, in usual diets, it
ies assessing a long-term (>6 months) modest has been estimated that more than 85% of so-
reduction of salt intake have shown only small dium is consumed as sodium chloride.
increases in renin activity and little or no change Experimental models of hypertension provide
in sympathetic tone or plasma lipid levels. convincing evidence of a genetic susceptibility to
salt sensitivity. The most intensely studied experi-
S A LT SENSI T I V I T Y mental model of salt-sensitive hypertension has
OF BL O OD PR E SSUR E been the Dahl rat, developed by Lewis K. Dahl
nearly 50 years ago and inbred by John Rapp. In
Blood-pressure responses to salt are heteroge- consomic rats (in which otherwise genetically
neous and are normally distributed within popu- identical animals differ by one chromosome),
lations. Although salt sensitivity is a continuous, transfer of any one of several chromosomes from
Results of epidemiologic studies and random- tion was associated with increased all-cause and
ized trials suggest that potassium consumption cardiovascular mortality (median follow-up, 9.9
influences the effect of sodium on blood pres- years).45 Notably, the patients in that study had
sure and the risk of cardiovascular disease. Low multiple coexisting conditions, including renal
potassium intake is associated with an increased impairment and cardiovascular disease, at base-
risk of hypertension, and a high ratio of sodium line. Clinical trials of a low-sodium diet in com-
intake to potassium intake is a more potent risk bination with high-dose diuretic agents and
factor for hypertension and cardiovascular dis- fluid restriction in patients with congestive heart
ease than each factor alone.34,38 A high potassi- failure have been reported to show an increase
um intake offers the greatest benefit when so- in hospital readmissions and deaths.46,47 How-
dium intake is high. ever, in patients with chronic kidney disease,
All observational studies share intrinsic weak- modest salt reduction is associated with im-
nesses and methodologic limitations. None were proved clinical outcomes and greater reductions
designed to address the relation between daily so- in blood pressure in response to pharmacologic
dium intake and the risk of cardiovascular disease. inhibition of the reninangiotensin system.48
In several long-term, prospective, randomized
clinical trials, reduced salt intake was reported to MECH A NISMS OF S A LT-INDUCED
result in a decreased incidence of cardiovascular H Y PER TENSION A ND TA RGE T- ORG A N
events.39-41 In contrast, on the basis of a meta- DA M AGE
analysis of seven randomized trials (involving a
total of 6250 participants) with at least 6 months In parallel with these observational studies and
of follow-up, a 2011 Cochrane analysis conclud- clinical trials, mechanisms by which a high salt
ed that reducing dietary salt intake did not de- intake may increase blood pressure and lead to
crease the risk of death or cardiovascular dis- adverse cardiovascular outcomes have been studied
ease.42 One of the trials in the analysis included in the laboratory. Hypertension can be produced in
patients with heart failure who were simultane- response to a high dietary sodium intake in a num-
ously receiving aggressive treatment with diuretic ber of well-recognized experimentally induced
agents. In addition, trials involving persons with conditions, all of which have the common denom-
normal blood pressure and those involving per- inator of a diminution in the renal capacity to
sons with hypertension were analyzed separately, excrete sodium.49 This natriuretic handicap may
potentially resulting in lack of statistical power. be due to an intrinsic renal defect. Alternatively,
On the basis of a meta-analysis that excluded the stimuli that result in increased renal tubular re-
study in which patients received concomitant di- absorption of sodium chloride may reset the kid-
uretic therapy and that combined the normoten- neys so that a higher level of renal-arterial perfu-
sive and hypertensive study populations, He and sion pressure is required to maintain a net
MacGregor concluded that decreased salt intake sodium balance.
was associated with a significant reduction in As suggested by Guyton (cited in Cowleys
cardiovascular events and a nonsignificant re- review50), impaired natriuresis may result in a
duction in all-cause mortality.43 small increase in blood volume, and in response,
Results from trials in discrete patient popula- whole body autoregulation may explain the rise
tions suggest that caution is warranted in rec- of total peripheral resistance. Whether this se-
ommending rigorous sodium restriction for quence of events occurs in either the Dahl salt-
specific patient groups. An observational study sensitive rat or in humans with salt-sensitive
involving 2807 adults with type 1 diabetes (mean hypertension is not clear. What is clear, however,
age, 39 years) showed that dietary sodium was is that salt can activate a number of neural, en-
inversely associated with all-cause mortality and docrine or paracrine, and vascular mechanisms,
the development of end-stage renal disease (me- all of which have the potential to increase arte-
dian follow-up, 10 years).44 In that study, reduced rial pressure (Table 2); for a list of relevant
survival was also observed among adults with references, see the Supplementary Appendix,
high sodium intakes. In a related study involving available with the full text of this article at
638 patients with long-standing type 2 diabetes NEJM.org.
(mean age, 64 years), low urinary sodium excre- In rats, a high-salt diet leads to accumulation
Sodium (mg/day)
effect of salt reduction on blood pressure are
2500
based on a surrogate end point. They also note
that a reduction in sodium intake may have an 2000
adverse effect on other end points, such as level of
1500
lipids, catecholamines, renin, and aldosterone, and
that in the general population, a J-shaped curve 1000
may characterize the relationship between salt 0
consumption and cardiovascular morbidity and
00
02
04
06
08
10
20
20
20
20
20
20
mortality. Critics also note that low salt intake
9
9
9
0
19
20
20
20
20
20
increases the risk of cardiovascular events in spe-
cific patient groups (e.g., patients with congestive
Figure 2. Average Daily Sodium Consumption in the United States, 19992010.
heart failure who are aggressively treated with
Data are from the National Health and Nutrition Examination Survey. Current
diuretic agents and patients with diabetes). In recommendations of the U.S. Department of Health and Human Services
response to concerns that a low level of sodium for the general population and various subgroups are shown.
intake may adversely affect blood lipids, insulin
resistance, and the risk of cardiovascular dis-
ease, the Institute of Medicine is undertaking a pressure responses to antihypertensive drug ther-
study to evaluate the results, study design, and apy, including drug therapy in patients with
methodological approaches that have been used resistant hypertension.85,86 Most, but not all, clin-
to assess the relationship between sodium and ical trials have shown that reduced salt intake is
health outcomes.84 also associated with decreased risks of cardio-
vascular events and death. Consequently, recom-
C ONCLUSIONS mendations for reducing the currently high levels
of salt consumption in the general population
Although it has been difficult to separate salt seem justifiable, although in terms of safety, the
need from salt preference, current levels of salt lower limit of salt consumption has not been
consumption exceed salt need and are associated clearly identified. It may be premature to dis-
with adverse clinical outcomes. High salt intake count the apparently paradoxical cardiovascular
is associated with high blood pressure and in- outcomes associated with low salt intake, par-
creased rates of cardiovascular disease. Experi- ticularly in specific clinical conditions (e.g., type
mental studies continue to provide information 1 or type 2 diabetes and congestive heart failure
about mechanisms for these adverse effects of that is aggressively treated with diuretic agents).
salt. In clinical trials, a reduction in salt intake is Less-rigorous targets for salt reduction may be
associated with reduced blood pressure, more so appropriate for these and other patient groups.
in persons with hypertension than in those with
normal blood pressure. Although not discussed No potential conflict of interest relevant to this article was
reported.
in the present review, it should be noted that re- Disclosure forms provided by the authors are available with
duced salt intake is associated with greater blood- the full text of this article at NEJM.org.
REFERENCES
1. Daniels D, Fluharty SJ. Salt appetite: a sodium intake. Physiol Behav 2008;94: 4. Frost CD, Law MR, Wald NJ. By how
neurohormonal viewpoint. Physiol Behav 709-21. much does dietary salt reduction lower
2004;81:319-37. 3. Leshem M. Biobehavior of the human blood pressure? Analysis of observational
2. Morris MJ, Na ES, Johnson AK. Salt love of salt. Neurosci Biobehav Rev 2009; data within populations. BMJ 1991;302:
craving: the psychobiology of pathogenic 33:1-17. 815-8.
There is overwhelming evidence that our current high-salt During evolution, the dietary intake of salt by the human
intake is the major factor increasing blood pressure (BP) and, race and its forebears was o0.25 g/day (0.1 g sodium). The
thereby, a major cause of cardiovascular disease and kidney addition of salt to foods began only about 5000 years ago
disease worldwide. A reduction in salt intake to the when the Chinese discovered that salt could be used to
recommended level of o56 g/day is very beneficial, and preserve foods. Salt then became of great economic
could prevent millions of deaths each year and make importance, as it was possible to preserve foods during the
major savings for healthcare services. Several countries, winter and allowed for the development of settled commu-
e.g., Finland and the UK, have already reduced the amount nities. Salt was the most taxed and traded commodity in the
of salt being consumed by a combined policy of getting the world, with the intake reaching a peak around the 1870s.
food industry to decrease the amount of salt added to foods, However, with the invention of the deep freezer and the
clear labelling on food products, and increasing public refrigerator, salt was no longer required as a preservative. Salt
awareness of the harmful effects of salt on health. Many intake had been declining, but with the recent large increase
other developed countries, e.g., Australia, Canada, and the in the consumption of highly salted processed foods, salt
US, are also stepping up their activities. The major challenge intake is now increasing again. The average salt intake in
now is to spread this out worldwide, particularly to most countries around the world is E912 g/day, with many
developing countries where E80% of global BP-related Asian countries having mean intakes over 12 g/day.1
disease burden occurs. In many developing countries, most Over several million years of evolution, humans have
of the salt consumed comes from salt added during cooking become genetically programmed to a salt intake of o0.25 g/
or from sauces; therefore, public health campaigns are day. The recent dietary changes (in evolutionary terms) to the
needed to encourage consumers to use less salt. A modest high-salt intake present a major challenge to the physiological
reduction in salt intake across the whole population will systems in excreting these large amounts of salt through the
result in major improvements in public health and have huge kidneys. The consequence is that the high-salt intake causes a
economic benefits in all countries around the world. World gradual rise in blood pressure (BP),2,3 thereby increasing the
Action on Salt and Health (WASH) is a coalition of health risk of cardiovascular disease (CVD)4 and renal disease.57
professionals from different countries who know very well Furthermore, a high-salt intake may have direct effects on
the harm of high BP and has a major role in implementing stroke,8 left ventricular hypertrophy,9 progression of renal
changes in their own countries. We welcome nephrologists disease, and proteinuria,5 independent of but additive to the
to join (http://www.worldactiononsalt.com). effect of salt on BP. There is also increasing evidence that salt
Kidney International (2010) 78, 745753; doi:10.1038/ki.2010.280; intake is indirectly related to obesity through soft drink
published online 18 August 2010 consumption,10,11 associated with an increased risk of renal
KEYWORDS: cardiovascular disease; cost-effectiveness; dietary salt; renal stones and osteoporosis,12 linked to the severity of asthma,13
disease; salt reduction activities and is probably a major cause of stomach cancer.14 The
evidence on these harmful effects of salt has been compre-
hensively examined in several recent review articles.3,15 In this
paper, we provide a brief update on the evidence, particularly
that relating salt to CVD and kidney disease. Additionally, we
provide an update on the activities of reducing salt intake
occurring around the world.
0.12
1.0
0.8 0.08
0 0
Lower Higher Lower Higher TOHP II
0.10
salt salt salt salt
0.08 Control
Stroke CVD
Figure 1 | Relative risk for stroke and total cardiovascular Salt
disease (CVD) associated with a 5 g/day increase in salt 0.06
reduction
intake. Adapted from Strazzullo et al.19
0.04
that dietary salt is the major cause of raised BP and is largely 0.02
responsible for the rise in BP with age, and that a reduction
in salt intake lowers BP, thereby reducing BP-related
0
diseases.3 On the basis of the falls in BP from a meta- 2 4 6 8 10 12 14 16
analysis of randomized salt reduction trials,17 it was Follow-up (years)
estimated that a reduction of 6 g/day in salt intake would Figure 2 | Cumulative incidence of cardiovascular disease
reduce stroke by 24% and coronary heart disease by 18%. (CVD) by salt intervention group in the Trial of Hypertension
This would prevent E35,000 stroke and coronary heart Prevention (TOHP) I and II, adjusted for age, sex, and clinic.
Adapted from Cook et al.4
disease deaths a year in the UK18 and E2.5 million deaths
worldwide.
Prospective cohort studies have shown that salt intake was
related to CVD. A recent meta-analysis of 13 cohort studies 25% lower incidence of cardiovascular events after adjusting
with 177,025 participants for a follow-up period between 3.5 for confounding factors (Figure 2).4 Another outcome trial of
and 19 years showed that an increase of 5 g/day in salt intake over 2.5 years in elderly Taiwanese veterans (N 1981)
was associated with a 23% increase in the risk of stroke and a showed that switching from the usual salt (i.e., sodium
14% increase in total CVD.19 After excluding one study chloride) to potassium-enriched salt (49% sodium chloride,
which had serious methodological problems,20 the pooled 49% potassium chloride, 2% other additives), with a
analysis showed that a 5 g/day increase in salt intake was subsequent reduction of 17% in salt intake and an increase
associated with a 17% increase in CVD risk (Figure 1).19 of 76% in potassium intake, as measured by urinary sodium/
Outcome trials have demonstrated that a reduction in salt creatinine and potassium/creatinine ratio, resulted in a 40%
intake results in a decrease in CVD. A follow-up study of decrease in CVD mortality.21
individuals who took part in two large randomized salt
reduction trials, Trial of Hypertension Prevention (TOHP) I SALT REDUCTION IS ONE OF THE MOST COST-EFFECTIVE
and II, has shown a significant effect of salt reduction on INTERVENTIONS TO REDUCE CVD
CVD.4 In the original TOHP trials, over 3000 participants Several studies have demonstrated that a reduction in
with an average baseline BP of 127/85 mm Hg were population salt intake is very cost-effective in both developed
randomized to a reduced salt group (for 18 months in and developing countries.2228 Murray et al.25 carried out a
TOHP I and 3648 months in TOHP II) or to a control global analysis and showed that non-personal health inter-
group. Compared with the control group, individuals in the ventions, including government action to stimulate a
intervention group reduced their salt intake by 2530% from reduction in the salt content of processed foods, were cost-
an average of E10 g/day. This resulted in a fall in BP of 1.7/ effective ways to limit CVD and could avert over 21 million
0.9 mm Hg at 18 months (TOHP I) and 1.2/0.7 mm Hg at 36 disability-adjusted life years per year worldwide. Studies in
months (TOHP II). After the original trials were completed, Norway,22 Canada,23 and the US2628 have all demonstrated
participants were not given further dietary advice. A follow- that salt reduction not only saves lives, but also saves money.
up study at 1015 years post-trial showed that individuals For example, a recent study in the US showed that even a very
who were originally allocated to the reduced-salt group had a modest reduction in salt intake of only 10%, which could be
15% 20%
15% 20%
10 Reduction Reduction
0.40 Reduction Reduction
9 in salt intake in smoking
deaths averted (106)
0.30 prevalence
Annual cost
(20062015)
7
6
5 0.20
4
3 0.10
2
1 0.00
0
Figure 3 | Number of cardiovascular disease (CVD) deaths averted and the financial costs associated with implementation of salt
reduction and tobacco control in 23 low- and middle-income countries. Adapted from Asaria et al.24
easily achieved, as demonstrated in the UK, would prevent been shown to aggravate proteinuria and glomerulosclerosis,
hundreds of thousands of strokes and heart attacks over the and accelerate progression.29 In both partially nephrecto-
lifetimes of adults aged 4085 years, who are alive today, and mized, spontaneously hypertensive rats and in the remnant
could save 4$32 billion in medical expenses in the US kidney model of renal failure, salt restriction slowed down
alone.28 A larger decrease in salt intake would result in a the deterioration of renal impairment to a greater extent than
larger health improvement and greater cost savings.26 A thiazide diuretics.30
modest reduction in salt intake could be as beneficial as In humans, epidemiological studies have shown a direct
smoking cessation.26 association between salt intake and urinary albumin excre-
In developing countries, where the healthcare resources tion, independent of BP.31,32 A randomized double-blind trial
are very limited, a reduction in salt intake is one of the most in 40 Black hypertensive individuals demonstrated that a
cost-effective interventions to reduce CVD. Asaria et al.24 reduction in salt intake from E10 to 5 g/day reduced 24-h
estimated the effects and cost of strategies to reduce salt urinary protein by 19% (Po0.01; Figure 4).7 A more recent
intake and to control tobacco use for 23 low- and middle- double-blind trial in a larger number of individuals including
income countries that account for 80% of chronic disease 71 Whites, 69 Blacks, and 29 Asians with mildly raised BP
burden in the developing world. They demonstrated that, demonstrated that even a smaller reduction in salt intake, i.e.,
over 10 years (from 2006 to 2015), a 15% reduction in from an average of 9.7 to 6.5 g/day reduced 24-h urinary
population salt intake could avert 8.5 million CVD deaths albumin excretion significantly in all three ethnic groups.33
and a 20% reduction in smoking prevalence could avert 3.1 Other studies in patients with proteinuria or diabetes showed
million CVD deaths. The modest reduction in salt intake that the antiproteinuric effects of an angiotensin-converting
could be achieved by a voluntary reduction in the salt content enzyme inhibitor or angiotensin receptor blocker were
of processed foods and condiments by manufacturers, plus a abolished by increasing salt intake.6,34 A retrospective analysis
sustained mass-media campaign aimed to encourage dietary of 57 chronic kidney disease patients, with an average
change within households and communities. The cost for observation period of 3 years showed that a lower salt intake
implementing such salt reduction programmes was estimated reduced proteinuria and slowed down the progression of
to be US$0.09 per person per year. The cost for tobacco renal disease despite a similar BP control between the two
control including both price and non-price measures was groups on a high-and low-salt intake.5
US$0.26 per person per year (Figure 3).24 These figures In patients who are on dialysis, a reduction in salt intake
clearly suggest that a reduction in salt intake is more or at the reduces the amount of fluid that they drink between dialyses.
very least just as cost-effective as tobacco control in terms of This particularly applies to hemodialysis patients, where
reducing CVD on its own, the leading cause of death and controlling BP is a major problem. A lower salt intake
disability worldwide. decreases weight gain between dialyses and BP is easier to
control.
EVIDENCE RELATING SALT TO KIDNEY DISEASE Salt intake is one of the major dietary determinants of
AND KIDNEY STONES urinary calcium excretion.12,35 As calcium is the main
Dietary salt intake increases BP, which, in itself, can cause component of most urinary stones, salt intake is therefore
kidney disease through hypertensive glomerulosclerosis or an important cause of renal stones. Randomized trials have
through malignant hypertension. However, more frequently, demonstrated that a reduction in salt intake reduces calcium
salt causes an increase in BP in those who have underlying excretion36,37 and reduces reoccurrences of kidney stones.37
kidney diseases. This will accelerate the rate of deterioration
of renal function. There is also increasing evidence from WORLDWIDE ACTIONS ON SALT
animal experiments demonstrating that salt may have a direct The World Health Organization (WHO) has set a worldwide
effect on kidney function, independent of BP. For instance, in target of a maximum intake of 5 g/day for adults.38 Many
animal models of renal impairment, a high-salt intake has countries have developed their own guidelines and set targets
sodium
Table/cooking (15%) 1.4 g 40% reduction 0.9 g
100 Natural (5%) 0.5 g No reduction 0.5 g
Food industry (80%) 7.6 g 40% reduction 4.6 g
Total: 9.5 g Target: 6.0 g
90
The UK salt reduction campaigns started in 2003/2004 in bread has been decreased by around 20% from E1.5%
and have been successful. The salt content in many food to E1.2%.
categories has already been reduced, e.g., a one-third The salt reduction campaign has resulted in a significant
reduction in the average amount of salt in branded, pre- reduction in the average salt intake of the Finnish popula-
packed, sliced bread, and a reduction of over 40% in branded tion53,55 from E14 g/day in 1972 to o9 g/day in 2002.53 The
breakfast cereals.46 Table and cooking salt sales have also been reduction in salt intake was accompanied by a fall of over
reduced by E4050%. The average salt intake, as measured 10 mm Hg in both systolic and diastolic BP, a pronounced
by 24-h urinary sodium, in the general adult population in decrease of 7580% in both stroke and coronary heart disease
the UK has fallen from 9.5 to 8.6 g/day by May 2008.49 This mortality, and a remarkable increase of 56 years in life
change reverses the previous increasing trend for salt intake expectancy.53 The reduction in salt intake was a major
and marks the beginning of a substantial reduction in salt contributory factor for these results, particularly the fall in
intake, as increasing reductions in salt added to foods are BP as both body mass index and alcohol consumption had
being made by the food industry. The UK salt reduction increased during that period. An increase in potassium intake
campaigns, which cost just d15 million, led to E6000 fewer via the use of reduced-sodium, potassium- and magnesium-
CVD deaths per year, saving the UK economy Ed1.5 billion enriched salt, an increased consumption of fruit and
per annum.41,50 vegetables, a reduction in fat intake, and a decrease in
Clear labelling of the salt content of food is essential for smoking rate in men also played an important part in the fall
consumers to choose products with less salt. A front of pack in CVD.
signpost labelling system51 has been developed in the UK,
which uses a combination of the traffic light system, where Japan
there is a colour-coding of green, amber, and red for low, In the late 1950s, deaths from stroke in Japan were among the
medium, and high amounts of salt, fat, saturated fat, and highest in the world, and salt intake was also very high. It was
sugar, and the Guideline Daily Amount system where the found that the number of stroke in different parts of Japan
amount of salt per portion is expressed as a percentage of the was directly related to the amount of salt consumed. The
daily recommended maximum. This type of standardised Japanese government initiated a campaign to reduce salt
label is already being implemented by many supermarkets intake. Over the following decade, the national salt intake was
and is much preferred by consumers,52 as they can see at a reduced from an average of 13.5 to 12.1 g/day and in the
glance whether a product has a little or a lot of salt. Signpost north, salt intake fell dramatically from 18 to 14 g/day.
labelling has already been shown to have a dramatic effect on Paralleling this reduction in salt intake, there were falls in BP
the purchase of foods, particularly of those in the red (high) both in adults and children, and an 80% reduction in stroke
category. Despite this, perhaps owing to intense lobbying mortality56 despite large increases in fat intake, cigarette
from the food industry, the European Parliament has recently smoking, alcohol consumption, and obesity. It would appear
voted against making this scheme mandatory in Europe, and that the Western influence, which was rapidly overtaking
the shape of the final nutritional labelling legislation is Japan at that time, had little effect on BP, provided salt intake
uncertain. was reduced. Overall the reduction in salt intake appeared to
be associated with the fall in deaths from stroke.
Finland
Finland was one of the first countries to initiate a systematic The US
approach to reduce salt intake, in the late 1970s, through In the US, there has been consistent advice to reduce salt
mass-media campaigns, cooperation with the food industry intake to o6 g/day since the 1980s. However, little action has
and implementing salt labelling legislation.5355 Since the been taken until recently. In 2005, in a petition to the Food
1980s, many food companies have reduced the salt content of and Drug Administration (FDA), the Center for Science in
their food products by replacing conventional table salt with the Public Interest called for tougher regulations on salt.57 In
a sodium-reduced, potassium- and magnesium-enriched 2007, the American Medical Association published a report
mineral salt known as Pansalt. In the early 1990s, the calling for a major reduction in the salt content of processed
Ministry of Trade and Industry, and the Ministry of Social and restaurant foods.58 The FDA has been urged to cease the
Affairs and Health set salt labelling legislation for all food rule that allows salt and its component sodium to be treated
categories, which made a substantial contribution to the salt as generally recognized as safe.57,58 In November 2008,
intake of the Finnish population. Foods that are high in salt, Michael Bloomberg, the Mayor of New York City, announced
e.g., 41.3% in bread, 1.8% in sausages, 1.4% in cheese, are his plan to cut salt levels in processed foods by 20% over
required to carry a high-salt content warning and if a food 5 years. In January 2010, the National Salt Reduction
product contains a low level of salt, e.g., o0.7% in breads, Initiative (NSRI), a New York City-led partnership of more
1.2% in sausages, 0.7% in cheese, the product is allowed to than 40 cities, states, and national health organizations,
display a low-salt label. These different measures have been unveiled its proposed targets to guide a voluntary reduction
very effective and led to a marked reduction in the salt of salt levels in packaged and restaurant foods with an aim of
content of many food categories, e.g., the average salt content reducing Americans salt intake by 20% over 5 years and 40%
Figure 5 | Action groups on salt reduction around the world. The red dot indicates the location of World Action on Salt and Health
(WASH) members.
over 10 years.59 To achieve these reductions, the salt content media, and public. WASH is supported by more than 400
of processed and packaged foods will need to decrease by members from 81 countries (Figure 5).
25% in 5 years and by 50% in 10 years. Similar to the UK WASH works to reduce salt in the diet worldwide by
strategy, the US NSRI has developed voluntary 2- and 4-year exerting pressure on multi-national food companies to
interim targets for cutting salt levels in restaurant and reduce the salt content of their products. In 2009, WASH
packaged foods.59 and the International WASH members conducted a survey of
More recently, the Institute of Medicine (IOM) released over 260 branded products from KFC, McDonalds, Kelloggs,
the report Strategies to Reduce Sodium Intake in the United Nestle, Burger King, and Subway in different countries and
States.60 The report recommends setting mandatory national revealed huge variations in salt content in global brands
standards for the salt levels that food manufacturers, unrelated to local traditional taste preferences. Not one
restaurants, and food service companies can add to their product surveyed had the same salt content around the world
products in the US, and that all of the relevant health and some displayed huge differences from one country to
agencies should be involved in achieving this. The FDAs role, another. For instance, Kelloggs All Bran contained 2.15 g of
as envisioned in the report, will be to set the new standards, salt per 100 g in Canada, but only 0.65 g of salt per 100 g just
which can be phased in gradually allowing tastes to adjust over the border in the US, less than a third of the Canadian
over time, although no time scale has been given. Following level.61 This illustrates once again how easy it would be for
IOMs report, 16 food companies in the US, including food the food industry to reduce the amount of salt they add to
manufacturers Kraft Foods and HJ Heinz Co, have pledged to foods, particularly as they could do this straightaway to their
cut the amount of salt in some of their products. branded products. Pressure from campaigns such as this has
resulted in several large multi-national manufacturers pled-
WASH (World Action on Salt and Health) ging to reduce the amount of salt added to foods across the
Following the success of the UK campaign groupCASH, world. Those already committed to salt reduction worldwide
a world action groupWASH was established in 2005.61 include Pepsico, Unilever, Kelloggs, Heinz, Campbells, and
The aim of WASH is to set up similar groups modelled Kraft.
on CASH suited to each individual country to reduce salt WASH members in each country are encouraged to set up
intake with an appropriate strategy relevant to the needs of their own country division of WASH to work together on a
that particular country, and to stimulate actions from the local level to lower salt intake specifically in their own
government and/or department of health, the food industry, population. For example, in 2007, an Australian Division of
World Action on Salt and Health (AWASH) was established.62 salt intake to reach national targets or the internationally
AWASH has launched a national campaign to lower the salt recommended target of o5 g/day by 2020.
intake of the Australian population to 6 g/day by 2012. In
March 2010, the Australian government announced salt Developing countries
targets for bread and breakfast cereals. AWASH is calling on Approximately 80% of CVD deaths occur in developing
the government to take urgent action by setting salt target countries (i.e., low- and middle-income countries).69 Salt
levels for all processed and take-away foods. intake is very high70 and the healthcare resources are very
constrained in many of these countries. A reduction in salt
Other developed countries intake is probably the most important strategy to combat the
In Ireland, the Food Safety Authority set a goal to reduce salt epidemic of CVD in these countries.24
intake in the population to o6 g/day.63 Their strategic Despite the compelling evidence for a reduction in
approach includes consumer awareness efforts and actions by population salt intake, little action is happening in develop-
the food industry to lower the salt content of their food ing countries. Although action groups have been set up in
products, where to-date substantial reductions have been several countries, e.g., China, Pakistan, Bangladesh, Nepal,
made, e.g., the salt content of breakfast cereals has been Cuba, Kenya, and Ghana, many developing countries have
reduced by 11% since 2003. not developed dietary guidelines or strategies to reduce salt
In Portugal, in view of the high-salt intake and the high intake. It is important that each country determines what its
prevalence of stroke, a group of doctors formed an action salt intake is and where the major sources of salt are in the
group, the Portuguese Action Against Salt and Hypertension diet, and then implements a strategic approach to lower salt
(PAASH). They carried out a pilot study in almost 500 people intake in the population to the target level. In developing
and showed that the average salt intake was 11.9 g/day,64 countries, the major sources of salt consumption are additions
which is twice the recommended level for adults. PAASH during cooking and in sauces (e.g., soy sauce), spice mixes,
started a mass-media campaign about the harmful con- seasonings, pickles, etc. rather than pre-packaged prepared
sequences of consuming too much salt, along with an foods. Public health campaigns are needed to encourage
educational campaign directed at food manufacturers, health people to use less salt in their own food preparations.
authorities, and policy makers. This has led to the Portuguese In many developing countries, more and more processed
Parliament approving a law restricting the salt content of foods are being consumed, as their diets are becoming
bread and processed foods.65 westernized. The food industry in these countries is poorly
In Canada, the first Chair in Hypertension Prevention and regulated with very little or no food-content labelling,
Control was appointed in 2006.66 One of the major aims of making informed eating almost impossible. Therefore, many
the Chair was to get a reduction in salt intake among the developing countries need a combined policy of getting the
Canadians. The Chair is supported by a number of health public to use less salt at home and getting the food industry
organizations and scientists. Together, they lobby the to add less salt to foods and to adopt a clear labelling system
government and the food industry for policies to reduce such as the signpost labelling system.51 At the same time,
the level of salt in foods.15 In October 2007, 17 health developing countries need to ensure that all imported food
organizations endorsed a national collaborative policy products are low in salt.
statement urging reductions in salt in processed and Additionally, on behalf of WASH, many members from
packaged foods, and the Canadian government established both developed and developing nations have supported the
an expert Sodium Working Group to oversee the imple- last three World Salt Awareness Weeks held in February each
mentation of salt reduction strategies. year. Activities from both developed and developing
Many other developed countries, e.g., New Zealand, countries have ranged from involvement with hypertension
France, Italy, Holland, Belgium, Switzerland, Sweden, Den- and renal experts, representatives from the health and
mark, Slovenia, Bulgaria, Croatia, Serbia, and Barbados, are education ministries, consumers, the media, and the food
stepping up their activities to reduce salt intake. The WHO is industry. The Salt Awareness Weeks have generated signifi-
starting salt reduction strategies through its regional cant publicity and helped raise awareness of salt reduction
directorates.67 The European Union is also following suit strategies in many countries.
and 11 countries have signed up to make a 16% reduction in
salt intake over 4 years. In February 2010, the Pan American CONCLUSIONS
Health Organization/WHO Regional Expert Group on There is now overwhelming evidence for a reduction in salt
cardiovascular disease prevention through dietary salt intake in populations worldwide. Reducing salt from the
reduction68 produced a policy statement outlining the current intake of 912 g/day to the recommended level of
recommendations for a population-based approach to reduce o56 g/day will have major beneficial health effects along with
dietary salt intake in the Americans. The Policy Statement major cost savings in all countries around the world. Clearly,
provides countries with a roadmap for concerted actions by the time has now come for action and implementation.
governments, non-governmental organizations, and the food All countries should adopt a coherent and workable
industry. The goal is a gradual and sustained drop in dietary strategy to reduce salt intake. In most developed countries,
E80% of salt is hidden in foods, i.e., added by the food 9. Kupari M, Koskinen P, Virolainen J. Correlates of left ventricular mass in a
population sample aged 36 to 37 years. Focus on lifestyle and salt intake.
industry.42 It is therefore vital to persuade the food industry Circulation 1994; 89: 10411050.
to make a gradual and sustained reduction in the amount of 10. He FJ, Markandu ND, Sagnella GA et al. Effect of salt intake on renal
excretion of water in humans. Hypertension 2001; 38: 317320.
salt they add to foods. Several countries, e.g., Finland, Japan, 11. He FJ, Marrero NM, MacGregor GA. Salt intake is related to soft drink
and the UK, have successfully carried out salt reduction consumption in children and adolescents: a link to obesity? Hypertension
programmes and salt intake has already fallen.49,53 Other 2008; 51: 629634.
12. Cappuccio FP, Kalaitzidis R, Duneclift S et al. Unravelling the links
countries, e.g., Australia, Canada, the US, are also stepping between calcium excretion, salt intake, hypertension, kidney stones and
up their activities to reduce salt intake by following the UK bone metabolism. J Nephrol 2000; 13: 169177.
model. The global food industry is now poised to reduce the 13. Burney P. A diet rich in sodium may potentiate asthma. Epidemiologic
evidence for a new hypothesis. Chest 1987; 91: 143S148S.
amount of salt being added to foods and cannot justify the 14. Joossens JV, Hill MJ, Elliott P et al. Dietary salt, nitrate and stomach cancer
inequality of reducing salt levels in some countries and mortality in 24 countries. European Cancer Prevention (ECP) and the
INTERSALT Cooperative Research Group. Int J Epidemiol 1996; 25:
excluding others throughout the world. 494504.
In developing countries, where most of the salt consumed 15. Mohan S, Campbell NR. Salt and high blood pressure. Clin Sci (Lond) 2009;
comes from salt either added during cooking or from sauces, 117: 111.
16. World Health Organisation. World Health Report 2002: Reducing Risks,
public health campaigns are needed to encourage consumers Promoting Healthy Life. World Health Organisation: Geneva, Switzerland,
to use less salt. As there is an increasing trend in the 2002. http://www.who.int/whr/2002 (access verified 28 May 2010).
consumption of processed foods in many developing 17. He FJ, MacGregor GA. Effect of modest salt reduction on blood pressure:
a meta-analysis of randomized trials. Implications for public health. J Hum
countries, a combined policy of educating the public, Hypertens 2002; 16: 761770.
working with the food industry, and clear labelling, is 18. He FJ, MacGregor GA. How far should salt intake be reduced?
Hypertension 2003; 42: 10931099.
needed. A modest reduction in population salt intake
19. Strazzullo P, DElia L, Kandala NB et al. Salt intake, stroke, and
worldwide would result in a major improvement in public cardiovascular disease: meta-analysis of prospective studies. BMJ 2009;
healthsimilar to the provision of clean water and drains in 339: b4567.
20. Alderman MH, Madhavan S, Cohen H et al. Low urinary sodium is
the late nineteenth century in Europe. associated with greater risk of myocardial infarction among treated
WASH would like to extend its membership to new hypertensive men. Hypertension 1995; 25: 11441152.
members, particularly nephrologists, who are well aware of 21. Chang HY, Hu YW, Yue CS et al. Effect of potassium-enriched salt on
cardiovascular mortality and medical expenses of elderly men. Am J Clin
the dangers of salt. Joining WASH does not require any time Nutr 2006; 83: 12891296.
or financial commitments from the members, but only their 22. Selmer RM, Kristiansen IS, Haglerod A et al. Cost and health consequences
support to salt reduction worldwide. We value information of reducing the population intake of salt. J Epidemiol Community Health
2000; 54: 697702.
sharing, feedback, and input from your country and 23. Joffres MR, Campbell NR, Manns B et al. Estimate of the benefits of a
encourage members to participate in activities such as World population-based reduction in dietary sodium additives on hypertension
and its related health care costs in Canada. Can J Cardiol 2007; 23:
Salt Awareness Week, in particular, by setting up an action 437443.
group in your own country. To join, please e-mail WASH 24. Asaria P, Chisholm D, Mathers C et al. Chronic disease prevention: health
(wash@qmul.ac.uk) or visit the WASH website for more effects and financial costs of strategies to reduce salt intake and control
tobacco use. Lancet 2007; 370: 20442053.
information (http://www.worldactiononsalt.com). 25. Murray CJ, Lauer JA, Hutubessy RC et al. Effectiveness and costs of
interventions to lower systolic blood pressure and cholesterol: a global
DISCLOSURE and regional analysis on reduction of cardiovascular-disease risk. Lancet
All the authors declared no competing interests. 2003; 361: 717725.
26. Bibbins-Domingo K, Chertow GM, Coxson PG et al. Projected effect of
dietary salt reductions on future cardiovascular disease. N Engl J Med
REFERENCES 2010; 362: 590599.
1. Brown IJ, Tzoulaki I, Candeias V et al. Salt intakes around the world: 27. Palar K, Sturm R. Potential societal savings from reduced sodium
implications for public health. Int J Epidemiol 2009; 38: 791813. consumption in the U.S. adult population. Am J Health Promot 2009; 24:
2. Intersalt Cooperative Research Group. Intersalt: an international study of 4957.
electrolyte excretion and blood pressure. Results for 24 h urinary sodium 28. Smith-Spangler CM, Juusola JL, Enns EA et al. Population strategies to
and potassium excretion. BMJ 1988; 297: 319328. decrease sodium intake and the burden of cardiovascular disease: a cost-
3. He FJ, MacGregor GA. Reducing population salt intake worldwide: effectiveness analysis. Ann Intern Med 2010; 152: 481487, W170-483.
from evidence to implementation. Prog Cardiovasc Dis 2010; 52: 29. Dworkin LD, Benstein JA, Tolbert E et al. Salt restriction inhibits renal
363382. growth and stabilizes injury in rats with established renal disease. J Am
4. Cook NR, Cutler JA, Obarzanek E et al. Long term effects of dietary Soc Nephrol 1996; 7: 437442.
sodium reduction on cardiovascular disease outcomes: observational 30. Benstein JA, Feiner HD, Parker M et al. Superiority of salt restriction over
follow-up of the trials of hypertension prevention (TOHP). BMJ 2007; diuretics in reducing renal hypertrophy and injury in uninephrectomized
334: 885. SHR. Am J Physiol 1990; 258: F1675F1681.
5. Cianciaruso B, Bellizzi V, Minutolo R et al. Salt intake and renal outcome in 31. du Cailar G, Ribstein J, Mimran A. Dietary sodium and target organ
patients with progressive renal disease. Miner Electrolyte Metab 1998; 24: damage in essential hypertension. Am J Hypertens 2002; 15: 222229.
296301. 32. Verhave JC, Hillege HL, Burgerhof JG et al. Sodium intake affects urinary
6. Heeg JE, de Jong PE, van der Hem GK et al. Efficacy and variability of the albumin excretion especially in overweight subjects. J Intern Med 2004;
antiproteinuric effect of ACE inhibition by lisinopril. Kidney Int 1989; 36: 256: 324330.
272279. 33. He FJ, Marciniak M, Visagie E et al. Effect of modest salt reduction on
7. Swift PA, Markandu ND, Sagnella GA et al. Modest salt reduction reduces blood pressure, urinary albumin, and pulse wave velocity in White, Black,
blood pressure and urine protein excretion in black hypertensives: and Asian mild hypertensives. Hypertension 2009; 54: 482488.
a randomized control trial. Hypertension 2005; 46: 308312. 34. Houlihan CA, Allen TJ, Baxter AL et al. A low-sodium diet potentiates
8. Perry IJ, Beevers DG. Salt intake and stroke: a possible direct effect. J Hum the effects of losartan in type 2 diabetes. Diabetes Care 2002; 25:
Hypertens 1992; 6: 2325. 663671.
Nutrition
journal homepage: www.nutritionjrnl.com
Review
a r t i c l e i n f o a b s t r a c t
Article history: There has been much concern regarding the role of dietary fructose in the development of
Received 21 October 2009 metabolic diseases. This concern arises from the continuous increase in fructose (and total added
Accepted 26 February 2010 caloric sweeteners consumption) in recent decades, and from the increased use of high-fructose
corn syrup (HFCS) as a sweetener. A large body of evidence shows that a high-fructose diet
Keywords: leads to the development of obesity, diabetes, and dyslipidemia in rodents. In humans, fructose has
Fructose
long been known to increase plasma triglyceride concentrations. In addition, when ingested in
Metabolic diseases
large amounts as part of a hypercaloric diet, it can cause hepatic insulin resistance, increased total
and visceral fat mass, and accumulation of ectopic fat in the liver and skeletal muscle. These early
effects may be instrumental in causing, in the long run, the development of the metabolic
syndrome. There is however only limited evidence that fructose per se, when consumed in
moderate amounts, has deleterious effects. Several effects of a high-fructose diet in humans can be
observed with high-fat or high-glucose diets as well, suggesting that an excess caloric intake may
be the main factor involved in the development of the metabolic syndrome. The major source of
fructose in our diet is with sweetened beverages (and with other products in which caloric
sweeteners have been added). The progressive replacement of sucrose by HFCS is however unlikely
to be directly involved in the epidemy of metabolic disease, because HFCS appears to have basically
the same metabolic effects as sucrose. Consumption of sweetened beverages is however clearly
associated with excess calorie intake, and an increased risk of diabetes and cardiovascular diseases
through an increase in body weight. This has led to the recommendation to limit the daily intake of
sugar calories.
! 2010 Elsevier Inc. All rights reserved.
Pure, white, and deadly: the dark side of sugar was suspected Fructose metabolism has been reviewed extensively else-
many years ago, when an association between sugar consump- where [46] and will be only briefly outlined here. In the gut,
tion and coronary heart diseases was recognized and empha- fructose is transported by specific transporters, GLUT5 [7,8]. In
sized by John Yudkin [1]. Sugar, a natural sweetener obtained some subjects, fructose absorption is quantitatively limited, and
from either sugar cane or beets, is a disaccharide composed of some malabsorption occurs when large amounts of fructose are
one glucose molecule linked through an a14 glycoside bond to ingested. This can cause abdominal discomfort and diarrhea, and
a fructose molecule. Fructose, besides contributing to half the production of volatile fatty acids from colonic fructose fermen-
total content of sugar, can also be found as a hexose in fruits and tation [9,10]. Fructose absorbed from the gut into the portal vein
honey. More recently, sweeteners started to be produced from is nearly completely metabolized in the liver through metabolic
corn through starch isolation and hydrolysis to glucose, followed pathways distinct from those of glucose; furthermore, the initial
by enzymatic isomerization of part of the glucose into fructose steps of its metabolism are insulin-independent, and hence,
[2,3]. The resulting mixture, known as high-fructose corn syrup fructose is largely metabolized without requiring insulin secre-
(HFCS), has several industrial advantages over sugar, the most tion and without increasing plasma glucose. This is due to the
important being its low price, and has progressively replaced fact that 1) part of the fructose appears to be directly metabo-
sugar consumption in North America over the past 30 years. lized in enterocytes, where it is converted and into lactate and
glucose, and 2) the bulk of absorbed fructose is taken up by liver
* Corresponding author. Tel.: ++41 21 692 55 41; fax: ++41 21 692 55 95. cells, where it is rapidly converted into fructose 1-phosphate and
E-mail address: luc.tappy@unil.ch (L. Tappy). triose-phosphates through the sequential actions of fructokinase
0899-9007/$ see front matter ! 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.nut.2010.02.014
L. Tappy et al. / Nutrition 26 (2010) 10441049 1045
and aldolase B and triokinase. Fructokinase and aldolase B are lipid metabolism. In rats, intrahepatic fat content and blood very
not inhibited by ADP and citrate and hence are not regulated by low density lipoprotein (VLDL)-triacylglycerol concentrations
the cellular energy status. In that, fructose differs from glucose, increase within 6 wk of a high-fructose diet, while intramuscular
because the ADP and citrate concentrations exert a negative fat content increases within about 3 mo. Interestingly, the
feedback control on the initial steps of glycolysis. As a conse- development of insulin resistance follows a similar time course:
quence of this absence of feedback inhibition, virtually all the hepatic insulin resistance is observed early after switching to
fructose ingested with a meal (whether under its pure, unbound a high-fructose diet, while the appearance of muscle insulin
form, or bound to glucose in sucrose) is rapidly converted into resistance is delayed [19]. This suggests that fructose-induced
hepatic triose-phosphates [11]. These substrates are subse- insulin resistance is closely linked to ectopic lipid deposition
quently oxidized within the liver cells or converted into glucose and tissue-specific lipotoxicity [25].
and lactate to be released into the bloodstream, or converted into In humans, the adverse metabolic effects of fructose are less
hepatic glycogen. A small, but significant amount of trisose- clearly documented. There is ample evidence that increasing the
phsophates is also converted into triacylglycerol in liver cells fructose content of the diet increases plasma triglycerides,
through the process of de novo lipogenesis (TG) [12]. Although through several mechanisms, among which are a stimulation of
quantitatively far less important than the other pathways of hepatic de novo lipogenesis [26], and a decreased VLDL-
fructose disposal, de novo lipogenesis appears to be closely triacylglycerol clearance [27]. Based on animal experiments, an
associated with the adverse metabolic effects of fructose. increased VLDL-triacylglycerol secretion is also likely to be
Because fructose metabolism is not dependent on insulin involved [28]. The increase in plasma triglyceride induced by
secretion, at least for its initial steps, and because fructose fructose is significantly blunted in females, which suggests that
ingestion causes only a limited rise in glycemia, fructose was female sex hormones may exert a protective effect [29,30]. In
initially proposed as a natural substitute of sucrose for diabetic humans, a high-fructose, hypercaloric diet leads to ectopic fat
patients. It however became rapidly apparent that an increased deposition in liver cells and skeletal muscle in humans already
dietary intake of fructose had serious adverse metabolic effects after a 1-wk period [31]. There is strong evidence that fructose
in both rodents and humans. Thus it was recognized that a high- decreases hepatic insulin sensitivity in humans [26,32,33], and
fructose intake is associated with increased plasma triglyceride that a 10-wk high-fructose diet impairs glucose homeostasis in
concentrations, hepatic steatosis, impaired glucose tolerance and overweight individuals [34]; however, few studies have directly
insulin resistance, and even high blood pressure [4,5]. evaluated the effect of a high-fructose diet on whole body insulin
Besides these metabolic effects, fructose effects on mineral sensitivity. It has been reported that supplementation with
metabolism have also been considered. Fructose forms various amounts of fructose for up to 4 wk failed to decrease
complexes with metal ions and hence may modulate the intes- whole body insulin sensitivity in healthy individuals [31,35].
tinal absorption and bioavailability of minerals [13]. Fructose has Whether this may be different in obese insulin-resistant subjects
indeed been reported to decrease copper absorption in rats [14]. remains to be evaluated.
A diet containing up to 20% energy as fructose had however no In animal models, high-fructose diets almost invariably lead
adverse effect on copper balance in humans [15]. Fructose also to the concomitant development of excess body fat and insulin
increases iron absorption in rats [15] but does not appear to alter resistance; an increased body fat mass, body fat distribution, and
zinc bioavailability [16]. The effects of fructose on calcium ectopic lipid deposition in liver and muscle can all play a role in
metabolism have also been documented. The results indicated the development of the metabolic syndrome. In humans, it has
that rats receiving glucose-sweetened beverages had lower been recognized for several decades that abdominal fat distri-
phosphate and calcium intake and increased urinary calcium bution is associated with metabolic and cardiovascular diseases
excretion compared to the rats receiving fructose-sweetened [36], and that intravisceral fat contributes much more to insulin
beverages. These results suggest that fructose is not directly resistance than subcutaneous fat [37]. In this regard, the inter-
involved in the negative association that was observed between relationship between body fat distribution and intrahepatic fat
sugar intake and bone health [17]. Sweetened beverage content bears a special interest. Many obese patients indeed do
consumption however accounts for an important portion of not display significant metabolic disorders and have been coined
total fructose intake. In some populations, mainly children and healthy obese or metabolically fit obese [38]. When
teenagers, a high consumption of sweetened beverages may comparing these healthy obese subjects to obese subjects with
be associated with a lowered intake of milk, and hence, of metabolic complications, it appears that both intravisceral fat
calcium [18]. and intrahepatic fat content are tightly related with insulin
Several important questions regarding the role of fructose in resistance. Furthermore, intrahepatic fat content appears more
the pathogenesis of metabolic diseases remain, however, only specifically related to the metabolic syndrome [39,40]. In
partially addressed, as follows. humans, a high-fructose diet has been shown to increase
visceral fat [34] and intrahepatic fat [31], which raises the
What are the mechanisms responsible for fructose-induced possibility that the effects on hepatic fat metabolism and on
metabolic alterations? visceral fat deposits may be central in the adverse metabolic
effects of fructose. Based on these considerations, it is tempting
There is overwhelming evidence that, in rodents, a high- to speculate that a stimulation of hepatic de novo lipogenesis
sucrose or a high-fructose diet will lead to the development of induced by fructose leads to intrahepatic fat deposition,
obesity, diabetes, and dyslipidemia, and to a substantial decrease increased VLDL-triacylglycerol secretion, and hepatic insulin
in both liver and muscle insulin sensitivity [19]. Oxidative stress resistance; this may also secondarily lead to visceral fat depo-
and endoplasmic reticulum stress appear to be involved in these sition through mechanisms that remain to be determined. With
processes [2023]. Furthermore, the adverse metabolic effects time, the increased plasma VLDL-TG and an inhibition of lipid
observed with high-sucrose diets appear directly related to the oxidation induced by a high-fructose diet may favor ectopic fat
fructose component of sucrose [24]. The adverse effects of fruc- deposition in muscle, muscle lipotoxicity, and whole body
tose on glucose metabolism are closely linked to alterations of insulin resistance (Fig. 1).
1046 L. Tappy et al. / Nutrition 26 (2010) 10441049
Excess Fructose Calories stimulated lipogenic gene expression, but failed to suppress lipid
oxidation genes or to produce hepatic steatosis or dyslipidemia
[43]. In humans, it has been shown that fructose, when
Hepatic de novo
substituted for starch, increases plasma triglycerides [44]. It has
lipogenesis
also been shown that, in obese hyperinsulinemic women,
VLDL-triacylglycerol administration of equivalent amounts of pure fructose, sucrose,
secretion
and HFCS led to identical increases in plasma triglyceride,
+
VLDL- triacylglyerol although the intake of fructose was about 50% less with sucrose
clearance and HFCS compared to pure fructose [30]. In normal weight and
obese women overfed for 4 d with 50% glucose or fructose above
their energy requirement, de novo lipogenesis was found to be
hypertriacylglycerolemia identical under both conditions. There was also no significant
difference in plasma glucose, triacylglycerol, or insulin concen-
Hepatic steatosis trations [45]. In healthy young normal weight subjects, a
Ectopic lipids
Visceral fat 6-d overfeeding with 3.5 g/kg fat-free mass/d (corresponding
in skeletal muscle
roughly to 30% energy requirements) fructose or glucose
increased both VLDL-triacylglycerol concentrations and intra-
hepatic fat, with large interindividual variations, however [46].
Insulin resistance In overweight subjects who received glucose or fructose drinks
Impaired glucose tolerance corresponding to 30% of their energy requirements with ad
Metabolic syndrome
libitum food intake, body weight increased to the same extent
Fig. 1. Putative mechanisms linking excess fructose consumption to the metabolic with both sugars, while impaired glucose tolerance developed
syndrome. In liver cells, fructose stimulates de novo lipogenesis, leading to only in subjects receiving fructose. Interestingly, fructose
increased hepatic fatty acids, which can be deposited as ectopic liver fat (hepatic significantly increased visceral adipose tissue as estimated by
steatosis) or be secreted as VLDL-triacylglycerols. In addition, fructose impairs the
magnetic resonance imaging, while the increase with glucose
extrahepatic clearance of VLDL-triacylglycerols. When excess fructose calories are
consumed, this leads to hepatic steatosis and to hyper-triacylglycerolemia, which in was of smaller magnitude and failed to reach statistical
turn favors visceral fat accumulation and ectopic lipid deposition in skeletal muscle. significance [34].
The metabolic syndrome develops in the long run as a consequence of hepatic and Because all these experiments involved hypercaloric condi-
muscle lipotoxicity and of visceral obesity. tions, one might wonder whether the observed effects can be
specifically attributed to sugars or are the mere consequence of
hypercaloric feeding. Similar short-term overfeeding experi-
Fructose is also known to increase plasma uric acid, and this
ments were performed with a saturated fat supplementation
effect may be involved in the development of insulin resistance.
corresponding to 30% energy requirements: they demonstrated
In rats, fructose-induced hyperuricemia results in inhibition of
that fat overfeeding also increased intrahepatic fat content, but
NO synthase. Because insulin-induced glucose utilization
failed to increase plasma VLDL-triacylglycerol. Altogether, one
involves not only the stimulation of key metabolic pathways in
may conclude that excess calories, whether as simple sugars or as
insulin-sensitive cells but also a NO-dependent increase in
saturated fat, increases intrahepatic fat content, and that sugars
muscle blood flow [41], it was proposed that inhibition of the
specifically increase plasma VLDL-triglyceride. Fructose however
vascular effects of insulin by uric acid was involved in fructose-
appears to exert this latter effect more potently than glucose.
induced insulin resistance. In support of this hypothesis the
Such short-term experiments remain however difficult to
development of insulin resistance was prevented by lowering
extrapolate to dietary alterations of longer durations, for, with
uric acid concentrations with an uricosuric agent in fructose-fed
time, significant changes in body composition are likely to occur.
rats [42].
Are the effects of free fructose different from bound
Are the effects of fructose different from those of glucose? fructose?
In our everyday diet, naturally occurring, free fructose In North America, HFCS has replaced a substantial portion of
(essentially with fruits and honey) is a modest component of sucrose over the past three decades. The concomitant increases
energy intake. Furthermore, there is some evidence that in the HFCS intake and in the prevalence of obesity have raised
consumption of fructose with fruits or honey does not produce concern regarding a possible causal role of HFCS [47]. In rodents,
the same adverse metabolic effects as added fructose. This may feeding a diet rich in HFCS increased body weight and body fat
be due to the presence of natural antioxidants and/or dietary and caused dyslipidemia and insulin resistance. As for fructose,
fibers with fruits and honey [20,21]. The vast majority of fructose HFCS feeding elicited an endoplasmic reticulum stress response
in our diet corresponds to added sugars, the two main sources in hepatocytes. These effects of HFCS appear therefore very
being sucrose (containing 50% fructose) and HFCS (containing comparable to those of sucrose [48,49]. HFCS differs from sucrose
42%55% fructose). As a consequence, the intakes of fructose and by a slightly higher fructose content, and by providing glucose
glucose always vary simultaneously, and therefore, high-fructose and fructose as monosaccharides. There is however no evidence
consumers are also high-glucose consumers. Few studies have that providing fructose as a hexose has different metabolic
however compared the effects of high-fructose versus high- effects than when fructose is consumed bound to glucose in
glucose diets. In rodents, a high-fructose diet was shown to sucrose. In patients with type 2 diabetes, the glucose and insulin
increase the expression of lipogenic genes, to suppress peroxi- responses to administration of 35 g of sucrose or HFCS were quite
some proliferator activated receptor-alpha (PPAR-a)-dependent similar [50]. Based on the fact that fructose inhibits less ghrelin
lipid oxidation genes, and to cause intrahepatic fat deposition and increases less leptin than glucose [51], one may have
and hypertriglyceridemia. In contrast, a high-glucose diet also hypothesized that HFCS would have a different effect than
L. Tappy et al. / Nutrition 26 (2010) 10441049 1047
sucrose on these hormones and hence would have a lesser Perspectives and conclusions
satiating effect. It was however documented that it was not the
case, and that HFCS and sucrose produced similar leptin There is considerable evidence that a high-fructose intake can
increases and ghrelin suppression in healthy female volunteers indeed produce adverse metabolic alterations, the most prom-
[52]. Furthermore, HFCS, sucrose, or equimolar glucose-fructose inent ones being an increase in plasma triglycerides, hepatic
mixtures elicited similar satiety responses [53] or energy intake insulin resistance, and hepatic steatosis. These effects are
at a subsequent meal [54]. There is therefore no hint that the consistently observed in rodents fed a high-fructose diet and are
effects of free fructose may differ from those of fructose bound to generally concomitant with an increased body mass. In humans,
glucose. many of these alterations (hepatic steatosis, hepatic insulin
resistance) are observed when fructose is experimentally
Do changes in fructose consumption explain the current administered in amounts largely exceeding usual fructose intake,
epidemics of metabolic disorders? and under hypercaloric conditions. When administered as part of
a weight-maintenance diet, fructose can indeed increase plasma
There is compelling evidence that a hypercaloric, high- triglyceride. A recent meta-analysis [79] concluded that a fruc-
fructose diet can induce, not only in animal models, but also in tose intake >50 g/d was already associated with altered plasma
humans, a whole range of metabolic alterations, the most triglyceride concentrations. Fructose consumption and sweet-
prominent being a disturbance of hepatic lipid metabolism and ened beverages are closely linked, and there is overwhelming
of plasma lipid profile. This, together with the observation that evidence that sweetened beverage intake favors weight gain and
total sugar and fructose intakes have increased significantly over that excess body weight in turn increases the risk of diabetes and
the past three decades, has led to the speculation that a high- coronary heart diseases. It appears therefore sound at this stage
fructose intake may bear a direct, causal effect in the current to advise limiting consumption of sugar calories to less than 140
epidemics of obesity and related metabolic disorders [47,55]. kcal/d for men and 100 kcal/d for women (corresponding to
Several surveys, using various epidemiologic methods to assess about one can of sweetened beverage/d), as recently proposed by
dietary intakes, have consistently reported that total sugar the American Heart Association [80]. Several important ques-
consumption, added sugar intake, and fructose have indeed tions remain nonetheless unaddressed, and further studies are
increased worldwide over the past three decades [56], reviewed clearly required to better delineate the potential adverse meta-
in [6]. In the US, average fructose consumption has increased bolic effects of fructose when included in a weight-maintenance,
from about 37 g/d in the late 1970s [56] to 49 g/d in the 1999 non-hypercaloric diet, to document the interactions between
2004 period [57]. Not only fructose, but also total energy intake, fructose and other nutrients, and to evaluate whether there exist
increased during the same period. subgroups of individuals (obese insulin-resistant patients,
Consumption of sweetened beverages accounts for the major offsprings of patients with type 2 diabetes) in whom the adverse
portion of total fructose consumption, the remaining being metabolic effects of fructose may possibly be enhanced.
mainly added sugar. As a result, glucose intake always covaries
with fructose intake, and epidemiologic studies cannot differ-
Acknowledgments
entiate between the effects of fructose per se and those specifi-
cally attributable to glucose. In this regard, it is therefore more
The work performed on this topic in the authors laboratory
appropriate to evaluate the potential effect of fructose by
has been supported by Grants 067787, 109737, and 121995 from
comparing the evolutions over time of sweetened beverages and
the Swiss National Foundation for Science.
metabolic diseases. While doing so, one has however to keep in
mind that sweetened beverage consumption impacts on both
sugar consumption and total energy intake at the same time. References
Several studies have assessed the relationships between
[1] Yudkin J. Dietary carbohydrate and ischemic heart disease. Am Heart J
sweetened beverages consumption and energy intake, and
1963;66:8356.
a meta-analysis clearly documented that sweetened beverages [2] Wolf A, Bray GA, Popkin BM. A short history of beverages and how our body
consumption was associated with an excess energy intake [18]. treats them. Obes Rev 2008;9:15164.
Many cross-sectional studies have shown a positive association [3] White JS. Straight talk about high-fructose corn syrup: what it is and what
it aint. Am J Clin Nutr 2008;88:1716S21S.
between consumption of sweetened beverages, mainly in teen- [4] Le KA TL. Metabolic effects of fructose. Curr Opin Clin Nutr Metabol Care
agers, and body weight [5865] (reviewed in [66]). Interven- 2007;10:2104.
tional studies further showed that adding sweetened beverages [5] Havel PJ. Dietary fructose: implications for dysregulation of energy
homeostasis and lipid/carbohydrate metabolism. Nutr Rev 2005;63:
to the usual diet led to an increase in body weight [67,68], while 13357.
reducing sweetened beverages intake in overweight subjects [6] Tappy L, Le KA. Metabolic effects of fructose and the worldwide increase in
decreased body weight [6973]. Several large studies also obesity. Physiol Rev 2010;90:2346.
[7] Douard V, Ferraris RP. Regulation of the fructose transporter GLUT5 in
reported that sweetened beverage consumption increased the health and disease. Am J Physiol Endocrinol Metab 2008;295:E22737.
risk of developing type 2 diabetes, but this effect was essentially [8] Corpe CP, Burant CF, Hoekstra JH. Intestinal fructose absorption: clinical
linked to body weight changes [7476]. A significant association and molecular aspects. J Pediatr Gastroenterol Nutr 1999;28:36474.
[9] Kneepkens CM, Vonk RJ, Fernandes J. Incomplete intestinal absorption of
was observed between sweetened beverage consumption and an fructose. Arch Dis Child 1984;59:7358.
increased incidence of heart disease. Here again, this was mainly [10] Ravich WJ, Bayless TM, Thomas M. Fructose: incomplete intestinal
explained by a higher body weight in those consuming sweet- absorption in humans. Gastroenterology 1983;84:269.
[11] Mayes PA. Intermediary metabolism of fructose. Am J Clin Nutr
ened beverages. The relationship nonetheless remained signifi-
1993;58(suppl):754S65S.
cant after adjusting for body weight and may be explained either [12] Parks EJ, Skokan LE, Timlin MT, Dingfelder CS. Dietary sugars stimulate
by the high-glycemic index or by the high-fructose content of fatty acid synthesis in adults. J Nutr 2008;138:103946.
sweetened beverages [77]. Finally it was reported that the [13] ODell BL. Fructose and mineral metabolism. Am J Clin Nutr 1993;58:
771S8S.
consumption of sweetened beverages was higher in patients [14] Johnson MA. Interaction of dietary carbohydrate, ascorbic acid and copper
with non-alcoholic steatohepatitis than in healthy controls [78]. with the development of copper deficiency in rats. J Nutr 1986;116:80215.
REVIEW
CURRENT
OPINION Resistant starch: a promising dietary agent for the
prevention/treatment of inflammatory bowel disease
and bowel cancer
Janine A. Higgins a,b,c and Ian L. Brown a,b,c,d
Purpose of review
Resistant starch represents a diverse range of indigestible starch-based dietary carbohydrates. Resistant
starch has been investigated in the past for its effects on bowel health (pH, epithelial thickness, and
apoptosis of colorectal cancer cells); reduction in postprandial glycemia; increased insulin sensitivity; and
effects on the gut microbiome. This review highlights advances as resistant starch gains clinical relevance
as a potential treatment/preventive tool for diseases such as colorectal cancer (CRC) and diabetes.
Recent findings
Recent articles have evaluated the comparative physiological effects of different types of resistant starch
and investigated the effects of resistant starch on blood lipids, body weight, and defining resistant starch-
induced changes to the micriobiome that may be important in health and disease. The most novel and
relevant recent data describe a role for resistant starch in ameliorating inflammation; the use of resistant
starch for optimal bowel health and prevention of CRC; and, further, that the systemic effects of resistant
starch may be important for the treatment of other forms of cancer, such as breast cancer.
Summary
This review describes advances in resistant starch research highlighting the gastrointestinal effects that are
now being linked to systemic, whole body effects with clinical relevance. These effects have important
implications for overall health and the prevention or amelioration of various chronic diseases.
Keywords
cancer, inflammation, lipids, microbiome, resistant starch, weight control
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Resistant starch for CRC and weight management Higgins and Brown
body weight with and without the gut and excising on satiety over 2 h [14 ]. This is important infor-
&
individual fat pads for weighing solutions that mation as different RS4 sources can be quickly and
cannot be employed in a human population. Recent inexpensively engineered to achieve specific results
data provide contradictory evidence which is similar in terms of fermentation profile. It remains an
to that previously published. In rats, resistant starch important area of investigation to define short chain
had no effect on total body, colon, liver, or epidi- fatty acid (SCFA) profile and the subsequent meta-
dymal fat pad weights [4]. In old mice, resistant bolic effects of different forms of RS4.
starch also had no effect on total body weight [5]. Previous studies have shown that resistant
However, other studies show decreased body weight starch consumption improves insulin sensitivity
[6,7] and liver weight with no change in the weight in rats and humans. New data using RS2 in pigs
of epididymal fat pads following resistant starch [15] and overweight men with high waist circum-
feeding [7]. In a study which lends credence to ference [16] corroborate these earlier findings. In a
the idea that individual responses to resistant starch result that warrants further investigation, a sex
are highly variable in part because of the plant difference was noted in the latter study. The
source of resistant starch, RS2 added to wheat bread improvement in insulin sensitivity in response to
had no effect on body weight, but, added to maize resistant starch consumption shown in men was not
bread, RS2 caused decreased weight gain in rats [8]. apparent in an equivalent cohort of women [16]. In
A novel area of research was pursued recently by insulin-resistant adults, resistant starch intake had
examining the effect of resistant starch and exercise no effect on whole body insulin sensitivity, but did
on weight regain following weight loss in obese rats improve glucose uptake in forearm muscle and
[9 ]. Resistant starch was almost as effective as
&&
increased free fatty acid suppression in adipose tis-
exercise at ameliorating the rate of weight regain sue [17 ], indicating that resistant starch improves
&
0267-1379 ! 2013 Wolters Kluwer Health | Lippincott Williams & Wilkins www.co-gastroenterology.com 191
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Nutrition
peripheral insulin sensitivity, but not to a magni- in different segments of the bowel and correlating
tude that is detectable on the whole body level. As these with improvements in bowel function and
insulin resistance is associated with the develop- decreases in pathogenic bacteria. However, with
ment and complications of diabetes and cardio- the advent of cheaper and faster microbiome
vascular disease (CVD), it remains important to analysis, more recent work has focused on fermen-
characterize this parameter in different populations, tation and the abundance of different bacterial
especially obese individuals who are at highest risk species. This work currently stands in a descriptive
for these diseases. phase with the future open to describe which bac-
terial populations and what interactions between
INFLAMMATION different species and strains confer which pheno-
Only one current publication specifically addresses typic and health benefits.
the issue of resistant starch ingestion and inflam- In a study confirming previous rodent work,
mation. As inflammation is a characteristic feature resistant starch increased intestinal nutrient flow
of many chronic diseases such as obesity, diabetes, and caused selective proliferation of Bifidobacterium
some cancers, and CVD, this is an area that has the [19] and Bacillus species [20] in pigs. In rats, RS2
potential to expand rapidly. RS3 ingestion in wild- selectively increases propionate concentrations and
type mice decreased ileal and colonic inflammatory Ruminococcus bromii abundance, whereas RS4 feed-
lesions and increased IL-10-expressing cells and ing increases butyrate and propionate concen-
peroxisome proliferator-activated receptor gamma trations and is associated with proliferation of
expression relative to a low fat diet [18 ]. In IL-10-/- Lactobacillus gasseri and Parabacteroides distasonis
&&
mice (model of cholitis), under the same conditions, [21]. A human study has recently confirmed that
RS3 caused lower ileal and colonic inflammatory RS4, in the form of butyrylated starch, leads to
lesions and lower IFNg in lesions [18 ]. These are higher butyrate delivery to the colon than RS2,
&&
promising data which provide evidence that dietary which, overall, increases both butyrate and propio-
resistant starch may play a role in attenuating some nate concentrations in the colon compared to an
of the inflammatory responses in chronic disease equivalent dose of RS2 [22].
and provide a therapeutic role for resistant starch A problem in this arena of research has been
consumption in the future. the variability of response to resistant starch
consumption. Since a particular type of resistant
LIPIDS starch can have a specific effect on bacteria, the
The effect of resistant starch on circulating lipids has description of the source of resistant starch in a
long been an area that is understudied and subject to particular experiment is essential for the interpret-
contradictory evidence. Data from rodents show ation of the data. Individual SCFA responses to
quite convincingly that chronic resistant starch resistant starch are highly variable [23] as is the
ingestion reduces circulating cholesterol and trigly- amount of energy that is salvaged from fermenta-
ceride concentrations. However, data from human tion [24 ]. Two recent publications give a clue to
&
studies have always been equivocal, possibly due to the source of this variability. The first describes
the length of resistant starch feeding and the study in-vitro experiments showing that resistant starch
populations examined. Most recently, pig studies from the same class (RS3) but derived from differ-
show that RS2 feeding reduces fasting serum trigly- ent plant sources in conjunction with different
ceride and cholesterol concentrations [7,15]. In bacteria yield very different SCFA levels [25]. The
addition, resistant starch caused lower liver triacyl- second indicates that R. bromii is a key species for
glycerol [15] and cholesterol concentrations [7], fermentation of resistant starch in the human
and lower adipose triglyceride concentrations [7]. colon and that R. bromii facilitates resistant starch
Hepatic expression of lipogenic genes was lower, fermentation by other species, even by bacteria
whereas expression of genes of the lipolytic pathway that are weak RS2 and RS3 fermenters in isolation
was higher in response to resistant starch [15]. (Eubacterium rectal, Bacteroides thetaiotaomicron)
Thus, this is the field that requires more clinical [24 ]. As every individual has a unique microbiome
&
data before a sound judgment can be made regard- footprint and consumes resistant starch from a
ing any effect of resistant starch on lipid concen- variety of classes and plant sources, individual
trations. responses to resistant starch consumption would
be expected to be very high. Perhaps, in future
BOWEL EFFECTS: FERMENTATION AND studies, assessing the basal microbiome of the
MICROBIOME patient and using very well characterized resistant
Previous work in the area of bowel health has starch sources would glean some more consistent
focused on measuring pH and SCFA concentrations and interpretable data.
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Resistant starch for CRC and weight management Higgins and Brown
concentrations in the serum and along the length feeding significantly reduced mean tumor size,
of the colon and the highest rates of apoptosis, lowered circulating estradiol concentrations, and
relative to RS2 and control starches [29], and that, decreased proliferation and increased apoptotic
in the raw form, RS2 causes the greatest fermenta- protein and mRNA expression, relative to controls
tion and lowest aberrant crypt formation, but, under [34 ]. There was a positive correlation between
&&
cooked conditions, RS4 yields a greater amount of circulating estrogen and mean tumor size. These
fermentation and a lower aberrant crypt formation are exciting new data which pave the way for future
[30]. In a study comparing colon damage from resistant starch research beyond the realm of CRC.
moderate fat and protein (Western diet) feeding
in rats, RS2 and RS4 had equivalent effects, causing
less single-stranded DNA breaks and higher SCFA CONCLUSION
concentrations in the portal vein and cecum than Much recent data corroborate earlier evidence for
rats fed a low fat or Western diet containing diges- the reduction in glycemia, improvement in insulin
tible starch [4]. sensitivity, and positive bowel effects, including
The only clinical trial to examine the role of amelioration of CRC, that occur with resistant
resistant starch in CRC prevention was designed to starch ingestion. The most prominent recent work
test the effects of both aspirin and resistant starch on includes the first data from breast cancer studies in
polyp number and size in individuals with familial rodents, the positive effects of resistant starch in
adenomatous polyposis, which is a rare and aggres- preventing weight regain after weight loss in obese
sive form of CRC. This study showed no effect of rodents, and an initial study defining the effects of
resistant starch supplementation on either polyp resistant starch on systemic inflammation. These
load or maximal polyp size over 19 years [31]. three areas are of potential clinical relevance and
The length of the study and the high dose of warrant further examination. Should the descriptive
0267-1379 ! 2013 Wolters Kluwer Health | Lippincott Williams & Wilkins www.co-gastroenterology.com 193
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Nutrition
13. Bodinham CL, Smith L, Wright J, et al. Dietary fibre improves first-phase
data regarding the microbiome effects of resistant insulin secretion in overweight individuals. PloS One 2012; 7:e40834; Epub
starch translate in prospective, intervention studies, 2012/07/21..
14. Haub MD, Louk JA, Lopez TC. Novel resistant potato starches on glycemia and
further uses for resistant starch in the treatment of & satiety in humans. J Nutr Metab 2012; 2012:478043; Epub 2012/06/02..
colitis and other inflammatory bowel diseases might This is the only study which describes the glycemic and satiety effects of several
forms of RS4 vs. RS2.
develop in the future. Resistant starch sources can be 15. He J, Chen D, Zhang K, Yu B. A high-amylopectin diet caused hepatic
manipulated to achieve the optimal desired physio- steatosis associated with more lipogenic enzymes and increased serum
insulin concentration. Br J Nutr 2011; 106:14701475.
logical response or to target drug delivery to the 16. Maki KC, Pelkman CL, Finocchiaro ET, et al. Resistant starch from high-
colon. Resistant starch can be incorporated into amylose maize increases insulin sensitivity in overweight and obese men.
J Nutr 2012; 142:717723.
foods such as breakfast cereals, pizza bases, breads, 17. Robertson MD, Wright JW, Loizon E, et al. Insulin-sensitizing effects on
candy, pasta, yogurt, cheeses, and smoothies. Thus, & muscle and adipose tissue after dietary fiber intake in men and women with
metabolic syndrome. J Clin Endocrinol Metab 2012; 97:33263332; Epub
resistant starch is a versatile carbohydrate that 2012/06/30..
potentially offers an inexpensive, easy, and practical This study eloquently describes the tissue-specific effect of resistant starch on
insulin sensitivity in a patient population which is at greatest risk of diabetic
way to impact health and prevent or treat many complications and CVD.
chronic diseases in a broad population. 18. Bassaganya-Riera J, DiGuardo M, Viladomiu M, et al. Soluble fibers and
&& resistant starch ameliorate disease activity in interleukin-10-deficient mice
with inflammatory bowel disease. J Nutr 2011; 141:13181325.
Acknowledgements This is the first comprehensive investigation of resistant starch and inflammation
which paves the way for future studies that have ramifications for the prevention
None. and/or treatment of inflammatory bowel disease.
19. Regmi PR, Metzler-Zebeli BU, Ganzle MG, et al. Starch with high amylose
content and low in vitro digestibility increases intestinal nutrient flow and
Conflicts of interest microbial fermentation and selectively promotes bifidobacteria in pigs. J Nutr
2011; 141:12731280.
There are no conflicts of interest. 20. Han GQ, Xiang ZT, Yu B, et al. Effects of different starch sources on Bacillus
Supported in part by the National Institutes of Health spp. in intestinal tract and expression of intestinal development related genes
of weanling piglets. Mol Biol Rep 2012; 39:18691876.
(DK 038088 and DK43607; Colorado CTSI Grant UL1 21. J Abell GC, Christophersen CT, McOrist AL, Clarke JM. Dietary resistant and
RR025780). Contents are the authors sole responsibility butyrylated starches have different effects on the faecal bacterial flora of azox-
ymethane-treated rats. Br J Nutr 2011; 105:14801485 [Epub 2011/01/25].
and do not necessarily represent official NIH views. 22. Clarke JM, Topping DL, Christophersen CT, et al. Butyrate esterified to starch
is released in the human gastrointestinal tract. Am J Clin Nutr 2011;
94:12761283.
23. McOrist AL, Miller RB, Bird AR, et al. Fecal butyrate levels vary widely among
REFERENCES AND RECOMMENDED individuals but are usually increased by a diet high in resistant starch. J Nutr
READING 2011; 141:883889.
Papers of particular interest, published within the annual period of review, have 24. Ze X, Duncan SH, Louis P, Flint HJ. Ruminococcus bromii is a keystone
been highlighted as: & species for the degradation of resistant starch in the human colon. ISME J
& of special interest 2012 [Epub 2012/02/22].
&& of outstanding interest This publication identifies one key species of bacteria that could alter the
Additional references related to this topic can also be found in the Current fermentation profile of many types of resistant starch and, therefore, alter SCFA
World Literature section in this issue (p. 237). production.
25. Purwani EY, Purwadaria T, Suhartono MT. Fermentation RS3 derived from
1. Asp NG, Bjorck I, Holm J, et al. Enzyme resistant starch fractions and dietary sago and rice starch with Clostridium butyricum BCC B2571 or Eubacterium
fibre. Scand J Gastroenterol Suppl 1987; 129:2932. rectale DSM 17629. Anaerobe 2012; 18:5561.
2. Englyst KN, Liu S, Englyst HN. Nutritional characterization and measurement 26. Cruz-Bravo RK, Guevara-Gonzalez R, Ramos-Gomez M, et al. Fermented
of dietary carbohydrates. Eur J Clin Nutr 2007; 61 (Suppl 1):S1939. nondigestible fraction from common bean (Phaseolus vulgaris L.) cultivar Negro
3. Topping D. The contribution of resistant starch to health benefits of dietary 8025 modulates HT-29 cell behavior. J Food Sci 2011; 76:17503841.
fibre. J Japan Assoc Dietary Fiber Res 2012; 16 (Suppl A):s11s19. 27. OCallaghan NJ, Toden S, Bird AR, et al. Colonocyte telomere shortening is
4. Conlon MA, Kerr CA, McSweeney CS, et al. Resistant starches protect greater with dietary red meat than white meat and is attenuated by resistant
against colonic DNA damage and alter microbiota and gene expression in rats starch. Clin Nutr 2012; 31:6064.
fed a Western diet. J Nutr 2012; 142:832840. 28. Winter J, Nyskohus L, Young GP, et al. Inhibition by resistant starch of red
5. Zhou J, Keenan MJ, Keller J, et al. Tolerance, fermentation, and cytokine & meat-induced promutagenic adducts in mouse colon. Cancer Prev Res (Phila)
expression in healthy aged male C57BL/6J mice fed resistant starch. Molec 2011; 4:19201928; Epub 2011/09/03.
Nutr Food Res 2012; 56:515518; Epub 2011/12/17. This study provides mechanisic data on the role of resistant starch in ameliorating
6. Zhu L, Gu M, Meng X, et al. High-amylose rice improves indices of animal the proliferation of CRC cells.
health in normal and diabetic rats. Plant Biotechnol J 2012; 10:353362. 29. Clarke JM, Young GP, Topping DL, et al. Butyrate delivered by butyrylated
7. Ha AW, Han GJ, Kim WK. Effect of retrograded rice on weight control, gut starch increases distal colonic epithelial apoptosis in carcinogen-treated rats.
function, and lipid concentrations in rats. Nutr Res Pract 2012; 6:1620; Carcinogenesis 2012; 33:197202.
Epub 2012/03/14.. 30. Zhao Y, Hasjim J, Li L, et al. Inhibition of azoxymethane-induced preneoplastic
8. Brites CM, Trigo MJ, Carrapico B, et al. Maize and resistant starch enriched lesions in the rat colon by a cooked stearic acid complexed high-amylose
breads reduce postprandial glycemic responses in rats. Nutr Res 2011; cornstarch. J Agric Food Chem 2011; 59:97009708; Epub 2011/07/26.
31:302308. 31. Burn J, Bishop DT, Chapman PD, et al. A randomized placebo-controlled
9. Higgins JA, Jackman MR, Brown IL, et al. Resistant starch and exercise prevention trial of aspirin and/or resistant starch in young people with familial
&& independently attenuate weight regain on a high fat diet in a rat model of adenomatous polyposis. Cancer Prev Res 2011; 4:655665.
obesity. Nutr Metab 2011; 8:49; Epub 2011/07/09.. 32. Pu H, Chen L, Li X, et al. An oral colon-targeting controlled release system
This is the first study to comprehensively show the independent and additive effects based on resistant starch acetate: synthetization, characterization, and pre-
of resistant starch consumption on weight maintenance after weight loss in obesity. paration of film-coating pellets. J Agric Food Chem 2011; 59:57385745.
10. Keenan MJ, Martin RJ, Raggio AM, et al. High-amylose resistant starch 33. Chen L, Pu H, Li X, Yu L. A novel oral colon-targeting drug delivery system
increases hormones and improves structure and function of the gastrointest- based on resistant starch acetate. J Control Release 2011; 30:e51e52.
inal tract: a microarray study. J Nutrigenet Nutrigenom 2012; 5:2644; Epub 34. Kondegowda NG, Meaney MP, Baker C, Ju YH. Effects of nondigestible
2012/04/21. && carbohydrates on the growth of estrogen-dependent human breast cancer
11. Hallstrom E, Sestili F, Lafiandra D, et al. A novel wheat variety with elevated (MCF-7) tumors implanted in ovariectomized athymic mice. Nutr Cancer
content of amylose increases resistant starch formation and may beneficially 2011; 63:5564.
influence glycaemia in healthy subjects. Food Nutr Res 2011; 55:22. This is the first article which describes an effect of resistant starch on a cancer
12. Karupaiah T, Aik CK, Heen TC, et al. A transgressive brown rice mediates outside of the colon, namely breast cancer. These data provide evidence that the
favourable glycaemic and insulin responses. J Sci Food Agric 2011; systemic effects of resistant starch ingestion may be useful to treat many types of
91:19511956. cancer and chronic diseases.
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
www.thelancet.com
Nutrition specific Optimum fetal and child nutrition and development Nutrition sensitive
interventions programmes and approaches
and programmes 1Agriculture and food security
1Adolescent health and Breastfeeding, nutrient- $$' + $ Low burden of 1% "().$)(
preconception nutrition rich foods, and eating practices, parenting, infectious diseases 1'". "development
1)'$"dietary routine stimulation 1)'$"#$)"")
supplementation 1Womens empowerment
1 '%$*)' $) 1Child protection
supplementation or %%(*' )., including $$' + $ Access to and use of 1Classroom education
fortification availability, economic resources (maternal, health services, a safe and 1Water and sanitation
1'() $$ access, and use of food household, and hygienic environment 1
")$# ".&"$$ $('+ (
complementary feeding community levels)
1
)'.(*&&"#$)) %$
for children
1
)'.diversification Building an enabling environment
Knowledge and evidence 1 %'%*(+"*) %$(
1 $+ %*'($
Politics and governance 1Advocacy strategies
stimulation
Leadership, capacity, and financial resources 1
%' /%$)"$+') "%%' $) %$
1Treatment of severe acute
Social, economic, political, and environmental context (national and global) 1Accountability, incentives regulation,
malnutrition
1
((&'+$) %$$ legislation
management 1'( &&'%'##(
1*)' ) %$ $)'+$) %$( $ 1Capacity investments
emergencies 1
%#() '(%*'#% " () %$
Figure 1: Framework for actions to achieve optimum fetal and child nutrition and development
2 www.thelancet.com
Executive Summary
An unfinished agenda for undernutrition productivity losses, losses via poorer cognition, and
The publication of The Lancet Maternal and Child losses via reduced schooling.20 We cannot aord for
Undernutrition Series 5 years ago stimulated a nothing to change.
tremendous increase in political commitment to
reduction of undernutrition at global and national Burden of nutritional conditions
levels. Most development agencies have revised their Undernutrition in LMICs
strategies to address undernutrition focused on the Stunted linear growth has become the main indicator of
1000 days during pregnancy and the first 2 years of life, childhood undernutrition, because it is highly prevalent
as called for in the 2008 Series. One of the main drivers in nearly all LMICs, and has important consequences for
of this new international commitment is the Scaling Up health and development. It should replace underweight
Nutrition (SUN) movement.18,19 National commitment as the main anthropometric indicator for children. The
in LMICs is growing, donor funding is rising, and civil prevalence of stunting in children younger than 5 years
society and the private sector are increasingly engaged. in LMICs in 2011 was 26%, a decrease from 40% in
However, this progress has not yet translated 1990, and 32% in 2005, the estimate in the previous
into substantially improved outcomes globally. nutrition Series.1,6 The number of stunted children has
Improvements in nutrition still represent a massive also decreased globally, from 253 million in 1990, to 178
unfinished agenda. The 165 million children with million in 2005, to 165 million in 2011. This represents
stunted growth have compromised cognitive an average annual rate of reduction of 21%.6
development and physical capabilities, making yet The World Health Assembly (WHA) called for a 40%
another generation less productive than they would reduction in the global number of children younger
otherwise be.6 Countries will not be able to break out than 5 years who are stunted by 2025 (compared with
of poverty and sustain economic advances without the baseline of 2010).21 This aim would translate into a
ensuring that their populations are adequately 39% reduction per year and imply reducing the number
nourished. Undernutrition reduces a nations economic of stunted children from 171 million in 2010, to about
advancement by at least 8% because of direct 100 million in 2025.6 At the present rate of decline,
www.thelancet.com 3
Executive Summary
stunting is expected to reduce to 127 million, a 25% from 56 to 61 million, whereas Asia is projected to show
a substantial decrease in stunting prevalence.
The prevalence of wasting was 8% globally in 2011,
the largest number of children aected by stunting, aecting 52 million children younger than 5 years, an
69 million, live in south-central Asia. In Africa, only small 11% decrease from an estimated 58 million in 1990.6
improvements are anticipated on the basis of present The prevalence of severe wasting was 29%, aecting
trends, with the number of aected children increasing 19 million children.6 70% of the worlds children with
wasting live in Asia, mostly in south-central Asia, where
Attributable Proportion of Attributable Proportion of an estimated 15% (28 million) are aected.6
deaths with total deaths deaths with total deaths of
UN of children NIMS children
Deficiencies of essential vitamins and minerals
prevalences* younger than prevalences younger than are widespread and have substantial adverse eects
5 years 5 years
on child survival and development.6 Deficiencies of
Fetal growth restriction (<1 month) 817 000 118% 817 000 118%
vitamin A and zinc adversely aect child health and
Stunting (159 months) 1 017 000* 147% 1 179 000 170%
survival, and deficiencies of iodine and iron, together
Underweight (159 months) 999 000* 144% 1 180 000 170%
Wasting (159 months) 875 000* 126% 800 000 115% with stunting, contribute to children not reaching their
Severe wasting (159 months) 516 000* 74% 540 000 78% developmental potential. Much progress has been made
Zinc deficiency (1259 months) 116 000 17% 116 000 17% in addressing vitamin A deficiency but eorts must
Vitamin A deficiency (659 months) 157 000 23% 157 000 23% continue at present coverage levels to avoid regressing
Suboptimum breastfeeding 804 000 116% 804 000 116%
(023 months)
because dietary intake of vitamin A is still inadequate.
Joint eects of fetal growth restriction 1 348 000 194% 1 348 000 194% Additionally, micronutrient deficiencies have an
and suboptimum breastfeeding in important part to play in maternal health.6
neonates
Joint eects of fetal growth restriction, 3 097 000 447% 3 149 000 454% Breastfeeding practices are far from optimum,
suboptimum breastfeeding, stunting, despite improvements in some countries. Suboptimum
wasting, and vitamin A and zinc
deficiencies (<5 years) breastfeeding results in an increased risk for mortality
in the first 2 years of life and results in 800 000 deaths
Data are to the nearest thousand. *Prevalence estimates from the UN. Prevalence estimates from Nutrition Impact
Model Study (NIMS). annually.6
Table 1: Global deaths in children younger than 5 years attributed to nutritional disorders
Maternal, newborn, and child nutrition
New evidence further reinforces the importance of the
Key messages on disease burden due to nutritional conditions
nutritional status of women at the time of conception
Iron and calcium deficiencies contribute substantially to maternal deaths and during pregnancy, both for the health of the mother
Maternal iron deficiency is associated with babies with low weight (<2500 g) at birth
and for ensuring healthy fetal growth and development.
Maternal and child undernutrition, and unstimulating household environments,
contribute to deficits in childrens development and health and productivity in adulthood 32 million babies are born small-for-gestational-age
Maternal overweight and obesity are associated with maternal morbidity, preterm (SGA) annuallyrepresenting 27% of all births in LMICs.
birth, and increased infant mortality Fetal growth restriction causes more than 800 000
Fetal growth restriction is associated with maternal short stature and underweight deaths each year in the first month of lifemore than
and causes 12% of neonatal deaths
a quarter of all newborn deaths.6 This new finding
Stunting prevalence is slowly decreasing globally, but aected at least 165 million
contradicts the widespread assumption that babies who
Suboptimum breastfeeding results in more than 800 000 child deaths annually are born SGA, by contrast with preterm babies, are not
Undernutrition, including fetal growth restriction, suboptimum breastfeeding, at a substantially increased risk of mortality. Neonates
stunting, wasting, and deficiencies of vitamin A and zinc, cause 45% of child deaths, with fetal growth restriction are also at substantially
resulting in 31 million deaths annually
increased risk of being stunted at 24 months and of
Prevalence of overweight and obesity is increasing in children younger than 5 years
globally and is an important contributor to diabetes and other chronic diseases development of some types of non-communicable
in adulthood diseases in adulthood.6
Undernutrition during pregnancy, aecting fetal growth, and the first 2 years of life is Undernutrition (fetal growth restriction, suboptimum
a major determinant of both stunting of linear growth and subsequent obesity and breastfeeding, stunting, wasting, and deficiencies
non-communicable diseases in adulthood
of vitamin A and zinc) causes 45% of all deaths of
4 www.thelancet.com
Executive Summary
www.thelancet.com 5
Executive Summary
zinc supplementation in children aged 659 months, 90% coverage in 34 countries is Int$96 billion per
management of severe acute malnutrition (SAM), and year (table 2).7 Of the $96 billion, $37 billion (39%)
management of moderate acute malnutrition. is for micronutrient interventions, $09 billion (10%)
Continued investment in nutrition-specific inter- for educational interventions, and $26 billion (27%)
ventions and delivery strategies to reach poor segments for management of SAM. The remaining $23 billion
of the population at greatest risk can make a substantial (24%) accounts for provision of food for pregnant
dierence. If these ten proven nutrition-specific women and children aged 623 months in poor
interventions were scaled-up from existing population households. Since many interventions are being scaled
coverage to 90%, an estimated 900 000 lives could be
saved in 34 high nutrition-burden countries (where 90% the cost per discounted life-year saved is about
of the worlds stunted children live, figure 3) and the $370 ($213 per undiscounted life-year saved).
prevalence of stunting could be reduced by 20% and that More than half the $96 billion is accounted for by
of severe wasting by 60%. This would reduce the number two large countries which will rely heavily on domestic
of children with stunted growth and development by resources (India and Indonesia). Consumables (drugs,
33 million.7 On top of existing trends, this improvement or other items such as for transport or administration)
would comfortably reach the WHA targets for 2025. account for a little less than half of the $96 billion, and
all but the poorest countries can be expected to cover
Cost of scaling up proven interventions most of the expenditures on personnel. Therefore,
We estimate that the cost of scaling-up this package $34 billion from external donors could make a
of ten essential nutrition-specific interventions to substantial dierence to child nutrition
Delivery platforms: Community delivery platforms, integrated management of childhood illnesses, child health days, school-based
delivery platforms, financial platforms, fortification strategies, nutrition in emergencies
# =Interventions modelled
$s=Other interventions reviewed
6 www.thelancet.com
Executive Summary
Afghanistan
Pakistan
Iraq Bangladesh
Egypt
Chad Nepal Myanmar
The promise of emerging interventions and delivery insecurity, and scarcity of access to adequate care
strategies and platforms resourcesand include nutrition goals and actions. They
Delivery strategies are crucial to achieving coverage can therefore help enhance the eectiveness, coverage,
with nutrition-specific interventions and reaching and scale of nutrition-specific interventions.
populations in need. A range of channels can provide Our review of potentially nutrition-sensitive
opportunities for scaling up and reaching large programmes in agriculture, social safety nets, early child
population segments, such as fortification of staple
foods and conditional and unconditional cash transfers.7
Number of lives Cost per life-year
Community delivery platforms for nutrition education saved* saved
and promotion, integrated management of childhood Optimum maternal nutrition during pregnancy
illness, school-based delivery platforms, and child health Maternal multiple micronutrient supplements to all 102 000 $571 (3981191)
days are other possible channels. Calcium supplementation to mothers at risk of low intake (49 000146 000)
Maternal balanced energy protein supplements as needed
Innovative delivery strategiesespecially community- Universal salt iodisation
based delivery platformsare promising for scaling Infant and young child feeding
up coverage of nutrition interventions and have the Promotion of early and exclusive breastfeeding for 6 months and 221 000 $175 (132286)
continued breastfeeding for up to 24 months (135 000293 000)
potential to reach poor and dicult to access populations Appropriate complementary feeding education in food secure
through communication and outreach strategies.7 These populations and additional complementary food supplements in
food insecure populations
could also lead to potential integration of nutrition with Micronutrient supplementation in children at risk
maternal, newborn, and child health interventions, Vitamin A supplementation between 6 and 59 months age 145 000 $159 (106766)
helping to achieve reductions in inequities. Preventive zinc supplements between 12 and 59 months of age (30 000216 000)
Management of acute malnutrition
Management of moderate acute malnutrition 435 000 $125 (119152)
Unlocking the potential of nutrition-sensitive Management of severe acute malnutrition (285 000482 000)
programmes Data are number (95% CI) or cost in 2010 international dollars (95% CI).
In addition to nutrition-specific interventions, packages are scaled up at once. Cost per life-year saved assumes that a life saved of a child younger than 5 years saves on
average 59 life-years, based on WHO data (2011188) that life expectancy at birth on average in low-income countries is 60,
acceleration of progress in nutrition will also require and that most deaths of children younger than 5 years occur in the first year of life. To convert to cost per discounted life-
increases in the nutritional outcomes of eective, large- year saved multiply these estimates by 59/32 (ie, 184). Intervention has eect on maternal or child morbidity, but no
direct eect on lives saved. Cost per life-year saved by management of severe acute malnutrition only, costs for
scale, nutrition-sensitive development programmes.8 supplementary feeding for moderate acute malnutrition are currently unavailable.
Nutrition-sensitive programmes address key underlying
Table 2: Eect of packages of nutrition interventions at 90% coverage
determinants of nutritionsuch as poverty, food
www.thelancet.com 7
Executive Summary
development, and schooling confirms that programmes the exception of eects on vitamin A intake and status
in these sectors are successful at addressing several of from homestead food production programmes and
the underlying determinants of nutrition, but evidence distribution of biofortified vitamin A-rich orange sweet
of their nutritional eect is still scarce.
Targeted agricultural programmes have an important programmes are more successful when they incorporate
role in support of livelihoods, food security, diet strong behaviour change communications strategies
quality, and womens empowerment, and complement and a gender-equity focus. Although firm conclusions
global eorts to stimulate agricultural productivity have been hindered by a dearth of rigorous programme
and thus increase producer incomes while protecting evaluations, weaknesses in programme design and
consumers from high food prices.8 implementation also contribute to the limited evidence
on nutrition outcomes, however, is inconclusive, with of nutritional outcomes so far.
8 www.thelancet.com
Executive Summary
www.thelancet.com 9
Executive Summary
which integrate complementary inputs that address other fragmented in terms of messaging, priorities, and
constraints to optimum nutritionsuch as maternal funding.5 Much progress has been made since then,
depression, or scarcity of access to water, sanitation, and largely driven by the new evidence introduced in the
hygiene servicesand are strengthening links with health 2008 Series, which identified the first 1000 days of life
services. Rigorous impact evaluations are underway, as the window for outcomes, pinpointed a package
many of which are based on strong programme theory of highly eective interventions for reduction of
and impact pathway analysis. They are also addressing undernutrition, and proposed a group of high-burden
key weaknesses encountered in previous evaluations and countries as priorities for increased investment.
are assessing outcomes on a range of nutrition and child The launch of the SUN movement in 2010 represented
development outcomes as well as several household and a major step toward improved stewardship of the global
gender outcomes along the impact pathway. The body nutrition architecture.18,19 SUN brings together more
of evidence generated by these enhanced programmes than 100 entities across the organisational spectrum
and evaluations in the next 510 years will be of crucial of the nutrition community. Up to now, more than 30
importance to inform future investments in nutrition- countries (representing 35% of the global child stunting
sensitive programmes from many sectors. burden) have joined SUN, committing to scaling-up direct
nutrition interventions and advancing nutrition-sensitive
Building an enabling environment to deliver development. Although it is too soon to evaluate SUNs
nutrition results eect on rates of reduction of undernutrition, it is clear
The nutrition landscape has shifted fundamentally since that through SUN, many countries have made advances
2008. The 2008 Series showed that the stewardship in building multistakeholder platforms across sectors,
of the nutrition system was dysfunctional and deeply aligning nutrition-relevant programmes within a common
results framework, and mobilising national resources.
Key messages on enabling environments for nutrition Additionally, nutrition has been greatly elevated on the
global agenda. Nearly every major development agency
accelerated with deliberate action has published a policy document on undernutrition, and
donors have increased ocial development assistance
human suering should prioritise investment in scale-up of nutrition-specific
interventions, and should maximise the nutrition sensitivity of national
to basic nutrition by more than 60% between 2008 and
development processes 2011, in a very dicult fiscal climate. Nutrition is now
more prominent on the agendas of the UN, the G8 and
G20, and supporting civil society.
governance, and capacity and resources
Nowadays, the impetus for improving nutrition is
defeating. Political calculations are at the basis of eective coordination between even stronger than it was 5 years ago. The WHA targets
sectors, national and subnational levels, private sector engagement, resource for reducing stunting, wasting, low birthweight,
mobilisation, and state accountability to its citizens anaemia, and overweight, and increasing exclusive
breastfeeding in the first 6 months of life can be
squanderedconversion to results needs a dierent set of strategies and skills
achieved by 2025 with sucient support.21 Central to
fundamentally important for creating and sustaining momentum and for conversion this scaled-up support is the creation of an enabling
of that momentum into results on the ground. environment to build commitment and ensure that it is
translated into outcomes.
national and global support to a long-term process of strengthening systemic and
organisational capacities
Improvement of data, research, and accountability for
improvements in nutrition, but eorts to realise this have to date been hindered by a results
scarcity of credible evidence and trust. Both these issues need substantial attention if The availability of timely and credible nutrition data,
the positive potential is to be realised presented in accessible ways, can help governments
and other actors to be responsive to challenging
contextual analyses about how to shape and sustain enabling environments, is
circumstances, and help civil society organisations
essential as the focus shifts toward action
to hold them accountable for the eectiveness of
10 www.thelancet.com
Executive Summary
their interventions.9 Advances in health management the food industryremains high and is linked, partly,
information systems and the growing availability to the decades-long tussle related to the marketing
of newer technologies can help with the real-time of breastmilk substitutes in developing countries and
monitoring of nutrition outcomes and programme around continued marketing of sugar-sweetened
coverage and quality, and should be researched. beverages and fast foods worldwide.
Additionally, although much progress has been made This troubled history has made it more dicult for the
to work out the costs of addressing undernutrition, private sector to be a major contributor to the collective
continued work to contextualise and specify these costs creation and sustenance of momentum for reduction
for dierent countries is essential, along with stronger of malnutrition. In view of the needs and substantial
designation of donor and government spending to resources, influence, and convening power of the
improve tracking of investments and results in nutrition. private sector, it might represent a missed opportunity.
Improved data for micronutrient deficiencies and Opportunities exist for collaboration around advocacy,
other nutritional conditions are needed at national and monitoring, value chains, technical and scientific
subnational levels. This improvement should involve collaboration, and staple-food fortification that are
the development and use of improved biomarkers uncontentious and deserve further exploration. Know-
that could be used to describe nutritional conditions ledge in this area must be expanded rapidly to guide the
and increase knowledge of how they aect health and private sector toward more positive eects for nutrition.
development. Such information is needed to guide Regulatory and fiscal eorts are essential when the
intervention programmes in countries and priorities for private sector is involved in marketing of products that
support globally. are detrimental to optimum nutrition. The experience
Although substantial progress has been made gained with the International Code of Marketing
to establish the needs around nutrition, no of Breastmilk Substitutes should be applied to the
systematic process exists for bringing together the promotion of other harmful, widely-consumed food
implementation-related evidence for how to scale products that are being marketed for young children.
up the vast array of nutrition-specific and nutrition-
sensitive interventions with quality and equity (so- Mobilisation of resources
called implementation science). This evidence is High-burden countries, together with donors,
essential to ensure that future investments are directed multilaterals, and the private sector, have a responsibility
toward proven pathways to outcomes. to increase allocations to nutrition-specific and
Beyond this evidence, service providers, governments, nutrition-sensitive programmes. Meeting the estimated
donors, and the private sector need strong national $96 billion financing gap will require an increase in
monitoring and assessment platforms to hold them donor spending, alongside an equal or greater increase
accountable for the quality and eectiveness of of spending by LMICs and the establishment of nutrition
their investments in nutrition.9 Boosting nutrition budget lines in all high-burden countries.7 To achieve this
commitment and accountability can be achieved aim will be politically challenging, hence the need to build
through assessing and implementing innovative new leadership, commitment, and accountability at national
instruments and mechanisms, including computer- and international levels.9 However, the financing gap is
based monitoring systems, commitment indices, and unlikely to be closed by these sources alone. Innovation
social accountability mechanisms. is needed across all sectors to leverage private-sector and
public-sector resources and generate additional funding.
Engagement and regulation of the private sector The nutrition sector can draw on several innovative ideas
The scale, know-how, reach, financial resources, and from other sectors, including advance market contracts
existing involvement of the private sector in actions that to promote investment, market levies, and taxes in the
aect nutrition status is well known.9 Yet there are still eort. Additional resources must be directed not only to
too few independent and rigorous assessments of the interventions, but also to the creation of environments
eectiveness of involvement of the commercial sector to enable advancement of nutrition, including capacity
in nutrition. Distrust of the private sectorespecially and leadership at all levels of government.9 A political
www.thelancet.com 11
Executive Summary
economy approach to prioritisation of such investments 8 Ruel MT, Alderman H, and the Maternal and Child Nutrition Study Group.
is crucial if sustainable, supportive environments for accelerate progress in improving maternal and child nutrition. Lancet
long-term nutrition agendas are to be created.
6736(13)60843-0.
Nutrition is crucial to both individual and national 9 Gillespie S, Haddad L, Mannar V, Menon P, Nisbettt N, and the Maternal
development. The evidence in this Series furthers the building commitment and accelerating progress. Lancet
evidence base that good nutrition is a fundamental
driver of a wide range of development goals. The building momentum for impact. Lancet
post-2015 sustainable development agenda must put
addressing all forms of malnutrition at the top of its Lancet
goals. 6736(13)61100-9.
Now is our crucial window of opportunity to scale- Lancet
6736(13)61054-5.
up nutrition.22 National and international momentum
13 Taylor A, Dangour AD, Reddy KS. Only collective action will end
to address human nutrition and related food security undernutrition. Lancet
org/10.1016/S0140-6736(13)61084-3.
and health needs has never been higher. We must work 14 Nabarro D. Global child and maternal nutritionthe SUN rises. Lancet
together to seize this opportunity.
6736(13)61086-7.
References
action. Lancet
S0140-6736(13)61053-3.
regional exposures and health consequences. Lancet371: 24360.
16 Scaling Up Nutrition. Progress report from countries and their partners in
Undernutrition Study Group. What works? Interventions for maternal and
child undernutrition and survival. Lancet371: 41740.
Improving nutrition through multisectoral appraoches. Washington, DC,
International Bank for Reconstruction and Development, International
consequences for adult health and human capital. Lancet371: 340-57.
Development Association of The World Bank, 2013.
4 Bryce J, Coitinho D, Darnton-Hill I, et al, for the Maternal and Child
Announcements/Scaling_Up_Nutrition-A_Framework_for_Actionpdf
action at national level. Lancet371: 51026.
(accessed April 2, 2013).
5 Morris SS, Cogill B, Uauy R, for the Maternal and Child Undernutrition
Food
Nutr Bull31: 17886.
has it proven so dicult and what can be done to accelerate progress?
Lancet371: 60821. 20 Horton S, Steckel R. Global economic losses attributable to malnutrition
Group. Maternal and child undernutrition and overweight in low-income
2013.
and middle-income countries. Lancet
dx.doi.org/10.1016/S0140-6736(13)60937-X. 21 WHO. Proposed global targets for maternal, infant and young child
7 Bhutta ZA, Das JK, Rizvi A, et al, The Lancet Nutrition Interventions Review
2012.
22 Dube L, Pingali P, Webb P. Paths of convergence for agriculture, health, and
be done and at what cost? Lancet
wealth. Proc Natl Acad Sci USA 109: 12294301.
doi.org/10.1016/S0140-6736(13)60996-4.
Acknowledgments
Maternal and Child Nutrition Study Group:
Funding: Funding for the preparation of the Series was provided to the Johns
School of Public Health, USA), Harold Alderman (International Food Policy Research Hopkins Bloomberg School of Public Health through a grant from the Bill &
Institute, USA), Zulfiqar A Bhutta (Aga Khan University, Pakistan), Stuart Gillespie Melinda Gates Foundation. The sponsor had no role in analysis or interpretation of
(International Food Policy Research Institute, USA), Lawrence Haddad (Institute of the evidence.
Development Studies, UK), Susan Horton (University of Waterloo, Canada),
Anna Lartey (University of Ghana, Ghana), Venkatesh Mannar (The Micronutrient
Initiative, Canada), Marie Ruel (International Food Policy Research Institute, USA),
Second printing with corrections.
Cesar Victora (Universidade Federal de Pelotas, Brazil), Susan Walker (The University
of the West Indies, Jamaica), Patrick Webb (Tufts University, USA)
12 www.thelancet.com
Series
Maternal and child undernutrition is highly prevalent in low-income and middle-income countries, resulting in Lancet 2008; 371: 24360
substantial increases in mortality and overall disease burden. In this paper, we present new analyses to estimate the Published Online
eects of the risks related to measures of undernutrition, as well as to suboptimum breastfeeding practices on January 17, 2008
DOI:10.1016/S0140-
mortality and disease. We estimated that stunting, severe wasting, and intrauterine growth restriction together were
6736(07)61690-0
responsible for 22 million deaths and 21% of disability-adjusted life-years (DALYs) for children younger than 5 years.
See Comment pages 179, 180,
Deficiencies of vitamin A and zinc were estimated to be responsible for 06 million and 04 million deaths, and 181
respectively, and a combined 9% of global childhood DALYs. Iron and iodine deficiencies resulted in few child deaths, See Perspectives page 197
and combined were responsible for about 02% of global childhood DALYs. Iron deficiency as a risk factor for maternal This is the first in a Series of five
mortality added 115 000 deaths and 04% of global total DALYs. Suboptimum breastfeeding was estimated to be papers about maternal and child
responsible for 14 million child deaths and 44 million DALYs (10% of DALYs in children younger than 5 years). In an undernutrition
analysis that accounted for co-exposure of these nutrition-related factors, they were together responsible for about 35% *Members listed at end of paper
of child deaths and 11% of the total global disease burden. The high mortality and disease burden resulting from Johns Hopkins Bloomberg
School of Public Health,
these nutrition-related factors make a compelling case for the urgent implementation of interventions to reduce their Baltimore, MD, USA
occurrence or ameliorate their consequences. (Prof R E Black MD,
Prof L E Caulfield PhD); USDA,
Introduction Undernutrition encompasses stunting, wasting, and ARS Western Human Nutrition
Research Center, Davis, CA, USA
Maternal and child undernutrition remain pervasive and deficiencies of essential vitamins and minerals (col- (Prof L H Allen PhD);
damaging conditions in low-income and middle-income lectively referred to as micronutrients) as one form of the Aga Khan University, Karachi,
countries. A framework developed by UNICEF recognises condition known as malnutrition, with obesity or Pakistan (Prof Z A Bhutta, MD);
the basic and underlying causes of undernutrition, over-consumption of specific nutrients as another form. World Health Organization,
Geneva, Switzerland
including the environmental, economic, and The term hunger, which literally describes a feeling of (M de Onis MD, C Mathers PhD);
sociopolitical contextual factors, with poverty having a discomfort from not eating, has also been used to describe Harvard School of Public
central role (figure 1). Although addressing general Health, Boston, MA, USA
deprivation and inequity would result in substantial (M Ezzati PhD); and Mexico
Key messages National Institute of Public
reductions in undernutrition1 and should be a global Health, Cuernavaca, Mexico
priority, major reductions in undernutrition can also be Maternal and child undernutrition is the underlying cause (Prof J Rivera PhD)
made through programmatic health and nutrition of 35 million deaths, 35% of the disease burden in Correspondence to:
interventions. This paper is the first in a Series of five children younger than 5 years and 11% of total global Robert Black, Johns Hopkins
papers that focus on the disease burden attributable to Bloomberg School of Public
DALYs Health, Baltimore, MD, USA
undernutrition and the interventions aecting household The number of global deaths and DALYs in children less rblack@jhsph.edu
food availability and use, maternal and child care, and than 5 years old attributed to stunting, severe wasting,
control of infectious diseases. The first two papers and intrauterine growth restriction constitutes the largest
quantify the prevalence of maternal and child percentage of any risk factor in this age group
undernutrition and consider the short-term consequences Vitamin A and zinc deficiencies have by far the largest
in terms of deaths and disease burden, as measured by remaining disease burden among the micronutrients
disability-adjusted life-years (DALYs). They also discuss considered
the long-term educational and economic eects and Iodine and iron deficiencies have small disease burdens,
associations with adult chronic diseases, particularly as partly because of intervention programmes, but sustained
countries go through the demographic, epidemiological, eort is needed to further reduce their burden
and nutritional transitions.26 The third paper estimates Suboptimum breastfeeding, especially non-exclusive
the potential benefits of implementing health and breastfeeding in the first 6 months of life, results in
nutrition interventions that current evidence indicates 14 million deaths and 10% of disease burden in children
are eective and applicable in low-income and younger than 5 years
middle-income countries. The final two papers consider Maternal short stature and iron deficiency anaemia
the current state of such interventions and how they increase the risk of death of the mother at delivery,
could be implemented fully through actions at national accounting for at least 20% of maternal mortality
and global levels.
undernutrition, especially in reference to food insecurity, even though there are no MDG indicators for these
wherein people do not have physical and economic conditions. This Series also examines the consequences of
access to sucient, safe, nutritious, and culturally low body-mass index and short stature in women.
acceptable food to meet their dietary needs.7,8 This paper consists mainly of new analyses of the
Undernutrition is an important determinant of maternal prevalence of nutritional conditions, risk factors, and
and child health.912 consequent disease burden; if this was not possible or
The Millennium Development Goals (MDGs) state as necessary, previously published results are presented.
the first goal to halve between 1990 and 2015 the proportion The burden of disease attributable to maternal and child
of people who suer from hunger.7 One indicator to undernutrition is presented for three world regions
monitor progress for this target is the proportion of Africa, Asia, and Latin Americathat include primarily
children who are underweightie, low weight compared low-income and middle-income countries; only 1% of
with that expected for a well-nourished child of that age deaths in children younger than 5 years occur outside
See Online for webtable 1 and sex. This anthropometric indicator can indicate these regions. UN regions and subregions (webtable 1)
wasting (ie, low weight-for-height, indicating acute weight were used.
loss), or much more commonly, stunting (ie, low
height-for-age, indicating chronic restriction of a childs Prevalence and consequences
potential growth). Those two conditions can have dierent Maternal short stature and low body-mass index in
determinants and respond to dierent interventions.13 pregnancy and lactation
Therefore, consideration of wasting and stunting is more Maternal undernutrition, including chronic energy and
useful than consideration of underweight. This series micronutrient deficiencies, is prevalent in many regions,
primarily uses these two indicators, but also presents especially in south-central Asia, where in some countries
information on underweight because weight-for-age has more than 10% of women aged 1549 years are shorter
See Online for webtable 2 been used in previous analyses.911 Reduction of fetal than 145 cm (webtable 2). Maternal undernutritionie,
growth restriction and micronutrient deficiencies is also body-mass index of less than 185 kg/mranges
essential to achieving the MDGs and deserves high priority, from 10% to 19% in most countries. A serious problem of
maternal undernutrition is evident in most countries in
Short-term consequences: Long-term consequences:
sub-Saharan Africa, south-central and southeastern Asia,
Mortality, morbidity, disability Adult size, intellectual ability,
and in Yemen, where more than 20% of women have a
economic productivity, body-mass index of less than 185 kg/m. With a
reproductive performance, prevalence of low body-mass index around 40% in
metabolic and cardiovascular disease
women, the situation can be considered critical in India,
Maternal and child Bangladesh, and Eritrea. Maternal short stature and low
undernutrition body-mass index have independent adverse eects on
pregnancy outcomes.
Immediate
The nutritional status of a woman before and during
Inadequate dietary intake Disease
causes pregnancy is important for a healthy pregnancy
outcome.14,15 Maternal short stature is a risk factor for
caesarean delivery, largely related to cephalopelvic
disproportion. A meta-analysis of epidemiological studies
Unhealthy household
Household food
Inadequate care environment and lack
found a 60% (95% CI 5070) increased need for assisted
insecurity
of health services delivery among women in the lowest quartile of stature
(146 cm to 157 cm, depending on the region) compared
with women in the highest quartile.16 If operative delivery
to ensure a healthy birth is not available to women who
Income poverty: need it, both mother and baby are at risk.17 Even if
Underlying employment, self-employment,
causes dwelling, assets, remittances, operative delivery is accessible, aordable, and safe,
pensions, transfers etc anaesthesia and laparotomy increase the risk of maternal
morbidity.18 Low maternal body-mass index does not
seem to increase the risk of pregnancy complications and
Lack of capital: financial, human, assisted delivery.16 Rather, there seems to be a synergistic
physical, social, and natural
positive eect of short stature and higher maternal
Basic
causes body-mass index on increasing these complications.19,20
Social, economic, Low maternal body-mass index is associated with
and political context intrauterine growth restriction.11 Previous analyses
estimated the disease burden of low maternal body-mass
Figure 1: Framework of the relations between poverty, food insecurity, and other underlying and immediate index as a risk factor for perinatal conditions,11 whereas
causes to maternal and child undernutrition and its short-term and long-term consequences the estimates presented in this paper consider intrauterine
Stunting=height-for-age less than 2 SD. Severe wasting=weight-for-length or weight-for-height less than 3 SD. Underweight=weight-for-age less than 2 SD.
Table 1: Childhood stunting, severe wasting, and underweight estimates and numbers aected in 2005 based on the WHO Child Growth Standards by
UN regions and subregions
Table 3: Prevalence of intrauterine growth restriction-low birthweight and its components by UN region in 2004
Zinc deficiency
The International Zinc Nutrition Consultative Group has
proposed a method for assessment of the populations
risk of zinc deficiency based on indirect indicatorsie,
the prevalence of stunting, one of the clinical
manifestations of zinc deficiency, and the adequacy of No data
Low
absorbable zinc in the food supply at the country level.65,66 Medium
The populations risk of zinc deficiency was estimated High
for the 178 countries for which information is available
from the UN Food and Agriculture Organization. The Figure 6: National risk of zinc deficiency in children under 5 years
latest prevalence of stunting for 131 countries was
obtained from the WHO Global Database on Child burden for the age group 659 months because this age
Growth and Malnutrition.24 Data from other sources were group has a more consistent benefit in supplementation
used to classify 35 additional countries by stunting trials than infants younger than 6 months.
prevalence. Each country was classified into three
categories of risk of zinc deficiency on the basis of the Iron deficiency anaemia
combination of stunting prevalence and adequacy of zinc According to a WHO review of nationally representative
in the food supply.65,66 Countries at high risk of zinc surveys from 1993 to 2005, 42% of pregnant women and
deficiency are those with a stunting prevalence of more 47% of preschool children worldwide have anaemia.75 For
than 20% and estimated prevalence of inadequate zinc these analyses, 60% of this anaemia was assumed to be
intake of more than 25%; countries at low risk of zinc due to iron deficiency in non-malaria areas and 50% in
deficiency are those with stunting prevalence of less malaria areas.76
than 10% and inadequate zinc intake of less than 15%; The major cause of iron deficiency anaemia is low
countries at medium risk of zinc deficiency are those consumption of meat, fish, or poultry, especially in poor
with all other combinations of the categories of stunting people (panel 1).77 In young children the peak prevalence
prevalence and adequacy of zinc in the food supply. All of iron deficiency anaemia occurs around 18 months of
the trials of zinc supplementation included in the disease age and then falls as iron requirements decline and iron
burden risk estimates were done in countries classified intake is increased through complementary foods.
as having medium-high zinc deficiency. The national Women of childbearing age are at high risk for negative
prevalence of zinc deficiency is high in south Asia, most iron balance because of blood loss during menstruation
of sub-Saharan Africa, and parts of Central and South and the substantial iron demands of pregnancy.
America (figure 6). In countries classified as having zinc Previous analyses examined the relation between
deficiency, all children were considered to be at risk. anaemia in pregnancy and risk of maternal mortality.12
Zinc deficiency in children results in increased risk of An odds ratio of 08 (95% CI 070091) for maternal
diarrhoea, pneumonia, and malaria, as evidenced by mortality was found for a 10 g/L increase in mean
many randomised placebo-controlled trials done in haemoglobin in late pregnancy. These analyses also
various populations in all regions of the world.6771 The assessed the eects of anaemia on child cognition: the
relative risk of deficiency for outcomes in the at-risk combined analysis of the five available trials found 173
populations can be calculated by inverting the reduction (95% CI 104241) lower IQ points per 10 g/L decrease
in morbidity and mortality seen in the trials of zinc in haemoglobin.12 A separate meta-analysis of iron
supplementation. The pooled relative risk for morbidity supplementation trials an overall benefit of 12 IQ points,
associated with zinc deficiency is 109 (95% CI 101118) in children receiving iron, but there was no eect in
for diarrhoea, 125 (109143) for pneumonia, and 156 children younger than 27 months.78
(129189) for malaria.72,73 For mortality in infants aged
159 months the relative risk is estimated to be 127 Iodine deficiency
(096163) for diarrhoea, 118 (090154) for The prevalence of sequelae of iodine deficiency, such as
pneumonia, and 111 (094130) for malaria, based on a goitre, congenital hypothyroidism, and developmental
meta-analysis (S Sazawal, unpublished) of randomised disability, have been estimated and assigned disability
controlled trials in Bangladesh,74 Nepal,71 and Zanzibar.69 weights to calculate disease burden.79 More recently the
These values were used in the estimates of disease prevalence of iodine deficiency has been measured by an
regarded as roughly 1 lost year of so-called healthy life. standards,23 and YLD estimates for iron deficiency
Methods and data for estimating mortality and DALYs in anaemia have been updated using the most recent data
the global burden of disease project are described from the WHO Global Database on Anaemia.123 All
elsewhere.54 estimates of death and disease burden attributable to
WHO has made an incremental revision of the global undernutrition in this analysis are for the year 2004.
burden of disease for 2004, which included updating The 136 diseases and injuries in the global burden of
mortality and causes of death, and revising YLD estimates disease project follow the rules of the international
for selected causes if new data were available. Estimates classification of disease (ICD) system. The ICD considers
of child deaths by cause have been revised to take into four groups of nutritional deficiencies, and their sequelae,
account new estimates of the total numbers of neonatal, as possible direct causes of death: protein-energy
infant, and child deaths for 2004,118 updated malnutrition, iron deficiency anaemia, vitamin A
country-specific estimates of deaths in children younger deficiency, and iodine deficiency. For this reason,
than 5 years by cause based on models developed by the estimates were made of the number of deaths and
WHO Child Epidemiology Reference Group,119,120 and amount of disability that could be directly attributed to
updated UNAIDS and WHO estimates for HIV/AIDS, these conditions. As described above, in addition to these
tuberculosis, and vaccine-preventable child deaths.121,122 direct deaths, undernourished children are at an
YLD estimates for protein-energy malnutrition have been increased risk of illness and death from many infectious
updated to take into account latest WHO estimates of diseases, also included in the global burden of disease
stunting and wasting according to the new WHO categories. Indeed, substantially more children die as a
result of the synergy between infectious diseases and
undernutrition than from the direct sequelae of the
Panel 2: Specific assumptions for calculating the population nutritional conditions themselves.11 Like the ICD, global
attributable fraction for multiple nutritional risks burden of disease also includes low birthweight as a
The hazardous eects of one risk are not mediated direct cause of death; nearly all of these deaths in
through other risks developing countries are due to preterm births, not
This assumption might not hold for zinc deficiency because intrauterine growth restriction among births at term.119
some of its hazardous eects are mediated through reducing
growth. On the basis of the eects of zinc supplementation on Estimating deaths and burden of disease attributable to
linear growth,127 and the model used to estimate intervention nutritional risk factors
eects,28 it is estimated that around 46% of the excess risk for To estimate the number of deaths from various infectious
zinc deficiency is mediated through stunting. There would be diseases attributable to dierent forms of undernutrition,
no mediation via stunting for vitamin A deficiency128 or iron the methods of the comparative risk assessment project
deficiency,129 because of their lack of eects on growth. were used.124,125 In brief, for each risk factor and for each of
the diseases that are aected by it, population attributable
The proportional risks for one risk do not depend on fractions were computed using the following equation,
exposure to other risks which is for a risk factor with n levels of exposure.
This assumption seems to be valid in that there has been no Population attributable fractions estimate the proportional
eect modification found in factorial design trials of reduction in mortality that would be seen if risk factor
vitamin A, zinc supplementation, or both,130,131 and no exposure had been reduced to an alternative (counterfactual)
dierence in eect of vitamin A supplementation by baseline distribution that would result in the lowest population risk,
nutritional status by various anthropometric measures.37,132 irrespective of whether currently attainable in practice.
Likewise, no significant dierences have been seen with the This optimum exposure, referred to as the theoretical-
eects of zinc supplementation on infectious disease minimum-risk exposure distribution, results in estimates
morbidity in children with better or worse nutritional status.133 of deaths and disease burden that could be avoided if all
Exposures to these risks are uncorrelated levels of suboptimum nutrition were addressed, in a
The data from the cohort studies used to estimate odds ratios consistent and comparable way across risk factors.
show that there was a small positive correlation between n n
both stunting and wasting. *Deaths (138 739) and DALYs (14 486 400) directly attributed to protein energy malnutrition included Included in
The nutritional status measuresunderweight, wasting. Total takes into account the joint distribution of stunting and severe wasting.
stunting, and wastingwere treated as risk factors whose
Table 5: Global deaths and disease burden measured in disability-adjusted life-years (DALYs) in children
exposures and relative risks applied to all children in the under 5 years of age attributed to nutritional status measures in 2004
age group 159 months, and intrauterine growth
restriction-low birthweight only for the first month of life.
The analyses for vitamin A and zinc deficiencies and Deaths Percentage of deaths in Disease burden Percentage of DALYs in
suboptimum breastfeeding, for which either exposure or children under 5 years (1000 DALYs) children under 5 years
increased risk were age-dependent, had to be done for age Vitamin A deficiency 667 771 65 22 668 53
groups other than all children under 5 or neonatal, the Zinc deficiency 453 207 44 16 342 38
two groups that are readily available from global burden Iron deficiency 20 854 02 2 156 05
of disease databases. The division of cause-specific deaths Iodine deficiency 3 619 003 2 614 06
between 4 weeks and 59 months by age (eg, for the interval
15 months, 611 months, 1223 months) were from data Table 6: Global deaths and disability-adjusted life-years (DALYs) in children under 5 years of age
attributed to micronutrient deficiencies in 2004
for total (all-cause) deaths by month of age estimated
using demographic techniques from complete birth
histories in demographic and health surveys. (table 5). The burden due to severe wasting from infectious
Exposures and relative risks for the calculation of diseases is modest compared with the other measures,
infectious disease deaths and burden from largely because most children with non-optimum weight
nutrition-related risk factors were largely newly derived, for height have a Z score of 1 to 3 and substantially fewer
as described in this paper; however, the estimation of the have a score less than 3. The burden associated with all
direct sequelae was done using previous methods with levels of wasting was slightly less than that of stunting.
the results updated to 2004. A summary of the source of Stunting, severe wasting, and intrauterine growth
the inputs is provided in webtable 7. restriction-low birthweight together were responsible for See Online for webtable 7
Cause-specific deaths can be caused by multiple risk 21 million deaths (21% of worldwide deaths in children
factors acting simultaneously, and hence can be prevented under 5) and 910 million DALYs (21% of global DALYs
by intervention on each of the risks. For example, some for children under 5; 7% of global total DALYs). Of the
deaths from diarrhoea could be prevented by removing UN subregions, the disease burden attributed to these
exposure to stunting, zinc deficiency, vitamin A deficiency, anthropometric measures is highest in south-central
or suboptimum breastfeeding. Such deaths are attributable Asia (webtable 8), where India alone has 06 million See Online for webtable 8
to all four risks when considered individually. As a result deaths and 246 million DALYs attributed to stunting,
of multi-causality, the population attributable fractions for severe wasting, and intrauterine growth restriction-low
multiple risk factors that aect the same disease outcome birthweight. The undernutrition burden is also high in
overlap and cannot be combined by simple addition; eastern, middle and western Africa; these subregions,
rather their combined eect is generally less than the which together have an under-5 population of 111 million
crude sum of individual ones, because some deaths are (93% of India), 11 million deaths, and 423 million
attributed to multiple exposures.126 Under specific DALYs, showing higher relative and absolute eects
assumptions (panel 2), the combined (joint) population among African children.
attributable fraction that avoids double-counting can be Among the deficiencies of vitamins and minerals
estimated from those of individual risks.126 examined, the largest disease burdens were attributed to
vitamin A and zinc deficiencies (table 6). Vitamin A
Deaths and burden of disease attributable to nutritional deficiency in newborn babies, infants, and children
risk factors resulted in about 6% of under-5 deaths, 5% of under-5
Of the nutritional status measures for children, DALYs, and 17% of total DALYs. Zinc deficiency resulted
underweight was responsible for the largest disease burden in about 4% of under-5 deaths and DALYs and 1% of total
DALYs. The regional patterns for disease burden maternal deaths and DALYs due to iron deficiency
attributed to vitamin A and zinc deficiencies are similar. anaemia increases the total global disease burden
The highest burden for each is in South-central Asia, attributed to undernutrition to 11%.
See Online for webtable 9 followed by several subregions of Africa (webtable 9).
Iron deficiency resulted in a relatively small number of Discussion
under-5 deaths and DALYs (table 6). This burden is directly The attribution of more than a third of child deaths and
attributed to the sequelae of iron deficiency anaemia in more than 10% of total global disease burden to maternal
children including cognitive impairment.76 In addition, and child undernutrition demonstrates the importance of
iron deficiency anaemia in pregnancy is considered a risk these prevalent risk factors. The 21% of global deaths and
factor for maternal mortality; 115 000 deaths per year from DALYs in children younger than 5 years old attributed to
maternal causes, resulting in 34 million DALYs, have stunting, severe wasting, and intrauterine growth
been attributed to this risk factor.12 Iodine deficiency restriction-low birthweight, largely because of their
results in a very small number of deaths, which is synergistic relationship with infectious diseases, constitutes
attributed to an excess risk of death with congenital the largest percentage for any risk factor in this age
hypothyroidism, and a modest number of DALYs that group.125
primarily result from multiple sequelae from iodine The estimated total burden attributed to the com-
deficiency, including cognitive and motor impairment and bination of underweight and intrauterine growth
hearing loss.79 restriction-low birthweight is about the sameie, 22% of
In an analysis taking into account the joint distribution deaths, and disease burden in children younger than 5
of nutritional status risk factors (intrauterine growth yearsas the burden attributed to stunting, severe
restriction, stunting, severe wasting, and deficiencies of wasting, and intrauterine growth restriction-low
vitamin A or zinc) and the fact that some of the eect of birthweight. This result is lower than our previous
zinc deficiency is mediated through stunting, it is estimate that indicated that 35% of child deaths could be
estimated that there were 28 million child deaths attributed to childhood underweight and maternal low
(28% of under-5 deaths) and 1140 million DALYs (27% of body-mass index operating through intrauterine growth
under-5 DALYs) attributable to these forms of childhood restriction to aect low birthweight.11 There are several
undernutrition. This constitutes 85% of the total global reasons for this dierence. The new estimates were done
disease burden. for a dierent base year (2004 vs 2000) with modest
There was a large disease burden attributed to reductions in overall child mortality and in some major
suboptimum breastfeeding, including 14 million deaths causes of death, such as diarrhoea, malaria, and measles,
(12% of under-5 deaths) and 435 million DALYs, which that are aected by the nutritional risk factors. Further,
is 10% of global under-5 DALYs and 3% of total DALYs. there have been declining trends in the prevalence of
Most of the attributable deaths (106 million) and DALYs stunting in most regions.134 More importantly the current
(370 million) were due to non-exclusive breastfeeding in estimates used the new WHO growth standards for all
the first 6 months of life, accounting for 77% and 85%, nutritional status prevalence and mortality and morbidity
respectively, of deaths and DALYs attributed to risk estimates. These factors, along with the use of
suboptimum breastfeeding. Again the highest disease dierent data sets for much of the risk analysis, resulted
burden estimates are for south-central Asia and several in somewhat lower relative risks of mortality than in
See Online for webtable 10 sub-regions of Africa (webtable 10). This risk factor was previous analyses,10,11 especially for children with Z scores
combined with anthropometric status and deficiencies of between 1 and 2. Thus, some of the dierence in
vitamin A and zinc in an analysis allowing for co-exposure estimated disease burden is due to changes in methods,
and avoiding double counting of disease burden. If one but it also is likely to indicate progress in preventing and
assumes that the risks of suboptimum breastfeeding and managing undernutrition and reducing some important
other nutritional factors are independent, the combined causes of death.
mortality eects of all risk factors were 36 million child The methods used in these analyses indicate several
deaths (35% of under-5 deaths) and 1405 million DALYs advances from those used previously.11,12,56,73 Most of the
(35% of under-5 DALYs); this is 10% of the total global risk relations were re-examined, as were the exposure
disease burden. These estimates change only slightly if definitions and prevalences, with the most current
the burden of disease directly attributed to iron and evidence. The latest information on total and cause-specific
iodine deficiencies are added. These results are robust to child mortality was used. The new WHO growth standards
assumptions about some of the eects of suboptimum were used for all analyses. Importantly, the prevalence of
breastfeeding being mediated through other nutritional stunting is about 15% higher with the new standards than
exposures. For example, if 25% of the hazardous eects with the previous growth reference.135 The attributable
of suboptimum breastfeeding are mediated through deaths and DALYs were estimated for intrauterine growth
other nutritional risks, the total number of attributable restriction, stunting, and severe wasting, separately and
child deaths is reduced to 35 million, still about together, as well as for all wasting and underweight.
35% deaths and DALYs in this age group. Adding the Co-exposures were considered, and the resulting estimate
for the anthropometric and micronutrient factors, as well iodised salt; this and other interventions should be
as suboptimum breastfeeding, provides a total disease continued to prevent resurgence of the related disease
burden without double counting that results from simple burden. For iron deficiency the disease burden in
addition of the burdens from individual factors. children is relatively small, accounting for only 05% of
These estimations using a risk factor method show that under-5 DALYs, despite inclusion in the calculations of
449 000 child deaths can be attributed to severe wasting. life-long developmental disability in 20% of children who
Estimates by others136 have ranged up to 17 million deaths have ever had severe anaemia.76 This postulated
using a methodology based on actual mortality rates in permanent disability has not been seen in all studies and
studies done an average 18 years (maximum 25 years) its existence might need to be reconsidered in the
ago, when overall child mortality rates were much higher future.137 Nevertheless, children with iron deficiency
than today. The case fatality rates implied by these two anaemia should be treated to prevent any possible current
calculations would be 2% versus 9%. True case-fatality and future adverse eects. Iron deficiency anaemia is an
rates for all cases of severe wasting identified in a important contributor to maternal mortality, increasing
cross-sectional population survey are unknown, but are the risk of dying with blood loss during delivery.12
almost certainly lower than these of severely malnourished Suboptimum breastfeeding has large mortality
children presenting to hospitals or feeding centres in consequences worldwide, similar to those of stunting.
emergency settings. Futhermore, over time case fatality The estimate reported here is the same (14 million
rates have decreased because of better treatment. Since deaths) as a previous estimate using a dierent
the prevalence of HIV infection has increased in some categorisation of breastfeeding practices (combining
populations, some of the deaths reported to have been partial and predominant breastfeeding) and dierent
due to protein-energy malnutrition in civil registration mortality risks.94,138 The timing of initiation of
systems could actually have been due to AIDS, but the breastfeeding was not considered in these estimates.
magnitude of this possible misclassification is unknown. There is epidemiological evidence to suggest that
Finally, direct application of case fatality (versus beginning breastfeeding within the first day post partum
proportional risk) overlooks the fact that some of the would have additional benefit with regard to mortality
exposed children would have non-zero risk of death even in exclusively breastfed infants, rearming
regardless of their severe wasting status. Irrespective of recommendations to begin breastfeeding immediately
whether the true number of deaths attributable to severe after delivery.89 More than three quarters of the burden
wasting is closer to a half million or to 15 million a year, attributed to suboptimum breastfeeding is due to
therapeutic feeding interventions are important com- non-exclusive breastfeeding in the first 6 months of life
ponents of nutrition programmes in settings where when even provision of water or teas leads to an increased
severe wasting is common. risk of death. These estimates do not consider the adverse
Deficiencies of vitamin A and zinc were confirmed to eect of transmission of HIV in breast milk. Recent
be important risk factors in terms of their eects on total evidence indicates that conditions in which breast milk
child mortality. The current estimate of disease burden substitutes are recommended for babies of HIV-positive
due to vitamin A deficiency is similar to previous mothers are rarely met, and there is a net benefit of
estimates despite increases in the population coverage breastfeeding in terms of HIV-free survival even in
with vitamin A supplementation, because of inclusion of populations with a high prevalence of HIV infection.139141
risk in the 05 month age group for the first time. In Exclusive breastfeeding results in lower rates of HIV
other words, the true decline in the burden in children transmission than partial breastfeeding with rates
above 6 months of age has been numerically compensated of 1%142,143 and 4%140 being reported from studies in Africa.
by the finding that vitamin A is important in an even With around 2 million babies born to HIV-positive
younger age group. That the estimates of disease burden mothers every year,144 these transmission rates would
due to zinc deficiency are about 40% lower than result in 20 00080 000 HIV-infected infants in the first
previously reported73 is because of the lower relative risks 6 months of life. Antiretroviral treatment for pregnant or
based on analysis of additional studies.72 The lactating women will substantially reduce these resulting
WHO/UNICEF recommendation that zinc be used infections, and such treatment is increasing in Africa.144
along with oral rehydration therapy for all childhood Because HIV transmission continues with partial
diarrhoea is just beginning to be implemented in many breastfeeding after the first 6 months of life,145 early
countries, and preventive use of zinc supplementation breastfeeding cessation has been considered an option.
or fortification is uncommon. The high residual burden Evidence is still incomplete; however, a recent trial found
of those micronutrient deficiencies indicates that no dierence in HIV-free survival at 24 months of age in
interventions need to be expanded, especially in children who were randomly assigned to abrupt weaning
south-central Asia and Africa. at 4 months of age or continued breastfeeding.146
Iodine and iron deficiencies currently result in only Inadequate complementary feeding contributes to
small disease burdens. For iodine deficiency, there has stunting as do infectious diseases, especially repeated
been substantial success with promotion of the use of episodes without adequate case management. These
exposures, which are arguably the key determinants of nutritional risks were addressed, a substantial number of
stunting, were not modelled as risk factors in this paper. child deaths would still require interventions related to
However, the eects on stunting and related disease breastfeeding practices.
burden of interventions to improve complementary Additional research will be important for improving
feeding and to reduce diarrhoea incidence are calculated understanding of the distribution of nutritional de-
in a subsequent paper in the Series.28 Results presented ficiencies in populations and their risk relations. Some
here on the eects of diarrhoea on the risk of stunting are specific research topics are listed in panel 3. Despite the
incorporated in these calculations of intervention eects, remaining research needs, these results, which show
but other infectious diseases and perhaps asymptomatic the high prevalence of various forms of maternal and
infections are also likely to contribute to stunting. child undernutrition and their health consequences,
The nutrition-related risk factors individually result in present a compelling justification for implementation
large disease burdens, but often coexist both in individuals of eective nutrition-related interventions. The potential
and in populations. Stunting and severe wasting in eects of these interventions are considered in the third
children 159 months of age are largely uncorrelated, and paper in this Series, with particular attention to their
the estimation of the joint eects of these two risk factors contribution to achievement of MDG 1 with reduced
resulted in a slightly smaller total disease burden than undernutrition and also MDGs 4 and 5 regarding child
simply adding their eects because even in the absence of and maternal mortality.
statistical correlation, some children would be exposed to Contributors
both conditions. Intrauterine growth restriction-low RB conceptualised and coordinated the analyses and preparation of the
birthweight was considered for these estimations to aect paper. Primary responsibility for specific topics were as follows: LHA,
maternal undernutrition and consequences of selected micronutrient
only infants younger than 1 month, but in reality could deficiencies; ZB, complementary feeding; LEC, analysis of intrauterine
aect later ages. Vitamin A, zinc, iron, and iodine growth restriction-low birthweight and childhood undernutrition risks
deficiencies and the anthropometric measures as risk for mortality; MdO, stunting, wasting and low birthweight prevalence by
factors are largely uncorrelated, but there is still overlap in country and region; ME, analysis of risk factor attributable mortality and
disease burden; CM, update of country, regional, and global
their risk because there is multicausality, such as diarrhoea disease-specific mortality and burden of disease for 2004; and JR,
for vitamin A, zinc, and stunting. In addition to simple exposure to zinc deficiency by country. All authors contributed to the
multicausality, about half of the eect of zinc deficiency is final paper.
mediated through stunting; the rest is a more direct eect Other contributors
on morbidity and mortality, probably as a result of reduced Analyses of national surveys for prevalence of underweight, stunting and
immune function.147 The risk related to suboptimum wasting based on new WHO Child Growth Standards: Monika Blssner,
Yang Hong, Allen Shoemaker (WHO, Switzerland). Provision of
breastfeeding might in part be due to micronutrient estimates of prevalence of breastfeeding status: Ana Betrn, Jeremy Lauer,
deficiencies resulting from inadequate dietary intake, but (WHO, Switzerland). Review of relative risks of suboptimum
is also due to avoidance of infection. Calculating the breastfeeding: Ai Koyanagi, Sunil Sazawal (Johns Hopkins Bloomberg
combined eects of suboptimum breastfeeding and the School of Public Health, USA). Provision of panel 1 (equity gaps in regard
to stunting and dietary diversity): Satoru Shimokawa (Cornell University,
other nutrition-related risk factors found that, even after USA), Cesar Victora (University of Pelotas, Brazil). Provision of data for
accounting for overlapping eects from multicausality of analysis of mortality risks related to anthropometric indicators:
diarrhoea and pneumonia, about half of the disease Keith West, Luke Mullany, Parul Christian, Larry Moulton, James Tielsch
(Johns Hopkins Bloomberg School of Public Health, USA), Linda Adair
burden attributable to suboptimum breastfeeding was
(University of North Carolina School of Public Health, USA), Kre Mlbak
added to the other risk factors. Therefore, even if all other (Statens Serum Institut, Denmark), Zulfiqar A Bhutta (Aga Khan
University, Pakistan), A M Menezes (University of Pelotas, Brazil),
Michel Garenne (Institute Pasteur, France), Shams El Arifeen
Panel 3: Research needs (International Centre for Diarrhoeal Diseases Research, Bangladesh),
M Andersen (University of Copenhagen, Denmark). Provision of data for
Development of methods to assess nutritional status and analysis of socioeconomic confounding: Saul S Morris (London School of
its determinants Hygiene and Tropical Medicine, United Kingdom), Keith West (Johns
Hopkins Bloomberg School of Public Health, USA). Review of risk related
Prevalence of deficiencies of vitamin A, zinc, iron, and to maternal body-mass index: Anju Aggarwal (Johns Hopkins Bloomberg
iodine in subnational populations School of Public Health, USA). Provision of data on vitamin A
Consequences of nutritional deficiencies for mortality supplementation coverage: Nita Dalmiya (UNICEF, USA). Assistance in
from HIV/AIDS, malaria, and other important infectious classifying countries as vitamin A deficient: Keith West (Johns Hopkins
Bloomberg School of Public Health, USA). Analysis of the contribution of
diseases diarrhoea to stunting: William Checkley, Gillian Buckley (Johns Hopkins
Consequences of nutritional deficiencies for immune Bloomberg School of Public Health, USA). Analyses of mortality and
competence, brain development, cognitive ability, and burden of disease attributable to nutrition risk factors: Rodrigo Dias
(Harvard University Initiative for Global Health, USA). Analyses of age
other possible eects
patterns of deaths in children younger than 5 years: Kenneth Hill,
Overlap of micronutrients and their joint eects on Thomas Laakso (Harvard University Initiative for Global Health, USA).
mortality and morbidity Analyses of contribution of undernutrition to mortality by cause at birth
Development of international fetal and newborn growth and during childhood, confounding by socioeconomic factors and of
correlation between stunting and wasting: Ping Chen, Carmen Carrillo
standards
(Johns Hopkins Bloomberg School of Public Health, USA).
Maternal and Child Undernutrition Study Group 12 Stoltzfus RJ, Mullany L, Black RE. Iron deficiency anaemia. In:
Series steering committeeRobert E Black (Johns Hopkins Bloomberg Ezzati M, Lopez AD, Rodgers A, Murray CLJ, eds. Comparative
School of Public Health, USA), Zulfiqar A Bhutta (Aga Khan University, quantification of health risks: global and regional burden of disease
Pakistan), Jennifer Bryce (Johns Hopkins Bloomberg School of Public attributable to selected major risk factors. Geneva: World Health
Health, USA), Saul S Morris (London School of Hygiene and Tropical Organization, 2004: 163209.
Medicine, UK), Cesar G Victora (Federal University of Pelotas, Brazil). 13 Waterlow JC. Classification and definition of protein-calorie
Other membersLinda Adair (University of North Carolina, USA), malnutrition. BMJ 1972; 3: 56669.
Tahmeed Ahmad (ICDDR,B, Bangladesh), Lindsay H Allen (USDA ARS 14 Kramer MS. Determinants of low birth weight: methodological
Western Human Nutrition Research Center, USA), Laura E Caulfield assessment and meta-analysis. Bull World Health Organ 1987;
65: 663737.
(Johns Hopkins Bloomberg School of Public Health), Bruce Cogill
15 Kramer MS, Victora C. Low birth weight and perinatal mortality. In:
(UNICEF, USA), Denise Coitinho (WHO, Switzerland), Simon Cousens
Semba RD, Bloem MW, eds. Nutrition and health in developing
(London School of Hygiene and Tropical Medicine, UK), Ian Darnton-Hill
countries. Humana Press, 2001.
(UNICEF, USA), Mercedes de Onis (WHO, Switzerland); Kathryn Dewey
16 WHO. Maternal anthropometry and pregnancy outcomes: A WHO
(University of California, Davis, USA), Majid Ezzati (Harvard School of
Collaborative Study World Health Organ Suppl 1995; 73: 3237.
Public Health, USA), Caroline Fall (University of Southhampton, UK),
17 Ronsmans C, Holtz S, Stanton C. Socioeconomic dierentials in
Elsa Giugliani (Federal University of Rio Grande de Sul, Brazil),
caesarean rates in developing countries: a retrospective analysis.
Batool A Haider (Aga Khan University, Pakistan), Pedro Hallal (Federal Lancet 2006; 368: 151623.
University of Pelotas, Brazil), Betty Kirkwood (London School of Hygiene 18 Villar J, Valladares E, Wojdyla D, et al. Caesarean delivery rates and
and Tropical Medicine, UK), Reynaldo Martorell (Emory University, pregnancy outcomes: the 2005 WHO global survey on maternal and
Rollins School of Public Health, USA), Colin Mathers (WHO, perinatal health in Latin America. Lancet 2006; 367: 181929.
Switzerland), David Pelletier (Cornell University, USA), 19 Cnattinguis R, Cnattinguis S, Notzon F. Obstacles to reducing
Per Pinstrup-Andersen (Cornell University, USA), Linda Richter (Human cesarean rates in a low-cesarean setting: the eect of maternal age,
Sciences Research Council, South Africa), Juan A Rivera (Mexico National height and weight. Obstet Gynecol 1998; 92: 50106.
Institute of Public Health), Harshpal Singh Sachdev (Sitaram Bhartia 20 Dempsey J, Ashiny Z, Qiu C, et al. Maternal pre-pregnancy
Institute of Science and Research, India), Meera Shekar (World Bank, overweight status and obesity as risk factors for cesarean delivery.
USA), Ricardo Uauy (Institute of Nutrition, Chile). Matern Fetal Neonatal Med 2005; 17: 17985.
Conflict of interest statement 21 Allen LH. Maternal micronutrient malnutrition: eects on breast
milk and infant nutrition, and priorities for intervention. SCN News
We declare that we have no conflict of interest. As corresponding author,
1994; 11: 2124.
R Black states that he had full access to all data and final responsibility to
22 WHO. WHO Child Growth Standards: Length/height-for-age,
submit for publication.
weight-for-age, weight-for-length, weight-for-height and body mass
Acknowledgments index-for-age: methods and development. Geneva: World Health
Funding for the preparation of the Series was provided by the Bill & Organization, 2006.
Melinda Gates Foundation. Meetings were hosted by the UNICEF 23 de Onis M, Garza C, Onyango AW, Martorell R. WHO Child Growth
Innocenti Research Centre and the Rockefeller Foundation Bellagio Standards. Acta Paediatr Suppl 2006; 450: 1101.
Conference Center. The sponsors had no role in the analysis and 24 de Onis M, Blossner M. The World Health Organization Global
interpretation of the evidence nor in writing the report and the decision Database on Child Growth and Malnutrition: methodology and
to submit for publication. We thank Barbara Ewing and Mary Rybczynski applications. Int J Epidemiol 2003; 32: 51826.
for administrative assistance with the Series. 25 de Onis M, Blossner M, Borghi E, Morris R, Frongillo EA.
Methodology for estimating regional and global trends of child
References malnutrition. Int J Epidemiol 2004; 33: 126070.
1 Haddah H, Alderman H, Appleton S, Song L, Yisehac Y. Reducing 26 de Onis M, Blossner M, Borghi E, Frongillo EA, Morris R. Estimates
child malnutrition: How far does income growth take us? of global prevalence of childhood underweight in 1990 and 2015.
World Bank Economic Review 2003; 17: 10731. JAMA 2004; 291: 260006.
2 Ezzati M, Vander Hoorn S, Lawes CMM, et al. Rethinking the 27 UN Department of Economics and Social Aairs, Population
diseases of auence paradigm: global patterns of nutritional risks Division. World Population Prospects, the 2004 revision. New York:
in relation to economic development. PLoS Medicine 2005; 2: e133. United Nations, 2005.
3 Prentice AM, Moore SE. Early programming of adult diseases in 28 Bhutta ZA, Ahmad T, Black RE, et al, for the Maternal and Child
resource poor countries. Arch Dis Child 2005; 90: 42932. Undernutrition Study Group. What works? Interventions for
4 Caballero B, Popkin BE, eds. The nutrition transition: diet and maternal and child undernutrition and survival. Lancet 2008;
disease in the developing world. London: Academic Press, 2002. published online Jan 17. DOI: 10.1016/S0140-6736(07)61693-6.
5 Salomon JA, Murray CJL. The epidemiologic transition 29 Gwatkin DR, Rutstein S, Johnson K, Suliman EA, Wagsta A.
re-examined: compositional models for causes of death by age and Initial country-level information about socio-economic dierences
sex. Population and Development Review 2002; 28: 20528. in health, nutrition, and population 2nd edn. Washington DC:
6 Grantham-McGregor S, Cheung YB, Cueto S, et al. Developmental World Bank, 2003.
potential in the first five years for children in developing countries. 30 Measure DHS. Demographic and Health Surveys. About DHS
Lancet 2006; 369: 6070. homepage. http://www.measuredhs.com/aboutdhs (accessed
7 UN Millennium Project 2005. Halving hunger: it can be done. Nov 21, 2007).
London and Sterling, VA: Task Force on Hunger, 2005. 31 Victora CG. The association between wasting and stunting: an
8 The World Bank. Repositioning nutrition as central to development. international perspective. J Nutr 1992; 122: 110510.
A strategy for large-scale action. Washington DC: The World Bank, 32 Adair LS, Popkin BM, VanDerslice J, et al. Growth dynamics in the
2006. first two years of life: a prospective study in the Philippines.
9 Caulfield LE, de Onis M, Blossner M, Black RE. Undernutrition as Eur J Clin Nutr 1993; 47: 4251.
an underlying cause of child deaths associated with diarrhea, 33 WHO/CHD Immunisation-Linked Vitamin A Supplementation
pneumonia, malaria, and measles. Am J Clin Nutr 2004; 80: 19398. Study Group. Randomised trial to assess benefits and safety of
10 Pelletier DL, Frongillo EA Jr, Habicht JP. Epidemiologic evidence vitamin A supplementation linked to immunisation in early
for a potentiating eect of malnutrition on child mortality. infancy. Lancet 1998; 352: 125763.
Am J Public Health 1993; 83: 113033. 34 Arifeen S, Black RE, Antelman G, et al. Exclusive breastfeeding
11 Fishman SM, Caulfield L, de Onis M, et al. Childhood and maternal reduces acute respiratory infection and diarrhea deaths among
underweight. In: Ezzati M, Lopez AD, Rodgers A, Murray CLJ, eds. infants in Dhaka slums. Pediatrics 2001; 108: E67.
Comparative quantification of health risks: global and regional 35 Garenne M, Maire B, Fontaine O, Dieng K, Briend A. Risks of dying
burden of disease attributable to selected major risk factors. associated with dierent nutritional status in pre-school aged
Geneva: World Health Organization, 2004: 39161. children. Dakar: ORSTOM, 1987 (reprinted by CEPED, Paris, 2000).
Maternal and child malnutrition in low-income and middle-income countries encompasses both undernutrition and Lancet 2013; 382: 42751
a growing problem with overweight and obesity. Low body-mass index, indicative of maternal undernutrition, has Published Online
declined somewhat in the past two decades but continues to be prevalent in Asia and Africa. Prevalence of maternal June 6, 2013
http://dx.doi.org/10.1016/
overweight has had a steady increase since 1980 and exceeds that of underweight in all regions. Prevalence of stunting
S0140-6736(13)60937-X
of linear growth of children younger than 5 years has decreased during the past two decades, but is higher in south
This publication has been
Asia and sub-Saharan Africa than elsewhere and globally aected at least 165 million children in 2011; wasting corrected. The corrected version
aected at least 52 million children. Deficiencies of vitamin A and zinc result in deaths; deficiencies of iodine and first appeared at thelancet.com
iron, together with stunting, can contribute to children not reaching their developmental potential. Maternal on June 20, 2013
undernutrition contributes to fetal growth restriction, which increases the risk of neonatal deaths and, for survivors, See Comment page 371
of stunting by 2 years of age. Suboptimum breastfeeding results in an increased risk for mortality in the first 2 years This is the first in a Series of
of life. We estimate that undernutrition in the aggregateincluding fetal growth restriction, stunting, wasting, and four papers about maternal and
child nutrition
deficiencies of vitamin A and zinc along with suboptimum breastfeedingis a cause of 31 million child deaths
*Members listed alphabetically
annually or 45% of all child deaths in 2011. Maternal overweight and obesity result in increased maternal morbidity
Members listed at end of paper
and infant mortality. Childhood overweight is becoming an increasingly important contributor to adult obesity,
diabetes, and non-communicable diseases. The high present and future disease burden caused by malnutrition in Johns Hopkins University,
Bloomberg School of Public
women of reproductive age, pregnancy, and children in the first 2 years of life should lead to interventions focused on Health, Baltimore, MD, USA
these groups. (Prof R E Black MD,
Prof P Christian DrPH,
Introduction continuing stunting of growth and deficiencies of essential Prof J Katz ScD); Universidade
Federal de Pelotas, Pelotas,
Maternal and child malnutrition, encompassing both nutrients along with obesity in national populations and Rio Grande do Sol, Brazil
undernutrition and overweight, are global problems with within families. We also want to assess national progress (Prof C G Victora MD); The
important consequences for survival, incidence of acute in nutrition programmes and international actions con- University of the West Indies,
Tropical Medicine Research
and chronic diseases, healthy development, and the sistent with our previous recommendations.
economic productivity of individuals and societies.
Maternal and child undernutrition, including stunting,
wasting, and deficiencies of essential vitamins and Key messages
minerals, was the subject of a Series15 in The Lancet in Iron and calcium deficiencies contribute substantially to maternal deaths
2008, which quantified their prevalence, short-term and Maternal iron deficiency is associated with babies with low weight (<2500 g) at birth
long-term consequences, and potential for reduction Maternal and child undernutrition, and unstimulating household environments,
through high and equitable coverage of proven nutrition contribute to deficits in childrens development and health and productivity in adulthood
interventions. The Series identified the need to focus on Maternal overweight and obesity are associated with maternal morbidity, preterm
the crucial period of pregnancy and the first 2 years of birth, and increased infant mortality
lifethe 1000 days from conception to a childs second Fetal growth restriction is associated with maternal short stature and underweight
birthday during which good nutrition and healthy growth and causes 12% of child deaths
have lasting benefits throughout life. The 2008 Series also Stunting prevalence is slowly decreasing globally, but aected at least 165 million
called for greater national priority for nutrition pro- children younger than 5 years in 2011; wasting aected at least 52 million children
grammes, more integration with health programmes, Suboptimum breastfeeding results in more than 800 000 child deaths annually
enhanced intersectoral approaches, and more focus Undernutrition, including fetal growth restriction, suboptimum breastfeeding,
and coordination in the global nutrition system of inter- stunting, wasting, and deficiencies of vitamin A and zinc, cause 45% of child deaths,
national agencies, donors, academia, civil society, and the resulting in 31 million deaths annually
private sector. 5 years after that series, we intend not Prevalence of overweight and obesity is increasing in children younger than 5 years
only to reassess the problems of maternal and child globally and is an important contributor to diabetes and other chronic diseases
undernutrition, but also to examine the growing problems in adulthood
of overweight and obesity for women and children and Undernutrition during pregnancy, aecting fetal growth, and the first 2 years of life is
their consequences in low-income and middle-income a major determinant of both stunting of linear growth and subsequent obesity and
countries (LMICs). Many of these countries are said to non-communicable diseases in adulthood
suer the so-called double burden of malnutrition, with
Institute, Mona Campus, The present Series is guided by a framework (figure 1) undernutrition in childhood and obesity and related
Kingston, Jamaica that shows the means to optimum fetal and child growth diseases in adulthood. Thus, we organise this paper to
(Prof S P Walker PhD);
The Aga Khan University and
and development, rather than the determinants of consider prevalence and consequences of nutritional
Medical Center, Department of undernutrition as shown in the conceptual model conditions during the life course from adolescence to
Pediatrics, Karachi, Pakistan developed by UNICEF and used in the 2008 Series.1 This pregnancy to childhood and discuss the implications for
(Prof Z A Bwhutta PhD); World new framework shows the dietary, behavioural, and adult health. In the second paper, we describe evidence
Health Organization,
Department of Nutrition for
health determinants of optimum nutrition, growth, and supporting nutrition-specific interventions and the health
Health and Development, development and how they are aected by underlying eects and costs of increasing their population coverage.
Geneva, Switzerland food security, caregiving resources, and environmental In the third paper we examine nutrition-sensitive inter-
(M de Onis MD); Imperial College conditions, which are in turn shaped by economic and ventions and approaches and their potential to improve
of London, St Marys Campus,
School of Public Health,
social conditions, national and global contexts, resources, nutrition. In the fourth paper we examine the features of
MRC-HPA Centre for and governance. This Series examines how these an enabling environment that are needed to provide
Environment and Health, determinants can be changed to enhance growth and support for nutrition programmes and how they can be
Department of Epidemiology development. These changes include nutrition-specific favourably changed. Finally, in a Comment6 we will
and Biostatistics, London, UK
(Prof M Ezzati PhD); Institute of
interventions that address the immediate causes of examine the desired national and global response to
Child Health, University College suboptimum growth and development. The framework address nutritional and developmental needs of women
London, London, UK (Prof shows the potential eects of nutrition-sensitive inter- and children in LMICs.
S Grantham-McGregor FRCP);
ventions that address the underlying determinants of
The University of the West
Indies, Mona, Jamaica malnutrition and incorporate specific nutrition goals Prevalence and consequences of nutritional
(Prof S Grantham-McGregor); and actions. It also shows the ways that an enabling conditions
Emory University, Atlanta, GA, environment can be built to support interventions and Adolescent nutrition
USA (Prof R Martorell PhD); and
programmes to enhance growth and development and Adolescent nutrition is important to the health of girls
London School of Hygiene and
Tropical Medicine, London, UK their health consequences. In the first paper we assess and is relevant to maternal nutrition. There are 12 billion
(Prof R Uauy PhD) the prevalence of nutritional conditions and their health adolescents (aged 1019 years) in the world, 90% of whom
Correspondence to: and development consequences. We deem a life-course live in LMICs. Adolescents make up 12% of the population
Prof Robert Black, Johns Hopkins perspective to be essential to conceptualise the nutritional in industrialised countries, compared with 19% in LMICs
University, Bloomberg School of eects and benefits of interventions. The nutritional (appendix p 2 shows values for ten countries studied in
Public Health, Baltimore,
MD 21205, USA
status of women at the time of conception and during depth).7 Adolescence is a period of rapid growth and
rblack@jhsph.edu pregnancy is important for fetal growth and development, maturation from childhood to adulthood. Indeed, some
and these factors, along with nutritional status in the first researchers have argued that adolescence is a period
See Online for appendix
2 years of life, are important determinants of both with some potential for height catch-up in children with
Nutrition specific Optimum fetal and child nutrition and development Nutrition sensitive
interventions programmes and approaches
and programmes Agriculture and food security
Adolescent health and Breastfeeding, nutrient- Feeding and caregiving Low burden of Social safety nets
preconception nutrition rich foods, and eating practices, parenting, infectious diseases Early child development
Maternal dietary routine stimulation Maternal mental health
supplementation Womens empowerment
Micronutrient Child protection
supplementation or Food security, including Feeding and caregiving Access to and use of Classroom education
fortification availability, economic resources (maternal, health services, a safe and Water and sanitation
Breastfeeding and access, and use of food household, and hygienic environment Health and family planning services
complementary feeding community levels)
Dietary supplementation
for children
Dietary diversification Building an enabling environment
Knowledge and evidence Rigorous evaluations
Feeding behaviours and
Politics and governance Advocacy strategies
stimulation
Leadership, capacity, and financial resources Horizontal and vertical coordination
Treatment of severe acute
Social, economic, political, and environmental context (national and global) Accountability, incentives regulation,
malnutrition
Disease prevention and legislation
management Leadership programmes
Nutrition interventions in Capacity investments
emergencies Domestic resource mobilisation
Figure 1: Framework for actions to achieve optimum fetal and child nutrition and development
stunting from early childhood.8 Adolescent fertility is In India, for example, 558% of adolescents aged
three times higher in LMICs than in high-income coun- 1519 years and 567% of women aged 2024 years were
tries. Pregnancies in adolescents have a higher risk of anaemic;14 corresponding values for Guatemala were
complications and mortality in mothers9 and children10 210% and 204 %, respectively.15
and poorer birth outcomes than pregnancies in older
women.10,11 Furthermore, pregnancy in adolescence will Maternal nutrition
slow and stunt a girls growth.12,13 In some countries, as Prevalence of low BMI (<185 kg/m) in adult women
many as half of adolescents are stunted (height-for-age has decreased in Africa and Asia since 1980, but remains
Z score [HAZ] <2), increasing the risk of poor perinatal higher than 10% in these two large developing regions
outcomes in their ospring (appendix p 2). We used (figure 2). During the same period, prevalence of
age-specific, low body-mass index (BMI) cutos (BMI overweight (BMI 25 kg/m) and obesity (BMI 30 kg/m)
Z score [BMIZ] <2) from the WHO reference for has been rising in all regions, together reaching more
children aged 519 years to examine ten selected than 70% in the Americas and the Caribbean and more
countries; in these locations as many as 11% (India) of than 40% in Africa by 2008.16,17
adolescent girls are thin. In these countries, prevalence of Few studies have examined the risk of maternal
high BMI for age, defined as BMIZ >2 (obesity), is as mortality in relation to maternal anthropometry with a
high as 5% (Brazil; appendix p 2). Adolescents have as prospective design. In one study in Nepal18 of about
high a prevalence of anaemia as women aged 2024 years. 22 000 women, mid upper arm circumference during
300
200
100
Asia Europe
600
500
400
Prevalence (%)
300
200
100
Oceania Global
600
500
400
Prevalence (%)
300
200
100
0
BMI 185 BMI 25 BMI 30 BMI 185 BMI 25 BMI 30
Figure 2: Trends in thinness (BMI <185 kg/m), overweight (BMI 25 kg/m), and obesity (BMI 30 kg/m), using population weighted average prevalences
for women aged 2049 years UN regions and globally, 19802008
Error bars are 95% CIs. BMI=body-mass index.
pregnancy was inversely associated with all-cause mater- Maternal vitamin deficiencies
nal mortality up to 42 days post-partum after adjusting Anaemia and iron
for numerous factors. An inverse association exists Anaemia (haemoglobin <110 g/L), which might be
between maternal height and the risk of dystocia (dicult attributable to low consumption or absorption in the diet
labour), as measured by cephalopelvic disproportion or or to blood loss, such as from intestinal worms, is highly
assisted or caesarean deliveries.1923 prevalent during pregnancy. This Series focuses on
Figure 2 also shows trends for overweight and obesity anaemia amenable to correction with iron supplemen-
in women aged 2049 years in dierent UN regions. tation.42 To establish the importance of iron deficiency as
Oceania, Europe, and the Americas had the highest a cause of maternal anaemia we used the results of trials
proportion of overweight and obese women; however of iron supplementation to work out the shift in the
northern and southern Africa, and central and west Asia population haemoglobin distribution. A meta-analysis of
also had high prevalences (appendix p 3). the eects of iron supplementation trials showed that,
Maternal obesity leads to several adverse maternal and among pregnant women with anaemia at baseline, iron
fetal complications during pregnancy, delivery, and supplementation led to a 102 g/L increase in haemo-
post-partum.24 Obese pregnant women (pre-pregnant globin.43 The corresponding figure for children was
BMI 30 kg/m) are four times more likely to develop 80 g/L.44 We applied these shifts to the present distri-
gestational diabetes mellitus and two times more likely to butions of haemoglobin estimated by Stevens and
develop pre-eclampsia compared with women with a colleagues42 and calculated the proportion of pregnant
BMI 185249 kg/m2).2528 During labour and delivery, women with anaemia whose blood haemoglobin would
maternal obesity is associated with maternal death, increase to at least 110 g/L. We likewise calculated the
haemorrhage, caesarean delivery, or infection;2931 and a proportion of severe anaemia that would increase to at
higher risk of neonatal and infant death,32 birth trauma, least 70 g/L. These results constitute the prevalence of
and macrosomic infants.3337 In the post-partum period, iron amenable or iron deficiency anaemia (IDA) or severe
obese women are more likely to delay or fail to lactate and IDA, defined as the proportion of anaemia or severe
to have more weight retention than women of normal anaemia that would be reduced if only iron was provided,
weight.38 Obese women with a history of gestational holding other determinants of anaemia unchanged. With
diabetes have an increased risk of subsequent type this approach in 2011, Africa had the highest proportion
two diabetes, metabolic syndrome, and cardiovascular of IDA for pregnant women followed closely by Asia
disease.39 The early intrauterine environment has a role in (table 1,4547 appendix p 4). Likewise, Africa had the highest
programming phenotype, aecting health in later life. prevalence of severe IDA, but in all regions prevalence
Maternal overweight and obesity at the time of pregnancy was less than 1%.
increases the risk for childhood obesity that continues Our previous analyses48 showed that anaemia in preg-
into adolescence and early adulthood, potentiating the nancy increased the risk of maternal mortality. An
transgenerational transmission of obesity.40,41 updated analysis49 with ten studies (four more than the
Table 1: Prevalence of vitamin A deficiency (19952005), iodine deficiency (2013), inadequate zinc intake (2005), and iron deficiency anaemia (2011)
previous analysis) showed that the odds ratio (OR) for suggesting that benefits can be detected in later childhood
maternal deaths was 071 (95% CI 060085) for a when more complex tasks can be measured.62
10 g/L greater mean haemoglobin in late pregnancy.
Only two of the ten studies adjusted for socioeconomic Vitamin A
confounding variables; one showed no attenuation and Maternal vitamin A deficiency can cause visual impair-
the other a 20% attenuation of the eect. ment and possibly other health consequences. Deficiency
There is strong biological plausibility for a causal link is assessed in pregnant women as either a history of night
between maternal IDA and adverse birth outcomes includ- blindness or serum or plasma retinol concentrations of
ing low birthweight and increased perinatal mortality.5052 A less than 070 mol/L (subclinical vitamin A deficiency).
meta-analysis53 that included 11 trials indentified a WHO provides prevalence estimates for 19952005 from
significant 20% reduction in the risk of low birthweight 64 countries, which we used for estimates for the UN
associated with antenatal supplementation with iron alone world regions (table 1).45 Globally, the prevalence of night
or combined with folic acid (relative risk [RR] 080, 95% CI blindness in pregnant women is estimated to be 78%
071090). A previous Cochrane review54 had much the (95% CI 7087), aecting 97 million women. An
same findings. Dibley and colleagues55 pooled data from estimated 153% (74232) of pregnant women globally
demographic and health surveys from Indonesia for 1994, (191 million women) have deficient serum retinol con-
1997, 200203, and 2007 and showed that risk of death of centrations. The degree to which night blindness and low
children younger than 5 years was reduced by 34% when serum retinol overlap is not accounted for in this esti-
the mother consumed any iron-folic acid supplements mation, but night blindness is known to be associated
(hazard ratio [HR] 066; 95% CI 053081). Dibley and with a four-times higher odds of low serum retinol
colleagues further showed that the protective eect was (OR 402, 95% CI 2274).63 Night blindness is reduced
greatest for deaths on the first day of life (040; 021077), by vitamin A supplementation in pregnancy.63,64 Maternal
but the protective eect was also shown for neonatal night blindness has been associated with increased low
deaths (069; 95% CI 049097) and post-neonatal deaths birthweight64 and infant mortality,65 yet trials of vitamin A
(074; 056099). A randomised controlled trial56 from in pregnancy have not showed significant eects on
China showed a significant 54% (RR 046, 95% CI these outcomes.6669
021098) reduction in neonatal mortality with antenatal
iron and folic acid supplementation compared with folic Zinc
acid alone as control. In Nepal, mortality from birth to Zinc is a key micronutrient with a ubiquitous role in
7 years was reduced by 31% (HR 069, 95% CI 049099) biological functions, including protein synthesis, cellular
in the ospring of mothers who had received iron and folic division, and nucleic acid metabolism. Estimates revised
acid during pregnancy compared with controls who in 2012 suggest that 17% of the worlds population is at
received vitamin A only.57 risk of zinc deficiency, on the basis of an analysis of
Randomised controlled trials from high-income national diets.47 Excess losses of zinc during diarrhoea
countries have shown benefits of iron supplementation also contribute to zinc deficiency. The eect of subclinical
for improved maternal mental health and reduced zinc deficiency (defined as low plasma zinc concentration
fatigue.58 Evidence from LMICs for the eect of mater- without obvious signs of zinc deficiency) in women of
nal IDA on mothers mental health and mother-child reproductive age and during pregnancy on health and
interactions is scarce. In a small South African trial,59 development outcomes is poorly understood, although
iron supplementation of women with IDA from zinc deficiency has been suggested as a risk factor with
10 weeks post-partum to 9 months led to lower maternal adverse long-term eects on growth, immunity, and
depression and perceived stress at 9 months compared metabolic status of surviving ospring.70 Zinc deficiency
with placebo. At 9 months, iron supplemented mothers due to a rare genetic abnormalityacrodermatitis entero-
had better maternal-child interactions.60 By contrast, in pathicain pregnancy results in a high risk of preterm
Bangladesh higher levels of maternal iron supple- and prolonged labour, post-partum haemorrhage, and
mentation decreased the quality of maternal-child fetal growth restriction.70,71 A review of supplementation
interaction at age 34 months and had no eect on trials with zinc in pregnancy showed a significant 14%
maternal distress (anxiety and depression).61 reduction in preterm births in women in low-income
There is some evidence for whether maternal IDA settings, but no significant eect on low birthweight.72
aects child development. Infants of mothers identified
as having IDA at 68 weeks post-partum had lower Iodine
developmental levels at 10 weeks and 9 months compared Maternal iodine deficiency is of concern in regard to
with infants of control mothers without IDA.60 In Nepal, adverse eects on fetal development, yet few countries
children whose mothers received iron and folate supple- have nationally representative data from large-scale
mentation during pregnancy had better general intelli- surveys of urinary iodine concentration in pregnant
gence and cognitive functioning at age 79 years women. Because of the correlation between urinary
compared with children of mothers receiving placebo, iodine concentration in pregnant women and children
aged 612 years (r 069),73 status assessment in school- Substantial evidence suggests that calcium supple-
age children is used to estimate country, regional, and mentation in pregnancy is associated with a reduction
global prevalence of iodine deficiency. Global estimates in gestational hypertensive disorders and preterm
of iodine deficiency suggest that 285% of the worlds birth.89,90 However, the eect varies according to the
population or 19 billion individuals are iodine- baseline calcium intake of the population and pre-
deficient (table 1).46,74 This figure represents largely mild existing risk factors. A review of 1591 randomised con-
deficiency (defined as urinary iodine concentration of trolled trials suggested that calcium supplementation
5099 ug/L). during pregnancy was associated with a reduction in the
Severe iodine deficiency in pregnancy causes cretin- risk of gestational hypertension and a 52% reduction
ism, which can be eliminated with iodine supplemen- in the incidence of pre-eclampsia, along with a
tation before conception or in the first trimester of 24% reduction in preterm birth and an increase in birth-
pregnancy.75 Furthermore, two meta-analyses showed weight of 85 g. There was no eect on low birthweight
average deficits of 125135 intelligence quotient (IQ) or perinatal or neonatal mortality. The eect was mainly
points in children associated with iodine deficiency of noted in populations with low calcium intake.92 The
their mothers in pregnancy; however, they controlled eects of calcium supplementation interventions are
for only a limited number of socioeconomic con- described in the accompanying report by Zulfiqar A
founders.76,77 A review of the eects of iodine supple- Bhutta and colleagues.93
mentation in deficient populations showed a small The US Institute of Medicine has defined adequate
increase in birthweight.78 Eects of mild or moderate vitamin D status as having serum 25-hydroxyvitamin D
iodine deficiency on brain development are not well ([OH]D) concentrations greater than 50 nmol/L in both
established.78,79 The index of iodine deficiency (urinary the general population and pregnant women;94 serum
iodine concentration) is a population measure and not concentrations of less than 25 nmol/L denote vita-
an individual one,80 therefore some individuals in min D deficiency whereas concentrations of less than
regions of mild to moderate deficiency might have more 50 nmol/L denote vitamin D insuciency.95 Although
severe deficiency. few nationally representative surveys exist for vitamin D
status, an estimated 1 billion people globally residing in
Folate diverse geographies, many in LMICs, might be vitamin D
The global prevalence of folate deficiency has not been insucient or deficient.96103
estimated because of the scarcity of suitable population- Vitamin D has an essential role in fetal development,
based data.81 A substantial proportion of neural tube ensuring fetal supply of calcium for bone development,
defects (congenital malformations of the spinal cord and enabling immunological adaptation required to main-
brain) are related to inadequate consumption of folic tain normal pregnancy, preventing miscarriage, and
acid around the time of conception, in some populations promoting normal brain development.99,104108 Poor
associated with genetic factors that increase the need for maternal vitamin D status has been associated with
dietary folic acid. A Cochrane review82 in 2010 included severe pre-eclampsia (new-onset gestational hyper-
five trials of folic acid (a synthetic form of folate) tension and proteinuria after 20 weeks of gestation) in
supplementation and identified a 72% (RR 028; 95% CI turn leading to an increased risk of perinatal morbidity
015052) reduction in the risk of neural tube defects. and mortality.99,109,110 Maternal vitamin D deficiency,
A more recent systematic review had much the same especially in early pregnancy, has been associated
findings and estimated that in 2005 56 000 deaths were with risk of pre-eclampsia (OR 209, 95% CI
attributable to insucient dietary folic acid.83 These 150290), preterm birth (158, 108231), and small-
deaths were not added to the total deaths associated with for-gestational age (SGA; 152, 108225).109,111 A
undernutrition in the present analysis, because of the systematic review of three trials of vitamin D in
uncertainty about this estimate. pregnancy showed an overall reduction of low
birthweight of borderline significance (relative risk
Calcium and vitamin D [RR] 048; 95% CI 023101).111 These associations
Calcium is an essential nutrient for several body func- need to be better quantified before they can be included
tions, including enzymatic and hormonal homoeostasis. in the global disease burden related to undernutrition.
Evidence for the association between maternal dietary
calcium intake and maternal bone density and fetal Eect of maternal stature or BMI on fetal
mineralisation is inconsistent.84 Epidemiological evi- growth restriction or postnatal growth
dence does show an inverse association between calcium Maternal characteristics
intake and development of hypertension in pregnancy.85,86 The 2008 Maternal and Child Undernutrition Series
Gestational hypertensive disorders are the second examined the association of maternal nutritional status
leading cause of maternal morbidity and mortality and (BMI and short stature) and fetal growth restriction,
are associated with increased risk of preterm birth and defined as low birthweight at term. Here, we use data
fetal growth restriction.87,88 from nine (height) and seven (BMI) population-based
cohort studies and WHO perinatal facility-based data for studies that use dierent reference populations for SGA.
24 countries to examine associations separately for term Previous studies have also not separated SGA into term
and preterm SGA in LMICs.112121 Maternal stunting and preterm. In a pooled-analysis130 of 22 population-
(height <145 cm) put infants at risk of term and preterm based cohort studies in LMICs in Asia, sub-Saharan
SGA (appendix p 5). Low maternal BMI in early pregnancy Africa, and Latin America, the RR for neonatal (128 day)
also put infants at higher risk of SGA (appendix p 5). BMI mortality associated with SGA (<10th centile)
of 25 or greater was somewhat protective against term and was 183 (95% CI 134250) and for post-neonatal
preterm SGA (appendix p 5). Notably, most women in the (29365 day) mortality was 190 (132273), compared
BMI category of greater than 25 kg/m were very mildly with appropriate-for-gestational-age (AGA) infants. The
overweight with very few obese women, which might RR for term SGA was 306 (95% CI 221423) for
explain the protective eect. neonatal mortality and 198 (139281) for post-neonatal
Maternal stature is a composite indicator representing
genetic and environmental eects on the growing period 350 Prevalence preterm SGA
of childhood. In a study122 involving 109 Demographic Prevalence term SGA
200
00720074) and to those born to the shortest mothers
(<145 cm) was 0128 (01260130). The correspond- 150
ing absolute risk for a child being stunted was
0194 (01920196) for the tallest mothers and 100
0682 (06730690) for the shortest. The association
with wasting was significant but much weaker.122 50
mortality, relative to term AGA infants. Infants born Causes of childhood growth faltering are multifactorial,
preterm and SGA were at highest risk with RR of 1542 but fetal growth restriction might be an important
(9112612) for neonatal mortality and 522 (283964) contributor to stunting and wasting in children. To
for post-neonatal mortality, relative to term AGA. quantify the association between fetal growth restriction
By applying the attributable fractions of deaths to the and child undernutrition, we did a systematic scientific
total neonatal and post-neonatal deaths for 2011 obtained literature search to identify longitudinal studies that had
from the UN Interagency Group on Mortality Esti- taken measurements of birthweight, gestational age, and
mation,131 we estimated that the number of deaths child anthropometry. Data could be obtained from 19 birth
attributed to SGA in 2011 was 817 000 in neonates and cohort studies which were submitted to a meta-analysis to
418 000 in infants aged 111 months using standard examine the odds of stunting, wasting, and underweight
methods.1 The largest number of attributed deaths were in children 1260 months of age associated with SGA and
in Asia (appendix p 22). For the calculation in the group preterm birth.132 Four mutually exclusive exposure cate-
who were both SGA and preterm, the counterfactual was gories were created: AGA and preterm; SGA and term;
AGA and preterm so that the SGA eect was separated SGA and preterm; and term AGA (as the reference group).
from that of being preterm. In the estimation of the The meta-analysis showed that SGA alone and preterm
deaths attributed to several nutritional conditions, we alone were associated with increased overall ORs of 243
attributed only the neonatal deaths to SGA; for children (95% CI 222266) and 193 (171218), respectively,
aged 159 months, we attributed deaths to wasting and whereas both SGA and preterm increased the OR to 451
stunting, to which SGA contributed (table 2). (342593) for stunting (appendix p 23). These raised
ORs were also noted in the Asian, African, and Latin that is attributable to birth size, independent of that
American UN regions (data not shown). attributable to poor postnatal growth.137
We estimated population attributable risk for childhood Evidence for eects of fetal growth restriction on
stunting for the risk categories of SGA and preterm cognition and behaviour after early childhood is less
birth. Because risk estimates were derived as ORs using consistent. Birthweight was associated with attained
logistic regression analysis, we used a method to schooling in the COHORTS analyses;138 however, this
approximate the risk ratio proposed by Zhang and Yu133 was unadjusted for gestational age. Size at birth was
for estimation of the population attributable risk. Using not related to womens educational achievement in
the approximated RR estimates across all 19 cohorts, Guatemala139 and term low birthweight was not associated
population attributable risk for SGA-term for stunting with IQ and behaviour in school-aged children in Brazil140
was 016 (012019), that for SGA-preterm was 004 and Jamaica,141 or behaviour in South Africa.142 In Taiwan,
(002005), and that for AGA-preterm was 004 term infants of low birthweight had lower academic
(002006). The combined population attributable risk achievement at age 15 years than did infants of normal
related to SGA for stunting was 020 and that for preterm birthweight,143 and in Thailand birthlength was associated
birth was 008. Thus, overall we estimate that about a with IQ at age 9 years independent of postnatal growth to
fifth of childhood stunting could have its origins in the age 1 year;144 however, in both cases eect sizes were small.
fetal period, as shown by being born SGA.
Most studies of fetal growth restriction and childhood Childhood nutrition
cognitive and motor development in LMICs involve term Stunting, underweight, and wasting
infants of low birthweight or examine birthweight Panel 12,145153 describes the determinants of stunting and
adjusted for gestational age. Consistent evidence exists overweight in children. Estimates of the prevalence of
for associations of fetal growth restriction with lower stunting, underweight, and wasting worldwide and for
psychomotor development levels in early childhood (up UN subregions are based on analyses jointly done by
to age 36 months) with small to moderate eect sizes UNICEF, WHO, and the World Bank154 of 639 national
compared with infants of normal birthweight.134,135 A surveys from 142 countries in the WHO database, using
study136 from Bangladesh showed much the same standard methods.1,155,156 In 2011, globally, 165 million
associations for both mental and motor development. children younger than 5 years had a height-for-age
Evidence suggests that there is an eect on development Z score (HAZ) of 2 or lower (stunted) on the basis of
Stunting (HAZ <2) Wasting (WHZ <2) Severe wasting (WHZ <3) Underweight (WAZ <2)
UN155 NIMS158 UN NIMS UN NIMS UN NIMS
Pro- Number Pro- Number Pro- Number Pro- Number Pro- Number Pro- Number Pro- Number Pro- Number
portion (millions) portion (millions) portion (millions) portion (millions) portion (millions) portion (millions) portion (millions) portion (millions)
Africa 356% 563 355% 566 85% 134 79% 125 35% 55 28% 44 177% 279 184% 290
(333 (525 (344 (543 (74 (116 (73 (115 (29 (4564) (25 (4052) (157 (247 (174 (275
380) 600) 366) 578) 96) 152) 86) 136) 41) 33) 197) 311) 191) 301)
Asia 268% 958 295% 1035 101% 361 112% 392 36% 129 36% 127 193% 691 219% 766
(232 (828 (264 (925 (79 (282 (95 (334 (24 (84 (27 (94 (146 (521 (188 (659
305) 1088) 313) 1098) 123) 440) 127) 446) 48) 173) 47) 164) 241) 861) 240) 841)
Latin 134% 71 146% 78 14% 07 15% 08 03% 02 04% 02 34% 18 37% 20
America (94 (4894) (136 (7382) (09 (0510) (13 (0710) (02 (0102) (04 (0203) (2345) (1224) (35 (1822)
and the 177) 155) 19) 18) 04) 06) 41)
Carib-
bean
Oceania 355% 05 347% 04 43% 01 51% 01 07% 00 15% 00 140% 02 139% 02
(160 (0208) (278 (0305) (30 (0001) (33 (0001) (05 (0000) (09 (0000) (80 (0103) (107 (0102)
614) 395) 62) 68) 11) 24) 232) 168)
LMICs 28% 1597 299% 1683 88% 503 93% 526 33% 185 31% 173 174% 990 194% 1091
(256 (1459 (279 (1573 (74 (421 (84 (474 (25 (140 (26 (144 (143 (817 (173 (972
304) 1734) 310) 1746) 103) 584) 104) 585) 40) 231) 38) 215) 204) 1163) 205) 1154)
High- 72% 51 17% 12 03% 02 24% 17
income (41 (2989) (08 (0625) (00 (0009) (1734) (1224)
countries 126) 35) 13)
Global 257% 1648 80% 515 29% 187 157% 1007
(235 (1508 (68 (433 (22 (142 (130 (833
279) 1788) 93) 596) 36) 232) 184) 1180)
Data are % (95% CI). HAZ=height-for-age Z score. WHZ=weight-for-height Z score. WAZ=weight-for-age Z score. LMICs=low-income and middle-income countries.
Table 3: Prevalence and numbers of children younger than 5 years with stunting, wasting, severe wasting, and underweight using estimates from UN and NIMS, by UN regions for 2011
the WHO Child Growth Standards (table 3)a However, with the increase in population in Africa, this is
35% decline from an estimated 253 million in 1990. The the only major world region with an increase in the
prevalence decreased from an estimated 40% in 1990, to number of stunted children in the past decade.
an estimated 26% in 2011an average annual rate of The complex interplay of social, economic, and
reduction of 21% per year (figure 4154). East and west political determinants of undernutrition (figure 1)
Africa, and south-central Asia have the highest preva- results in substantial inequalities between population
lence estimates in UN subregions with 42% (east subgroups. In our analysis, using previously described
Africa) and 36% (west Africa and south-central Asia); methods,158 of 79 countries with population-based
the largest number of children aected by stunting, surveys since the year 2000 (figure 5), stunting
69 million, live in south-central Asia (appendix p 9). prevalence among children younger than 5 years was
The surveys in the WHO database and other population- 247 times (range 100764) higher in the poorest
representative data were also analysed with a Bayesian quintile of households than in the richest quintile. Sex
hierarchical mixture model to estimate Z-score distri- inequalities in child nutrition tend to be substantially
butions of anthropometric indices by the Nutrition smaller than economic inequalities (appendix p 24). In
Impact Model Study (NIMS).157 These distributions were 81 countries with data, stunting prevalence is slightly
used to assess trends in stunting and underweight in higher (114 times, range 083153) in boys than in
children, the present prevalence of these measures, and girls. This finding is consistent with the higher
the present prevalence of wasting. These methods have mortality in children younger than 5 years in boys than
the advantage of estimating the full distribution of in girls in most countries in the world. Place of
anthropometric variables and therefore measure the full residence is also an important correlate of the risk of
extent of mild-to-severe undernutrition without restrictive stunting (appendix p 24). In 81 countries with data,
assumptions. They also allow for non-linear time trends. stunting was 145 times higher (range 094 to 294) in
NIMS analyses resulted in much the same estimates of rural than in urban areas.
the prevalence of stunting, underweight, and wasting According to UN estimates, globally in 2011, more than
as those of the UN in 2011 (table 3, appendix p 10). 100 million children younger than 5 years, or 16%, were
The complete trend analysis showed that the largest underweight (weight-for-age Z score [WAZ] <2 on the
reductions in stunting since 1985 have been in Asia, basis of the WHO Child Growth Standards), a 36%
whereas Africa had an increase until the mid-1990s and decrease from an estimated 159 million in 1990.154
subsequently a modest reduction in the prevalence.157 Estimated prevalences in NIMS were slightly higher at
Global Asia
300 Numbers Prevalence 60 300 60
270 2531 55 270 55
2263 Global target 50 50
240 240
2010 =100 million 45 45
Numbers (millions)
Numbers (millions)
Prevalence (%)
Numbers (millions)
210 210
40 40
Prevalence (%)
Prevalence (%)
180 35 180 35
150 30 150 30
120 25 120 25
20 20
90 90
533 558 581 599 606 15 15
60 457 479 502 60
10 10
30 5 30 137 121 104 90 74 62 5
51 42
0 0 0 0
1990 1995 2000 2005 2010 2015 2020 2025 1990 1995 2000 2005 2010 2015 2020 2025
Year Year
Figure 4: Trends in prevalence and numbers of children with stunted growth (HAZ <2), by selected UN regions and globally, 19902010, and projected to
2025 on the basis of UN prevalence estimates
HAZ=height-for-age Z score. Data from UNICEF, WHO, World Bank.154
110 million (194%).155 Prevalences were highest in south- Standards) was 8% (52 million) globally in 2011, an 11%
central Asia and western Africa where 30% and 22%, decrease from an estimated 58 million in 1990.154 70% of
respectively, were underweight (appendix p 9). the worlds children with wasting live in Asia, most in
The UN estimate for wasting (weight-for-height Z south-central Asia, where an estimated 15% (28 million)
score [WHZ] <2 on the basis of WHO Child Growth are aected. Much the same regional pattern occurs for
severe wasting (WHZ <3), with a global prevalence in 2 years of life are a crucial period linking growth and
2011 of 3% or 19 million children. The highest percen- development; growth from birth to 24 months but not
tages of children with severe wasting are in south-central from 24 to 36 months was associated with child develop-
Asia (51%) and central Africa (56%). ment in Guatemala,167 and weight gain in the first
Suboptimum growth, according to anthropometric 2 years predicted school outcomes in five cohorts.138
measures indicative of stunting, wasting, and under- Analyses from the COHORTS group that are presented
weight, has been shown to increase the risk of death from here suggest that growth in the first 2 years of life, but
infectious diseases in childhood.159,160 This association not at later ages, is associated with achieved school
has been recently re-examined with the pooled analysis grades in adults.168 However, some evidence suggests
of individual-level data from ten longitudinal studies that growth after 24 months of age might also be
involving more than 55 000 child-years of follow-up and associated with lower cognitive ability, but with a
1315 deaths in children younger than 5 years.161 As with smaller eect size than for early growth.169 In a Malawi
previous analyses, all degrees of stunting, wasting, and cohort, height gain from 18 to 60 months predicted
underweight had higher mortality and the risk increased mathematics ability at 12 years. Height gain at 1 month
as Z scores decreased (appendix p 11). Undernutrition and change from 1 to 6 months and 6 to 18 months
can be deemed the cause of death in a synergistic were not significant predictors.170
association with infectious diseases; if the undernutrition We analysed changes in stunting prevalence between
did not exist, the deaths would not have occurred.1 All 1996 and 2008 in Bangladesh, Brazil, and Nigeria,
anthropometric measures of undernutrition were asso- according to wealth and urban or rural status (panel 2,
ciated with increased hazards of death from diarrhoea, figure 6).
pneumonia, and measles; the association was also noted
for other infectious diseases, but not malaria. We Overweight and obesity
calculated the population attributable fractions for stunt- The prevalence of overweight worldwide and for UN
ing, underweight, wasting, and its subset of severe regions is based on the joint analyses done by UNICEF,
wasting using the UN and NIMS prevalence data with WHO, and the World Bank.154 In 2011, globally, an
standard methods.1 These fractions were multiplied by estimated 43 million children younger than 5 years, or
the corresponding age-specific and cause-specific deaths162 7%, were overweight (ie, WHZ greater than two Z scores
to estimate the number of deaths attributable to each above the median WHO standard), on the basis of the
anthropometric measure (table 2).162 For the percentage of WHO Child Growth Standards (appendix p 9)a 54%
total deaths the denominator was 6934 million.131 Details increase from an estimated 28 million in 1990. This
of these estimates for UN subregions and causes of death trend is expected to continue and reach a prevalence of
are in appendix pp 1213. Stunting and underweight have 99% in 2025 or 64 million children (figure 7). Increasing
the highest proportions of attributed child deaths, about trends in child overweight are taking place in most world
14% for both; wasting accounts for 126% (severe wasting regions, not only in high-income countries, where preva-
74%) of child deaths. Table 2 and appendix pp 1415 lence is the highest (15% in 2011). However, most
show estimations using the NIMS prevalence data. In overweight children younger than 5 years (32 million in
these estimates stunting and underweight each account 2011) live in LMICs. In Africa, the estimated prevalence
for 17% of child deaths and wasting for 115% (severe increased from 4% in 1990 to 7% in 2011, and is expected
wasting 78%). to reach 11% in 2025 (figure 7). Prevalence of over-
Stunting is a well established risk factor for poor child weight is lower in Asia (5% in 2011), but the number of
development with numerous cross-sectional studies aected children is higher compared with Africa (17 and
showing associations between stunting and motor and 12 million, respectively).
cognitive development. Several longitudinal studies Dierences in childhood overweight prevalence
show stunting before age 23 years predicts poorer cog- between the richest and poorest quintiles are small in
nitive and educational outcomes in later childhood and most countries (figure 5), and in general prevalence
adolescence.135,163 Eect sizes for the longitudinal studies tends to be higher in the richest quintile than in the
comparing children with HAZ of 2 or lower with non- poorest. In 78 countries with data, prevalence in the
stunted children (HAZ 1) are moderate to large.163 richest quintile was on average 131 (range 055360)
Length-for-age Z score (LAZ) at age 2 years was con- times higher than in the poorest quintile. Overweight is
sistently associated with higher cognitive Z scores in much the same between the sexes (appendix p 24) and
children aged 49 years (017019 per unit change LAZ) slightly more prevalent in urban than in rural areas
across four cohorts with moderate (2432%) or high (appendix p 25). In 81 countries with data, urban preva-
(6786%) stunting prevalence.164 Associations with lence was 108 times higher on average (range 044146)
underweight have also been reported.163 than rural prevalence.
Stunted children show behavioural dierences in Childhood overweight results in both immediate and
early childhood including apathy, more negative aect, longer-term risks to health. Among the immediate risks
and reduced activity, play, and exploration.165,166 The first are metabolic abnormalities including raised cholesterol,
70
no cognitive benefit when followed up at age 9 years. 60
Four additional randomised controlled trials184187 50
examined the combined eects of iron and folate
40
supplementation in children younger than 36 months
30
of age. One showed benefits to motor milestones,184
20
others showed no benefits to motor185,186 or language186
10
milestones, and one showed no benefit to cognitive
function.187 Thus, there is some evidence that iron 0
96
99
04
07
96
06
03
08
20
20
20
20
19
19
19
younger than 4 years, but no consistent evidence for an Bangladesh Brazil Nigeria
eect on mental development. However, many of the
supplementation trials produced only small dierences
Figure 6: Changes in stunting over time, in Bangladesh, Brazil, and Nigeria
in iron status between the treated and control groups, (A) Stunting prevalence by wealth quintile. The longer the line between
possibly limiting their ability to aect development. It two groups, the greater the inequality. (B) Stunting prevalence by urban or
is also possible that mental development takes longer to rural location.
Global Africa
65 Numbers Prevalence 637 14 65 14
60 60
552
55 12 55 12
50 478 50
10 10
45 45
412
Numbers (millions)
Numbers (millions)
40 40
Prevalence (%)
Prevalence (%)
353 8 8
35 317 35
284 296
30 30
6 6
25 25 214
20 20 173
4 4
15 15 138
110
10 10 87
2 69 2
46 56
5 5
0 0 0 0
50 50
10 10
45 45
Numbers (millions)
Numbers (millions)
40 40
Prevalence (%)
Prevalence (%)
8 8
35 35
30 30
6 6
25 147 231 25
144 141 143 205
20 184 20
165 4 4
15 15
10 2 10 2
5 5 36 37 38 39 38 38 38 38
0 0 0 0
1990 1995 2000 2005 2010 2015 2020 2025 1990 1995 2000 2005 2010 2015 2020 2025
Year Year
Figure 7: Trends in prevalence and numbers of overweight (WHZ >2) children, by selected UN regions and globally, 19902010, and projected to 2025, on the
basis of UN prevalence estimates
WHZ=weight-for-height Z score.
improve than the duration of the trials or that the has declined, probably because of large-scale vitamin A
eects of iron deficiency early in life are irreversible. supplementation programmes in many countries, sub-
clinical vitamin A deficiency aects high proportions of
Vitamin A children in Africa and southeast Asia (table 2).
Clinical assessment of ocular symptoms and signs of Many randomised controlled trials have been done
xerophthalmia and biochemical assessment of serum to examine the eect of supplementation every
concentration of retinol are the two common methods 46 months and fortification on survival of children
in population surveys for estimation of prevalence of aged 6 months and older; these studies provide the
vitamin A deficiency. WHO provides prevalence esti- best evidence for deaths attributable to vitamin A
mates of vitamin A deficiency in preschool children deficiency.188191 Meta-analyses of these trials show a
(<5 years) for 19952005 from 99 countries.45 Globally, mortality reduction of 23%,188 30%,189 and 24%190 in
09% (95% CI 0315) or 517 million preschool age children aged 659 months. With publication of a large
children are estimated to have night blindness and programme eectiveness study from India, a revised
333% (311354) or 90 million to have subclinical meta-analysis shows a mortality eect of 11%, still a
vitamin A deficiency, defined as serum retinol concen- statistically significant benefit.192 In our calculations we
tration of less than 070 mol/L. Vitamin A deficiency use only the eects in the trials on particular causes of
using night blindness prevalence can be defined as a death, not the eects on overall deaths because the
global problem of mild public health importance,45 causes of death at the time of the trials and nowadays
although the prevalence in Africa (2%) is higher than are probably dierent (eg, diarrhoea and measles
elsewhere. Although prevalence of clinical symptoms account for a much smaller proportion of child deaths
now than 1020 years ago).162 To derive the risk of method allows the adjusted RRs to be applied to the
vitamin A deficiency the inverse (1/risk reduction) of present prevalence of inadequate zinc intake in countries
the cause-specific mortality reduction identified in the to estimate the attributable deaths. The population
trials was deemed to be the risk of deficiency and attributable fractions for diarrhoea and pneumonia were
adjusted with the assumption that all the eect was in multiplied by the number of these deaths in 2011.162 The
the subset of the trial population with low serum retinol number of child deaths attributed to zinc deficiency in
(appendix p 16). Much the same adjustment was done 2011 is 116 000 (table 2).
for the eect of vitamin A deficiency on diarrhoea Zinc deficiency also has a small negative eect on
incidence (appendix p 17). This allows the adjusted RR growth. A meta-analysis of randomised controlled trials of
to be applied to the present prevalence of low serum zinc supplementation showed a significant benefit for
retinol to estimate the attributable deaths or disease linear growth in children aged 05 years.201 The eect was a
episodes. Although results of trials193195 in south Asia gain of 037 cm in zinc-supplemented children. Trials that
show a newborn supplementation eect on mortality in used a dose of zinc of 10 mg per day for 24 weeks, rather
the first 6 months of life, the evidence from Africa is than lower doses, showed a larger benefit of 046 cm.
less clear, and further studies are underway. Therefore,
we do not estimate deaths attributable to vitamin A Breastfeeding practices
deficiency in the first 6 months of life. Child deaths Present recommendations are that babies should be
attributable to vitamin A deficiency for 2011 are put on the breast within 1 h after birth, be exclusively
estimated to be 157 000 (table 2). breastfed for the first 6 months, and for an additional
18 months or longer, be breastfed along with comple-
Zinc mentary foods. There are no recently published system-
Zinc deficiency in populations could be assessed by a atic compilations of data for breastfeeding patterns so
shift of the population distribution of serum zinc con- we analysed data from 78 countries with surveys done
centrations to lower values as recommended by the in LMICs during 200010 (appendix p 20). Early
International Zinc Nutrition Consultative Group;196 how- initiation of breastfeeding (within 1 h) is highest in
ever, insucient data exist to classify countries or sub- Latin America (mean 58%, 95% CI 5067), intermediate
national populations. Instead the proportion of the in Africa (50%, 4555) and Asia (50%, 4258), and
national population estimated to have an inadequate lowest in eastern Europe (36%, 2350). Except for
zinc intake on the basis of national food availability and eastern Europe, where the lowest rates of breastfeeding
dietary requirements is used.70 According to this method, are recorded, globally about half of children younger
an estimated 17% of the worlds population has an than 1 month, and three in every ten children aged
inadequate zinc intake; substantial regional variation 15 months are exclusively breastfed. Breastfeeding in
exists, with Asia and Africa having the highest preva- 623 month olds is most frequent in Africa (mean 77%,
lences (table 1). A systematic review197 showed that zinc 95% CI 7381) followed by Asia (62%, 5471) and Latin
supplementation resulted in a 9% reduction (RR 091, America (60%, 5069), with lower occurrence in eastern
95% CI 082101) of borderline significance in all- Europe (33%, 2442).
cause child mortality. A separate analysis of available The risks of increased mortality and morbidity due to
trials showed a significant 18% reduction (RR 082, deviation from present breastfeeding recommendations
070096) in all-cause mortality in children aged 14 are well documented.1 Updated systematic reviews of
years.198 There were suggestive benefits on diarrhoea- these risks have results that are much the same as our
specific (RR 082, 95% CI 064105) and pneumonia- previous estimates (appendix p 21).1,202,203 The number of
specific (085, 065111) mortality.197 This and previous child deaths attributed to suboptimum breastfeeding in
analyses have included cause-specific mortality eects 2011 is 804 000 or 116% of all deaths (table 2).
even when they were not statistically significant when Three prospective case-cohort studies provide data for
the eect on all-cause mortality was statistically the association of early breastfeeding initiation (within
significant.199,200 These trials were not powered for cause- 24 h) with neonatal mortality.204 Although early initiation
specific mortality eects. Supporting evidence for the was associated with lower neonatal mortality (RR 056,
cause-specific mortality eects comes from randomised 95% CI 046079), in babies who were exclusively
controlled trials that showed significant reductions in breastfed the mortality risk was not significantly
diarrhoea incidence (RR 087, 95% CI 081094) and reduced (069, 027175). The possible benefit of early
pneumonia incidence (081, 073090).197 To derive the initiation of breastfeeding was therefore not deemed to
risk of zinc deficiency, we adjusted the inverse of the be additive to the eects of exclusive breastfeeding in
cause-specific mortality (appendix p 18) and incidence our analyses.
(appendix p 19) reductions noted in the trials with the A systematic review shows that breastfeeding is con-
assumption that all the eect was in the subset of the sistently associated with an increase in IQ of about three
trial population at risk of zinc deficiency, as estimated points,205 even after adjustment for several confound-
from the availability of food in national diets.70 This ing factors including maternal IQ. Evidence for the
protection aorded by breastfeeding against risk factors attributed jointly to fetal growth restriction, suboptimum
for non-communicable diseases is less consistent. A breastfeeding, stunting, wasting, and deficiencies of
series of meta-analyses,205 based mainly on studies in vitamin A and zinc. The overall results were the same
adults from high-income settings, showed no evidence of using UN or NIMS prevalence estimates. As part of this
protection against total cholesterol levels or diastolic total, the joint distribution of suboptimum breastfeeding
blood pressure. When the meta-analysis was restricted to and fetal growth restriction in the neonatal period
high-quality studies, breastfeeding was associated on contributes 13 million deaths or 19% of all deaths of
average with a 1 mm reduction in systolic blood pressure, children younger than 5 years.
and with a 12% reduction in the risk of overweight or
obesity. Studies of diabetes or glucose levels were too few Eects of fetal and early childhood
to allow a firm conclusion. undernutrition on adult health
In The Lancets 2008 Series on maternal and child
Joint eects of nutritional conditions on child undernutrition,2 consequences of early childhood
mortality nutrition on adult health and body composition were
To estimate the population attributable fraction and the assessed by reviewing the scientific literature and doing
number of deaths attributable to several risk factors, we meta-analyses of five birth cohorts from LMICs (India,
used Comparative Risk Assessment methods.206 We the Philippines, South Africa, Guatemala, Brazil), an
calculated attributable deaths from four specific causes eort that gave rise to the COHORTS collaboration.208 On
of mortality (diarrhoea; measles; pneumonia; and other the one hand, the conclusions were that small size at
infections, excluding malaria) associated with dierent birth and at 2 years of age (particularly height) were
nutritional status measures. In the neonatal period all associated with reduced human capital: shorter adult
deaths were deemed to be associated only with sub- height, less schooling, reduced economic productivity,
optimum breastfeeding and fetal growth restriction. The and for women, lower ospring birthweight. On the
Comparative Risk Assessment methods allow the other hand, larger child size at 24 months of age was a
estimation of the reduction in death that would take risk factor for high glucose concentrations, blood
place if the risk factors were reduced to a minimum pressure, and harmful lipid levels once adjustment for
theoretical level or counterfactual level of exposure.206 For adult BMI was made, suggesting that rapid weight gain,
the assessment the following formula was used: especially after infancy, is linked to these conditions.
The COHORTS group later did pooled analyses that use
n conditional growth variables that remove the correlation
P ( RR i i 1) between growth measures across ages, allowing inferences
i =1
PAF = n to be made about the relative association between growth
P (RR i i 1) + 1 during specific age intervals and outcomes.149 Also,
i =1 because gains in height and weight are correlated, these
analyses use conditionals that separate linear growth from
For the purpose of this analysis, RRi equals the RR of weight gain. The conclusions were that heavier birthweight
mortality for the ith exposure category. Pi equals the and faster linear growth from 0 to 2 years lead to large
proportion of children in the ith exposure category, for gains in human capital, but have little association with
dierent levels of fetal growth restriction, stunting, adult cardiovascular risk factors. Also, faster weight gain
wasting, deficiencies of vitamin A and zinc, and sub- independent of linear growth has little benefit for human
optimum breastfeeding practices in their relevant age capital. After the age of 2 years, and particularly after the
groups in specific UN regions. age of 4 years, rapid weight gains show adverse eects on
Because specific causes of deaths could potentially be adult cardiovascular risk. The COHORTS analyses control
caused by more than one factor, the population for confounding and the findings are broadly much the
attributable fractions for multiple risk factors that aect same across the five cohorts.
the same disease outcome overlap and cannot be Additional longitudinal studies report eects on later
combined by simple addition. For this reason, the joint mental health with higher levels of depression and
population attributable fraction was estimated with the anxiety and lower self-esteem in adolescents who were
formula: Joint PAFi=1product (1PAFi), where PAFi stunted by age 2 years compared with non-stunted;209
equals the population attributable fraction of the dierent increased depression in adolescents who had severe
risk factors.207 All analyses were done for relevant age acute malnutrition in infancy;210 increased risk of suicidal
groups of that risk factor and then the results from the ideation associated with lower HAZ at 24 months,211 and
age groups were aggregated. For stunting and wasting higher levels of hyperactivity in late adolescence and
we did calculations separately with UN and NIMS attention deficit in adults.209,212 Although observational,
prevalence estimates. the findings are persuasive and consistent with the little
The resulting 31 million deaths constitute 45% of quasi-experimental evidence available. Follow-up studies
global deaths in children younger than 5 years in 2011, of a community randomised nutrition trial in Guatemala
have shown long-term eects of exposure to improved fetal growth restriction and maternal nutritional con-
nutrition during the first 23 years, but not after 3 years, ditions, short stature, and low BMI, should lead to more
on education,213 and wages.214 Eects of improved nutri- emphasis on nutritional interventions before and during
tion in the first several years of life on risk factors for pregnancy. These interventions would have benefits for
chronic diseases were minor but in some cases bene- health of adolescents and women, could reduce compli-
ficial.215 Exposure to improved nutrition during child- cations of pregnancy and delivery for the mother, and
hood aected the growth of the next generation in girls enhance fetal growth and development.
and their future children.2 Stunted linear growth has become the main indicator
Famines are another source of information about long- of childhood undernutrition, because it is highly
term eects of poor nutrition in early life. The Dutch prevalent in all developing regions of the world, and has
famine of 194445 (brief, intense, but aecting a pre- important consequences for health and development. It
viously well-nourished population) suggests eects of should replace underweight as the main anthropometric
prenatal exposure on schizophrenia, no eects on human indicator for children. Underweight indices include chil-
capital (height, cognitive function), and weak and in- dren who are short, but who can have an increased
consistent eects on cardiovascular risk factors.216 The WHZ, being therefore at increased risk of long-term
195961 Chinese famine (prolonged, severe, and aecting adverse health outcomes. Linear growth assessment in
an already malnourished population) suggests eects of primary care is an essential component of country eorts
exposure during pregnancy and the first 2 years of life to reduce childhood stunting. More experience is needed
on height, wealth, income, mental health, and inter- in the operational aspects of the assessment and inter-
generational eects on birthweight.217220 Surprisingly, pretation of linear growth by health workers and in the
some findings suggest protective eects of famine expo- eective intervention responses.
sure, which might be explained by high mortality and Prevalence of stunting in children younger than 5 years
intense selection of the hardiest.219221 in developing countries in 2011 was about 28%, a
decrease from 40% or more in 1990 and the 32% estimate
Discussion in our 2008 nutrition Series for 2004.1 The number of
Nutrition has profound eects on health throughout the stunted children globally has decreased from 253 million
human life course and is inextricably linked with cog- in 1990 to 165 million in 2011. Another reported estimate
nitive and social development, especially in early child- was 167 million in developing countries for 2010.157 The
hood. In settings with insucient material and social 13-year Comprehensive Implementation Plan (201225)
resources, children are not able to achieve their full on Maternal, Infant and Young Child Nutrition, endorsed
growth and developmental potential. Consequences at the 2012 World Health Assembly, includes six global
range broadly from raised rates of death from infectious nutrition targets, the first of which calls for a 40%
diseases and decreased learning capacity in childhood to reduction of the global number of children younger than
increased non-communicable diseases in adulthood. 5 years who are stunted by 2025 (compared with the
Nutrition and growth in adolescence is important for a baseline of 2010).222 This goal would translate into a 39%
girls health and adult stature. Women with short stature relative reduction per year and imply reducing the
are at risk of complications in delivery, such as obstructed number of stunted children from the 171 million in 2010
labour. Nutritional status at the time of conception and to about 100 million. However, at the present rate of
during pregnancy is crucial for fetal growth. Babies with decline, prevalence of stunting is expected to reach 20%,
fetal growth restriction, as shown by being SGA, are at or 127 million, in 2025. In Africa, only small reductions
increased risk of death throughout infancy. We estimate in prevalence are anticipated on the basis of present
that 32 million babies are born SGA, 27% of births trends. However, in view of the rising number of births,
in LMICs, and about 800 000 neonatal deaths and the actual number of stunted children will increase from
400 000 post-neonatal infant deaths can be attributed to 56 to 61 million. By contrast, Asia is projected to show a
the increased risk associated with having fetal growth substantial decrease in stunting prevalence and in the
restriction. The use of more appropriate methods than absolute number of children aected.
those in our previous review1 to understand the preva- Stunted, underweight, and wasted children have an
lence and risk of SGA presented here resulted in increased risk of death from diarrhoea, pneumonia,
estimates that are more than double our previous measles, and other infectious diseases. Recalculated
estimate of attributable neonatal deaths related to term risks of cause-specific mortality161 confirmed previous
low birthweight. This new finding contradicts the estimates1,159,160 and were used in estimation of the deaths
widespread assumption that SGA infants, by contrast attributable to these conditions. The attributable fraction
with preterm babies, are not at a substantially increased of deaths for these nutritional status measures has
risk of mortality. Additionally, babies who are SGA have remained nearly the same as in previous calculations,
an increased risk of growth faltering in the first 2 years of but the number of attributed deaths has declined because
life; our estimates suggest that 20% of stunting might be of the decrease in the prevalence of these measures and a
attributable to fetal growth restriction. The links between decline in the aected causes of death. By our estimates
more than 1 million deaths can be attributed to stunting is highly beneficial. Data from developing regions show
and about 800 000 to wasting, about 60% of which are that present practices are far from optimum, despite
attributable to severe wasting. These attributable deaths improvements in some countries. Our previous estimates
cannot be added because of the overlap of these and suggested that 14 million deaths were attributed to
other nutritional conditions, but are instead included in suboptimum breastfeeding, especially related to non-
the calculation of the deaths attributed to the joint exclusive breastfeeding in the first 6 months of life. Our
eects of all nutritional conditions. Our estimate of new estimates are 804 000 deaths, a substantial reduction
about 1 million child deaths due to underweight is higher since our estimate for the year 2004. Although present
than the recently published 860 000 deaths in the Global estimates of the risks associated with suboptimum
Burden of Disease (GBD).200 The GBD Study did not breastfeeding practices are almost unchanged from what
estimate deaths attributed to stunting or wasting. was used previously, the reduction in infectious disease
Deficiencies of vitamins and minerals have important deaths in the past 10 years results in a lower number of
health consequences, both through their direct eects, preventable deaths.173 Our number is higher than the
such as iron deficiency anaemia, xerophthalmia due to estimated 545 000 deaths attributed to suboptimum
vitamin A deficiency, and iodine deficiency disorders, breastfeeding in the GBD Study.200 Insucient methods
and because they increase the risk of serious infectious are provided by GBD to understand the reasons for
diseases. In the latter category, vitamin A and zinc dierences, but one reason is probably the lower
deficiencies have been shown to have the greatest eects estimates of aected diarrhoea and pneumonia deaths
among the micronutrients. Our estimate of the nearly than the UN estimates we used.223
157 000 child deaths attributed to vitamin A deficiency is To work out the total deaths attributed to nutritional
smaller than our previous estimate (668 000).1 Notably, conditions, we calculated the joint distribution of
we did not include the risk of mortality in the first stunting, wasting, fetal growth restriction, deficiencies of
6 months of life, as we did last time for Asia, because of vitamin A and zinc, and suboptimum breastfeeding. The
more uncertainty about this risk; several trials of neonatal resulting 31 million deaths constitute 45% of the
vitamin A supplementation are underway in Asia and 69 million global child deaths in 2011. The total number
Africa to assess the benefits. Likewise, our estimate of of attributed deaths is reduced from the 35 million we
116 000 child deaths attributed to zinc deficiency is much estimated for 2004, despite the population attributable
smaller than the 450 000 in our previous estimates, partly fraction increasing from 35% to 45%. In this period the
because we now estimate the risk of pneumonia and mortality in children younger than 5 years decreased
diarrheal mortality beginning at 12 rather than 6 months from 10 million to 69 million deaths and there were
of age. In both cases there have also been changes in our even larger decreases in the causes of deathsuch as
methods of estimating the prevalence of the conditions measles, diarrhoea, and pneumonia162that are most
and the risk associations, and combining these factors to aected by nutritional conditions. The increase in the
get attributed fractions of deaths. These methods are percentage of attributed deaths compared with the
much the same as those reported by the GBD Study.200 previous series is largely because of the more appropriate
Our estimates are 20% higher for zinc and 30% higher inclusion of the eects of SGA (about 800 000 deaths)
for vitamin A, than the GBD estimates. The reduction instead of term low birthweight (337 000 deaths) and the
from our estimates for the year 2004 is partly explained inclusion of all wasting (WHZ <2) instead of severe
by decreases in the prevalences of these deficienciesfor wasting (WHZ <3) in the joint calculations.
vitamin A mostly because of large-scale implementation The number of children aected by excessive body-
of high-dose supplementation programmes, and a weight relative to length or height is increasing globally.
reduction in deaths from diarrhoea and measles aected Although the prevalence of overweight in high-income
by vitamin A deficiency and in deaths from diarrhoea countries is more than double that in LMICs, three
and pneumonia aected by zinc deficiency. Some of quarters of the global total live in LMICs. The recorded
this mortality reduction (eg, for diarrhoea) could be trends in the prevalence of early childhood overweight are
attributable to the vitamin A supplementation, but might probably a consequence of changes in dietary and physical
be also related to improvement in nutritional status activity patterns over time. These behavioural changes
measures (eg, stunting), water and sanitation, and illness are aected by many social and environmental factors,
treatment, and for measles is largely attributable to including interpersonal (family, peers, and social net-
improved coverage with measles vaccine. Although the works), community (school, workplace, and institutions),
number of deaths attributed to the present prevalence of and governmental (local, state, and national policies).224
vitamin A deficiency is relatively small, importantly, a Studies show that the trend toward childhood obesity can
reduction in present coverage of vitamin A interventions start as early as age 6 months.225,226 In LMICs, rapid weight
would probably result in an increase in mortality, because gain after 2 years of age is particularly associated with
in most LMICs dietary intake is still inadequate. adult fat mass.2,168 Intrauterine, infant, and preschool
Breastfeeding exclusively for the first 6 months and periods are deemed to be possible crucial periods for
then along with complementary food to 24 months of age programming long-term regulation of energy balance.227229
If trends are not reversed, increasing rates of childhood health consequences, with a high child mortality
overweight and obesity will have vast implications not burden related to both stunting and wasting. New
only for future health-care expenditures but also for the evidence in this paper supports the focus on pregnancy
overall development of nations. These findings confirm and the first 2 years of life, the crucial 1000 days, which
the need for eective interventions and programmes to we called for in the previous series, but adds more
reverse anticipated trends. The early recognition of exces- emphasis to the nutritional conditions in adolescence,
sive weight gain relative to linear growth is essential. at the time of conception, and during pregnancy, as
Routine assessment of both weight and length in all important for maternal health and survival, fetal growth
children needs to become standard clinical practice from and subsequent early childhood survival, growth, and
very early childhood. development. Fetal growth restriction and poor growth
We have reported previously and confirmed in a new early in infancy are now recognised as important
analysis that anaemia that is reduced by supplemental determinants of neonatal and infant mortality, stunting,
iron in pregnancy is a risk factor for more than a and overweight and obesity in older children and
quarter of maternal deaths. Anaemia is probably a adults. Preventive eorts should continue to focus on
particularly important risk factor for haemorrhage, the the 1000 days, while therapeutic eorts continue to
leading cause of maternal deaths (23% of total deaths).223 target severe wasting.
Additionally, there is now sound evidence that calcium
deficiency increases the risk of pre-eclampsia, now the
second most important cause of maternal death (19% of Panel 3: Research priorities
total deaths).223 Thus, addressing deficiencies of these How much of the eect of the underlying economic and social determinants of
two minerals could result in substantial reduction of optimum growth and development is mediated through known, measurable
maternal deaths. proximate determinants?
Nutrition is crucial for optimum child development What are the consequences of calcium or vitamin D deficiencies for maternal health,
throughout the first 1000 days of life and beyond. fetal growth, birth outcomes, and infant health?
Maternal nutrition aects fetal growth and brain develop- What is the importance of zinc deficiency or other micronutrient deficiencies in the
ment with clear evidence for the need to continue and risk of preterm delivery?
strengthen programmes to prevent maternal iodine What is the role of nutrition in adolescence, at the time of conception, and in
deficiency. Fetal growth restriction and postnatal growth pregnancy in healthy fetal growth and development?
aect motor and cognitive development, with the largest What factors are associated with regional variation in fetal growth restriction and do
eects from stunting before age 23 years. Growth later its consequences on growth and mortality vary by setting?
in childhood might also aect development though with What is the interaction of fetal growth restriction and post-partum infections and diet
a smaller eect size. Iron deficiency anaemia aects in contributing to stunting?
motor development; eects on cognitive development What is the role of vitamin A deficiency in newborn babies or the first 6 months of life
have been more clearly shown in children older than in neonatal and infant mortality?
5 years than in younger children. Are there benefits from iron supplementation for mental development in children
Promotion of good early nutrition is essential for younger than 5 years with iron deficiency anaemia?
children to attain their developmental potential; however, Do deficiencies of any micronutrients other than vitamin A and zinc increase the risk
poor nutrition occurs with other risks for development, of mortality from infectious diseases?
in particular inadequate stimulation during early child- Can the risk of stunting be explained by specific dietary factors, micronutrient
hood. Interventions to promote home stimulation and deficiencies, and clinical or subclinical infections?
learning opportunities in addition to good nutrition will How can the rapid decreases in stunting noted in selected countries be explained by
be needed to ensure good early development and longer- changes in intervention coverage and in distal determinants?
term gains in human capital, with integration and coord- Are there benefits of early initiation of breastfeeding that are independent of the
ination of programmes and policies across sectors. practice of exclusive breastfeeding?
Panel 3 lists research priorities. Additionally, there is a How can developmental dierences associated with nutritional deficiencies at
general need for better data on micronutrient deficiencies various ages of childhood be best quantified and related to functioning and
and other nutritional conditions at national and sub- productivity in adulthood?
national levels. This work should involve the development What are the optimum physical growth rates in the first year of life to avoid both
and use of improved biomarkers that could be used to undernutrition and obesity?
describe nutritional conditions and increase knowledge What are the eects of physical growth rates and stunting on cognitive abilities and
about how they aect health and development. Such psychological functioning?
information is needed to guide intervention programmes Can health workers eectively and eciently assess and interpret height-for-age
in countries and priorities for support globally. measures of linear growth in health and nutrition programmes?
In the 5 years since our last series there have How important is optimum growth and nutritional status in the fetal and early
been some improvements in nutritional conditions, childhood period (first 1000 days of life) in the development of adult obesity and
especially for child growth. Nonetheless, the extent of non-communicable diseases?
these conditions remains high with serious detrimental
*Mark A Boyd, Basil Donovan 5 Ra F, Rachlis A, Stellbrink HJ, et al, on behalf of the SPRING-2 study group.
The Kirby Institute for infection and immunity in society, Once-daily dolutegravir versus raltegravir in antiretroviral-naive adults
with HIV-1 infection: 48 week results from the randomised, double-blind,
University of New South Wales, Sydney, NSW 2052, Australia non-inferiority SPRING-2 study. Lancet 2013; 381: 73543.
mboyd@kirby.unsw.edu.au 6 Walmsley S, Antela A, Clumeck N, et al. Dolutegravir (DTG; S/GSK1349572)
+ abacavir/lamivudine once daily statistically superior to tenofovir/
MAB has received research support from AbbVie, Gilead Sciences, and Merck; emtricitabine/efavirenz: 48-week resultsSINGLE (ING114467).
and honoraria for serving on advisory boards and speaking from AbbVie, 52nd International Conference on Antimicrobials and Chemotherapy;
Boehringer Ingelheim, Bristol-Myers Squibb, Gilead Sciences, Janssen Sept 912, 2012; San Francisco, CA, USA. Presentation H-556b.
Pharmaceuticals, and Merck. BD declares that he has no conflicts of interest. 7 Cahn P, Pozniak AL, Mingrone H, et al, on behalf of the extended SAILING
1 Lennox JL, DeJesus E, Lazzarin A, et al; STARTMRK investigators. Safety and Study Team. Dolutegravir versus raltegravir in antiretroviral-experienced,
ecacy of raltegravir-based versus efavirenz-based combination therapy in integrase-inhibitor-naive adults with HIV: week 48 results from the
treatment-naive patients with HIV-1 infection: a multicentre, double-blind randomised, double-blind, non-inferiority SAILING study. Lancet 2013;
randomised controlled trial. Lancet 2009; 374: 796806. published online July 3. http://dx.doi.org/10.1016/S0140-6736(13)61221-0.
2 Steigbigel RT, Cooper DA, Kumar PN, et al; BENCHMRK Study Teams. 8 Underwood MR, Johns BA, Sato A, Martin JN, Deeks SG, Fujiwara T.
Raltegravir with optimized background therapy for resistant HIV-1 The activity of the integrase inhibitor dolutegravir against HIV-1 variants
infection. N Engl J Med 2008; 359: 33954. isolated from raltegravir-treated adults. J Acquir Immune Defic Syndr 2012;
3 Sax PE, DeJesus E, Mills A, et al, for the GS-US-236-0102 study team. 61: 297301.
Co-formulated elvitegravir, cobicistat, emtricitabine, and tenofovir versus 9 Madruga JV, Cahn P, Grinsztejn B, et al; DUET-1 study group. Ecacy and
co-formulated efavirenz, emtricitabine, and tenofovir for initial treatment safety of TMC125 (etravirine) in treatment-experienced HIV-1-infected
of HIV-1 infection: a randomised, double-blind, phase 3 trial, analysis of patients in DUET-1: 24-week results from a randomised, double-blind,
results after 48 weeks. Lancet 2012; 379: 243948. placebo-controlled trial. Lancet 2007; 370: 2938.
4 DeJesus E, Rockstroh JK, Henry K, et al, for the GS-236-0103 Study Team. 10 Nichols G, Grossberg R, Lazzarin A, et al. Antiviral activity of dolutegravir in
Co-formulated elvitegravir, cobicistat, emtricitabine, and tenofovir subjects with failure on an integrase inhibitor-based regimen: week 24
disoproxil fumarate versus ritonavir-boosted atazanavir plus phase 3 results from VIKING-3. Abstracts of the 11th International Congress
co-formulated emtricitabine and tenofovir disoproxil fumarate for initial on Drug Therapy in HIV Infection. J Int AIDS Soc 2012; 15 (suppl 4): 18112.
treatment of HIV-1 infection: a randomised, double-blind, phase 3,
non-inferiority trial. Lancet 2012; 379: 242938.
Lancet 2013; 382: 53651 Acceleration of progress in nutrition will require eective, large-scale nutrition-sensitive programmes that address
Published Online key underlying determinants of nutrition and enhance the coverage and eectiveness of nutrition-specific
June 6, 2013 interventions. We reviewed evidence of nutritional eects of programmes in four sectorsagriculture, social safety
http://dx.doi.org/10.1016/
nets, early child development, and schooling. The need for investments to boost agricultural production, keep prices
S0140-6736(13)60843-0
low, and increase incomes is undisputable; targeted agricultural programmes can complement these investments by
This online publication has
been corrected. The corrected supporting livelihoods, enhancing access to diverse diets in poor populations, and fostering womens empowerment.
version first appeared at However, evidence of the nutritional eect of agricultural programmes is inconclusiveexcept for vitamin A from
thelancet.com on June 20, 2013 biofortification of orange sweet potatoeslargely because of poor quality evaluations. Social safety nets currently
This is the third in a Series of provide cash or food transfers to a billion poor people and victims of shocks (eg, natural disasters). Individual studies
four papers about maternal and show some eects on younger children exposed for longer durations, but weaknesses in nutrition goals and actions,
child nutrition
and poor service quality probably explain the scarcity of overall nutritional benefits. Combined early child development
*Members listed at end of paper
and nutrition interventions show promising additive or synergistic eects on child developmentand in some cases
Poverty, Health and Nutrition
nutritionand could lead to substantial gains in cost, eciency, and eectiveness, but these programmes have yet to
Division, International Food
Policy Research Institute, be tested at scale. Parental schooling is strongly associated with child nutrition, and the eectiveness of emerging
Washington, DC, USA school nutrition education programmes needs to be tested. Many of the programmes reviewed were not originally
(M T Ruel PhD, H Alderman PhD) designed to improve nutrition yet have great potential to do so. Ways to enhance programme nutrition-sensitivity
Correspondence to: include: improve targeting; use conditions to stimulate participation; strengthen nutrition goals and actions; and
Dr Marie T Ruel, Poverty, Health
optimise womens nutrition, time, physical and mental health, and empowerment. Nutrition-sensitive programmes
and Nutrition Division,
International Food Policy can help scale up nutrition-specific interventions and create a stimulating environment in which young children can
Research Institute, Washington, grow and develop to their full potential.
DC 20006, USA
m.ruel@cgiar.org
Introduction increasing incomes, and urbanisationshifts that raise
The food system is threatened by food and oil price concerns about diet quality and food safety, while
volatility, diversion of resources from production of food threatening water, land, and other finite natural
to biofuels, climate change and related water short- resources.58 In view of these challenges, protection of
ages, persistent conflicts and emergencies, and natural nutrition, let alone acceleration of progress, will entail
disasters aecting agriculture production and yields.14 more than bringing nutrition-specific interventions to
These challenges are compounded by changes in demand scale. It will require a new and more aggressive focus on
for food that are brought about by growing populations, coupling eective nutrition-specific interventions (ie,
those that address the immediate determinants of
Key messages nutrition) with nutrition-sensitive programmes that
address the underlying causes of undernutrition
Nutrition-sensitive interventions and programmes in agriculture, social safety nets, (panel 19,10).
early child development, and education have enormous potential to enhance the scale Nutrition-sensitive programmes draw on comple-
and eectiveness of nutrition-specific interventions; improving nutrition can also help mentary sectors such as agriculture, health, social pro-
nutrition-sensitive programmes achieve their own goals. tection, early child development, education, and water
Targeted agricultural programmes and social safety nets can have a large role in and sanitation to aect the underlying determinants of
mitigation of potentially negative eects of global changes and man-made and nutrition, including poverty; food insecurity; scarcity of
environmental shocks, in supporting livelihoods, food security, diet quality, and access to adequate care resources; and to health, water,
womens empowerment, and in achieving scale and high coverage of nutritionally and sanitation services.11 Key features that make pro-
at-risk households and individuals. grammes in these sectors potentially nutrition-sensitive
Evidence of the eectiveness of targeted agricultural programmes on maternal and are: they address crucial underlying determinants of
child nutrition, with the exception of vitamin A, is limited; strengthening of nutrition nutrition; they are often implemented at large scale and
goals and actions and rigorous eectiveness assessments are needed. can be eective at reaching poor populations12 who
(Continues on next page) have high malnutrition rates; and they can be leveraged
to serve as delivery platforms for nutrition-specific
consistent across all reviews (appendix pp 34). First, these 0 2000 4000 6000 8000 10 000
GDP per person (2005 international $)
reviews note that there is little evidence of eectiveness of
homestead food production programmes on maternal or Figure 1: Prevalence of stunting in children aged 05 years and GDP per person
child nutritional status (anthropometry or micronutrient Most observations for prevalence of stunting are from 200008. The fitted curves are locally weighted regressions
of prevalence of stunting in children aged 05 years and poverty (<$125 per person, per day), against GDP per
status), with the possible exception of vitamin A status. For person. The adjustment to international dollar units converts income expressed in nominal dollars to one that is
child anthropometry, a few studies reported an eect on at expressed in terms of international dollars, which have the same estimated purchasing power as a dollar in the
least one indicator,5357 but eects were generally small. USA, accounting for local prices. The size of the circles represents the estimated population of stunted children
Although meta-analysis might not be the method of choice aged 05 years, in about 2005, on the basis of multiplication of stunting prevalence by UN estimates of the
population of children aged 05 years. Data are sourced principally from the Demographic and Health Surveys,18
for synthesis of evidence from such diverse programmes, with observations for some countries sourced from WHO.19 GDP=gross domestic product. BGD=Bangladesh.
the results of a four-study meta-analysis52 showed no CIV=Cte dIvoire. DOM=Dominican Republic. DRC=Democratic Republic of the Congo. ETH=Ethiopia.
overall eect of targeted agricultural programmes on IDN=Indonesia. IRQ=Iraq. MDG=Madagascar. MMR=Myanmar (Burma). KEN=Kenya. NGA=Nigeria. PAK=Pakistan.
underweight, wasting, or stunting. Another four-study PHN=Philippines. SDN=Sudan. VTN=Vietnam.
meta-analysis51 for vitamin A status, however, reports a
small overall dierence in serum retinol between
intervention and control areas (008 mol/L); a cluster-
80
randomised eectiveness assessment of a biofortified Egypt
orange sweet potato intervention in Uganda also showed a
95 percentage point reduction in the prevalence of low Jordan Bosnia-Herzegovina
Women overweight prevalence, BMI>25 (%)
References are provided in appendix pp 56. *Approved for release by National Governments after intensive multi-location testing for agronomic traits and micronutrient
performance. Completed though not necessarily reported.
Table: Release schedule for biofortified crops and status of related nutrition studies
Costs will dier between the two approaches as will the 6000
distribution of benefits for dierent outcomes; as such, no
5000
method dominates in all situations. *
*
4000
In-kind household food distributions
the energy gap in the targeted population (panel 5,99,100 Figure 3: Estimated eect of Mexicos Programa de Apoyo Alimentario
figure 3). In addition to general family rations, food programme on annual weight gain in women, by initial BMI
distribution programmes often provide micronutrient- BMI=body-mass index. *p<005. p<001.
fortified foods (eg, corn-soy or wheat-soy blend) to
mothers and young children. In Haiti, distribution of family rations or cash. Disasters, particularly sudden
such food rations to all mothers and children within the onset emergencies such as earthquakes and hurricanes,
first 1000 days of life had a greater eect on child growth often disrupt normal market channels, which might
than did targeting of underweight children younger than dampen the logistical advantages of cash compared with
5 years.101 In view of the severity of food insecurity in this food transfers. Although food aid deliveries overall
population, no unintended eects on overweight or declined from 15 million metric tonnes (t) in 1999 to
obesity were identified. Complementation of this pro- 41 million t in 2011, emergency deliveries have remained
gramme with the distribution of iron-fortified micro- almost constant; they now account for more than 67% of
nutrient powders reduced anaemia prevalence by half in total food aid.103 Even when targeted towards overall
as little as 2 months.102 household subsistence, aid during disasters can prevent
major deteriorations in child undernutrition.104,105 Age-
Transfer programmes in emergencies based targeting of fortified foods can help prevent under-
Transfer programmes in emergencies also usually nutrition and complement eorts to tackle cases of severe
combine nutritionally enhanced complementary foods for acute malnutrition with specially formulated products.
pregnant and lactating women and their young child with The nutritional eects of emergency deliveries can be
enhanced by inclusion of lipid-based nutrient supplements transfers on child cognitive development, which were
in the package of assistance to families.106108 mediated by increased preschool participation, improved
diets, and reduced anaemia (Gilligan D, Roy S, Inter-
Early child development national Food Policy Research Institute, personal
Stunting and impaired cognitive development share communication).
several of the same risk factors, including deficiencies in Other trialsin India, Pakistan, and Bangladesh116118
protein, energy, and some micronutrients, intrauterine although successful at improving child development or
growth retardation, and social and economic conditions, nutrition outcomes, or both, failed to show additive or
such as maternal depression and poverty.109 Some of the synergistic eects between nutrition and stimulation
key phases of brain growth and development also interventions. In India,116 beneficiaries of the Integrated
encompass the first 1000 days of life, the period of peak Child Development Services (ICDS) programme were
susceptibility to nutritional insults. Therefore, some key allocated to groups receiving breastfeeding and comple-
interventions can protect children from both nutritional mentary feeding counselling, or this package plus
and developmental risks; these include core maternal responsive feeding and psychosocial stimulation skills.
and child nutrition interventions, psychosocial stimu- Compared with ICDS only, both intervention packages
lation and responsive parenting, and interventions to improved child dietary intake and haemoglobin and
alleviate poverty, food insecurity, maternal depression, reduced morbidity, but only the nutrition intervention
and gender inequity.21 increased length gain, and only the full package including
Evidence of the eect of early child development inter- stimulation benefited development outcomes.116 A factorial
ventions, with or without a nutrition component, on child design trial in Pakistan117 showed no evidence of additive
development outcomes has been extensively reviewed in or synergistic eects of a nutrition (counselling and
two previous series in The Lancet.20,21 We focus on evidence micronutrient powders) and stimulation intervention
of how child stimulation and nutrition interventions can (monthly group meetings and home visits for children
have complementary eects on nutrition outcomes aged 024 months) on child development or nutrition
(appendix pp 79). The most comprehensive, long-term outcomes. Preliminary results show eects of all three
study110 of interventions that provided both child stimu- intervention packages on developmental scores compared
lation and food supplementation to stunted children aged with control, with larger eect sizes among the two
924 months in Jamaica showed an additive eect of the stimulation groups. In Bangladesh,118 psychosocial
two interventions on cognitive development, but not on stimulation with or without food supplements among
growth. At adolescence, the additive eects on cognition severely underweight children aged 624 months on
were not sustained, but the group having received discharge from hospital had an eect on mental
stimulation had long-term benefits, which ranged from development and a small eect on WAZ, but no additive
improved development outcomes to educational attain- or synergistic eects were noted between the two
ment and social behaviour.111,112 In Bangladesh, addition of interventions. One intervention that closely ties feeding
stimulation and home visits to standard nutrition and practices with child stimulation is responsive feeding.
health care for severely malnourished children improved Few studies of this approach, however, have been designed
development outcomes and weight-for-age Z score to distinguish messaging on complementary feeding
(WAZ).113 Another trial in Bangladesh, which added from those on psychosocial care114 and few so far have
responsive parenting (including feeding) to an informal shown a clear association with nutritional outcomes.119
nutrition and child development education programme, Reduction of maternal depression is another way to
showed benefits on several feeding and parenting address risk factors common to both nutrition and child
behaviours, child self-feeding, and development out- development.120,121 Eorts are being made to link basic
comes; addition of iron-fortified micronutrient powders to health services with a wide range of social support for
the intervention improved weight gain and WAZ but had women. Ecacy trials.122,123 show that the benefits might
no additional eect on development outcomes.114 A zinc accrue to both mothers and their newborn babies,
supplementation and responsive stimulation intervention justifying ongoing eorts to bring these initiatives to scale.
in underweight children in Jamaica showed synergistic
eects on child development between the two interven- Schooling
tions: greater benefits on development outcomes were Although children are beyond the crucial 1000 days
identified in the zinc and stimulation group, compared window when they enter school, their schooling
with little or no eect in the groups receiving either.115 This experience might be a strong determinant of the nutrition
synergistic eect, however, was not noted for morbidity of the next generation. Parental schooling has been
(reduced only in zinc group) or growth (no eect in either consistently associated with child nutritional status, with
group). A randomised controlled trial of cash transfers to maternal education often, but not always, having a larger
households linked to preschool enrolment in Uganda explanatory power than paternal education, controlling
provides an example of joint benefits on cognition and for income and schooling choices.124,125 The positive global
nutrition; findings showed significant eects of cash trends in schooling are, therefore, encouraging for
nutrition. Data from developing countries show an health and nutrition curriculum in developing countries
average increase in years of schooling from 260 to 762 on undernutrition or health knowledge, let alone on
for boys and from 150 to 664 for girls between 1950 parenting skills decades later, are absent.
and 2010.126 The girl to boy ratio shows a substantial Schools are also suitable venues to introduce pro-
improvement over this period, from 577% to 859%. grammes to combat obesity. Such programmes can focus
Enrolment data show parity in girls primary schooling in on healthy diets and promotion of physical activity. A
most countries;127 moreover, in many countries more systematic review133 of 22 studies in low-income and
girls than boys are now in secondary school. Still, only middle-income countries noted that 82% of such
about a fifth of adolescent girls in sub-Saharan Africa programmes had a favourable eect on physical activity,
and two fifths of girls in south Asia are enrolled in diet, or both.
secondary education.128
We assessed the level of parental education necessary Discussion
for a meaningful reduction in child undernutrition by In 2008, The Lancet Maternal and Child Undernutrition
analysing 19 datasets from the Demographic and Health Series included conditional cash transfer programmes
Survey (collected since 1999) and derived estimates of the and dietary diversification approaches as general
risk of child stunting associated with maternal and nutrition support strategies, and noted small positive
paternal primary and secondary education, controlling for eects of conditional cash transfers on child anthrop-
household wealth, rural versus urban residence, and child ometry in three Latin American countries, and an
age and sex.The analysis showed that the risk of stunting absence of a statistically significant of eect of dietary
is significantly lower among mothers with at least some diversification strategies on child nutrition outcomes.134
primary schooling (odds ratio [OR] 089, 95% CI In the present series, we discuss evidence regarding the
085093), and even lower (p<0001) among mothers nutritional contribution of programmes in four sectors
with some secondary schooling (075, 071079). Paternal and the potential for enhancing their nutrition-sensitivity.
education at both the primary and secondary levels also Although the concept of nutrition sensitivity is not new,
reduced the risk of stunting although the respective ORs investments in development and implementation of
(096, 093101; and 085, 081089) are smaller than nutrition-sensitive programmes have intensified in the
for maternal schooling. Despite this overall association, past few years, prompted by the 2008 series, and
there is appreciable heterogeneity in eect sizes for both spearheaded by the Scaling Up Nutrition movement.9 It For more on the Scaling Up
maternal and paternal education in individual countries, is important to recognise, when interpreting the results Nutrition movement see http://
scalingupnutrition.org
probably indicative of dierences in both quality of of our review, that most of the programmes included
education and quality of data. were retrofitted and tagged as nutrition sensitive without
Schooling directly increases individual earnings and having been originally designed as such.
national income and, through these pathways, can aect Targeted agricultural programmes have an important
nutrition in the long term. Thus, programmes to role in supporting livelihoods, improving household
increase schooling via the supply of inputs or through food security and healthy diets, and in fostering
fee waivers or cash transfers can be expected to reduce womens empowerment. Yet, our review shows
the risk of undernutrition for the next generation. There inconclusive evidence of eects on child nutritional
is a lack of clarity, however, about which aspects of status, with the possible exception of benefits on
schooling, beyond the income eect, benefit nutrition. vitamin A intake and, to a lesser extent, vitamin A
At least five overlapping pathways have been suggested, status. These findings are probably the result of a
but not formally tested. Schooling might: 1) transmit combination of factors, including: weaknesses in
information about health and nutrition directly; 2) teach programme design and implementation (especially the
numeracy and literacy, thereby assisting caregivers in nutrition, behaviour-change communication, and
acquiring information and possibly nutrition know- health components);60 inclusion of households with
ledge;129,130 3) expose individuals to new environments, children outside of the 1000 days window with little
making them receptive to modern medicine; 4) impart potential to benefit in linear growth; and the fact that
self-confidence, which enhances womens roles in other pressing constraints to nutritionsuch as infec-
decision making, and their interactions with health-care tious diseases, helminths, and environmental enter-
professionals; and 5) provide women with the opathy associated with scarcity of access to appropriate
opportunity to form social networks, which can be of water, sanitation, and hygienemight not be addressed
particular importance in isolated rural areas. The by the programmes. Additionally, the assessments of
question remains as to whether schooling could do even most of the programmes we reviewed had crucial
more to directly aect nutrition, both in the short term weaknesses such as an absence of valid comparison
for school children and in the long term as they and control groups, a possibly too-short duration of
transition into their parental role. Although nutrition intervention, small sample sizes, the inclusion of the
modules are available in some school health education wrong age group in eectiveness assessments, and the
programmes,131,132 assessments of the eect of a school failure to control for potential confounding factors in
Our review shows the potential of programmes in the actions from several sectors, or co-locate programmes
four sectors reviewed to improve the lives of poor managed by dierent sectors so that they reach and
households and individuals, both in the short-term and the saturate the same communities, households, and indi-
long-term. It also shows, however, that more needs to be viduals. In view of the complexity of integration,33,137,138
done to increase the nutrition sensitivity of programmes especially across many sectors, it is important to carefully
so that their potential to deliver on maternal and child assess whether investments should focus on joint
nutrition outcomes is unleashed. The nutrition sensitivity planning, implementation, monitoring, and assessment,
of promising programmes can be enhanced in several or on eective programme co-location (panel 6).
ways. First, targeting on the basis of nutritional
vulnerability (eg, age, physiological status) in addition to Conclusions
geographic targeting on the basis of poverty, food Nutrition-sensitive programmes hold great promise for
insecurity, or location can help reach households and supporting nutrition improvements and boosting the
individuals most likely to benefit from the programme. scale, coverage, and benefits of nutrition-specific actions.
Alternatively, targeting of nutritionally vulnerable indi- New incentives are needed to support innovations in
viduals could be used as a second level of targeting for nutrition-sensitive programmes and unleash their
subgroups of programme beneficiaries who meet pre- potential to tackle nutrition while also achieving their own
established criteria. For example, targeting agricultural goals. New nutrition-sensitive agriculture44 and social
programmes to households with pregnant or lactating safety net programme designs, methods, and packages of
women or children younger than 2 years might be neither interventions are being tested and are strengthening links
logistically feasible nor optimum for community develop- with health services. Rigorous impact evaluations, many
ment; however, geographic community-level targeting of which are based on strong programme-theory and
could be used as a first targeting criteria, with a second impact pathway analysis, are addressing key weaknesses
level focusing on reaching mothers and young children encountered in previous evaluations and are assessing
with a specific package of preventive nutrition and health impacts on a range of nutrition and child development
interventions. Another key target group for nutrition- outcomes and several household and gender outcomes
sensitive programmes is adolescent girls; conditions or along the impact pathway. Evidence generated by these
other incentives can be used to keep girls in school, help enhanced programmes and assessments in the next
delay first pregnancy, address HIV risk factors,135 and 510 years will be of crucial importance to inform future
improve adolescent girls' nutrition knowledge and micro- investments in agriculture and social safety net pro-
nutrient status to prepare them for motherhood. grammes to improve nutrition.
Second, evidence shows that nutrition improvements The potential benefits of integration of early child
are not automatic even with programmes that are success- development and nutrition programming include cost
ful at reducing poverty, food insecurity, and sex inequalities. savings and gains in both child development and nutrition
To reach their full potential, programmes such as those outcomes. Leveraging health, agriculture, or social safety
reviewed need careful identification of nutrition goals and net platforms for joint early child development and
appropriate design and eective implementation of nutrition programming during the first 1000 days of life
interventions to achieve them. A third way to enhance the would help focus on the crucial period of peak vulnerability
nutrition sensitivity of programmes is to engage women for both nutrition and development. Current work is
and include interventions to protect and promote their exploring such approaches. Benefits from psychosocial
nutritional wellbeing, physical and mental health, social interventions on cognition, however, extend well beyond
status, decision making, and their overall empowerment the first 2 years, and therefore, continued child develop-
and ability to manage their time, resources, and assets. A ment support is required throughout the entire preschool
fourth promising, yet underused approach to enhance the period. Early child development programmes, possibly
nutrition sensitivity of programmes is to use them as linked to conditions in transfer programmes or delivered
delivery platforms for various nutrition-specific inter- through preschool or community settings, could oer
ventions. Nutrition behaviour-change communications, psychosocial stimulation and parenting interventions,
which are incorporated in several agriculture, social safety while also providing relevant nutrition interventions for
nets, early child development, and school health pro- children 2 years and older, focusing on micronutrients,
grammes are one example of such use. Other opportunities healthy diets, and obesity prevention. With improved
include addition of the distribution of micronutrient- guidance and curricula for nutrition education in schools,
fortified products to nutritionally vulnerable adolescent a new emphasis on using schools to improve nutrition
girls, mothers, and young children,101,136 or of preventive knowledge and practices and preparing school children
health inputs to agriculture, social safety net, early child for their future parenting roles should also emerge.
development, or school programmes. The immense potential of programmes addressing
Finally, a crucial question that remains to be addressed the underlying determinants of undernutrition to
when designing nutrition-sensitive programmes is the complement and enhance the eectiveness of nutrition-
degree to which programmes should indeed integrate specific interventions is real, but is yet to be unleashed.
Investments in nutrition-sensitive programmes can 2 IFPRI. 2011 Global food policy report. Washington, DC:
have a pivotal role in prevention of the excess stunting, International Food Policy Research Institute, 2012.
3 Nelson GC, Rosegrant MW, Koo J, et al. Climate change. Impact on
wasting, and impaired child development that the scale- agriculture and costs of adaptation. Washington, DC: International
up of nutrition-specific interventions cannot resolve on Food Policy Research Institute, 2009.
its own. 4 Webb P. Medium-to long-run implications of high food prices for
global nutrition. J Nutr 2010; 140: S14347.
Contributors 5 FAO, WFP, IFAD. The state of food insecurity in the world, 2012.
MTR and HA conceptualised the report, reviewed the literature, and Economic growth is necessary but not sucient to accelerate
wrote all drafts of the report. reduction of hunger and malnutrition. Rome: Food and Agriculture
Maternal and Child Nutrition Study Group Organization, 2012.
Robert E Black (Johns Hopkins Bloomberg School of Public Health, 6 World Bank. Global economic prospects: commodities at the
crossroads. Washington, DC: World Bank, 2009.
USA), Harold Alderman (International Food Policy Research Institute,
USA), Zulfiqar A Bhutta (Aga Khan University, Pakistan), 7 OECD-FAO. OECD-FAO agricultural outlook, 2012. DOI:10.1787/
agr_outlook-2012-en.
Stuart Gillespie (International Food Policy Research Institute, USA),
Lawrence Haddad (Institute of Development Studies, UK), 8 Pinstrup-Andersen P, Watson DDI. Food policy for developing
countries. The role of government in global, national and local food
Susan Horton (University of Waterloo, Canada), Anna Lartey
systems. Ithaca: Cornell University Press, 2011.
(University of Ghana, Ghana), Venkatesh Mannar (The Micronutrient
9 Scaling Up Nutrition. Progress report from countries and their
Initiative, Canada), Marie Ruel (International Food Policy Research
partners in the Movement to Scale Up Nutrition. New York: UN, 2011.
Institute, USA), Cesar Victora (Universidad de Federal de Pelotas,
10 Shekar M, Ruel-Bergeron J, Herforth A. Module A. Introduction.
Brazil), Susan Walker (The University of the West Indies, Jamaica),
In: Improving nutrition through multisectoral approaches.
Patrick Webb (Tufts University, USA). Washington, DC, International Bank for Reconstruction and
Series Advisory Committee Development, International Development Association of The
Marc Van Ameringen (Gain Health Organization, Switzerland), World Bank, 2013.
Mandana Arabi (New York Academy of Sciences, USA), Shawn Baker 11 Black RE, Victora CG, Walker SP, and the Maternal and Child
(Helen Keller International, USA), Martin Bloem (United Nations World Nutrition Study Group. Maternal and child undernutrition and
Food Programme, Italy), Francesco Branca (WHO, Switzerland), overweight in low-income and middle-income countries. Lancet
2013; published online June 6. http://dx.doi.org/10.1016/S0140-
Leslie Elder (The World Bank, USA), Erin McLean (Canadian
6736(13)60937-X.
International Development Agency, Canada), Carlos Monteiro
12 Grosh M, Del Ninno C, Tesliuc E, Ouerghi A. For protection and
(University of So Paulo, Brazil), Robert Mwadime (Makerere School of
promotion: the design and implementation of eective safety nets.
Public Health, Uganda), Ellen Piwoz (Bill & Melinda Gates Foundation, Washington, DC: World Bank, 2008.
USA), Werner Schultink (UNICEF, USA), Lucy Sullivan (1000 Days,
13 Maxwell D, Webb P, Coates J, Wirth J. Rethinking food security in
USA), Anna Taylor (Department for International Development, UK), humanitarian response. Food Policy 2010; 35: 9197.
Derek Yach (The Vitality Group, USA). The Advisory Committee
14 Connolly MA, Gayer M, Ryan MJ, Salama P, Spiegel P,
provided advice in a meeting with Series Coordinators for each paper at Heymann DL. Communicable diseases in complex emergencies:
the beginning of the process to prepare the Series and in a meeting to impact and challenges. Lancet 2004; 364: 197483.
review and critique the draft reports. 15 Alderman H. Safety nets can help address the risks to nutrition
Other contributors from increasing climate variability. J Nutr 2010; 140: S14852.
Derek Headey, research fellow at IFPRI, did the analysis and prepared 16 Haddad L, Alderman H, Appleton S, Song L, Yohannes Y.
the panel on the relation between income and nutrition. Joe Green, Reducing child malnutrition: how far does income growth take us?
independent consultant, did the analysis of the role of education. World Bank Econ Rev 2003; 17: 10731.
Mara van den Bold, research analyst at IFPRI, reviewed the literature on 17 Alderman H, Linnemayr S. Anemia in low-income countries is
womens empowerment, provided background text, and wrote the panel unlikely to be addressed by economic development without
additional programs. Food Nutr Bull 2009; 30: 26569.
on womens empowerment and nutrition. Jef Leroy, research fellow at
IFPRI, did the analysis and wrote the panel on the cash transfer 18 Demographic Health Surveys. Measure DHS Stat-Compiler. http://
www.statcompiler.com/ (accessed Nov 22, 2012).
programme in Mexico. Erick Boy, senior research fellow at IFPRI,
provided background text and other material for the section on 19 WHO. Global database on child Growth and malnutrition. World
Health Organization, 2012. http://www.who.int/nutgrowthdb/
biofortification. Sivan Yosef, programme manager at IFPRI, provided
database/en/ (accessed Dec 12, 2012).
research and editing assistance during all steps of the process.
20 Engle PL, Black MM, Behrman JR, Cabral de Mello M, Gertler PJ,
Conflicts of interest Kapiriri L. Strategies to avoid the loss of developmental potential in
REB serves on the Boards of the Micronutrient Initiative, Vitamin more than 200 million children in the developing world. Lancet
Angels, the Child Health and Nutrition Research Initiative, and the 2007; 369: 22942.
Nestle Creating Shared Value Advisory Committee. VM serves on the 21 Engle PL, Fernald LCH, Alderman H, et al. Strategies for reducing
Nestle Creating Shared Value Advisory Committee. All other authors inequalities and improving developmental outcomes for young
declare that they have no conflicts of interest. As corresponding author, children in low-income and middle-income countries. Lancet 2011;
Marie Ruel states that she has full access to all data and final 378: 133953.
responsibility for the decision to submit for publication. 22 Bhutta ZA, Das JK, Rizvi A, et al, The Lancet Nutrition Interventions
Review Group, and the Maternal and Child Nutrition Study Group.
Acknowledgments Evidence-based interventions for improvement of maternal and
Funding for the preparation of the Series was provided to the Johns child nutrition: what can be done and at what cost? Lancet 2013;
Hopkins Bloomberg School of Public Health through a grant from the published online June 6. http://dx.doi.org/10.1016/S0140-
Bill & Melinda Gates Foundation. The sponsor has no role in the 6736(13)60996-4.
analysis and interpretation of the evidence or in writing the paper and 23 Almond D, Currie J. Killing me softly. J Econ Perspect 2011;
the decision to submit for publication. We thank John Hoddinott, senior 25: 125.
research fellow at IFPRI, who reviewed an early version of the paper, and 24 Hoddinott J, Maluccio J, Behrman JR, et al. The consequences of
to four anonymous reviewers for their insightful comments. early childhood growth failure over the life course. IFPRI
Discussion Paper 01073. Washington, DC: International Food Policy
References Research Institute, 2011.
1 Pinstrup-Andersen P, ed. The African food system and its 25 Alderman H. The economic cost of a poor start to life.
interactions with human health and nutrition. Ithaca: Cornell J Dev Origins Health Dis 2010; 1: 1925.
University Press, 2010.
S ECU R I T Y
IS YOUR
FOOD T
CONTAMINATED?
H
By Mark Fischetti C
G
E
IVEN THE BILLIONS OF FOOD ITEMS THAT ARE PACKAGED, PURCHASED AND CONSUMED
every day in the U.S., let alone the world, it is remarkable how few of them O
are contaminated. Yet since the terrorist attacks of September 11, 2001, F
food defense experts have grown increasingly worried that extremists
might try to poison the food supply, either to kill people or to cripple the F
O
of edible products is becoming ever more centralized, speeding the spread
D
of contaminantsnatural ones or those introduced purposelyfrom farms or processing
plants to dinner tables everywhere. Mounting imports pose yet another rising risk.
Can we protect food from tainting? And TIGHTEN SECURITY tions are unlikely to work because the
if toxic substances or pathogens do slip PREVENTING A TERRORIST or a disgruntled chain differs from commodity to com-
into the supply chain, can they be quickly employee from contaminating milk, juice, modity, says David Hennessy, an eco-
detected to limit their harm to consum- produce, meat or any other type of comes- nomics professor at Iowa State Universi-
ers? Tighter production procedures can tible is a daunting problem. The food sup- tys Center for Agricultural and Rural De-
go a long way toward protecting the pub- ply chain comprises a maze of steps, and velopment. Protecting dairy products is
lic, and if they fail, smarter monitoring virtually every one of them presents an different from protecting apple juice,
technologies can at least limit injury. opportunity for tampering. Blanket solu- which is different from protecting beef.
IN BRIEF
Terrorist plots, greater imports and more Safety procedures adapted from mili- New technologies such as micro uidic Streamlining government regulations
centralized food production are raising tary methods of vulnerability analysis chips, RFID tags and edible markers can could help, but closer oversight by brand-
the chances that food will be tainted by can tighten the physical security of food- detect contamination, help trace it back name sellers of their suppliers will do
natural or man-made contaminants. production facilities. to its source and speed recalls. more to ensure safer fare.
COMING TECHNOLOGIES
Microfluidic detectorsBotulinum bacteria produce the most poisonous toxin known. They and
similar agents, such as Escherichia coli could be detected during ood rocessing by micro uidic
chi s sel contained diagnostic labs the si e o a finger esearch grou s ha e cra ted chi s
in which channels are lined with antibodies that can detect botulism or E. coli during milk
production. Samples could be tested either before or after the milk is piped into tanker trucks
that leave the dairy and again before or after it is pasteurized at a production plant. Similar
chi s could detect other to ic substances at uid rocessing lants that roduce a le
juice, soup, baby formula and other liquid foods.
RFID tags allets or cases o a ew select oods now s ort radio re uency identification
(RFID) tags that, when read by a scanner, indicate which farm or processing plant the batch
came from. Future tokens that are smaller, smarter and cheaper could adorn individual packages
and log every facility they had passed through and when. Researchers have been develop-
ing tags that could be read through uid (traditional designs cannot) and thus
could be embedded inside the wall of sour cream or yogurt containers. Active
tags could record the temperatures a package had been exposed to, but they
are more expensive to produce.
Edible tagsManufacturers often combine crops from many growers, such as spinach leaves, into a retail
Orchard
ac age so tags a ed to bags might not hel in estigators trac contamination bac to a s ecific source 132
Edible inks can now be used to print markers that indicate site of origin directly onto a package of foodor
60 microns
in some cases, onto the food itself.
ing these agents, tracing them back to dio-frequency identification (RFID) tags the tags on their items to see whether
the spot of introduction, and tracking are furthest along, in part because the they received a product that was later
which grocery stores and restaurants Defense Department and Walmart have identified as suspicious.
ended up with tainted products are para- required their main suppliers to attach As RFID tags get smaller and cheap-
mount. Putting such systems in place is the tokens to pallets or cases of food- er, they may be placed directly on indi-
just as important as prevention, Schnei- stuffs. The Metro AG supermarket chain vidual itemson every bag of spinach,
der says. in Germany has done the same. The ulti- jar of peanut butter, container of shrimp
Here new technology does play a ma- mate intent is for automated readers to and sack of dog food. That way if a re-
jor role, with various sensors applied at scan the tags at each step along the sup- call is issued, the items can be found as
different points along the chain. You ply chainfrom farm, orchard, ranch or they run past a scanner at the checkout
cant expect one technology to counter all processor, through packaging, shipping counter, says Jean-Pierre mond, of the
the possible taintings for a given food, and wholesaleand to report each items electro-optical systems laboratory at the
notes Ken Lee, former director of Ohio location to a central registry. That way if Georgia Tech Research Institute.
State Universitys department of food sci- a problem surfaces, investigators can Universities and companies are devel-
ence and technology. quickly determine where the batch origi- oping all kinds of other tags, some that
A variety of hardware is being devel- nated and which stores or facilities are very inexpensive and others that cost
oped [see box above], although little has might have received goods from that more but supply extensive information.
been deployed commercially thus far. Ra- batch and when. Retailers can also read Some tokens, for example, can sense
61
POLICY SOLUTIONS
N
22 ,55 5
Canada
C
Form 151
er U.S.S.R.: 4 19,
U.S. pe
:
E 31
:
ro
Japa Eu
E. As n: 827 394
rica:
ia: 1,10 exi N. Af
M 18,0
China: 6 6 c
0 7 o: Sub
,679 -Sa M. East:
hara 940
S. E. Asia: 12,0 n Afri
O
13 ca: 1,837
Am
C.
,16
3 erica: 4,970 S. Asia: 4,793
F
s: 5
land Caribbean: 562
Paci fic Is
Australia,
S. A
F me
rica
: 14, 5 87
O
D
U.S. FOOD IMPORTS, 2013
Millions of U.S. dollars Total U.S. Imports: $114,025 million
whether food has been exposed to warm tuce. For some growers, mond notes, that nies via a secure Internet site hosted by
temperatures and thus might be more could equal the profit they are going to VeriSign, the data security firm. The com-
likely to harbor Escherichia coli or Salmo- make on that case. To be embraced widely, pilation allowed the participants to more
nella bacteria. Other tags could track how therefore, he says tags will have to provide quickly resolve order discrepancies, to log
long items spent in transit from node to additional value to suppliers or buyers. how long food sat idle and to reveal ways
node in the supply chain, which could in- The University of Florida conducted a to raise shipping efficiency.
dicate unusually long delays that might project with Publix Super Markets and
SOURCE: U.S. DEPARTMENT OF AGRICULTURE
raise suspicion about tampering. So-called produce suppliers in Florida and Califor- CONTROL SUPPLIERS
active packaging could detect contamina- nia to assess such possibilities. In initial COSTS WILL NOT DROP until new techno-
tion directly and warn consumers not to trials, tags tracked crates and pallets that logies are deployed widely. But food de-
eat the product they are holding. were shipped from the growers to several fense analysts say adoption is unlikely to
The big impediment for any marker, of of Publixs distribution centers. Informa- occur until clear, streamlined regulations
course, is the price. Today it costs around tion gleaned from scanning tags at vari- are enacted. And that prospect, in turn,
15cents to put an RFID tag on a case of let- ous points was available to all the compa- seems likely to be remote until the high-
S A F E T Y AT H O M E
Microwave Refrigerator I
An infrared sensor gauges internal A built-in reader scans RFID tags on food and checks E
food temperature and compares it for recalls over a wireless Internet connection. (A
with safety guidelines, indicating homeowner could hold nonrefrigerator items under N
when the proper value has been it, too.) The reader also notes expiration dates written C
reached. Instead of entering a into the tags and tracks when containers such as milk
cooking time, a user enters the cartons are removed and put back, to see if they have E
food type or target temperature. been out for too long and therefore might be spoiled.
A red light warns of trouble.
F
est levels of government are reformed. proposed unifying responsibility for pre- the latest security procedures and detec-
There are more than a dozen differ- vention of food contamination and re- tion, track and trace technologiesand F
ent federal agencies that oversee some as- sponse to outbreaks of foodborne illness drop those that do not. O
pect of food safety, Lee points out, noting in a new Food Safety Administration. Brand-name providers, after all, have
O
that simple coordination among them is More effective government is good, the most to lose. If a natural or man-
difficult enough, and efficient approval of but experts say food suppliers also need made toxin is found in, say, a bag of Dole D
sensible requirements is even harder to to improve vigilance. The strongest tool spinach or a container of Safeway milk,
come by. The FDA regulates pizza with for stopping intentional contamination consumers will shun that particular la-
cheese on it, but the USDA regulates pizza is supply-chain verification, says Shaun bel. If a brand-name company wants to
if it has meat on it, quips Jacqueline Kennedy, director of the National Center protect its products, Kennedy says, it
Fletcher, a professor of entomology and for Food Protection and Defense. Each should validate every participant in the
plant pathology at Oklahoma State Uni- brand-name provider such as Dole and chain, all the way back to the farm.
versity. The requirements for organic each grocery store conglomerate such as
farmers are different from those for non- Safeway must verify that every company Mark Fischetti is a senior editor
organic farmers. involved in its supply chain implements at Scientific American.
Spurred by recalls and outbreaks, Con-
gress passed the Food Safety Moderniza-
tion Act in 2011, which directed the FDA to M O R E TO E X P L O R E
refocus its efforts on prevention of con- Terrorist Threats to Food. World Health Organization, May 2008.
tamination rather than recalls of bad food, The Case of Botulinum Toxin in Milk: Experimental Data. Oliver G. Weingart et al. in Applied and Environmental
strengthen inspections of both domesti- Microbiology, Vol. 76, No. 10, pages 32933300; May 2010.
cally produced and imported foods [see Emerging Global Food System Risks and Potential Solutions. Shaun Kennedy in Improving Import Food Safety.
box on opposite page], and make better Edited by Wayne Ellefson, Lorna Zach and Darryl Sullivan. John Wiley & Sons, 2013.
National Center for Food Protection and Defense: www.ncfpd.umn.edu
use of risk analysis in setting its priorities.
In February, the Obama administration s c i e n t i f i c a m e r i c a n . c o m /m a g a z i n e /s a
63
Summary
Lancet Infect Dis 2014; Background In May, 2013, an outbreak of symptomatic hepatitis A virus infections occurred in the USA. Federal, state,
14: 97681 and local public health ocials investigated the cause of the outbreak and instituted actions to control its spread. We
Published Online investigated the source of the outbreak and assessed the public health measures used.
September 4, 2014
http://dx.doi.org/10.1016/
S1473-3099(14)70883-7 Methods We interviewed patients, obtained their shopping information, and did genetic analysis of hepatitis A virus
See Comment page 907
recovered from patients serum and stool samples. We tested products for the virus and traced supply chains.
Division of Viral Hepatitis,
National Center for HIV, Findings Of 165 patients identified from ten states, 69 (42%) were admitted to hospital, two developed fulminant hepatitis,
Hepatitis, STD and TB and one needed a liver transplant; none died. Illness onset occurred from March 31 to Aug 12, 2013. The median age of
Prevention (M G Collier MD, patients was 47 years (IQR 3558) and 91 (55%) were women. 153 patients (93%) reported consuming product B from
Y E Khudyakov PhD,
N P Nelson MD, A Suryaprasad MD,
retailer A. 40 patients (24%) had product B in their freezers, and 113 (68%) bought it according to data from retailer A.
J Drobeniuc MD, Hepatitis A virus genotype IB, uncommon in the Americas, was recovered from specimens from 117 people with
S D Holmberg MD, F Xu MD), hepatitis A virus illness. Pomegranate arils that were imported from Turkeywhere genotype IB is commonwere
Epidemic Intelligence Service identified in product B. No hepatitis A virus was detected in product B.
(E Epson MD), Division of
Foodborne, Waterborne, and
Environmental Diseases, Interpretation Imported frozen pomegranate arils were identified as the vehicle early in the investigation by combining
National Center for Emerging epidemiologywith data from several sourcesgenetic analysis of patient samples, and product tracing. Product B was
and Zoonotic Infectious Diseases removed from store shelves, the public were warned not to eat product B, product recalls took place, and postexposure
(S E Lance DVM, M E Wise PhD),
Centers for Disease Control and
prophylaxis with both hepatitis A virus vaccine and immunoglobulin was provided. Our findings show that modern
Prevention, Atlanta, GA, USA; public health actions can help rapidly detect and control hepatitis A virus illness caused by imported food. Our findings
New Mexico Department of show that postexposure prophylaxis can successfully prevent hepatitis A illness when a specific product is identified.
Health, Santa Fe, NM, USA
Imported food products combined with waning immunity in some adult populations might make this type of intervention
(D Selvage MHS,
M Adams-Cameron MPH); necessary in the future.
Colorado Department of Public
Health and Environment, Funding US Centers for Disease Control and Prevention, US Food and Drug Administration, and US state and local
Denver, CO, USA (E Epson,
public health departments.
A Cronquist RN, R H Jervis MPH);
California Department of Public
Health, Sacramento, CA, USA Introduction retailer A and who had hepatitis A illness with symptom
(K Lamba MPH, A C Kimura MD); Infection with the hepatitis A virus causes an acute viral onset dates within 1 week of each other. The Department
Nevada Division of Public and
illness characterised by fever, abdominal pain, high contacted states in the same distribution area for retailer A,
Behavioral Health, Carson City,
NV, USA (R Sowadsky MSPH); concentrations of aminotransferase enzymes, and and the Colorado Department of Public Health and
Arizona Department of Health jaundice. Hepatitis A vaccine or immunoglobulin given Environment identified four additional retailer A shoppers
Services, Phoenix, AZ, USA within 2 weeks of exposure can prevent disease.1 In the with hepatitis A illness. None of these patients had typical
(R Hassan MSPH); Hawaii State
USA, hepatitis A vaccination has been recommended for risk factors for hepatitis A (eg, international travel, injection
Department of Health, Honolulu,
HI, USA (S Y Park MD); Texas all children aged 1223 months since 2006,2 and drug use). Initial interviews using a hypothesis generating
Department of State Health consequently the number of acute cases has greatly questionnaire, which asks about all possible food exposures,
Services, Austin, TX, USA fallen.3 Hepatitis A virus is transmitted by person-to- identified some similar eating patterns, including fruits
(E Garza MPH) Food and Drug
Administration, Silver Spring,
person contact or ingestion of contaminated food or and vegetables consumed in smoothies. Four patients from
MD, USA (A J Elliott MS, water, and is primarily a human pathogen. International two states had had product Ba frozen mix of cherries,
D S Rotstein DVM, J Beal MPH, travel was reported in about half of all reported US cases.4 strawberries, raspberries, blueberries, and pomegranate
T Kuntz MS, S E Lance, The most recent common source outbreak of foodborne arilsin their freezers and used it in smoothies. Retailer A
R Dreisch MS); and Oakridge
Institute for Science and
hepatitis A in the USA happened in 2005.5 confirmed that people with hepatitis A purchased product B
Education, Oakridge, TN, USA On May 13, 2013, the New Mexico Department of Health from retailer A. This report describes the investigation of
(A J Elliott) notified the Centers for Disease Control and Prevention the multistate outbreak associated with product B and the
(CDC) of two people in the same city who had shopped at public health actions taken to halt the outbreak.
of product E
4
0
31
14
21
28
12
19
26
16
23
30
14
21
28
g4
1
g1
ril
ly
ay
ne
ne
ch
ril
ay
ly
ne
ril
ly
ay
ne
ne
Au
ril
ay
ly
Ju
Ap
Au
Ju
Ju
Ju
Ju
Ju
Ap
Ap
Ap
M
ar
Ju
Ju
Ju
M
departments about who should receive what postexposure back to their originating firms through investigation of
prophylaxis according to the Advisory Committee for purchasing and invoice documents.
Immunization Practices guidelines;1 however, final Strawberries and blueberries came from multiple
decisions were made locally based on local resources. importers from more than one location, and so were not
considered the likely vehicle. Cherries from the USA and
Results strawberries from Argentina in product B came from a
We identified 165 people from ten states meeting the single supplier but in several dierent lots and so were
confirmed case definition; eight of whom met the unlikely to be the vehicle. Product B cherries and
secondary case definition. Dates of illness onset ranged strawberries were also used in other widely distributed
from March 31, 2013, to Aug 12, 2013 (figure 1). 69 (42%) products made by manufacturer C and no other outbreak-
of 165 patients were admitted to hospital, two (1%) associated cases were reported among consumers of other
developed fulminant hepatitis, and one secondary case manufacturer C products. Frozen pomegranate arils
patient with fulminant hepatitis needed a liver transplant; imported from Turkey by importer D from a single lot
no deaths were reported. Most patients were aged were identified in product B (figure 2). The same lot of
4064 years and most were female (table). Few patients frozen pomegranate arils were also used in product E
were aged 18 years or younger, none of whom were produced by manufacturer F, and one person who had
vaccinated (two children were from one family who consumed product E had an infection of hepatitis A virus
refused postexposure prophylaxis). IB matching the outbreak strain. The first lot to contain
Product B was a mixture of frozen strawberries, red the Turkish arils was produced on Jan 24, 2013, and the
raspberries, blueberries, cherries, and pomegranate arils. bags had a shelf life of 2 years.
The workers processing product B wore gloves, limiting Of 157 patients, 152 (97%) reported consuming
bare hand contact with the product, and no workers product B or had purchased it according to member card
reported symptoms of acute hepatitis. Through lot information. Of five people who did not recall consuming
numbers of product B recovered from four patients product B, three were members of retailer A but did not
freezers, the five ingredients of product B were traced purchase product B, one person did not consume or
purchase product B but consumed product E (which
contained the same arils), and one person did not
Patients (n=165)
consume either product B or E. According to member
Female patients 91 (55%) card information, 113 patients (72%) bought product B,
Age, years and 40 had product B in their freezers (of whom, we
Median (IQR) 47 (3558) could not find purchase records for three). 71 (63%) of
018 11 (7%) 113 patients bought product B more than once between
1839 42 (25%) Jan 24 and May 31, 2013. The person identified in New
4064 95 (58%) Jersey was a secondary case exposed to a primary case in
65 17 (10%) Colorado, the two people from Wisconsin consumed
Ethnic origin product B while living in California before relocating to
White 123 (74%) Wisconsin, and the person from New Hampshire
Hispanic or Latino 13 (8%) traveled to Nevada during the illness incubation period
Asian 6 (4%) but did not recall consuming product B.
Black 2 (1%) Of the 165 patients, 120 had clinical specimens available
Other 3 (2%) for testing, including 119 serum samples and one stool
Unknown 18 (11%) specimen. 117 (98%) tested positive for hepatitis A virus
State of residence genotype IB. Genotype IA virus was recovered from two
Arizona 24 (15%) patients who reported consuming product B, and one
California 80 (48%) patient was hepatitis A virus negative. Phylogenetic analysis
Colorado 29 (18%)
of the outbreak specimens showed that the sequences from
Hawaii 8 (5%)
99 genotype IB specimens (85%) were identical in the 315
New Hampshire 1 (<1%)
bp segment of VP1/P2B (figure 3); the remaining 18
outbreak specimens belonged to minor genotype IB
New Jersey 1 (<1%)
strains. To ensure that other hepatitis A cases related to this
New Mexico 11 (7%)
outbreak were detected, CDC issued a national request
Nevada 6 (4%)
through Epi-X for state and local health departments to
Utah 3 (2%)
submit specimens from people with hepatitis A symptom
Wisconsin 2 (1%)
onset from March 1 to Sept 5, 2013. Specimens from an
Data are n (%) unless otherwise stated. additional 131 people were submitted as a result; of these,
47 (36%) were PCR negative for hepatitis A virus, 82 (62%)
Table: Patient characteristics
were positive for genotype IA, of which 66 belonged to the
Importer U, Chile
Raspberry, multiple lots
Manufacturer C
Importer D, Turkey
Raspberries Strawberries Pomegranates Pomegranate, single lot
Blueberries Cherries
adequate processing methods are dicult to verify, the 5 Bialek SR, George PA, Xia GL, et al. Use of molecular
eectiveness of vaccination for postexposure prophylaxis epidemiology to confirm a multistate outbreak of hepatitis A
caused by consumption of oysters. Clin Infect Dis 2007; 44: 83840.
in people older than 40 years needs to be studied. 6 NNDSS hepatitis A, acute. http://wwwn.cdc.gov/NNDSS/script/
We encountered challenges during this investigation. casedef.aspx?CondYrID=703&DatePub=1/1/2012 (accessed
First, we were unable to confirm pomegranate aril July 23, 2014).
7 Nainan OV, Xia G, Vaughan G, Margolis HS. Diagnosis of hepatitis A
contamination through food testing because of the virus infection: a molecular approach. Clin Microbiol Rev 2006; 19: 6379.
diculty of isolating hepatitis A virus from food.9,29 Second, 8 Felsenstein J. Maximum-likelihood estimation of evolutionary trees
public health surveillance in the USA for hepatitis A is from continuous characters. Am J Hum Genet 1973; 25: 47192.
passive and specimens are not routinely submitted for 9 Food and Drug Administration bacteriological analytical manual
(BAM). Detection of hepatitis A in foods. http://www.fda.gov/
public health analysis. Cases could have been missed if not Food/FoodScienceResearch/LaboratoryMethods/ucm374006.htm
reported or the outbreak link was not considered. Third, (accessed July 23, 2014).
we were unable to inspect importer D because of political 10 Multistate outbreak of hepatitis A virus infections linked to
pomegranate seeds from Turkey (final update). http://www.cdc.gov/
unrest in the region. hepatitis/outbreaks/2013/a1b-03-31/index.html (accessed July 23, 2014).
In conclusion, public notification, product recall, and 11 FDA investigates multistate outbreak of hepatitis A illnesses
hepatitis A postexposure prophylaxis within 2 weeks of associated with pomegranate seeds from Turkish importer. http://
www.fda.gov/Food/RecallsOutbreaksEmergencies/Outbreaks/
exposure can prevent illnesses during outbreaks; these ucm354698.htm (accessed July 23, 2014).
eorts are helped by retailers with reliable sales data 12 Hutin YJ, Pool V, Cramer EH, et al. A multistate, foodborne
who are willing to work with public health agencies. The outbreak of hepatitis A. N Engl J Med 1999; 340: 595602.
foreign origin of the contaminated ingredient in this 13 Niu MT, Polish LB, Robertson BH, et al. Multistate outbreak of
hepatitis-A associated with frozen strawberries. J Infect Dis 1992;
outbreak underscores the importance of the new import- 166: 51824.
related regulations in the Food Safety Modernization 14 Petrignani M, Harms M, Verhoef L, et al. Update: A food-borne
Act. Under this law, importers are required to have a outbreak of hepatitis A in the Netherlands related to semi-dried
tomatoes in oil, January-February 2010. Euro Surveill 2010; 15: 19572.
programme to verify food product safety, and will be 15 Shieh YC, Stewart DS, Laird DT. Survival of hepatitis A virus in spinach
required to implement risk-based preventive controls. during low temperature storage. J Food Protect 2009; 72: 239093.
Contributors 16 McCaustland KA, Bond WW, Bradley DW, Ebert JW, Maynard JE.
All authors collected, analysed, and interpreted data, and wrote and Survival of hepatitis A virus in feces after drying and storage for
1 month. J Clin Microbiol 1982; 16: 95758.
revised the report.
17 Sow H, Desbiens M, Morales-Rayas R, Ngazoa SE, Jean J. Heat
Hepatitis A outbreak investigation team inactivation of hepatitis A virus and a norovirus surrogate in soft-
Ronald Balajadia, Steany J Cavallo, Lei Chen, Elizabeth R Daly, shell clams (Mya arenaria). Foodborne Pathog Dis 2011; 8: 38793.
Paul Ettestad, Julianna Ferreira, Eric Garza, Gemechu Gerbi, Debra Gilliss, 18 Normann A, Badur S, Onel D, et al. Acute hepatitis A virus
Rikita I Hatia, Michael Hernandez, James Kazmierczak, Maja Kodani, infection in Turkey. J Med Virol 2008; 80: 78590.
Ben Kupronis, Yulin Lin, Trudy Murphy, Jennifer F Myers, Vi Peralta, 19 Nejati A, Makvandi M, Samarbafzadeh A, Neisi N,
Tasha Poissant, Suzanne Rouleau, Eyasu Teshale, Daulati Thakare, Moradzadegan H. Molecular epidemiology of hepatitis A virus in
John W Ward, Joli R Weiss, Kathleen Winter, and Guo-liang Xia. patients in the Ahwaz region of Iran. J Med Virol 2012; 84: 58286.
20 Faber MS, Stark K, Behnke SC, Schreier E, Frank C. Epidemiology
Declaration of interests of hepatitis A virus infections, Germany, 2007-2008.
We declare no competing interests. Emerg Infect Diseases 2009; 15: 176068.
Acknowledgments 21 Nordic outbreak investigation team. Joint analysis by the Nordic
This study was funded by the US Centers for Disease Control and countries of a hepatitis A outbreak, October 2012 to June 2013:
Prevention, the Food and Drug Administration, and other state and frozen strawberries suspected. Euro Surveill 2013; 18: 20520.
local public health departments. We thank the following people for their 22 Swinkels HM, Kuo M, Embree G, et al. Hepatitis A outbreak in
contributions to the investigation: Nicole Comstock, Alycia Downs, British Columbia, Canada: the roles of established surveillance,
Kathleen Harriman, Cynthia Jorgensen, Bill Keene, Kenneth Komatsu, consumer loyalty cards and collaboration, February to May 2012.
Euro Surveill 2014; 19: 20792.
Paul Peterson, Devin Raman, and Phil Spradling. The findings and
conclusions in this report are those of the authors and do not 23 Carvalho C, Thomas HL, Balogun K, Tedder R, Pebody R,
Ramsay M, Ngui SL. A possible outbreak of hepatitis A associated
necessarily represent the ocial position of the Centers for Disease
with semi-dried tomatoes, England, JulyNovember 2011.
Control and Prevention. Euro Surveill 2012; 17: 20083.
References 24 Donnan EJ, Fielding JE, Gregory JE, et al. A multistate outbreak of
1 Novak R, Williams I, Bell B. Advisory Committee on Immunization hepatitis A associated with semidried tomatoes in Australia, 2009.
Practices (ACIP) Centers for Disease Control and Prevention Clin Infect Dis 2012; 54: 77581.
(CDC). Update: prevention of hepatitis A after exposure to 25 Gallot C, Grout L, Roque-Afonso AM, et al. Hepatitis A associated
hepatitis A virus and in international travelers. Updated with semidried tomatoes, France, 2010. Emerg Infect Diseases 2011;
recommendations of the Advisory Committee on Immunization 17: 56667.
Practices (ACIP). MMWR Recomm Rep 2007; 56: 108084. 26 Holmberg SD. Hepatitis A epidemiology goes global. Clin Infect Dis
2 Fiore AE, Wasley A, Bell BP. Advisory Committee on 2012; 54: 78283.
Immunization Practices (ACIP). Prevention of hepatitis A 27 Armstrong GL, Billah K, Rein DB, Hicks KA, Wirth KE, Bell BP.
through active or passive immunization: recommendations of the The economics of routine childhood hepatitis A immunization in
Advisory Committee on Immunization Practices (ACIP). the United States: the impact of herd immunity. Pediatrics 2007;
MMWR Recomm Rep 2006; 55: 123. 119: e2229.
3 Viral hepatitis surveillance United States, 2011. http://www.cdc. 28 Klevens RM, Kruszon-Moran D, Wasley A, et al. Seroprevalence of
gov/hepatitis/Statistics/2011Surveillance/index.htm (accessed hepatitis A virus antibodies in the U. S.: results from the National
July 23, 2014). Health and Nutrition Examination Survey. Public Health Rep 2011;
4 Klevens RM, Miller J, Iqbal K, et al. The evolving epidemiology of 126: 52232.
hepatitis A in the United States: incidence and molecular 29 Koopmans M, Duizer E. Foodborne viruses: an emerging problem.
epidemiology from population-based surveillance. Int J Food Microbiol 2004; 90: 2341.
Arch Intern Med 2010; 170: 181118.
20
Fruit lovers infected with hepatitis A have been betrayed by poor hygiene overseas.
The danger of uncooked foreign food has been exposed after berries contaminated with
a hard-to-detect virus infected at least nine Australians.
Frozen mixed berries from Chile and China sold by the Nanna and Creative Gourmet
brands have been recalled nationwide, prompting experts and industry voices to
promote buying local.
More than a dozen cases of hepatitis A linked to the imported berries have been
confirmed around the country.
Dont eat it, says Australian Blueberry Growers Association president Greg McCulloch.
They might think its just a germ or something, but hepatitis A is actually a germ in s**t,
which might make them think twice.
Food standards in China may not match those to which Australian farmers
adhere. Photo: Getty
Eating foreign imports without cooking or boiling is dangerous because only heat
cleanses food tainted by faecal matter. Australia is the latest to be stricken by frozen
berry imports, after similar infections in Ireland, Italy and several Scandinavian countries
since 2013.
Poor hygiene is the most likely cause of the contamination, The University of Melbourne
agricultural expert Dr Said Ajlouni agrees.
The virus probably came from infected people who went to the toilet and came back to
the factory without appropriate hand washing, he says.
Because we dont have any control over good hygiene and sanitation practices in these
countries, they can be very risky.
Raw human sewage, which may contain viruses, bacteria and industrial chemicals, is
widely used as fertiliser for a wide variety of crops in China and Latin America, says a
2008 report by Sri Lanka-based International Water Management Institute (IWMI).
It would be unfair to boycott foreign fresh food based on one incident, says Dr Ajlouni,
but he confirms that buying local is safer.
Farmers are calling for an overhaul of labelling laws to help consumers identify
Australian grown and packaged foods following the berry scare.
Victorian Farmers Federation president Peter Tuohey says loopholes in labelling laws
have allowed many imported foods to be passed off as if they are Australian made.
Weve been pushing for truth-in-labelling laws for years and it hasnt happened yet, he
says.
Australian berries are grown under stringent standards, says another industry voice.
The fresh Australian berry is certainly grown under some of the strictest food safety
standards in the world, says Raspberries and Blackberries Australia spokesman
Jonathan Eccles.
Consumers should be confident that when theyre buying a fresh berry it is safe.
The impediment to safety is price. Most local berries are sold fresh because Australias
high labour costs make it impossible to compete in the frozen fruit market.
For example, it costs berry grower Phil Rowe more to transport his fresh berries from his
farm in West Gippsland to Melbourne than to ship a frozen load from Chile.
If you prefer the frozen variety, homegrown berries can just as easily be stored in the
freezer, says the Sunny Creek Farm owner.
The recalled products are 1kg packs of Nannas Frozen Mixed Berries, and 300g and
500g packs of Creative Gourmet Mixed Berries.
Hepatitis symptoms usually begin two to seven weeks after exposure and include fever,
loss of appetite, nausea, fatigue, abdominal pain, diarrhoea and jaundice and can last
for months.
EFSA Journal 2011;9(10):2390
ABSTRACT
On 21 May 2011, Germany reported an ongoing outbreak of Shiga-toxin producing Escherichia coli
(STEC), serotype O104:H4. From an initial case control study, the outbreak was associated with the
consumption of fresh salad vegetables. Subsequent investigations showed that the risk of infection was
significantly associated with the consumption of fresh sprouted seeds rather than with other fresh
vegetables. A tracing back and tracing forward study showed that all of the clusters for which there
was sufficient information, could be attributed to consumption of sprouted seeds from a single
sprouted seed producer in Germany. Investigation of the production site showed no evidence of
environmental contamination. Employees were found to be infected, but since they had not become ill
prior to the outbreak, it was concluded that they were not the source of the food contamination. Hence,
contaminated seeds used for the sprout production were the most likely source. Several varieties of
seeds were used, and sprouts thereof were sold as a mixture. Subsequently, a cluster of patients with
bloody diarrhoea was reported, after having participated in a local event in France on 8 June.
Consumption of sprouted seeds was also associated with occurrence of the disease in this cluster.
Furthermore, the STEC isolates responsible for the outbreaks in France and Germany were found to be
indistinguishable. It was therefore concluded that there was a common source for both outbreaks. A
comparison of the back tracing information on the seeds from the French and German outbreaks led to
the conclusion that a specific consignment (lot) of fenugreek seeds imported from Egypt was the most
likely link between the outbreaks, although it could not be excluded that other lots from the same
exporter and importer were also implicated. STEC O104 is a very rare serogroup in humans in the EU
and worldwide. Sporadic cases in the EU have been previously linked to travel to North Africa, the
Middle East and Central Asia. On 26 July, the Robert Koch Institute declared the outbreak finished. A
total of 3911 cases have been reported to the ECDC and WHO, linked to the outbreaks, to date.
KEY WORDS
E. coli, outbreak, sprouted seeds, fenugreek, STEC, VTEC, EHEC, EAggEC, ESBL
1
On request from EFSA, Question No EFSA-Q-2011-00958, issued on 16 September 2011.
2
Correspondence: emrisk@efsa.europa.eu
3
Acknowledgement: EFSA wishes to thank Alfredo Caprioli, Alexandra Fetsch, Matthias Greiner, Stefan Gross, Arie Havelaar,
Petra Luber, Anika Schielke and Armin Weiser for the preparatory work on this scientific output and EFSA staff: Pierre-Alexandre
Beloeil, Ernesto Liebana Criado, Hubert Deluyker, Olaf Mosbach-Schulz, Tobin Robinson, Didier Verloo for the support provided
to this scientific output. This Scientific Report has been reviewed by John Cowden, Tine Hald and James Mclauchlin.
Suggested citation: European Food Safety Authority; Shiga toxin-producing E. coli (STEC) O104:H4 2011 outbreaks in Europe:
Taking Stock. EFSA Journal 2011;9(10):2390. [22 pp.] doi:10.2903/j.efsa.2011.2390.
Available online: www.efsa.europa.eu/efsajournal
SUMMARY
On 21 May 2011, Germany reported an ongoing outbreak of Shiga-toxin producing Escherichia coli
(STEC), serotype O104:H4. As of 27 July, 3126 cases of diarrhoeal disease caused by STEC E. coli
O104:H4 (probable and confirmed), including 17 deaths linked to the outbreak in Germany and
occurring in the EU (including Norway) have been reported to the European Centre for Disease and
Control (ECDC). In addition, in the EU 773 cases of haemolytic uraemic syndrome (HUS) caused by
this bacterium were also reported, including 29 deaths, and linked to the German outbreak. At this
time of reporting, a further 119 suspect cases including 4 deaths are also associated with this outbreak.
In addition, outside the EU 8 cases of STEC and 5 cases of HUS, including 1 death have been reported
in the USA, Canada and Switzerland through the international health regulations (IHR), all with recent
travel history to Germany. Based on the most recently reported information (last updated on 26 July
2011), the clinical onset of the last outbreak-related case in Germany was 4 July 2011.
Shortly after the onset of the outbreak in Germany, case-control studies conducted by the Robert Koch
Institute (RKI) demonstrated that clinical disease was statistically significantly associated with the
consumption of fresh salad vegetables. The high proportion of adult women among cases, was
consistent with fresh salad vegetables as the source of infection. Later, a detailed cohort study
demonstrated an association with sprouted seeds.
A tracing back and tracing forward study showed that most of the clusters, and all of the clusters for
which there was sufficiently detailed food consumption data, could be attributed to consumption of
sprouted seeds from one producer in Germany. Investigation of the production site showed no
evidence of environmental contamination. Some employees were found to be infected, but reported
not to have become ill prior to the outbreak and hence it was concluded that they were not the source
of the contamination. This left the seeds used for the sprout production as the prime suspect vehicle of
infection; however it was not possible to identify a single seed source since different species were used
to produce the sprouted seeds, which were sold as several different mixtures.
On the 24 June, France reported to the Rapid Alert System for Food and Feed (RASFF4), a cluster of
patients with bloody diarrhoea, after having participated in an event in the Commune of Bgles near
Bordeaux on the 8 June. At the time of issue of this report, there were 2 confirmed STEC cases and 9
cases with HUS reported to ECDC, with 4 further cases of non-HUS suspected. Eleven of these
patients, 7 women and 4 men, between 31 and 64 years of age, had attended the same event in Bgles.
Infection with E. coli O104:H4 was confirmed for 12 of the 15 cases. Epidemiology studies on the
French outbreak also implicated sprouted seeds as the outbreak vehicle.
The phenotypic and genotypic characterisation of the E. coli O104:H4 indicated that the isolates from
the French and German outbreaks were common to both incidents. Hence, it was concluded that the
same strain was involved in the outbreaks both in Germany and in France, strongly indicating a
common source.
A tracing back initiated to find the common food source for both outbreaks revealed fenugreek seeds
to be common to the 2 outbreaks. The comparison of the back tracing information on the seeds from
the French and German outbreaks led to the conclusion that a specific consignment (lot) of fenugreek
seeds imported from Egypt was the most likely link between the outbreaks, although it could not be
excluded that other lots imported by the same importer/exporter might be implicated.
Data concerning the trace back and trace forward were exchanged through the RASFF, allowing the
Member States and European institutions to receive up to date information.
The actual cause or route of contamination of the seed has not been demonstrated. However, based on
epidemiological and microbiological investigations within the EU as well as from previous sprouting
4
RASFF: http://ec.europa.eu/food/food/rapidalert/index_en.htm
seed-related outbreaks, it is likely that contamination occurred during seed production. This part of the
investigation would need to extend beyond the point of EU import to include the site(s) of production.
The published data for STEC O104 are scarce as this is a very rare serogroup infecting humans in
Europe and globally. According to the information reported to ECDC, there were 10 reported cases of
STEC O104 infection in the EU Member States and Norway during 2004-2010 from: Austria (1 case
in 2010), Belgium (1 case in 2008), Denmark (1 case in 2008), Finland (1 case in 2010), France (1
case in 2004), Norway (1 case in 2006, 3 cases in 2009), and Sweden (1 case in 2010). Moreover, a
paediatric case of HUS which occurred in Italy in 2009 has now been associated with STEC O104, so
giving a total of 11 cases.
Five of the 10 cases between 2004 and 2010 were related to travel outside the EU; the countries of
origin of the infection being Afghanistan (2008), Egypt (2010), Tunisia (2009, 2010) and Turkey
(2009). Only 3 of the STEC O104 strains isolated from these cases were of serotype O104:H4 (in
Finland in 2010, in Italy in 2009, and in France in 2004). As for the outbreak strain, the STEC
O104:H4 strains isolated in Italy and in Finland were both positive for genetic markers of
enteroaggregative adhesion, but differed from the 2011 epidemic strain in that they were negative for
extended-spectrum beta-lactamase production. The Finnish case was travel-related, with infection
acquired in Egypt; whereas the Italian case had a recent history of travel to Tunisia. The origin of the
source of infection for the French case was not reported.
In addition to those cases reported to ECDC, a review of the scientific literature revealed that STEC
O104:H4 has been isolated twice in Germany in 2001 and once in Korea in 2005. The German isolates
differed from the 2011 outbreak strain.
E. coli O104:H4 was not isolated from any batches of the suspect fenugreek seeds. The inability to
demonstrate the presence of E. coli O104:H4 in the suspect seeds is not unexpected. It is possible that
contaminated seeds were no longer in stock when sampling took place, or even if present were
contaminated at a level which made isolation of the organism impossible. However, this does not
mean that enterobacteriaceae would not have been present in seeds and sprouted seeds. Previous
studies have shown enterobacteriaceae to be present on the surface of the tissue of the plants and that
they can also be internalised within the plant (e.g. at primary production, through irrigating with
contaminated water or application of organic fertilizer not properly treated and still containing enteric
pathogens). In this regard, it is important to underline that a negative laboratory test does not prove the
absence of a pathogen, which may be unevenly distributed within the food matrix, perhaps at low
levels. This is particularly true when dealing with seeds: the matrix is made up of particles (seeds)
which individually may become contaminated and dispersed in large lots. Similarly, favourable
physico-chemical conditions to support survival or growth may not be homogeneously distributed.
The preparation of fresh sprouted seeds seldom includes a step where bacterial contamination is
eliminated. Hence, food preparation of fresh sprouted seeds is based on the understanding that they are
sold as ready-to-eat, i.e. safe to eat as is, or following only minimal preparation. For fresh produce,
this assumes and relies on a production process which prevents contamination and an ability to detect
contamination when it occurs. These conditions have proven not to be satisfied in this case. The fact
that the sampling and bacteriological methods might fail to detect the presence of pathogens such as
STEC O104:H4 and Salmonella spp. emphasises the importance of good production and handling
practices, as the public health protection associated with such a criterion is questionable given the
difficulty of tracing pathogens in seeds.
TABLE OF CONTENTS
Abstract .................................................................................................................................................... 1
Summary .................................................................................................................................................. 2
Table of contents ...................................................................................................................................... 4
Background as provided by EFSA ........................................................................................................... 5
Terms of reference as provided by EFSA ................................................................................................ 5
Assessment ............................................................................................................................................... 6
1. Introduction ..................................................................................................................................... 6
2. Characterisation of the E. coli O104:H4 strain isolated in the current outbreak ............................. 7
3. Previous worldwide occurrence of E. coli O104:H4 ....................................................................... 8
4. Background on the outbreaks in the EU .......................................................................................... 9
5. Epidemiological investigations into the vehicle, source and cause of the outbreak ........................ 9
6. Microbiological investigations in food and the environment ........................................................ 12
7. Discussion ...................................................................................................................................... 14
7.1. Public health aspects ............................................................................................................. 14
7.2. Epidemiological aspects........................................................................................................ 15
7.3. Microbiological aspects ........................................................................................................ 15
Conclusions ............................................................................................................................................ 17
References .............................................................................................................................................. 19
Appendix I .............................................................................................................................................. 21
Glossary .................................................................................................................................................. 22
On the 21 of May 2011, Germany reported an ongoing outbreak of Shiga-toxin producing Escherichia
coli- bacteria (STEC56), serotype O104:H4. As of 27 July, 3126 cases of diarrhoeal disease caused by
STEC E. coli O104:H4 (probable and confirmed), including 17 deaths linked to the outbreak in Germany
and occurring in the EU (including Norway) have been reported to the ECDC. In addition, 773 cases of
haemolytic uraemic syndrome (HUS) caused by this bacterium were also reported, including 29 deaths, and
linked to the German outbreak. At this time of reporting, a further 119 suspect cases including 4 deaths are
also associated with this outbreak7.
On the 24 of June, France reported8 a cluster of patients with bloody diarrhoea, after having participated in
an event in the Commune of Bgles near Bordeaux on the 8 of June. At the time of issue of this report,
there were 2 confirmed STEC cases and 9 cases with HUS reported to ECDC, with a further 4 suspected
cases of non-HUS. Eleven of these patients had attended the same event in Bgles. Infection with E. coli
O104:H4 was confirmed for 12 of the 15 cases. Epidemiology studies on the French outbreak also
implicated sprouted seeds as the outbreak vehicle.
Both events have resulted in numerous investigations to establish the extent, vehicle, source, and cause of
this outbreak. These investigations have led to the conclusion that fenugreek seeds imported into Europe
from Egypt were the most likely source of contamination, and sprouted seeds the most likely vehicle. The
objective of this review is to provide an overview of key information that was gathered and to reference
where this information is presented in detail.
The Emerging Risks unit (EMRISK) is requested to coordinate the following activities in collaboration
with the Scientific Support and Assessment Unit, Biological Hazards Unit, Biological Monitoring Unit and
Communications;
1) In collaboration with the French and German authorities, and the EU Reference Laboratory on E.
coli, prepare a Scientific Report summarising the strain characteristics, epidemiological
investigations, and analytical methods for foods linked to this outbreak. Make extensive reference
to the reports produced by the national authorities and European institutions concerning this
outbreak.
2) In collaboration with the units involved in EFSAs outbreak response, produce a discussion paper
summarising EFSAs contribution to resolving the outbreak, and highlighting lessons learnt for
EFSA, for collaboration at MS level and at the level of EU coordination. This paper will address
both EFSAs scientific work and risk communications activities undertaken in coordination with
ECDC, Member States and the European Commission.
This report only addresses the first of the terms of reference (ToR). The second ToR will be addressed in a
separate document.
5
European Food Safety Authority; Urgent advice on the public health risk of Shiga-toxin producing
Escherichia coli in fresh vegetables. EFSA Journal 2011; 9(6):2274. [50 pp.]
6
In the EU and as reflected in EFSAs work on zoonoses, Shiga-toxin producing Escherichia coli are referred to as
VTEC (verotoxin-producing E. coli), but the term STEC is used for this outbreak as it is in line with terminology used
by WHO and other organisations.
7
http://www.ecdc.europa.eu/en/activities/sciadvice/Lists/ECDC%20Reviews/ECDC_DispForm.aspx?List=512ff74f-
77d4-4ad8-b6d6-bf0f23083f30&ID=1166&RootFolder=/en/activities/sciadvice/Lists/ECDC Reviews&MasterPage=1
8
RASFF Alert Notification 2011.08428 RASFF Alert Notification 2011.0842, Available at
https://webgate.ec.europa.eu/rasff-window/portal/index.cfm?event=notificationsList
ASSESSMENT
1. Introduction
On the 21 of May 2011, Germany reported an ongoing outbreak of Shiga-toxin producing Escherichia
coli- bacteria (STEC910), serotype O104:H4. As of 27 July, 3126 cases of diarrhoeal disease caused by
STEC E. coli O104:H4 (probable and confirmed), including 17 deaths linked to the outbreak in
Germany and occurring in the EU (including Norway) have been reported to the European Centre for
Disease Prevention and Control (ECDC). In addition, 773 cases of haemolytic uraemic syndrome
(HUS) caused by this bacterium were also reported, including 29 deaths, and linked to the German
outbreak. At this time of reporting, a further 119 suspect cases including 4 deaths are also associated
with this outbreak 11. Further details on the German cases are available from the Robert Koch
Institute12 (RKI, 2011).
On the 24 of June, France reported13 a cluster of patients with bloody diarrhoea, after having
participated in an event in the Commune of Bgles near Bordeaux on the 8 of June. At the time of
issue of this report, there were 2 confirmed STEC cases and 9 cases with HUS reported to ECDC, with
a further 4 suspected cases of non-HUS. Eleven of these patients, 7 women and 4 men, between 31
and 64 years of age, had attended the same event in Bgles. Infection with E. coli O104:H4 was
confirmed for 12 of the 15 cases. Epidemiological studies on the French outbreak also implicated
sprouted seeds as the outbreak vehicle.
Both events have resulted in investigations to find the source of this outbreak which resulted in the
conclusion that contaminated fenugreek seeds imported into Europe from Egypt were the most likely
source. The objective of this review is to provide an overview of key information that was gathered
and to reference where this information is presented in detail. The paper covers what is known about:
9
European Food Safety Authority; Urgent advice on the public health risk of Shiga-toxin producing
Escherichia coli in fresh vegetables. EFSA Journal 2011; 9(6):2274. [50 pp.]
10
In the EU and as reflected in EFSAs work on zoonoses, Shiga-toxin producing Escherichia coli are referred to as
VTEC (verotoxin-producing E. coli) but the term STEC is used for this outbreak as it is in line with terminology used
by WHO and other organisations.
11
http://www.ecdc.europa.eu/en/activities/sciadvice/Lists/ECDC%20Reviews/ECDC_DispForm.aspx?List=512ff74f-
77d4-4ad8-b6d6-bf0f23083f30&ID=1166&RootFolder=/en/activities/sciadvice/Lists/ECDC Reviews&MasterPage=1
12
http://www.rki.de/cln_169/nn_217400/EN/Home/PM__EHEC.html
13
RASFF Alert Notification 2011.084213 RASFF Alert Notification 2011.0842, Available at
https://webgate.ec.europa.eu/rasff-window/portal/index.cfm?event=notificationsList
For a general recent brief review of characteristics of E. coli O104:H4 that are particularly relevant to
this outbreak, see EFSA (2011a).
Germany
The outbreak strain is a Shiga toxin producing Escherichia coli (STEC) that belongs to serotype
O104:H4. Detailed information on the outbreak strain characteristics can be found at the Robert Koch-
Institut (RKI) website14. Briefly, its microbiological characteristics are as follows:
In conclusion, the STEC O104:H4 outbreak strain shows an unusual combination of virulence factors
of STEC and EAggEC which has only been reported for a group of strains of serotype O111:H2
(Morabito et al., 1998) isolated in France during an outbreak of HUS (Boudailliez et al., 1997). As in
the present outbreak, the association of the O111 strains with severe disease supports the view that this
unusual combination of virulence factors might confer a very high degree of virulence. It is also
noteworthy that the outbreak described by Boudailliez et al. (1997), was ascribed to a person-to-
person transmission, without any evidence of foodborne transmission.
Whole genome sequencing data confirms that the German outbreak strain is an EAggEC which can be
distinguished from those of other O104:H4 strains because it contains a prophage encoding Shiga
toxin 2 and a distinct set of additional virulence and antibiotic-resistance factors which have been
acquired by horizontal genetic exchange (Rasko et al., 2011).
14
http://www.rki.de/cln_160/nn_217400/EN/Home/EHECO104,templateId=raw,property=publicationFile.pdf/EHECO1
04.pdf
France
The strain of E. coli O104:H4, isolated from 5 HUS cases, all of whom consumed sprouted seeds
served at an event in Bgles (near Bordeaux, South-west France) on 8 June 2011, exhibits the
following characteristics (ANSES, 2011):
o The strain contains the stx2 and aggR genes, but not the genes coding for intimin (eae),
haemolysin A (hlyA) and EAST1 toxin (astA);
o The antimicrobial susceptibility pattern of the strain is as follows: resistant to ampicillin,
cefotaxime, ceftazidime, cotrimoxazole, streptomycin, trimethoprim, sulfamethoxazole,
tetracyclin, and nalidixic acid. Sensitive to chloramphenicol, ciprofloxacin, gentamicin,
imipenem, and kanamycin;
o The strain contains extended-spectrum beta-lactamase (ESBL) blaCTX-M-15 (group 1) gene and
the penicillinase blaTEM gene.
Thus, information on the microbiological characterisation of the isolates implicated in the French
outbreak indicate that many characteristics (stx2 positive, eae negative, hlyA negative, multi-resistance
pattern to antimicrobials) are common with the German outbreak strain. In addition, the 2 molecular
techniques (Repetitive sequence based Polymerase Chain Reaction (Rep-PCR) and pulsed-field gel
electrophoresis (PFGE)) used to compare the outbreak strain in France and isolates of E. coli O104:H4
from 2 imported cases in France linked to the E. coli O104:H4 outbreak in Germany in May and June
2011, showed that the isolates were indistinguishable by these methods (Gault et al., 2011).
ECDC and EFSA reported on the previous occurrence of STEC O104:H4 (ECDC, EFSA 2011). The
text below is extracted from this document.
The published data for STEC O104 is scarce as this is a very rare serogroup in humans in Europe and
globally. According to the information reported to ECDC, there were 10 reported cases infected with
STEC O104 in the EU Member States and Norway in the period 2004-2009. These STEC O104 cases
have been reported from Austria (1 case in 2010), Belgium (1 case in 2008), Denmark (1 case in
2008), Finland (1 case in 2010), France (1 case in 2004), Norway (1 case in 2006, 3 cases in 2009),
and Sweden (1 case in 2010) (ECDC and EFSA, 2011). Moreover, a paediatric case of HUS which
occurred in Italy in 2009 has now been associated with STEC O104 (Scavia et al., 2011), so giving a
total of 11 cases.
Based on the known data, 56% of the cases were male and the age ranged from <1 year to 76 years.
Two of the cases (18%) developed HUS. Five of the cases were travel related; the countries of origin
of the infection being Afghanistan (2008), Egypt (2010), Tunisia (2009 and 2010) and Turkey (2009).
Only 3 of the STEC O104 strains isolated from these cases were of serotype O104:H4 (in Finland in
2010, in Italy in 2009, and in France in 2004). Similar to the outbreak strain, the STEC O104:H4
strains isolated in Italy and in Finland were both positive for genetic markers of enteroaggregative
adhesion, but negative for ESBL (extended-spectrum beta-lactamase) production. As mentioned, the
Finnish case was travel-related, with the infection acquired in Egypt; whereas the Italian case had a
recent history of travel to Tunisia (Scavia et al., 2011). The origin of the source of infection for the
French case was not reported.
From a review of the scientific literature, STEC O104:H4 has been isolated twice in Germany in 2001
(Mellmann et al, 2008) and once in Korea in 2005 (Bae et al, 2006). The case in Korea was a 29-year
old woman who developed HUS. The source of the infection was unknown. The STEC O104:H4
from Germany, with a MLST ST678, was isolated from an HUS case (Mellmann et al., 2008). Like
the current outbreak strain, the isolate was stx2-positive, eae-negative, and positive for
enteroaggregative adhesion and EAggEC genetic markers. However, it was negative for ESBL
production15, and some other differences with the outbreak strain were reported (Bielaszewska et al.,
2011).
This section complements the ECDC/EFSA rapid risk assessment16 published on 29 June, 2011, and
the ECDC rapid risk assessment update published on 8 July17.
As of 27 July, 3126 cases of STEC caused by E. coli O104:H4 (probable and confirmed), including 17
deaths in the EU (including Norway), linked to the outbreak in Germany, have been reported to the
ECDC. In addition, in the EU 773 cases of HUS have been reported, including 29 deaths, also linked
to the German outbreak18. A further 119 cases of STEC and 4 deaths are suspected at this time of
reporting. Switzerland has so far reported 5 non-HUS STEC cases through the international health
regulations (IHR), the United States reported 5 HUS cases including 1 death and 2 non-HUS cases, all
with recent travel history to Germany. Canada has also reported 1 non-HUS case with travel history to
Germany.
As based on the last reported information (26 July 2011), the clinical onset of the last outbreak-related
case in Germany was the 4 July 201119.
On the 24 June, France reported to the RASFF20 a cluster of patients with bloody diarrhoea, after
having participated in an event in the Commune of Bgles near Bordeaux on the 8 June. At the time of
issue of this report, there were 2 confirmed STEC cases and 9 cases with HUS reported to ECDC, with
4 further cases of non-HUS suspected. Eleven of these patients, 7 women and 4 men, between 31 and
64 years of age, had attended the same event in Bgles. Infection with E. coli O104:H4 was confirmed
for 12 of the 15 cases. Epidemiology studies on the French outbreak also implicated sprouted seeds as
the outbreak vehicle.
An important development since 29 June comes from the results of the screening of children and staff
in a school in Germany where 3 cases of HUS and a further 4 cases of STEC related diarrhoea had
been identified. As most recently reported (5 August 2011)21, the epidemic strain of E. coli was
isolated from 22 of the 30 children tested, indicating a high level of occurrence of asymptomatic
carriage. Three teachers at the school were found to be asymptomatic carriers of STEC as were 2
employees of the catering company supplying the school.
5. Epidemiological investigations into the vehicle, source and cause of the outbreak
Early after the onset of the outbreak in Germany, case-control studies conducted by the RKI
demonstrated that the onset of clinical disease was linked to the consumption of fresh salad vegetables.
Circumstantial evidence, such as the high proportion of adult women among the ill, was also
15
http://www.ehec.org/index.php?hid=43&lang=de&pid=HUSEC
16
http://ecdc.europa.eu/en/publications/Publications/2011June29_RA_JOINT_EFSA_STEC_France.pdf
17
http://www.ecdc.europa.eu/en/publications/Publications/110712_TER_Risk_Assessement_Ecoli.pdf
18
http://www.ecdc.europa.eu/en/activities/sciadvice/Lists/ECDC%20Reviews/ECDC_DispForm.aspx?List=512ff74f-
77d4-4ad8-b6d6-bf0f23083f30&ID=1166&RootFolder=/en/activities/sciadvice/Lists/ECDC Reviews&MasterPage=1
19
http://www.rki.de/cln_110/nn_205760/DE/Content/InfAZ/E/EHEC/Info-
HUS,templateId=raw,property=publicationFile.pdf/Info-HUS.pdf
20
RASFF: http://ec.europa.eu/food/food/rapidalert/index_en.htm
21
http://www.kreis-paderborn.org/kreis_paderborn/presse/2011/entries/erster-todesfall-ehec.php
consistent with vegetables being the vehicle of infection. Later, a detailed cohort study demonstrated
an association with sprouted seeds22.
In an attempt to identify the food source through microbiological examination, the responsible
competent authorities of the German Federal States performed risk-based inspections of clusters
(where a common meal or place of eating was identified) and took samples from foods and the
environment at clusters, suppliers, whole sale markets, and at retail23. So far (26 July, 2011), none of
the food samples taken during this monitoring activity, for which cross-contamination (from infected
persons in the household) could be excluded, has been found positive for the STEC serotype O104:H4
(BfR, 2011b).
Consequently, it was decided that a co-ordinated initiative to identify the food source and stop the
outbreak was needed. For this purpose, the German Task Force EHEC (Enterohaemorrhagic
Escherichia coli) was established by the German Federal Ministry of Food, Agriculture and Consumer
Protection (BMELV). Members of the Task Force EHEC were experts from the Federal States in
Germany, from the Federal Office of Consumer Protection and Food Safety (BVL), the Federal
Institute for Risk Assessment (BfR), the RKI and EFSA scientists. The Task Force EHEC was hosted
in the crisis centre at the BVL (Task Force EHEC, 2011).
In a first phase, 2 different strategies were followed to identify the STEC contaminated food. The first
approach examined 5 clusters in Germany (associated with hotels, restaurants or canteens), which
were prioritized due to the availability of precise information on the food intake of the affected cases
(e.g. tour parties, family celebrations etc.). A detailed trace back analysis was conducted for all salad
ingredients (including sauces and dressings etc.) and toppings that had been served to the customers.
The second approach was a trace forward analysis of the supply chains of a horticultural farm in
Lower Saxony (Establishment A) in order to find out if any of the clusters were linked to sprouting
seeds originating from there. The production site was suspected early on by the local authorities as a
possible source of infection for several clusters. Furthermore, STEC O104:H4 was isolated from the
faeces of employees of Establishment A, including 2 asymptomatic cases.
A total of 41 clusters were identified for which there was sufficient information on food intake in
order to carry out the tracing. The analysis of the sprouted seed supply chains showed that for all 41
clusters, there was a common link to sprouted seeds from Establishment A (the trace forward
approach). The results of the detailed trace back analysis at 5 clusters confirmed the findings of the
trace forward analysis. It was shown that sprouted seeds from Establishment A had been delivered to
all 5 clusters. Thus, the epidemiological investigations provided strong evidence that sprouted seeds
from Establishment A were the vehicle of infection of the German STEC O104:H4 outbreak24.
Further epidemiological investigations revealed that the disease occurrence coincided with
consumption of either of 2 different sprouted seed mixtures, the Keimspross-Mischung / Milde
Mischung (germ sprout or mild blend) or the Wrz-Mischung (spicy blend). The mild blend
contained 4 and the spicy 3 different kinds of sprouted seeds. Only lentil and fenugreek sprouts were
present in both mixtures. Whereas lentil sprouts were also used in other sprouted seed mixtures,
fenugreek sprouts were used for Keimspross-Mischung / Milde Mischung and Wrz- Mischung
only25. Hence, fenugreek sprouts were almost exclusively used for the 2 mixtures and only small
amounts were sold unmixed.
The sprouting procedure of fenugreek seeds was examined in detail. Fenugreek seed sprouting was
initiated twice per week, every week. The sprouting process took 3.5 days. After sprouting, the
22
http://www.rki.de/cln_117/nn_217400/EN/Home/EHEC__Report,templateId=raw,property=publicationFile.pdf/EHEC
_Report.pdf
23
www.bvl.bund.de/taskforce_en
24
www.bvl.bund.de/taskforce_en
25
www.bvl.bund.de/taskforce_en
fenugreek seeds were mixed with other sprouted seeds to generate the Keimspross-Mischung / Milde
Mischung and the Wrz-Mischung. It was assumed that fresh sprouted seeds were eaten after a
maximum of 14 days following their production. Starting with the known dates of exposure of the
cases from the selected 5 clusters and the delivery dates for the sprouted seeds to the place of exposure
(hotels, restaurants or canteens) the possible production dates at Establishment A were assessed. A
comparison with the total epidemiological curve of the outbreak confirmed the isolated production
cycles as consistent with the whole outbreak. Establishment A used lot nos. 48088 and 8266 of
fenugreek seeds for their production during April to mid-May 2011. The earlier one (lot no. 48088)
was totally used up when sampling by the local authorities took place at Establishment A (Task Force
EHEC 2011). Both lots were traced back to a single importer in Germany, who imported the seeds
from Egypt.
Three of the personnel of Establishment A developed diarrhoeal symptoms consistent with an STEC
infection, with earliest onset of disease on 6 May 2011. Two of the clinically ill staff worked part-
time, but included days when fenugreek seeds and sprouts were handled. The outbreak strain STEC
O104:H4, however, could only be detected in 1 of the 3 staff members. Moreover, within the course of
the investigation of the environment of Establishment A, 2 more staff members were identified
shedding the outbreak strain, both without any diarrhoeal symptoms. It was therefore questioned
whether the personnel could be the origin of the infection. While the onset of disease of the first staff
member who became ill occurred early in the outbreak, it occurred within days of the first occurrences
of illness in other people having eaten sprouted seeds originating from Establishment A. Personnel
from Establishment A were given a weekly package of sprouted seeds for private use, however, this
may explain the early exposure of personnel from Establishment A, but does not support that they
introduced the contamination into the production chain (BfR, 2011a). This does not exclude that they
may have amplified the spread of the infection in Establishment A, or that an unidentified carrier or
infected individual was responsible.
It should be noted that the tracing back investigations in Germany were done for all the types of seeds
used in Keimspross-Mischung / Milde Mischung and Wrz-Mischung. Additionally, an inventory
of all sprouted seed producers in Germany was established to gather information on the simultaneous
use of the suspected seeds. Hence, intensive sampling and microbiological analysis of seeds and
sprouts were performed. In the case of the occurrence of additional clusters, this information could
then be used to refine the identification of the source of infection, as well as, if seed contamination
were detected, a more rapid recall of all material.
The occurrence of the outbreak in France, without any direct link to the outbreak in Germany, but
caused by the same STEC strain as in the German outbreak (Gault et al., 2011) and the fact that
sprouting seeds were also strongly suspected as the source of the French outbreak (ANSES, 2011),
considerably strengthened the hypothesis that, among possible sources, seeds were the most likely
vehicle of the contamination. A cohort study of those attending the event was initiated to identify
which specific foods were significantly associated with the onset of clinical disease in the French
cluster and preliminary results revealed a significant association between consumption of sprouted
seeds (fenugreek, rocket and white mustard) and infection. In addition, the study showed that none of
the food handlers at the event or people who served the dishes had recently travelled to Germany, and
none of the guests had recent contact with someone coming back from Germany26.
Following this outbreak in France, EFSA was asked by the European Commission to support the
Member States and coordinate activities to investigate the source of the outbreaks in France and
Germany in order to allow risk managers to take the necessary and appropriate risk mitigating
measures towards prevention of further outbreaks. For this purpose EFSA set up a task force
composed of experts from the European Commission, affected EU Member States, the ECDC, WHO
and FAO, as well as EFSA staff members. Data concerning the trace back and trace forward were
26
http://www.invs.sante.fr/behweb/2011/03/pdf/n3.pdf
exchanged through the RASFF27, allowing the MS and European institutions to receive up to date
information. The study established one common link between all 41 of the identified German clusters
and the French cluster through the import of fenugreek from Egypt by a single importer in Germany.
The Importer had bought several large lots of these seeds over the last 48 months and distributed more
than one lot at the same time.
The technical report published by this ESFA task force on 5 July (EFSA, 2011b) concluded that a
specific lot (no. 48088) of fenugreek seeds imported from Egypt was the most likely common link;
however the implication of other lots cannot be excluded. The exact point of contamination in the food
chain has not been established, but took place at some point prior to leaving the Importer. The
contamination of seeds with the STEC O104:H4 strain reflects a production or distribution process
which allowed for contamination with faecal material of human and/or animal origin. Where exactly
this took place has not been established. Typically such contamination occurs during production at the
farm level, due to the use of natural fertiliser (e.g. animal manure or slurry) or contaminated irrigation
water. In general, such types of contamination lead to bigger (e.g. transnational) outbreaks, while local
contamination during storage or transport is more likely to lead to more confined outbreaks.
A sample size of 50 g (rather than the 25 g required for other matrices) is recommended to increase the
sensitivity of the subsequent diagnostic testing (ANSES, 2011). The seeds are ground then suspended
in buffered peptone water and incubated at 37C for 24 h. 5 ml of the suspension (including some
ground seed debris) are vortexed to loosen bacteria adhering to the seed debris. The suspension is
centrifuged and then the supernatant is used for DNA extraction (ANSES, 2011).
The EU Reference Laboratory (EU-RL) for E. coli has proposed a method for the detection and
identification of STEC O104:H4 in food by Real Time PCR28. The method aims at extending the
scope of the CEN/ISO Technical Specification 13136 for the detection, in food and feed, of STEC
belonging to the 5 serogroups most commonly involved in severe human disease (O157, O103, O111,
O145, O26).
According to the approach described in the CEN/ISO TS 13136, the sample is subjected to the Real-
Time PCR screening for the presence of the stx and eae genes followed by, in the stx-positive and eae-
negative samples, the detection of the O104 and H4 antigen-associated genes. Finally, a confirmation
step aiming at isolating the STEC strain responsible for the positive PCR reactions is performed. It is
noted that if a sample is co-infected with an eae positive strain and O104, the latter would be missed
by not testing for it. The Real Time PCR screening is performed on DNA extracted from enrichment
cultures obtained by incubating a portion of the food sample in an appropriate medium, as described in
the CEN/ISO TS 13136. The antimicrobial resistance characteristics of the outbreak strain are
exploited for the isolation by plating PCR-positive enrichment cultures onto media supplemented with
antibiotics. The suspected colonies are tested for stx genes and then for the serotype-specific genes
wzxO104 and fliCH4.
27
RASFF: http://ec.europa.eu/food/food/rapidalert/index_en.htm
28
http://www.iss.it/vtec/work/cont.php?id=152&lang=2&tipo=3
Microbiological investigations
After identification of Establishment A and sprouting seeds as the food vehicle for the outbreak,
further investigations were carried out to identify where STEC O104:H4 may have been introduced
into the production and/or distribution chain (the cause). This included studies to assess whether the
contamination could have come from the environment (esp. through water), the seeds used for
sprouting, personnel or by cross-contamination. Microbiological analysis of the water has thus far
proved negative (BfR, 2011a). Several batches of the different seed varieties that were used by
Establishment A and were still present at the production site were sampled for microbiological
analysis, but none were positive for STEC O104:H4. Exhaustive inspection of all parts of the supply
chain showed no likely source of infection after the sprouting at Establishment A.
A recent overview of investigations carried out at MS level and transmitted through the RASFF,
shows that a total of 10392 samples were tested, 41 samples were found to be stx positive, and 8 STEC
O104 positive (i.e. 5 from food, 1 from irrigation and processing water, and 2 from the environment,
see Appendix I). Thus far E. coli O104:H4 has not been isolated from foods where cross-
contamination from cases or infected household members could not be excluded. In total, there have
been 5 reported isolations from foods, from locations linked to cases. One was from a small remaining
part of a cucumber which had been collected from a communal organic waste bin. A second sample
was from leftover seed sprouts (one of the mixtures with fenugreek seeds produced by Establishment
A) in an open package that was collected from a communal waste bin. In both cases, the vegetable
leftovers were taken by household members out of the waste bins. A third sample was from a bell
pepper. The other 2 samples were from raw and cooked salmon. These last 3 samples were all
obviously contaminated during the incubation period in the laboratory carrying out the analysis
(personal communication).
7. Discussion
STEC strains with similarities to the outbreak strain (serotype O104:H4, Stx2 production, positive for
the genetic markers of enteroaggregative adhesion) have been sporadically isolated in Europe since
2001. The O104:H4 outbreak strain is a typical EAggEC strain that has acquired bacteriophages
encoding for Shiga toxin (Scheutz et al., 2011; Rasko et al., 2011). If it proves to be of human origin,
as for the other EAggEC, it is unlikely that it would have maintained itself in the EU population for
such a long time without being recognised as causing severe human health consequences. Instead, it
has appeared causing sporadic cases typically with a recent travel history outside the EU. Therefore,
we could assume that the O104:H4 clone, during the intervals between the sporadic cases, has
maintained itself outside the EU. If it is a pathogen with a human reservoir (rather than zoonotic), it is
conceivable that such a reservoir may be present in a population where the acquired immunity towards
EAggEC might limit the severity of the clinical manifestation of the infection, and/or the surveillance
systems in place may not be adequate to reveal the disease prevalence. If this is correct, the strain may
be imported into the EU by returning travellers or migrants, resulting in a few sporadic cases. The
Finnish and Italian cases being linked with North Africa support this hypothesis. However, the
zoonotic potential of this strain has yet to be characterized, so it remains possible that reservoirs of this
pathogen could be maintained in both wild and domestic animals, including those within the EU.
This outbreak, originating from contaminated seeds rather than, sporadic individual cases, allowed the
agent to enter into the food chain and resulted in significant exposure and a major outbreak, which had
not been the case for previous sporadic cases reported in the EU. While the epidemiological evidence
strongly points to contaminated sprouting seeds as the main vehicle for the outbreaks observed in
Germany and in France, as well as the related cases observed outside Germany, but with previous
recent travel to Germany; the most recent cases observed in Germany and in other countries may result
from vehicles or modes of transmission other than directly from contaminated seeds and sprouts
thereof. Since in Germany thousands of people were infected, it is not surprising that secondary cases
occur (Hauri et al., 2011; Kuijper et al., 2011) (see also Aldabe et al., 2011, for reporting on household
transmission in the French outbreak), or even secondary clusters when infected food handlers are
involved (see Grieg et al., 2007 for a review of outbreaks where food handlers are implicated). In
addition, it is well known that enteric pathogens for which there is a high risk of infection from low
levels of exposure cause outbreaks with person-to-person transmission, as occurs for shigellosis and,
for secondary cases, for STEC O157 (Snedeker et al., 2009).
The limited information available on the infectivity makes it difficult to quantify the role that person-
to-person transmission may play in the future. From a bacteriological perspective, a key issue that
needs to be considered is that the outbreak strain is an enteroaggregative E. coli that produces Stx, and
not a "true" STEC. However, EAggEC are also known to have caused foodborne outbreaks in Europe:
links with restaurants have been frequently reported as well as a possible role of food handlers, as in a
recent outbreak in Italy (Scavia et al., 2008).
In addition, if such asymptomatic food handlers contaminate food, as probably happened in a school in
Germany, new clusters may appear with different vehicles of infection29.
In summary, following a dramatic decrease in the number of new cases during 28 June 7 July, new
cases and clusters continued to occur despite the identification of the original vehicle (sprouting seeds)
29
http://www.kreis-paderborn.org/kreis_paderborn/presse/2011/entries/erster-todesfall-ehec.php
and the communication of public health advice to avoid consuming sprouted seed. New cases and
clusters were most probably due to secondary STEC infections, through different vehicles or by
person-to-person transmission. The emerging evidence of a substantial proportion of subclinical
infections is important as they are generally unrecognised and unreported, and may contribute to
further transmission of the infection. This may have an impact on the epidemiology of the disease and
should be taken into account when defining public health measures to prevent further spread.
Regarding seeds already imported, trace-forward activities were advised. During the period that these
activities are under way, the advice to consumers not to grow sprouted seeds for consumption and not
to eat sprouts or sprouted seeds, unless they have been cooked thoroughly30, was maintained. Once
completed the trace forward operation, and the implicated lot(s) removed from the market, such advice
should be revised.
The investigations conducted thus far have demonstrated that there was a single hazard (STEC
O104:H4) involved in the outbreak. In Germany, the clusters were all linked to sprouted seeds
produced by one establishment. The tracing back demonstrated a probable common source for the
German and French outbreaks, a particular lot of fenugreek seeds imported from Egypt.
The cause of the outbreak i.e. how, when and where the seeds became contaminated has not (yet) been
demonstrated. Assuming this took place during the primary production process, this part of the
investigation would need to extend beyond the point of EU import to include the site(s) of production.
The inability thus far to demonstrate the presence of E. coli O104:H4 in the suspected seeds is not
unexpected. It is, of course, possible that contaminated seeds were no longer in stock when sampling
took place. However, it is important to realise that a negative laboratory test does not prove the
absence of a pathogen. This is because sample-based sampling always involves some statistical
uncertainty, where there is a risk for not collecting samples of the contaminated lots or part of a lot
that is contaminated. This risk increases when the contamination occurs at a very low prevalence (i.e.
few units are contaminated). Besides the statistical uncertainty, the test sensitivity plays a role for
detecting a positive sample. The test sensitivity is a measure of how good a detection method is to
detect microorganisms in a contaminated sample. The rule is that no method is 100% effective. So
even though samples of contaminated material have been collected, it may be that the amount of
pathogens in the sample is so low that the diagnostic method is unable to detect them.
The uncertainties described above may be particularly true when dealing with seeds, because the
matrix is made up of particles, which may become singularly contaminated and dispersed in lots of
large dimensions. In addition, the contamination may be heterogeneously distributed within the food
matrix as well as, and maybe as a result of, varying physico-chemical conditions for supporting the
survival or growth of pathogens. Previous studies have shown enterobacteriaceae to be present on the
tissue surface of plants as well as internalised within various the plant tissues including seeds. This
may happen at primary production, through irrigating with contaminated water or application of
untreated or insufficiently treated fertilizer still containing enteric pathogens) (EFSA, 2011a).
The difficulties of detecting STEC in seeds and sprouts thereof with the applied sampling plans and
analytical methods when the contamination is low and unevenly distributed have been observed in
30
http://www.efsa.europa.eu/en/topics/topic/ecolioutbreak2011.htm
previous outbreaks. For example, in an outbreak of Salmonella enterica serotype Enteritidis phage
type 4b associated with bean sprouts in the Netherlands in 2000 (Van Duynhoven et al., 2002), it was
noted that the outbreak occurred even though at the implicated production facility, every new batch of
seeds and samples from each batch of sprouted seeds harvested each day was routinely tested for
Salmonella spp.
The proposed molecular analytical method is characterised by a high sensitivity. However, as seeds
are generally contaminated at very low level (0.1 to 1.8 cfu/g as previously assessed for Salmonella
(Liao and Fett, 2003)), test portions would then need to weigh more than the recommended 50g, which
may not be the case for leftover samples collected during outbreak investigation. In addition, bacteria
on or in seeds may undergo stress induced by very low water activity and therefore exhibit low
culturability during the enrichment phase which precedes the PCR step (i.e. viable but non-culturable
(VBNC) state).
Furthermore, seeds, as a biological matrix, present particular characteristics (e.g. very low water
activity, biofilm etc.) which must be considered when defining the analytical phases preceding the
PCR step. The method proposed by the EU-RL includes an annex specific for seed testing. However,
it has been recently suggested, elaborating from methods typically used in phytopathology for testing
for bacterial pathogens from seeds, to perform a preliminary step of seed soaking, to facilitate the
revivification of stressed bacteria (ANSES, 2011). Further investigations are still needed to determine
an optimal test portion of seeds, fully define preliminary analytical steps and assess the impact on the
sensitivity of the test method. Potential growth and PCR-inhibitory effects of the seeds may also have
to be taken into account.
Work has been initiated by the EU-RL to evaluate the test methods for seeds in use by coordinating a
dedicated working group and organising an inter-laboratory study on seeds naturally contaminated
with STEC other than O104. Seed lots naturally contaminated with STEC represent a valuable
material for conducting studies to evaluate and validate specific laboratory methods. Indeed, it is very
difficult to define spiking strategies that could reliably mimic the contamination naturally occurring in
field conditions.
According to the protocol for seed testing proposed by the EU-RL, the reporting of a negative result or
a presumptive positive result after the first screening step (detection of stx genes) will require 26-28
hours. The isolation of STEC O104:H4 from PCR-positive enrichment samples could require 48-72
hours. It is also possible that the positive results obtained in the PCR screening step, both for STEC in
general or specifically for STEC O104, are not confirmed by the isolation of the STEC strain. The
meaning of such a result will depend on the epidemiological setting driving the sampling: in a crisis
setting and with samples presumably related to an outbreak.
CONCLUSIONS
Initial epidemiological studies of the German outbreaks indicated fresh salad vegetables as the
probable vehicle of infection. Further, more detailed studies implicated fresh sprouting seeds.
Tracing forward and backward the dissemination of the implicated sprouting seeds demonstrated that
all of the clusters for which there was sufficient information, could be linked to a single sprouted seed
producer in Germany.
No evidence for environmental contamination was found at the premises of the sprouted seed
producer. Employees found to be infected were not thought to be the origin of the contamination due
to the comparative dates of onset of illness with those of consumers. As a consequence, and taking
into account experience from other outbreaks of enteric pathogens through consumption of
contaminated sprouting seeds, contaminated seeds were hypothesized as being the most likely source.
Fenugreek sprouts were mostly sold as mixtures of sprouts, so that at this stage, it was not possible to
identify the specific sprout (and seed) type responsible.
The STEC strain responsible (O104:H4) for both the German and French outbreaks was found to be
indistinguishable, strongly indicating that there was a common source for both outbreaks.
The comparison of the back tracing information from the French and German outbreaks led to the
conclusion that a specific lot of fenugreek seeds imported from Egypt was the most likely source of
both outbreaks, although it could not be excluded that other lots imported by the same importer from
the same exporter may be implicated.
The cause of the outbreak, i.e. how, when and where the seeds became contaminated has not (yet)
been demonstrated. Assuming this took place during the primary production process, this part of the
investigation would need to extend beyond the point of EU import to include the site(s) of production.
Prior to the 2011 outbreak, STEC O104 was a very rare serogroup in humans in the EU and
worldwide. Sporadic cases in the EU have been linked to travel to North Africa, the Middle East and
Central Asia.
Work has been initiated by the EU-RL to evaluate the test methods for seeds in use by coordinating a
dedicated working group and organising an inter-laboratory study.
The preparation of fresh sprouted seeds does not always include a step where pathogen contamination
is eliminated. The consumption of fresh sprouted seeds however, is based on the understanding that
they are sold as ready to eat. For fresh produce this assumes and relies on a production process which
prevents contamination and ability to detect contamination when it occurs. These conditions have
proven not to have been met in the case of the 2011 outbreaks.
Previous food-borne outbreaks due to sprouting seeds have led public authorities in some countries
(FSANZ, 2010) to carry out risk analyses to identify critical steps and effective measures in the
production and processing of sprouted seeds to prevent contamination of the food chain (EFSA-Q-
2011-0087731).
Prevention of further exposure to the lots of contaminated seeds and sprouts thereof
In the short term all efforts were focused on the prevention of any further exposure of the consumer to
sprouts of seeds from the lots of concern. EFSA and ECDC strongly recommended advising consumers not
to grow sprouted seed for their own consumption and not to eat sprouted seeds unless they have been
cooked thoroughly (ECDC and EFSA, 2011).
Furthermore, it was important that a trace back investigation be initiated on the incriminated lot(s) of
fenugreek seeds in the country from which they were exported to the EU. In addition, Member States and
third countries were encouraged to initiate or complete forward tracing of companies receiving the suspect
lot(s).
To complement public health measures regarding food handling, information to the public should
stress the need for proper hand washing, in addition to the specific advice related to sprouted seeds.
Since there is evidence of asymptomatic carriers of STEC in humans, screening of humans involved in
food handling is relevant. The monitoring and/or exclusion of STEC carriers from food handling
should be considered as a mitigation option.
31
http://registerofquestions.efsa.europa.eu/roqFrontend/?wicket:interface=:1::::
32
http://www.ecdc.europa.eu/en/activities/epidemicintelligence/Pages/Activities_EpidemicIntelligence.aspx
GLOSSARY
EAggEC Enteroaggregative Escherichia coli
Recent steep price increases of major crops (cereals, oilseeds) were triggered by a combination of
production remaining somewhat below trend and strong growth of demand.
A low and declining level of stocks has added to the price rise, as has probably a significant
increase in investments in agricultural derivative markets.
The OECD-FAO Agricultural Outlook expects prices to come down again, but not to their
historical levels. On average over the coming ten year period, prices in real terms of cereals, rice
and oilseeds are projected to be 10% to 35% higher than in the past decade.
The acute price hike adds to inflationary pressures in developed countries. Poor consumers in
developing countries, and food importing developing countries overall, will have to spend an
even higher share of their limited income on food.
In the short term, humanitarian aid is required, where appropriate in the form of cash or vouchers
so as to strengthen, rather than undermine domestic markets in recipient countries.
In the medium term, there is a need to foster growth and development in poor countries, to
improve the purchasing power of food buyers. Agricultural trade policies require further reform
in order to ensure an effective supply response. Investments in productivity growth, particularly
in less developed countries, should also strengthen the supply side of global agriculture. On the
demand side, policies that encourage increased production and use of biofuels from agricultural
commodity feed stocks warrant review.
Introduction
World prices of wheat, coarse grains, rice and oilseed crops all nearly doubled between the 2005 and
2007 marketing years and continued rising in early 2008. These increases in agricultural commodity prices
have been a significant factor driving up the cost of food and have led to a fuller awareness and a
justifiably heightened concern about problems of food security and hunger, especially for developing
countries.
The causes of this price spike are complex and due to a combination of mutually reinforcing factors,
including droughts in key grain-producing regions, low stocks for cereals and oilseeds, increased feedstock
use in the production of biofuels, rapidly rising oil prices and a continuing devaluation of the US dollar, the
currency in which indicator prices for these commodities are typically quoted. This turmoil in commodity
markets has occurred against the backdrop of an unsettled global economy, which in turn appears to have
contributed to a substantial increase in speculative interest in agricultural futures markets.
Tight market conditions for essential agricultural commodities pose policy challenges for national
governments as well as for international organisations. In order to take the right policy decisions, we need
to understand what caused the current price spike, what the implications may be for prices and price
volatility in the future, and how various countries and members of society may be affected. This note aims
to improve this understanding and thereby to contribute to sound policy formulation.
Figure 1. Food commodity price trends 1971 2007, with projections to 2017
700 500
Wheat Coarse Grains
450
600
400
500 350
300
400
USD/t
USD/t
250
300
200
200 150
100
100
50
0 0
1971
1973
1975
1977
1979
1981
1983
1985
1987
1989
1991
1993
1995
1997
1999
2001
2003
2005
2007
2009
2011
2013
2015
2017
1971
1973
1975
1977
1979
1981
1983
1985
1987
1989
1991
1993
1995
1997
1999
2001
2003
2005
2007
2009
2011
2013
2015
2017
1600 1200
Rice Oilseeds
1400
1000
1200
800
1000
USD/t
USD/t
800 600
600
400
400
200
200
0 0
1971
1973
1975
1977
1979
1981
1983
1985
1987
1989
1991
1993
1995
1997
1999
2001
2003
2005
2007
2009
2011
2013
2015
2017
1971
1973
1975
1977
1979
1981
1983
1985
1987
1989
1991
1993
1995
1997
1999
2001
2003
2005
2007
2009
2011
2013
2015
2017
Prices rise when markets get tight. Between 2005 and 2007 there were coinciding spells of
unfavourable weather in major producing regions in the world, pushing crop yields in those areas below
long term average levels. World cereal output in 2007 was just 3% larger than in 2005 while there was a
decline in overall oilseeds output; nevertheless, vegetable oil production rose by 7% due to rapid growth in
palm oil output.
At the same time, there was strong demand growth. Demand for wheat and coarse grains grew almost
twice as much as did production, and demand for vegetable oil increased two percentage points more than
output. More than half of the increase in use of both coarse grains and vegetable oil was due to higher use
in the biofuels industry. Use of cereals and vegetable oil for food also continued to grow, as did cereal use
for feed. This food and feed demand growth came primarily from countries outside the OECD area and
accounted for the remaining nearly 50% of the total increase in demand.
The production shortfall, relative to trend, would in itself have been enough to send prices higher,
although under normal conditions stocks would have buffered the market and dampened the price rise.
But stocks were already low and they kept declining in 2006 and 2007 because of bad weather and low
yields in major exporting countries. Supply shortfalls, the absence of a sufficient buffer, the continued
increase in food and feed use, and the high growth in relatively price-insensitive demand for biofuels all
coincided to make the price increases exceptional. More recently, there has also been a significant increase
in investments in agricultural derivative markets from non-traditional sources, whether for portfolio
diversification or speculation. It is likely that this has contributed to the rise in short term futures prices and
is an additional factor in the current spike in spot market prices.
Some of the factors behind the current price hike are transitory while others may be more permanent.
Making that distinction is an important ingredient in projecting market developments over the coming ten
year period, as is done in the OECD-FAO Agricultural Outlook. This differentiation is also important in
designing good policy to deal with the implications of the price increases.
The recent negative yield shocks in key agricultural commodity producing regions that have
contributed to price increases should be viewed as temporary. Barring any underlying climate change or
water constraints that could lead to permanent reductions in yield, normal higher output can be expected in
the very short term.
Macroeconomic conditions that favour economic growth, increases in purchasing power, and stronger
demand for agricultural commodities are expected to continue, at least for many non-OECD economies.
This is a permanent factor in future price determination, but not a new one: strong GDP growth in
developing countries has been a feature of commodity markets for many years. Thus, this factor should
slow the decline in real prices in the future, but not lift average prices to permanently higher levels.
Feedstock demand for biofuel production is expected to increase further, albeit at a slower rate than in
the past three years, and under current policy settings appears to represent a permanent factor in price
formation. Unlike strong income growth in developing countries, this is a new source of demand which is
seen as one of the factors lifting prices to higher average levels in the future.
Stocks of wheat, coarse grains and vegetable oil have fallen to low levels relative to use, reducing the
buffer against shocks in supply and demand. Stocks are not expected to be fully replenished over the
coming ten years, implying that tight markets may be a permanent factor in the period to 2017. This should
not lead to permanently higher prices, but provides the background for more price volatility in the future.
The surge of investment in futures commodity markets from non-traditional sources may have short
term price effects. But relative to the ten year outlook period these may prove temporary, given adjustment
in markets and participants behaviour: funds can move rapidly in and out of commodity markets as profit
opportunities dictate. Given their size, this may well be a new and permanent element in future price
volatility.
A more general point concerning price volatility relates to the thinness of markets, or the share of
imports and exports relative to the size of global consumption or production. When markets are thinner and
prices in domestic markets do not follow those in international trade because of insulating policies or
market imperfections, world market prices must change more to accommodate an external shock to traded
quantities, all else equal. Such market characteristics are expected to remain a permanent feature in the
volatility of prices.
Finally, the nature and composition of demand are factors that may increase the future variability in
world prices. First, industrial demand for grains and oilseeds and in particular policy-driven demand for
biofuels production is generally considered less responsive to prices than traditional food and feed demand.
Second, food demand becomes less responsive to price changes as incomes rise and the commodity share
in the food bill falls. Such changes are permanent factors that may lead to greater volatility in future world
prices.
The current outlook for crop production in 2008 is generally positive. The April forecast for global
cereal production suggests an increase of over 3% from 2007, with the bulk of the growth in wheat due to a
substantial recovery in production in major exporting countries. Rice harvests in major producing countries
are just beginning. Wheat production in major exporting countries is expected to be above 2005 levels. The
increased supply should bring some relief to the market in the course of 2008, but with use forecast to
again exceed supply and stocks forecast to be further drawn down, the overall market situation is likely to
remain tight throughout the year.
Over the 2008-2017 period covered by the OECD-FAO Agricultural Outlook, a strong combination of
supply response and continued growth in demand is expected to keep prices above historical levels, but
well below the peaks experienced today.
Demand for cereals for use as feed stocks in biofuel production is projected, under current policies, to
almost double between 2007 and 2017, but the largest part of future growth in total use is explained by
rising food and feed demand, particularly in countries outside the OECD area that are experiencing strong
economic growth. Little rice is used for feed and almost none in biofuel production. Demand for rice,
almost all for food use, is expected to increase by less than 1% per year and is dominated entirely by
growth in developing countries. Biofuel use of vegetable oils is forecast to account for more than a third of
the expected growth in vegetable oil use from 2005 to 2017, and other uses are also expected to grow
substantially. Income growth drives much of this expansion in demand, with countries outside the OECD
area increasing their consumption of vegetable oils by 50%.
Based on these market developments, the Agricultural Outlook forecasts relatively tight markets to
continue, with prices down from current peaks but remaining higher on average than prices experienced
over the past decade, as shown in Figure 2. On average over the coming ten year period, nominal prices for
cereals, rice and oilseeds are expected to be 35% to be 60% higher than on average in the past ten years.
Prices in real terms are projected to be 10% to 35% higher than in the past decade. Productivity gains and
increasing competition in trade from countries outside the OECD area will eventually overtake stronger
demand. As that happens, prices will resume their decline in real terms, though more gradually than in the
past (see Figure 1).
Figure 2. Nominal prices fall but stay above average levels of the past
900
800
700
600
US$/tonne
500
400
300
200
100
0
Wheat Maize Rice Oilseeds
The recent spikes in food commodity prices surprised most economic forecasters, reminding us of the
inherent vulnerability of projections to unanticipated developments. The price projections discussed here
assume normal weather, unchanged policies, and stable economic performance. Looking at alternatives to
these assumptions provides additional insights regarding the factors influencing future prices. Key results
of four alternative scenarios are illustrated in Figure 3 and summarised below:
If it is also assumed that oil prices stay at their 2007 level over the next decade, projected prices
for wheat and maize fall by a further 10% and for vegetable oil by a further 7%.
If the rates of growth in GDP in Brazil, China, India, Indonesia and South Africa are then
reduced to half the rate assumed in the Outlook with exchange rates and inflation also changed
consistently projected wheat and coarse grains prices would fall by a modest 1 to 2% more but
vegetable oil prices by a further 10%.
Finally, if it is also assumed that cereals and oilseeds yields are 5% higher than expected,
projected wheat and maize prices would be a further 6 and 8% lower, respectively, but there is
little further change in vegetable oil prices.
All these assumptions tend in the same direction, to lower prices, and taken together would lead to
prices for wheat, coarse grains and vegetable oils that are 20 to 35% lower in 2017 than what is now
projected. It is unlikely that these factors would all combine in the configuration described here, but it
serves to illustrate the relative significance of some of the factors that will determine future price levels. Of
course, alternative scenarios that could push prices back up could also be envisaged an equivalent
production shortfall in a major exporting country again this year, for example, would be expected to lead to
prices that remain near current levels.
-5%
-10%
-15%
-20%
-25%
-30%
-35%
-40%
Scenario 1: Biofuel production constant at 2007 level
Scenario 3: Scenario 2 + Lower income growth in EE5 countries (half annual growth rate)
Scenario 4: Scenario 3 + Yield for wheat, oilseeds and coarse grains 5 % higher than over the projection period
The impact of high food prices on developing countries depends on the interplay of various factors. In
general, commercial producers of these commodities will benefit directly from higher prices, as will in
many cases the people they employ (assuming, of course, that governments do not prevent higher prices on
world markets from being transmitted to domestic markets). Livestock producers, on the other hand, are
squeezed by both higher feed and energy costs and relatively flat prices. For farm households producing
mainly for their own consumption or for local markets insulated from price fluctuations on national and
international markets, the impacts will be mitigated. But for the urban poor and the major food importing
developing countries, the impacts will be strongly negative as an even higher share of their limited income
will be required for food. Each 10% increase in the prices of all cereals (including rice) adds nearly USD
4.5 billion to the aggregate cereals import bill of those developing countries that are net importers of
cereals.
The impact of high agricultural commodity prices on developed countries is relatively modest, overall.
The agricultural commodity price component of final food product prices is relatively small (often 35% or
less), as is the proportion of disposable income spent on food (10-15% for most OECD countries). Of
course these averages mask much more significant impacts on lower income consumers who spend a larger
share of their expenditure on food. In addition, and to the extent that high prices persist and hence do not
reduce the future rate of inflation, indirect economic impacts might also be important.
In the short term, humanitarian aid is required. Before recent price increases, hundreds of millions of
people were going hungry because they could not afford food. With higher prices, the numbers of people
suffering from extreme hunger has increased even further the 1st UN MDG has become an even greater
challenge. Immediate and substantial aid is required, where appropriate in the form of cash or vouchers so
as to strengthen, rather than undermine domestic markets in the recipient countries.
In the medium term, there is a real need to improve the purchasing power of poor food buyers so they
can acquire enough food even at the higher prices, relative to past averages, that are expected to prevail in
the future. Fundamentally that requires to foster growth and development in poor countries. In some of the
poorest countries, investment in agriculture, including in agricultural research, extension and education,
may be the best way to cut poverty and stimulate economic activity. In other situations, investment in
agriculture may also be helpful, but there may equally be a need to diversify the structure of the economy.
In many cases, investments in improving the overall environment in which agriculture operates may be
most appropriate improving basic governance systems, macroeconomic policy, infrastructure,
technology, education, health, etc. In other words, a tailored approach is needed, one that builds upon the
capacity and potential of individual countries, rather than a generalized rush to develop agriculture.
Agricultural trade policies require further reform. Trade restricting policies whether they restrict
exports or imports have undesirable and often unintended impacts, especially in the medium and long
term. Subsidies that distort markets are equally unhelpful. Export taxes and embargos may in the short
term provide some relief to domestic consumers, though such measures do not distinguish between low
and high income consumers, and they also impose a burden on domestic producers and limit their supply
response. Export restrictions contribute to global commodity market uncertainty and drive international
market prices further up. On the import side, protecting domestic producers of agricultural commodities
by providing high price support and border protection restricts growth opportunities for producers abroad
and imposes a burden on domestic consumers. A swift and ambitious conclusion of the Doha Round of
WTO negotiations could make an important contribution to exploiting the potential of markets to balance
global supply and demand.
High agriculture commodity prices also have an impact on close substitutes, such as fish, and could
contribute to even further pressure on already depleted fish stocks, as well as to increased demand for fish
from aquaculture. Policies that ensure the sustainable and responsible use of ocean resources have a key
role to play, both within national boundaries and on the high seas; concerted action to control illegal
fishing is needed. Options to improve the business environment for private investment in aquaculture
might also be explored.
The current hike in food prices is an issue of a truly global nature. It has complex causes and impacts,
and requires a complex response at the international level. Current developments on global food markets
are having dramatic implications for food security among poor people. At the same time, speculative
factors and inward looking policy actions contribute to the nervousness and volatility of markets. What is
needed now is an objective, effective and coherent global response to avoid making a difficult situation
worse.
Now, more than ever, it is important to counter growing calls for trade protectionism. Closing markets
to either imports or exports will have exactly the wrong result. More secure global food supplies will only
come from competitive producers around the world being provided the freedom to respond to current
market opportunities. Continuation or introduction of policies that create distortions and that undermine the
appropriate market responses should be avoided. The OECD will continue to monitor commodity market
developments and government initiatives over the coming months, and report periodically on its findings.
2 June 2011
UN-HLTF on
Global Food Security
6 PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES
1. Introduction
1.1 Scope
1. Under the Food Security pillar of the Seoul Multi-year Action Plan on Development, the G20
request that FAO, IFAD, IMF, OECD, UNCTAD, WFP, the World Bank and the WTO work with key
stakeholders to develop options for G20 consideration on how to better mitigate and manage the risks
associated with the price volatility of food and other agriculture commodities, without distorting market
behaviour, ultimately to protect the most vulnerable. This report has been prepared by FAO, IFAD,
IMF, OECD, UNCTAD,WFP, the World Bank, the WTO, IFPRI and the UN HLTF.
2. The approach taken in this report reflects the view of the collaborating international
organisations that price volatility and its effects on food security is a complex issue with many
dimensions, agricultural and non-agricultural, short and long-term, with highly differentiated impacts on
consumers and producers in developed and developing countries. The report begins with a discussion of
volatility and of the ways in which volatility affects countries, businesses, consumers and farmers.
Lessons learned from recent experiences are briefly reviewed as well as the factors determining likely
levels of volatility in future. This report offers suggestions for a systematic and internationally
coordinated response building on the lessons learned as a result of the 2007-2008crisis.
4. In a purely descriptive sense volatility refers to variations in economic variables over time,
(more technical definitions of volatility and related terms are put forward in Annex A) Here we are
explicitly concerned with variations in agricultural prices over time. Not all price variations are
problematic, such as when prices move along a smooth and well-established trend reflecting market
fundamentals or when they exhibit a typical and well known seasonal pattern. But variations in prices
become problematic when they are large and cannot be anticipated and, as a result, create a level of
uncertainty which increases risks for producers, traders, consumers and governments and may lead to
sub-optimal decisions. Variations in prices that do not reflect market fundamentals are also problematic
as they can lead to incorrect decisions. These implications of volatility will be explored in detail in
Chapter 2.
5. Behind concerns about volatility lie concerns about price levels and behind both, lie concerns
about food security. While producers benefit (or at least those who are net producers and whose asset
base and knowledge enable them to respond effectively), consumers, especially poor consumers, are
severely adversely affected by high prices2. Food accounts for a very high share of the total budget of the
poorest households. And because poor households often consume foods that are less processed the effect
of rises in commodity prices is felt more strongly. These households find their nutrition status (especially
of pregnant women, children and those affected by long-term diseases such as HIV), as well as their
capacity to purchase education, health care, or other basic needs compromised, when food prices are
high.
6. Producers are more concerned about low prices, which may threaten their living standards as
well as their longer term viability when income is too low to provide for the farm family or for the
6
PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES 7
operational needs of the farm. Uncertainty may result in less than optimal production and investment
decisions3. In developing countries, many households are both producers and purchasers of agricultural
products. For this group the impacts of price volatility are complex, with net outcomes depending on a
combination of many factors.4
7. No attempt is made here to define extreme or excessive price volatility. Suffice it to say that
volatility becomes an issue for concern and for possible policy response when it induces risk averse
behaviour that leads to inefficient investment decisions and when it creates problems that are beyond the
capacity of producers, consumers or nations to cope. What constitutes excessive volatility depends very
much on the situation of the individual or nation. Poor consumers in less developed countries without
access to adequate social support are most immediately affected by price surges. Small resource limited
farmers face particularly severe problems when prices fall. The episode of volatility that occurred during
the 2007-2008 period, resulted in poor, vulnerable consumers and producers and poorer developing
countries dependent on food imports experiencing severe economic, social and political stress because of
high prices and fears of scarcity. Lessons learned concerning appropriate national and international
response are instructive as we enter 2011 with many commodity prices again increasing sharply.
8. When looked at in the long term there is little or no evidence that volatility in international
agricultural commodity prices, as measured using standard statistical measures is increasing and this
finding applies to both nominal and real prices5. Volatility has, however, been higher during the decade
since 2000 than during the previous two decades and this is also the case of wheat and rice prices in the
most recent years (2006-2010) compared to the nineteen seventies.6 Another conclusion that emerges
from the study of long term trends in volatility is that periods of high and volatile prices are often
followed by long periods of relatively low and stable prices. Finally, it is well established that
agricultural markets are intrinsically subject to greater price variation than other markets, for reasons that
are outlined in the introduction to Chapter 2.
9. International commodity prices since 1970 are presented in Figure 1 and commodity price
movements during the past decade as shown in Figure 2.
Figure 1. Agricultural commodity prices in real terms Figure 2. Monthly commodity price indices
(2005=100) (2002-04=100)
IMF food Index Beef Butter Maize Rice Wheat
1000
Food Prices Meat Dairy Sugar Wheat Rice
800
900
700
800
600 700
600
500
500
400
400
300
300
200
200
100
100
0
avr.-04
avr.-05
avr.-06
avr.-07
avr.-08
avr.-09
avr.-10
oct.-04
oct.-05
oct.-06
oct.-07
oct.-08
oct.-09
oct.-10
juil.-04
juil.-05
juil.-06
juil.-07
juil.-08
juil.-09
juil.-10
janv.-04
janv.-05
janv.-06
janv.-07
janv.-08
janv.-09
janv.-10
10. Since 1990, as shown in Figure 3, the implied volatility for major crops has increased
significantly.7 Implied volatility reflects the expectations of market participants on how volatile prices
will be and is measured as a percentage of the deviation in the futures price (six months ahead) from
underlying expected value (for a more detailed explanation of implied volatility see Annex A). Broadly
7
8 PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES
speaking, increases in implied volatility reflect how market conditions and unpredictable events translate
to higher uncertainty ahead for traders and other market participants.
30
20
10
0
90 92 94 96 98 00 02 04 06 08 10
11. Irrespective of any conclusion that might be drawn concerning the long term trends, there is no
doubt that the period since 2006 has been one of extraordinary volatility. Prices rose sharply in 2006 and
2007, peaking in the second half of 2007 for some products and in the first half of 2008 for others. For
some products the run-up between the average of 2005 and the peak was several hundred percent. On the
rice market the price explosion was particularly pronounced. The price rises caused grave hardship
among the poor and were a major factor in the increase in the number of hungry people to more than one
billion.8 Prices then fell sharply in the second half of 2008, although in virtually all cases they remained
at or above the levels in the period just before the run-up of prices began. Market tensions emerged again
during 2010 and there have been sharp rises in some food prices. By early 2011, the FAOs food price
index was again at the level reached at the peak of the crisis in 2008 and fears emerged that a repeat of
the 2008 crisis was underway.
12. The trends and fluctuations described in the previous paragraphs relate to international prices.
Domestic price movements can be different. The extent to which global prices are transmitted to
domestic markets depends on how strongly integrated the latter are with the former. Measures such as
import duties, export taxes, non-tariff barriers or domestic policies such as price support all influence the
extent to which price changes in domestic markets mirror those on international markets. Market
structure is also important. In monopsonistic markets, whether private or state controlled, higher
international prices may not always result in better prices for producers Countries that insulate their own
markets export instability onto international markets, especially if they are major players in terms of
consumption or production. The degree of processing of final consumption goods also affects price
transmission. Lack of domestic infrastructure and generally undeveloped or inefficient market structures
can also significantly obstruct price transmission due to high transport and transactions costs.
8
PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES 9
13. Developing country markets often lack the capacity to absorb domestic shocks, and can be
subject to high domestic price volatility even during periods of calm international markets. Attention also
needs to be paid to volatility at local and national levels, and to its consequences for poor rural people
including small farmers. The causes may relate to climate shocks, pests or other natural calamities,
exacerbated by the fact that farmers may have poor access to technologies and generally poor
management of soil and water. Poor infrastructure, high transport costs, absence of credit or insurance
markets and various policy and governance failures may compound the initial difficulty. A relatively
minor climatic incident in these conditions can become a serious food crisis at local or regional level.
Again those most affected will be poor consumers and rural dwellers, mainly smallholders in less
developed countries or regions, heavily dependent on their own production.
14. During the 2007-2009 price spike and subsequent decline, there were quite significant
differences among regions and products in the speed and degree to which world price movements were
felt in regional or local markets. Many factors explain these differences including policy responses,
exchange rate movements, competition policy, market structure and degree of market openness.9 These
differences are important because they suggest that both price levels and degrees of volatility may differ
significantly from place to place at any given time and, therefore, that the level of hardship and
disruption being experienced may also differ. The international community needs timely and
differentiated information about the situation in different places in order to respond appropriately.
15. Are recent events random resulting from an unusual coincidence of different factors or are
there reasons to believe that the world is entering into a period of recurrent episodes of extreme price
volatility? It is not possible to have a view on the appropriate policy responses to volatility without first
exploring this question in some detail. In this context too, it is worth recalling that behind the expressed
concerns about volatility is a concern about price levels, particularly the impact of high prices on the
food security of the most vulnerable households and countries and of low prices on vulnerable producers.
16. Most agricultural commodity markets are characterized by a high degree of volatility. Three
major market fundamentals explain why that is the case. First, agricultural output varies from period to
period because of natural shocks such as weather and pests. Second, demand elasticities are relatively
small with respect to price and supply elasticities are also low, at least in the short run. In order to get
supply and demand back into balance after a supply shock, prices therefore have to vary rather strongly,
especially if stocks are low. Third, because production takes considerable time in agriculture, supply
cannot respond much to price changes in the short term, though it can do so much more once the
production cycle is completed. The resulting lagged supply response to price changes can cause cyclical
adjustments (such as the often referenced hog cycle) that add an extra degree of variability to the
markets concerned. Business cycle fluctuations in demand for agricultural non-food commodities (such
as cotton) from rapidly growing, industrializing economies may also be contributing to increased
volatility.
17. As of Spring 2011, world price levels as reflected in various measures, including the FAOs
world food price index, have once again reached the levels of 2007/08, giving rise to concerns that a
repeat of the earlier crisis is underway. Several of the same factors known to have contributed to the
2007/08 crisis are also present weather-related crop losses, export restrictions, high oil prices, and a
depreciating US dollar, against a background of a continuing tight supply-demand balance. The debate
on the impact of financial investment in commodity markets also continues. On the other hand, the
2010/11 situation differs from the earlier episode in some important respects. Firstly, the 2010 harvests in
many food importing countries in Africa were above average or very good, so that prices in the region
have been more stable. Stocks were higher at the outset which has also helped to mitigate the price rises.
Finally, the price increases have been differently distributed among commodities. Meats, sugar and dairy
9
10 PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES
products are all affected, and these are commodities that are less important in the food bills of the most
vulnerable. It should be noted also that while the index of prices for cereals has come close to its 2008
level on average, and prices of vegetable oils are also very high, contrary to the 2007/08 situation the
price rises have not affected rice. As rice is the staple food of many millions of the worlds most
vulnerable consumers, this means that the incidence of current price increases is somewhat different.
Nevertheless, there are serious risks to food security and the situation needs to be kept under close
review by national governments, and by international organisations and non-governmental agencies. .
18. Growing population and income in emerging and developing countries will add significantly to
the demand for food in the coming decades. By 2050 the worlds population is expected to have reached
about 9 billion people and the demand for food to have increased by between 70% and 100%. This alone
is sufficient to exert pressure on commodity prices. According to the latest OECD/FAO medium term
outlook projections, prices of crops and most livestock products will be higher in both real and nominal
terms during the decade to 2019 than they were in the decade before the 2007/08 price spikes. If the rate
of growth of agricultural production does not keep pace with demand, upward pressure on prices will
result. A demand or supply shock in a situation where the supply-demand balance is already tight, can,
for the reasons explained in the previous paragraph, result in increased volatility around the upward
trend.
19. The demand for food and feed crops for the production of biofuels is another significant factor.
During the 2007-2009 period biofuels accounted for a significant share of global use of several crops
20% for sugar cane, 9% for vegetable oil and coarse grains and 4% for sugar beet. Projections encompass
a broad range of possible effects but all suggest that biofuel production will exert considerable upward
pressure on prices in the future. For example, according to one study international prices for wheat,
coarse grains, oilseeds and vegetable oil could be increased by 8%, 13%, 7% and 35% respectively10.
Moreover, as long as governments impose mandates (obligations to blend fixed proportions of biofuels
with fossil fuels, or binding targets for shares of biofuels in energy use), biofuel production will
aggravate the price inelasticity of demand that contributes to volatility in agricultural prices.
20. Agricultural commodity prices are becoming increasingly correlated with oil prices. Oil prices
affect agricultural input prices directly and indirectly (through the price of fuel and fertiliser, for
example). In addition, depending on the relative prices of agricultural crops and oil, biofuel production
may become profitable (without government support) in some OECD countries. Financial investment in
commodities may also have contributed to an increasing correlation between oil and non-oil commodity
prices because of the significant share of such investment that tracks indexes containing a basket of
different commodities. High and volatile oil prices (if that is what is expected) could therefore contribute
to higher and more volatile agricultural prices, through higher input costs, higher demand for the
commodities used in the production of biofuels (sugar, maize, vegetable oils), through competition for
land with commodities that are not used directly for the production of fuel, and possibly through
financial investment in commodity baskets.
21. Low stocks relative to use, and uncertainty about stock levels in some parts of the world
contributed to the 2007/2008 price spike. Stocks can be drawn down in response to a supply or demand
shock, but once they have been depleted, supply can no longer be increased until new production comes
on board. Even expectations of depleted stocks may lead prices to rise sharply. The low stock levels
observed in recent years have been attributed to the partial dismantling of price support and intervention
purchase schemes in some OECD countries, as well as to correction of the quality of information on
private and government held stocks in important producing and consuming countries. Stocks were rebuilt
during 2009 and the first part of 2010 but currently stocks are again being depleted. If stock levels
10
PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES 11
remain low in major markets, and projections based on existing knowledge of market conditions and
policy settings suggest that they may, the risk of volatility in prices will remain high.
22. Climatic factors have indisputably contributed to the price rises in 2007/2008 and again in
2010. In 2008, an already tight market situation for wheat was aggravated by drought in Australia, which
is an important supplier of wheat to world markets. Canada, another important supplier, also experienced
weather related low yields for several crops. More recently, drought followed by fire in the Russian
Federation, fears about the Australian and Argentinean crops, and several downward revisions of US
crop forecasts in late 2010 and early 2011 have brought strong market reactions and soaring prices. It is
not clear whether these weather-related events are transitory in nature, cyclical (El Nino and La Nina) or
the harbingers of long term climate change. Experts concur broadly that climate change will, in the
longer term, lead to worsening conditions in some arid and semi-arid regions where agricultural
production is already difficult, while temperate regions in particular, but not exclusively, may benefit. It
is also thought that climate change will lead to more frequent extreme events such as droughts, heat
waves and floods. Clearly, climate change will provoke some adjustment of production patterns around
the world, as well as increased risks of local or regional supply problems that could add to future
volatility.
23. Stronger demand for food crops and animal products in conjunction with slow growth in
agricultural productivity and low stocks results in upward pressure on prices. Recent years have also seen
some shift in production patterns, particularly of food and feed grains, and world markets are more
dependent on supplies from the Black Sea region and other, newer, agricultural production regions than
in the past. Yields in these regions are less stable and supply more variable than in some other parts of
the world where natural conditions are better and where application of the most up-to-date technologies
and management practices have increased and stabilised yields. As the geographical distribution of
production changes, supply may therefore become more variable, in turn leading to increased price
volatility.
24. The same underlying factors that are leading to increased demand for food growth in
population, affluence leading to increased demand for animal protein, urbanisation, and biofuels are
also increasing pressure on finite resources such as land and water. While such resource constraints are,
thus far, more local than global in nature, growing concern is evident and the associated uncertainty may
imply upward pressure on prices and continuing or increased volatility.
25. During the 2007-2008 period, some policy measures put in place by a number of governments
contributed directly and indirectly to the crisis (export restrictions, hoarding), increasing the amplitude of
price movements and in some cases provoking price increases that were otherwise inexplicable in terms
of the market fundamentals. Inappropriate policy responses also contributed to volatility and could
continue to do so unless the international community is able to take steps to avoid such actions.
Additionally, private and public actors responding to the general nervousness of the markets, or for
speculative reasons, engaged in hoarding or precipitated purchases in an already tense market situation.
How to avoid repetition of these types of damaging private and public reactions is addressed later in this
report.
11
12 PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES
beyond the scope of this report but if the future is marked by increased exchange rate volatility this will
also have repercussions for the volatility of international prices of commodities.
27. There is no doubt that investment in financial derivatives markets for agricultural commodities
increased strongly in the mid-2000s, but there is disagreement about the role of financial speculation as a
driver of agricultural commodity price increases and volatility. While analysts argue about whether
financial speculation has been a major factor, most agree that increased participation by non-commercial
actors such as index funds, swap dealers and money managers in financial markets probably acted to
amplify short term price swings and could have contributed to the formation of price bubbles in some
situations. Against this background the extent to which financial speculation might be a determinant of
agricultural price volatility in the future is also subject to disagreement. It is clear however that well
functioning derivatives markets for agricultural commodities, could play a significant role in reducing or
smoothing price fluctuations indeed, this is one of the primary functions of commodity futures markets.
This topic will be taken up in more detail in Chapter 3.
28. This catalogue of factors points to a likelihood of higher real prices and a risk of increased
volatility in future years. While it is not possible to forecast future prices or future returns, there may be
ways in which the international community could be alerted to a risk that a period of excessive volatility
is in the offing.11 Various tools and mechanisms that could assist in this respect are described in later
sections.
29. There are also a number of factors that could act to mitigate increased price pressure and
increased volatility. Most important will be the supply response? Analysts have consistently
underestimated the capacity of producers to respond to positive market signals, as well as the potential
for higher yields and increased acreage. Successful conclusion to the WTO Doha Development Agenda
negotiations would be an important step, along with complementary policies that improve supply
capacity and ensure the benefits of open and competitive markets are widely spread. Deeper integration
of global and regional markets, better defined safeguard mechanisms and improvements in the
competitive environment will bring increased trade volume and more suppliers and buyers to markets
that are currently very shallow. Local or regional supply shocks could more easily be absorbed leading to
lower volatility on domestic and international markets and food could more easily flow from surplus
areas to rapidly urbanising food-importing areas
30. The extent of potential future increases in prices and volatility cannot be estimated accurately,
but the risks are sufficiently large to warrant serious reflection about what can be done to mitigate it
when the nature of the underlying causes makes mitigation possible and to manage the consequences,
when, as is inevitable, episodes of high volatility occur. The remainder of this report takes up this crucial
topic.
31. At the macro level it is useful to distinguish between long and short run effects of commodity
price volatility and between importing and exporting countries. The assumption is that countries likely to
be most concerned by macro-economic impacts of agricultural price volatility are developing or
emerging economies that are dependent on agricultural commodities for a large share of their export
revenues, or whose food imports are significant in balance of payment or government finance terms. For
exporting countries heavily dependent on agricultural commodities, exceptionally low prices will have
immediate balance of payments impacts, but beyond that, uncertainty may curtail investment and affect
capacity utilization and there is some evidence of long-lasting significant negative effects on growth.
Importing countries faced with exceptionally high prices may also experience deterioration in the balance
of payments and deterioration in their public finances. Food price increases can have major repercussions
on the whole economy. For low-income food-importing countries, high food prices can result in inflation
12
PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES 13
and high import bills which in turn worsen the current account balance. As countries have to export more
to pay for imports, such deficits may result in the depreciation of the exchange rate. Fiscal measures,
such as cuts in import tariffs and in taxes on food, subsidization of food consumption, and increased
demands on risk management instruments entail increased budgetary costs that will have to be met by
increased government borrowing and budgetary discipline.
32. Looked at from the demand side, significantly higher food prices are disastrous for the poor
especially in developing countries where up to three-quarters of their total income may be spent on basic
foodstuffs. Immediate impacts are obvious, but there are also longer term costs imposed on the poorest
and most vulnerable as spending is switched to less nutritious foods and away from other basic needs
such as education or health. Typically the effects are felt more strongly by women and children.
Particularly severe are the effects on children stunting and cognitive loss often occurring as a result of
inadequate nourishment during the first 1 000 days after conception. The consequences are tragic for
individuals and for future prosperity in the countries where they live. This is not only an issue of concern
for the affected countries but also for humanitarian agencies and the international community.
33. Food price inflation can also be a serious issue in middle income countries, where many
consumers expend as much as half of their budget on basic foods. Even in the developed countries
significantly higher food prices can create hardship for the least well-off, who tend also to devote a larger
share of household spending to food. Nevertheless, consumers in developed countries face wider choices
in terms of their ability to adjust spending on different types of foods and most developed countries have
safety net mechanisms that are well suited to delivering targeted assistance to the most affected.
34. Looked at from the supply side, high prices benefit net producers of these commodities and
signal a need for increased production. Price increases affect mainly grains and oilseeds which is a high
proportion of total costs in intensive production systems. Profitability of livestock enterprises will be
affected especially if these costs cannot be fully passed on to consumers. Volatile feed prices are also
problematical for livestock producers; such uncertainty is detrimental to investment and production
decisions, particularly where the physical production cycle is long.
35. Low or volatile prices pose significant problems for farmers and other agents in food chains
who risk losing their productive investments if price falls occur while they are locked into strategies
dependent on higher price levels to be viable. Farmers who have already planted their crop are the classic
example. Poor smallholders who do not have access to credit may have difficulty financing the crucial
inputs needed to plant again and stay in business. This kind of problem may be particularly severe for the
female smallholders who are in the majority in many countries. Many farmers in developing (and even
some in more advanced) economies may not be operating on a sufficiently large scale to be able to carry
over income from one season to another. Thus, both the welfare of the family and the viability of the
farm may be threatened by excessive volatility. Uncertainty may also result in sub-optimal investment
decisions in the longer term.
36. It is beyond the scope of this paper to undertake an exhaustive account of the ways in which
national governments and international institutions responded to the price volatility during 2007-2008. It
is generally agreed, however, that policy responses were mainly ad hoc in nature, that some decisions
were taken hastily, and that measures were somewhat inconsistent and largely uncoordinated at
international level. The speed and strength of the price rises also took the international agencies by
surprise and, they too, had to resort to ad hoc measures. Developed countries relied mainly on already
existing safety net mechanisms while developing countries took new measures or adjusted the parameters
of existing instruments.
13
14 PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES
37. Of 81 developing countries surveyed by the FAO, 43 reduced import taxes and 25 either
banned exports or increased taxes on them.12 A large number of developing countries implemented
measures to provide relief or partial relief from high prices to consumers 45 in all. Measures consisted
of cash transfers, direct food assistance or increases in disposable income (by reducing taxes or other
charges), or some combination of these measures. A significant number of countries also granted support
to producers in order to offset rapidly rising input costs, as prices for fertilizer also surged as did feed
costs for livestock producers. Several countries went to the international markets to procure supplies of
basic foodstuffs for their populations, believing that high prices would persist and that scarcity was
imminent, notwithstanding the fact that they did not have any immediate or short term need to do so.
38. The extremely rapid run-up in food prices eroded the capacity of the national and international
relief organisations to purchase food in the most hard hit countries and regions. With prices doubling or
tripling within a few months, their purchasing power was dramatically reduced. While response to
appeals made, for example, by the World Food Programme were both rapid and generous, crucial weeks
and months were lost as international organisations and humanitarian NGOs scrambled to raise funds or
divert monies from other uses to address the crisis. This situation revealed deficiencies in international
readiness to deal with such a widespread problem.
39. The events of 2007-2008 also revealed serious deficiencies in the quality of the information
base, and in particular concerning short-term forecasts and the level of stocks. More timely, complete and
accurate information and improved capacity to identify and analyse early warning signs might have
calmed the markets, re-assured populations and resulted in better readiness.
40. The different measures taken by individual governments in response to the crisis had different
degrees of effectiveness.13 The scale of the price increases was such that for many countries reducing
import tariffs had a relatively modest impact because the initial tariffs were low or the scale of the price
increases was so large. In any event, this instrument was quickly exhausted as tariffs were reduced to
zero. Some of these countries suffered steep falls in tariff revenues and deterioration in their fiscal
situation. Export taxes and restrictions differed between countries in their effectiveness in keeping
domestic prices lower and in some cases had only a relatively minor effect. Export restrictions by major
food exporters had strong destabilising effects on international markets. As more countries followed the
first movers, volatility was exacerbated and the upward price movement was amplified. Export
restrictions proved extremely damaging to third countries, especially the poorest import dependent
countries, and to the efforts of humanitarian organisations to procure supplies, despite various ad hoc
exemptions and exceptions which were put in place in order to mitigate the worst of these beggar thy
neighbour effects.
41. Targeted assistance to those most in need, either using cash transfers or direct food assistance,
may be the most effective and equitable way of reaching those affected by a food price crisis and several
countries have successfully used this kind of instrument. However, many countries did not have the
administrative frameworks in place to be able to implement safety-net measures at short notice. Neither
did they have the fiscal capacity. They therefore made blanket market and trade interventions that
sometimes proved ineffective or costly or both. Such measures, when they delivered some relief did so
irrespective of need. This revealed the importance of contingency planning to better equip countries to be
able to deliver targeted assistance where it is most needed.
42. Estimated numbers of hungry people in the world rose from 820 million in 2007 to more than a
billion in 2009, proof that neither national nor international responses were able to fully cope with the
scale of the problem. Deficiencies in information, communication, and in readiness contributed, as did
uncoordinated measures that may have actually aggravated the problem for people and countries less
well able to cope. The numbers of hungry people have since dropped to about 900 million. These events
have drawn increased attention to the fact that a significant proportion of humanity (about 16%) remains
14
PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES 15
chronically under-nourished, even during periods of relatively normal prices and low volatility. The
overarching goal of actions with respect to food price volatility should be to ensure that the most
vulnerable people have access to sufficient, nutritious food. All policy interventions should have as their
ultimate aim, the elimination of all food insecurity, whatever its cause.
43. Sections 4 and 5 of this report address policy solutions that aim to reduce price volatility and to
deal with its consequences, particularly for the most vulnerable and food insecure, respectively. A key
element in any long term solution is investment in increasing the productivity and resilience of
developing country agriculture. This can contribute to improving food security in two ways. It can reduce
food price volatility, for example through increased productivity and improved technical management of
production and of risk, and it can help farmers and households to cope better with the effects of
volatility, once it occurs. The set of recommendations put forward here, if implemented, would probably
constitute the single most important contribution to an enduring solution to global food insecurity. While
the benefits would accrue in the longer term, actions are needed immediately.
44. Agriculture is a source of livelihood for about 86% of rural people 1.3 billion smallholders
and landless workers it provides farm-financed social welfare when there are urban shocks, and a
foundation for viable rural communities.14 Long run projections are fraught with difficulty and estimates
can vary widely. According to FAO, the rate of growth in agricultural production is expected to fall to
1.5% between now and 2030 and further to 0.9% between 2030 and 2050, as compared with 2.3% per
year since 1961. Population estimates suggest that by 2050 the planet will be home to 9.1 billion persons,
up from the current population of 6.8 billion. This represents a 34% increase over the next 41 years.15
These particular estimates suggest that in the future, with the supply of food not growing at the same
pace with demand, upward pressure on prices could be a principal attribute of world food markets. In
addition to high price levels, shocks, due to climatic or other reasons, can create wide price movements,
as the food market may lack the capacity to absorb them. This adds to vulnerability and underlines the
importance for supply to keep up with growing demand.
45. Investing in agricultural productivity growth and resiliency in low income countries is
paramount to addressing local food price volatility. FAO estimates indicate that agricultural production
would need to grow globally by 70% over the same period, and more specifically by almost 100% in
developing countries, to feed the growing population. In the medium and longer term only investment in
developing countries agricultural sectors will result in sustainable increases in productivity, healthy
markets, increased resilience to international price spikes and improved food security. Investments in
infrastructure, extension services, education, as well as in research and development, can increase food
supply in developing countries and improve the functioning of local agricultural markets, resulting in less
volatile prices. In this way, markets can work for the poor people who bear the burden of food price
volatility.
46. Waste, due to post harvest losses, inadequate storage and infrastructure as well as under-
developed markets in general are a huge issue in agriculture in developing countries, amounting to a
significant proportion of production. The investments proposed here would contribute to reducing or
even eliminating avoidable waste of this type and in so doing would provide a significant boost in the
quantity of food actually reaching consumers. The increase in production needed to meet the anticipated
future demand would therefore be less than what is currently estimated. Similar efforts to avoid waste
which in the developed countries occurs mainly at retail level and in homes and restaurants would also
have large beneficial effects. This aspect is discussed in section 4.5.
47. The investments required in developing countries to support this expansion in agricultural
output amount to an average annual net investment of USD 83 billion (in 2009 USD).16 This total
15
16 PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES
includes investment needs in primary agriculture and necessary downstream services such as storage and
processing facilities, but does not include public goods like roads, large scale irrigation projects,
electrification and others that are also needed.
48. Most of the investment, both in primary agriculture and downstream sectors, will have to come
from private sources, primarily farmers themselves purchasing implements and machinery, improving
soil fertility, etc. For a better functioning agricultural system and improved food security, three kinds of
public investments are also needed.
49. Farmers and prospective farmers will invest in agriculture only if their investments are
profitable. Many types of public goods, such as the above mentioned, that make private investments
financially viable can only be provided by the governments. Private sector investment also needs to be
encouraged at all stages in the value chain upstream of the farm, in seed and fertilizer production and
distribution, and downstream in processing, marketing and distribution. Underlying competition
problems that have led to the development of cartels or of monopsonistic/monopolistic market structure
should also be tackled.
50. However, the capacity of the poorer developing countries to fill the investment gap is limited.
The share of public spending on agriculture has fallen to an average of around 7% in developing
countries, even less in Africa, and the share of official development assistance going to agriculture has
fallen to as little as 3.8%. Commercial bank lending to agriculture in developing countries is also small
less than 10% in sub-Saharan Africa. Private investment funds targeting African agriculture are an
interesting recent development but current investments are still small.17
51. Investments in agricultural research and development have been shown to have very high rates
of return and have an important role to play in fighting hunger and poverty. Bridging the gap between
research and development in the main cereals and staples that are most important for small farmers in
regions with high prevalence of hunger is an important challenge.
52. The agricultural sector is very green-house-gas (GHG) intensive: it accounts for about 13%-
33% of global GHG-emissions, but only for about 4% of global output. Agriculture will therefore be
called upon to contribute significantly to mitigation. At the same time, agriculture will be affected in
ways that are not fully understood or fully predictable, but there is little doubt that some regions,
principally arid and semi-arid zones, will come under increasing pressure. Climate change will lead to
more frequent extreme events such as droughts, heat waves and floods. These incidents will affect not
just production and the volatility of production, they may also create new difficulties related to water
quality, storage and related food safety issues. Complex demands for mitigation and adaption will
therefore be made on the sector during a period when significantly increased production is needed in
response to projected needs. In-depth research on the link between climate change and agricultural
production will be needed and would best be undertaken in coordination with the IPCC, while building
agricultural resilience through increased funding for climate change adaptation measures in developing
countries will be extremely important for example, under the newly created Green Climate Fund.18 19
16
PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES 17
53. Agricultural research is increasingly being delivered by the private sector with technologies
being developed for larger, commercial farming operations. The adoption of such technologies requires
increased management skills and effective learning, so that small farms too can have access to innovative
inputs. There is need to improve agricultural technologies specific for, and well targeted to small-scale
agriculture and for appropriate production policies and practices aimed at increasing smallholder
productivity in a sustainable manner.20
54. At present, much public research is carried out by the International Research Centres of the
Consultative Group on International Agricultural Research (CGIAR). There is general recognition of the
utility and benefits provided by this system of international research bodies and affiliated organizations
which have enormously contributed to the global pool of available agricultural technology and
knowledge. A new CGIAR Multi-Donor Trust Fund is established to harmonize donor investments in
key global challenges on agriculture and is being hosted and managed by the World Bank. New results-
oriented research programs focus on climate change mitigation and adaptation policies and technologies
and include a broad group of partners. There is a need to increase and sustain the financing of such
bodies in order that they may continue to invest today in the techniques and innovations that will be
needed to deal with the food security and climate challenges that have been defined elsewhere in this
report.
55. Increasing public investment in transport and productive infrastructure, as well as in human
capital, is central in stimulating productivity and reducing post-harvest wastage. Improving
infrastructure, in particular rural roads and market facilities such as warehouses, storage facilities and
market-information systems are important in reducing transport costs and integrating smallholders to
markets. Investing in, and improving irrigation facilities, and market institutions and mechanisms will
result in increased quantities of food produced, better quality and more stable prices.21 Improving
extension, education and health, targeting small producers but also other value chain actors, are key
elements of a sound policy approach to increase productivity and enhance food security and the well-
being of farmers. Annex B to this report contains a more complete treatment of the role of smallholders,
describing their role in production and consumption, how they are impacted by volatility and further
developing some of the policy recommendations made here to apply more specifically to small scale
agriculture.
56. All these responses to increase the resilience of agriculture and stabilize prices require public
interventions. Government expenditure on agriculture can have a significant positive impact on
productivity. Foreign direct investment also has a positive impact on productivity growth, but only if
carried out responsibly in combination with efficient bureaucracy, a lack of corruption, and democratic
political structures.22 23
57. More and better support for public investment in agriculture public goods will allow private
sector actors, including smallholders and small-scale market agents, to respond more profitably to rising
prices, both increasing local food supply and boosting the incomes of the poor.
58. Priority interventions include support for generation, adaptation, and adoption of improved
technology; improved agricultural water management, tenure security and land markets; and
strengthening agricultural innovation systems. Not only must there be far more investment in public
goods in these areas to facilitate smallholder and other private sector supply response, but investment
must be better.
59. The Global Agriculture and Food Security Program (GAFSP), launched in April 2010, provides
an important avenue for public investment. The GAFSP has pledged USD 925 million from a number of
donors. To date, investment programmes are assisted by the World Bank, the African Development
Bank, the Inter-American Development Bank, the International Fund for Agricultural Development
17
18 PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES
(IFAD) and FAO. However, there are unfunded country proposals to GAFSP, amounting to
approximately USD 800 million.
Recommendation 1
G20 governments commit to take comprehensive action to strengthen the longer term productivity,
sustainability and resilience of the food and agriculture system world-wide, encompassing several elements.
Improve food and agriculture innovation systems, encompassing public and private investments in scientific
research and development, technology transfer, and education, training and advisory services and ensure that
successful practices are scaled up.
Strengthen the CGIAR system to support technological innovation and global dissemination of technology, in
particular to improve productivity performance in less developed countries taking into account the needs of
smallholder and especially women farmers.
Support the development of technologies and provide the appropriate incentives to address challenges specific
to climate change and sustainable resource use (land and water).
Increase public (ODA and national governments) investment in developing country agriculture, and in activities
strongly linked to agricultural productivity growth, such as agricultural institutions, extension services, roads,
ports, power, storage, irrigation systems and information and communication technology, where appropriate.
link public investment to the provision of sustainable public-private-civil society partnerships.
Support comprehensive national food security strategies that are country-owned and led, evidence-based and
inclusive of civil society and farmer organizations. In this respect, follow up on previous G 20 commitments,
such as the Pittsburgh summit commitment, to fund the Global Agriculture and Food Security Program.
Provide the enabling environment for farmers and other private sector actors to scale up investments, above
and beyond ODA and national government spending, to achieve the increased productivity and enhanced
resilience on which long term food security will depend. To elicit the needed level of private sector investment,
less developed countries in particular will need to support introduction of effective governance systems and
institutions, stable macroeconomic conditions, sound structural policies, human capital development and public
services.
Annex C to this report contains selected project descriptions furnished by the international organisations, to
illustrate the kinds of practical, action-oriented initiatives that are needed in order to implement the
recommendations made here.
60. There are many factors that contribute to high and volatile agricultural prices, making
necessary a combination of policy responses. In order to meet their objectives, policies need to be
legitimate and broadly owned by relevant stakeholders, particularly those policies that aim to restore trust
in markets and avoid panic-driven behaviour. The goal of the policies recommended is not to eliminate
agricultural price volatility, but rather to reduce uncertainty, and perhaps also the amplitude of variations
by smoothing out the extremes. Most importantly, price volatility should reflect market fundamentals as
accurately as possible and not convey incorrect signals as a result of missing or wrong information,
speculation, panic or other disruptive factors.
61. A lack of reliable and up-to-date information on crop supply, demand, stocks and export
availability contributed to recent price volatility.24 Information on the current situation and outlook for
global agriculture shapes expectations about future prices and allows markets to function more
efficiently. Lack of accurate information on market fundamentals may reduce efficiency and accentuate
price movements.25 At a regional level (mainly in Africa, but also in Haiti, Afghanistan and some Central
American countries), a few successful efforts, such as the Famine Early Warning System Network, have
18
PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES 19
increased the availability of information to governments and market participants. Better information and
analysis of global and local markets and improved transparency could reduce the incidence and
magnitude of panic-driven price surges.
62. Recent events have revealed weaknesses in the capacity of nations and international
organizations to produce consistent, accurate and timely agricultural market data and analysis, especially
in response to weather shocks. Action is needed to increase capacity to undertake more frequent and
systematic monitoring of the state of crops, and to develop mechanisms for improved short-run
production forecasts, able to translate crop growth, meteorological and remote sensing data into yield and
production expectations. Greater use could be made of satellite data and geo-information systems and, in
this context, international co-ordination and exchange of technologies and information could be
enhanced.
63. Information on stocks is an essential component of a global food market information system,
yet reliable data on stocks of grains and oilseeds are often not collected or, if collected are not reported
publicly. The reasons for poor stock data are multiple: some countries no longer hold public stocks
because the policy measures that created them have been removed or reformed; stocks can be very
dispersed among farmers, traders and other actors and difficult to track; and some information on stocks
is commercially sensitive. Generally, international agencies estimate net changes in stocks as a residual
on the basis of data on production, consumption and trade. As a result it is not possible to have complete
confidence in world food stock estimates. International cooperation could redress this situation and
ensure that reliable information on global stocks becomes widely available. This would in turn better
inform market participants and help avoid mis-informed panic-induced price surges. A first step could be
an audit and assessment of available information, identifying gaps and proposing ways in which they
could be filled.
64. Monitoring food prices, both on cash and futures markets, is essential in a food market
monitoring system. In a similar manner, assessing changes in oil prices and analysing their impact on
food markets is important.26 Better information about domestic price movements is necessary to better
understand how international price changes affect domestic markets in developing countries. Such
information is important for early warning systems, such as the FAO Global Information Early Warning
System and for Vulnerability Analysis and Mapping of WFP.27 It is also crucial for policy making and
designing effective risk management instruments for developing countries.
65. For developing countries, enhanced market information and early warning systems would
enable both governments and the private sector to plan ahead. Governments would be able to more
accurately assess needs, make budgetary provision for producer and consumer safety nets and better
position emergency food security reserves. Better market information and analysis could reduce
uncertainties and assist producers, traders and consumers to make better decisions.
66. Over the last decade a great deal of baseline information on food security vulnerability has
been developed. WFP support of national Food Security Monitoring Systems already provides a
monitoring and decision support tool to help governments manage and respond to risk related to price,
weather or other hazards. The reliability and timeliness of such early warning systems needs to be
improved, and capacity to develop and utilise them could be strengthened at both the national and
regional level. The focus should be on countries that are particularly vulnerable to price shocks and food
emergencies.
67. The experience of the 2007-08 food price crisis and the current excess price volatility in many
international food markets have exposed weaknesses in relation not only to the provision of market
information at the global level but also to the coordination of policy responses to food price volatility.
There is need to ensure better preparedness and more rapid and consistent policy responses in times of
19
20 PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES
crisis. Building on and complementing existing systems, improvements in global market information and
policy guidance could be achieved through a collaborative food information and policy initiative, the
Agricultural Market Information System (AMIS). Such initiative would improve data reliability,
timeliness and frequency, as well as enhance policy coordination in times of crisis.
68. AMIS could be built on the model of JODI (the Joint Oil Data Initiative), launched in 2000 to
improve information about oil markets. However, it would have the additional function of issuing global
food price surge alerts and promoting policy coherence. AMIS would involve the major food producing,
exporting and importing countries. It would also involve a joint Secretariat comprising of international
organizations with capacity to collect, analyse and disseminate information on a regular basis regarding
the food situation and outlook, as well as food policies.
69. The structure of AMIS would include two groups to effectively perform two important
functions: a Global Food Market Information Group would be responsible for food market information
collection and analysis, while the promotion of international policy coordination would be the objective
of a Rapid Response Forum.
70. Increased and regular exchange of information and collaboration between market experts from
participating countries and organizations in the AMIS Global Food Market Information Group could
result in more complete and reliable data on consumption, production, trade and stocks, increasing
market transparency and curbing food price volatility that is not based on underlying market conditions.
71. Through the comprehensive coverage of global major food markets and the close monitoring of
prices in combination with food security assessments across vulnerable countries AMIS will also provide
a mechanism for global early warning. This will increase the scope for more automated systems of
evaluating food security implications of changing market situations whereby an indicator of different
degrees of severity can be calculated routinely and where appropriate trigger an alert.
72. The AMIS Rapid Response Forum would provide policy advice and promote policy
coordination when the market situation and outlook indicates a high food security risk. Through the
participation of policy experts from the major producing and importing countries AMIS Rapid Policy
Response Forum will be able to mobilise political support to achieve agreement on appropriate policy
response and actions in times of crisis.
73. The Rapid Response Forum will meet in response to a food crisis alert. Its actions would be as
follows:
receive and assess information and analyses from the AMIS Secretariat on the current global
market situation and outlook and issue regular statements on the ensuing implications for food
security; receive information and assessments for particularly vulnerable countries.
provide appropriate policy guidance and promote policy coordination when the market
situation and outlook indicates a high food security risk. Such guidance will encourage the
implementation of efficient and effective policies, the avoidance of potentially damaging policy
choices, and will ensure that humanitarian responses are rapid and appropriate.
work closely with the Committee on World Food Security (CFS) to promote greater policy
convergence and strengthen policy linkages at global level.
20
PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES 21
Recommendation 2
Building upon existing mechanisms, establish an Agricultural Market Information System encompassing four elements.
G20 governments commit to instruct statistical or other relevant agencies to provide timely and accurate data on
food production, consumption, and stocks. Where the mechanisms and institutions are not in place nationally to
do so, G20 governments should undertake to create them.
International Organizations, with broad involvement of countries (G20 and other relevant players) commit to
undertake monitoring, reporting and analysing of current conditions and policy developments in major markets as
well as to enhance global food security by encouraging information sharing, improving data reliability and
increasing transparency, and introducing a global early warning system.
G-20 governments support the establishment of a Rapid Response Forum, with broad involvement of countries
(G20 and other relevant players) building on the proposed Agricultural Markets Information System to promote
policy coherence and coordination in times of crisis.
International Organizations support the improvement of national or regional systems to monitor stocks,
production, forecasts (with improved modelling and weather forecasting), food and nutrition security and
vulnerability, in order to enhance Early Warning Systems in vulnerable developing countries and regions.
Concrete proposals on the implementation of this Agricultural Market Information System are detailed in a
comprehensive scoping proposal made available as a separate document.
74. Buffer stocks attempt to influence prices rather than to provide emergency relief in a crisis. At
the international level buffer stocks have been an important characteristic of commodity markets in the
past. However, the various international commodity agreements which provided for stockholding or
supply controls to stabilise prices have either collapsed or been replaced by agreements whose main role
is market information provision.
75. Historically, international buffer stock mechanisms are widely judged to have had limited
success in reducing the volatility of prices. They have been more effective in moderating downward price
movements than price surges. In the case of a price surge, a buffer stock agency can only release in the
market what it has previously bought, and once its stock is exhausted there are no further means to curb
price increases.28
76. Attempting to stabilise prices using buffer stocks is potentially very costly. Stabilising world
prices around a level either lower or higher than that determined by market fundamentals requires
significant resources. Attempts to defend a price ceiling and reduce the average world level of food
prices over time can lead to substantial costs. Buffer stocks set to defend against price spikes are also
vulnerable to speculative attacks. If speculators perceive that the stocks held by the stabilization agency
are insufficient to maintain the target lower price level, they will compete to buy the entirety of the stock
in order to take advantage of likely profits.29
77. Futures markets perform several functions: they provide the instruments to transfer price risk,
they facilitate price discovery and arguably, increasingly in recent years they are offering commodities as
an asset class for financial investors, such as fund and money managers who had not previously been
present in these markets.
78. Commercial participants utilize futures contracts to hedge, or insure their crops or inventories
against the risk of fluctuating prices. For example, processors of agricultural commodities, who need to
obtain inventories, buy futures contracts to guard against future price rises. If the price rises, they use the
21
22 PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES
increased value of the futures contract to offset the higher cost of the physical quantities they need to
purchase.
79. Speculators also trade in the futures markets; they buy and sell futures contracts and take on the
risk of future price fluctuations to gain a risk premium. They are non-commercial participants as they
have no involvement in the physical commodity trade in contrast to commercial participants, such as
farmers, traders and processors.30
80. Since the beginning of the last decade, commodity derivative markets, including those for
agricultural commodities, have experienced significant inflows of funds from non-traditional investors,
such as commodity index funds, swap dealers and money managers. These financial investors hold large
futures positions including in basic food commodities such as wheat, maize and soybeans as well as in
cocoa, coffee and sugar.
81. Another essential function of futures markets is to facilitate price discovery. Price discovery is
the continuous process by which futures prices are reassessed by buyers and sellers as new information
becomes available. Market participants continuously update their expectations as both public and private
information become available. They adjust their market behaviour and through their transactions,
information is incorporated into the price.
82. Speculators are necessary for the performance of both these functions. They buy and sell
futures contracts and take on the risk of price fluctuations to earn a profit on price movements. By doing
so, they provide the market liquidity which enables commercial hedgers to find counterparties in a
relatively costless manner. Too little non-commercial participation results in low liquidity and potentially
in large seasonal price swings. 31 Too much non-commercial participation can cause frequent and erratic
price changes. This is the case when speculators assume that past developments carry information on
future price movements, giving rise to trend chasing. This will result in buying after prices rise and
selling after prices fall, independently of any changes in market fundamentals.32
83. The debate on whether speculation stabilizes or destabilizes prices resumes with renewed
interest and urgency during high price episodes. Some analysts purport that the influx of financial
investors in commodity futures markets has scant impact on market prices.33 Other analysts stress that the
large amount of money invested in commodity futures by financial investors has amplified price
movements to an extent which cannot be explained by market fundamentals.34 More research is needed
to clarify these questions and in so doing to assist regulators in their reflections about whether regulatory
responses are needed and the nature and scale of those responses.35
84. Despite these differences, there is widespread agreement that for agricultural commodity
derivatives markets to function well, and as intended in terms of hedging and price discovery,
appropriate regulation needs to be in place across all relevant futures exchanges and markets. In
particular, there is need for greater transparency about transactions across futures markets and especially
across over-the-counter (OTC) markets, where transactions take place off the regulated commodity
exchanges. Comprehensive trading data need to be reported to enable regulators and participants to
monitor information about the frequency and the volume of transactions to understand what is driving
commodity prices.36 Such data exist for some commodity exchanges (though perhaps in an aggregate
form which makes the identification of various participants difficult), but are currently unavailable for
off-exchange trading.
85. The specific nature of the regulatory framework for futures exchanges and OTC markets,
whether for agriculture or other commodities, is an issue best addressed by financial market regulators.
And indeed significant regulatory changes have already been decided or are under consideration in
several important jurisdictions. In the United States, the Dodd-Frank Wall Street Reform and Consumer
Protection Act (2010) has mandated a tightening of financial market regulation to improve transparency
22
PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES 23
and reduce systemic default risk in the over-the-counter (OTC) derivatives trade. In the European Union,
with the same objectives, the Commission has adopted a proposal for regulation of OTC derivatives
trading and is currently reviewing several key directives that regulate financial markets including the
Market Abuse Directive and the Markets in Financial Instruments Directive.
86. In addition to the long-established markets in the United States and Europe, agricultural
commodity futures exchanges also exist in some emerging-market members of the G20, including Brazil,
China, India and South Africa.37 Price developments in most of the contracts traded on these exchanges
closely follow the developed-country exchanges where price discovery provides global benchmarks.
Trading on local platforms allows exchange-rate risk to be avoided and reduces basis risk stemming from
a variety of factors: climatic conditions and different seasonal timings (South Africa), restrictions on
international and domestic trade (China), differences in quality specifications and difficulty in delivering
to overseas markets (Brazil). Some exchanges, e.g. in India, offer exchange trading for commodities
(e.g. cardamom and mentha oil) for which contracts exist nowhere else. All of these futures exchanges
are established venues for price-risk management through futures contracts on internationally traded
commodities and they have a highly although not necessarily heavily regulated environment.38
87. More generally, debate is on-going at national and international level about the possible merits
of the following actions in terms of transparency and improved market functioning:39
Establish a trade depository to register OTC contracts, in line with earlier decisions in the G20
Summit in 2009 in Pittsburgh.
Use of speculative position limits on commodity futures contracts to ensure control of undue
market influence.
Use of maximum limits to daily price changes to reduce volatility.
Use of limits on inventories held in delivery warehouses by non-commercial entities to limit
market manipulation possibilities.
Introduction of provisions for high volume and frequency trading into the regulatory regime.40
Ensuring that changes in regulation are adopted across commodity exchanges and across
countries in order to avoid the migration of participants and regulatory arbitrage.41
88. Beyond regulatory concerns, new futures instruments for mitigating commodity price risk
exposure might be explored. For example, a global wheat contract that would specify export delivery
points in the major producing regions has been proposed.42 The potential advantages of a global futures
contract with compulsory delivery include: identifying cheapest to deliver sources by designating
delivery points all over the world; acting as a global signalling system of both price and regional supply
availabilities; and attracting well-informed commercial entities while deterring non-commercial entities
from investing on such contracts. The development of such new instruments lies with the futures
exchanges and the market participants.
Recommendation 3
G20 governments recognize the need to improve information and transparency in futures and over-the-counter
markets and encourage appropriate rules to enhance their economic functions paying attention to the need for
harmonization across exchanges in order to avoid regulatory arbitrage.
Proposed changes should be considered in light of the on-going review of regulatory oversight of all financial
markets and not solely agricultural commodity markets, in particular by G20 Finance Ministers and Central Bank
Governors.
The G20 supports the efforts made by the United States, the European Commission and others in addressing
transparency and efficiency issues in futures markets.
23
24 PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES
89. Price volatility may originate from either domestic or international markets.43 Trade is an
excellent buffer for localised fluctuations originating in the domestic market. Though stockholding is a
necessary component of a well functioning market, in particular to smooth out seasonal fluctuations and
time lags in trade, year-to-year variations in domestic production can be more effectively and much less
expensively buffered by adjustments in the quantities imported or exported.44 To the extent that shocks
tend to be specific to individual regions of the globe, and to partly cancel out on a worldwide level, world
output of a given agricultural product is far less variable than output in individual countries. International
trade is therefore a potentially powerful engine to even out supply fluctuations across the globe, and as a
result to reduce market volatility. To fulfil this beneficial pooling function to the maximum degree, trade
has to be able to flow between nations and the tendency which has emerged, in recent crises, for
countries to try to insulate themselves from international markets needs to be reversed.
90. More generally and in the longer term context, trade is an essential component of any food
security strategy. There is significant potential for increased production in many parts of the world, but
not all countries everywhere can or should aspire to supplying all their own needs. Doing so is
excessively costly, and will reduce choice and quality, without providing the reliability needed to achieve
food security. Moreover, the impact that climate change will have on food production is uncertain, but
many experts concur that it will lead to a worsening of conditions for agricultural production in some
countries or regions already facing difficult climatic and natural conditions. Experts also agree that there
will be an increase in the incidence of extreme events such as drought, heat-waves, and floods. A
combination of better functioning and deeper markets for agricultural commodities and improved supply
capacity and resilience will allow countries in the most vulnerable zones to overcome these problems.
91. Trade policy restrictions are not the only impediment to the free flow of goods and services.
Market and transportation infrastructure, the capacity to meet sanitary and phyto-sanitary requirements
and many other factors will determine a countrys capacity to export. Initiatives such as the Aid-for
Trade programme being implemented by the WTO and the OECD are contributing to overcoming some
of these domestic barriers to trade. A stepping up of this effort could bring significant benefits to
developing countries with export potential in agriculture.
92. Policies that distort production and trade in agricultural commodities potentially impede the
achievement of long run food security, by stimulating or conserving production in areas where it would
not otherwise occur and by distorting, obscuring or impeding the transmission of price signals to
competitive producers elsewhere. The efficient functioning of food supply chains domestically and
internationally also requires attention to be paid to competition policy issues upstream and downstream
of the farm sector, as a necessary complement to agricultural and trade policy reforms.
93. Despite on-going reforms there are still significant barriers to trade in agricultural commodities
among developing countries and between developing and OECD countries. They contribute to the
thinness of international markets that has been blamed for some of the volatility experienced in recent
years. Average tariffs on agricultural and food are high for middle income and high income countries,
25% and 22% respectively.45 Protectionism on agricultural products is not only higher than on non
agricultural products (by a factor of four), it is also much more volatile.46 Agricultural trade policies are
designed to insulate domestic prices from world markets and lead to pro-cyclic effects: protection
decreases when prices are high, increasing demand on world markets, and increases when world prices
are low effectively operating as a variable levy. Therefore, large country trade policies increase world
price volatility and create negative externalities for smaller countries.47 A conclusion of the Doha Round
will reduce the scope to implement destabilizing policies on world markets by reducing the bound level
of tariffs and subsidies.
24
PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES 25
94. OECD monitors support and protection of agriculture in its own member countries and in some
emerging countries that are major players in global food production and consumption. While trends in
the indicators measuring support and protection all point towards a continuing reduction in the levels and
the intensity of distorting interventions, much still needs to be done. Latest data for the OECD countries
indicate that government support still accounts for 22% of the total receipts of agricultural producers and
that more than half of that support is delivered in ways that are highly distorting of trade and
competition48.
96. Export prohibitions or restrictions relating to foodstuffs must also conform with the provisions
of the Agreement on Agriculture, that requires WTO Members to give due consideration to the effects of
such prohibition or restriction on importing Members' food security, give notice in writing, as far in
advance as practicable, and consult, upon request, with other WTO Members. These provisions do not
apply to a developing country Member, unless the measure is taken by a developing country Member
which is a net-food exporter of the specific foodstuff concerned.
97. These disciplines are considered to have been insufficient and weak during the 2007-2009
period, when export restrictions exacerbated or even, according to most experts, caused severe disruption
and a collapse in confidence on international markets. Export restrictions have also contributed to the
price increases and general market nervousness currently being experienced.49 It has been estimated that
if countries are free to implement export taxes a 10 percentage point increase in world prices can be
amplified to between 20 and 50 percentage points. In addition, the risk of export restrictions, and the
asymmetry between international disciplines (e.g. in WTO agreements) on export restrictions (unbound)
and import restrictions (bound) is a severe barrier to increasing trust in international markets. To be sure
that international trade is a reliable source of food supply net food importers should benefit from much
stronger guarantees from their trading partners. A first best option would be to a ban on export
restrictions. Countries would address domestic food security issues with direct and targeted support.
However, it is most unlikely that a ban on export restrictions would be agreed and, even if agreed, that it
would be enforced during a food crisis. On the other hand, reinforced rules, in particular in terms of
transparency, are both possible and useful.
98. As agricultural markets become more open, alternative measures addressing market
imperfections, many of which are policy driven, are needed. This would ensure that the potential new
opportunities created can actually be exploited by competitive suppliers. In the case of less developed
countries in particular, investments in improving supply capacity, including Aid for Trade will be
important.
25
26 PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES
Recommendation 4
G20 governments demonstrate leadership in on-going WTO DDA negotiations, moving immediately to strengthen
international disciplines on all forms of import and export restrictions, as well as domestic support schemes, that
distort production incentives, discourage supply in response to market demand, and constrain international trade of
food and agriculture products. Specifically,
substantially improve market access, while maintaining appropriate safeguards for developing countries,
especially the most vulnerable ones;
substantially reduce trade distorting domestic support, especially by developed countries; and,
eliminate export subsidies.
Taking existing WTO rules into account and the state of play in the DDA negotiations G20 governments should:
develop an operational definition of a critical food shortage situation that might justify consideration of an export
restricting measure. An export ban would be defined as a time-limited measure of last resort, allowed only
when other measures, including triggering domestic safety net measures for the poorest, have been exhausted,
and taking into account, in particular, the food security needs of least developed countries and net food
importing developing countries.
widen, strengthen and enforce consultation and notification processes currently in place at the WTO. The
intention to impose an export restriction would have to be notified in advance of the action being applied and a
fast track consultation process could be put in place to discuss whether the measure can be avoided and
how. Consultation should be on-going and regular with a view to ensuring that the measure, once in place, is
removed at the earliest possible moment.
99. With respect to export restrictions nations have agreed to commit to make humanitarian
exemptions, first, at the G8 Summit in LAquila in July 2009, and then at the World Summit on Food
Security in Rome in November 2009, where all FAO member states agreed to remove food export
restrictions or extraordinary taxes for food purchased for non-commercial humanitarian purposes, and to
consult and notify in advance before imposing any such new restrictions. If honoured these
commitments would allow food to be shipped rapidly to where it is needed in an emergency.
100. Some nations that imposed export restrictions during 2008 and 2010 made exemptions for
purchases of humanitarian food, including those by the WFP. However, others have not made such
exemptions, forcing in-country and international humanitarian agencies to purchase food from more
distant sources. And most exemptions, if made, are on a case-by-case basis after concern has been raised
and the exemption requested. Valuable emergency response time and resources are lost, as procurement
teams have to spend time negotiating, or find alternative suppliers from other regions.
Recommendation 5
G20 governments strengthen the commitments made at the LAquila and Rome Summits, calling on all nations
to allow purchases of humanitarian food, especially by WFP, to be exempted from food export restrictions
and/or extraordinary taxes, so that humanitarian food can be purchased, exported and/or transited regardless
of any prohibitions, restrictions or extraordinary taxes imposed; and resolve to bring this commitment and call to
the UN General Assembly and to the WTO.
101. Between 2000 and 2009, global output of bio-ethanol quadrupled and production of biodiesel
increased tenfold; in OECD countries at least this has been largely driven by government support
policies.50 Moreover, trade restrictions by favouring domestic sources of raw material for biofuels do not
maximise expected environmental benefits. Biofuels overall now account for a significant part of global
use of a number of crops. On average, in the 2007-09 period that share was 20% in the case of sugar
26
PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES 27
cane, 9% for both oilseeds and coarse grains (although biofuel production from these crops generates by-
products that are used as animal feed), and 4% for sugar beet.51 With such weights of biofuels in the
supply-demand balance for the products concerned, it is not surprising that world market prices of these
products (and their substitutes) are substantially higher than they would be if no biofuels were produced.
Biofuels also influence products that do not play much of a role as feedstocks, for example wheat,
because of the close relations between crops on both the demand side (because of substitutability in
consumption) and the supply side (due to competition for land and other inputs).
102. At the international level, crop prices are increasingly related to oil prices in a discrete manner
determined by the level of biofuel production costs. Increases in the price of oil enhance ethanols
competitiveness relative to petrol and strengthen its demand. Since both energy and food/feed utilise the
same input, for example grain or sugarcane, increases in the production of ethanol reduce the supply of
food and result in increases in its price. This relationship between the prices of oil, biofuels and crops
arises due to the fact that, in the short run, the supply of crops cannot be expanded to meet the demand by
both food and energy consumers.
103. If oil prices are high and a crops value in the energy market exceeds that in the food market,
crops will be diverted to the production of biofuels which will increase the price of food (up to the limit
determined by the capacity of conventional cars to use biofuels - in the absence of flexfuel cars and a
suitable distribution network). Changes in the price of oil can be abrupt and may cause increased food
price volatility. Support to the biofuel industry also plays a role. Subsidies to first-generation biofuel
production lower biofuel production costs and, therefore, increase the dependence of crop prices on the
price of oil. Such policies warrant reconsideration.52 53
Recommendation 6
G20 governments remove provisions of current national policies that subsidize (or mandate) biofuels production or
consumption. At the same time, governments should:
Open international markets so that renewable fuels and feed stocks can be produced where it is economically,
environmentally and socially feasible to do so, and traded more freely.
Accelerate scientific research on alternative paths to reduced carbon emissions and to improved sustainability
and energy security.
Encourage more efficient energy use, including in agriculture itself, without drawing on finite resources,
including those needed for food production.
Failing a removal of support, G20 governments should develop contingency plans to adjust (at least temporarily)
policies that stimulate biofuel production or consumption (in particular mandatory obligations) when global markets
are under pressure and food supplies are endangered.
Some ideas for how this could be done are explored in Annex D.
27
28 PRICE VOLATILITY IN FOOD AND AGRICULTURAL MARKETS: POLICY RESPONSES
104. Waste has been identified as an important issue which affects the underlying supply-demand
balance for food.
In OECD countries and increasingly in emerging economies waste occurs in the distribution
system, in the restaurant sector and at home,54including parts of food products which are not
economical to use; food that does not meet cosmetic standards, plate waste; food that is
discarded because it spoiled, and inefficient use of food , contributing to obesity.
105. Most losses are avoidable to some degree and some types of waste could be almost entirely
eliminated. Reducing waste could be an important part of a strategy to improve food security while
reducing environmental and resource pressures. If food waste can be reduced, the increase in production
estimated to be needed to meet the increase in demand over the next 40 years would be smaller.
Reducing waste would also help to reduce the pressure on land, water stress, soil degradation, and
greenhouse gas emissions.55 In developing countries, the measures proposed in Chapter 3 to improve the
overall resilience and productivity of agriculture should address much of the problem of waste from post-
harvest losses. In developed countries, possible avenues for policy action could include engaging with
the private sector, to increase awareness and develop voluntary agreements, reviewing regulations that
may inadvertently generate avoidable waste, supporting research to improve storage, prolong shelf life
and better detect deterioration, implementing public education campaigns, and investment in better
assessment and monitoring.
5. Policy options to deal with the consequences of price volatility, particularly for the most
vulnerable
5.1. Coping with volatility in the short run: buffer stocks, emergency food reserves, international
and national safety nets
National buffer stocks
106. Buffer stocks are an important policy instrument in a number of emerging economies and
developing countries, though they have been virtually abandoned in developed countries. Some rice
producing Asian countries rely on a combination of rice reserves, import or export monopolies, and
domestic procurement to stabilise prices within a pre-determined band. These measures aim to stabilise
domestic rice prices and, in some cases, have stimulated agricultural growth. In Africa, the experience
with maize buffer stocks is mixed.
107. The operational costs of buffer stocks are significant. Appropriate storage infrastructure is
extremely costly to acquire, and buying the food stock and holding it is also very expensive. Domestic
procurement, food releases from buffer stocks and trade programmes require continuing budgetary
allocations to cover any operational losses occurring in domestic and international trading. Losses
incurred on behalf of policy-dictated objectives for price stabilization may be viewed as direct subsidies.
Although expenditures associated with the acquisition and holding of stocks for food security purposes
can qualify under the WTO Green Box,56 from a WTO point of view, such price stabilisation
mechanisms could also be considered as trade distorting support. In times of price increases, such costs
can escalate to significant levels, rendering buffer stocks ineffective in containing price surges.
28
Food Sec.
DOI 10.1007/s12571-015-0427-z
ORIGINAL PAPER
Abstract Fish provides more than 4.5 billion people with at cessing and trading), fish contribute substantially to the in-
least 15 % of their average per capita intake of animal protein. come and therefore to the indirect food security of more than
Fishs unique nutritional properties make it also essential to 10 % of the world population, essentially in developing and
the health of billions of consumers in both developed and emergent countries. Yet, limited attention has been given so
developing countries. Fish is one of the most efficient con- far to fish as a key element in food security and nutrition
verters of feed into high quality food and its carbon footprint is strategies at national level and in wider development discus-
lower compared to other animal production systems. Through sions and interventions. As a result, the tremendous potential
fish-related activities (fisheries and aquaculture but also pro- for improving food security and nutrition embodied in the
strengthening of the fishery and aquaculture sectors is missed.
The purpose of this paper is to make a case for a closer inte-
gration of fish into the overall debate and future policy about
C. Bn (*) food security and nutrition. For this, we review the evidence
Institute of Development Studies, University of Sussex, Sussex, UK from the contemporary and emerging debates and controver-
e-mail: c.bene@cgiar.org
sies around fisheries and aquaculture and we discuss them in
M. Barange the light of the issues debated in the wider agriculture/farming
Plymouth Marine Laboratory, Plymouth, UK literature. The overarching question that underlies this paper
e-mail: MABA@pml.ac.uk is: how and to what extent will fish be able to contribute to
feeding 9 billion people in 2050 and beyond?
R. Subasinghe
Food and Agriculture Organization, Rome, Italy
e-mail: Rohana.Subasinghe@fao.org
Keywords Fish . Food security and nutrition . Micro-nutrient
P. Pinstrup-Andersen deficiency . Fisheries . Aquaculture
Cornell University Ithaca, Ithaca, USA
e-mail: pp94@cornell.edu
G. Merino
AZTI, Pasaia, Gipuzkoa, Spain
e-mail: gmerino@azti.es
Introduction
G.<I. Hemre Fish is critically important to food security and good nutrition
Institute of Nutrition and Seafood Research, Bergen, Norway (Allison 2011; Thilsted 2012; Beveridge et al. 2013). Fish and
e-mail: Gro-Ingunn.Hemre@nifes.no
other aquatic foods are high in protein and contain many es-
M. Williams sential micronutrients. The fishery and aquaculture sectors are
17 Agnew Street, Aspely, Queensland, Australia the source of income for millions of women and men in low-
e-mail: meryljwilliams@gmail.com income families (Bn 2006), thus contributing directly and
indirectly to their food security (Bn et al. 2007; Allison
Present Address:
C. Bn 2011; World Bank/FAO/WorldFish 2012). Yet, the potential
International Center for Tropical Agriculture, Cali, Colombia contributions of fish to food security and nutrition (FSN) are
C. Bn et al.
all but ignored in the international debate, as if a firewall stood of long-chain, poly-unsaturated fatty acids (LC-PUFAs), is
between the discussions about the role of fish and the broader unique. Fish is also an important source of essential
debates about FSN issues. A recent review of international micronutrients vitamins D, A and B, and minerals (calcium,
development and research agencies working on FSN revealed phosphorus, iodine, zinc, iron and selenium), which makes it
for instance that Bfish is strikingly missing from strategies for particularly attractive in the current fight against malnutrition
reduction of micronutrient deficiency, precisely where it could in low income and food deficient countries (LIFDCs). Some
potentially have the largest impact^ (Allison et al. 2013: 45). countries (e.g., Zambia, Brazil, or Chile) have already recog-
The purpose of this paper is to make a case for a closer nized this potential and have included fish in their national
integration of fish into the overall FSN debate. As such the school-feeding programmes. Should this be more systemati-
discussion is directed not to those in the fisheries and aqua- cally considered in countries with high levels of malnutrition
culture communities who have been relentlessly advocating to complement or substitute the technology-dependant (and
for fish as an entry point in this debate (see e.g., Bn et al. expensive) bio-fortification programmes?
2007; Hall et al. 2011; Thilsted 2012), but to the vast majority In addressing these issues we must first acknowledge that
of the international experts who are influencing the wider fish production sectors and associated value chains have been
debate and policies on FSN. remarkably dynamic in the last three decades what some
The World Committee on Food Security (CFS) took an refer to as the blue revolution. Every second fish we con-
important step in the right direction in 2012 by requesting that sume is now produced in aquaculture. This new situation is a
the High Level Panel on Food Security (HLPE) undertook an result of the stagnation of capture fisheries following decades
in-depth study of the role of sustainable fisheries and aquacul- of expansion, some of it carried out in an unsustainable man-
ture for FSN. The report (HLPE 2014) presents the strongest ner (FAO 2012), combined with impressive growth rates in
case yet for incorporating fish into the debates on how to the aquaculture industry, subject to some major adjustments to
achieve FSN for all and suggests a set of activities and policies overcome the early challenges such as fish disease and nega-
to be pursued. tive environmental impacts. Critics of aquaculture have also
Building on this report and a large body of literature, our pointed to the use of fishmeal and fish oil produced from wild
ambition is to bring fish onto the table. A number of papers fish, to feed farmed fish. These concerns are valid and need to
have succeeded in presenting a very comprehensive evidence- be taken (along with their solutions) into account in this
based argument for supporting fish as a central element in the discussion.
FSN debate (see e.g., Prein and Ahmed 2000; Roos et al.
2003; 2007; Kawarazuka and Bn 2011; Beveridge et al.
2013). Here, we aim to review the evidence from the contem- Understanding fish production in relation to food security
porary and emerging debates and controversies around fisher- and nutrition
ies and aquaculture. In doing this we do not go into the micro-
level issues concerning social, gendered and micro-economic The fastest growing food-supply industry in the world
aspects of fish, FSN at the local levels for those we refer
readers to the HLPE report (HLPE 2014). In 2011, 173 Mt of fish were extracted from the global
Some of the underlying questions in the present study are marine and inland water ecosystems, of which 7 to 10
closely linked to the issues discussed in the wider agriculture/ Mt were discarded prior to landing and 12 Mt were lost
farming literature. The latest estimate suggests for instance at the post-harvest stage. The production from capture
that, in 2009, fish accounted for 17 % of the global popula- fisheries and aquaculture available amounted therefore
tions intake of animal protein and 6.5 % of all protein con- to 154 Mt, of which about 131 Mt were utilized directly
sumed (FAO 2014). In the context of this Special Issue on for human consumption (Fig. 1) (HLPE 2014). Thanks
BFeeding 9 Billion^, the key questions are: a) Can we main- to aquaculture and fisheries, the global supply of fish
tain these consumption rates, given the projected growth in has grown by a factor of 8 since 1950. By comparison,
human population and the growing environmental challenges even after the Green Revolution the world rice produc-
facing the earth? and b) Could fish - one of the most efficient tion increased only by a factor of 3. In effect, fish
converters of feed into high quality food be a substitute for production has been the fastest growing food industry
other sources of animal protein? Finally, in the context of the in the world for the last 40 years and is expected to
current discourse about future resource scarcity, should fish be remain so in the near future. The fish supply per capita
considered a more environmentally friendly source of protein has more than tripled in the last half century, from
than the other livestock production systems? 6 kg/year in 1950 to 18.8 kg/year in 2011 Fig. 2.
Fish is more than just a source of animal protein. Fish In fact, the world fish supply has effectively been grow-
contains several essential amino acids, especially lysine and ing faster than the worlds population (FAO 2012).
methionine. The lipid composition of fish, with the presence These global figures mask, however, some important
Feeding 9 billion by 2050 Putting fish back on the menu
Fish 73%
Pig 20%
17 mill tonnes Fishmeal, Fish oil (b) Poultry 05%
Others 02%
6 mill tonnes Ornamental, bait, etc.
regional variations: fish consumption is the lowest in capita intake (HLPE 2014). This share can exceed 50 %
Africa (9.1 kg per capita in 2009), while Asia accounts in some countries. In West Africa, Asian coastal coun-
for almost two-thirds of total consumption (20.7 kg per tries, and many small island states, the proportion of
capita). total dietary protein from fish can reach 60 % or more
(e.g., Gambia, Sierra Leone, Ghana, Cambodia,
Fish as the largest source of animal protein Bangladesh, Indonesia, Sri Lanka, or the Maldives)
(FAO 2014).
Fish is a major source of animal protein, overshadowing The geography of fish as a source of protein is also signif-
most other sources. In 2010 it represented a source icant in the FSN discussion. Twenty-two of the 30 countries
twice as important as poultry, and three times larger where fish contribute more than one-third of the total animal
than cattle (Fig. 3). Today capture fisheries and aqua- protein supply were officially referred to as LIFDCs in 2010
culture provide 3 billion people with almost 20 % of (Kawarazuka and Bn 2011). In other words, almost three-
their average per capita intake of animal protein, and a quarters of the countries where fish is an important source of
further 1.3 billion people with about 15 % of their per animal protein are poor (income-wise) and food-deficient. Yet
Source: FAO Statistics and Information Branch of the Fisheries and Aquaculture Department.
C. Bn et al.
in these LIFDCs, even if fish is a substantial proportion of the offer to address multiple micronutrient deficiencies is now
food intake, undernourishment can still occur as total food being recognized (e.g., Roos et al. 2007; Kawarazuka and
intake is often insufficient. Bn 2011; Thilsted 2012). For instance, the high level of
iodine found in fish can help prevent iodine deficiency which
Fish - beyond protein is known to cause cretinism (stunted growth and mental
retardation).
Some point out, however, that the main contribution of fish to
FSN may not be in relation to its protein content, but its lipid
and micro-nutrient content. Current issues in relation to fish contribution to food
The lipid composition of fish is unique, having LC-PUFAs, security
with many beneficial effects for child development and adult
health (Thilsted et al. 1997; Larsen et al. 2011; Richardson and Fishmeal and fish oil in aquaculture
Montgomery 2005). Among fish species that are cheaper in
developing countries, small pelagic fish such as anchovy and In addition to being used directly as human food, fish also
sardine are some of the richest in LC-PUFAs (USDA 2011), contributes indirectly to human nutrition when it is used as
especially compared to large freshwater fish such as carp and fishmeal for aquaculture and poultry/livestock feeds (Tacon
tilapia. When its rich nutrient content is preserved, fish can and Metian 2009). In 2011, 23 Mt of fish essentially small
provide protective effects against a wide range of health is- pelagic fish species such as anchovy, herring, mackerel and
sues. LC-PUFAs for instance provide protection against dis- sardine have been destined to non-direct human consump-
eases such as stroke, high blood pressure or coronary heart tion, of which 75 % (17 Mt) was reduced to fishmeal and fish
disease (Miles and Calder 2012; Rangel-Huerta et al. 2012). oil for aquaculture, poultry and other livestock feeding (cf.
Complementing its fatty acid content, fish is also known to Fig. 1). In 2010, 73 % of the total world fishmeal was used
be an important source of essential micronutrients vitamins to feed farmed fish, followed by pigs (20 %), poultry (5 %)
D and B, and minerals (Roos et al. 2003; 2007; Bonham et al. and others (2 %) (Shepherd and Jackson 2013).
2009). Lipid-rich fish also contain vitamin A. Recent research From a FSN perspective, the use of fishmeal for farmed
showed that fish species consumed whole with bones, heads, fish (and livestock) raises important issues. Leaving aside the
and viscera play a critical role in micronutrient intakes of debate on the role of small pelagic fish in supporting larger
people as these parts are where most micronutrients are con- fish, birds and marine mammals in the ecosystem (Smith et al.
centrated. Some of these small fish (such as mola 2010), is fishmeal the most efficient way to use fish (especial-
(Amblypharyngodon mola), darkina (Esomus danricus), sar- ly low-cost small pelagic fish rich in LC-PUFA) or would
dines and pilchards, anchovy, seabass, tilapia) contain high these fish contribute more to food security if a larger share
levels of minerals such as calcium, phosphorus, iodine, zinc, of them was eaten directly by people? Indeed, despite some
iron and selenium, which are low in other foods. The potential substantial improvement in the last decade, the rate of conver-
contribution that fish (even in small quantity) can therefore sion of fishmeal to fish is still a source of concern (Troell et al.
Feeding 9 billion by 2050 Putting fish back on the menu
2014). On average, for every kg of farmed fish produced, expressed a more nuanced but nonetheless concerned view
0.7 kg of wild fish is needed (Tacon and Metian 2008). This about the state of world marine fisheries (e.g., FAO 2012),
average figure, however, masks important differences: while acknowledging the granularity in the state of resources world-
for omnivorous farmed fish, the rate is down to an acceptable wide. The current consensus is that global fisheries would be
0.2 to 1.41 kg of wild fish per 1.0 kg of farmed fish, for more productive if the levels of overfishing were reduced
carnivorous farmed fish, the figure is higher: 1.35 to 5.16 kg (Srinivasan et al. 2010), and the environmental sustainability
to produce one kg of farmed fish (Boyd et al. 2007). of fisheries were recognized to be a sine qua none condition
for FSN (HLPE 2014).
Fish losses and implications on food security and nutrition An important point in this debate, however, is that improv-
ing FSN through fisheries would depend on stock recovery
The global discards of fish (fish caught but dumped overboard and also on access to and distribution of the harvest, as well as
due to low quality, damage or spoilage, non-targeted species gender consideration (de Schutter 2012; Williams 2010).
or below regulation size) were estimated to be around 7.3 Mt Indeed overfishing per se is only one aspect of the problem.
in 2005, 80 % of which coming from industrial fleets. In Other economic activities, such as oil drilling, coastal devel-
contrast, small-scale fisheries generate less wastage in the opment, pollution, or dams and water flow management have
form of discards (about 2 Mt a year, that is, 4 % of their significant negative impact on aquatic habitats (Halpern et al.
estimated landing1) (Kelleher 2005). 2008).
With increasing fish scarcity and increasing fish prices,
species previously deemed commercially inferior are progres- Indirect contribution to food security and nutrition
sively integrated into consumer eating habits and markets. through livelihood support
Most shrimp trawlers, which used to discard up to 95 % of
their catch, are now landing more bycatch for human con- A critical pathway to enhance FSN is through the income that
sumption (Bn et al. 2007). In small-scale fisheries where people generate from engaging in remunerated activities. In
discarding fish is rare, substantial quantities and quality are this respect aquaculture and fisheries (especially through the
lost due to post-harvest mishandling during transport, storage, number of small-scale operators engaged in fishing, aquacul-
processing, on the way to markets and waiting to be sold. ture, processing, and trading businesses see e.g., Bn et al.
Especially in developing countries, where access to electricity 2010) play a critical role in low income and emergent coun-
and cold chain can still be an issue, fish post-harvest losses can tries. Altogether it is estimated that between 660 and 820
be significant. Estimated at 1012 Mt in 2005, the total ac- million people (fishers, fish-farmers, fish traders, workers in
counts for 10 % of global capture and culture fisheries pro- fish processing factories, and their families) depend on fish-
duction (Bn et al. 2007). While physical losses in small- related activities as a source of income (Allison et al. 2013;
scale fisheries are less than 5 %, economic losses can be sub- HLPE 2014). This represents more than 10 % of the world
stantial. In Africa, some estimates (FAO Focus nd) put post- population. For most of these households, the revenues gen-
harvest losses in some cases to levels as high as 2025 %. In erated may not necessarily be very high (Neiland and Bn
aquaculture, waste streams in value chains have been large in 2004; Allison et al. 2011; Bn and Friend 2011), but it is
the past but are tending to decline rapidly as competitive pres- often the main component of their livelihood, which allows
sures force innovation (Arthur et al. 2013). them to secure accessibility to food (Heck et al. 2007; Bn
et al. 2009; Eide et al. 2011).
Sustainability of fisheries and implications for food security
Feeding 9 billion by 2050: where does fish stand?
Because total fish production (availability) is an important
dimension of food security, a key issue in this debate is the
The global human population is expected to exceed 9 billion
environmental sustainability of capture fisheries. The extent to
by 2050 (UN estimates), increasing the pressure on the food
which capture fisheries have exceeded sustainable levels has
sectors to maximize production and reduce waste. Production
generated strong expert and public opinion debates (Worm
increase must occur in a sustainable way and in a context
et al. 2006; Pauly et al. 2005; Hilborn 2013) and many media
where key resources, such as land and water, are likely to be
headlines, scientific papers and environmental campaigns
scarcer and where climatic change impact will intensify. The
have been in the last two decades framed around the idea that
fish-production sector is no exception.
world fisheries resources are in crisis due to overfishing (see
In this context two key questions emerge. First, will fish-
e.g., Pauly et al. 1998; Myers and Worm 2003). FAO has
eries and aquaculture be able to maintain the current global
1
Overall, small scale fisheries land approximately 40 Mt annually, com- fish consumption rate of 18 kg per capita per year, and the
pared to 50 Mt by large scale fisheries (HLPE 2014). equivalent regional values, and if not, how will society
C. Bn et al.
address the needs of expected winners and losers (Barange Demand projection
et al. 2014)? So far technological and institutional innovations
have ensured that the combined production of fish through Modelling exercises have been conducted recently with the
fisheries and aquaculture has been faster than the world pop- objective of estimating the projections of fish demand and
ulations demand of fish. The question is now whether we can supply. These modelling exercises include the World Bank-
keep up this pace with another 2 billion people added in the FAO-IFPRI Fish 2030 analysis elaborated on the IMPACT
next 35 years, and how the four dimensions of food security model developed by the International Food Policy Research
(availability, accessibility/affordability, utilization and stabili- Institute (IFPRI) (World Bank 2014); the OECD-FAO
ty) will be balanced to ensure that fish go to those who need Agricultural outlook model built on the combined
them most (Merino et al. 2012). multimarket, partial equilibrium AgLink-CoSiMo model
The second key question is whether sustainable fisheries (OECD-FAO 2013; Lem et al. 2014); and a series of peer-
and aquaculture will be able to help address the bigger food reviewed articles relying on various types of modelling and
security issue that will affect the world in the coming decades? projection tools (Rice and Garcia 2011; Merino et al. 2010,
In particular, could aquaculture become a substitute form of 2012; Barange et al. 2014). The time horizon of these different
protein for some of the less efficient food production systems, analyses is not always the same. For instance, the OECD-FAO
or even be used to compensate for the decline in farming outlook model works over a 10 year-projection period (i.e., up
systems productivity that is predicted as a consequence of to 2023 as per the last iteration of the model), while the World
the impact of climate change? Bank-FAO-IFPRI model runs until 2030; Merino et al. 2010
used a 20 year simulation, calibrated on a 19972004 data set
(meaning technically that their projection runs until 2024),
How much fish do we need? while Merino et al. (2012) and Barange et al. (2014) proposed
a projection up to 2050. A strict comparison of the different
Global drivers of fish demand projections is therefore difficult.
A bigger issue is that very few of these studies integrated
World population is often presented as a key driver for the information on the drivers of changes (the combination of
growth in seafood demand and for fisheries development. In urbanization and increase in income), to estimate with accu-
reality, a more important driver for fish (and other animal- racy the future demand for fish across the world. Many studies
source food) consumption is income (Speedy 2003). assume constant consumption rates in the future (e.g., Barange
Demand for fish as food is particularly high in the wealthier et al. 2014) or fixed nutritional targets (Rice and Garcia 2011).
strata of societies, including in the low-income countries, and Others worked directly with projected fish consumption (e.g.,
as income will continue to increase in highly populated coun- OECD-FAO 2013; Merino et al. 2012), that is, by dividing the
tries such as China and India, demand levels are likely to rise projected supply by the projected population. None of these
more strongly (Garcia and Rosenberg 2010). Overall, a large approaches therefore offers an appropriate basis for estimating
increase in the number of people moving into the middle class, the actual demand for fish. The World Bank-FAO-IFPRI study
particularly but not exclusively in Asia, is likely to result in a uses regional fish consumption rates to estimate the global
very large expansion in the demand for fish. demand but the report does not clarify how these figures are
Income is, however, not the only driver of fish demand. It is constructed.2 Aggregating the regional figures at the global
recognized that urbanization is also an important factor in- level, the report estimates that the world demand for fish will
creasing animal-source food consumption in general and fish be around 152 Mt in 2030 (World Bank 2014).
consumption in particular. Delgado et al. (1997) suggest that Merino et al. (2012) estimated the expected production
changes in food preference driven by urbanization alone have, capacity of marine ecosystems exploited under maximum sus-
in the past, accounted for an extra 5.79.3 kg per capita con- tainable yield principles, and projected aquaculture production
sumption of meat and fish per year. Similarly, Betru and requirements to achieve a range of food consumption targets.
Kawashima (2009) and Toufique and Belton (2014) present They concluded that between 125 and 210 Mt of fish by 2050
data from Ethiopia and Bangladesh, indicating that urbaniza- will be necessary to maintain fish consumption at around 15
tion affects animal food consumption rates independently of 20 kg per capita per year. Starting from a different angle, Rice
income.
These different factors explain the rapid increase in 2
The report explains that BFor the subsequent years in the simulation
demand for meat, milk, and fish in the emerging econ- [after 2000 which was used as the base year for the calibration], these
omies of Asia. In China, for example, the demand for intercept values [between supply and demand] are changed according to
the exogenous growth rates specified for each of the supply and demand
fish is likely to increase from 24.4 kg per person per
functions^ (World Bank 2014: 23, our emphasis). It is not clear from the
year in 2000 to 41 kg per person per year by 2030 report how the exogenous growth rates in regional demand have been
(World Bank 2014). computed in the model.
Feeding 9 billion by 2050 Putting fish back on the menu
and Garcia (2011) sought to estimate the need for additional stronger impact on fishing and aquaculture distribution and
fish necessary to supply 20 % of the dietary protein require- productivity (Cheung et al. 2009). At the same time however,
ment to feed a 9 billion population by 2050. On this basis, an the interconnectedness of aquatic systems allows many spe-
additional production of 75 Mt of fish from fisheries and cies to change spatial distribution more easily as ecosystems
aquaculture would be needed above the 2006 production level shift, to remain in their zones of preference. Clearly, therefore,
(144 Mt), that is, approximately 215 Mt. This represents an the impact of global climate change on ocean capture fisheries
almost 50 % increase in production with respect to the 2006 will be important. Biological predictions based on ocean-
level. While this figure is above other projections, it has the atmosphere general circulation models (OA-GCMs) have al-
advantage that it starts from an actual estimate of the future ready demonstrated that the physical and chemical properties
needs (as opposed to the estimate of apparent consumption of the oceans will be modified, affecting the productivity,
rates), but without factoring in future production potentials distribution, seasonality and efficiency of food webs, from
or market responses to the fishmeal/fish oil demand (see primary producers to fish (Steinacher et al. 2010; Cheung
below). et al. 2009, 2011). Some of these earlier GCM models how-
In summary, the current understanding of the global drivers ever, were limited by their coarse resolution, too low to cap-
of fish demand (urbanization and increase in living standards ture the processes that dominate the dynamics of the worlds
in developing and emerged countries) is relatively well coastal and shelf regions, where most fisheries operate. More
established but not all of the current models have integrated advanced models are now available (e.g., Merino et al. 2012;
these drivers comprehensively. In comparison, the efforts to Blanchard et al. 2012; Barange et al. 2014).
better understand the ability of the world to produce fish in the Overall, and with few exceptions, the conclusion of all
future (i.e., the supply side of the equation) have been more these models is that although climate change will alter the
elaborate. present geographical distribution of shelf-sea ecosystems pro-
ductivity, in most of the regions and EEZs, the overall poten-
tial impact is projected to be low to moderate. Barange et al.
The future of fisheries production and the impact of climate
(2014), for instance, used a high resolution shelf-sea physical-
change
biological model that allowed them to scale down the analysis
and gave greater confidence in predicting the consequences at
A consensus has emerged in the literature that the doom-and-
national scales. They conclude that by 2050, estimates of na-
gloom rhetoric that had driven the discussion surrounding the
tional fish production should remain on average within 10 %
state of marine fisheries in the late 2000s (Garcia and Grainger
of the present yields.
2005; Caddy and Seijo 2005) was exaggerated (Grafton et al.
2010; Hilborn 2010) and that although the situation remains
The contribution of aquaculture to future fish supply
concerning in respect to many stocks, we are not likely to face
the global collapse that had been announced by some biolo-
The second element on the supply side is aquaculture.
gists (e.g., Myers and Worm 2003; Worm et al. 2006; Pauly
Discussion of the rise of aquaculture has so far largely focused
2009). Instead, the downward trend of overfished stocks may
on its contribution to global aquatic animal food supplies,
have been reined in (Fig. 4).3 Reflecting this, most of the
ignoring the resultant changes in species composition of the
projections proposed in the recent literature estimate that the
fishes consumed, how it is farmed, and the implications for
global fisheries landings are likely to be stable in the short to
food and nutrition security (Kawarazuka and Bn 2011;
medium term. The OECD-FAO model for instance estimates
UNHRC 2012). As a consequence, our understanding mainly
that capture fisheries will be 5 % higher by 2024 than it is was
concerns the question of the ability of the aquaculture industry
in 2013, that is, around 96 Mt (OECD-FAO 2013) while the
to maintain its rate of growth. In this regard, most of the recent
World Bank-FAO-IFPRI model estimates that this will be
analyses agree that the era of exponential growth is over and
around 93 Mt in 2030. These figures are at a global scale
while the sector will still continue to grow, the projected rate
however, and some regional outlooks are for good stock re-
of growth is expected to decelerate. The main causes of this
building, while others are for a worsening in overfishing.
slower growth are likely to be freshwater scarcity, lower avail-
Yet another key factor for which much uncertainty remains
ability of locations for optimal production, and high costs of
is the impact of climate change. Unlike most terrestrial ani-
fishmeal, fish oil and other feeds (FAO 2012). Nonetheless,
mals, aquatic animal species are poikilothermic (cold-
the World Bank-FAO-IFPRI model suggests this rate will still
blooded) and changes in aquatic habitat temperatures will
be sufficient to maintain a steady rise, reaching the point
more rapidly and significantly influence distribution, prey
where it will equal global fisheries production by 2030
availability, metabolism, growth and reproduction, with
around 93 Mt (World Bank 2014) (Fig. 5). Technical innova-
3
According to the FAO, the number of stocks fished at unsustainable tion, improved farm and animal health management, and im-
levels decreased from 32.5 % in 2008 to 28.8 % in 2011 (FAO 2014). proved and more efficient germplasm will be responsible for
C. Bn et al.
this increased growth, in combination with the continued ex- Putting all the pieces together
pansion of fish-farming. Combined with a projected capture
fish production that will remain fairly stable over the 2000 Given the very rough projections for fish demand and the
2030 period (see above), the global fish supply is projected to more elaborate projections for fish supply just reviewed
rise to 187 Mt by 2030 (World Bank 2014). These figures are above, will fisheries and aquaculture be able to maintain their
consistent with the projections proposed by OECD-FAO in current contribution to food security in the future? The answer
which the global fish production will reach 181 Mt in 2022, that emerges from the literature is that as far as food availabil-
of which 161 Mt would be destined for direct human con- ity and demands are concerned, it is a conditional yes. All the
sumption (OECD-FAO 2013). projection models currently available seem to agree that the
A key element in this discussion is the importance of overall fish consumption rate could be maintained, in other
fishmeal and fish oil and how markets and technological in- words, that the fisheries and aquaculture aggregated growth
novations will respond to price signals (Merino et al. 2010, will keep up with population growth rates. In fact, the OECD-
2012; FAO 2012; World Bank 2014; Troell et al. 2014).4 The FAO even estimates that the world fish consumption will in-
consensus is that the use of fishmeal in aquaculture feeds is crease by another 10 % and reach 20.9 kg per capita per year
expected to decrease in percentage with time, thanks to in- by 2023. The World Bank-FAO-IFPRI report is slightly more
creasingly effective replacements, including plant proteins,
waste products from fish and terrestrial animals and use of
better/improved breeds of aquatic animals with better feed
conversion (Tacon et al. 2011). Formulated feeds are a signif-
icant factor in production costs, and this is a strong incentive
to develop technology that will make feeds more affordable
and sustainable. Overall, the proportion of fish used for non-
direct human consumption has decreased from 30 % in the
early 1990s to 15 % in 2010 and the World Bank analysis
concluded that the projected expansion of aquaculture will
be achieved with only an 8 % increase in the global fishmeal
supply during the 20102030 period (World Bank 2014).
4
For instance, one of the 6 selected scenarios of the World Bank-FAO-
IFPRIs analysis is specifically focusing on these issues (Scenario 2
BExpanded Use of Fish Processing Waste in Fishmeal and Fish Oil
Data source: FAO FishStat and IMPACT projection model.
Production^) (World Bank 2014) and one of the selected issues of the
2012 FAO SOFiA report was on BDemand and supply of aquafeed and Fig. 5 Global fish production: data and projection 19842030 from the
feed ingredients for farmed fish and crustaceans: trends and future IMPACT model (World Bank 2014). Data source: FAO FishStat and
prospects^ (FAO 2012: 171182). IMPACT projection model
Feeding 9 billion by 2050 Putting fish back on the menu
conservative and estimates that the per capita fish growth will increase fish consumption while low-income peo-
consumption will remain at around 18 kg per year in 2030. ple will experience reductions in their fish consumption. This
Merino et al. (2012) and Barange et al. (2014) reached the is an important distributional and food security consequence.
same conclusion for 2050, but this is essentially due to the To avoid that, fish production needs to expand faster than
underlying assumptions of their models. population growth.
All these analyses, however, stress the fact that for this
outcome to occur, several specific conditions must be satis-
fied: capture fisheries will need to be exploited according to
sustainable principles; very significant technological develop- Can fish contribute further to the 2050 global food
ment will need to take place in aquaculture feeds to reduce security challenge?
fishmeal dependency and in farm management and germ-
plasm to improve the overall efficiency of aquaculture; and The arguments presented above are mainly constructed
discards, waste and losses will need to be reduced. Some of around a sectorial perspective. At least three other trans-
these conditions could be particularly challenging and a busi- sectorial (or systemic) arguments should be considered in re-
ness-as-usual approach is expected to fail.5 lation to the wider debate of feeding 9 billion by 2050.
In addition to whether or not the world (and the markets) First, in terms of animal protein availability, with 18.2 kg
will be responsible enough to ensure that these conditions are per capita per year, fish is providing 115, 133, and 189 % more
satisfied, the question of who will be the winners and losers is protein per capita than pig, poultry and beef respectively. In
vital to the other key FSN issues, which are: access to liveli- fact, fish (combining capture fisheries and aquaculture) has
hoods in fish value chains and affordability of fish. been the main contributor to the 61 % increase in the world
Unfortunately, the consensus is that in the coming decades, per capita consumption of animal protein for the period 1969
the current situation regarding the imbalance between con- 2009 (Table 1). As economic development is expected to con-
sumers in developed and developing countries is unlikely to tinue driving an increasing trend in animal protein demand,
change. While the present figures indicate the lowest fish (OECD-FAO 2013) and aquaculture is projected to remain the
consumption per capita in Africa (9.1 kg in 2009), the fastest growing food commodity sector, this sector will soon
different models project that this imbalance will deteriorate become even more central in the future food security of the
further. The OECD expects that per capita consumption in world population.
Africa will decrease by a further 10 % by 2024, while that
of Asia will show the highest growth rate (+14 %, OECD- Table 1 World per capita meat and fish food supply (kg per capita per
FAO 2013). The World Bank-FAO-IFPRI projection year) 19692009
(World Bank 2014) is even more alarming with per capita
1969 1979 1989 1999 2009
fish consumption expected to decline in sub-Saharan
Africa by 1 % annually to 5.6 kg in 2030.6 Yet at the same Pig meat 9.4 11.5 13.1 15.1 15.8
time developing countries will account for more than 91 % 25.2 % 27.7 % 28.2 % 28.1 % 26.3 %
of the total increase in fish consumption. Even so, their Poultry meat 3.8 5.6 7.3 10.7 13.6
annual per capita fish consumption will still remain below 10.2 % 13.5 % 15.7 % 19.9 % 22.7 %
that of more developed regions (19.8 vs. 24.2 kg) (World Bovine meat 10.8 10.7 10.3 9.7 9.6
Bank 2014). 29.0 % 25.8 % 22.2 % 18.1 % 16.0 %
A last important point needs to be mentioned. In view of Mutton and goat meat 1.8 1.5 1.7 1.8 1.9
the importance of income growth and urbanization as drivers 4.8 % 3.6 % 3.7 % 3.4 % 3.2 %
of fish demand, the objective of merely maintaining fish pro- Meat, other 0.8 0.8 0.7 0.8 0.9
duction growth rate at the same level as the growth in world 2.1 % 1.9 % 1.5 % 1.5 % 1.5 %
population would not be enough to prevent fish price from
Capture 10 10.4 11 10.5 10
increasing. The urban population with the highest income
26.8 % 25.1 % 23.7 % 19.6 % 16.7 %
5 Aquaculture 0.7 1 2.4 5.1 8.2
Merino et al. (2012) estimate that the fish used in aquaculture feed to
produce one unit of output would have to be reduced by at least 50 % 1.9 % 2.4 % 5.2 % 9.5 % 13.7 %
from current levels to secure sustainability. If not, demand will push the Capture and Fisheries 10.7 11.4 13.4 15.6 18.2
price of fishmeal products up to levels where the short term economic 28.7 % 27.5 % 28.8 % 29.1 % 30.3 %
incentive to exploit small pelagic (the main source of this fishmeal) be-
yond their maximum sustainable yield would be high, potentially leading Total 37.3 41.5 46.5 53.7 60
to increases in fishing capacity and rapid depletion of resources.
6
Some other regions of the world are also expected to face lower fish Source: FAOStat
consumption rate per capita: Japan, Latin America, Europe, and Central Figures in percent are the respective contribution of each sector to the
Asia. total figure
C. Bn et al.
Second, in term of efficiency, fish in aquaculture systems policies and programmes and (b) should be more systemati-
are very efficient converters of feed into protein in fact far cally included in countries nutritional programmes and inter-
more efficient than most terrestrial livestock system (Fig. 6a). ventions aiming at tackling micronutrient deficiencies espe-
For instance, poultry converts about 18 % of their consumed cially among children and women (HLPE 2014: 18
food and pigs about 13 %, as compared with 30 % in the case Recommendation 1a and 1b).
of fish (Hasan and Halwart 2009). Production of 1 kg of beef
protein requires 61.1 kg of grain, production of 1 kg of pork
protein requires 38 kg of grain, while fish only requires Conclusion
13.5 kg (Hall et al. 2011). Most of this difference comes from
two biological characteristics of fish which give them great Fish is already making a major contribution to human food
advantages over land-based livestock in growth performance: supply and to the support of FSN for more than 660 million
(i) the fact that fish are poikilotherms and therefore do not fish-workers and their families. It also provides more than 4.5
expend energy maintaining a constant body temperature; and billion consumers with at least 15 % of their average per capita
(ii) the fact that, because finfish are physically supported by intake of animal protein. In addition, because fish is more nu-
the aquatic medium, fewer resources are used on bony skeletal tritious than staple foods such as cereals, providing in particular
tissues, and a larger part of the food they eat is effectively essential fatty acids and micronutrients, it can play an extreme-
allocated to body growth. ly important role in improving the nutritional status of individ-
Third, in terms of carbon footprint, aquatic animal produc- uals, in particular those at risk such as children and women.
tion systems have a lower carbon footprint per kilogram of Yet limited attention has been given so far to fish as a key
output compared with other terrestrial animal production sys- element in FSN strategies at national level and in wider de-
tems (Hall et al. 2011). As a consequence, nitrogen and phos- velopment discussions and interventions. Part of the problem
phorous emissions (kg of nitrogen and phosphorus produced might have been that specialists in fisheries debates have con-
per tonne of protein produced) from aquaculture systems are centrated predominantly on questions of biological sustain-
much lower than those in beef and pork production systems ability and on the economic efficiency of fisheries, neglecting
and slightly higher than that of poultry (Fig. 6b). In fact, some issues linked to its contribution to reducing hunger and mal-
aquaculture production systems such as farming of bivalves nutrition and to supporting livelihoods (Kawarazuka and Bn
absorb nitrogen and phosphorous emissions from other systems. 2010, 2011). On the other end of the spectrum, and with too
All these reasons are important arguments for giving fish few exceptions, most (non-fishery) food security experts and
far greater attention in the food security debate and in the decision-makers seem unfamiliar with these facts and, there-
current discussion about how to feed 9 billion by 2050. The fore, unaware of the critical role that aquaculture is likely to
CFS has started to recognize this and the report that was play in the future. The problem is particularly pronounced in
commissioned in 2012 indeed recommends that fish (a) be- the current debate about how to make the food system more
comes an integral element in inter-sectoral national FSN nutrition sensitive, i.e., how to change and improve the food
1400
70
1200
kg / tonne of protein produced
60
1000
50
800
40
600 Nitrogen
30 Phosphorus
400
20
200
10
0
0
Source: Data for fish are derived from Hall et al. (2011). Data for beef, pork and chicken are derived from Flachowsky (2002) in Potrk
(2003).
Fig. 6 a Feed and protein conversion efficiency of beef, pork and fish; b Nitrogen and phosphorous emissions for animal production systems. Source:
Data for fish are derived from Hall et al. (2011). Data for beef, pork and chicken are derived from Flachowsky (2002) in Potrk (2003)
Feeding 9 billion by 2050 Putting fish back on the menu
systems in order to advance nutrition. As a consequence, fish Barange, M., Merino, G., Blanchard, J. L., Scholtens, J., Harle, J.,
Allison, E. H., et al. (2014). Impacts of climate change on marine
has so far been only marginally included in the international
ecosystem production in fisheries-dependent societies. Nature
FSN debate. Many nutritional programmes are still not aware Climate Change, 4, 211216.
of, or not recognizing and building on the potential of fish for Bn, C. (2006). Small-scale fisheries: assessing their contribution to
the reduction of micronutrient deficiency. rural livelihoods in developing countries. FAO Fisheries Circular,
In this paper, we make the case that fish deserves more No. 1008. Rome: Food and Agriculture Organization (FAO).
Bn, C., & Friend, R. (2011). Poverty in small-scale inland fisheries: old
attention in food policies due to its importance in the food issues, new analysis. Progress in Development Studies, 11(2), 119
basket, its unique nutritional properties, its higher efficiency 144.
of production and carbon footprint compared to other forms of Bn, C., Macfadyen, G., & Allison, E. H. (2007). Increasing the contri-
animal production systems. We acknowledge some chal- bution of small-scale fisheries to poverty alleviation and food secu-
rity. FAO Fisheries Technical Paper, No. 481. Rome: FAO.
lenges, however, especially in making fish more affordable
Bn, C., Steel, E., Kambala Luadia, B., & Gordon, A. (2009). Fish as the
for the poor or in maintaining or restoring the environmen- Bbank in the water^ - evidence from chronic-poor communities in
tal sustainability of the sectors. We also stress that the ability to Congo. Food Policy, 34, 104118.
meet the potential average fish demands of 9 billion people Bn, C., Hersoug, B., & Allison, E. H. (2010). BNot by rent alone^:
masks inequalities and inequities in who eats the fish and who analysing the pro-poor functions of small-scale fisheries in develop-
ing countries. Development Policy Review, 28(3), 325358.
benefits from the value chains. At present, people in Africa, Betru, S., & Kawashima, H. (2009). Patterns and determinants of meat
the poor in many societies, and women and minority groups, consumption in urban and rural Ethiopia. Livestock Research for
including small-scale fishing and aquaculture communities Rural Development, 21(9/143).
are in tension with large corporations and production units Beveridge, M. C. M., Thilsted, S. H., Phillips, M. J., Metian, M., Troell,
M., & Hall, S. J. (2013). Meeting the food and nutrition needs of the
over access to fish and fish-related employment (HLPE
poor: the role of fish and the opportunities and challenges emerging
2014). We have shown that the best available projections for from the rise of aquaculture. Journal of Fish Biology, 83, 10671084.
fish supply and demand are relatively positive in terms of the Blanchard, J., Jennings, S., Holmes, R., Harle, J., Merino, G., Allen, I.,
capacity to meet future demands, although more solid model- et al. (2012). Potential consequences of climate change on primary
ling is still needed to better incorporate demand projections. production and fish production in 28 large marine ecosystems.
Philosophical Transactions of the Royal Society, B: Biological
Climate change impacts on fish production also create uncer- Sciences, 367(1605), 29792989.
tainty in the projections. Bonham, M. P., Duffy, E. M., Robson, P. J., Wallace, J. M., Myers, G. J.,
In conclusion, fish should certainly be on the menu. Davidson, P. W., et al. (2009). Contribution of fish to intakes of
micronutrients important for foetal development: a dietary survey
of pregnant women in the Republic of Seychelles. Public Health
Acknowledgments This paper was part of a workshop sponsored by Nutrition, 12(09), 13121320.
the OECD Co-operative Research Programme on Biological Resource
Boyd, C. E., Tucker, C., McNevin, A., Bostock, K., & Clay, J. (2007).
Management for Sustainable Agricultural Systems.
Indicators of resource use efficiency and environmental perfor-
mance in fish and crustacean aquaculture. Reviews in Fisheries
Open Access This article is distributed under the terms of the Creative
Science, 15, 327360.
Commons Attribution License which permits any use, distribution, and
reproduction in any medium, provided the original author(s) and the Caddy, J. F., & Seijo, J. C. (2005). This is more difficult than we thought!
source are credited. The responsibility of scientists, managers and stakeholders to miti-
gate the unsustainability of marine fisheries. Philosophical
Transactions of the Royal Society London B: Biological Sciences,
360(1453), 5975.
References Cheung, W. W. L., Lam, V., Sarmiento, J., Kearney, K., Watson, R.,
Zeller, D., & Pauly, D. (2009). Large-scale redistribution of maxi-
mum fisheries catch potential in the global ocean under climate
Allison, E. H. (2011). Aquaculture, fisheries, poverty and food security. change. Global Change Biology, 16, 2435.
Working Paper 201165. Penang: WorldFish Center. Cheung, W. W. L., Dunne, J., & Sarmiento, J. L. P. D. (2011). Integrating
Allison, E. H., Bn, C., & Andrew, N. L. (2011). Poverty reduction as a ecophysiology and plankton dynamics into projected maximum
means to enhance resilience in small-scale fisheries. In R. S. fisheries catch potential under climate change in the Northeast
Pomeroy & N. L. Andrew (Eds.), Small-scale fisheries management Atlantic. ICES Journal of Marine Science, 68, 10081018.
frameworks and approaches for the developing world (pp. 216 de Schutter, O. (2012). The right to food - Note to the General-Secretary
238). Wallingford: CABI. from the Special Rapporteur on the right to food. New York: United
Allison, E. H., Delaporte, A., & Hellebrandt de Silva, D. (2013). Nation, Sixty-seventh session General Assembly.
Integrating fisheries management and aquaculture development Delgado, C. L., Crosson, P., & Courbois, C. (1997). The impact of live-
with food security and livelihoods for the poor. Report submitted stock and fisheries on food availability and demand in 2020. MSSD
to the Rockefeller Foundation. Norwich: School of International Discussion Paper, No.19. Washington, DC: IFPRI.
Development, University of East Anglia. Eide, A., Bavinck, M., & Raakjr, J. (2011). Avoiding poverty:
Arthur, R., Bn, C., Leschen, W., & Little, D. (2013). Fisheries and Distributing wealth in fisheries. In J. Svein & A. Eide (Eds.),
aquaculture and their potential roles in development: an assessment Poverty mosaics: Realities and prospects in small-scale fisheries
of the current evidence. London, UK: Marine Resources Assessment (pp. 1325). Dordrecht: Springer.
Group Limited (MRAG). (http://r4d.dfid.gov.uk/pdf/outputs/ FAO. (2012). The state of world fisheries and aquaculture 2012. Rome:
fisheries/61091-Fisheries_and_Aqua_Evidence_Review.pdf). FAO.
Editorial
Nutrigenomics and the future of dietetics
The role of the genome and the interaction of the environ- genomes of numerous other organisms have been
ment on the genome and its products are of increasing sequenced and the focus has shifted to determining the
importance to health care. Genetic principles are now rec- products of these sequences and their roles in the function
ognised as being fundamental to health and disease, a real- of organisms.
isation that is leading to a change in focus from disease
One of the earliest practical applications derived from
management to disease prevention and health promotion.1
the Human Genome Project has been in the field of phar-
Not surprisingly, it is the integration of genetic principles
macogenomics, the study of genetic influences on drug
and technologies into the food and nutrition sciences, the
metabolism.9 For any given pharmaceutical, there are indi-
emerging field called nutrigenomics, that will be critical to
viduals within a population who benefit from the drugs
realising the full potential of applying genomics to health
action, those who derive little benefit and those who expe-
care. Nutrigenomics is concerned with how dietary com-
rience harmful side-effects. The underlying basis for this
ponents interact with genetic sequences and their products
heterogeneous response within a population is the genetic
to impact the functioning of the organism, particularly with
heterogeneity with respect to key drug-metabolising
respect to health and disease.25
enzymes. Specific genetic variants of these enzymes have
Dietetics professionals will play a central role in dis- been associated with particular drug responses. Diagnostic
ease management and prevention in the years ahead in tests are now available that allow clinicians to detect which
ways never before possible.6,7 Clinicians will possess genetic variants an individual carries and predict that indi-
knowledge and skills unique within health care. Those in viduals response to a particular drug. Pharmacogenomics
the community health and public policy arenas will under- points up the advisability of matching pharmaceutical
stand the molecular basis for ethnic differences in nutrient therapy with the drug-metabolising capabilities of the indi-
requirements and will be in a position to develop effective vidual, abilities that derive from the individuals genetic
dietary approaches tailored to each ethnic group. Dietetics make-up.
professionals in the food sciences will use nutrigenomics
In many respects, nutrigenomics parallels the develop-
to identify and isolate the bioactive components in food
ment of pharmacogenomics. As with pharmaceuticals, a
and will develop health-promoting functional foods. The
number of key enzymes and metabolic reactions are
need for education of consumers and health care profes-
involved with the digestion, absorption and metabolism of
sionals will be extensive and ongoing. Clearly, nutrige-
food. Genetic variations in these key processes lead to a
nomics will have a significant impact on the field of
populations heterogeneous response to diets in general and
dietetics.
specific food components in particular. The ability to
The concept of nutrigenomics is not new. What is new match foods to the genetically determined capability of
is the appreciation of its integral role in health and disease. using that food for positive benefit is crucial to an
The biological world has long benefited from the practical individuals optimal health.10,11
applications of this concept. From primitive unicellular
However, nutrigenomics differs from pharmacogenom-
microorganisms to the most highly evolved mammals,
ics in several important aspects. Many decades of pharma-
genetic mechanisms exist that allow information from the
cological research have led to a strong foundation of
environment surrounding a cell to be transmitted to the
scientific knowledge, which has given rise to the practical
genetic material deep within that cell. In elaborate signal-
applications of pharmacogenomics. In contrast, nutrition
ling cascades that impact gene expression, bioactive com-
research is in its infancy. Unlike drugs, food is a complex
ponents from food serve to inform the organism about the
mixture of numerous components. Furthermore, an indi-
nutritional sufficiency of its environment. These same bio-
viduals genetic material has been interacting with food
active dietary components supply an organism with essen-
components for its entire lifetime, which may have lasting
tial nutrients that, because of genetic mutations, the
physiological consequences as compared with the limited
organism cannot synthesise and depends upon the environ-
druggenome interactions. These important differences
ment to supply. The evolutionary process itself has been
will likely slow the entry of nutrigenomics into wide-
significantly impacted by the goodness-of-fit between the
spread application compared with the rapid entry of
genetic make-up of an organism and the food present in its
pharmacogenomics.
environment.
By David Tribe
Senior Lecturer in Food Biotechnology and Microbiology, Agriculture and Food Systems at
University of Melbourne
More Events
Field testing genetically modified crops is a necessary step of the process, but who makes
sure its all safe? Andreas Beer/Flickr, CC BY-NC-SA
With the advent of crop genetic engineering in the 1980s, public controversy and intense
public scrutiny over genetically modified (GM) foods meant that the trial and error method of
discovering whether new GM foods were safe became unacceptable.
In Australia, this occurred during the major policy reform of all Australian food safety
regulation, through the revision of the national Food Standards Code (1994-2002).
During this period, safety assessment of GM foods became the responsibility of Food
Standards Australia New Zealand (FSANZ). Section 1.5.2 of the Food Standards Code
defines the requirements for the compulsory pre-market assessment and labelling of foods
produced by gene technology.
The presence of thousands of potentially toxic natural substances in plant foods meant that
the safety assessment of GM foods had to be done by different procedures than those that had
worked well with food additives and synthetic pesticides.
Regulators realised that if the hundreds of thousands of natural substances present in plant
foods such as soybeans or maize were subjected to the same standards used with food
additives and synthetic pesticides, all food GM or non-GM would fail regulatory
standards.
a GM food and
an otherwise comparable non-GM food that can be presumed to be already safe
because of its history of safe use.
If no meaningful differences are detectable in a new crop variety, the GM food can be
assumed to be at least as safe as its non-GM counterpart.
The standards agency FSANZ provides many further details of how safety assessment of GM
foods is carried out, before they are allowed to enter the commercial marketplace.
Box 4.1 from their introductory booklet on the topic gives the main features of the process.
Codex has established international guidelines for the safety assessment of foods derived
from modern biotechnology. The Australian protocols for the pre-market safety assessment of
GM foods are congruous with Codex principles.
Codex emphasises the use of sound scientific evidence for food risk assessment, and the
separation of risk policy and risk management from scientific risk assessment. In Australia,
the management of GM food safety follows these principles.
GM safety assessment and food standards development are carried out by FSANZ.
Genetically modified foods have now been in the marketplace for nearly two decades without
any harmful effects identified.
The absence of any evidence pointing to the lack of food safety in the recent farmer court
case in Western Australia (fought over the unintended presence of genetically modified
canola plants on another farmers property), underlines the overall conclusion that there is no
credible evidence of any food safety risk with the GM foods that have been approved so far.
More tangible problems of food safety such as the mould toxins that can ruin staple grains
and cause cancer in developing countries should now get more attention
GM oilseed crops and the Australian oilseed industry
Ruth Holtzapffel, Hilary Johonson and Osman Mewett
Executive Summary
Oilseed crops are The worlds major oilseed crops are soybean, rapeseed (including
important to Australian canola), peanut, oil palm and sunflower. Australia produces
agriculture. between 2 and 3 million tonnes of oilseeds each year, with canola
and cottonseed being the major crops. These two crops account for
92% of Australias total oilseed production with peanuts, soybeans
and sunflower seeds accounting for the remaining 8%.
The gross value of oilseed production averaged around 7% of the
total gross value of Australian broadacre crop production over the
three years to 2005/06. Over the same period, Australian exports of
oilseeds averaged around 8.5% of the total value of Australian
exports of grain and oilseeds.
Some oilseeds also provide benefits to wheat and barley crops
through their use in rotation cropping cycles.
Oilseeds are widely used Vegetable oils are widely consumed in our diets, including in
in food, feed and margarines and processed and fried foods. Meal derived from
industrial applications oilseeds is an important source of protein in the stockfeed industry
and vegetable oils are also used in the manufacture of soaps,
industrial lubricants and paints.
and have been Oilseeds have been selected over many generations to contain
extensively modified to desirable fatty acids to suit their different uses. For example, canola
meet these needs. was developed by selective breeding to contain low levels of anti-
nutritional compounds. Other modifications have included breeding
canola and sunflower varieties with high levels of monounsaturated
fats to improve the frying quality of their oils. Chemical
modifications can also be made to purified vegetable oils to make
margarines or to improve their stability for cooking and industrial
purposes.
Modifications are The 2005 BRS report Whats in the Pipeline? highlighted the
continuing to produce significant amount of research currently being conducted on
oilseeds that are easier genetically modified (GM) oilseed crops. This includes a number of
and cheaper to grow traits designed for easier, cheaper and more sustainable production
(e.g. herbicide tolerance, insect resistance, increased yields). GM
cotton varieties with herbicide tolerance and insect resistance have
been widely adopted in Australia and provide benefits to the
environment and farmers.
GM oilseed crops 5
produce healthier A significant amount of research is also being directed into
oils developing healthier oilseeds, particularly those which produce high
levels of polyunsaturated or long chain omega-3 fatty acids; higher
levels of essential amino acids and/or vitamins; or decreased
allergenicity. These improvements and modifications are expected
to provide health benefits for people throughout the world and to
add value to oilseed crops. There are commercial drivers for this
research due to public health policies recommending consumption
of long chain omega-3 fatty acids, industry pressures to reduce
processing costs and the increasing appeal of healthy foods
for consumers.
and may produce Research is also underway to produce oilseed crops designed to act
pharmaceutical and as biofactories, producing pharmaceutical or industrial compounds
GM oilseed crops 6
industrial products in instead of food, feed or fibre. Examples include pharmaceuticals
the future. (antibodies, vaccines or enzymes) and industrial compounds
(biofuels, bioplastics, lubricant oils or enzymes). These
developments are reviewed in more detail in the 2007 BRS report
Plant molecular farming in Australia and overseas.
GM oilseed crops that Combining a literature review and extensive consultations with
have been developed, or representatives of the oilseed industry and State Government
are in development, in Agencies, BRS has identified a range of GM oilseed crops that are
Australia include already in commercial production overseas or are under
herbicide tolerant and development in Australia and overseas.
insect resistant cotton, In Australia, GM herbicide tolerant and insect resistant cotton
and herbicide tolerant varieties and GM herbicide tolerant canola have been approved for
canola commercial release. Research is continuing into improving the fatty
acid profiles (e.g. high oleic and/or low linolenic acid levels) and
developing novel fatty acid compositions (e.g. long chain omega-3
fatty acids) for Australian oilseed crops.
while overseas GM Overseas countries have approved the commercial release of a
cotton, canola and number of GM oilseed crops, namely cotton, canola and soybean
soybean have been varieties incorporating herbicide tolerance, insect resistance or a
commercialised. combination of the two. Overseas researchers are developing
oilseed crops with improved environmental stress tolerance,
improved disease resistance, increased nitrogen use efficiency,
increased seed size, reduced pod shatter, increased yield and
increased oil content. Further research overseas is investigating
ways in which to produce oilseed crops with altered fatty acid
profiles or novel fatty acid composition, increased essential amino
acid content (e.g. methionine and lysine), production of essential
vitamin precursors (e.g. Vitamin A and E), increasing the
nutritional value of stockfeed (e.g. by reducing anti-nutritional
compounds such as phytic acid), and as expression platforms for
pharmaceutical and industrial products (e.g. plant-made vaccines or
bioplastics).
Oilseeds play a vital role The stakeholder consultations highlighted the important place that
as rotational break oilseeds have in Australian agriculture. In particular, canola
crops for wheat and provides value as rotational break crops to our major crop exports:
barley wheat and barley. Many people consulted were interested in the
potential for lower production costs (first generation traits), higher
value uses for oilseed crops such as specialty oils (second
generation traits), or crops as biofactories (third generation traits) as
potential ways to increase the value of oilseed crops. There was also
a desire for public breeding programmes to continue to ensure that
oilseed traits relevant to Australia are developed in the future.
and there are Canola and other oilseeds provide significant benefits to the
opportunities for Australian agricultural sector. A number of opportunities for
expansion and expansion and diversification of the Australian oilseed industry have
diversification. been identified, including: supplying increased quantities of local
oilseed meal to the stockfeed industry; developing stable oils for
frying that compete with palm and soybean oil imports; developing
value-added oils such as long chain omega-3 fatty acids to supply a
niche market for omega-3-containing oilseed crops; producing high
volumes of oil for use in biodiesel production; and
GM oilseed crops 7
meeting the needs of niche markets for GM and non-GM oilseeds as
they develop.
Experience with GM Experience with GM cotton crops in Australia and GM cotton and
oilseed crops in canola crops overseas has shown that there are benefits for farmers
Australia and overseas and the environment from growing these crops. For instance, in
show economic, Australia the GM insect resistant cotton Bollgard II led to an 85%
environmental and reduction in insecticide use in comparison to conventional cotton
agronomic benefits. over the first three seasons following its introduction in the 2002/03
season. The implementation of other Integrated Pest Management
practices has also contributed to this reduction.
Results from studies overseas show net economic benefits for many
farmers growing GM crops. For GM cotton, the level of this benefit
varies between countries and regions within countries due to
differences in environmental and climatic conditions and in some
countries, the way that GM crops are developed and sold.
In Canada, a number of agronomic benefits have been associated
with the adoption of GM canola, including: improved yields;
decreased herbicide use and increased weed management options;
and increased ease in adopting minimum and no-till cultivation
practices.
A study by the Canola Council of Canada released in 2005 indicated
that Canadian canola farmers who chose to grow GM varieties were
better off when compared to those who continued to cultivate
conventional canola varieties. Canadian growers have not lost
market share in their main export markets despite the majority of
their canola crop being comprised of GM varieties. There is no
evidence that GM canola is having difficulty finding ready markets
throughout the world. No price premium on bulk non-GM canola
shipments has been identified.
Current challenges faced Consultations with the Australian oilseed industry identified a
by the oilseed industry in number of current issues, perhaps the most immediate of which is
Australia include that posed by the current drought. Canola production in the 2006/07
drought, competition season is estimated to be the lowest in a decade with New South
from soybean and palm Wales the worst affected state.
oils The development of soybean and palm oil varieties that produce oil
of similar quality to canola is of concern to the oilseed industry as
both these oils are available in large volumes and at low prices on
the world market. Developments such as these may affect the
competitiveness of the Australian oilseed industry.
and the inability to The stakeholder consultations also identified a number of important
adopt GM food crops challenges that will need to be addressed if GM oilseeds are to be
such as canola due to widely grown in Australia. The most important of these are the
moratoria. current State and Territory Government moratoria that restrict the
commercial growing of GM crops in many parts of Australia.
Other challenges identified by stakeholders were the need to
address the perceived lack of public acceptance of GM
technologies; and supply chain management issues such as
segregation and coexistence. Many stakeholders believed that the
introduction of GM oilseeds was necessary for the future viability
GM oilseed crops 8
Chapter 1 Introduction
This report is the result of a study funded by the Department of Agriculture, Fisheries and Forestry
(DAFF) using funds provided under the National Biotechnology Strategy to investigate the potential
impacts of genetically modified (GM) oilseed crops on the Australian oilseed industry. The aim of
this work is to enhance the understanding of applications of biotechnology in the agriculture and
food sectors. This report identifies potential effects of GM oilseeds on Australian agriculture.
Research for this report was based on a combination of a desktop review of published literature,
review of web-based information on Australian and international trials of GM crops and
consultation with the Australian oilseed industry. These consultations focused on farmers,
researchers, seed suppliers, food and feed manufacturers.
Definitions of the term GM vary, but for this report, the term refers to plants that have acquired
new genes by laboratory gene technology methods, as defined in the Gene Technology Act 2000
(Cth).
GM oilseed crops 3
is relatively low for peanut, cottonseed and rapeseed. These are mainly produced for Australian
consumption (Table 1.1).
BOX 1 Canola versus Rapeseed whats the difference?
Canola is the name for Brassica oilseed crops that have less than 2% erucic acid (a fatty acid) and
less than 30 micromoles of glucosinolates per gram of seed solids. Canola varieties were developed
in Canada in the 1950s from Brassica napus and B. campestris (Canola Council of Canada, 2005a;
Tribe and Kalla, 2005). It is also known as double zero or double low rapeseed (Gunstone,
2004; Canola Council of Canada, 2005a).
Prior to the development of canola-type varieties, rapeseed oil was used both as a food oil and in a
variety of industrial purposes, with a long history of use as lamp oil (Tribe and Kalla, 2005).
Following feeding experiments with rats that suggested that high dietary exposure to erucic acid
could lead to fatty deposits and lesions in heart muscle, it was decided to decrease the levels of
erucic acid in rapeseed oil (FSANZ, 2003a). Glucosinolate is an anti-nutrient that can reduce
uptake of iodine in rats (Duncan, 1991).
The term rapeseed is used in many countries to describe oilseeds produced from various Brassica
species, including B. napus, B. rapa and B. campestris. Where the oil produced from these species
has low erucic acid and the meal has low glucosinolate levels, it can be called canola. Canada and
Australia are the main users of this term. When countries report rapeseed production levels, the
figures will often also include Indian mustard (B. juncea).
Australia mainly grows canola varieties of rapeseed and the majority of GM research is occurring
in these varieties. Therefore, the term canola is used frequently throughout this report. However,
when describing world production figures, rapeseed is used as this is the most common term.
Soybean is produced mainly in the United States, South America (predominantly Argentina and
Brazil) and China (Appendix A). From Table 1.1 and Appendix D, it is clear that Australia grows
relatively small quantities of soybean and is a net importer of this oilseed.
Although Australia is also a relatively minor rapeseed producer, in the three years to 2005-06 it
accounted for 19% of world canola exports (if intra-European Union trade is excluded) (Foster and
French, 2007). The countries that produce the largest amounts of rapeseed (the European Union,
China and India) are usually net importers as they have large domestic markets. Production, supply
and demand figures for world rapeseed (Appendix B) indicate that Canada is Australias main
competitor for rapeseed markets. The major global import markets for rapeseed are shown in
Figure 1.1.
Although cotton fibre is an important export commodity for Australia, the export market for
cottonseed is small (Tribe and Kalla, 2005). Australia supplies approximately 28% of this market
(Appendix C). The United States and Australia are the major exporters of cottonseed. The major
import markets for cottonseed are shown in Figure 1.2. Australia is not a significant exporter of
oilseeds other than canola and cottonseed.
GM oilseed crops 4
Japan (41%)
Mexico (20%)
Pakistan (13%)
China, Peoples Republic of, (8%)
United States, (7%)
EU-25 (4%)
Canada (3%)
Bangladesh (2%)
Figure 1.1 Relative sizes of global import markets for rapeseed (based on production
20032006).
Mexico (30%)
Japan (17%)
EU-25 (17%)
Korea, Republic of (12%)
Turkey (10%)
South Africa, Republic of (7%)
Saudi Arabia (4%)
Others (2%)
Figure 1.2. Relative sizes of global import markets for cottonseed (based on production
20032006).
GM oilseed crops 5
Section 1.3 Oilseeds are important to the Australian economy
The gross value of oilseed production averaged $766 million over the three years to 2005/06;
around 7% of the total gross value of Australian broadacre crop production (ABARE pers. comm.,
2006). Over the same period, Australian exports of oilseeds averaged $478 million, around 8.5% of
the total value of Australian exports of grain and oilseeds (ABARE, 2006). Oilseed production also
provided benefits to wheat and barley crops through their use in rotation cropping cycles.
Appendix D presents the amounts of oilseeds, wheat and barley produced in Australia,
demonstrating the much larger volume of cereal production.
Meal derived from oilseeds is an important source of protein in the stockfeed industry. Canola and
soybean meals are used extensively and cottonseed meals are also important. Sunflower meal is
used to a minor extent (see Table 3.1 Chapter 3). The stockfeed industry currently imports
significant volumes of soybean meal, as there is insufficient production in Australia. For instance,
imports of soybean meal can be up to 300 kt per year. There could be opportunities within the
oilseed industry in Australia to expand to meet some of this demand with meal from other oilseeds,
provided that there was also a market for the oil (Stakeholder Consultations).
Although oilseed crops make a relatively small contribution to the Australian economy as a
domestic and export commodity, they have a high value as break crops for cereals, such as wheat
and barley, that are of greater value to the Australian economy. A break crop is used in a crop
rotation system in order to help control weeds and plant diseases. When planted as part of a crop
rotation, they can increase the yield of subsequent cereal crops (Angus et al., 1991; Angus et al.,
1999; Evans et al., 2003). The size of this impact by canola on the following wheat crop is variable.
For instance, over a four year period (19901994) the two-year gross margin of canola-wheat was
approximately 50% higher than for wheat-wheat, however, in 1995-1996, which were both good
cropping seasons, the two-year gross margin was only about 25% higher for canola-wheat than for
wheat-wheat (Evans et al., 2003). This benefit for the following cereal crop is a good reason to
encourage the continuation of oilseed crops in Australian crop rotations where they contribute to
the competitiveness and sustainability of Australian agriculture (Stakeholder Consultations).
GM oilseed crops 6
Table 1.1. World oilseed production and exports (2005/06)
Rapeseed Cottonseed Soybean Sunflower Palm fruit Peanuts
a
(MMT) (MMT) (MMT) (MMT) (MMT) (MMT)
b
Total oilseed 48.42 42.31 224.12 29.98 9.71 33.09
production c
130
Total oilseed 6.35 1.05 66.35 2.01 0.10b 1.83
exports
(13%) (2%) (30%) (7%) (1%) (6%)
(% production)
Total meal 25.37 14.24 143.90 11.10 5.06b 5.72
production
Total meal 2.23 0.45 48.48 3.43 3.76b 0.15
exports
(9%) (3%) (34%) (31%) (74%) (3%)
(% production)
Total oil 16.59 4.67 33.87 10.46 34.80d 4.93
production 4.20 b
GM oilseed crops 7
products and may also lead to the development of new applications due to the unique properties
present in some of the more unusual fatty acids (EPOBIO, 2006).
GM oilseed crops 8
Other oilseeds have been similarly modified, with a GM high oleic acid cottonseed variety
developed and trialled in Australia by the Commonwealth Scientific and Industrial Research
Organisation (CSIRO) (Liu et al., 2002a; Liu et al., 2002b; OGTR, 2003f). Conventionally bred,
high oleic sunflowers (e.g. Monosun and Sunola) have also been developed and are available
commercially; they are produced in a segregated supply chain in Australia.
Some oilseed modifications can be made using conventional breeding techniques, however, some
cannot and for these, GM techniques are required if these desirable traits are incorporated into the
plants. An example of this is the production of omega-3 fatty acids in oilseeds discussed in Chapter
2.
Stable Unstable
Figure 1.3 Relationship between selected oils and their nutritional and functional
attributes.
Source: Figure modified from Carr (2005).
As shown in Figure 1.4, the changes being made to the fatty acid profiles of oilseed crops result in
fatty acid compositions that are increasingly similar to one another, which is likely to lead to
increased substitutability between these oils, dependant on relative availability and price, as
discussed below in Section 1.5.
GM oilseed crops 9
Canola
Cottonseed
Soybean
HO Canola
HO Cottonseed
HO Soybean
Figure 1.4 Relative fatty acid compositions of conventional and high-oleic (HO)
cultivars of major oilseed crops.
Source: Singh et al (2005)
GM oilseed crops 10
been used in the past to manufacture soap (Nyberg, 1970). The degree to which substitution occurs
depends on both the composition and price of the oils in question.
Soybean and canola oils are readily interchangeable and this substitutability will increase with the
growing availability of newly developed soybean varieties with oil profiles closer to that of canola,
particularly high oleic varieties. As demonstrated in Figure 1.4, the current trend towards
developing high oleic acid cultivars is resulting in increased similarity of the oil profiles for
cottonseed, soybean and canola. Soybean dominates the world oilseed market due to its lower
production costs, higher quality meal and increases in oil quality, resulting from both conventional
breeding and GM technology, which will increase the value of its oil too. This could be a significant
threat to the Australian oilseed industry in the future (Stakeholder Consultations). This
is discussed further in Chapter 3 of this report.
Palm oil is widely used in the food services industry as it is very cheap; there are large volumes
available; and it has good properties for cooking applications. Its high level of saturated fatty acids
means that it has a high level of thermal stability and a long shelf life. As a result, it does not
require hydrogenation and there are no trans-fatty acids associated with it. However, palm oil
contains higher levels of saturated fats than many other vegetable oils (Gunstone, 2004). Health
professionals recommend minimising the consumption of saturated fatty acids to decrease the risk
of developing coronary heart disease (European Commission, 2001; NHMRC, 2003; USDHHS and
USDA, 2005).
Substitution of palm oil with other sources of oil such as high oleic oils would allow Australian
oilseeds to gain a larger domestic market (Stakeholder Consultations). The main impediment to this
strategy would be the very low cost of palm oil; Australian production of alternatives is unlikely to
be price competitive. There are also limits to such substitution. As an example, there are some
popular biscuits manufactured in Australia for which there is no substitute for palm oil because the
texture of these biscuits is not acceptable when another oil is substituted (Stakeholder
Consultations).
GM oilseed crops 11
Section 1.6 Current opportunities and challenges for the
Australian oilseed industry
A range of people from the oilseed industry were interviewed using the questions presented in
Appendix E. These people were farmers, seed producers, stockfeed manufacturers, food
manufacturers and people working for relevant State Government agencies and had expertise in
different aspects of the Australian oilseed industry. In total, 21 people, representing 17
organisations were interviewed. Six other organisations were approached but declined to
participate. The organisations and people targeted for consultation were selected with the aim of
gaining a balanced view of issues facing the oilseed industry and the effects that GM oilseeds may
have on it. Thus people expressed both positive and negative opinions about GM oilseeds and the
range of views has been reflected in this report. The issues raised are summarised below and are
explored in more detail in Part 2 of this report.
opportunities for expansion of the oilseed industry in Australia
regulatory issues
potential for developing niche markets
potential for health benefits for consumers
agronomic, environmental and economic benefits from GM cotton
potential benefits from adopting GM canola
effects of seasonal variability on the oilseed industry
the necessity for canola to compete with improved soybean oil traits
consumer attitudes towards GM foods
segregation and coexistence strategies
effects of the State and Territory Government moratoria on growing GM food crops
liability issues.
GM oilseed crops 12
Full article go to http://www.ipsnews.net/2013/06/biofortification-may-hold-keys-to-hidden-hunger/
IPS News Agency
ROME, Jun 21 2013 (IPS) - The U.N. Food and Agriculture Organisation (FAO), which works to end
malnutrition among more than two billion people worldwide, is expressing strong support for enriching the
micronutrient content of plants.
In technical terms, it is called biofortification: a nutrition-specific intervention designed to enhance the
micronutrient content of foods through the use of agronomic practices and plant breeding.
The breeding is taking place at HarvestPlus, an international programme supported by the Consultative
Group on International Agricultural Research (CGIAR) and at national agricultural research centres, mostly
in developing countries.
The first nutritious crop, developed by African scientists and released in partnership with the Internal Potato
Center (CIP), was the orange sweet potato, which has been effective in providing up to 100 percent of daily
vitamin A needs for young children, according to CGIAR.
Six additional nutritious crops are now being developed through the use of conventional breeding methods:
vitamin A-rich cassava and maize, iron-rich beans and pearl millet, and zinc-rich wheat and rice.
The first three crops are targeted to Africa and the rest to South Asia.
New varieties of the first four crops were launched in 2012, says CGIAR, with wheat and rice expected to
follow later this year.
While it takes time to produce the amount of seed necessary to meet demand, up to half a million farmers
will be growing these nutritious crops by year end, it predicts.
Asked how far plant breeding can go in resolving hunger and nutrition problems worldwide, Dr. Erick Boy,
head of nutrition at HarvestPlus, told IPS, Our focus is on hidden hunger, caused by not getting enough
minerals and vitamins in the diet that is the major hunger problem the world faces today.
The six new varieties of staple crops we are developing are more nutritiousthey contain higher amounts
of vitamin A, zinc, or iron, he added.
Lack of these nutrients is what causes widespread suffering and health problems, especially for women and
children.
Boy said these crops will be distributed to more than three million farming households in seven countries in
Africa and Asia by 2015.
Not bad for a programme that started from scratch to develop these crops beginning only in 2003, he
noted.
When eaten regularly, these nutritious crops could provide on average 50 percent of vitamin A, zinc, or iron
requirements. According to CGIAR, more than two billion people worldwide do not get enough of these
crucial nutrients in their diets.
Deficiencies can lead to lower IQ, stunting, and blindness in children; increased susceptibility to disease for
both children and adults; and higher health risks to mothers and their infants during childbirth.
According to the World Bank, malnourished children are more likely to drop out of school and have lower
incomes as adults, thus reducing overall economic growth.
In its latest annual flagship publication The State of Food and Agriculture (SOFA) released here, FAO
explains that unlike food fortification, which occurs during food processing, biofortification involves
enriching the micronutrient content of plants.
Questions remain about the readiness of consumers to purchase biofortified foods, especially when they look
or taste different from traditional varieties. But, FAO says, early evidence suggests that consumers are
willing to buy them and may even pay a premium.
In Uganda, FAO discovered consumers were willing to pay as much for the orange-fleshed varieties of
sweet potato as for the white varieties, even in the absence of a promotional campaign.
Similar results were found for nutritionally-enhanced orange maize in Zambia, where consumers did not
confuse it with ordinary yellow or white maize. They were also willing to pay a premium when its
introduction was accompanied by nutrition information.
Asked why the project targets Asia and Africa and not Latin America, CGIARs Dr. Boy said, Our focus is
on subSaharan Africa and South Asia because if you look at any map of hidden hunger, these are the regions
marked in red.
Latin American countries have done a better job of improving nutrition over the past two decades, he added.
There are still, however, pockets where hidden hunger is a problem.
So we are also working in this region. In fact, I am in Guatemala now to work with stakeholders to buy in
to our high-iron beans and high zinc-maize initiative there. We anticipate that we could have varieties of two
to three crops that are rich in iron and zinc to LAC farmers by 2015, Boy added.
Meanwhile, in early June, the UK government granted 30 million [46.4 million dollars] to HarvestPlus to
develop and deliver six nutritious crops to several million farming households in Africa and Asia.
The grant was announced at a high-level international meeting in London that brought together a range of
partners to make strong political and financial commitments to improve nutrition globally.
In his opening remarks, British Prime Minister David Cameron said, It has to be about doing things
differentlyFor science, its about harnessing the power of innovation to develop better seeds, [and] more
productive and nutritious crops.
European Journal of Nutrition & Food Safety
5(1): 59-71, 2015, Article no.EJNFS.2015.003
ISSN: 2347-5641
SCIENCEDOMAIN international
www.sciencedomain.org
Authors contribution
This work was carried out in collaboration between all authors. Author IH headed the project. Author
MEW conducted the analysis and interpretation and drafted the manuscript. Author FW was involved
in the early work and assisted with the interpretation of the results. Author TH was in charge of data
management. All authors were involved in the critical revision of the manuscript for important
intellectual content. All authors read and approved the final manuscript.
Article Information
DOI: 10.9734/EJNFS/2015/12734
Editor(s):
(1) Prof. Hans Verhagen, Senior Scientific Advisor Nutrition and Food Safety, National Institute for Public Health and the
Environment (RIVM), P.O. Box 1, 3720 BA, Bilthoven, The Netherlands.
ABSTRACT
Aims: Using national food consumption data this paper addresses issues whether buyers of
organic food make healthier food choices and pursue a healthy lifestyle concerning smoking,
physical exercise and body weight compared to non-buyers.
Study Design: The German National Nutrition Survey II (NVS II) is a nationwide food consumption
study providing additional information on social demographics, health, and lifestyle aspects as well
as anthropometric measurements. Using data of several assessment tools, a comparison was
conducted between buyers and non-buyers of organic food.
Place and Duration of the Study: From November 2005 to November 2006, data collection took
place in about 500 randomly chosen sample points across Germany.
Methodology: 13,074 participants aged 18-80 years were divided into groups of buyers (44.9%)
and non-buyers (55.1%) of organic food. According to their purchase frequency, the organic food
buyers were further differentiated into intensive, occasional or infrequent purchase groups. A diet
history method was applied to assess food consumption, trained staff measured BMI while
questionnaires were used for social demographic description and healthy lifestyle factors.
____________________________________________________________________________________________________
Results: Buyers of organic food consumed 17% more fruit and 23% more vegetables as well as
less meat/sausages (25%) and soft drinks (58%) than non-buyers did (P< .001, resp.). These
results are more pronounced for women and for intensive buyers. Additionally, buyers of organic
food exhibit healthier lifestyle characteristics in respect to smoking behaviour, physical activity, and
body weight compared to non-buyers.
Conclusion: German buyers of organic food demonstrate health behaviours complying better with
the recommendations for a healthy lifestyle compared with non-buyers. Independent of the
discussion whether organically produced food exerts additional health effects, buyers of organic
food make healthier food choices than non-buyers, thereby benefiting for their overall health.
Keywords: Food consumption survey; organic food; food choice; healthy lifestyle.
as lifestyle characteristics between buyers and Social class was defined as an interrelation of
non-buyers of organic food in Germany. several factors, so an index was generated
based on the degree of education, the
2. METHODOLOGY employment status of the households principle
earner, and the monthly net income of the house-
2.1 Data Collection and Description of hold. Individuals were classified according to
Variables these characteristics by a point system [24]. The
answers to four questions, (1) arrangement of
food items perceived as important for a healthy
The baseline study NVS II was commissioned by
diet, (2) classification of probiotic yoghurt, (3)
the Ministry of Food, Agriculture and Consumer
classification of ACE-beverages and (4)
Protection and the dietary assessment referred
comprehension of the 5-a-day campaign, from
to in the present study was undertaken from
the questionnaire were combined to a NVS II
2005 until 2006 [24]. At about 500 randomly
nutrition knowledge index. Information about
chosen sample points across Germany, the local
smoking (yes/no) and physical activity (yes/no)
register offices randomly selected 14-80 year old
was retrieved according to statements from a
participants of which 18-80 year old participants
questionnaire. Body Mass Index (BMI) was
were integrated in this evaluation. Computer
calculated dividing weight into kilograms by
assisted personal interviews were conducted at
height in square meters. In order to classify the
study centers to obtain basic socio-demographic
BMI, the WHO definition was employed [30].
information followed by a computer assisted diet
history interview. The diet history method is a
The consumption of fruit, vegetables, meat and
dietary assessment instrument asking directly for
sausages as well as sweets and soft drinks
the habitual food consumption [26,27]. As
serves as indicator for a healthy food pattern.
software DISHES (Diet Interview Software for
Food consumption data were evaluated and
Health Examination Studies) was used, a
compared with the dietary guidelines of the
program developed by the Robert Koch-Institute,
German Nutrition Society [31,32].
Berlin [28,29]. The standardized interviews lead
participants through the sequence of meals of a
day followed by questions on the frequencies of 2.2 Data Analysis
consumed food. Participants also completed a
questionnaire on health-related aspects and For descriptive analysis, Pearsons Chi-square
leisure time activities. Additionally, anthropo- test was applied to 2x2 contingency tables,
metric measurements were conducted at the testing for independence of two criteria. For
study centers [24]. During data assessment, multiple comparisons, the Bonferroni correction
participants and interviewers were not aware of was subsequently used. Food consumption was
the future research question regarding organic calculated as arithmetic mean ( standard error
purchase. This diminishes a possible response of mean). The presentation of the means instead
bias. of medians was chosen because several food
groups were consumed by less than 50% of the
There was no specific differentiation for the participants. Consumption data did not show
consumption of organically or conventionally normal distribution and could not be normalized
produced food. By means of a questionnaire, by log-transformation, thus non-parametric tests
participants were classified into buyers and non- were carried out. The Mann-Whitney-Wilcoxon
buyers based on the question: Do you buy test was applied to compare food consumption of
organic food? (yes/no). Furthermore, buyers buyers and non-buyers of organic food. Kruskal-
were asked to give detailed information on the Wallis test was used to compare more than two
frequency of their purchase of twelve organic groups.
food items. The frequencies were coded with
points and a ratio was achieved by dividing the Statistically significant differences were defined
sums of points by the number of given answers at the p-level of 0.05. Calculations were
to correct for the possibility that no answer was performed with SAS program version 9.2. (SAS
given or a person did not eat/drink a food item. Institute Inc., Cary, North Carolina, USA).
Based on the ratios buyers were divided in either
the intensive, occasional or infrequent purchase
group.
Eisinger-Watzl et al.; EJNFS, 5(1): 59-71, 2015; Article no.EJNFS.2015.003
18-24 years 25-34 years 35-50 years 51-64 years 65-80 years
100%
20.3 *
22.0
80%
20%
P=.851
11.4
11.5
5.8 11.2
0% ***
Buyer Non-buyer
Fig. 1. Buyer (n= 5875) and non-buyer (n= 7199) of organic food by age group
*** P < .001, * P <.05
Comparisons are based on Chi square test, Bonferroni correction
Eisinger-Watzl et al.; EJNFS, 5(1): 59-71, 2015; Article no.EJNFS.2015.003
buyers and 44% of all male organic buyers guideline for meat and meat products is 300 to
consumed at least 250 g/d fruit (non-buyers: 45% 600 g per week [31,32]. Of all organic food
for women and 36% for men). There were buyers, 48% exceed the recommendation
significant differences between female organic compared to 64% of the non-buyers (P < .001).
buyers and non-buyers as well as between the This was the case for 33% of the intensive
male organic buyers and non-buyers (P < .001, buyers compared to 48% of the occasional and
resp.). 53% of the infrequent buyers (P < .001, resp.).
The proportion of women exceeding the
For women and men, mean consumption of recommendation was 35% for organic food
vegetables was 23% higher for the organic food buyers and 45% for non-buyers (P < .001) while
buyer group than for the non-buyer group the proportion of men exceeding the
(P < .001) (Fig. 3, by sex). The German dietary recommendation was 68% for buyers of organic
guideline for vegetable consumption of at least food and 81% for non-buyers (P <.001).
400 g per day [31,32] was reached by 18% of all
organic food buyers and by 10% of all non- All men regardless of buying organic food and in
buyers (P < .001). The proportion of meeting the which frequencies, consumed comparable
recommendation within the three frequency amounts of sweets while organic food buying
groups of organic food purchase was 27% for women ate fewer sweets (45 g/d) than non-
intensive, 19% for occasional and 14% for buying women (48 g/d) (P < .05) (Fig. 4). Female
infrequent buyers (P < .001, resp.). Differentiated intensive buyers consume the least sweets
by sex, 20% of the women and 16% of the men (41 g/d), followed by female occasional buyers
of the organic food buyers met the (45 g/d) and female infrequent buyers (48 g/d)
recommendation compared to 11% female and (P < .001, resp.). Buyers of organic food
10% male non-buyers (differences between consumed 58% less soft drinks (like soda and ice
female buyers and non-buyers as well as male tea) than non-buyers (P < .001) (Fig. 4, by sex).
buyers and non-buyers: P < .001, resp.).
However, a high proportion of all participants did
Organic food buyers consumed 25% less meat not drink soft drinks at all: 77% of organic food
and sausages than non-buyers of organic food buyers and 65% of non-buyers (P <.001).
(P < .001) (Fig. 3, by sex). The German dietary Considering only the consumers of soft drinks,
40% 32.0
*** 27.6
25.8 23.6 25.3
20%
15.6
9.9 *** 10.3 11.6
9.0
4.3 *** 8.1 4.5 3.9 5.0
0%
Buyer Non-buyer Intensive buyer Occasionally Infrequent
buyer buyer
Fig. 2. Buyer (incl. sub-groups) and non-buyer of organic food by social class, intensive
(n= 661), occasional (n= 3428), infrequent (n=1786)
*** P < .001; Comparisons are based on Chi square test, Bonferroni correction.
Eisinger-Watzl et al.; EJNFS, 5(1): 59-71, 2015; Article no.EJNFS.2015.003
the recorded quantities of the buyers were still 3.3 Health Aspects and Anthropometric
37% less than the quantities of the non-buyers of Measurements
organic food (P < .001). The average soft drink
consumption of women buying intensive, Buyers of organic food more often were non-
occasional and infrequent organic food were smokers and vegetarians and indicated more
163 ml/d, 226 ml/d and 260 ml/d, while women often to be physically active (Table 1). In
belonging to the non-buyer group consumed addition, organic food buyers more often were
313 ml/d (P< .001, resp.). The corresponding interested in nutrition information and achieved a
consumption in men was 345 ml/d (intensive), higher score in the NVS II nutrition knowledge
310 ml/d (occasional), 329 ml/d (infrequent) and index. More frequently, they stated to be satisfied
505 ml/d for non-buyers (P < .001, resp.). with their personal health status. Within the buyer
group, the proportion of normal body weight was
Altogether, organic food buyers more often met
higher and the proportions of overweight and of
the food-based dietary guidelines of the German
obese participants lower than in the non-buyer
Nutrition Society [31]. Consequently, the
group. This was also the case for women and
absolute nutrient intake as well as nutrient
men respectively (with the exception of the male
densities of buyers of organic food corresponded
overweight participants) (Fig. 5).
better to the recommendations than those of the
non-buyers [33] (data not presented).
450
dietary guidelines
400 (DGE 2013)
350
300
***
250 *** *
g/day
200 ***
150
100
***
50 ***
0
Buyer Non- Buyer Non- Buyer Non- Buyer Non- Buyer Non- Buyer Non-
buyer buyer buyer buyer buyer buyer
Fig. 3. Mean consumption ( SE) of fruit, vegetables and meat/sausages of buyer and non-
buyer of organic food, by sex Buyer: women (n=3577), men (n= 2298); Non-buyer: women
(n= 3536), men (n= 5961)
*** P < .001; Comparisons are based on Mann-Whitney-Wilcoxon test.
250
200
150
***
100
50
P=.790
0 * ***
Buyer Non- Buyer Non- Buyer Non- Buyer Non-
buyer buyer buyer buyer
Fig. 4. Mean consumption ( SE) of sweets and soft drinks of buyer and non-buyer of organic
food, by sex buyer: women (n=3577), men (n= 2298); non-buyer: women (n= 3536),
men (n= 5961)
*** P < .001, * P <.05; Comparisons are based on Mann-Whitney Wilcoxon test.
100%
17.9 *** 22.5 17.7 *** 22.2 18.2 *** 22.6
80%
27.7 ***
35.5 *** 31.5
60% 39.2 47.6 d 46.6
40%
52.6 ***
45.1 *** 44.8
20% 37.2 33.5 *** 30.0
c
1.4 a 1.1 1.9 b 1.6 0.6 0.7
0%
Buyer Non-buyer Buyer Non-buyer Buyer Non-buyer
Fig. 5. Organic food buyer and non-buyer by body weight (BMI) and sex organic food buyer:
women (n=3577), men (n= 2298);Non-buyer: women (n= 3536), men (n= 5961)
*** P < .001, aP=.178, b P=.277, cP=.675, dP=.473;
Comparisons are based on Chi square test, Bonferroni correction
Eisinger-Watzl et al.; EJNFS, 5(1): 59-71, 2015; Article no.EJNFS.2015.003
4. DISCUSSION AND CONCLUSION not indicate any health factors [41]. Data of the
NHANES study in the United States show the
With the data of the NVS II, the nationwide impact of three lifestyle factors (healthy diet, non-
representative nutrition survey for Germany, a smoking and being physically active) with a risk
comparison between buyers and non-buyers of reduction of 65% for cardiovascular disease and
organic food was conducted focusing on food 83% for cancer [17]. Other studies also confirm
consumption and lifestyle factors. The results these findings but are even more specific for
show that, compared to non-buyers, German special diseases like heart disease [42] and
buyers of organic food eat more fruit and diabetes mellitus [43]. The four lifestyle factors
vegetables, less meat and sausages and (healthy diet, non-smoking, high level of physical
therefore demonstrate a more favorable food activity and normal body weight) also proved
choice. This food pattern is more in line with responsible for a risk reduction of 66% regarding
national and international recommendations the endpoint all-cause mortality [44]. Ford et al.
[31,34-38]. These results apply to both, women [17], found strong relationships for three lifestyle
and men, while women do even better. The more factors (healthy diet, non-smoking and being
intense the purchase of organic food the more physically active) and all-cause mortality with an
favorable the food choice turns out to be. 82% risk reduction compared to participants
Additionally, buyers of organic food decide to without an indication of health factors.
lead a healthier lifestyle as indicated by lower
body weight, higher level of physical activity and The outcome is not to be confused with the
less smokers than non-buyers. question whether health effects of organically
produced food compared to conventionally
The French web-based cohort study from a large produced food are evident. Related research
sample of volunteers identified five clusters, two concentrates on differences in health effects of
of them comprising occasional and regular foods from the two farming systems and is
organic food consumers [23]. The comparison of discussed elsewhere [45-51]. Instead, the data
food consumption, assessed with three 24-h set of the NVS II allows the comparison of food
recalls, between non-consumers and consumers consumption and lifestyle factors of buyers and
of organic food revealed that food patterns of non-buyers of organic food. The results show
organic food consumers show a higher overall better health behaviour of buyers of
agreement with dietary guidelines, which gets organic food.
even better with increasing frequency of
consumption. In addition, the presented results The presented finding that for the German
regarding body weight were confirmed by the population from 2005-2006 45% are buyers and
French study. 55% were non-buyers differs from results of the
kobarometer. The kobarometer is an
A higher percentage of normal body weight in the annual assessment initiated by the Federal
organic food group was also found in the Ministry of Food, Agriculture and Consumer
Norwegian study with pregnant women [39]. In Protection of Germany. For the years 2005 and
1998, a comparison in the Netherlands between 2007 the data collection indicated about 75% of
a sample of 271 customers of health food stores the interviewed participants (about 1.000
and a nationwide household sample (n=581) participants starting at age 14) buying organic
showed that organic food buyers displayed the food in different frequencies [52,53]. In contrast,
lowest BMI [40]. the German household study with a sample
of 13,000 totaled 50% as buyers for 2008 [21]
Healthy food choices, normal body weight, being showing an increase from 36% of the same
physically active, as well as non-smoking household panel in 2005 [54]. Thus, the finding
contribute substantially to a healthy lifestyle [18]. of the present study that 45% of for the German
There is a robust understanding of links between population bought organic food in the year 2006
healthy lifestyle factors and a lower risk of is in a comparable range with the household
developing diabetes, heart disease, and cancer. study 2008 regarding sample size and
In the EPIC-study, the lifestyle factors healthy percentage of organic purchase.
diet, non-smoking, high level of physical activity
and normal body weight are associated with a Socio-demographics are used to test how the
reduced risk of developing diabetes (93%), profile of the organic buyers of the NVS II
myocardial infarction (81%), stroke (50%) and compares to other international and German
cancer (36%) compared to participants who do studies. The finding that women are generally
Eisinger-Watzl et al.; EJNFS, 5(1): 59-71, 2015; Article no.EJNFS.2015.003
more engaged in health and nutrition has often public health point of view, organic food buyers
been demonstrated [10,14,23,55,56]. lead a way of life closer to the recommendations
Participants with a higher social status, i.e. for a healthy lifestyle and thereby profit by
higher level of education and income, have also scientifically well-established health benefits and
more often been linked to intensive organic risk reductions.
purchase behavior [2,23,54,57,58], while age is a
factor with no clear association [10,14,39,54]. In ACKNOWLEDGMENTS
summary, concerning socio-demographic
variables the findings of the present study are in Franziska Lindner (KIT, Dept. Mathematics,
concordance with other studies on organic Karlsruhe) was a consultant for methods of
buyers. analysis.
Organic agriculture has a history of being contentious and is considered by some as an inefficient approach to food production.
Yet organic foods and beverages are a rapidly growing market segment in the global food industry. Here, we examine the per-
formance of organic farming in light of four key sustainability metrics: productivity, environmental impact, economic viability
and social wellbeing. Organic farming systems produce lower yields compared with conventional agriculture. However, they are
more profitable and environmentally friendly, and deliver equally or more nutritious foods that contain less (or no) pesticide
residues, compared with conventional farming. Moreover, initial evidence indicates that organic agricultural systems deliver
greater ecosystem services and social benefits. Although organic agriculture has an untapped role to play when it comes to the
establishment of sustainable farming systems, no single approach will safely feed the planet. Rather, a blend of organic and
other innovative farming systems is needed. Significant barriers exist to adopting these systems, however, and a diversity of
policy instruments will be required to facilitate their development and implementation.
O
rganic agriculture has a history of being contentious. the latest innovations in feeding and handling livestock. Organic
Emblematic of this, and representing the prevailing attitudes farming systems range from strict closed-cycle systems that go
of many farmers and scientists in the 1970s and 1980s, are beyond organic certification guidelines by limiting external inputs
the unsympathetic words uttered in 1971by then US Secretary of as much as possible to more standard systems that simply follow
Agriculture Earl Butz: Before we go back to organic agriculture in organic certification guidelines.
this country, somebody must decide which 50 million Americans Rudolf Steiners 1924 course on biodynamic agriculture sparked
we are going to let starve or go hungry1. At the turn of the twenty- the evolution of organic agriculture in Europe1. Organic agricul-
first century, sceptics considered organic agriculture to be ideologi- ture was established in its own right in the 1930s and 1940s, being
cally driven and inefficient 2,3. They argued that organic agriculture developed in Britain by Lady Eve Balfour and Sir Albert Howard,
relies on more land to produce the same amount of food as con- in Switzerland by Hans Mueller, in the United States by J. I. Rodale
ventional agriculture and that adopting organic agriculture on too and in Japan by Masanobu Fukuoka1. By the 1970s, organic foods
large a scale could potentially threaten the worlds forests, wetlands had grown in popularity, prompting the first organic certification
and grasslands2,3. They also asserted that organic agriculture has too standards to be drafted in Europe and the United States, and com-
many shortcomings and poor solutions to agricultural problems2,4. mencing an ongoing evolution of certifiers that now includes 283
Organic agriculture is still considered by some critics as being an organic certification bodies worldwide operating in 170 countries7.
inefficient approach to food security 5,6 and a farming system that This proliferation of certifiers reflects both a complex history of
will become less relevant in the future6. sometimes competing independent standards and the demand for
Yet the number of organic farms, the extent of organically farmed access to certifiers around the world.
land, the amount of research funding devoted to organic farming and Many farms in both developed and less-developed countries
the market size for organic foods have steadily increased7. Sales of implement organic practices but are not certified organic. However,
organic foods and beverages are rapidly growing, increasing almost growers are increasingly turning to certified organic farming sys-
fivefold between 1999and 2013to US$72 billion (ref.7; Fig.1); this tems as a way to provide verification of production methods,
2013 figure is projected to double by 2018. Moreover, recent interna- decrease reliance on non-renewable resources, capture high-value
tional reports recognize organic agriculture as an innovative farm- markets and premium prices, and boost farm income. Although
ing system that balances multiple sustainability goals and will be of requirements vary slightly between certifying agencies, they pro-
increasing importance in global food and ecosystem security 810. mote soil quality, crop rotations, animal and plant diversity, bio-
Here, we review the performance of organic farming systems logical processes, and animal welfare, while generally prohibiting
in the context of sustainability metrics and global challenges, and irradiation, sewage sludge, genetic engineering, the prophylactic
examine some of the barriers to the adoption of organic farming use of antibiotics, and virtually all synthetic pesticides and fertiliz-
systems and the policies needed to overcome them. ers. Standards continue to evolve with changing technologies and
socioecological conditions; some requirements are based on scien-
Organic practices and certification tific evidence, whereas others are driven by ideology.
Organic agriculture, sometimes called biological or ecological As most certification standards originated in temperate devel-
agriculture, combines traditional conservation-minded farming oped countries, they are not always applicable in other regions, espe-
methods with modern farming technologies. It emphasizes rotating cially in less-developed countries. High demand for organic foods
crops, managing pests naturally, diversifying crops and livestock, in Europe and North America has resulted in the import of organic
and improving the soil with compost additions and animal and foods from large farms in less-developed countries7. Although pre-
green manures (Fig. 2). Organic farmers use modern equipment, mium prices for exported foods may be beneficial to farmers, the
improved crop varieties, soil and water conservation practices, and inaccessibility of many of these foods to local consumers raises
Department of Crop and Soil Sciences, Washington State University, Pullman, Washington 99164, USA.
*e-mail: reganold@wsu.edu
80 45
Sales, other Food, fibre and fuel for
Sales of organic food (US$billion)
Mineralization Mineralization
Soil Soil Soil Soil
Crop N Crop N
organic N mineral N organic N mineral N
Immobilization Crop N N export Immobilization Crop N N export
uptake in harvest uptake in harvest
Figure 3 | Hypothetical nitrogen stocks and flows of two contrasting cropping systems. a,b, Cropping systems relying mainly on mineral nitrogen inputs
(a) have relatively higher nitrogen losses to air and water than cropping systems with emphasis on biological N fixation, manure and other organic matter
amendments, cover crops and perennial crops, and low reliance on mineral N fertilizer, such as organic and integrated systems (b). The width of the
arrows is relative to the size of the nitrogen flux; boxes representing nitrogen stocks are not scaled to the pool size. Figure adapted from ref.100, 2015
The National Academies100. Arrows represent nitrogen inputs (green), losses (orange) and transformations (blue).
may present some clear advantages when it comes to synthetic pes- product 49,57,58. In a meta-analysis of environmental quality param-
ticide residues, the human health impacts of pesticide exposure eters, organic farms were found to have lower nitrate leaching,
from food are not clear 26, and organically certified pesticides need nitrous oxide emissions and ammonia emissions per unit of field
to be better identified and taken into account 28. area, but higher leaching and emissions per unit product 48. Severe
At least 15 reviews or meta-analyses26,27,3244 of the scientific degradation of freshwater and marine ecosystems around the world
literature comparing the nutrition of organic and conventional is linked to excessive use of nitrogen and phosphorous fertiliz-
foods have been published in the past 15 years. Twelve of these ers12,59, leading to eutrophication of freshwater and the production
studies27,3234,3639,4144 found some evidence of organic food being of hypoxic zones in coastal waters. Lower nutrient pollution from
more nutritious (for instance, having higher concentrations of organic compared with conventional systems can be illustrated by
vitamin C, total antioxidants and total omega-3 fatty acids, and differences in their nitrogen cycling and losses (Fig.3).
higher omega-3to -6 ratios). Whether or not these are nutritionally Organic systems are usually more energy efficient than their
meaningful differences continues to be debated26,43. The other three conventional counterparts4648,51,54,58. For example, in Germany, Italy,
studies26,35,40 concluded that there were no consistent nutritional dif- Sweden and Switzerland, organic farms were found to use signifi-
ferences between organic and conventional foods. However, one of cantly less energy on a per-hectare basis than their conventional
the three studies found that conventional chicken and pork had a counterparts, and 70% of organic farms and 30% of conventional
33% higher risk for contamination with antibiotic-resistant bacteria farms had significantly lower energy consumption per unit of out-
compared with organic alternatives26. put 45. The generally lower energy use4648,54 and higher soil organic
matter 4549 of organic systems make them ideal blueprints for devel-
Environment. Reviews and meta-analyses generally support the oping methods to limit fossil fuel emissions and build soil carbon
perception that organic farming systems are more environmentally stores, important tools in addressing climate change.
friendly than conventional farming systems4558. For example, such
aggregate studies have found that organic farming systems con- Economics. Whether organic agriculture can continue to expand
sistently have greater soil carbon levels, better soil quality and less globally will primarily be determined by its financial performance
soil erosion compared with conventional systems4551. In addition, compared with conventional agriculture17,60. The main factors that
organic farms generally have more plant diversity, greater faunal determine the profitability of organic agriculture include crop yields,
diversity (insects, soil fauna and microbes, birds) and often more labour and total costs, price premiums for organic products, the
habitat and landscape diversity 4655. Most functional groups, such as potential for reduced income during the organic transition period
herbivores, pollinators, predators and producers (plants), are more (usually three years), and potential cost savings from the reduced
diverse in organic farming systems5153. Moreover, in a study cov- reliance on non-renewable resources and purchased inputs61.
ering eight western and eastern European countries, insecticides To the best of our knowledge, only one meta-analysis has ana-
and fungicides had consistently negative effects on biodiversity, lysed the financial performance of organic and conventional agri-
with insecticides also reducing the biological control potential in culture20. The analysis combines findings from 40years of studies
farming systems56. covering 55 crops grown on five continents. When actual price
As organic agriculture uses virtually no synthetic pesticides, premiums (higher prices awarded to organic foods) were included,
there is little to no risk of synthetic pesticide pollution of ground organic agriculture proved significantly more profitable (22to 35%
and surface waters46. With respect to nitrate and phosphorous leach- greater net present values) and had higher benefit/cost ratios (20to
ing and greenhouse gas emissions, organic farming systems score 24%) than conventional agriculture. When organic premiums
better than conventional farming when expressed per unit produc- were taken away, net present values (27to 23%) net returns
tion area46,49,51,57,58; however, given the lower land-use efficiency of accounting for the time value of money and benefit/cost ratios
organic farming in developed countries, this positive effect is less (8 to 7%) of organic agriculture were significantly lower than
pronounced and in some cases reversed when expressed per unit conventional agriculture20.
Conventional Organic
Minimize Minimize
pesticide Biodiversity pesticide Biodiversity
residues residues
Reduce Reduce
worker Minimize worker Minimize
exposure to water exposure to water
pesticides pollution pesticides pollution
Employment Employment
Profitability Profitability
of workers of workers
Figure 4 | Assessment of organic farming relative to conventional farming in the four major areas of sustainability. Lengths of the 12 flower petals
are qualitatively based on the studies discussed in this Review1523,2529,3256,58,6274 and indicate the level of performance of specific sustainability metrics
relative to the four circles representing 25, 50, 75 and 100%. Orange petals represent areas of production; blue petals represent areas of environmental
sustainability; red petals represent areas of economic sustainability; green petals represent areas of wellbeing. The lengths of the petals illustrate that
organic farming systems better balance the four areas of sustainability.
Although price premiums were 29 to 32%, breakeven premi- awarded to organic products. Price premiums and European sub-
ums necessary for organic profits to match conventional profits sidies for organic farms are often justified on the grounds that they
were only 5to 7%, even with organic yields being 10to 18% lower. compensate farmers for providing ecosystem services or avoiding
The size of organic premiums awarded, and the difference between damage to the environment.
organic premiums and breakeven premiums, were consistent dur-
ing the 40-year study period. The fact that organic premiums were Wellbeing. How well organic, conventional and other farming
significantly higher than breakeven premiums suggests that organic systems are performing in areas such as social equity (for instance,
agriculture can continue to expand even if premiums decline. The issues of gender, race, ethnicity and class) and quality of life for farm
study also found that total costs were not significantly different, but families and communities remains unclear due to limited research.
labour costs were significantly (7to 13%) higher with organic farm- Available data indicate that both organic and conventional farming
ing practices20. Although one of the successes of conventional agri- systems need to make significant progress to meet social sustain-
culture has been its ability to produce more with less labour, some ability goals10. However, organic farming has been shown to have
have found the extra labour of organic agriculture to be beneficial some sociocultural strengths, such as positive shifts in community
in providing rural employment and development opportunities62,63. economic development, increased social interactions between farm-
Few economic studies have accounted for negative externalities ers and consumers71,72, and greater employment of farm workers and
(such as environmental costs) or positive externalities (such as eco- cooperation among farmers62,63.
system services), with associated monetary values, in organic and Although organic farming often requires additional manual
conventional comparison studies. Putting a price on the negative work on the farm, it reduces the exposure of farm workers to pesti-
externalities caused by farming, such as soil erosion or nitrate leach- cides and other chemicals. Such exposure can be particularly prob-
ing into groundwater, would make organic agriculture even more lematic in less-developed countries, where illnesses and death have
profitable, given that its environmental impact is less than that of resulted from occupational and accidental exposure (due in part
conventional agriculture4558. Indeed, it has been estimated that a to the fact that it is impractical and expensive for workers to use
switch to organic production would lower the external costs of agri- safety equipment)73,74.
cultural production in the United Kingdom by 75%, from 1,514 Organic certification programmes have adopted social wellbe-
million yr1 to 385 million yr1 (ref.64). ing goals. Guidelines of the International Federation of Organic
A number of studies (for example, refs 65,66) have compared Agriculture Movements (IFOAM) stipulate that organic farmers
ecosystem services in organic and conventional farming systems. should be able to support themselves and other workers with fair
A few of these studies have accounted for the monetary value of incomes, while maintaining safe and dignified working conditions75.
ecosystem services; these studies generally show that conventional Furthermore, organically certified animals must be raised humanely
practices decrease the ability of farms to provide some economically under conditions that allow for the expression of their natural behav-
significant ecosystem services relative to organic practices6769. For iours and needs75. For example, European Union, US and Japanese
example, in a study comparing 14 organic arable fields with 15 con- rules on organic production require livestock to have access to open
ventional ones in New Zealand70, the total economic value of three air or grazing whenever possible, and that sick animals be treated as
ecosystem services (biological pest control, soil formation and the needed, even with the use of antibiotics if required7678.
mineralization of plant nutrients) in the organic fields was signifi- Organic farming can improve food security by diversifying
cantly greater at US$232ha1yr1 compared with the conventional on-farm crop and livestock operations, which diversifies income
fields at US$146ha1yr1. Factoring in such differences in economic sources and improves variety in diets79. Organic farming necessi-
comparison studies would probably make up for price premiums tates diverse crop and livestock rotations, encourages the integration
to sustainable agricult
riers ure
Bar
stin g policies Lack
exi of i
and nfo
rm
ests at
er io
nt
n
i
an
d
ste
dk
l ve
no
Production Environment
wle
erfu
dge
Pow
Legal instruments
Measures to ensure open and
Financial instruments Sustainable competitive markets
n ge s
Increased farmer access to capital agriculture Reduction of commercial influence
c c h a ll e
Elimination of crop subsidies and in government
biofuel incentives Development of national targets and trade
M is p
omi
Increased costs of scarce resources Wellbeing Economics Enhanced security of land tenure
erc
co n
Targeted input subsidies for
ep
resource-poor farmers
re
ti o
he
Investment in rural infrastructure
ns
ot
d
an
cu
nd
a
l tu re
ral tu
b uc
ias e
s fra str
in
Weak
Knowledge-based instruments
Public investment in agroecological research
Farmer engagement in research and development
Outreach to improve farmer knowledge and capacity
Enhancement of women's educational and leadership opportunities
Increased transparency throughout the food chain
Figure 5 | Policy instruments for overcoming barriers to farmers adopting more sustainable farming systems. For any farm to be sustainable, it must
meet four goals, shown in the centre: (1) produce adequate amounts of high-quality food (production); (2) enhance the natural-resource base and
environment (environment); (3) be financially viable (economics); and (4) contribute to the wellbeing of farmers and their communities (wellbeing)10.
Despite the appeal of a sustainable agriculture philosophy, the task for farmers to achieve agricultural sustainability is challenging. Just because a farm
is organic does not mean that it is sustainable. However, research shows that organic farming systems better balance the four sustainability goals than
their conventional counterparts and are more likely to achieve agricultural sustainability (overlapping area). Yet, significant barriers to adopting organic
agriculture exist, including powerful vested interests and existing policies, lack of information and knowledge, weak infrastructure and other economic
challenges, and misperceptions and cultural biases. In fact, many of these same barriers exist for other innovative systems, such as agroforestry,
conservation agriculture, integrated farming and mixed croplivestock systems. A diversity of policy instruments is needed to overcome these barriers, and
can be categorized as financial, legal and knowledge-based instruments. Examples of these instruments are shown in the figure.
fastest growing segments of American agriculture and helps farmers organic and other innovative farming systems to move towards truly
receive a higher price for their product as they strive to meet grow- sustainable production systems. This is no small task, but the con-
ing consumer demand93. Moreover, organic agriculture has been sequences for food and ecosystem security could not be bigger. To
able to provide jobs, be profitable, benefit the soil and environment, make this happen will require mobilizing the full arsenal of effective
and support social interactions between farmers and consumers. policies, scientific and socioeconomic advances, farmer ingenuity
Although organic agriculture has an untapped potential role in and public engagement.
global food and ecosystem security, no one farming system alone
will safely feed the planet. Rather, a blend of organic and other inno- Received 7 September 2015; accepted 17 December 2015;
vative farming systems, including agroforestry, integrated farming, published online 3 February 2016
conservation agriculture, mixed crop and livestock, and still undis-
covered systems, will be needed for future global food and ecosys- References
tem security. For example, integrated farming systems that blend 1. Lockeretz, W. in Organic Farming: An International History (ed. Lockeretz, W.)
mostly organic with some conventional practices have been shown 18 (CABI, 2007).
2. Trewavas, A. Urban myths of organic farming. Nature 410, 409410 (2001).
to be more sustainable than conventional farming systems94,95 and
3. Emsley, J. Going one better than nature? Nature 410, 633634 (2001).
are likely to play a central role. Achieving global food and ecosys- 4. Kirchmann, H. & Thorvaldsson, G. Challenging targets for future agriculture.
tem security requires more than just achieving sustainable farming Eur. J.Agron. 12, 145161 (2000).
systems worldwide. We need to reduce food waste, improve food 5. Connor, D.J. & Mnguez, M.I. Evolution not revolution of farming systems
distribution and access, stabilize the human population, eliminate will best feed and green the world. Glob. Food Secur. 1, 106113 (2012).
the conversion of food into fuel, and change consumption patterns 6. Pickett, J.A. Food security: intensification of agriculture is essential, for which
current tools must be defended and new sustainable technologies invented.
towards a more plant-based diet.
Food Energ. Secur. 2, 167173 (2013).
Equal adherence to all four sustainability goals of production, 7. Willer, H. & Lernoud, J. (eds) The World of Organic Agriculture: Statistics and
environment, economics and social wellbeing does not limit but Emerging Trends 2015 (FiBL-IFOAM, 2015).
encourages farmers and researchers to innovate. The challenge fac- 8. International Assessment of Agricultural Science and Technology for
ing policymakers is to create an enabling environment for scaling-up Development Agriculture at a Crossroads: Global Report (Island, 2009).
FJ
18,2
Customer value perception of
organic food: cultural differences
and cross-national segments
Barbara Seegebarth
eived 2 July 2015
Institute of Marketing, Technische Universitt Braunschweig,
ised 4 November 2015 Braunschweig, Germany
epted 4 November 2015
Stefan Henrik Behrens, Christiane Klarmann and Nadine Hennigs
Marketing and Management, Leibniz University of Hannover,
Hannover, Germany, and
Lisa Luebbehusen Scribner
Cameron School of Business, University of North Carolina Wilmington,
Wilmington, North Carolina, USA
Abstract
Purpose Due to consumer concerns about food-related diseases and an increase in the use of
genetically modified food, more and more green consumers integrate environmental considerations
into daily purchases, asking for healthier, safer and higher quality food. Marketing managers still face
the challenge of broadening the understanding of how and why consumers purchase organic food.
Specifically, a deeper understanding of the value dimensions consumers perceive in the context of
organic food products is required to develop and implement successful management strategies which
might transfer positive consumer perceptions to actual buying behavior and satisfaction. The paper
aims to discuss these issues.
Design/methodology/approach Drawing on theoretical and empirical insights on organic food
consumption in different markets, this research investigates antecedents of organic food products and
differences regarding the relative importance of the value-based drivers across two Western nations.
Findings The results from survey data indicate significant differences in the value perceptions,
especially the functional and individual value perceptions, and recommendation behavior related to
organic food for consumers from the USA and Germany. In addition, the segmentation approach
provides evidence for consumer segments that cross-national borders: the convinced opponents, the
silent/private consumers, the prestige-seekers and the passionate evangelists.
Originality/value Consequently, instead of a country-based segmentation approach, marketers
should emphasize the different types of consumers across national borders in order to address the
differences in customer value perception in the organic food market.
Keywords Food marketing, Customer perceived value, Cross-national market segmentation,
Organic food consumption
Paper type Research paper
Introduction
The organic food market has become a rapidly growing sector in developed
agricultural economies (Lockie et al., 2002; Chen, 2007). Particularly in Europe, America,
Latin America and Australia, more and more green consumers integrate
environmental considerations into daily purchases, asking for healthier, safer and
sh Food Journal higher quality foods (Lockie et al., 2002; Chen, 2007; Siderera et al., 2005). This
118 No. 2, 2016
96-411
phenomenon is largely due to consumer concerns about food-related diseases and an
merald Group Publishing Limited
070X
increase in the use of genetically modified food (Chen, 2009; Hamzaoui Essoussi and
10.1108/BFJ-07-2015-0235 Zahaf, 2008; Vindigni et al., 2002). Being suspicious of conventionally produced
products, the new consumer orientation leads to a high demand for organic products so Customer
that even discounters offer organic products in large quantities at exceptionally low value
prices (Stolz et al., 2010). Consumers often associate organic food with an environment
friendly agricultural production as well as certain intrinsic quality and safety
perception of
characteristics (Brennan and Kuri, 2002; Vindigni et al., 2002). Reasoning that the term organic food
organic refers to the production process and not only to the product itself (Zakowska-
Biemans, 2011), organic food is a produce of organic farming, which is a type of
farming that sets very strict limits on the amount of artificial synthetic inputs allowed
[]. These inputs include the ones used in production (fertilisers, pesticides, herbicides,
antibiotics, etc.) and processing (food additives, including artificial flavourings,
preservatives and colourings) (First and Brozina, 2009, p. 186). However, from a
rational point of view, organic food may not necessarily be healthier
or tastier than its conventionally produced counterpart (Torjusen et al. 1999 quoted
by Chen, 2009).
During the last decades and with regard to the immense future potential of the
market for environmental and ethical goods, the interest of both marketing researchers
and practitioners in ethical consumerism in general and organic food consumption in
particular has increased significantly. With special focus on consumer demand for
organic food, existing studies have investigated the role of risk and value perceptions,
personal, health and environmentally friendly behavior in organic food purchases, as
well as demographic characteristics of organic food consumers across nations
(e.g. Chen, 2009; Lockie et al., 2002; Magnusson et al., 2003; Saba and Messina, 2003).
The consumption of organic food has been shown to provide the consumer with deeper
values and has become one of the fastest growing food trends not only in the EU but
also in the USA (Nie and Zepeda, 2011; Thgersen, 2010). As a result, marketing
managers have realized the importance of customer ethics and values and how meeting
ethical demands is critical if they wish to gain a competitive advantage (Browne et al.,
2000). However, in the context of environmental and ethical purchases, empirical
studies often reveal a substantial gap between positive consumer attitudes, buying
intentions and effective buying behavior (e.g. Cowe and Williams, 2000; De Pelsmacker
et al., 2005; King and Bruner, 2000; Strong, 1996). Therefore, the question remains of
how to meet the consumers expectations toward organic food products and support
the transfer of positive perception to actual buying behavior: what motivates
consumers to buy organic food products? What are the perceived value dimensions
related to organic food consumption? Does the organic character of a product increase
the willingness to pay a price premium or help to build product loyalty?
With regard to consumption values that directly explain why consumers choose to
buy or avoid particular products (Sheth et al., 1991), the aim of the present study is to
examine the value of organic food consumption as perceived by consumers in
important markets such as the USA and Germany. Specifically, a deeper understanding
of the value dimensions consumers perceive in the context of organic food products is
provided as a general basis to develop and implement successful management
strategies which might transfer positive consumer perceptions to satisfaction and
actual buying behavior.
Incorporating relevant theoretical and empirical findings, this study contributes to
the field of inquiry on organic food consumption by:
Related outcomes
Apart from the empirical investigation of the dimensions of customer value perception
in the context of organic food, the analysis of behavioral outcomes are of particular
importance in order to determine if and to what extent positive value perceptions are
translated into actual organic food consumption. Typically, when dealing with ethical
purchases such as fair trade or organic food, results reveal a substantial gap between
consumer attitudes, buying intentions and effective behavior (e.g. De Pelsmacker et al.,
2005; Greenwald and Banaji, 1995; King and Bruner, 2000; Maison, 2002; Shaw
and Clarke, 1999; Strong, 1996; Ulrich and Sarasin, 1995). Even if studies suggest
that consumers prefer to buy ethical products (e.g. Cryer and Ross, 1997;
FJ Fernandez-Kranz and Merino-Castello, 2005; Mohr et al., 2001), only a niche market
18,2 actually buys them (Cowe and Williams, 2000). To assess if and to what extent positive
attitudes toward organic food products result in actual behavior, for the purposes of
this study and in accordance to existing studies on behavioral loyalty dimensions
(e.g. Zeithaml et al., 1996), three basic components of actual consumer behavior are of
particular interest: consumers purchase intention, recommendation behavior and
willingness to pay a higher price.
Indicating positive customer experiences and signaling that customers perceive
products as favorable (Gupta and Zeithaml, 2006), purchase intentions as an antecedent
of behavioral loyalty creates sales and profits (Chaudhuri and Holbrook, 2001) enhances
cash flows, and secures long-term survival ( Jacoby and Chestnut, 1978). Against the
backdrop of our research context, we suggest that the individuals perceived value is an
important driver for the consumers intention to (re)buy organic food products. Likewise,
loyal customers are willing to engage in positive word-of-mouth communication and
recommend a given brand or product to relevant others (Reichheld and Sasser, 1990;
Zeithaml et al., 1996). Furthermore, the perceived value of a brand or product has been
shown to positively affect a consumers willingness to pay premium prices (Keller, 1993).
Therefore, we suggest that in the context of organic food consumers are willing to
pay a considerably higher price and continue to purchase organic food even with a
price increase.
Country comparison
As it is widely discussed that the market for organic products continues to grow
globally (Chen, 2007; Nie and Zepeda, 2011; Thgersen, 2010), it becomes not only more
important to understand why consumers buy organic foods at an individual level, but
also why consumers across nations buy organic food products for different reasons
(Bartels and Reinders, 2010; First and Brozina, 2009). Drawing from prior findings in
cultural studies and based on our conceptual framework, one main research question
will be analyzed by using a cross-national data set that allows comparisons between US
and German consumers perceptions:
RQ1. Do similar perceptions of organic food exist across countries, or are there
significant differences in perceptions of organic food products across
countries?
Based on this, with reference to Clarks (1990) general hypothesis that is often used in
cross-national studies, we formulate:
H0. Consumers in different countries exhibit the same perceptions of organic food.
H1. Consumers in different countries exhibit differing perceptions of organic food.
Research methodology
To measure the underlying value dimensions of consumers intentions to buy organic
food (Davies et al., 1995) in the USA and Germany, this study employed existing
multi-item measures to draw on new conclusions from an online survey. Related to
organic food, respondents were pleased to evaluate their financial, functional, social
and individual values adapted from Sweeney and Soutar (2001), their willingness to
pay a premium price (Lockie et al., 2002, 2004), their recommendation intention
(Wangenheim and Baon, 2004; Yu and Dean, 2001) and the purchase intention
(Michaelidou and Hassan, 2008; Tarkiainen and Sundqvist, 2005). Besides, some Customer
general statements related to the organic food market have been included with value
reference to health orientation (Tarkiainen and Sundqvist, 2009; Lockie et al., 2002) and
general attitudes to organic products (Chen, 2007).
perception of
A German version of the latent variables was created using back translation organic food
according to equivalent meaning. All items were rated on a five-point Likert scale
(1 strongly disagree; 5 strongly agree) and we operationalized all latent variables
via reflective measurement models. Data were collected at a US and two German
universities. The study sample included 206 US and 240 German respondents. Sample
characteristics are shown in Table I.
Even if the sample used is not representative for the US and Germany, it offers a
balanced set of data for the investigation of customer value perception and related
behavior in the organic food market.
Germany USA
Variable n % n %
Value perception
Financial value Cronbachs 0.825
OF is reasonably priced 0.795
OF offers good value for money 0.874
OF is a good product for the price 0.849
OF is economically 0.726
Functional value Cronbachs 0.822
OF has consistent quality 0.749
OF is well made 0.853
OF has an acceptable standard of quality 0.813
OF is of consistent quality 0.822
Individual value Cronbachs 0.890
OF is something that I would enjoy 0.864
OF is enticing to me 0.819
OF is something that I would feel comfortable using 0.768
OF makes me feel good 0.862
OF gives me pleasure 0.876
Social value Cronbachs 0.897
OF helps me to feel accepted by others 0.865
OF improves the way I am perceived 0.914
OF makes a good impression on other people 0.857
OF gives me social approval 0.863
Related outcomes
Purchase intention Cronbachs 0.934
I intend to buy organic food in the near future 0.923
I am going to buy organic food in the future again 0.925
I will continue to buy organic food 0.923
The next time I purchase healthy food I intend for it to be organic food 0.883
Willingness to pay Cronbachs 0.756
I am willing to pay a higher price for organic food than for
non-organic food 0.897
I am willing to pay a lot more for organic food than for non-organic food 0.897
Recommendation behavior Cronbachs 0.912
ble II. I talk positively about organic food 0.896
asurement I recommend organic food to friends 0.946
constructs I encourage friends to buy organic food 0.925
(grouping) variable and the factors that were identified against the backdrop of our Customer
conceptual model were the dependent variables (factor mean scores). As presented in value
Table III our research hypothesis H1 can be confirmed because the participants
differed significantly ( p o 0.01) in the degree to which they agreed with the statements
perception of
related to the value-based dimensions as well as behavioral outcomes. organic food
With reference to the customer value perception of organic food, the results show that
American consumers rate the functional and the individual value components related to
organic food products significantly higher than the respondents from Germany.
More than Germans, consumers in the USA state that they buy organic food because it is
healthier and has better quality than conventional food products. Additionally, American
consumers associate organic food consumption with pleasure, relaxation and fun.
In contrast to this, even if not significant, German consumers place emphasis on the
financial value of organic food. In their opinion, organic food products offer excellent
value for money. In terms of behavioral outcomes, the results reveal significant
differences in the perception of the willingness to pay and recommendation behavior.
While German consumers are more willing to pay a premium price for organic food than
Americans are, US consumers are more active in positive word-of-mouth activities and
recommend organic food products to relevant others.
In sum, with regard to our initial research question and hypotheses, supportive of
H1, the ANOVA results provide evidence to confirm significant cross-national
differences between the antecedents and behavioral outcomes of organic food value as
perceived by consumers from Germany and the USA Hence, the perception and
consumption of organic food products are multifaceted and comprise a combination of
aspects whose importance differs across countries. Nevertheless, what is the optimal
basis on which appropriate marketing strategies should be developed in the organic
food market? Would a country-based segmentation approach cover the differences in
customer value perception or would an approach aimed at different types of consumers
across national borders be preferable? Hence, in the next step, a cluster analysis was
chosen to identify different types of organic food consumers based on the assumption
that common structures in consumer perception and behavior exist cross-nationally.
Value perception
Financial value 0.117 0.100 5.257 0.022
Functional value 0.262 0.223 27.451 0.000
Individual value 0.331 0.283 45.933 0.000
Social value 0.043 0.036 0.689 0.407
Related outcomes
Purchase intention 0.006 0.005 0.014 0.907
Willingness to pay 0.100 0.085 3.816 0.051 Table III.
Recommendation behavior 0.375 0.320 60.630 0.000 ANOVA results
FJ hierarchical clustering procedure was employed to obtain the number of possible
18,2 clusters and seed points for a k-means cluster analysis. The respondents were
partitioned using a hierarchical procedure and the variance ration criterion was used to
identify the correct number of clusters. Wards method of minimum variance was
chosen to determine the cluster differences in each stage of combinations and to
maximize homogeneity and heterogeneity within and between clusters. The results
strongly suggested the presence of four clusters which were validated using non-
hierarchical k-means clustering. After the clusters were identified, a discriminant
analysis was used to verify the cluster groupings. Overall, 94.1 percent of the cases
were assigned to the correct groups; thus, the results of the cluster analysis were valid
for the useful classification of consumer subgroups based on their organic food value
perceptions that differed significantly from one another. Thus, as shown in Table IV,
this four-cluster solution produced the most interpretable and stable results.
For the development of a profile for each market segment, more detailed information
was obtained by examining the questionnaire variables that were cross-tabulated by
cluster segment. Comparisons among the five clusters were conducted with respect to a
variety of descriptive variables, including demographic and socioeconomic
characteristics. Based on the variables from which they were derived, the four
clusters can be described as follows:
Cluster 1: the convinced opponents. This cluster constitutes 24.3 percent of the
sample, has a mean age of 24.3 years, and consists of 65.4 percent male and 34.6 percent
female respondents. With regard to nationalities, this group primarily consists of
German consumers (66.4 percent). Compared to the rest of the clusters, the members of
this group showed the lowest ratings for all factors included in our analysis. The
convinced opponents are significantly more likely than the members of other clusters to
state that organic food products are too expensive and of worse quality than
conventional food products. Furthermore, they are convinced that organic fraud is
common. These consumers are not willing to pay a premium price for organic food and
do not consider purchases of organic food products as an alternative. They have no
intention to buy organic food and are absolutely not willing to recommend organic food
products to relevant others. In sum, consumers in this group can be regarded as
unquestionable antagonists of the organic food movement.
Cluster 2: the silent/private consumers. The second cluster forms 26.6 percent of the
sample, has a mean age of 24.1 years, and is comprised of 46.2 percent male and
53.9 percent female respondents. With regard to country composition, this cluster is
composed of 57.3 percent German and 42.7 percent US consumers. The consumers in
Value perception
Financial value 0.876 0.178 0.077 0.983 79.409 0.000
Functional value 0.842 0.206 0.132 0.911 93.044 0.000
Individual value 1.081 0.203 0.048 1.079 179.466 0.000
Social value 0.566 0.843 0.624 0.870 57.466 0.000
Related outcomes
ble IV. Purchase intention 1.199 0.460 0.095 0.993 214.768 0.000
means Willingness to pay 0.988 0.238 0.053 0.984 117.102 0.000
ster results Recommendation behavior 1.103 0.194 0.129 1.261 285.473 0.000
this cluster are more likely than members of Cluster 1 to perceive organic food as a Customer
valuable alternative to conventional food products. In their opinion, health-oriented value
aspects are particularly important; they perceive organic food to be healthier and less
contaminated with chemical additives. They state that they are satisfied consumers of
perception of
organic food and intend to buy organic food products again in the future even if they organic food
have to pay a premium price. Nevertheless, consumers in this group are not likely to
recommend organic food products to others; statements related to the social value
dimension reveal the lowest ratings of all groups. In sum, this group consists of people
who consume organic food for personal reasons and have absolutely no interest in
aspects related to buying to impress others.
Cluster 3: the prestige-seekers. Constituting 28.6 percent of the sample, this cluster
consists of 55.6 percent male and 44.4 percent female respondents with a mean age of
23.7 years. With regard to nationality, this group comprises 47.6 percent US and
52.4 percent German consumers. The consumers in this cluster state that they perceive
organic food to be expensive, but in fashion. Therefore, even if they show only medium
ratings for most of the value components as well as the behavioral outcomes, they
reveal particular interest in the social value dimension. The main reason for prestige-
seekers to buy organic food products is to make a good impression on relevant others.
Consequently, in contrast to Cluster 2, this group is very interested in conspicuous
consumption and wants to be regarded as responsible, progressive consumers.
Cluster 4: the passionate evangelists. The cluster represents 20.5 percent of the sample,
has a mean age of 23.3 years, and comprises 47.8 percent male and 52.2 percent female
respondents. Concerning the nationalities of this group, this group primarily consists of US
consumers (60 percent). Taken as a whole, the members of this cluster place significantly
more emphasis than other consumers on all aspects included in our analysis and exhibit
the highest ratings for all factors. They value the healthiness and quality of organic food
products and are absolutely willing to pay a premium price. Additionally, they state that
organic food consumption makes them happy and they consider themselves as satisfied
and loyal consumers. Moreover, they can be regarded as loyal advocates as evidenced by
the highest ratings for the willingness to recommend organic food to relevant others. All in
all, the passionate evangelists represent an enthusiastic group of loyal consumers who are
absolutely convinced that organic food is the best alternative in the food industry.
In sum, the cluster analysis results provide evidence that even with the ANOVA
showing significant cross-country differences, there are similarities in consumer
perceptions that cross-national borders.
Conclusion
Main contribution
In times when consumers are increasingly concerned about certain risks associated
with conventional or genetically modified food products, environmental and ethical
aspects provide consumers with a feeling of security and well-being. Widely perceived
to be reliable, fair, sustainable and healthy, organic food consumption reflects the
association of responsibility and environmental friendliness to the individual decision-
making process and related buying behavior. As a consequence of the increasing
importance of organic products, a deeper understanding of the key drivers of
consumers organic food perception and consumption behavior is a major challenge for
practitioners and researchers in the domain of food marketing. Following a holistic
perspective, the aim of the present study was to examine the various dimensions of
FJ customer perceived value that lead the customer to choose and consume organic food
18,2 products and to identify differences and similarities regarding the relative importance
of the value-based drivers across US and German consumers. Our empirical results are
supportive of the hypothesized significant differences in the value perceptions of
consumers from the USA and Germany, particularly in terms of the functional and
individual value perception as well as the recommendation behavior related to organic
food. Nevertheless, the cluster analysis provides evidence for consumer segments that
cross-national borders: the convinced opponents, the silent/private consumers, the
prestige-seekers and the passionate evangelists. Consequently, instead of a country-
based segmentation approach, the segmentation based on different types of consumers
across national borders would be preferable to address the differences in customer
value perception in the organic food market.
Managerial implications
From a managerial perspective, our study may form the basis of a deeper
understanding of perceived value and related consumer behavior in the context of the
rapidly growing and highly competitive market of organic food. Managers have to be
aware of what influences consumers buying behavior apart from safety issues and
moral judgment. Our results show that organic food producers should include the value
dimensions in marketing and communication strategies to increase sales and overall
profits. Nevertheless, our results show differences in the importance of these
dimensions in a cross-national context and with regard to different consumer segments.
Thus, in view of the US market, organic food producers should focus on the quality
and health-related aspects. We recommend that organic food marketers integrate
information on the principles and the benefits of organic food in their communication
strategies. Regarding signals of high quality of organic food, communication about
trustworthy labels, food certification standards or details about production conditions, the
origin of organic food products or additional information on animal welfare represent
initial communication strategies. When addressing German consumers, the
communication should focus on the financial aspects, as being worth it. However, the
promoted price for high-quality organic food has to be carefully calculated regarding
consumers perceptions that organic food is worth the price in comparison to
conventionally produced food, but it is not perceived as a luxury good (Yiridoe et al., 2005).
With regard to the rising demand for organic food and referring to our results, we suggest
marketers increase the availability of organic foods by setting up new stores or by
increasing the availability of organic food products in supermarkets and/or discounters.