Use of metagenomics to understand the genetic basis

of malnutrition
Tahmeed Ahmed, Rashidul Haque, Abul Mansur Shamsir Ahmed, William A Petri Jr, and
Alejandro Cravioto

Childhood malnutrition is not just due to lack of nutrients, it can also be caused by
enteric infections leading to intestinal inflammation and malabsorption of
nutrients. Human genetic polymorphisms can alter host genes that affect nutrient
absorption and metabolism. Changes in intestinal microbial ecology and the
microbiome (the collective genome of the intestinal microbiota) can also affect the
harvest of nutrients from the diet. A substantial proportion of malnourished children
fail to recover due to inappropriate treatment. However, there may be other causes
for treatment failure, including changes in the microbiome and infection with an
enteropathogen, and a genetic predisposition to malnutrition may exist. It is,
therefore, logical to undertake the following: 1) investigate genetic predisposition to
malnutrition, 2) determine the genetic markers and biomarkers that can help
identify children at risk of malnutrition, and 3) look for new treatment modalities
that can improve the clinical management of children with malnutrition.
© 2009 International Life Sciences Institute

INTRODUCTION human genome, enteropathogenic infections, and intesti-

Poor nutrition is linked to more than one-third of all
child deaths worldwide. Nearly one-third of children in
the developing world are malnourished. Malnutrition in
the first 2 years of life leads to irreversible damage to
cognitive functions and physical capacity, and it is trans-
mitted across generations as malnourished mothers give
birth to low-birthweight children. Management of mal-
nutrition today is a failure; according to the World Bank,
only in East Asia and Latin America has there been a
reduction in malnutrition in the last decade. Malnutrition
today is not primarily due to a calorie deficit. This can be
seen in any urban slum in the world, where well-
nourished and malnourished children with equal access
to calories are found side-by-side. Malnutrition is the
result of an inadequate response of the host and the host’s
gut microbiome to caloric deficit. Thus, for the treatment
of malnutrition to be effective and the response to feeding
optimal, the contributions of the gut microbiome, the
nal inflammation need to be considered.
MAGNITUDE OF CHILDHOOD MALNUTRITION
The magnitude of childhood malnutrition is huge yet
hardly recognized in the planning of public health pro-
grams. Severe acute malnutrition (SAM), characterized
either by severe wasting (weight-for-height less than -3
standard deviations) or the presence of bipedal edema or
a mid-upper-arm circumference of less than 11 cm, may
be quite obvious and easy to identify. In developing coun-
tries, about 3.5% of 556 million children under the age of
5 years suffer from SAM.1 Mild and moderate types of
childhood malnutrition are even more prevalent, but
their significance in childhood morbidity and mortality is
less well recognized. Currently, 20% of children in the
developing world who are under the age of 5 years are
underweight, with a weight-for-age of less than -2 SD. A

Affiliations: T Ahmed, R Haque, Shamsir A, and A Cravioto are with the International Centre for Diarrhoeal Disease Research, Bangladesh
(ICDDR,B), Dhaka, Bangladesh. WA Petri Jr is with the with the Division of Infectious Diseases and International Health, Department of
Medicine, University of Virginia Health System, Charlottesville, Virginia, USA.
Correspondence: T Ahmed, Nutrition Program, ICDDR, B, 68 Shaheed Tajuddin Ahmed Sarani, Mohakhali, Dhaka, Bangladesh. E-mail:
tahmeed@icddrb.org, Phone: +88-02-9899206, Fax: +88-02-8823116.
Key words: genetic markers, malnutrition, metagenomics

doi:10.1111/j.1753-4887.2009.00241.x
Nutrition Reviews® Vol. 67(Suppl. 2):S201–S206 S201
staggering number of 178 million children (32% of child-
ren in the developing world) suffer from stunting, a form
of chronic malnutrition characterized by a height-for-age
of less than -2 SD. Although the prevalence of childhood
malnutrition is decreasing in Asia, countries in South
Asia still have both the highest rates of malnutrition and
the largest numbers of malnourished children. Indeed,
the prevalence of malnutrition in India, Bangladesh,
Afghanistan, and Pakistan (range, 38–51%) is much
higher than in sub-Saharan Africa (26%).2
The consequences of malnutrition are manifold –
increased susceptibility and incidence of infections,
impaired mental development, increased case fatality, and
a huge loss in national productivity. Mild and moderate
forms of malnutrition (underweight and stunting) pri-
marily account for the burden of malnutrition worldwide
and are directly or indirectly responsible for more than
55% of all deaths among children younger than 5 years.3
Maternal and child undernutrition is the underlying
cause of 3.5 million deaths, 3.5% of the disease burden in
children younger than 5 years and 11% of total global
disability-adjusted life years (DALYs).1 Intrauterine
growth restriction, stunting, severe wasting, and micro-
nutrient deficiencies are among the most important
nutritional problems. There is, however, a window of
opportunity for intervention: from pregnancy to 2 years
of age. After the age of 2 years, malnutrition will have
caused irreversible damage particularly to mental
development.
The millennium development goal (MDG) 1 is to
halve (from the 1990 level) the proportion of people who
suffer from hunger by the year 2015. One indicator for
monitoring progress towards this target is the proportion
of children who are underweight.4 Unfortunately, with
only 7 years remaining to achieve the MDGs, only a few
countries show indications of reaching the target for
MDG 1. This means there cannot be any more compla-
cency about childhood malnutrition. An urgent need
exists to reduce the prevalence of protein-energy malnu-
trition, to combat the rampant micronutrient malnutri-
tion, and to reduce the case fatality rate for children
affected by severe malnutrition. How these objectives can
be met is the question faced by countries in Asia and
Africa, which bear most of the burden of childhood mal-
nutrition and are also heavily constrained.
MALNUTRITION AND ENTERIC DISEASES
In 1968, Nevin Scrimshaw et al.5 illustrated the relation-
ship between malnutrition and diarrhea.Analysis of long-
itudinal data from five countries collected over a period
of 20 years reveals that the adjusted odds of stunting
increase by 1.13 for every five episodes of diarrhea (95%
CI 1.07–1.19), and by 1.16 for every 5% unit increase in
S202
longitudinal prevalence (95% CI 1.07–1.25).6 The pro-
portion of stunting attributed to five or more diarrheal
episodes before 24 months of age is 25% (95% CI 8–38%),
suggesting that the risk of stunting increases along with
the cumulative burden of diarrheal disease. But emerging
evidence links stunting not just with diarrhea but with
other enteric infections as well, either overt or subclinical.
Guerrant et al.7 postulate that a substantial proportion of
childhood malnutrition is due to impaired intestinal
absorptive function resulting from multiple and repeated
enteric infections. This reduced intestinal absorption has
potential negative implications, particularly now, a time
when global food prices have spiked and the resulting
food insecurity has further marginalized food and nutri-
ent intake in many developing countries. Enteric infec-
tions (some of which do not cause diarrhea) that are
known to cause stunting include infections with
Cryptosporidium parvum, enteroaggregative Escherichia
coli (EAEC), enterotoxigenic E. coli, and Entamoeba
histolytica.8–11 Infection and malnutrition produce a
vicious cycle from which recovery to a normal nutritional
status and mental development, while possible, is difficult
to achieve; within this cycle, there can be a further
worsening of the nutritional status or the child may die.
Malnutrition alone can cause blunting of the villus archi-
tecture and a reduction of brush borders, which ulti-
mately cause malabsorption of nutrients and a further
decline of nutritional status if not treated appropriately.
Thus, a package of interventions is required that aims to
improve the nutritional status of children younger than 5
years of age through breastfeeding, improved comple-
mentary feeding, and micronutrient supplementation;
further interventions are required to control the huge
burden of enteric infections, including appropriate
antimicrobial treatment, deworming, and good hygiene
practices.12
UNKNOWN FACTORS IN THE CAUSAL FRAMEWORK
FOR MALNUTRITION
The package of efficacious interventions mentioned
above, if implemented at scale, can reduce stunting at 36
months of age by 36%, mortality between birth and 36
months by about 25%, and disability-adjusted life-years
associated with stunting, severe wasting, intrauterine
growth restriction, and micronutrient deficiencies by
about 25%.12 Clearly other factors, hitherto unknown, are
involved in the causal framework for malnutrition. These
factors need to be identified so that new or novel solu-
tions can be designed to reduce the burden of malnutri-
tion. In the conventional and widely accepted conceptual
framework, the basic causes of malnutrition are the
social, economic, and political contexts that lead to a lack
of financial, human, social, and natural capital.1 The
Nutrition Reviews® Vol. 67(Suppl. 2):S201–S206
 Table 1 Household (HH) food security and child nutrition status in select
countries.
Country Number Food-insecure HH Food-secure HH
weight-for-age weight-for-age
SD score SD score†
- -
Bangladesh 1575 2.46 2.38
- -
Ethiopia 1247 2.54* 2.09
- -
Ghana 264 1.1 1.28
- -
Guatemala 207 2.16 1.75
- -
Kenya 1138 1.69 1.81
- -
Pakistan 2005 1.39* 1.67
- -
Philippines 2889 1.63* 1.51
- -
Zambia 784 0.63 0.71
* P < 0.01 between food-insecure and -secure households.
†
A household is defined as food secure if calorie intake per adult equivalent is above
2200, except for in Guatemala, where the top three quartiles of food expenditure per
capita is used as a cutoff.13

underlying causes include income poverty, which results
in household food insecurity, inadequate care, and an
unhealthy household environment. The immediate
causes include suboptimal nutrient intake and disease,
which ultimately lead to malnutrition. Although house-
hold food insecurity has often been cited as an important
determinant of childhood malnutrition, the evidence is
equivocal. Among eight countries, weight-for-age SD
scores were significantly worse among children of food-
insecure households in two countries – Ethiopia and the
Philippines (Table 1).13 However, WFA SD scores were
significantly better among children of food-insecure
households in Pakistan, while in other countries there
was no significant difference between food-secure and
-insecure households. This multicountry analysis thus
points to a lack of association between food insecurity
and child malnutrition. The nutritional status of an indi-
vidual is influenced not only by food but also by nonfood
factors, such as clean water, sanitation, and health care.
The effect of all of these factors must be considered in
efforts to rid communities of malnutrition. Food security
will result in good nutrition only if non-food factors are
dealt with effectively. Malnutrition among preschool chil-
dren is determined by a complex interaction between
illness and lack of food.
Further from the conventional framework is the
hypothesis that malnutrition is caused not solely by nutri-
ent deprivation, but also by the following factors
(Figure 1): 1) enteric infections that cause intestinal
inflammation and malabsorption of nutrients; 2) human
genetic polymorphisms that alter host genes, which ulti-
mately affect nutrient absorption and metabolism or
because of effects on intestinal microbial ecology; and 3)
changes in intestinal microbial ecology and the intestinal
microbiome, which affect the efficient harvesting of nutri-
ents from the diet.
CHILD MALNUTRITION IN BANGLADESH
Over the last two decades in Bangladesh, there has been
an almost 20 percentage-point reduction in the preva-
lence of stunting and underweight among children
younger than 5 years of age. Despite this good news, the
prevalence of underweight in Bangladesh is almost
double that in sub-Saharan Africa. The prevalence of
stunting (height-for-age < -2 SD) among 6–23-month-
old children, as observed in the baseline survey of the
National Nutrition Program (NNP) in 2004–2005, was
41.7%.14 Almost 50% of children in this age group were
underweight (weight-for-age < -2 SD), while 16.4% of
children were wasted (weight-for-height < -2 SD). The
Bangladesh Integrated Nutrition Project (BINP) was
launched to reduce malnutrition among Bangladeshi
women and children through community-based nutri-
tion interventions and comprehensive national and inter-
sectoral initiatives for improved nutritional status at the
population level. The community-based nutrition com-
ponent focused on growth monitoring and promotion,
Figure 1 Interaction of human genetic polymorphisms,
enteric infections, and gut microbiome in malnutrition.

Nutrition Reviews® Vol. 67(Suppl. 2):S201–S206 S203

breastfeeding, supplementary feeding, behavior change
communication, and community mobilization. The BINP
has now been scaled up into the NNP.
Some of the successes documented in the nutrition
sector in Bangladesh (mentioned above) can be attributed
to the NNP.15 However, not all children who participate in
the program exit successfully. A study on severely under-
weight children in rural Bangladesh revealed that 10% of
children (n = 2064) could not graduate from the program
after 3 months of food supplementation, i.e., they failed to
achieve a 500-g weight gain. Those who did not graduate
were more often female and came from poorer, less edu-
cated, and larger families compared to those children who
met the graduation criteria. Those who failed to graduate
were also more often reported to have been sick during
the supplementation period.16 One percent of all children
relapsed back into severe malnutrition or growth falter-
ing and thus became eligible for re-enrolment into the
supplementary feeding program after graduation. Among
children suffering from severe acute malnutrition, the
case-fatality rate in hospital settings has remained virtu-
ally unchanged at around 20%.17 Although a substantial
proportion of this case fatality rate can be ascribed to
faulty case management, it is not impertinent to think
there may be underlying causes for the treatment failure.
These causes would include changes in the microbiome
and intercedent infection with an enteropathogen, as well
as genetic predisposition to malnutrition. The following
steps are therefore considered logical: 1) investigate
genetic predisposition to undernutrition; 2) determine
the genetic markers and other biomarkers that can help
identify children at risk of malnutrition; and 3) look for
new treatment modalities that can improve the clinical
management of children with malnutrition.
METAGENOMICS AND CHILDHOOD MALNUTRITION
Metagenomics (also referred to as environmental and
community genomics) is the genomic analysis of micro-
organisms by direct extraction and cloning of DNA from
an assemblage of microorganisms. The development of
metagenomics resulted from the evidence that as-yet-
uncultured microbes represent the vast majority of
organisms in most environments on earth.18 Microbes are
a part of human life, living on all the surfaces and cavities
of the human body. There are at least 10 microbial cells
for every human cell. Moreover, the number of genes that
encode these microbes is thought to be some 100 times
greater than that found in the human genome. These
complex, dynamic microbial communities have consider-
able impact on human physiology, nutrition, immunity,
and development, and changes in them can be a major
factor for many diseases. However, unlike the human
S204
genome, the genomic material of these communities,
referred to as the human metagenome, is largely
unknown.19
Metagenomics is a new and exciting field involving
the analysis of genomes of all the microorganisms in an
ecological site, the majority of which cannot be grown in
usual laboratory conditions. To do this, the microbial
population is first extracted from the site and its DNA
is purified and sequenced using high-throughput tech-
niques. The computer analysis of a sequence enables
researchers to identify major functions, such as virulence
genes or genes that encode enzymes of industrial interest.
Sometimes, it is even possible to reconstruct the complete
genome of particularly abundant microorganisms – this,
in turn, provides a more comprehensive view of their
biological potential. Metagenomics embraces cultivation-
independent, genome-based characterizations of intact
microbial communities so their assembly, composition,
and operation are clearly defined.20 It should provide
answers to a number of key questions such as the follow-
ing: 1) Do the interpersonal differences in human micro-
biota defined by 16S rRNA gene sequencing accurately
portray the degree of diversity in the microbiome? 2)
How are microbiomes shaped by host genotypes, environ-
mental exposures, including exposure to our mothers and
other family members, nutritional status, and lifestyles,
including diet? 3) Do interpersonal differences in intesti-
nal microbiome structure and function affect susceptibil-
ity to malnutrition?
Most of our interactions with microbes are mutually
beneficial, not pathogenic. We are born “germ-free.”
Beginning at birth, microbes are added to the human
body. The process by which human body surfaces are
colonized is complex, dynamic, and driven by mecha-
nisms that are poorly understood. Recent work, however,
has emphasized how the acquisition of our microbiota is
likely to reflect a confluence of “legacy effects” (the
microbes we encounter following birth), and body
“habitat effects” (how human genotypes, immune
systems, diets, and other factors define the environment
of those parts of our body that become home to our
indigenous microbial communities).21 By the time
humans reach adulthood, microbial cells (primarily bac-
teria) outnumber human cells by as much as an order of
magnitude. Most of these bacteria reside in the distal
intestine, where their density approaches one trillion
organisms per milliliter of luminal contents. A few
notable facts have emerged from 16S rRNA-based surveys
of the distal intestinal microbiota of individuals living in
Western societies.22,23 There are marked interpersonal dif-
ferences in community composition at the species level.
Although representatives of 10 of the 70 known bacterial
divisions were identified in this body habitat, more than
90% of all phylogenetic types (phylotypes) belonged to
Nutrition Reviews® Vol. 67(Suppl. 2):S201–S206
members of just two of these divisions – the Bacteroidetes
and the Firmicutes.24
Transplanting the intestinal microbiota from normal
mice into germ-free recipients increases their body fat
without any increase of food composition, raising the
possibility that the composition of the microbial commu-
nity in the intestine affects the amount of energy
extracted from the diet. To investigate the relationship
between intestinal microbial ecology and body fat in
humans, Jeffrey Gordon and his colleagues studied 12
obese people, who were randomly assigned to either a
fat-restricted or to carbohydrate-restricted low-calorie
diet.25 The composition of these 12 individuals’ intestinal
microbiota was monitored over the course of 1 year by
sequencing 16S ribosomal RNA gene from stool samples.
The data set obtained from this study has shown that out
of 18,348 bacterial 16S rRNA sequences, most (70%) of
the 4074 identified species-level phylogenetic types (phy-
lotypes) were unique to each person. Despite this marked
interpersonal difference in species-level diversity,
members of the Bacteroidetes and Firmicutes divisions
dominated in microbiota (92.6% of all 16S rRNA
sequences). Before diet therapy, obese people had fewer
Bacteroidetes and more Firmicutes than did the lean con-
trols. Over time, the relative abundance of Bacteroidetes
increased and the abundance of Firmicutes decreased
irrespective of diet type.25 From this study, it is clear that
a division-wide change in microbial ecology is associated
with obesity. The intestinal habitat itself selects for spe-
cific ratios of divisions: microbiota transplanted to germ-
free recipients of a different species reconfigure to match
the community structure usually present in the recipient.
The coexistence of Bacteroidetes and Firmicutes in the
intestine implies minimized competition for resources
through cooperation or specialization. Why Firmicutes
predominate in the obese intestine is not quite clear. The
factors that drive shifts in representation at such broad
taxonomic levels must operate on highly conserved bac-
terial traits, as they are shared by a great variety of phy-
lotypes within the divisions.
The discovery that the obesity-associated intestinal
microbiota/microbiome has an increased capacity for
energy harvest makes it important to characterize the
microbiota, its microbiome in individuals who are mal-
nourished, and to define its energy and nutrient harvest-
ing capacity. In Bangladesh, in collaboration with the
University of Virginia, Charlottesville, and the University
of Washington at St. Louis, we are planning to undertake
a birth-cohort microbiome study in which the fecal
microbiota of healthy mono- and dizygotic twins during
the first 24 months life will be analyzed. A case-control
study to characterize the fecal microbiota and micro-
biomes of 1–2-year-old twins, discordant or concordant
for malnutrition, prior to, during, and after nutritional
Nutrition Reviews® Vol. 67(Suppl. 2):S201–S206
intervention is also being planned. DNA will be purified
separately using a bead-beating procedure from stool
samples for microbiome analyses. Culture-independent
enumeration studies of the microbiota based on 16S
rRNA gene sequence-based enumerations of fecal
samples (384 amplicons per individual) will be per-
formed. This will also allow us to investigate the associa-
tion of the microbiome with malnutrition.
CONCLUSION
There is an emerging need to gain insight into the influ-
ence of human genetic polymophisms, the gut micro-
biome, and enteric infections on malnutrition, in order to
provide a foundation for new treatment and prevention
programs on a population-wide basis. By defining the risk
factors and biomarkers for malnutrition, optimized strat-
egies for interventions for childhood malnutrition,
testing, evaluation, and surveillance can be developed.
Acknowledgments
Declaration of interest. The authors have no relevant
interests to declare.
REFERENCES
1. Black RE, Allen LH, Bhutta ZA, et al.; Maternal and Child
Undernutrition Study Group. Maternal and child undernutri-
tion: global and regional exposures and health conse-
quences. Lancet. 2008;371:243–246.
2. The World Bank. Repositioning Nutrition as Central to Develop-
ment. A Strategy for Large-Scale Action. Directions in Develop-
ment. Washington, DC: The World Bank; 2006.
3. Pelletier DL, Frongillo EA, Schroeder DG, et al. The effects of
malnutrition on child mortality in developing countries. WHO
Bull. 1995;73:443–448.
4. UN Millennium Project 2005. Halving Hunger: It Can Be Done.
London and Sterling, VA: Task Force on Hunger; 2005.
5. Scrimshaw NS, Taylor CE, Gordon JE. Interactions of Nutrition
and Infection. Geneva: World Health Organization; 1968.
6. Checkley W, Buckley G, Gilman RH, et al.; the Childhood Mal-
nutrition and Infection Network. Multi-country analysis of
the effects of diarrhoea on childhood stunting. Int J Epide-
miol. 2008;37:816–830.
7. Guerrant RL, Oriá RB, Moore SR, Oriá MOB, Lima AAM. Malnu-
trition as an enteric infectious disease with long-term effects
on child development. Nutr Rev. 2008;66:487–505.
8. Checkley W, Epstein LD, Gilman RH, Black RE, Cabrera L,
Sterling CR. Effects of Cryptosporidium parvum infection in
Peruvian children: growth faltering and subsequent catch-up
growth. Am J Epidemiol. 1998;148:497–506.
9. Steiner TS, Lima AA, Nataro JP, Guerrant RL. Enteroaggrega-
tive Escherichia coli produce intestinal inflammation and
growth impairment and cause interleukin-8 release from
intestinal epithelial cells. J Infect Dis. 1998;177:88–96.
10. Qadri F, Saha A, Ahmed T, Al Tarique A, Begum YA,
Svennerholm A-M. Disease burden due to enterotoxigenic
S205
 Escherichia coli in the first 2 years of life in an urban commu-
nity in Bangladesh. Infect Immun. 2007;75:3961–3968.
11. Mondal D, Petri WA Jr, Sack RB, Kirkpatrick BD, Haque R. Enta-
moeba histolytica-associated diarrheal illness is negatively
associated with the growth of preschool children: evidence
from a prospective study. Trans R Soc Trop Med Hyg.
2006;100:1032–1038.
12. Bhutta ZA, Ahmed T, Black RE, et al. What works? interven-
tions for maternal and child undernutrition and survival.
Lancet. 2008;371:417–440.
13. Haddad L, Bhattarai S, Immink M, Kumar S. Managing inter-
actions between household food security and preschooler
health. International Food Policy Research Institute, Washing-
ton, DC, 1996.
14. Ahmed T, Roy SK, Alam N, et al. Baseline survey 2004.
National Nutrition Programme. ISBN: 984-551-273-9; Special
publication No. 124, ICDDR,B, December 2005.
15. Sack DA, Roy SK, Ahmed T, Fuchs GJ. Responses to: “An evalu-
ation of the impact of a US$60 million nutrition programme
in Bangladesh.” Health Policy Plan. 2005;20:406–407.
16. Rahman F, Islam F. Socio-demographic and health character-
istics of Bangladeshi malnourished children requiring food
supplementation and risk factors for relapse back into the
supplementary feeding programme among under-2 years
children. In: Results of the Bangladesh Integrated Nutrition
Project (BINP) Operations Research Project (ORP), Vol 4.
ICDDR.B: Dhaka, 2000. BINP-OR 2000/HPQR-10, 11.
S206
17. Schofield C, Ashworth A. Why have mortality rates for severe
malnutrition remained so high? Bull WHO. 1996;74:223–
229.
18. Gill SR, Pop M, Deboy RT, et al. Metagenomic analysis of the
human distal gut microbiome. Science. 2006;312:1355–1359.
19. Turnbaugh PJ, Ley RE, Hamady M, Fraser-Liggett CM,
Knight R, Gordon JI. The human microbiome project. Nature.
2007;449:804–810.
20. Harris JK, Kelley ST, Pace NR. New perspective on uncultured
bacterial phylogenetic division OP11. Appl Environ Microbiol.
2004;70:845–849.
21. Rawls JF, Mahowald MA, Ley RE, Gordon JI. Reciprocal gut
microbiota transplants from zebra fish and mice to germ-free
recipients reveal host habitat selection. Cell. 2006;127:423–
433.
22. Eckburg PB, Bik EM, Bernstein CN, et al. Diversity of the
human intestinal microbial flora. Science. 2005;308:1635–
1638.
23. Ley R, Peterson DA, Gordon JI. Ecological and evolutionary
forces that shape microbial diversity and genome content in
the human intestine. Cell. 2006;124:837–848.
24. Backhed F, Ding H, Wang T, et al. The gut microbiota as an
environmental factor that regulates fat storage. Proc Nat
Acad Sci. 2004;101:15718–15723.
25. Ley RE, Turnbaugh PJ, Klein S, Gordon JI. Human gut microbes
associated with obesity. Nature. 2006;444:1022–1023.
Nutrition Reviews® Vol. 67(Suppl. 2):S201–S206