Paediatric Nutrition

Edited by

Colin Binns and Mi Kyung Lee

Printed Edition of the Special Issue Published in Nutrients

www.mdpi.com/journal/nutrients
Colin Binns and Mi Kyung Lee (Eds.)

Paediatric Nutrition
This book is a reprint of the special issue that appeared in the online open access journal
Nutrients (ISSN 2072-6643) in 2014 (available at:
http://www.mdpi.com/journal/nutrients/special_issues/paediatric-nutrition).

Guest Editors

Colin Binns
School of Public Health and Curtin Health Innovation Research Institute
Curtin University
Perth, WA 6845, Australia

Mi Kyung Lee
School of Health Professions
Murdoch University
Murdoch, WA 6150, Australia

Editorial Office
MDPI AG
Klybeckstrasse 64
Basel, Switzerland

Publisher
Shu-Kun Lin

Production Editor
Martyn Rittman

1. Edition 2014

0'3,‡%DVHO‡%HLMLQJ

ISBN 978-3-906980-51-5

© 2014 by the authors; licensee MDPI, Basel, Switzerland. All articles in this volume are Open
Access distributed under the Creative Commons Attribution 3.0 license
(http://creativecommons.org/licenses/by/3.0/), which allows users to download, copy and build
upon published articles even for commercial purposes, as long as the author and publisher are
properly credited, which ensures maximum dissemination and a wider impact of our publications.
However, the dissemination and distribution of copies of this book as a whole is restricted to
MDPI, Basel, Switzerland.
 III

Table of Contents

Mi Kyung Lee and Colin Binns

Guest Editors: Editorial .................................................................................................................. VII

1. General Review

Laura M. Lamberti, Christa L. Fischer Walker, Kit Y. Chan, Wei-Yan Jian and
Robert E. Black
Oral Zinc Supplementation for the Treatment of Acute Diarrhea in Children: A Systematic
Review and Meta-Analysis ............................................................................................................... 1
Reprinted from Nutrients 2013, 5(11), 4715-4740
http://www.mdpi.com/2072-6643/5/11/4715

Madoka Inoue and Colin W. Binns

Introducing Solid Foods to Infants in the Asia Pacific Region ....................................................... 29
Reprinted from Nutrients 2014, 6(1), 276-288
http://www.mdpi.com/2072-6643/6/1/276

2. Breastfeeding

Ekhard E. Ziegler, Steven E. Nelson and Janice M. Jeter

Iron Stores of Breastfed Infants during the First Year of Life ........................................................ 43
Reprinted from Nutrients 2014, 6(5), 2023-2034
http://www.mdpi.com/2072-6643/6/5/2023

Patricia Dominguez Castro, Richard Layte and John Kearney

Ethnic Variation in Breastfeeding and Complimentary Feeding in the Republic of Ireland .......... 55
Reprinted from Nutrients 2014, 6(5), 1832-1849
http://www.mdpi.com/2072-6643/6/5/1832

Vishnu Khanal, Jonia Lourenca Nunes Brites da Cruz, Rajendra Karkee and
Andy H. Lee
Factors Associated with Exclusive Breastfeeding in Timor-Leste: Findings from
Demographic and Health Survey 2009–2010 ................................................................................. 73
Reprinted from Nutrients 2014, 6(4), 1691-1700
http://www.mdpi.com/2072-6643/6/4/1691

Manal Dashti, Jane A. Scott, Christine A. Edwards and Mona Al-Sughayer

Predictors of Breastfeeding Duration among Women in Kuwait: Results of a Prospective
Cohort Study ................................................................................................................................... 83
Reprinted from Nutrients 2014, 6(2), 711-728
http://www.mdpi.com/2072-6643/6/2/711
Olof H. Jonsdottir, Inga Thorsdottir, Geir Gunnlaugsson, Mary S. Fewtrell,
Patricia L. Hibberd and Ronald E. Kleinman
Exclusive Breastfeeding and Developmental and Behavioral Status in Early Childhood ............ 101
Reprinted from Nutrients 2013, 5(11), 4414-4428
http://www.mdpi.com/2072-6643/5/11/4414

Jianghong Liu, Patrick Leung and Amy Yang

Breastfeeding and Active Bonding Protects against Children’s
Internalizing Behavior Problems ................................................................................................... 117
Reprinted from Nutrients 2014, 6(1), 76-89
http://www.mdpi.com/2072-6643/6/1/76

3. Infants

Melissa Thoene, Corrine Hanson, Elizabeth Lyden, Laura Dugick,

Leslie Ruybal and Ann Anderson-Berry
Comparison of the Effect of Two Human Milk Fortifiers on Clinical
Outcomes in Premature Infants ..................................................................................................... 131
Reprinted from Nutrients 2014, 6(1), 261-275
http://www.mdpi.com/2072-6643/6/1/261

Birna Thorisdottir, Ingibjorg Gunnarsdottir, Laufey Steingrimsdottir,

Gestur I. Palsson and Inga Thorsdottir
Vitamin D Intake and Status in 12-Month-Old Infants at 63–66° N ............................................ 147
Reprinted from Nutrients 2014, 6(3), 1182-1193
http://www.mdpi.com/2072-6643/6/3/1182

4. Children and Adolescents

Jessica S. Gubbels, Lieke G. M. Raaijmakers, Sanne M. P. L. Gerards and

Stef P. J. Kremers
Dietary Intake by Dutch 1- to 3-Year-Old Children at Childcare and at Home ........................... 159
Reprinted from Nutrients 2014, 6(1), 304-318
http://www.mdpi.com/2072-6643/6/1/304

Shu Chen, Colin W. Binns, Bruce Maycock, Yi Liu and Yuexiao Zhang
Prevalence of Dietary Supplement Use in Healthy Pre-School Chinese Children
in Australia and China ................................................................................................................... 175
Reprinted from Nutrients 2014, 6(2), 815-828
http://www.mdpi.com/2072-6643/6/2/815
 V

Lara Nasreddine, Farah Naja, Christelle Akl, Marie Claire Chamieh,

Sabine Karam, Abla-Mehio Sibai and Nahla Hwalla
Dietary, Lifestyle and Socio-Economic Correlates of Overweight, Obesity and
Central Adiposity in Lebanese Children and Adolescents ............................................................ 189
Reprinted from Nutrients 2014, 6(3), 1038-1062
http://www.mdpi.com/2072-6643/6/3/1038

Caroline M. Gallagher, Lucinda J. Black and Wendy H. Oddy

Micronutrient Intakes from Food and Supplements in Australian Adolescents ........................... 215
Reprinted from Nutrients 2014, 6(1), 342-354
http://www.mdpi.com/2072-6643/6/1/342

Mohsin Yakub, Kerry J. Schulze, Subarna K. Khatry, Christine P. Stewart,

Parul Christian and Keith P. West Jr.
High Plasma Homocysteine Increases Risk of Metabolic Syndrome in 6 to
8 Year Old Children in Rural Nepal .............................................................................................. 229
Reprinted from Nutrients 2014, 6(4), 1649-1661
http://www.mdpi.com/2072-6643/6/4/1649

Hae Dong Woo, Dong Woo Kim, Young-Seoub Hong, Yu-Mi Kim, Ju-Hee Seo,
Byeong Moo Choe, Jae Hong Park, Je-Wook Kang, Jae-Ho Yoo,
Hee Won Chueh, Jung Hyun Lee, Min Jung Kwak and Jeongseon Kim
Dietary Patterns in Children with Attention Deficit/Hyperactivity Disorder (ADHD) ................ 243
Reprinted from Nutrients 2014, 6(4), 1539-1553
http://www.mdpi.com/2072-6643/6/4/1539

Jianghong Liu, Alexandra Hanlon, Chenjuan Ma, Sophie R. Zhao,

Siyuan Cao and Charlene Compher
Low Blood Zinc, Iron, and Other Sociodemographic Factors Associated with
Behavior Problems in Preschoolers ............................................................................................... 259
Reprinted from Nutrients 2014, 6(2), 530-545
http://www.mdpi.com/2072-6643/6/2/530

5. Nutrition Assessment and Body Composition

Navnit Kaur Grewal, Annhild Mosdøl, Marte Bergsund Aunan,

Carina Monsen and Liv Elin Torheim
Development and Pilot Testing of 24-Hour Multiple-Pass Recall to
Assess Dietary Intake of Toddlers of Somali - and Iraqi-Born Mothers Living in Norway ......... 275
Reprinted from Nutrients 2014, 6(6), 2333-2347
http://www.mdpi.com/2072-6643/6/6/2333
Sissel J. Moltu, Daniel Sachse, Elin W. Blakstad, Kenneth Strømmen,
Britt Nakstad, Astrid N. Almaas, Ane C. Westerberg, Arild Rønnestad,
Kristin Brække, Marit B. Veierød, Per O. Iversen, Frode Rise, Jens P. Berg and
Christian A. Drevon
Urinary Metabolite Profiles in Premature Infants Show Early Postnatal Metabolic
Adaptation and Maturation............................................................................................................ 291
Reprinted from Nutrients 2014, 6(5), 1913-1930
http://www.mdpi.com/2072-6643/6/5/1913

Cindy Mari Imai, Ingibjorg Gunnarsdottir, Birna Thorisdottir,

Thorhallur Ingi Halldorsson and Inga Thorsdottir
Associations between Infant Feeding Practice Prior to Six Months and
Body Mass Index at Six Years of Age .......................................................................................... 309
Reprinted from Nutrients 2014, 6(4), 1608-1617
http://www.mdpi.com/2072-6643/6/4/1608

Masaharu Kagawa, Connie Wishart and Andrew P. Hills

Influence of Posture and Frequency Modes in Total Body Water Estimation
Using Bioelectrical Impedance Spectroscopy in Boys and Adult Males ...................................... 319
Reprinted from Nutrients 2014, 6(5), 1886-1898
http://www.mdpi.com/2072-6643/6/5/1913
 VII

Editorial
An agenda for research into nutrients in paediatrics

Food and nutrition has been central to human culture, philosophy and science since the beginning
of civilisation. However the building blocks of food and nutrition, the nutrients, remained
unknown until the late 19th century. Over the next 100 years advances in physics, chemistry and
physiology led to rapid developments in our knowledge, first with development of an
understanding of energy and the macronutrients, followed by the minerals and vitamins. The first
vitamins to be explored scientifically were thiamine, vitamin D and C and in 1935 ascorbic acid
was synthesised, beginning the 20th century rapid development of knowledge of nutrients[1].

Recommendations for intakes of nutrients were first made before chemical characterisation had
begun in earnest and the long road to the establishment of formal nutrient recommendations has
been described by Harper[2,3]. Recommendations on the intake of limes containing vitamin C to
prevent scurvy began in the British Navy at the end of the 18th century. Since then times of crisis,
mainly warfare, have stimulated further research into nutrients and improving the nutritional
status (and hence fighting ability) of the population. Early recommendations for intakes were
focussed on adult males, and children were simply not considered. After World War One and the
onset of the Great Depression, the League of Nations established several commissions to
investigate the provision of adequate nutrition for populations and the establishment of nutrient
intake recommendations. In the 1930s the special needs of children and nursing mothers were
considered for the first time in the recommendations of the British Committee on Nutrition [4].
This was followed by the reports of the League of Nations that included recommendations for
nutrients during periods of growth that were extrapolations of adult requirements of the known
macro and micronutrients[3]. In 1941 the Food and Nutrition Board of the National Academy of
Sciences was established and the first edition of the Recommended Dietary Allowances was
published. When commenting on this, Nutrition Reviews noted that the RDAs “emphasize once
more the truth of the opinion that the dietary requirements can be met by a well-chosen diet of
natural foods[5].” Since that time there have been numerous variations of the RDAs published by
national and international organisations under a variety of titles. The cost of development and the
size of documentation has grown exponentially and current volumes of the US Dietary Reference
Intakes occupy a whole shelf[6].

Setting nutrient requirements for children has required different approaches to setting the values
for adults and it is only in the last 50 years that extra effort has been placed on establishing
children’s nutrient requirements. Nutrients cannot be simply extrapolated on the basis of weight
from adults, but must provide for higher metabolic rates and energy expenditure, and growth and
development. In the history of nutrients protein was regarded as the key to infant and child
health. Low levels of protein were associated with poor growth and malnutrition (kwashiorkor)
in children. Although growth is a major consideration for children, it only requires a small
proportion of the total energy and protein requirements. In 1957 Hegsted concluded that “growth
was a minor determinant of protein or other nutrient needs after the first months of life. The
amount of new tissue protein deposited per day in growing children or during the adolescent
growth spurt is very small compared with the total maintenance requirement of protein. This
distinguishes humans and other primate species from common laboratory and domestic
species[7].” More recently the role of early life protein intake in the laying the basis for later
obesity through influencing insulin-like growth factor 1 (IGF-1) levels has been a focus of
research[8,9]. There are important lessons to be learned from the exploration of protein needs of
infants and children and the requirements follow a U-shaped curve. But it is even more complex
as in real life there increased requirements associated with response to illness, injury and periods
of rapid growth.
The establishment of iron requirements for infants and children and the development of
interventions to overcome deficiencies has proved to be complex. Iron deficiency has been
associated with poor growth, reduced cognitive development and ill health [10]. Yet breastmilk
contains only a low level of iron, albeit in a readily bioavailable form [11]. Lactoferrin is
important for transporting iron within the body, but is protective against infection by making iron
unavailable to micro-organisms that require iron for growth[12]. All infant feeding guidelines
recommend the introduction of complementary foods at around 6 months of age which provide
increased amounts of iron. Recent attempts to increase iron supplies for children in developing
countries by genetically modifying foods have not been entirely successful as they have resulted
in increased rates of infection, including malaria[13-15]. This has led to a re-evaluation of how
some nutrient requirements are set and deficiencies are met.
 IX

The current agenda for paediatric nutrients research will occupy nutritionists, biochemists,
paediatricians and epidemiologists for many years ahead. Some of the immediate needs to be
answered are to define the interactions and outcomes of nutrient levels with future health and
disease beyond childhood, epigenetics and nutrients, interactions of nutrients with the human
microbiome, sustainability and climate change, ethnicity gene interactions with nutrients.
The developmental origins of health and disease (DOHAD) hypothesis has added new emphasis
to early life nutrition. How nutrition and growth influence later chronic disease has been the
subject of many observational studies[16-18]. Future developments will probably rely on animal,
perhaps primate models and the use of laboratory studies, as longer term prospective human
studies are not feasible. The implications for setting of nutrient requirements may be far reaching
– no longer is deficiency or short term growth the criteria, but long term life-course outcomes
must be considered.

In recent years there has been considerable interest in the human microbiome and long term
health outcomes following the initiative of the National Institutes of Health to sponsor research in
this field[19]. It is well known that in the human body microbial cells outnumber human cells at
least ten-fold. But what is not known is how they influence nutrient requirements, particularly in
early life when the stable microbiome is being finalised. Malnutrition has an effect on the
establishment of a stable microbiome[20]. Dysfunction of the microbiome has been linked to
disorders as diverse as obesity, under-nutrition, diabetes and gastro-intestinal cancer[21-23]. It
can be anticipated that nutrient- microbiome interaction will become an important area for
nutrient research.

For infants the gold standard of nutrition and for nutrients is breastmilk. Many of the nutrients
contained in breastmilk are in relatively low concentrations, but in highly bioavailable forms[24].
In the development of infant formula nutrients such as iron have to be included in greater
concentrations than in breastmilk to adjust for the bioavailability[25]. Breastmilk is the most
sustainable of infant foods in an era where resources, climate change and sustainability are
paramount. A recent report of the Institute of Medicine explored current and emerging
knowledge on nutrients in the light of the increasing environmental constraints on the food
system[26]. This will continue to be an important area of research as climate affects different
aspects of nutrition, for example bioavailablity. All of these emerging issues meant that it is
likely that there will be further special issues of “Nutrients” devoted to paediatric concerns.

In this ‘Nutrients’ special collection we present a range of paediatric papers. The first section of
reviews contains papers on two important topics. UNICEF estimates that more than 500000
children die every year of diarrhoeal disease, and most could be prevented or treated with
relatively simple interventions[27]. The recent Global Burden of Disease study also confirms the
continuing burden of mortality and morbidity form diarrhoeal disease in children[28]. The
systematic review by Lamberti and colleagues confirms the value of zinc supplementation in the
management of diarrhoea and endorses the current WHO recommendations.

Six of the papers relate to breastfeeding and there are several more on early infancy reflecting
the importance of breastfeeding in early nutrition and in influencing life course nutrition. The
paper by Imai et al joins a large number of observational studies that show an association
between early nutrition and growth and later obesity. In this study infant feeding method and the
early introduction of solids are associated with a higher BMI. The ethical impossibility of
randomised controlled trials of breastfeeding and obesity means that we have to rely on the
weight of observational studies and recognise that residual confounding may persist. Other
papers document the optimal nutrients provided by breastmilk, supporting the continued
promotion of breastfeeding. There are two papers that document the provision of nutrients from
supplements in children, an area that will need continuing study to ensure that supplement use is
appropriate. The special issue concludes with several papers on methodology in nutrient and
body composition research.
The papers selected for this special collection illustrate the breadth of paediatric research, but
there are still many scientific challenges remaining. There is still great potential for improvement
in the health of children through nutrition and understanding of nutrient requirements,
metabolism and social context is important to realising these potential health gains.

Dr. Mi Kyung Lee BSc MA PhD and Prof Colin Binns MBBS MPH PhD
Guest Editors
 XI

References:
1. Carpenter, K.J. A short history of nutritional science: Part 3 (1912-1944). The Journal of
nutrition 2003, 133, 3023-3032.
2. Harper, A.E. Contributions of women scientists in the u.S. To the development of
recommended dietary allowances. The Journal of nutrition 2003, 133, 3698-3702.
3. Harper, A.E. Origin of recommended dietary allowances--an historic overview. The
American journal of clinical nutrition 1985, 41, 140-148.
4. British Committee On Nutrition. Report of committee on nutrition. British Medical
Journal 1933, 2, 1-16.
5. Feeding the army and the navy. Nutrition reviews 1942, 1, 1-2.
6. Food and Nutrition Board. Dietary reference intakes accessed 1 june 2014 Institute of
Medicine of the National Academies of Science: 2014
7. Hegsted, D.M. From chick nutrition to nutrition policy. Annual review of nutrition 2000,
20, 1-19.
8. Michaelsen, K.F.; Greer, F.R. Protein needs early in life and long-term health. The
American journal of clinical nutrition 2014, 99, 718S-722S.
9. Michaelsen, K.F. Effect of protein intake from 6 to 24 months on insulin-like growth
factor 1 (igf-1) levels, body composition, linear growth velocity, and linear growth
acceleration: What are the implications for stunting and wasting? Food and nutrition
bulletin 2013, 34, 268-271.
10. NHMRC. Nutrient reference values for australia and new zealand including
recommended dietary intakes page 199. NHMRC: Canberra, 2005.
11. National Health and Medical Research Council. Infant feeding guidelines for health
workers. Www.Nhmrc.Gov.Au. NHMRC: Canberra, 2012.
12. Lopez Alvarez, M.J. Proteins in human milk. Breastfeed Rev 2007, 15, 5-16.
13. Sangare, L.; van Eijk, A.M.; Ter Kuile, F.O.; Walson, J.; Stergachis, A. The association
between malaria and iron status or supplementation in pregnancy: A systematic review
and meta-analysis. PloS one 2014, 9, e87743.
14. Brabin, L.; Brabin, B.J.; Gies, S. Influence of iron status on risk of maternal or neonatal
infection and on neonatal mortality with an emphasis on developing countries. Nutrition
reviews 2013, 71, 528-540.
15. Quinn, E.A. Too much of a good thing: Evolutionary perspectives on infant formula
fortification in the united states and its effects on infant health. American journal of
human biology : the official journal of the Human Biology Council 2014, 26, 10-17.
16. Singhal, A. The global epidemic of noncommunicable disease: The role of early-life
factors. Nestle Nutrition Institute workshop series 2014, 78, 123-132.
17. Langley-Evans, S.C. Nutrition in early life and the programming of adult disease: A
review. J Hum Nutr Diet 2014.
18. Barker, D.J.; Eriksson, J.G.; Forsen, T.; Osmond, C. Fetal origins of adult disease:
Strength of effects and biological basis. Int J Epidemiol 2002, 31, 1235-1239.
19. Turnbaugh, P.J.; Ley, R.E.; Hamady, M.; Fraser-Liggett, C.M.; Knight, R.; Gordon, J.I.
The human microbiome project. Nature 2007, 449, 804-810.
20. Subramanian, S.; Huq, S.; Yatsunenko, T.; Haque, R.; Mahfuz, M.; Alam, M.A.; Benezra,
A.; DeStefano, J.; Meier, M.F.; Muegge, B.D., et al. Persistent gut microbiota immaturity
in malnourished bangladeshi children. Nature 2014.
21. Gordon, J.I.; Dewey, K.G.; Mills, D.A.; Medzhitov, R.M. The human gut microbiota and
undernutrition. Science translational medicine 2012, 4, 137ps112.
22. Turnbaugh, P.J.; Gordon, J.I. The core gut microbiome, energy balance and obesity. The
Journal of physiology 2009, 587, 4153-4158.
23. Jumpertz, R.; Le, D.S.; Turnbaugh, P.J.; Trinidad, C.; Bogardus, C.; Gordon, J.I.;
Krakoff, J. Energy-balance studies reveal associations between gut microbes, caloric load,
and nutrient absorption in humans. The American journal of clinical nutrition 2011, 94,
58-65.
24. National Health and Medical Research Council. Infant feeding guidelines. . National
Health and Medical Research Council.: Canberra: , 2012.
25. Domellof, M.; Braegger, C.; Campoy, C.; Colomb, V.; Decsi, T.; Fewtrell, M.; Hojsak, I.;
Mihatsch, W.; Molgaard, C.; Shamir, R., et al. Iron requirements of infants and toddlers.
Journal of pediatric gastroenterology and nutrition 2014, 58, 119-129.
26. Institute of Medicine. Sustainable diets: Food for healthy people and a healthy planet:
Workshop summary The National Academies Press.: Washington, DC., 2014.
27. UNICEF. State of the world's children: Every child counts revealing disparities,
advancing children’s rights. UNICEF: New York, 2014.
28. Walker, C.L.; Rudan, I.; Liu, L.; Nair, H.; Theodoratou, E.; Bhutta, Z.A.; O'Brien, K.L.;
Campbell, H.; Black, R.E. Global burden of childhood pneumonia and diarrhoea. Lancet
2013, 381, 1405-1416.
 1

1. General Review

Reprinted from Nutrients. Cite as: Lamberti, L.M.; Walker, C.L.F.; Chan, K.Y.; Jian, W.; Black, R.E.
Oral Zinc Supplementation for the Treatment of Acute Diarrhea in Children: A Systematic Review and
Meta-Analysis. Nutrients 2013, 5, 4715-4740.

Oral Zinc Supplementation for the Treatment of Acute

Diarrhea in Children: A Systematic Review and Meta-Analysis
Laura M. Lamberti 1, Christa L. Fischer Walker 1,*, Kit Y. Chan 2,3, Wei-Yan Jian 2
and Robert E. Black 1
1
Department of International Health, Johns Hopkins Bloomberg School of Public Health,
615 N. Wolfe St, Baltimore, MD 21205, USA; E-Mails: llambert@jhsph.edu (L.M.L.);
rblack@jhsph.edu (R.E.B.)
2
Department of Health Policy and Management, School of Public Health,
Peking University Health Science Centre, 38 Xueyuan Rd. in Haidian District, Beijing 10083,
China; E-Mails: k.chan@ed.ac.uk (K.Y.C.); jianweiyan@bjmu.edu.cn (W.-Y.J.)
3
Centre for Population Health Sciences, University of Edinburgh Medical School, Teviot Place,
Edinburgh, Scotland EH8 9AG, UK

* Author to whom correspondence should be addressed; E-Mail: cfischer@jhsph.edu;

Tel.: +1-410-502-3478; Fax: +1-410-955-7159.

Received: 4 September 2013; in revised form: 9 October 2013 / Accepted: 4 November 2013 /
Published: 21 November 2013

Abstract: Evidence supporting the impact of therapeutic zinc supplementation on the

duration and severity of diarrhea among children under five is largely derived from
studies conducted in South Asia. China experiences a substantial portion of the global
burden of diarrhea, but the impact of zinc treatment among children under five has not
been well documented by previously published systematic reviews on the topic. We
therefore conducted a systematic literature review, which included an exhaustive search
of the Chinese literature, in an effort to update previously published estimates of the
effect of therapeutic zinc. We conducted systematic literature searches in various
databases, including the China National Knowledge Infrastructure (CNKI), and
abstracted relevant data from studies meeting our inclusion and exclusion criteria. We
used STATA 12.0 to pool select outcomes and to generate estimates of percentage
difference and relative risk comparing outcomes between zinc and control groups.
 2

We identified 89 Chinese and 15 non-Chinese studies for the review, including studies
in 10 countries from all WHO geographic regions, and analyzed a total of 18,822 diarrhea
cases (9469 zinc and 9353 control). None of the included Chinese studies had
previously been included in published pooled effect estimates. Chinese and non-Chinese
studies reported the effect of therapeutic zinc supplementation on decreased episode
duration, stool output, stool frequency, hospitalization duration and proportion of
episodes lasting beyond three and seven days. Pooling Chinese and non-Chinese studies
yielded an overall 26% (95% CI: 20%í UHGXFWLRQLQWKHHVWLPDWHGUHODWLYHULVNRI
diarrhea lasting beyond three days among zinc-treated children. Studies conducted in
and outside China report reductions in morbidity as a result of oral therapeutic zinc
supplementation for acute diarrhea among children under five years of age. The WHO
recommendation for zinc treatment of diarrhea episodes should be supported in all
low- and middle-income countries.

Keywords: zinc; children; global health; China

1. Introduction

In response to mounting evidence supporting the efficacy and effectiveness of therapeutic zinc
supplementation for diarrhea among children under five years of age, the World Health
Organization (WHO) and the United Nation’s Children Fund (UNICEF) issued a global
recommendation in 2004, which advised zinc supplementation in addition to oral rehydration
solution (ORS) for the treatment of all diarrhea episodes among children <5 years of age [1,2].
Systematic reviews have quantified the association between therapeutic zinc supplementation and a
reduction in the duration and severity of childhood diarrhea episodes in low- and middle-income
countries (LMICs) [1,3,4]. Many of the studies contributing to this body of evidence were
conducted in South Asia [5–7], but literature stemming from East Asia has not been included in
past reviews. In 2011, Zhang published a systematic review which identified 11 Chinese studies
assessing zinc treatment for diarrhea and signified the need to update previous meta-analyses with
literature published in languages other than English [8].
We sought to conduct an extensive search for studies of oral therapeutic zinc supplementation
published in Chinese and any other language. We also aimed to combine evidence across regions in
order to generate global estimates of the effect of oral therapeutic zinc supplementation on selected
morbidity and mortality outcomes among children under five years of age.

2. Methods

We conducted a systematic literature search for studies published in any language between 1980
and November 2012 using the MeSH search terms “zinc” and “diarrhea” limited to “humans” in
the following databases: Biosis, Cumulative Index to Nursing and Allied Health (CINAHL),
Cochrane Central Register of Controlled Trials (CENTRAL), Embase, the WHO International
 3

Clinical Trials Registry Platform (ICTRP), Global Health, Latin American and Caribbean Health
Sciences Literature (LILACS), PubMed, Scopus, Web of Science, IndMed, Egyptian Universities
Library Consortium, Index Medicus for the Eastern Mediterranean Region (IMEMR), China
National Knowledge Infrastructure (CNKI), WanFang, and Chinese BioMedical (CBM) database.
Titles and abstracts were reviewed by two independent reviewers, and complete manuscripts
were obtained for further review of pertinent studies. Discrepancies were resolved in consultation
with a third reviewer. We restricted inclusion to individually randomized controlled trials (RCTs)
of children under five years of age with acute diarrhea, including dysentery, where diarrhea was
defined as the passage of at least three loose or watery stools in a 24-h period. We excluded cluster
RCTs, studies that exclusively enrolled a particular subgroup of children (e.g., HIV-infected
children; preterm infants), and studies of persistent diarrhea. We included RCTs assessing oral zinc
supplementation of any zinc salt in comparison to a control group receiving placebo supplement.
For studies conducted in China, where placebo supplements may not have been readily available,
we included trials in which cases received the same supportive therapy regardless of zinc
allocation. For all studies, administration of minerals (excluding iron), vitamins, and supporting
therapy beyond zinc were only considered acceptable if these were received by both the
intervention and control groups. Studies that used supplements that included iron, zinc-fortified
ORS, or zinc-fortified foods were excluded.
Included studies were reviewed for the following outcomes: diarrhea duration; the proportion of
diarrhea episodes lasting >3 and >7 days; duration of hospitalization; duration of fever; duration of
vomiting; proportion of cases vomiting; stool frequency (number per day); stool output (mL); and
death from diarrhea or any cause. Two independent reviewers entered data into structured tables,
and discrepancies were resolved in consultation with a third reviewer.
We conducted independent analyses for studies assessing diarrhea due to unspecified causes and
those assessing specific pathogens (e.g., rotavirus) that were laboratory confirmed prior to
enrollment. All data analyses were conducted in STATA 12.0 [9]. We fit Poisson and logistic
regression models to continuous and binary outcomes, respectively, weighting all outcomes by
sample size. These models generated pooled estimates and 95% confidence intervals lower bound
by zero for all outcomes and upper bound by one for proportions.
For continuous outcomes, we calculated the overall percentage difference between the pooled
estimates for the zinc and control groups. For binary outcomes, we calculated estimates of relative
risk (RR) with placebo as the reference group and conducted random effects meta-analyses to
combine RRs across studies [9].
We conducted hypothesis testing to assess the equivalence of pooled outcomes and of effect
estimates by placebo and non-placebo controlled trials. To compare effect estimates, we tested the
difference of mean percentage differences for continuous outcomes and the ratio of relative risks
(RRR) for binary outcomes [10]. We subsequently pooled placebo and non-placebo controlled
trials for outcomes with no statistically significant difference in effect size.
We assessed the association between the dose of oral zinc supplement and diarrhea duration by
regressing the mean percentage difference in diarrhea duration comparing the zinc and control
 4

groups onto a categorical variable which indicated whether zinc dose was lower than, equal to, or
greater than the WHO recommendation.
During the course of our analyses, we identified a zinc product called Licorzinc that appeared to
be unique to China. To determine whether outcomes for Chinese studies were generalizable
comparing Licorzinc to other better established zinc products, we conducted hypothesis testing to
assess the equivalence of the mean percentage difference in episode duration between zinc and
placebo. We also calculated the RRR to compare the RR of episodes lasting >3 days between
studies using Licorzinc and other zinc products.
We plotted funnel plots to assess our primary outcomes for publication bias. We also employed
the Child Health Epidemiology Reference Group (CHERG) grading system to assess the quality of
evidence for each outcome on a four-point scale (“high”, “moderate”, “low”, “very low”) [11].

3. Results

The systematic literature search of the non-Chinese databases uncovered 4038 titles, and 15
were included after subsequent review of abstracts and full manuscripts for inclusion and exclusion
criteria (Figure 1) [5–7,12–23]. Of the included studies, 13 were conducted in a hospital setting and
two assessed episodes occurring in the community. Included studies were conducted in sites
located within 10 countries: India (n = 6); Bangladesh (n = 5); Nepal (n = 1); Turkey (n = 1);
Brazil (n = 1); Pakistan (n = 1); Ethiopia (n = 1); Yemen (n = 1); and Poland (n = 1). These
studies enrolled a total of 3271 zinc-allocated and 3314 placebo-allocated diarrhea cases.
The systematic literature search for Chinese studies resulted in 1520 titles, of which 89 were
included (Figure 1) [24–112]. All included studies were conducted in a hospital setting, and 33
studies focused on diarrhea attributable to laboratory confirmed rotavirus. None of the included
studies identified through the Chinese database were placebo-controlled; for Chinese studies, zinc
and control groups received a range of supportive treatments, including fluid infusion, probiotics
and antivirals. The total enrolment of included Chinese studies was 6198 zinc group and 6039
control group diarrhea cases. Table 1 describes the trial setting, sample size, and zinc intervention
for all included studies.
 5

Figure 1. Results of systematic literature search and review.

 6

Table 1. Characteristics of included studies.

Specific Sample Size Tablet Length of
Author Year Trial Age Group
Country Causative Zinc Control Zinc Salt or Daily Zinc Dose Supplementation
[Reference] Published Setting (months)
Organisms Group Group Syrup (days)
3–5 mos: 22.5 mg
Al Sonboli [17] 2003 Brazil Hospital Unknown 3–60 37 37 Not Listed Tablet 5
6–60 mos: 45 mg
6–11 mos: 15 mg
Bahl [7] 2002 India Community Unknown 6–35 404 401 Zinc Gluconate Syrup 14
12–35 mos: 30 mg
Brooks [16] 2005 Bangladesh Hospital Unknown 1–6 91 93 Zinc Acetate Syrup 20 mg Duration of episode
Brooks [16] 2005 Bangladesh Hospital Unknown 1–6 91 93 Zinc Acetate Syrup 5 mg Duration of episode
Dutta [23] 2011 India Hospital Unknown 6–24 44 41 Not Listed Syrup 40 mg 14
Elnemr [21] 2007 Yemen Hospital Unknown 3–24 88 92 Zinc Acetate Syrup 20 mg 14
Faruque [12] 1999 Bangladesh Hospital Unknown 6–24 343 341 Zinc Acetate Syrup 14.2 mg 15
Fischer Walker [19] 2006 Pakistan Hospital Unknown 1–5 281 279 Zinc Sulfate Tablet 10 mg 14
Fischer Walker [19] 2006 India Hospital Unknown 1–5 186 187 Zinc Sulfate Tablet 10 mg 14
Fischer Walker [19] 2006 Ethiopia Hospital Unknown 1–5 87 90 Zinc Sulfate Tablet 10 mg 14
Larson [18] 2005 Bangladesh Hospital Unknown 3–59 267 266 Zinc Sulfate Tablet 20 mg 10
Patel [20] 2009 India Hospital Unknown 6–59 264 271 Zinc Sulfate Syrup 20 mg 14
3–5 mos: 10 mg
Patro [22] 2010 Poland Hospital Unknown 3–48 81 79 Zinc Sulfate Syrup 10
6–48 mos: 20 mg
Polat [15] 2003 Turkey Hospital Unknown 2–29 52 54 Zinc Sulfate Syrup 20 mg 10
Roy [13] 1999 Bangladesh Hospital Unknown 3–24 32 35 Zinc Acetate Syrup 20 mg 14
Sachdev [5] 1988 India Hospital Unknown 6–18 25 25 Zinc Sulfate Tablet 40 mg Not Listed
Sazawal [6] 1995 India Hospital Unknown 6–35 456 481 Zinc Gluconate Syrup 20 mg Not Listed
From enrolment
6–11 mos: 15 mg
Strand [14] 2002 Nepal Community Unknown 6–35 442 449 Not Listed Syrup until 7 days after
12–35 mos: 30 mg
episode subsided
 7

Table 1. Cont.
4–5 mos: 10.8 mg
Zhao [24] 2011 China Hospital Unknown 4–36 40 40 Licorzinc Tablet 6–12 mos: 14.4 mg Not Listed
13–36 mos: 21.6 mg
Not
Zhang [25] 2009 China Hospital Rotavirus 6–24 60 60 Zinc Gluconate 20 mg Duration of episode
Listed
1.5–5 mos: 10 mg
Lin [26] 2010 China Hospital Rotavirus 1.5–36 58 58 Zinc Gluconate Syrup Duration of episode
6–36 mos: 20 mg
Not
Zhou [27] 2010 China Hospital Rotavirus 6–24 42 40 Zinc Gluconate 20 mg 14
Listed
3–5 mos: 10 mg
Yang [28] 2011 China Hospital Unknown 3–36 42 40 Zinc Gluconate Tablet 10–14
6–36 mos: 20 mg
Not 5 mos: 10 mg
Liu [29] 2010 China Hospital Unknown 5–18 40 40 Zinc Gluconate 10–14
Listed 6–18 mos: 20 mg
Not
Chen [30] 2006 China Hospital Rotavirus 0–24 30 30 Zinc gluconate 10 mg Not Listed
Listed
Liu [31] 2011 China Hospital Unknown 6.8–22 90 90 Zinc Gluconate Tablet 20 mg Not Listed
Liu [32] 2009 China Hospital Unknown 6–36 112 108 Zinc Gluconate Tablet 20 mg 10
Fu [33] 2010 China Hospital Rotavirus 2–24 98 102 Zinc Gluconate Syrup 5 mg Not Listed
2–5 mos: 7.5 mg
Not
Zhou [34] 2008 China Hospital Unknown 2–48 40 40 Licorzinc 6–12 mos: 11.25 mg 10–14
Listed
13–48 mos: 18.75 mg
Not 4–5 mos: 7.2 mg
Chen [35] 2008 China Hospital Rotavirus 4–48 60 60 Licorzinc Not Listed
Listed 6–48 mos: 10.8 mg
 8

Table 1. Cont.
1.5–5 mos: 7.5 mg
Not 6–11 mos: 11.25 mg
Guan [36] 2012 China Hospital Rotavirus 1.5–45.6 45 45 Licorzinc 10–14
Listed 12–45.6 mos:
18.75 mg
Not 4–5 mos: 10 mg
Wu [37] 2010 China Hospital Rotavirus 4–13 46 46 Licorzinc Not Listed
Listed 6–13 mos: 20 mg
Zhou [38] 2010 China Hospital Unknown 6–24 65 60 Licorzinc Tablet 20 mg Not Listed
Luo [39] 2009 China Hospital rotavirus 6–36 55 50 Licorzinc Tablet 18.75 mg Not Listed
Not
Zhang [40] 2010 China Hospital Unknown 5–48 50 50 Licorzinc Not Listed * Not Listed
Listed
6–12 mos: 11–25 mg
Ju [41] 2007 China Hospital Unknown 6–36 40 38 Licorzinc Tablet Not Listed
13–36 mos: 15 mg
Wang [42] 2012 China Hospital Unknown 6–36 30 30 Licorzinc Tablet Not Listed * 3
3–11 mos: 20 mg
Hong [43] 2009 China Hospital Rotavirus 3–60 140 120 Zinc Sulfate Syrup 12–36 mos: 30 mg Not Listed
37–60 mos: 40 mg
Lin [44] 1994 China Hospital Unknown 0.5–24 46 58 Zinc Sulfate Syrup 10–14 mg/kg * Not Listed
5 mos: 50 mg
Yan [45] 2011 China Hospital Unknown 5–36 70 50 Zinc Sulfate Syrup Not Listed
6–36 mos: 100 mg
Not
He [46] 1997 China Hospital Unknown 6–36 52 58 Zinc Gluconate 20 mg Not Listed
Listed
3–5 mos: 10 mg
Wei [47] 2011 China Hospital Unknown 3–36 44 42 Zinc Gluconate Syrup 10–14
6–36 mos: 20 mg
0–5 mos: 10 mg
Yang [48] 2012 China Hospital Unknown 0–36 80 80 Zinc Gluconate Tablet 10
6–36 mos: 20 mg
 9

Table 1. Cont.
Not 0–5 mos: 10 mg
Pu [49] 2010 China Hospital Rotavirus 0–24 38 34 Zinc Gluconate Not Listed
Listed 6–24 mos: 20 mg
Not 3–5 mos: 10 mg
Zhang [50] 2011 China Hospital Rotavirus 3–36 53 53 Zinc Gluconate 10
Listed 6–36 mos: 20 mg
1.5–5 mos: 10 mg
Sun [51] 2008 China Hospital Unknown 1.5–36 45 45 Zinc Gluconate Syrup Not Listed
6–36 mos: 20 mg
3–5 mos: 10 mg
Zhang [52] 2011 China Hospital Unknown 3–36 90 90 Zinc Gluconate Syrup Not Listed
6–36 mos: 20 mg
Lin [53] 2010 China Hospital Rotavirus 6–54 28 20 Zinc Gluconate Tablet 6–54 mos: 20 mg 14
Not 3–5 mos: 10 mg
Liu [54] 2009 China Hospital Unknown 3–36 95 91 Zinc Gluconate 10–14
Listed 6–36 mos: 20 mg
Qiao [55] 2011 China Hospital Unknown 6–36 73 72 Zinc Gluconate Tablet 6–36 mos: 20 mg 14
Not 0–5 mos: 10 mg
Zhang [56] 2007 China Hospital Unknown 0–24 85 90 Zinc Gluconate 10
Listed 6–24 mos: 20 mg
0–5 mos: 10 mg
Zhao [57] 2012 China Hospital Unknown 0–24 70 70 Zinc Gluconate Syrup 10–14
6–24 mos: 20 mg
Not 0–5 mos: 10 mg
Cai [58] 2011 China Hospital Unknown 0–24 88 84 Zinc Gluconate 14
Listed 6–24 mos: 20 mg
Zhang [59] 2012 China Hospital Rotavirus 6–17 120 120 Zinc Gluconate Tablet 20 mg 10–14
Not 0–5 mos: 10 mg
Qiao [60] 2012 China Hospital Unknown 0–24 85 85 Zinc Gluconate 10
Listed 6–24 mos: 20 mg
3–5 mos: 10 mg
Zhong [61] 2012 China Hospital Rotavirus 3–48 50 50 Zinc Gluconate Tablet 10
6–48 mos: 20 mg
Not 0–5 mos: 10 mg
Wang [62] 2011 China Hospital Rotavirus 0–24 60 60 Zinc Gluconate 10
Listed 6–24 mos: 20 mg
Not 0–5 mos: 10 mg
Yang [63] 2008 China Hospital Rotavirus 0–36 164 168 Zinc Gluconate 10
Listed 6–36 mos: 20 mg
 10

Table 1. Cont.
Zhao [64] 2012 China Hospital Rotavirus 6–36 60 60 Zinc Gluconate Syrup 35 mg 10
Not
Ma [65] 2012 China Hospital Rotavirus 4–42 41 41 Zinc Gluconate 20 mg Not Listed
Listed
Not 0–5 mos: 10 mg
Chen [66] 2012 China Hospital Rotavirus 0–36 93 93 Zinc Gluconate 10
Listed 6–36 mos: 20 mg
4–5 mos: 10 mg
Hu [67] 2009 China Hospital Rotavirus 4–36 60 60 Zinc Gluconate Tablet 10
6–36 mos: 20 mg
1–12 mos: 70 mg
Yuan [68] 2011 China Hospital Unknown 1–36 100 100 Zinc Gluconate Tablet 14
13–36 mos: 140 mg
3–5 mos: 10 mg
Tan [69] 2011 China Hospital Unknown 3-36 50 35 Zinc Gluconate Tablet 10–14
6–36 mos: 20 mg
0–5 mos: 10 mg
Liu [70] 2010 China Hospital Unknown 0–36 89 77 Zinc Gluconate Syrup 10
6–36 mos: 20 mg
3–5 mos: 10 mg
Hu [71] 2011 China Hospital Unknown 3–60 108 100 Zinc Gluconate Tablet 14
6–60 mos: 20 mg
6–12 mos: 7.5 mg
Li [72] 2008 China Hospital Unknown 6–36 40 38 Zinc Gluconate Tablet 3
13–36 mos: 15 mg
Not 3–5 mos: 10 mg
Gao [73] 2012 China Hospital Unknown 3–36 74 74 Zinc Gluconate 14
Listed 6–36 mos: 20 mg
Wu [74] 2011 China Hospital Unknown 3–60 20 20 Zinc Sulfate Syrup 10 mg 10
Not
Wu [74] 2011 China Hospital Unknown 3–60 20 20 Zinc Sulfate 10 mg 10
Listed
3–5 mos: 10 mg
Liu [75] 2011 China Hospital Unknown 3–60 54 53 Zinc Gluconate Tablet 3–5
6–60 mos: 20 mg
 11

Table 1. Cont.
Not 5 mos: 10 mg
Chen [76] 2010 China Hospital Unknown 5–36 42 20 Zinc Gluconate 10–14
Listed 6–36 mos: 20 mg
Not 2–5 mos: 70 mg
Ma [77] 2012 China Hospital Unknown 2–36 63 63 Zinc Gluconate 10–14
Listed 6–36 mos: 140 mg
Not
Lu [78] 2012 China Hospital Unknown 6–18 120 140 Zinc Gluconate 140 mg 10–14
Listed
Ma [79] 2012 China Hospital Unknown 6–36 58 52 Zinc Gluconate Syrup 6–36 mos: 20 mg 10
0–5 mos: 10 mg
Ao [80] 2012 China Hospital Rotavirus 0–24 87 80 Zinc Gluconate Syrup Not Listed
6–24 mos: 20 mg
3–5 mos: 10 mg
Gu [81] 2011 China Hospital Unknown 3–60 56 60 Zinc Gluconate Syrup 10
6–60 mos: 20 mg
Not
Wen [82] 2006 China Hospital Unknown 0–24 30 29 Zinc Gluconate 20 mg 10–14
Listed
Not
Wang [83] 2011 China Hospital Unknown 3–36 60 60 Licorzinc 10–20 mg * Duration of episode
Listed
Not
Liu [84] 2012 China Hospital Rotavirus 8–30 90 90 Licorzinc 8–30 mos: 20 mg Not Listed
Listed
3–5 mos: 10 mg
Liu [85] 2012 China Hospital Unknown 3–60 100 100 Licorzinc Tablet Not Listed
6–60 mos: 20 mg
Not 2–5 mos: 10 mg
Tong [86] 2011 China Hospital Unknown 2–36 98 98 Licorzinc Not Listed
Listed 6–36 mos: 20 mg
1–5 mos: 10 mg
Qiu [87] 2010 China Hospital Rotavirus 1–24 53 52 Licorzinc Tablet 14
6–24 mos: 20 mg
3–5 mos: 10 mg
Kong [88] 2011 China Hospital Unknown 3–30 35 35 Zinc Gluconate Tablet 6–11 mos: 15 mg 14
12–30 mos: 20 mg
Not
He [89] 2007 China Hospital Rotavirus 5–22 60 63 Zinc Gluconate 20 mg Not Listed
Listed
 12

Table 1. Cont.
Kang [90] 2010 China Hospital Rotavirus 6–36 92 80 Zinc Gluconate Tablet 20 mg 14
Not
Su [91] 2012 China Hospital Rotavirus 6–36 97 97 Zinc Gluconate 20 mg Not Listed
Listed
2–5 mos: 10 mg
Huang [92] 2010 China Hospital Rotavirus 2–36 100 100 Not Listed Tablet Not Listed
6–36 mos: 20 mg
0–5 mos: 10 mg
Zhang [93] 2006 China Hospital Unknown 0–36 83 63 Licorzinc Syrup 10–14
6–36 mos: 20 mg
Wang [94] 2012 China Hospital Unknown 4–30 60 60 Zinc Gluconate Syrup 10 mg Not Listed
0.5–5 mos: 140 mg
Lin [95] 2008 China Hospital Unknown 0.5–34 60 60 Zinc Gluconate Tablet 10–14
6–34 mos: 280 mg
Yan [96] 2011 China Hospital Unknown 6–60 57 57 Zinc Gluconate Tablet 20 mg 10
0–5 mos: 10 mg
Yu [97] 2012 China Hospital Unknown 0–36 40 40 Zinc Gluconate Tablet 10–14
6–36 mos: 20 mg
4–5 mos: 10 mg
Zhang [98] 2011 China Hospital Rotavirus 4–36 128 128 Zinc Gluconate Syrup 14
6–36 mos: 20 mg
Not 2–5 mos: 10 mg
Xu [99] 2010 China Hospital Rotavirus 2–36 84 83 Zinc Gluconate 14
Listed 6–36 mos: 20 mg
3.5–5 mos: 10 mg
Tan [100] 2010 China Hospital Unknown 3.5–60 55 55 Zinc Gluconate Syrup 10–14
6–60 mos: 20 mg
Not 2.5–5 mos: 10 mg
Shen [101] 2012 China Hospital Rotavirus 2.5–40 46 42 Zinc Gluconate Duration of episode
Listed 6–40 mos: 20 mg
Wang [102] 2010 China Hospital Unknown 6–48 52 51 Zinc Gluconate Tablet 20 mg Not Listed
 13

Table 1. Cont.
1–5 mos: 5 mg
Chen [103] 2011 China Hospital Unknown 1–36 50 50 Zinc Gluconate Tablet Not Listed
6–36 mos: 10 mg
0–5 mos: 2.5 mg
Meng [104] 2012 China Hospital Unknown 0–24 90 90 Zinc Gluconate Tablet 6–12 mos: 5 mg Not Listed
13–24 mos: 10 mg
1–5 mos: 2.5 mg
Zhong [105] 2010 China Hospital Unknown 1–24 60 60 Zinc Gluconate Tablet 6–12 mos: 5 mg 5–7
13–24 mos: 7.5 mg
Xie [106] 2010 China Hospital Rotavirus 6–36 128 124 Zinc Gluconate Tablet 20 mg Not Listed
Not 0–5 mos: 10 mg
Fan [107] 2012 China Hospital Unknown 0–36 163 121 Not Listed 10
Listed 6–36 mos: 20 mg
Zhou [108] 2012 China Hospital Rotavirus 6–24 75 75 Zinc Gluconate Syrup 20 mg 10–14
0–5 mos: 10 mg
Zhao [109] 2008 China Hospital Unknown 0–36 44 43 Zinc Gluconate Tablet Not Listed
6–24 mos: 20 mg
Not
Wan [110] 2006 China Hospital Unknown 6–36 26 24 Not Listed Not Listed Not Listed
Listed
Not
Yang [111] 2012 China Hospital Unknown 6–60 60 60 Not Listed 20 mg Not Listed
Listed
Not 0–5 mos: 10 mg
Luo [112] 2012 China Hospital Unknown 0–36 168 196 Not Listed Not Listed
Listed 6–36 mos: 20 mg
* Study not included in dose analyses.
 14

The results of the studies identified through non-Chinese databases are summarized in Tables 2
and 3. Acute episodes were 4% (95% CI: 1%–8%) shorter in duration among children treated with
zinc compared to those receiving placebo (Table 2). Among children hospitalized for diarrhea, the
duration of hospitalization was reduced by 37% (95% CI: 21%–53%) comparing the zinc and
control groups (Table 2). Stool frequency was decreased by 6% (95% CI: 2%–10%) among
zinc-treated children. Zinc-treated children had a reduced relative risk (RR) of acute diarrhea
lasting beyond three and seven days and an increased risk of vomiting (RR: 1.83; 95% CI:
1.40–2.39) (Table 3).

Table 2. Pooled means of select outcomes for non-Chinese studies.

Study Pooled Mean
Percent Difference 3
Outcome Sites 1 (95% CI) 2
N Zinc Group Control Group (%)
Duration of
13 3.51 (3.43–3.60) 3.67 (3.59–3.76) í íí
Episode (days)
Duration of
1 2.00 (1.99–2.01) 3.17 (2.38–3.96) í íí
Hospitalization (days)
Stool Output (mL) 2 391.2 (388.5–393.8) 388.8 (386.2–391.5) 0.6 (í
Stool Frequency
6 5.04 (4.88–5.19) 5.36 (5.20–5.52) í íí
(Number per day)
1 2
Individual studies may contribute more than one study site (N) to each estimate; Estimates for •
3
study sites generated by Poisson regression model weighted by sample size; Percent difference
calculated by: 100 × ((Pooled Zinc Estimate í Pooled Control Estimate)/Pooled Control Estimate); 95%
CI calculated by: Percent Difference ± 1.96 × {|(meanzinc/meancontrol)| × sqrt[(std errorzinc)2/(meanzinc)2 +
(std errorcontrol)2/(meancontrol)2]} × 100.

Table 3. Pooled relative risk of select outcomes for non-Chinese studies.

Study Pooled Estimate
Pooled Relative Risk 3
Outcome Sites 1 Percentage (95% CI) 2
N Zinc Group Control Group RR (95% CI)
Episodes > 3 days (%) 3 29.7 (26.7–32.7) 39.5 (36.3–42.7) 0.78 (0.67–0.90)
Episodes > 7 days (%) 6 10.3 (8.9–11.7) 14.9 (13.2–16.5) 0.74 (0.55–0.99)
Vomiting (%) 3 18.8 (16.0–21.6) 9.4 (7.3–11.4) 1.83 (1.40–2.39)
1 2
Individual studies may contribute more than one study site (N) to each estimate; Estimates for •
3
study sites generated by logistic regression model weighted by sample size; Estimates for • VWXGLHV
generated by random effects meta-analysis.

Outcomes pooled across studies conducted in China showed reductions in the duration of
diarrhea, hospitalization, fever, vomiting, stool output and stool frequency among zinc-treated
children with acute diarrhea attributable to rotavirus and to non-specific causes (Table 4). The
reduction in the duration of diarrhea was 37% (95% CI: 35%–39%) among non-specific episodes
and 31% (95% CI: 29%–34%) among rotavirus episodes (Table 4). The RR of diarrhea lasting
beyond three days was reduced among zinc-treated patients with non-specific (RR: 0.73; 95% CI:
0.66–0.79) and rotavirus (RR: 0.70; 95% CI: 0.63–0.78) diarrhea (Table 5; Figures 2 and 3).
 15

Table 4. Pooled means of select outcomes for Chinese studies.

Specific Study Pooled Mean
Percent Difference 3
Outcome Causative Sites 1 (95% CI) 2
Pathogens N Zinc Group Control Group (%)
Duration of Unknown 40 2.96 (2.90–3.03) 4.68 (4.60–4.77) í íí
Episode (days) Rotavirus 24 3.45 (3.36–3.54) 5.01 (4.89–5.12) í íí
Duration of Unknown 10 4.65 (4.50–4.80) 6.43 (6.25–6.61) í íí
Hospitalization
Rotavirus 2 4.15 (3.79–4.51) 6.1 (5.66–6.54) í íí
(days)
Duration of Unknown 13 1.90 (1.80–1.99) 2.81 (2.70–2.92) í íí
Fever (days) Rotavirus 4 1.96 (1.78–2.14) 3.18 (2.95–3.41) í íí
Duration of Unknown 6 1.15 (1.05–1.25) 1.53 (1.41–1.64) í íí
Vomiting (days) Rotavirus 3 1.84 (1.64–2.04) 2.49 (2.26–2.72) í íí
Unknown 1 40 (38.1–41.9) 70 (68.0–72.0) í íí
Stool Output (mL)
Rotavirus 1 278.4 (256.8–300.0) 425.4 (382.1–468.7) í íí
Stool Frequency Unknown 1 4 (3.8–4.2) 8 (7.6–8.4) í íí
(Number per day) Rotavirus 2 3.74 (3.30–4.18) 4.27 (3.77–4.77) í í
1 2
Individual studies may contribute more than one study site (N) to each estimate; Estimates for • VWXG\ VLWHV
3
generated by Poisson regression model weighted by sample size; Percent difference calculated by: 100 × ((Pooled
Zinc Estimate í Pooled Control Estimate)/Pooled Control Estimate); 95% CI calculated by: Percent Difference ±
1.96 × {|(meanzinc/meancontrol)| × sqrt[(std errorzinc)2/(meanzinc)2 + (std errorcontrol)2/(meancontrol)2]} × 100.

Table 5. Pooled relative risk of select outcomes for Chinese studies.

Specific Study Pooled Estimate Percentage
Relative Risk 3
Outcome Causative Sites 1 (95% CI) 2
Pathogens N Zinc Group Control Group RR (95% CI)
Episodes > 3 Unknown 44 31.4 (29.4–33.5) 49.2 (46.6–51.8) 0.73 (0.66–0.79)
days (%) Rotavirus 29 31.8 (29.5–34.1) 50.3 (47.4–53.3) 0.70 (0.63–0.78)
Episodes > 7
Unknown 1 26.9 (-) 39.2 (-) 0.75 (0.42–1.37)
days (%)
1 2
Individual studies may contribute more than one study site (N) to each estimate; Estimates for •
3
study sites generated by Poisson regression model weighted by sample size; Estimates for •VWXGLHV
generated by random effects meta-analysis.

We did not identify any studies reporting diarrhea-specific or all-cause mortality for inclusion in
this review. Nor did we identify non-Chinese studies reporting duration of fever or vomiting, or
Chinese studies reporting the proportion of children vomiting.
The mean episode duration and proportion of episodes lasting >3 days were not statistically
significantly different comparing zinc-treated children in Chinese and non-Chinese studies. There
was no statistically significant difference between the estimated relative risk of an episode lasting
>3 days (RRR: 1.07; 95% CI: 0.90–1.27) comparing Chinese and non-Chinese studies; therefore,
we pooled this outcome across regions (RR: 0.74; 95% CI: 0.68–0.80) (Figure 3). The percentage
difference between the mean episode duration of zinc-treated and control group children was
statistically significantly larger for Chinese compared to non-Chinese studies (p < 0.05), so this
 16

outcome was not pooled across regions. We did not have sufficient power to compare other
commonly reported outcomes by region.

Figure 2. Forest plot for the effect of therapeutic zinc supplementation on Rotavirus
diarrhea episodes >3 days.

Zinc dose was not associated with the mean percent difference in diarrhea duration comparing
zinc and control children for non-Chinese (p = 0.50) or Chinese (p = 0.12) studies. Comparing
Chinese studies that used Licorzinc to those that used other zinc supplements, there were no
statistically significant differences in the mean percent difference in the duration of rotavirus
episodes (p = 0.56), the RR of non-specific episodes lasting >3 days (RRR: 0.99; 95% CI: 0.72–
1.35), or the RR of rotavirus episodes lasting >3 days (RRR: 0.93; 95% CI: 0.68–1.26).
The percentage difference in the mean duration of non-specific episodes comparing zinc and
control group children was statistically significantly higher for Licorzinc compared to “other zinc”
studies (p = 0.01).
Our assessment of publication bias yielded largely symmetrical funnel plots for all outcomes.
 17

Under the CHERG grading system, the studies included in this review were of moderate quality
(Table 6) [11]. Effect estimates were largely consistent in directionality for all outcomes.

Figure 3. Forest plot for the effect of therapeutic zinc supplementation on non-specific
diarrhea episodes lasting >3 days.
 18

Table 6. Quality assessment of studies measuring the association between therapeutic zinc supplementation and selected outcomes.
Directness
Number of
Design Limitations Consistency Generalizability to Generalizability to
Studies
Population of Interest Intervention of Interest
1
Diarrhea Duration (mean): Moderate outcome-specific quality
All but 4 studies showing decreased mean
53 non-specific Chinese studies not Mostly South Asia and
RCT duration of diarrhea among zinc-treated Generalizable
24 Rotavirus placebo-controlled (í China (í
children (+1)
Diarrhea Duration (>3 days): Moderate outcome-specific quality 1
All studies showing decreased risk of
47 non-specific Chinese studies not Mostly South Asia and
RCT diarrhea duration >3 days among Generalizable
29 Rotavirus placebo-controlled (í China (í
zinc-treated children (+1)
Diarrhea Duration (>7 days): Moderate outcome-specific quality 1
All but one study showing decreased risk
Chinese studies not Mostly South Asia and
7 non-specific RCT of diarrhea duration >7 days among Generalizable
placebo-controlled (í China (í
zinc-treated children (+1)
Hospitalizations Duration: Moderate outcome-specific quality 1
All studies showing decreased mean
11 non-specific Chinese studies not Only one non-Chinese
RCT duration of hospitalization among Generalizable
2 Rotavirus placebo-controlled (í study (í
zinc-treated children (+1)
Stool Output: Moderate outcome-specific quality 1
All but one study showing decreased
3 non-specific Chinese studies not Only South Asia and
RCT stool output among zinc-treated children Generalizable
1 Rotavirus placebo-controlled (í China (í
(+1)
Stool Frequency: Moderate outcome-specific quality 1
All but three studies showing decreased
7 non-specific Chinese studies not Mostly South Asia and
RCT stool frequency among zinc-treated Generalizable
2 Rotavirus placebo-controlled (í China (í
children (+1)
 19

Table 6. Cont.
Vomiting: Moderate outcome-specific quality 1
All studies showing increased vomiting
3 non-specific RCT None No Chinese studies (í Generalizable
among zinc-treated children (+1)
Vomiting Duration: Moderate outcome-specific quality 1
All but one study showing decreased
6 non-specific Chinese studies not No non-Chinese
RCT duration of vomiting among zinc-treated Generalizable
3 Rotavirus placebo-controlled (í studies (í
children (+1)
Fever Duration: Moderate outcome-specific quality 1
13 non-specific Chinese studies not All studies showing decreased duration of No non-Chinese
RCT Generalizable
4 Rotavirus placebo-controlled (í fever among zinc-treated children (+1) studies (í
1
Quality assessment scoring based on previously published criteria [11].
 20

4. Discussion

The findings of our systematic review confirm and highlight the benefits of therapeutic zinc
supplementation for diarrhea among children under five years of age in low- and middle-income
countries. The effects of zinc treatment, which include reductions in episode duration, stool output,
stool frequency and length of hospitalization, were consistent across Chinese and non-Chinese studies
and non-specific and rotavirus diarrhea. These results suggest that zinc therapy of diarrhea is largely
beneficial and important in both low- and middle-income settings.
The results of the large number of Chinese trials in rotavirus diarrhea are a substantial addition to
the global evidence base because there have been no non-Chinese trials. One study in India based on a
post-hoc subgroup analysis suggested that zinc treatment was not beneficial for rotavirus diarrhea [113];
however, the evidence from China demonstrates that therapeutic zinc supplementation reduces the
duration and severity of rotavirus episodes. As rotavirus is the predominant cause of severe acute
diarrhea worldwide and most likely the leading cause of diarrhea mortality [114], zinc treatment of
rotavirus diarrhea could potentially yield large reductions in hospitalizations and deaths.
In comparison to non-Chinese studies, the difference between the mean episode duration of zinc-
treated and control group children was statistically significantly higher for Chinese studies
(p < 0.05). It is possible that this difference resulted from lack of placebo-controlled groups and
blinding among Chinese studies. However, estimates of the effects of therapeutic zinc supplementation
on other outcomes were largely consistent across study locations and we were able to generate a
pooled global effect size for the proportion of episodes >3 days. The consistency of effect estimates
between studies conducted in and outside China suggests that the lack of placebo-controlled groups in
Chinese studies did not greatly bias the results.
Zinc dose did not affect the estimate of the effect of zinc supplementation on the duration of
diarrhea for non-Chinese or Chinese studies. Although Licorzinc was associated with slightly greater
reductions in the mean duration of non-specific diarrhea than other zinc products, zinc effect sizes
were generally comparable across Chinese studies regardless of type of zinc preparation.
There is a dearth of literature meeting our inclusion criteria that assessed diarrhea-specific and
all-cause mortality. Although a previous review published mortality effect estimates [4], the sole
study reporting diarrhea-specific deaths was cluster-randomized and thus violated our inclusion
criteria [115]. In addition, three studies of all-cause mortality were also excluded from our review; one
was on persistent diarrhea [116], and two others were review papers [3,117].
Using previously published scoring criteria, the studies included in our review yielded pooled
estimates of overall moderate quality [11]. The majority of studies contributing to this review were
conducted in China and South Asia; however, studies conducted outside Asia were consistent in the
directionality of effect estimates. The consistency and quality of all outcomes bolsters the evidence in
support of oral zinc supplementation for the treatment of acute diarrhea among children under five in
low- and middle-income countries.
 21

5. Conclusions

Oral therapeutic zinc supplementation reduces the morbidity of acute diarrhea among children
under five in and outside China. Global efforts should be made to support scale-up of the WHO
recommended regimen of therapeutic zinc in all regions.

Authors’ Contributions

LML conducted the systematic review of non-Chinese studies, analysis and led the initial
manuscript preparation. CLFW assisted with the analysis and the manuscript preparation. KC and
WYJ conducted the systematic review of Chinese studies. REB conceptualized the idea and assisted
with the interpretation of the analysis and the final manuscript preparation.

Acknowledgments

We would like to thank Wei-Ju Chen, Xun Luo and Wenze Zhang for reviewing and abstracting
data from Chinese publications. Financial support for this review was provided by the Bill and
Melinda Gates Foundation to the US Fund for UNICEF for the ongoing work of the Child Health
Epidemiology Reference Group (CHERG).

Conflicts of Interest

The authors declare no conflict of interest.

References

1. Fischer-Walker, C.; Lamberti, L.; Roth, D.; Black, R. Zinc and Infectious Diseases. In Zinc in
Human Health; Rink, L., Ed.; IOS Press: Amsterdam, The Netherlands, 2011; pp. 234–253.
2. WHO/UNICEF. Joint Statement: Clinical Management of Acute Diarrhoea; WHO/UNICEF:
New York, NY, USA, 2004.
3. Lazzerini, M.; Ronfani, L. Oral zinc for treating diarrhoea in children. Cochrane Database Syst. Rev.
2013, 1, doi:10.1002/14651858.CD005436.
4. Fischer Walker, C.L.; Black, R.E. Zinc for the treatment of diarrhoea: Effect on diarrhoea
morbidity, mortality and incidence of future episodes. Int. J. Epidemiol. 2010, 39, i63–i69.
5. Sachdev, H.P.; Mittal, N.K.; Mittal, S.K.; Yadav, H.S. A controlled trial on utility of oral zinc
supplementation in acute dehydrating diarrhea in infants. J. Pediatr. Gastroenterol. Nutr. 1988,
7, 877–881.
6. Sazawal, S.; Black, R.E.; Bhan, M.K.; Bhandari, N.; Sinha, A.; Jalla, S. Zinc supplementation in
young children with acute diarrhea in India. N. Engl. J. Med. 1995, 333, 839–844.
7. Bahl, R.; Bhandari, N.; Saksena, M.; Strand, T.; Kumar, G.T.; Bhan, M.K.; Sommerfelt, H.
Efficacy of zinc-fortified oral rehydration solution in 6- to 35-month-old children with acute
diarrhea. J. Pediatr. 2002, 141, 677–682.
8. Zhang, J.S. Efficacy and effectiveness of 20 child health interventions in China: Systematic
review of Chinese literature. J. Glob. Health 2011, 1, 87–95.
 22

9. Stata Statistical Software, release 12; StataCorp LP: College Station, TX, USA, 2011.
10. Altman, D.G.; Bland, J.M. Interaction revisited: The difference between two estimates. BMJ
2003, 326, 219.
11. Walker, N.; Fischer-Walker, C.; Bryce, J.; Bahl, R.; Cousens, S. Standards for CHERG reviews
of intervention effects on child survival. Int. J. Epidemiol. 2010, 39, i21–i31.
12. Faruque, A.S.; Mahalanabis, D.; Haque, S.S.; Fuchs, G.J.; Habte, D. Double-blind, randomized,
controlled trial of zinc or vitamin A supplementation in young children with acute diarrhoea.
Acta Paediatr. 1999, 88, 154–160.
13. Roy, S.K.; Tomkins, A.M.; Haider, R.; Behren, R.H.; Akramuzzaman, S.M.; Mahalanabis, D.;
Fuchs, G.J. Impact of zinc supplementation on subsequent growth and morbidity in Bangladeshi
children with acute diarrhoea. Eur. J. Clin. Nutr. 1999, 53, 529–534.
14. Strand, T.A.; Chandyo, R.K.; Bahl, R.; Sharma, P.R.; Adhikari, R.K.; Bhandari, N.; Ulvik, R.J.;
Molbak, K.; Bhan, M.K.; Sommerfelt, H. Effectiveness and efficacy of zinc for the treatment of
acute diarrhea in young children. Pediatrics 2002, 109, 898–903.
15. Polat, T.B.; Uysalol, M.; Cetinkaya, F. Efficacy of zinc supplementation on the severity and
duration of diarrhea in malnourished Turkish children. Pediatr. Int. Off. J. Jpn. Pediatr. Soc.
2003, 45, 555–559.
16. Brooks, W.A.; Santosham, M.; Roy, S.K.; Faruque, A.S.; Wahed, M.A.; Nahar, K.; Khan, A.I.;
Khan, A.F.; Fuchs, G.J.; Black, R.E. Efficacy of zinc in young infants with acute watery
diarrhea. Am. J. Clin. Nutr. 2005, 82, 605–610.
17. Al-Sonboli, N.; Gurgel, R.Q.; Shenkin, A.; Hart, C.A.; Cuevas, L.E. Zinc supplementation in
Brazilian children with acute diarrhoea. Ann. Trop. Paediatr. 2003, 23, 3–8.
18. Larson, C.P.; Hoque, A.B.; Khan, A.M.; Saha, U.R. Initiation of zinc treatment for acute
childhood diarrhoea and risk for vomiting or regurgitation: A randomized, double-blind,
placebo-controlled trial. J. Health Popul. Nutr. 2005, 23, 311–319.
19. Fischer Walker, C.L.; Bhutta, Z.A.; Bhandari, N.; Teka, T.; Shahid, F.; Taneja, S.; Black, R.E.
Zinc supplementation for the treatment of diarrhea in infants in Pakistan, India and Ethiopia.
J. Pediatr. Gastroenterol. Nutr. 2006, 43, 357–363.
20. Patel, A.; Dibley, M.J.; Mamtani, M.; Badhoniya, N.; Kulkarni, H. Zinc and copper
supplementation in acute diarrhea in children: A double-blind randomized controlled trial. BMC
Med. 2009, 7, 22.
21. Elnemr, M.A.M.; Abdullah, A.K. Effect of zinc supplementation on morbidity due to acute
diarrhoea in infants and children in Sanaa, Yemen: A randomized controlled double blind clinical
trial. Sultan Qaboos Univ. Med. J. 2007, 7, 219–225.
22. Patro, B.; Szymanski, H.; Szajewska, H. Oral zinc for the treatment of acute gastroenteritis in
Polish children: A randomized, double-blind, placebo-controlled trial. J. Pediatr. 2010, 157,
984–988. e1.
23. Dutta, P.; Mitra, U.; Dutta, S.; Naik, T.N.; Rajendran, K.; Chatterjee, M.K. Zinc, vitamin A, and
micronutrient supplementation in children with diarrhea: A randomized controlled clinical trial of
combination therapy versus monotherapy. J. Pediatr. 2011, 159, 633–637.
24. Zhao, L.; Li, X.; Sun, W. 40 Cases of oral licorzinc asist treatment of diarrhea. Chin. J. Esthet.
Med. 2011, 20, 455 (in Chinese).
 23

25. Zhang, L. The effect of supplement zinc orally assist treatment on autumal diarrhea. Mod. Hosp.
2009, 9, 54–56 (in Chinese).
26. Lin, C.; Li, Q.; Zhang, M. Observation ofthe therapeutic efficiacy of oral zinc supplement with
infantile rotaviral enteritis. Contemp. Med. 2010, 16, 8–10 (in Chinese).
27. Zhou, X.; Sun, Y.; Ouyang, J. Analysis on the therapeutic and preventative effects of zinc
supplement on rotavirus enteritis in children. J. Xianning Univ. 2010, 24, 401–403 (in Chinese).
28. Yang, H. Clinical analysis of oral zinc adjuvant treatment for infants with acute diarrhea.
China Mod. Med. 2011, 18, 7 (in Chinese).
29. Liu, G. Clinical observation on the therapeutic efficacy of oral zinc supplement in infants with
acute diarrhea. J. Qiaihar Med. Coll. 2010, 31, 1416 (in Chinese).
30. Chen, L.; Bao, Y.; Gao, S. The comparison of the therapeutic effects on children diarrhea treated
with smecta and oral zinc. J. Pediatr. Pharm. 2006, 12, 26–27 (in Chinese).
31. Liu, H.; Nie, X. Chinical observation on the therapeutic efficacy of oral zinc supplement in
infants with diarrhea. Pract. Clin. Med. 2011, 12, 86–88 (in Chinese).
32. Liu, X.; Lan, X. Therapeutic effect of oral Zinc preparation on infantile diarrhea. Int. Med.
Health Guid. News 2009, 15, 74–75 (in Chinese).
33. Fu, Y.; Zhang, W. Chinical analysis on the therapeutic efficacy of oral zinc supplement in infants
with rotavirus enteritis. Guide China Med. 2010, 8, 260–262 (in Chinese).
34. Zhou, G. Clinical efficacy of licorzinc in treatment of infantile acute diarrhea. Youjiang Med. J.
2008, 36, 573–574 (in Chinese).
35. Chen, J. Clinical efficacy of licorzinc in treatment of infantile autumal diarrhea. J. Huaihai Med.
2008, 26, 442 (in Chinese).
36. Guan, Y.; Yingwei, Y.; Yu, J.; Zhou, Y. Clinical efficacy of licorzinc in treatment of infantile
rotavirus enteritis. Stud. Trace Elem. Health 2012, 29, 21–22 (in Chinese).
37. Wu, W.; Li, F. 46 Cases of licorzinc in treatment of infantile rotavirus enteritis. Shaanxi J.
Tradit. Chin. Med. 2010, 31, 296–297 (in Chinese).
38. Zhou, Y. Clinical observation of licorzinc treatment for children with diarrhea. Contemp. Med.
2010, 16, 20–21 (in Chinese).
39. Luo, L.E.A. Clinical observations on curative effects of licorzinc particles in adjuvant treating
rotavirus enteritis. Clin. J. Med. Off. 2009, 37, 862–863 (in Chinese).
40. Zhang, S.; Wei, M. Observation of therapeutic effect of licorzinc granules adjuvant treatment for
children with autumn diarrhea. Gems Health 2010, 5, 127–128 (in Chinese).
41. Ju, Z.; Zhang, Y.; Teng, L. Observation of therapeutic effect of licorzinc granules adjuvant
treatment for children with autumn diarrhea. Shanxi Med. J. 2007, 36, 661 (in Chinese).
42. Wang, L. Observation of therapeutic effect of licorzinc granules treatment for 30 cases of
children with autumn diarrhea. Chin. Pediatr. Integr. Tradit. West. Med. 2012, 4, 168–169
(in Chinese).
43. Hong, Y.; Wang, Z.; Shen, C. Clinical analysis of zinc sulfate adjuvant treatment for children
with rotavirus enteritis. Zhejiang Clin. Med. J. 2009, 11, 845–846 (in Chinese).
44. Lin, G.; Wang, J.; Zhou, T. Observation of therapeutic effect of zinc sulfate treatment for 46
cases of autumn diarrhea. Stud. Trace Elem. Health 1994, 11, 61–62 (in Chinese).
 24

45. Yan, M.; Liu, Y. Clinical analysis on efficacy of zinc sulfate in adjuvant treating 120 children
with diarrhea. J. Taishan Med. Coll. 2011, 32, 303 (in Chinese).
46. He, S.; Quan, Y.; Li, S. Determination of zinc in hair from infant patients with diarrhea and
observation of curative effect with a supply of zinc. Guangdong Trace Elem. Sci. 1997, 4, 42–43
(in Chinese).
47. Wei, S. Observation of therapeutic effect of zinc gluconate treatment for children with acute
diarrhea. Youjiang Med. J. 2011, 39, 50–52 (in Chinese).
48. Yang, C.; Xu, L.; Huang, Z. Observation of therapeutic effect of zinc gluconate adjuvant
treatment for children with acute diarrhea. Chin. Community 2012, 14, 186 (in Chinese).
49. Pu, Z. Observation of therapeutic effect of zinc gluconate adjuvant treatment for children with
autumn diarrhea. J. Clin. Exp. Med. 2010, 9, 1016–1017 (in Chinese).
50. Zhang, W. Analysis of therapeutic effect of zinc gluconate adjuvant treatment for children with
rotavirus enteritis. China Mod. 2011, 49, 123–124 (in Chinese).
51. Sun, Y.; Sun, H.; Bian, X. Rotavirus enteritis suooprtive treated by glucoside zinc. J. Appl.
Clin. Pediatr. 2008, 23, 1532–1533 (in Chinese).
52. Zhang, Y.; He, J. Observation of therapeutic effect of zinc gluconate adjuvant treatment for
children with rotavirus enteritis. J. Med. Theory Pract. 2011, 24, 2319–2320 (in Chinese).
53. Lin, T.; Deng, J. Effect of zinc gluconate tablets on serum zinc and diarrhea of children with
rotariras enteritis. J. Hainan Med. Univ. 2010, 16, 491–493 (in Chinese).
54. Liu, D. Clinical observation of zinc gluconate’s effect on acute diarrhea of children. Sichuan Med.
J. 2009, 30, 696–697 (in Chinese).
55. Qiao, J.; Wu, B. Observation of therapeutic effect of zinc gluconate treatment for children with
acute diarrhea. Shanghai J. Prev. Med. 2011, 23, 581–582 (in Chinese).
56. Zhang, P.; Wu, J. Observation of therapeutic effect of zinc gluconate treatment for children with
autumn diarrhea. Chin. Pract. J. Rural. 2007, 14, 29–30 (in Chinese).
57. Zhao, S.; Gao, Z. Observation of therapeutic effect of zinc gluconate treatment for children with
autumn diarrhea. Chin. J. Clin. Res. 2012, 25, 463–464 (in Chinese).
58. Cai, A. The effect of zinc gluconate on infants with acutie diarrhea. Med. J. Qilu 2011, 26,
255–256 (in Chinese).
59. Zhang, J. Observation of curative effect of zinc gluconate in treatment of children rotavirus
enteritis. Guide China Med. 2012, 10, 218–219 (in Chinese).
60. Qiao, R. Observation of therapeutic effect of zinc gluconate treatment for children with autumn
diarrhea. Shanxi Med. J. 2012, 41, 380–381 (in Chinese).
61. Zhong, Z. Therapeutic effect of zinc gluconate on infantile rotavirus enteritis. Contemp. Med.
2012, 18, 130–131 (in Chinese).
62. Wang, Z. Observation of clinical efficacy of zinc gluconate on infantile rotavirus enteritis.
China Med. Pharm. 2011, 1, 94–95 (in Chinese).
63. Ying, Y.; Mei, Q. Clinical observation of therapeutic effect of zinc gluconate on infantile
rotavirus diarrhea. Chongqing Med. 2008, 37, 2442–2443 (in Chinese).
64. Zhao, Y. Clinical observation of therapeutic effect of zinc gluconate on infantile rotavirus
diarrhea. China Foreign Med. Treat. 2012, 13, 113 (in Chinese).
 25

65. Ma, Z. Validity and security of zinc gluconate on infantile rotavirus diarrhea. China J. Pharm.
Econ. 2012, 2, 273–274 (in Chinese).
66. Chen, Q.; Su, H. Efficacy and safety analysis of zinc gluconate in treatment of children with
rotavirus diarrhea. Contemp. Med. 2012, 18, 144–145 (in Chinese).
67. Hu, X.; Zhang, X.; Chen, A. Observation of clinical efficacy of zinc gluconate on 60 cases with
infantile acute duarrhea. Pract. Clin. Med. 2009, 10, 85 (in Chinese).
68. Yuan, C.; Guan, J. Therapeutic effect of zinc gluconate adjuvant treatment on infantile acute
diarrhea. Pract. Clin. Med. 2011, 12, 76–77 (in Chinese).
69. Tan, Z. Observation of therapeutic effect of zinc gluconate adjuvant treatment for children with
autumn diarrhea. Anhui Med. J. 2011, 32, 802–803 (in Chinese).
70. Liu, J. Clinical treatment and analysis in 69 Children with chronic diarrhea. Chin. Manip.
Rehabil. Med. 2012, 3, 245–246 (in Chinese).
71. Hu, Y.; Deng, H. Observation of therapeutic effect of zinc gluconate granules for children with
acute diarrhea. Chin. Foreign Women Health 2011, 19, 205 (in Chinese).
72. Li, X.; Wang, H.; Feng, G. Observation of therapeutic effect of zinc gluconate granules for
children with autumn diarrhea. China Pharm. 2008, 11, 218–219 (in Chinese).
73. Gao, R. Clinical observation of supplemental zinc to treat 74 children with acute diarrhea.
Asia Pac. Tradit. Med. 2012, 8, 58–59 (in Chinese).
74. Wu, H. The effect of zinc supplementation on children with diarrhea. Strait Pharm. J. 2011, 23,
156–158 (in Chinese).
75. Liu, J. Observation of therapeutic effects of zinc supplementation in treatment of children with
acute diarrhea. Chin. J. Mod. Drug Appl. 2011, 5, 27–28 (in Chinese).
76. Chen, L.; Wang, P. Clinical analysis of zinc supplementation to treat infant diarrhea.
Chin. J. Misdiagn. 2010, 10, 7582 (in Chinese).
77. Ma, K. Clinical observation of zinc supplementation to treat children with acute diarrhea.
Chin. Foreign Med. Res. 2012, 10, 32–33 (in Chinese).
78. Lu, J. Clinical observation of zinc supplementation in treatment of autumn diarrhea.
Med. Recapitul. 2012, 18, 1101–1102 (in Chinese).
79. Ma, M.; Yang, H. Clinical analysis of zinc preparation adjuvant treatment for children with acute
diarrhea. Matern. Child Health Care China 2012, 27, 3847–3848 (in Chinese).
80. Ao, Z.; Wang, J.; Lin, L. Formulation adjuvant therapy efficacy in children with diarrhea.
Hebei Med. 2012, 18, 1091–1093 (in Chinese).
81. Gu, Z.; Shen, H.; Zhao, P. Effect of zinc gluconate on acute diarrhea in children. J. Clin. Pediatr.
2011, 29, 249–251 (in Chinese).
82. Wen, Y.; Wang, W.; Yang, C. Zinc supplement in adjuvant treatment of children acute diarrhea.
Chin. J. Prim. Med. Pharm. 2006, 13, 1208 (in Chinese).
83. Wang, G.; Wu, X. Observation of therapeutic effect of zinc adjuvant treatment for infantile
autumn diarrhea. China Med. Pharm. 2011, 1, 66–98 (in Chinese).
84. Liu, R. Zinc therapy in the treatment of children with rotavirus enteritis. Jilin Med. J. 2012,
33, 5648 (in Chinese).
85. Liu, G. Observation of therapeutic effects of zinc supplementation in treatment of children with
acute diarrhea. J. Med. Theory Pract. 2012, 25, 2287–2288 (in Chinese).
 26

86. Tong, W. Observation of therapeutic effects of zinc supplementation in treatment of children

with diarrhea. China Health Ind. 2011, 8, 84 (in Chinese).
87. Qiu, Y. Serum zinc level of children with rotaviral diarrhea before and after zinc
supplementation and clinical efficacy of the therapy. Chin. J. Woman Child Health Res. 2010,
21, 616–617 (in Chinese).
88. Kong, X. Cinical analysis of combination of lysine hydrochloride and zinc gluconate granules
with ribavirin for children with diarrhea in autumn and winter. Inn. Mong. J. Tradit. Chin. Med.
2011, 30, 8 (in Chinese).
89. He, Y. Observation of therapeutic effects of oral zinc supplementation adjuvant treatment for 60
cases of rotavirus enteritis. Zhejiang J. Clin. Med. 2007, 9, 1635 (in Chinese).
90. Kang, C. Serum zinc levels and zinc supplementation treatment of young children with rotaviral
enteritis. J. Kunming Med. Univ. 2010, 31, 109–113 (in Chinese).
91. Su, M. Observation of effect of serum zinc levels and zinc supplementation treatment of young
children with rotaviral enteritis. China Health Ind. 2012, 6, 76–77 (in Chinese).
92. Huang, Z. Observation of therapeutic effects of zinc adjuvant treatment for 200 cases of infantile
rotavirus enteritis. Chin. J. Ethnomed. Ethnopharm. 2010, 19, 80 (in Chinese).
93. Zhang, J. Observation of therapeutic effect of licorzinc granules with smectite powder for
infantile diarrhea. Jiangxi Med. J. 2006, 41, 500–501 (in Chinese).
94. Wang, D. Observation of therapeutic effect of zinc treatment for infantile diarrhea. China Health
Care Nutr. 2012, 5, 249 (in Chinese).
95. Lin, T. Observation of therapeutic effect of zinc gluconate adjuvant treatment for 60 cases of
infantile acute diarrhea. Chin. Community 2008, 24, 29 (in Chinese).
96. Yan, P. Observation of therapeutic effect of zinc gluconate adjuvant treatment for children with
acute diarrhea. China Foreign Med. Treat. 2011, 30, 118–119.
97. Yu, J. Clinical observation of therapeutic effect of zinc gluconate adjuvant treatment for 80 cases
of children with diarrhea. Jilin Med. J. 2012, 26, 5644 (in Chinese).
98. Zhang, X. Effect of zinc gluconate among the treatment for infantile rotavirus diarrhea.
Chin. J. Midiagn. 2011, 11, 7848 (in Chinese).
99. Xu, X. Observation of therapeutic effect of zinc gluconate treatment for children with rotavirus
enteritis. Jiangsu Med. J. 2010, 36, 2327–2328 (in Chinese).
100. Tan, J. Observation of therapeutic effect of zinc gluconate treatment for children with autumn
diarrhea. Med. J. Chin. People Health 2010, 22, 1122 (in Chinese).
101. Shen, D. Clinical observation of therapeutic Effect of zinc gluconate on infantile rotavirus
enteritis. Guide China Med. 2012, 10, 46–47 (in Chinese).
102. Wang, J. Study of effect of zinc gluconate combined with Saccharomyces boulardii Sachets in
treatment for children with autumn diarrhea. Contemp. Med. 2010, 16, 141–142 (in Chinese).
103. Chen, L. Clinical observation of therapeutic effect of Lysine and zinc gluconate adjuvant
treatment for infantile acute diarrhea. Chin. J. Mod. Drug Appl. 2011, 5, 85–86 (in Chinese).
104. Meng, F. Observation of therapeutic effect of Lysine and zinc gluconate treatment for infantile
autumn and winter diarrhea. Chin. Community 2012, 14, 152 (in Chinese).
105. Zhong, L. Observation of therapeutic effect of Lysine and zinc gluconateadjuvant treatment for
infantile acute diarrhea. Chin. J. Aesthet. Med. 2010, 19, 258 (in Chinese).
 27

106. Xie, J. Serum zinc level of rotaviral enteritis in children and the significance of zinc treatment.
J. Pediatr. Pharm. 2010, 16, 18–20 (in Chinese).
107. Fan, J. Application of zinc preparation in treatment for infantile diarrhea. Chin. J. Misdiagn.
2012, 12, 1306 (in Chinese).
108. Zhou, X. Therapeutic effect of zinc in the prevention and treatment of children rotvirus enteritis.
Chin. Gen. Pract. 2012, 15, 1393–1394 (in Chinese).
109. Zhao, Y. Observation of effect of zinc preparation adjuvant treatment for infantile acute diarrhea.
Zhejiang J. Prev. Med. 2008, 20, 44 (in Chinese).
110. Wan, H. Observation of therapeutic effect of Treasured zinc and selenium in adjuvant treatment
for infantile diarrhea. Shandong Med. J. 2006, 46, 65 (in Chinese).
111. Yang, S. Observation of therapeutic effect of zinc preparation in adjuvant treatment for infantile
autumn diarrhea. Health World 2012, 2, 244 (in Chinese).
112. Luo, Y. Observation of therapeutic effect of zinc adjuvant treatment for 168 cases of infantile
acute diarrhea. Public Med. Forum Mag. 2012, 16, 3338 (in Chinese).
113. Patel, A.B.; Dibley, M.J.; Mamtani, M.; Badhoniya, N.; Kulkarni, H. Influence of zinc
supplementation in acute diarrhea differs by the isolated organism. Int. J. Pediatr. 2010, 2010,
doi:10.1155/2010/671587.
114. Fischer Walker, C.L.; Rudan, I.; Liu, L.; Nair, H.; Theodoratou, E.; Bhutta, Z.A.; O’Brien, K.L.;
Campbell, H.; Black, R.E. Global burden of childhood pneumonia and diarrhoea. Lancet 2013,
381, 1405–1416.
115. Baqui, A.H.; Black, R.E.; El Arifeen, S.; Yunus, M.; Chakraborty, J.; Ahmed, S.; Vaughan, J.P.
Effect of zinc supplementation started during diarrhoea on morbidity and mortality in
Bangladeshi children: Community randomised trial. BMJ 2002, 325, 1059.
116. Roy, S.K.; Tomkins, A.M.; Mahalanabis, D.; Akramuzzaman, S.M.; Haider, R.; Behrens, R.H.;
Fuchs, G. Impact of zinc supplementation on persistent diarrhoea in malnourished Bangladeshi
children. Acta Paediatr. 1998, 87, 1235–1239.
117. Fontaine, O. Effect of zinc supplementation on clinical course of acute diarrhoea. J. Health
Popul. Nutr. 2001, 19, 339–346.

© 2013 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).
28
 29

Reprinted from Nutrients. Cite as: Inoue, M.; Binns, C.W. Introducing Solid Foods to Infants in the
Asia Pacific Region. Nutrients 2014, 6, 276-288.

Introducing Solid Foods to Infants in the Asia Pacific Region

Madoka Inoue * and Colin W. Binns

School of Public Health and Curtin Health Innovation Research Institute, Curtin University,
GPO Box U1987 Perth, Western Australia 6845, Australia; E-Mail: C.Binns@curtin.edu.au

* Author to whom correspondence should be addressed; E-Mail: madoka.inoue@curtin.edu.au;

Tel.: +61-8-9266-1661; Fax: +61-8-9266-2958.

Received: 17 November 2013; in revised form: 20 December 2013 / Accepted: 24 December 2013 /
Published: 6 January 2014

Abstract: For infants’ optimal growth and development, the introduction of nutritionally
suitable solid foods at the appropriate time is essential. However, less attention has been
paid to this stage of infant life when compared with studies on breastfeeding initiation and
duration. The practice of introducing solid foods, including the types of foods given to
infants, in the Asia Pacific region was reviewed. In total nine studies using the same
questionnaire on infant feeding practices were analysed to gain a better understanding of
trends in the introduction of solid foods in this region. All studies showed less than optimal
duration of exclusive breastfeeding indicating an earlier time of introduction of solid foods
than recommended by the WHO. Most mothers commonly used rice or rice products as the
first feed. In many studies, the timing of introducing solid foods was associated with
breastfeeding duration. Compared with the Recommended Nutrient Intakes for infants aged
above six months, rice/rice products are of lower energy density and have insufficient
micronutrients unless they have been fortified. Although the timing of introducing solid
foods to infants is important in terms of preventing later health problems, the quality of the
foods should also be considered. Recommendations to improve the introduction of solid
foods include measures to discourage prelacteal feeding, facilitating breastfeeding
education and providing better information on healthier food choices for infants.

Keywords: complementary foods; infants; Asia pacific region; infant feeding practices
 30

1. Introduction

Appropriate nutrient intake, in quantity, bioavailability, and timing in infancy are essential for
optimal growth and development. Exclusive breastfeeding for six months and then the introduction of
nutritious complimentary or solid foods, while breastfeeding continues, contributes to the prevention
of acute and chronic diseases in early and later life [1]. Most reviews have concluded that “exclusive
breastfeeding” for the first six months of life provides sufficient nutrients for infants for around
six months, and then appropriate “complementary foods” should be introduced with continued
breastfeeding, preferably until around two years of age or longer [2–4]. Both breastfed and infant
formula fed infants should be introduced to safe and nutritious “complementary foods” at around
six months to prevent retardation of growth and to minimize the risk of nutrient deficiencies [5].
“Complementary foods” are defined as foods other than breastmilk, infant formula or follow-on
formula given to infants and these can be liquids, semi-liquids, and solids [6]. When these foods,
particularly solid foods, are introduced to infants, textures should be changed as appropriate to the age
of infants to give a variety of textural experiences. It is widely believed that foods with a pureed
texture should be the first solid foods introduced [7]. “Solid foods” can be defined as non-drinkable
food made by the food industry or by the family [8]. Complementary foods must also be nutritionally
adequate and provide the bioavailable nutrients required, in combination with breastmilk, to meet all
needs for growth and optimal health [4]. Since the nutrients in breastmilk are generally more
bioavailable than from other sources, breastmilk remains an important component of nutrition after the
introduction of solids. The term, “weaning” is often used to describe the infants who start taking solid
foods [9] but “weaning” usually indicates a transition period or process from breastmilk or infant
formula to solid foods [7]. It often refers to different events in different cultures and so it will not be
used in this paper.
The timing of the introduction of solid foods is important because the early introduction of solid
foods to infants by definition results in a shorter duration of “exclusive breastfeeding”, which in turn
increases the risk of morbidity. Several studies have found that the early introduction of solid
foods before six months of age has been associated with an increased risk of diarrheal disease or
gastro-intestinal infection in infancy [10,11], food allergies [12], and overweight or higher Body Mass
Index (BMI) in childhood [13]. The early introduction of solid foods may also change the composition
of gastro-intestinal bacteria, the microbiome, which has implications for health [14–16]. In contrast,
the late introduction of solid foods (after six months of age) predisposes to micronutrient deficiencies
including iron and zinc status, which affect cognitive and neurological development [17,18] and may
lead to other problems, including feeding difficulties [19].
In the most recent guidelines for infant feeding launched by the WHO/UNICEF in 2003, the
introduction of solid foods to infants is recommended at six months of age (180 days) [20] but other
international organizations recommendations may differ slightly from the WHO recommendation.
The nutrition committee of the European Society for Pediatric Gastroenterology, Hepatology and
Nutrition (ESPGHAN), recommended that the introduction of solid foods should not be commenced
before 17 weeks of age and not later than 26 weeks of age. [21]. The American Academy of
Paediatrics (AAP) stated that solid foods should not be commenced before six months of age [22].
Despite the WHO recommendations, many mothers have tended to introduce solid foods to their
 31

infants before six months. A cohort study of 401 mothers in Ireland found that the median age of
introducing solid foods to infants was 16 weeks (interquartile range = 14–17.7), while 22.6% of
the mothers introduced solid foods to their infants before 12 weeks postpartum [7]. In a British study
(n = 604), the median age of introducing solid foods reported by the mothers was 15 weeks
(interquartile range = 13–16), despite the national government recommendation to start solid foods for
infants from six months of age [11].
In the Asia Pacific region, the recommended timing for the introduction of solid foods for infants
varies between countries. For instance, in China, the Ministry of Health formerly recommended the
introduction of solid foods to infants after four months of age (16 weeks), but more recently they have
changed to six months of age [23]. In Australia, the National Health and Medical Research Council
Infant Feeding Guidelines state that “solid foods” should be commenced at around six months of age [4].
The introduction of solid foods to infants is influenced by many cultural factors and traditional beliefs.
For example, in Japan, a traditional ceremony is usually held at 100 days after birth and this ceremony
was the time when mothers started to introduce additional liquid foods, including fruit juice and
vegetable soup. Although the recent guidelines state that it is not necessary to provide any liquids other
than breastmilk before six months, the ceremony still remains as a traditional custom in Japan. While
changes in duration and exclusivity of breastfeeding have been extensively researched, less is known
about how the patterns of infant feeding, including introducing solid foods, are changing and how
these changes relate to differences in cultural practices between countries. The aim of this study was to
review the timing of the introduction of solid foods in the Asia Pacific region by comparing and
contrasting previous studies that used the same questionnaires and to describe the types of foods that
are introduced.

2. Methods

2.1. Study Details and Descriptions of Sample Recruitments

Nine studies (five countries) undertaken by Curtin University, School of Public Health, of infant
feeding practices conducted in the Asia Pacific region were reviewed including data on the first
introduction of solid foods to infants. The recruitment process of the samples in each study has been
published in detail elsewhere [24–30]. In Australia, the Perth Infant Feeding Study I (PIFS I) was
undertaken to obtain information about infant feeding practices and provide information to assist in
developing the Infant Feeding Guidelines and was repeated a decade later as the Perth Infant Feeding
Study II (PIFS II). The questionnaires and methodology developed for these initial studies were then
used in a similar way in other countries allowing for comparisons to be made. All of the Australian
mothers in these studies were recruited in hospitals after birth. The Vietnamese mothers were recruited
after giving birth in hospital, in community health centres, or at their home. In the two Chinese studies,
all mothers were recruited at either hospitals or health centres as birthing at home is uncommon in this
country. The mothers in the Zhejiang study were recruited from a city (Hangzhou), suburban, and rural
areas. The Maldivian mothers were recruited at clinics associated with hospitals, the Japanese mothers
were recruited at community health centres when they came for health examinations of their infants at
18 months of age and one study of mothers who have migrated to Australia used community samples
 32

recruited using “snowball” techniques. In the seven cohort studies the mothers were asked to complete
an initial questionnaire on their infant feeding practices while in hospital and were then followed up at
regular intervals for six months. Similar questionnaires were used in all of the studies and the three
cross-sectional studies have been included in this review. The mothers in the Maldives study are
representative of a conservative Islamic society, while the other countries are generally representative
of Asia Pacific cultures.

2.2. Study Questionnaire

All reviewed studies used the same questionnaires on infant feeding practices and demographic
details. Where necessary for cultural, translation, and ethical reasons several questions were modified.
Specific areas of the questionnaire include:
x Demographics: maternal age, occupation, marital status, method of birth, parity, family income,
husband’s/partner’s occupation, maternal smoking status, alcohol intake during lactation,
infants’ birth weight
x Infant feeding practices: timing of feeding changes, intention to breastfeed, expressing
breastmilk, breastfeeding problems, the reasons for ceasing breastfeeding, starting time of solid
food and infant formula, and types of the first solid food given to infants.

2.3. Ethical Consideration

All of the studies were approved by the Human Research Ethics Committees of Curtin University
and other relevant authorities. These included the local health authorities in Vietnam and Japan,
and participating hospitals in China, Australia, and the Maldives. Confidentiality was assured and
mothers were advised that their participation was voluntary and that they could withdraw at any time
without prejudice.

3. Results

The details of each study including the study year, country, sample size, response rate, and
methodology are shown in Table 1. Table 2 lists the sample characteristics reported by mothers in each
study. In all studies, the majority of mothers were married and mothers in industrialized countries,
particularly in cities, had more years of education than those in developing countries, but aboriginal
mothers had the lowest education levels. In Australia, there are similar sample characteristics in
maternal age in the PIFS I and II studies, while the Aboriginal mothers in PABS are younger. In both
Chinese studies (Xinjiang, a remote area that is located in the Northwest, and Zhejiang, an
industrialized province in Eastern China) most mothers were primiparous (75.8% and 88.6%,
respectively) and there were higher rates of caesarian section (44.1% and 67.0%, respectively), than in
the other studies. The Himeji study that was undertaken in the central part of Japan had the highest rate
of low birth weight (8.4%) and unemployment in mothers (71.4%) among the studies.
 33

Table 1. Details of the infant feeding studies in Asia Pacific region.

Response rate
* Authors Data collection periods Country Study name Sample size Study method
(%)
1 Binns et al. 1992/93 Australia Perth Infant Feeding Study I (PIFS I) 556 77 cohort
2 Binns et al. 2001/02 Australia Perth Aboriginal Breastfeeding Study (PABS) 425 93 cohort
Viet
3 Duong et al. 2002 Rural Viet Nam Infant Feeding Study 463 96 cohort
Nam
4 Binns et al. 2002/03 Australia Perth Infant Feeding Study Mark II (PIFS II) 587 68 cohort
5 Xu et al. 2002/03 China Xinjiang Infant Feeding Study 1219 97 cohort
6 Abdulraheem et al. 2004 Maldives Maldives Infant Feeding Study 251 81 cross-sectional
7 Qiu et al. 2004/05 China Zhejiang (Hangzhou) Infant Feeding Study 1520 96 cohort
8 Li et al. 2002 Australia Chinese Infant Feeding Study living in Perth 506 95 cross-sectional
9 Inoue et al. 2007 Japan Himeji Infant Feeding Study 1612 69 cross-sectional
* Sources: Study1: [31]; 2: [32]; 3: [29]; 4: [27]; 5: [28]; 6: [33]; 7: [34]; 8: [26]; 9: [25].

Table 2. The sample characteristics reported by mothers in each study.

Study 1 * 2 3 4 5 6 7 8 9
Variables
n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%)
Age (year)
<25 163 (29.3) 300 (70.6) 154 (26.2) 184 (15.1) 198 (78.8) 358 (23.6) 33 (6.5) 67 (4.2)
25–29 193 (34.7) 80 (18.8) 26.4 (4.97) 170 (29.0) 544 (44.6) (below 30) 800 (52.6) (below 30) 170 (29.0)
30–34 135 (24.3) 41 (9.6) (mean ± 178 (30.3) 307 (25.2) 53 (20.4) (above 473 (93.5) 722 (44.8)
338 (22.2)
35” 60 (10.8) (above 30) SD **) 84 (14.3) 66 (5.4) 30) (above 30) 411 (25.5)
No response 6 (1.0) 4 (0.9) 1 (0.2) 118 (9.7) 0 (0.0) 24 (1.6) 0 (0.0) 61 (3.8)
Marital status
Married/Defacto 370 (87.1) 540 (92.0) 239 (95.2) 1518 (99.9) 498 (98.2) 1535 (95.2)
Others N/A 52 (12.2) N/A 47 (8.0) N/A 12 (3.2) 2 (0.1) 8 (1.8) 46 (3.1)
No response 53 (0.7) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 28 (1.7)
<12 292 (52.5) 385 (90.8) (85.5) 249 (43.2) 781 (64.1) 222 (88.5) 915 (60.5) 5 (1.0)
• 248 (44.6) 34 (8.0) (14.5) 328 (56.8) 355 (29.1) 29 (10.4) 599 (39.6) 501 (99.0) N/A ***
No response 16 (2.9) 6 (1.2) 0 (0.0) 0 (0.0) 83 (6.8) 0 (0.0) 2 (0.1) 0 (0.0)
 34

Table 2. Cont.
Parity
Primiparous 170 (30.6) 287 (66.8) 216 (36.8) 924 (75.8) 1347 (88.6) 780 (44.0)
Multiparous 383 (68.9) 107 (25.2) N/A 371 (63.2) 199 (16.3) N/A 163 (10.7) N/A 899 (55.7)
No response 3 (0.5) 31 (7.3) 0 (0.0) 96 (7.7) 10 (0.7) 5 (0.3)
Method of birth
Vaginal 454 (81.7) 214 (50.4) 428 (92.4) 411 (70.0) 602 (49.4) 495 (32.6) 1370 (85.0)
Caesarean 97 (17.4) 210 (49.4) 29 (6.3) 171 (29.1) 537 (44.1) N/A 1019 (67.0) N/A 240 (14.9)
No response 5 (0.9) 1 (0.2) 6 (1.3) 5 (0.9) 80 (6.6) 6 (0.4) 2 (0.1)
Birth weight of infants
<2500 g 25 (4.5) 96 (22.6) (3.0) 13 (2.2) 37 (3.0) 27 (1.8) 135 (8.4)
•2500 g 531 (95.5) 329 (77.4) (97.0) 566 (96.4) 1131 (92.8) N/A 1479 (97.3) N/A 1368 (84.9)
No response 0 (0.0) 0 (0.0) 0 (0.0) 8 (1.4) 51 (4.2) 14 (0.9) 109 (6.8)
Maternal occupation
Unemployed 0 (0.0) 64 (10.9) 500 (41.0) 112 (44.6) 221 (14.5) 270 (53.4) 1151 (71.4)
Employed N/A N/A 463 (100.0) 512 (87.2) 644 (52.8) 139 (55.4) 1255 (82.6) 236 (46.6) 222 (26.2)
No response 0 (0.0) 11 (1.9) 75 (6.2) 0 (0.0) 44 (2.9) 0 (0.0) 39 (2.4)
Partner’s/husband’s occupation
Unemployed 315 (74.1) 0 (0.0) 12 (4.8) 34 (2.2) 0 (0.0)
Employed N/A 93 (21.9) 460 (99.4) N/A N/A 239 (95.2) 1420 (93.4) N/A 1492 (92.6)
No response 17 (4.0) 3 (0.6) 58 (4.4) 120 (7.4)
Maternal smoking status
Yes 299 (69.2) 196 (33.3) 192 (11.9)
No N/A 126 (29.2) N/A 321 (54.7) N/A N/A N/A N/A 1390 (86.2)
No response 6 (1.6) 70 (12.0) 30 (1.9)
Partner’s/husband’s smoking status
Yes 64 (46.0) ȥ 705 (57.8) 786 (48.4) #
No N/A 75 (54.0) ȥ N/A N/A 383 (31.4) N/A N/A N/A 712 (44.2)
No response 0 (0.0) 131 (10.8) 114 (7.1)
Source: 1: [31]; 2: [32]; 3: [29]; 4: [27]; 5: [28]; 6: [33]; 7: [34]; 8: [26]; 9: [25]; * Study numbers and sources are the same as above Table 1; ** SD = Standard Deviation;
*** N/A = Not applicable as the ethics issues arisen; # including other family members; ȥ n= 139.
 35

Table 3 presents the median age of introducing solid foods to infants in the Asia Pacific region.
Zhejiang is the earliest at 3.8 months, while Maldives and Japan were 5.5 months of age. In Vietnam,
some mothers (4.8%) introduced solid foods to their infants as early as one week postpartum while the
median age was approximately 4 months. Japanese mothers residing in Perth introduced solid foods
earlier than those who are living in Japan. For Australian mothers the timing of introducing solid food
to their infants changed over the decade between PIFS I and PIFS II with an increase in the mean age
from 4.0 to 4.4 months. The most common first solid foods given to infants are rice or rice products in
Asia Pacific region (Table 4) except in the Maldives where their traditional food which is made with
wheat flour and fish, and Chinese migrants to Australia (egg-yolk). It is also interesting to note that
over 40% of Vietnamese mothers used monosodium glutamate in the preparation of solid foods for
infants [8].

Table 3. The median age of the first introducing solid foods (in months) by the studies.
* Study name Median age (SD) **
1 Perth Infant Feeding Study I (PIFS I) 4.0
2 Perth Aboriginal Breastfeeding Study (PABS) 4.7
3 Rural Viet Nam Infant Feeding Study 4.0
4 Perth Infant Feeding Study Mark II (PIFS II) 4.4
5 Xinjiang Infant Feeding Study 4.0
6 Maldives Infant Feeding Study 5.5 (2.0)
7 Zhejiang (Hangzhou) Infant Feeding Study 3.8
8 Chinese Infant Feeding Study living in Perth N/A #
9 Himeji Infant Feeding Study 5.5 (1.1)
* Study numbers are the same as Table 1; ** SD = Standard Deviation; # N/A = Not available.

Table 4. The type of solid foods given to infants in the Asia Pacific region.
Study Study The most The second The third
number * location popular food popular food popular food
Rice cereal Milk Custards
1 Australia, 1992 Fruit gels, puree
(commercial) Yoghurt
Australia Rice cereal Commercial
2 Milk Custards Yoghurt
(Aboriginal mothers) (commercial) foods with meat
3 Viet Nam Rice porridge Rice-floured porridge Meat and egg
Fresh/processed fruits
4 Australia, 2002 Rice cereal N/A
and vegetables
5 China (Xinjiang) Rice paste Rice porridge Vegetable paste
Maldivian food
6 Maldives made with wheat Rice porridge Processed food
flour and fish
7 China (Hangzhou) Rice cereal Rice porridge Mashed egg, fish
Australia
8 Egg-Yolk Commercial infant food Fruit
(Chinese migrants)
9 Japan # Rice gruel Japanese noodles Puree vegetables
#
N/A = Not Applicable; Reference [35].
 36

Associations between the timing of introducing solid foods and breastfeeding duration were
explored in each study. In Australia, in the PIFS II study, mothers who introduced solids at or after
17 weeks had 11 weeks longer duration of breastfeeding than those who introduced solids before
17 weeks (p < 0.001). The Japanese study also found that the timing of the introduction of solid foods
was associated with the duration of “any breastfeeding” until six months of age (OR = 1.21,
95% CI = 1.10–1.33). Among Chinese migrants to Australia, mothers introduced solid foods to their
infants at similar times to other Australian infants, but this was delayed when compared with mothers
in home countries. In Viet Nam, significant factors associated with delayed introduction of solid food
at 24 weeks were “if mother was a farmer” (OR = 0.52, 95% CI = 0.18–0.95) and “completed
secondary school” (OR = 0.28, 95% CI = 0.10–0.54), whose “husband was satisfied with the infant’s
gender” (OR = 0.30, 95% CI = 0.17–0.53), her “mother-in-law preferred exclusive breastfeeding” (OR
= 0.18, 95% CI = 0.04–0.75), or her ‘friends practised exclusive breastfeeding’ (OR = 0.41, 95%
CI = 0.16–1.10).

4. Discussion

While the timing of introducing solid foods varies between countries, most infants in the Asia
Pacific region were introduced to solids earlier than recommended by the WHO. The mean age of
introducing solid foods to infants in China (Hangzhou) was 3.8 months, the earliest in these studies,
while Japan and Maldives were 5.6 months, closest to the WHO recommended age. Moreover, some
studies showed that the timing of the introduction of solid foods was related to not only breastfeeding
duration but also maternal occupation, education background, surrounding environments including
preferences of family or friends on infant feeding methods. While the timing of solid food introduction
is important in reducing problems related to infant health and development, the WHO has also
emphasized the importance of the quality of the foods. Solid foods given to infants are often of high
volume, with low energy and nutrient density together with a low meal frequency [36]. Our review
found that many countries in the Asia Pacific region used rice porridge/cereal (See Table 4) for
infants’ first foods since rice is culturally believed to help with digestion. Although some countries,
including Japan, excluded this question for ethical reasons, other reports still described that the most
common first solid foods was rice gruel [35,37]. These rice products are often of low energy and
micronutrient density, including iron, zinc and calcium. In a report by Dewey and Brown [36], the
WHO/UNICEF documented that energy requirements from solid foods for infants aged 6–8 months
should be 269 kcal per day (1125.5 kJ) and the infants would be able to obtain sufficient energy if they
were fed at least three meals with a minimum energy density of 1.0 kcal (4.2 kJ)/g. However, rice
porridge has only 37.8 kcal (158 kJ) per 100 g (0.378 kcal/g), a low energy food (See Table 5) [38].
While the WHO report recommended that infants aged 6–8 months, 9–11 months, and 12–24 months
should be fed at least 2–3 times, 3–4 times, and 3–4 times per day respectively, this is only applicable
when energy and nutrient density is appropriate for the infants age [39]. For infants who are fed rice
porridge to meet their energy requirements following the WHO recommendations, they would have to
be fed approximately seven times per day. Similarly, the supply of micronutrient composition in rice
products is less than the recommended nutrient intakes (Table 6). Several studies have shown that
breastfed infants have better absorption of micronutrients, including iron. However, after six months of
 37

age, the quantities of micronutrients in breastmilk become inadequate over time, particularly for
iron [40,41]. As this happens to both breast and bottle fed infants, the quality and timing of
introduction of solid foods is important in providing adequate micronutrient intakes. In both developed
and less developed countries, poor choices of solid foods may lead to nutritional deficiencies.

Table 5. Nutritional composition of rice porridge, rice cereal and egg yolk (value per 100 g).
Main nutrients Rice porridge Rice cereal * Hard-boiled egg yolk
Energy, including dietary fibre (kJ) 158 1537 1450
Protein (g) 0.7 6.8 16.1
Fat (g) 0.1 1.1 31.7
Calcium (mg) 2 6 115
Iodine (ug) 0.8 3.7 127.7
Iron (mg) 0.08 15.5 4.8
Zinc (mg) 0.12 7.8 2.7
Riboflavin (B2) (mg) 0.002 1.9 0.42
Pyridoxine (B6) (mg) 0.01 0 0.33
Vitamin C (mg) 0 33 0
Folate, natural (μg) 1 70 177
Source: [38]. * Note = this products was added vitamins B1, B2, B3, C, folate, iron and zinc.

Table 6. Recommended nutrient intakes for infants aged 7–12months and 12–24 months.
Main nutrients 7–12 months 12–24 months
Protein(g/day) NA NA
Calcium (mg/day) 400 500
Iodine (μg/day) 90 90
Iron (mg/day) 0.93 # 0.58 #
Zinc (mg/day) 4.1 * 4.1 *
Riboflavin (B2) (mg/day) 0.4 0.5
Pyridoxine (B6) (mg/day) 0.3 0.5
Vitamin C (mg/day) 30 30
Folate, natural (μg/day) 80 150
Source: [42]; Note: * = Moderate bioavailability; # = 95th percentile absolute requirements.

In developing countries, the inappropriate introduction of solid foods at an early age may be
reflected in the proportion of stunting and/or wasting in young children [43]. Breastfeeding and
nutritious solid foods play key roles in promoting appropriate nutrition for their growth and
development and thus the quality of solid foods need to be focused to reduce the prevalence of under
nutrition or malnutrition.
A meta-analysis on the impact of nutritional interventions on infant survival, disease prevention,
and stunting concluded that child stunting could be reduced by approximately one third, if nutritional
interventions were provided to infants before 36 months of age [44]. It is important to emphasize
appropriate nutritious solid foods given to infants at the appropriate time. Golden [45] estimated the
Recommended Nutrient Intakes (RNIs) for children who are moderately malnourished, and suggested
the importance of a balance in nutrients between the macro- and micro-nutrients. This study also
 38

recognised the importance of nutrient density in the developing world, as many of the earliest foods
introduced to infants are high volume with a low nutrient density.
Although our study showed that Maldives almost reached the WHO recommended age of
introducing solid foods (5.5 months), the stunting rate under five years old was still 19% between
2006 and 2010 [46]. A more recent study in the Maldives found that within the first seven days after
birth approximately 39% and 16% of infants (n = 458) were fed honey and dates, respectively,
suggesting that the earlier study may have underreported prelacteal and early life feeds [47]. These
prelacteal and early infancy feeds were related to specific cultural beliefs, but may also have had
detrimental effects on infant health and the incidence of stunting.
In Japan, the mean age for the introduction of solid foods is approximately 5.5 months, and
prelacteal feeds are still common, in contrast to the WHO recommendations for exclusive
breastfeeding. The first priority for mothers is to continue exclusive breastfeeding for the first six months
of life and then introduce nutritious complementary foods, appropriate nutrition during the 6–24 months
period is also critical for infants’ nutrition and development [2]. Parents should be provided with more
detailed information about introducing solid foods, including the quantity, timing, and quality of the
foods through breastfeeding education since nutritional status during the first two years of life is
critical in terms of their lifelong physical growth and mental development [5].
There are several limitations to consider when drawing conclusions from this study. Although these
studies used almost the same questionnaire on infant feeding practices and included WHO standard
infant feeding definitions, the sample selection and sizes used mean that the results may not be
representative of the whole of the country. Nevertheless, similar methodology used in each study
means that the main conclusions of this review can be used for nutrition education. The principal
finding of the review is that most countries do not achieve the WHO goal for the timing of the
introduction of solid foods. Increased promotion of optimum infant feeding guidelines is needed,
including guidance for the appropriate time and manner in which solid foods are intoduced. This is an
important public health message for infant nutrition in the Asia Pacific region.

5. Conclusions

The review of previous observational studies using the same questionnaire on infant feeding
practices in the Asia Pacific region has shown that many countries need further improvement in the
timing and the quality of first feeds with solid foods. This should be in conjunction with promoting the
optimal duration of exclusive breastfeeding. Rice and rice products are commonly used as the first
foods in this region and are of low energy density. Without fortification they provide insufficient
quantities of micronutrients. Education of not only the mothers, but also other family members, health
professionals and the community should be provided in order to facilitate understanding about the
importance of breastfeeding and the appropriate introduction of solid foods. Several strategies
including a general prohibition of prelacteal feeding in hospitals (except in specific medical
circumstances), a ban on distribution of free gifts of infant formula to mothers, and an expansion of the
roles of midwives should be explored. Further studies on this topic are required for a better
understanding and evaluation of growth and development, and will be able to contribute to the
development of more effective strategies in pediatric nutrition in this region.
 39

Acknowledgments

The authors are grateful to all mothers and staff who assisted in all of these studies.

Conflicts of Interest

The authors declare no conflict of interest.

References

1. Walker, A. Breast milk as the gold standard for protective nutrients. J. Pediatr. 2010, 156
(2 Suppl.), 3–7.
2. National Health and Medical Research Council. Dietary Guidelines for Children and Adolescents
in Australia Incorporating the Infant Feeding Guidelines for Health Workers; National Health
and Medical Research Council, Ed.; NHMRC: Canberra, Australia, 2003; pp. 1–450.
3. World Health Organization. Global Strategy on Infant and Young Child Feeding. 55th World
Health Assembly; World Health Organization: Geneva, Switzerland, 2002.
4. National Health and Medical Research Council. Infant Feeding Guidelines; National Health and
Medical Research Council: Canberra, Australia, 2012.
5. UNICEF; WHO; UNESCO; UNFPA; UNDP; UNAIDS; WFP; World Bank. Nutrition and
Growth. In Facts for Life, 4th ed.; United Nations Children’s Fund: New York, NY, USA, 2010;
pp. 61–75.
6. World Health Organization. Indicators for Assessing Infant and Young Child Feeding Practices:
Part 1, Definitions; World Health Organization: Geneva, Switzerland, 2008; pp. 1–20.
7. Tarrant, R.C.; Younger, K.M.; Sheridan-Pereira, M.; White, M.J.; Kearney, J.M. Factors
associated with weaning practices in term infants: A prospective observational study in Ireland.
Br. J. Nutr. 2010, 104, 1544–1554.
8. Van Duong, D.; Binns, C.W.; Lee, A.H. Introduction of complementary food to infants within the
first six months postpartum in rural vietnam. Acta Paediatr. 2005, 94, 1714–1720.
9. Allcutt, C.; Sweeney, M.R. An exploration of knowledge, attitudes and advice given by health
professionals to parents in Ireland about the introduction of solid foods. A pilot study. BMC
Public Health 2010, 10, 201.
10. Kramer, M.S.; Kakuma, R. Optimal Duration of Exclusive Breastfeeding (Edited 2009).
Cochrane Database Syst. Rev. 2009, 1, CD003517.
11. Wright, C.M.; Parkinson, K.N.; Drewett, R.F. Why are babies weaned early? Data from
a prospective population based cohort study. Arch. Dis. Child 2004, 89, 813–816.
12. Symon, B.; Bammann, M. Feeding in the first year of life—Emerging benefits of introducing
complementary solids from 4 months. Aust. Fam. Phys. 2012, 41, 226–229.
13. Huh, S.Y.; Rifas-Shiman, S.L.; Taveras, E.M.; Oken, E.; Gillman, M.W. Timing of solid food
introduction and risk of obesity in preschool-aged children. Pediatrics 2011, 127, e544–e551.
14. Martin, M.; Sela, D. Infant Gut Microbiota: Developmental Influences and Health Outcomes.
In Building Babies; Clancy, K.B.H., Hinde, K., Rutherford, J.N., Eds.; Springer: New York, NY,
USA, 2013; Volume 37, pp. 233–256.
 40

15. Thompson, A.L. Developmental origins of obesity: Early feeding environments, infant growth,
and the intestinal microbiome. Am. J. Hum. Biol. 2012, 24, 350–360.
16. Kau, A.L.; Ahern, P.P.; Griffin, N.W.; Goodman, A.L.; Gordon, J.I. Human nutrition, the gut
microbiome and the immune system. Nature 2011, 474, 327–336.
17. Christofides, A.; Schauer, C.; Zlotkin, S.H. Iron deficiency anemia among children: Addressing a
global public health problem within a canadian context. Paediatr. Child Health 2005, 10, 597–601.
18. Hokama, T.; Gushiken, M.; Nosoko, N. Iron deficiency anaemia and child development. Asia
Pac. J. Public Health 2005, 17, 19–21.
19. Northstone, K.; Emmett, P.; Nethersole, F. The effect of age of introduction to lumpy solids
on foods eaten and reported feeding difficulties at 6 and 15 months. J. Hum. Nutr. Diet. 2001, 14,
43–54.
20. World Health Organization; United Nations Children’s Fund. Global Strategy for Infant and
Young Child Feeding; World Health Organization: Geneva, Switzerland, 2003; pp. 1–30.
21. ESPGHAN Committee on Nutrition; Agostini, C.; Decsi, T.; Fewtrell, M.; Goulet, O.; Kolacek, S.;
Koletzko, B.; Fleicher Michaelson, K.; Moreno, L.; Puntis, J.; et al. Complementary feeding: A
commentary by the espghan committee on nutrition. J. Pediatr. Gastroenterol. Nutr. 2008, 46,
99–110.
22. American Academy of Pediatrics Committee on Nutrition. Pediatric Nutrition Handbook, 6th ed.;
AAP: Chiago, IL, USA, 2008.
23. Ministry of Health of People’s Republic of China. Report on the Nutrition Development of
0–6 Years Children in China (in Chinese); Ministry of Health of People’s Republic of China:
Beijing, China, 2012.
24. Binns, C.W.; Gilchrist, D.; Woods, B.; Gracey, M.; Scott, J.; Smith, H.; Zhang, M.; Roberman, B.
Breastfeeding by aboriginal mothers in perth. Nutr. Diet. 2006, 63, 8–14.
25. Inoue, M. Breastfeeding and Perceptions of Breast Shape Changes in Australian and Japanese
Women. Ph.D. Thesis, Curtin University, Perth, Australia, 2012.
26. Li, L.; Zhang, M.; Scott, J.A.; Binns, C.W. Infant feeding practices in home countries and
australia: Perth chinese mothers survey. Nutr. Diet. 2005, 62, 82–88.
27. Scott, J.A.; Binns, C.W.; Oddy, W.H.; Graham, K.I. Predictors of breastfeeding duration:
Evidence from a cohort study. Pediatrics 2006, 117, e646–e655.
28. Xu, F.; Binns, C.; Nazi, G.; Shi, L.; Zhao, Y.; Lee, A. A comparison of breastfeeding among han,
uygur and other ethnic groups in Xinjiang, pr China. BMC Public Health 2006, 6, 196.
29. Duong, D.V. Factors that Influence the Utilization of Maternity Services and Breastfeeding
Practices in Rural Viet Nam. Ph.D. Thesis, Curtin University, Perth, Australia, 2005.
30. Gilchrist, D.; Woods, B.; Binns, C.W.; Scott, J.A.; Gracey, M.; Smith, H. Aboriginal mothers,
breastfeeding and smoking. Aust. N. Z. J. Public Health 2004, 28, 225–228.
31. Scott, J.A.; Landers, M.C.; Hughes, R.M.; Binns, C.W. Factors associated with breastfeeding at
discharge and duration of breastfeeding. J. Paediatr. Child Health 2001, 37, 254–261.
32. Binns, C.W.; Gilchrist, D.; Gracey, M.; Zhang, M.; Scott, J.; Lee, A. Factors associated with the
initiation of breast-feeding by aboriginal mothers in Perth. Public Health Nutr. 2004, 7, 857–861.
33. Abdulraheem, R. The Infant Feeding Experiences of Mothers in Maldives. Master’s Thesis,
Curtin University of Technology, Perth, Australia, 2004.
 41

34. Qiu, L.; Xie, X.; Lee, A.; Binns, C.W. Infants’ first feeds in hangzhou, pr china. Asia Pac. J.
Clin. Nutr. 2007, 16, 458–461.
35. Sakashita, R.; Inoue, N.; Kamegai, T. From milk to solids: A reference standard for the transitional
eating process in infants and preschool children in japan. Eur. J. Clin. Nutr. 2004, 58, 643–653.
36. Dewey, K.G.; Brown, K.H. Update on technical issues concerning complementary feeding of
young children in developing countries and implications for intervention programs. Food Nutr.
Bull. 2003, 24, 5–28.
37. Ministry of Health Labour and Welfare. The Guidelines for Support of Breastfeeding and
Weaning (in Japanese); Ministry of Health Labour and Welfare Health Statistics, Ed.; Ministry of
Health Labour and Welfare: Tokyo, Japan, 2007; pp. 1–79.
38. Food Standard in Australia and New Zealand Nuttab 2010 Online Searchable Database.
Available online: http://www.foodstandards.gov.au/consumerinformation/nuttab2010/nuttab2010
onlinesearchabledatabase/onlineversion.cfm (accessed on 13 December 2013).
39. World Health Organization. Complementary Feeding: Report of the Global Consultation, and
Summary of Guiding Principles for Complementary Feeding of the Breastfed Child; World
Health Organization: Geneva, Switzerland, 2002.
40. Saarinen, U.M.; Siimes, M.A.; Dallman, P.R. Iron-absorption in infants—High bioavailability of
breast-milk iron as indicated by extrinsic tag method of iron-absorption and by concentration of
serum ferritin. J. Pediatr. 1977, 91, 36–39.
41. Andersson, M.; Hurrell, R.F. Prevention of iron deficiency in infancy, childhood and adolescence.
Ann. Nestlé 2010, 68, 120–131.
42. World Health Organization, Food and Agricultural Organization of the United Nations. Vitamin
and Mineral Requirements in Human Nutrition; World Health Organization: Geneva,
Switzerland, 2003; pp. 1–341.
43. Barennes, H.; Simmala, C.; Odermatt, P.; Thaybouavone, T.; Vallee, J.; Martinez-Ussel, B.;
Newton, P.N.; Strobel, M. Postpartum traditions and nutrition practices among urban lao women
and their infants in vientiane, lao pdr. Eur. J. Clin. Nutr. 2007, 63, 323–331.
44. Bhutta, Z.A.; Ahmed, T.; Black, R.E.; Cousens, S.; Dewey, K.; Giugliani, E.; Haider, B.A.;
Kirkwood, B.; Morris, S.S.; Sachdev, H.P.S.; et al. Maternal and child undernutrition 3—What
works? Interventions for maternal and child undernutrition and survival. Lancet 2008, 371, 417–440.
45. Golden, M.H. Proposed Recommended Nutrient Densities for Moderately Malnourished Children.
In Food and Nutrition Bulletin; Rosenberg, I., Ed.; United Nations University Press: Boston, MA,
USA, 2009; Volume 30, pp. S267–S342.
46. UNICEF. Information by Country: Maldives, Statistics. Available onine: http://www.unicef.org/
infobycountry/maldives_maldives_statistics.html (accessed on 18 January 2013).
47. Abdulraheem, R.; Binns, C.; Chih, H.J. First Feed Given to Infants in the Maldives. In
Proceedings of the 44th Asia Pacific Academic Consortium for Public Health (APACH),
Colombo, Sri Lanka, 14–17 October 2012; APACPH: Colombo, Sri Lanka, 2012.

© 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).
42
 43

2. Breastfeeding

Reprinted from Nutrients. Cite as: Ziegler, E.E.; Nelson, S.E.; Jeter, J.M. Iron Stores of Breastfed
Infants during the First Year of Life. Nutrients 2014, 6, 2023-2034.

Iron Stores of Breastfed Infants during the First Year of Life

Ekhard E. Ziegler *, Steven E. Nelson and Janice M. Jeter

Department of Pediatrics, University of Iowa, A136 MTF, 2501 Crosspark Rd., Coralville,
IA 52241-8802, USA; E-Mails: steven-nelson@uiowa.edu (S.E.N.); janice-jeter@uiowa.edu (J.M.J.)

* Author to whom correspondence should be addressed; E-Mail: ekhard-ziegler@uiowa.edu;

Tel.: +1-319-335-4570, Fax: +1-319-335-4856.

Received: 4 March 2014; in revised form: 21 April 2014 / Accepted: 9 May 2014 /
Published: 21 May 2014

Abstract: The birth iron endowment provides iron for growth in the first months of life.
We describe the iron endowment under conditions of low dietary iron supply. Subjects
were infants participating in a trial of Vitamin D supplementation from 1 to 9 months.
Infants were exclusively breastfed at enrollment but could receive complementary foods
from 4 months but not formula. Plasma ferritin (PF) and transferrin receptor (TfR) were
determined at 1, 2, 4, 5.5, 7.5, 9 and 12 months. At 1 month PF ranged from 38 to 752 μg/L
and was only weakly related to maternal PF. PF declined subsequently and flattened out at
5.5 months. PF of females was significantly higher than PF of males except at 12 months.
TfR increased with age and was inversely correlated with PF. PF and TfR tracked strongly
until 9 months. Iron deficiency (PF < 10 μg/L) began to appear at 4 months and increased
in frequency until 9 months. Infants with ID were born with low iron endowment. We
concluded that the birth iron endowment is highly variable in size and a small endowment
places infants at risk of iron deficiency before 6 months. Boys have smaller iron
endowments and are at greater risk of iron deficiency than girls.

Keywords: iron endowment; breastfed infant; iron stores; iron deficiency

1. Introduction

At birth, the body iron content of the infant is high (94 mg/kg fat-free mass) [1] due to a high
hemoglobin mass and a sizable amount of storage iron [2]. This birth iron endowment renders the
 44

infant independent of exogenous iron during the early months of life. Immediately after birth, the
hemoglobin mass begins to shrinks and its iron is transferred to the storage compartment [3]. The latter
therefore represents the entire birth iron endowment. Plasma ferritin (PF) concentration is proportional
to storage iron and thus provides a measure to follow the fate of the iron endowment. At birth, the size
of the endowment varies greatly, as we [4–6] and others [7–18] have shown. Although severe maternal
iron deficiency [11] and certain pregnancy complications [19] are known to reduce the size of the iron
endowment, under most circumstances iron status of the mother is not a determinant of the size of the
iron endowment. The cause(s) of the variation in its size remain largely unknown.
The iron endowment provides iron for growth and protects the breastfed infant against iron
deficiency in the first 4–6 months of life. It follows that the size of the endowment should determine
the degree of protection afforded the infant. Indeed, there is evidence that diminished size of the storage
iron compartment shortens the protection and places the infant at risk of iron deficiency [4–6,20].
Essentially, all breastfed infants in our studies [4–6] who developed iron deficiency by 6 months were
born with diminished iron endowment.
As the iron endowment is used up for growth, PF concentration declines, but the rate of decline is
modified by exogenous iron [4–6]. Few data exist regarding the unmodified iron endowment. Such
data may be valuable for examining longitudinal tracking of iron stores as well as defining gender-related
differences of iron stores. Only one cohort of infants in our earlier studies (control group in [5]) came
close to having an unmodified iron endowment, but as even these infants could receive supplemental
formula, their iron endowment could have been modified to a significant extent. Therefore, when in a
recently completed study [21] it was deemed necessary to prohibit the consumption of supplemental
formula until 9 months of age, the opportunity to evaluate the size of the iron endowment minimally
modified by dietary iron presented itself. The present report concerns data on the iron endowment of
these infants who, besides breast milk, received no source of iron until 4 months and thereafter
received iron only from complementary foods until 9 months. For this large group of breastfed infants,
iron status of the mothers soon after birth was known and detailed dietary information was available.
We define iron deficiency (ID) as a state of exhausted iron stores indicated by PF < 10 μg/L.
When there is, in addition, evidence of impaired hemoglobin synthesis, iron deficiency is considered
to be severe. Iron-deficiency anemia (IDA) carries a significant risk of impaired neurocognitive
development [22,23].

2. Experimental Section

Clinical Trial Registration: Registered at clinicaltrials.gov NCT00494104.

Briefly, the parent study [21] was a randomized double-blind trial of breastfed infants who received
one of four doses of Vitamin D supplementation (200 IU/day, 400 IU/day, 600 IU/day and 800 IU/day)
from 1 to 9 months. Plasma 25(OH)D concentration was the primary endpoint. In order to minimize
dietary Vitamin D intake, parents were asked not to feed supplemental formula.
Study design and intervals: The study was a prospective, randomized, double-blind study. Infants
were enrolled and randomized at 1 month (=within 4 days of 28 days). They visited the study center
every 28 days until 9 months (280 ± 4 days) and made a final visit at 12 months (364 ± 4 days). Infants
received the study Vitamin D drops from 1 to 9 months. The study protocol was reviewed and
 45

approved by the University of Iowa Institutional Review Board and parents provided written consent.
The trial was registered with ClinicalTrials.gov under NCT00494104.
Subjects were full-term infants considered normal by their parents and their physicians. Infants
were exclusively breastfed at enrollment. Starting at 4 months, they could receive complementary
foods but no formula until 9 months. Vitamin or mineral (iron) supplements were not permitted.
Infants were born between August 2006 and September 2010. Enrollment was limited to infants born
between June and November so they would be between 5.5 and 9 months old at the end of winter
(March to mid-May) when the primary assessment of Vitamin D status took place. Parents were asked
not to give any iron or Vitamin D supplements.
Procedures: Infants visited the study center every 28 days and had weight and length measured.
Parents completed an interim health and feeding questionnaire. Code-labeled Vitamin D supplements
were dispensed during visits and empty and half-empty containers were collected and weighed. Infants
had blood drawn at 1, 4, 5.5, 7.5, 9 and 12 months by heel prick using a disposable spring-loaded
device (Tenderfoot, International Technidyne) into heparin-coated tubes. Plasma ferritin (PF) was
determined using an immunoradiometric procedure (Ramco catalog no. F-11) with interassay coefficient
of variation of 6.5%. Soluble transferrin receptor (TfR) was measured by enzyme immuno assay
(Ramco catalog no TF-94).
Data analysis: Iron deficiency was defined as plasma ferritin <10 μg/L [4–6] and anemia as
hemoglobin <105 g/L before 9 months and <100 g/L at 9 and 12 months [24]. Body iron was
calculated as Body iron (mg/kg) =í ORJ 7I53)  í 2.8229)/0.1207, where TfR and PF are both
in μg/L [25]. Gender-related differences of PF, TfR and body iron assessed by t-tests and ANOVA
procedures. Tracking was determined by linear correlations between successive values. Associations of
PF and TfR were determined cross-sectionally on an age-specific basis by linear correlation analysis.
Percentiles of PF were determined by SAS univariate procedure. Statistical analyses were performed
using SAS version 9.1.3 (SAS Institute, Cary, NC, USA).

3. Results

Of 213 infants enrolled at 1 month and assigned to one of the Vitamin D supplement doses,
128 completed the intervention at 9 months and 120 were evaluated at 12 months. The main reason
why infants left the study was the parents’ desire to use supplemental formula. At enrollment,
177 mothers donated a venous blood sample. Table 1 summarizes feeding data as reported by the
parents and shows that parents adhered to feeding rules to a remarkable degree. Only one infant
received cereal at 2 and 3 months and 2 infants received formula at 3 months, but the amounts were in
each case small (”1 feeding/week) and infants continued in the study. Many infants did not receive
complementary foods until late. For example, at 5.5 months 63 infants (of 153) received no
complementary foods at all. Also, cereals, a rich source of iron, were received by only one-half of
infants at 5.5 months.
At 1 month, PF averaged 242 μg/L with a range from 38 to 752 μg/L (Figure 1, Table 2). Maternal
PF obtained at the same time averaged 42 μg/L (SD 34 μg/L) (Figure 1). It is evident that a fair
proportion of mothers were in less than optimal iron nutritional status, with 12 mothers having
PF < 10 μg/L. However, iron status of the mother was not an important factor determining infant iron
 46

stores. As illustrated in Figure 2, the relationship between maternal and infant PF levels was weak
(r = 0.081, p = 0.283). Maternal iron status explained only 6.4% of the variation in infant iron
endowment. As Figure 2 shows, when the mother’s iron stores were low (PF < 20 μg/L, her infant’s
PF could still range from 40 to 680 μg/L.

Table 1. Feedings as reported by parents. The table indicates number of infants receiving
the specified food during the month preceding the visit.
Age Total Any Table Cow
Cereal Fruits Vegetables Meats Formula
(month) subjects breast foods milk
1 213 213 0 0 0 0 0 0 0
2 194 194 1 0 0 0 0 0 0
3 181 181 1 0 0 0 0 2 0
4 165 165 4 2 0 0 0 1 0
5.5 153 152 76 29 40 0 5 1 0
7.5 138 138 79 92 97 12 20 4 1
9 128 124 86 96 99 30 46 9 1
12 120 92 43 70 69 44 92 23 43

Figure 1. Plasma ferritin concentrations of mothers and infants one month after birth.
 47

Table 2. Plasma concentrations of ferritin (PF) and transferrin receptor (TfR). Values are mean ± SD unless otherwise indicated.
Age (month) 1 2 4 5.5 7.5 9 12
Number determinations 201 190 165 152 138 126 118
Plasma ferritin (μg/L)
All 242 ± 125 184 ± 103 88 ± 57 44 ± 29 40 ± 28 26 ± 17 22 ±18
Range 38–752 43–710 10–373 3–137 5–144 4–90 5–137
Female 256 ± 131 201 ± 106 98 ± 58 51 ± 30 44 ± 29 30 ± 18 23 ± 17
Male 227 ± 119 169 ± 99 80 ± 55 39 ± 27 36 ± 27 24 ± 16 22 ± 19
p M vs. F 0.105 0.032 0.040 0.015 0.098 0.045 0.715
Number < 10 μg/L (M/F) 0/0 0/0 1/0 6/2 10/1 12/3 9/6
Number < 12 μg/L (M/F) 0/0 0/0 3/0 9/3 14/2 16/7 15/9
Transferrin receptor (mg/L)
All 3.21 ± 0.65 4.49 ± 1.10 6.52 ± 1.12 6.66 ± 1.17 7.05 ± 1.19 7.12 ± 1.41 6.97 ± 1.14
Female 3.08 ± 0.62 4.23 ± 0.97 6.19 ± 0.95 6.28 ± 1.01 6.74 ± 1.07 6.91 ± 1.28 6.94 ± 0.94
Male 3.34 ± 0.66 4.71 ± 1.15 6.79 ± 1.19 6.97 ± 1.21 7.30 ± 1.23 7.23 ± 1.50 6.90 ± 1.29
p M vs. F 0.0037 0.0027 0.0006 0.0003 0.005 0.200 0.85
Correl. coeff. PF vs. TfR 0.026 ía ía ía ía í í
Body iron (mg/kg)
All 13.7 ± 1.91 11.4 ± 2.32 7.20 ± 2.67 4.60 ± 2.90 3.03 ± 2.91 2.67 ± 2.48 2.12 ± 2.32
Female 14.1 ± 1.70 12.1 ± 2.01 7.94 ± 2.25 5.45 ± 2.57 4.63 ± 2.57 3.27 ± 2.24 2.26 ± 2.18
Male 13.3 ± 2.02 10.9 ± 2.43 6.58 ± 2.84 3.91 ± 2.59 3.33 ± 3.06 2.22 ± 2.58 2.01 ± 2.44
p M vs. F 0.0031 0.0004 0.0011 0.0010 0.0091 0.0188 0.579
a
correlation statistically significant (p < 0.05).
 48

Figure 2. Relationship between maternal and infant plasma ferritin one month after birth
(r = 0.081, p = 0.283).

As shown in Figure 3 and Table 2, infant PF decreased rapidly with age but leveled off after
5.5 months, indicating exhaustion of the iron endowment. In some infants, PF increased between 1 and
4 months, presumably indicating continuing recycling of iron from hemoglobin breakdown. Transient
increases of PF at other ages indicate acute phase reactions. From 2 to 5.5 months, average PF declined
by 1.1 (SD 0.40) % each day. The decline was strongly inversely correlated with gain in weight and
length (p ”0.0001). This is consistent with the notion that growth is the main cause of the decline of
PF. As PF decreased, the range of values narrowed progressively. This is illustrated by the percentile
values shown in Figure 4.

Figure 3. PF of individual infants from 1 to 12 months.

 49

Figure 4. PF of males and females. Differences were statistically significant except at

1 and at 12 months.

Concentrations of TfR increased with age until 7.5 months and then leveled off, mirror- imaging PF
values (Table 2). TfR was significantly inversely correlated with PF at most ages. Body iron was
highest at 1 month and declined progressively thereafter, in essence paralleling the course of PF.
Tracking: In spite of the marked decrease of PF, there was a strong tendency for infants to preserve
their rank. PF values correlated (tracked) strongly over time (Table 3). Although tracking is to be
expected given the wide range of PF values at 1 month, tracking continued at 9 and 12 months, which
suggests that other factors, including genetic, may be operating. TfR also showed tracking which was
less strong than that of PF but was still quite strong (Table 3).

Table 3. Pearson correlations among PF (a) values and TfR (b) values at different ages.
(a) PF
9
Age 4 months 5.5 months 7.5 months 12 months
months
1 month 0.670 a 0.638 a 0.496 a 0.546 a 0.465 a
4 months - 0.738 a 0.680 a 0.700 a 0.438 a
5.5 months - - 0.751 a 0.747 a 0.511 a
7.5 months - - - 0.804 a 0.515 a
9 months - - - - 0.579 a
(b) TfR
9
Age 4 months 5.5 months 7.5 months 12 months
months
1 month 0.321 a 0.357 a 0.319 a 0.297 a 0.146
4 months - 0.616 a 0.613 a 0.423 a 0.321 a
5.5 months - - 0.616 a 0.543 a 0.330 a
7.5 months - - - 0.605 a 0.515 a
9 months - - - - 0.525 a
a
correlation statistically significant (p < 0.05).
 50

Gender: Plasma ferritin showed marked gender-related differences as indicated in Table 2 and
Figure 5, with levels of female infants being significantly higher than levels of male infants at most
ages. Accordingly, female infants became iron deficient (PF < 10 μg/L) less frequently than male
infants (Table 2). Gender-related differences were also present in TfR but, contrary to PF, males had
higher levels, with differences being statistically significant except at 9 and 12 months. Body iron was
significantly higher in females than males except for 12 months.

Figure 5. Percentile values for PF from 1 to 12 months (males and females combined).

Infants who developed ID: PF values less than 10 μg/L indicating exhausted iron stores were
observed at almost all ages. The earliest age at which PF ”  —J/ RFFXUUHG ZDV  PRQWKV LQ RQH
infant. At 5.5 months, the PF of an additional seven infants dropped below 10 μg/L, meaning that eight
(5.3%) infants had exhausted their Fe stores before 6 months. At 28 days, the mean PF of these eight
infants was 125 ± 103 μg/L, which was significantly (p < 0.001) lower than the PF of all other infants
at 28 days (mean 245 ± 119 μg/L). At 7.5 months, an additional nine infants developed a PF
of ” —J/ PHDQLQJ WKDW    LQIDQWV KDG H[KDXVWHG WKHLU LURQ VWRUHV $IWHU
7.5 months, with the birth iron endowment exhausted, a PF ” —J/ ZDVUHIOHFWLYHRIORZGLHWDU\
iron intake. This occurred in 15 infants at 9 months and 15 infants at 12 months. A total of 36 infants
(19% of those randomized at 4 months) developed a PF of”—J/ DW OHDVWRQFHGXULQJWKHVWXG\
PF values < 12 μg/L were observed on 78 occasions (Table 2). Among infants with”PF  —J/
at 56 days of age (N = 12), one developed ID at 4 months and four more by 5.5 months. Thus, among
infants with a PF less than 65 μg/L at 2 months, 42% developed ID before 6 months.
When infants developed a PF of ”—J/WKHSDUHQWVZHUHLQIRUPHGDQGWKHVXJJHVWLRQZDVPDGH
by the investigators to increase the consumption of iron-containing foods. One such infant was placed
 51

on ferrous sulfate by his pediatrician and subsequent PF values were >10 μg/L. The parents of three
infants chose to withdraw from the study. In all remaining infants with ID, PF was monitored closely
and in seven infants the next PF was >10 μg/L. In cases where it remained <10 μg/L (N = 5)
hemoglobin was determined to check for IDA. Hemoglobin was in each case >105 g/100 mL. With
IDA ruled out, no iron treatment was recommended. The parents were always informed of findings and
feeding of iron-rich foods was recommended whenever PF was ”—J/

4. Discussion

We believe our cohort to be the largest group of breastfed infants in whom the birth iron
endowment was assessed while the intake of iron from exogenous sources was minimal. Thus,
the iron endowment was observed in its undisturbed state while not being modified by dietary iron.
The size of the iron endowment was assessed by determination of plasma ferritin concentration (PF).
At the earliest assessment at one and two months, the iron endowment showed enormous variation.
This wide range in the size of the iron endowment has been previously shown by us [4–6] and
others [7–18]. Only a small proportion of that variation was explained by iron status of the mother
whereas its bulk remained unexplained. Earlier studies found for the most part no association or only a
weak association between maternal iron status and infant iron status at birth [7,9,10,16–18]. Some
studies did find a reduction of infant iron status when maternal iron status was very poor [8,11,12].
Except for some maternal conditions that are known to reduce the size of the iron endowment [19], the
cause of the variation is unknown.
Because infants of the present cohort were not permitted to receive supplemental formula, a source
of dietary iron, we expected the iron endowment to be somewhat smaller than in the control infants of
our earlier study [5] who received complementary foods like the present cohort but did also receive
supplementary formula. However, PF values were nearly identical in that earlier group [5] and in
the present cohort. For example, at 5.5 months mean PF in the earlier group was 42 ± 29 μg/L and was
44 ± 29 μg/L in the present study. The explanation may be that infants who did receive formula
received only modest amounts, and many infants probably received no formula at all.
While the cause of the variation in the size of the iron endowment remains obscure, its consequences
are readily apparent. The small size of the iron endowment means that it becomes exhausted early and
places the infant at risk of ID. We [5,6] and others [19] have before noted this association between size
of the iron endowment and risk of ID early in life. In the present cohort a full 5.3% of infants
developed ID by 6 months. The incidence is comparable to the incidence observed earlier by us [5,6].
Others have reported somewhat lower [26] and also higher [27] incidences of ID by 6 months. The
present estimate of the incidence of ID in breastfed infants who receive only modest amounts of
dietary iron is solid thanks to the design of the study. On the other hand, the present study provides no
estimates of the incidence of IDA. The reasons are that hemoglobin was not determined with all PF
determinations but was determined only when PF remained low in spite of the parents being requested
to increase the amount of iron-containing foods. Without that intervention, some infants with ID might
have gone on to develop IDA on subsequent assessment.
The present study shows marked gender-related differences in the size of the iron endowment, with
girls having significantly greater size than boys. There was also a gender-related difference in TfR
 52

except in the opposite direction, with boys having higher values than girls. Similar gender-related
differences have been reported by Hay and colleagues [16,26], Domellöf et al. [28] and were seen in
our earlier studies [5,6]. PF and TfR were inversely correlated at most ages. Our data clearly show that
there are gender-related differences in the size of the iron endowment. This means that whatever is
causing gender-related differences at later ages is already operating in utero.
We showed strong tracking of PF and TfR that decreases in strength somewhat with increasing age
but is still strong at 12 months for PF. Tracking of PF has been reported before by Hay et al. [16] and
by us [5,6]. Tracking, together with gender-related differences, provides a strong suggestion that the
size of the iron endowment and iron status in general are controlled by genetic factors.
The strengths of the present study are that it involved a large cohort of breastfed infants in whom
the intake of dietary iron was kept to a practical minimum, permitting study of the natural course of the
birth iron endowment without modification by dietary iron. Another strength was that all observations
were made in longitudinal fashion and that detailed dietary information was recorded.

5. Conclusions

It was concluded that the birth iron endowment is highly variable in size and differs significantly
between males and females. Its size decreases as iron is used for growth. Some infants develop ID
by 6 months of age. A small iron endowment places infants at increased risk of iron deficiency.
The fact that PF and TfR track strongly during infancy suggests the operation of genetic factors.

Acknowledgments

This study was supported by the National Institutes of Health grant HD048870.
All laboratory determinations were performed by Joyce Guese, CLA, which is gratefully acknowledged.

Author Contributions

EEZ and SEN designed the study, analyzed the data and wrote the manuscript. JMJ had
responsibility for carrying out the study. All authors read the manuscript and agree with it.

Conflicts of Interest

The authors have no conflicts of interest to disclose.

Financial Disclosure

The authors have no financial relationships relevant to this article to disclose.

References

1. Widdowson, E.M.; Spray, C.M. Chemical development in utero. Arch. Dis. Child. 1951, 26,
205–214.
2. Dallman, P.R.; Siimes, M.A.; Stekel, A. Iron deficiency in infancy and childhood. Am. J.
Clin. Nutr. 1980, 33, 86–118.
 53

3. Siimes, M.A.; Addiego, J.E.; Dallman, P.R. Ferritin in serum: Diagnosis of iron deficiency and
iron overload in infants and children. Blood 1974, 43, 581–590.
4. Ziegler, E.E.; Nelson, S.E.; Jeter, J.M. Iron supplementation of breastfed infants from an early
age. Am. J. Clin. Nutr. 2009, 89, 525–532.
5. Ziegler, E.E.; Nelson, S.E.; Jeter, J.M. Iron status of breastfed infants is improved equally by
medicinal iron and iron-fortified cereal. Am. J. Clin. Nutr. 2009, 90, 76–87.
6. Ziegler, E.E.; Fomon, S.J.; Nelson, S.E.; Jeter, J.M.; Theuer, R.C. Dry cereals fortified with
electrolytic iron or ferrous fumarate are equally effective in breastfed infants. J. Nutr. 2011, 141,
243–248.
7. Rios, E.; Lipschitz, D.A.; Cook, J.D.; Smith, N.J. Relationship of maternal and infant stores as
assessed by determination of plasma ferritin. Pediatrics 1975, 55, 634–699.
8. Milman, N.; Ibsen, K.K.; Christensen, J.M. Serum ferritin and iron status in mothers and newborn
infants. Acta Obstet. Gynecol. Scand. 1987, 66, 205–211.
9. Wong, C.-T.; Saha, N. Inter-relationships of storage iron in the mother, the placenta and the
newborn. Acta Obstet. Gynecol. Scand. 1990, 69, 613–616.
10. Poulakka, J.; Jänne, O.; Vihko, R. Evaluation by serum ferritin assay of the influence of maternal
iron stores on the iron status of newborns and infants. Acta Obstet. Gynecol. Scand. Suppl. 1980,
95, 53–56.
11. Singla, P.N.; Tyagi, M.; Shankar, R.; Dash, D.; Kumar, A. Fetal iron status in maternal anemia.
Acta Peadiatr. 1996, 85, 1327–1330.
12. Sweet, D.G.; Savage, G.; Tubman, T.R.J.; Lappin, T.R.J.; Halliday, H.L. Study of maternal
influences on fetal iron status at term using cord blood tranferrin receptors. Arch. Dis. Child. Fetal
Neonatal Ed. 2001, 84, F40–F43.
13. Tamura, T.; Goldenberg, R.L.; Hou, J.; Johnston, K.E.; Cliver, S.P.; Ramey, S.L.; Nelson, K.G.
Cord serum ferritin concentrations and mental psychomotor development of children at five years
of age. J. Pediatr. 2002, 140, 165–170.
14. Lott, D.G.; Zimmerman, M.B.; Labbé, R.F.; Kling, P.J.; Widness, J.A. Eythrocyte zinc
protoporphyrin is elevated with prematurity and fetal hypoxemia. Pediatrics 2005, 116, 414–422.
15. Siddappa, A.M.; Rao, R.; Long, J.D.; Widness, J.A.; Gerogieff, A.M. The assessment of newborn
iron stores at birth: A review of the literature and standards for ferritin concentration. Neonatology
2007, 92, 73–82.
16. Hay, G.; Refsum, H.; Whitelaw, A.; Melbye, E.L.; Haug, E.; Borch-Iohnsen, B. Predictors of
serum ferritin and serum soluble transferrin receptor in newborns and their associations with iron
status during the first 2 y of life. Am. J. Clin. Nutr. 2007, 86, 64–73.
17. Ervasti, M.; Sankilampi, U.; Heinonen, S.; Punnonen, K. Early signs of maternal iron deficiency
do not influence the iron status of the newborn, but are associated with higher infant birthweight.
Acta Obstet. Gynecol. 2009, 88, 83–90.
18. Shao, J.; Lou, J.; Rao, R.; Georgieff, M.K.; Kaciroti, N.; Felt, B.T.; Zhao, Z.-Y.; Lozoff, B.
Maternal serum ferritin concentration is positively associated with newborn iron stores in women
with low ferritin status in late pregnancy. J. Nutr. 2012, 142, 2004–2009.
 54

19. Chockalingam, U.M.; Murphy, E.; Ophoven, J.C.; Weisdorf, S.A.; Georgieff, M.K. Cord
transferring and ferritin values in newborn infants ate risk for perinatal uteroplacental
insufficiency and chronic hypoxia. J. Pediatr. 1987, 111, 283–286.
20. Georgieff, M.K.; Wewerka, S.W.; Nelson, C.A.; de Regnier, R.-A. Iron status at 9 months of
infants with low iron stores at birth. J. Pediatr. 2002, 141, 405–409.
21. Ziegler, E.E.; Nelson, S.E.; Jeter, J.M. Vitamin D supplementation of breastfed infants: A
randomized dose-response trial. Pediatr. Res. 2014, in press.
22. Lozoff, B.; de Andraca, I.; Castillo, M.; Smith, J.B.; Walter, T.; Pino, P. Behavioral and
developmental effects of preventing iron-deficiency anemia in healthy full-term infants.
Pediatrics 2003, 112, 846–854.
23. Lozoff, B.; Beard, J.; Connor, J.; Felt, B.; Georgieff, M.; Schallert, T. Long-lasting neural and
behavioral effects of iron deficiency in infancy. Nutr. Rev. 2006, 64, S34–S43.
24. Domellöf, M.; Dewey, K.G.; Lönnerdal, B.; Cohen, R.J. The diagnostic criteria for iron
deficiency in infants should be reevaluated. J. Nutr. 2002, 132, 3680–3686.
25. Cook, J.D.; Flowers, C.H.; Skikne, B.S. The quantitative assessment of body iron. Blood 2003,
101, 3359–3364.
26. Hay, G.; Sandstad, B.; Whitelaw, A.; Borch-Iohnsen, B. Iron status in a group of Norwegian
children aged 6–24 months. Acta Paediatr. 2004, 93, 592–598.
27. Lind, T.; Lönnerdal, B.; Persson, L.-A.; Stenlund, H.; Tennefors, C.; Hernell, O. Effects of
weaning cereals with different phytate contents on hemoglobin, iron stores, and serum zinc: A
randomized intervention in infants from 6 to 12 months of age. Am. J. Clin. Nutr. 2003, 78,
168–175.
28. Domellöf, M.; Lönnerdal, B.; Dewey, K.G.; Cohen, R.J.; Rivera, L.L.; Hernell, O. Sex differences
in iron status during infancy. Pediatrics 2002, 110, 545–552.

© 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).
 55

Reprinted from Nutrients. Cite as: Castro, P.D.; Layte, R.; Kearney, J. Ethnic Variation in
Breastfeeding and Complimentary Feeding in the Republic of Ireland. Nutrients 2014, 6, 1832-1849.

Ethnic Variation in Breastfeeding and Complimentary Feeding

in the Republic of Ireland
Patricia Dominguez Castro 1,*, Richard Layte 2 and John Kearney 3
1
Trinity College Dublin, College Green, Dublin 2, Ireland
2
Economic and Social Research Institute, Sir John Rogerson’s Quay, Dublin 2, Ireland;
E-Mail: Richard.Layte@esri.ie
3
Dublin Institute of Technology, Kevin Street, Dublin 8, Ireland; E-Mail: John.Kearney@dit.ie

* Author to whom correspondence should be addressed; E-Mail: domingup@tcd.ie;

Tel.: +353-86-818-2276.

Received: 14 February 2014; in revised form: 3 April 2014 / Accepted: 18 April 2014 /
Published: 2 May 2014

Abstract: Early nutrition plays a pivotal role in long-term health. The World Health
Organization (WHO) recommends exclusive breastfeeding during the first six months of
life, with the gradual introduction of solids after this period. However, studies in the
Republic of Ireland (ROI) have shown poor compliance with guidelines. The ROI
continues to have one of the lowest breastfeeding rates worldwide. Our objective was to
analyse differences in breastfeeding and complimentary feeding behaviours between Irish
and non-Irish mothers residing in the ROI, as well as the role of acculturation on these
behaviours, using the national longitudinal study, Growing Up in Ireland (GUI). Mothers
(n = 11,134) residing in the ROI were interviewed when their infants were nine months of
age. The percentage of Irish mothers who initiated breastfeeding was 49.5%, as opposed to
88.1% among the non-Irish cohort (p < 0.001). Breastfeeding initiation reduced from
89.4% of non-Irish mothers who had arrived within the last year to five years ago to 67.5%
for those who had arrived 11 to >20 years ago (p < 0.001). Our results indicate that cultural
differences are an important factor in shaping patterns of infant feeding in the ROI.
Reviewing existing support and education policies for parents is required to achieve the
implementation of desirable infant feeding practices.

Keywords: infant feeding; breastfeeding; complimentary feeding; acculturation

 56

1. Introduction

Early nutrition plays a pivotal role in long-term health. Breastfeeding has been shown to have a
protective role in the development of being overweight, obesity and chronic diseases later in life [1–3].
The World Health Organization (WHO) recommends exclusive breastfeeding during the first
six months of life of the infant, with the gradual introduction of complimentary foods after this
period [4]. The European Society of Paediatric, Gastroenterology, Hepatology and Nutrition
(ESPGHAN) recommends not to introduce complimentary foods before 17 weeks and no later than
26 weeks, while also giving the advice to commence the introduction of solids near six months of
age [5]. Early complimentary feeding has been shown by studies to increase the risk of overweight and
obesity during childhood and adulthood [6–9]. Moreover, the transition from milk to solid foods can
have a life-long influence on dietary patterns [6,10–12]. The introduction of complimentary foods
cannot be studied in isolation from the type of milk feeding early in life, as milk type influences the
type of solid foods introduced and the timing of their introduction [13]. Studies on the predictors of
early complimentary feeding have shown that breastfeeding reduces the likelihood of early solid
food introduction [14–18].
In the Republic of Ireland (ROI), the Department of Health and Children updated their advice in
2003 to recommend adherence to the WHO advice of exclusive breastfeeding during the first six
months of the infants’ life [19]. The new Infant Feeding Guidelines released by the Food safety
Authority of Ireland (FSAI) in November, 2012, maintain the recommendation made by
ESPGHAN [20]. Despite these guidelines, the ROI continues to have one of the lowest breastfeeding
rates worldwide, and compliance with complimentary feeding guidelines is poor [21]. Irish studies
show rates of exclusive breastfeeding for six months of less than 1%, with 75% of infants being
introduced to complimentary feeding before 17 weeks, 22.6% of these being introduced prematurely
by 12 weeks [15,22].
Rates of breastfeeding in Ireland have increased since 2004, but they are still below national targets,
and a large percentage of this increase has been attributed to changes in maternal characteristics, such
as older age and an increase in non-national mothers [23–25]. Previous studies in the ROI have pointed
out different patterns of breastfeeding rates by maternal origin of birth [23,26,27]. However, we are not
aware of any studies in the ROI analysing different patterns in complimentary feeding introduction by
ethnic group. Given the fact that the time of complimentary feeding introduction seems to be linked to
the type of milk feeding early in life and given the low breastfeeding rates in the ROI, it could be
hypothesized that Irish born mothers are more likely to introduce complimentary foods early in the life
of their infant, thus increasing the risk of their infants suffering adverse health effects in the short and
long term. Moreover, acculturation of non-Irish mothers could play a role in their infant feeding
practices. The aim of this paper is to study variation in breastfeeding rates and the timing of the
introduction of complimentary feeding between Irish and non-Irish mothers living in the ROI, as well
as the role of acculturation on these behaviours using cross-sectional data from the national
longitudinal study of children in Ireland (Growing Up in Ireland).
 57

2. Methods

2.1. Study Design and Sample

Growing Up in Ireland (GUI) is a nationally representative cohort study of nine-month-old infants

residing in the Republic of Ireland. The main aim of the study is to study the factors affecting the lives
of infants in Ireland with the aim of creating evidence-based policy. The study sample consisted of
11,134 nine-month-old infants who participated in the first wave of the GUI study. These were
selected from the approximately 41,000 births over the period of 1 December 2007 to 30 June 2008.
The completed sample of 11,134 represents approximately one-third of all births in the ROI over the
field period. Families were invited to participate in the study when the child was nine months of age.
The sampling frame for the project was the Child Benefit Register for the Republic of Ireland.
Of 16,136 mothers selected from the sampling frame, 11,134 agreed to take part in the study, a
response rate of 69% [28].

2.2. Questionnaires and Measurements

Primary caregivers, defined as the person who spent more time with the child, and secondary
caregivers were interviewed at home and asked to complete a main questionnaire and a sensitive
questionnaire. Since only 0.5% of the primary care givers nominated were not the biological mothers,
we refer to responses from the primary care giver as those of the mother. Interviews were carried out
using a mixture of CAPI (computer-assisted personal interviewing) and CASI (computer-assisted
self-interviewing).
The wave one sample was selected from the Child Benefit Register for the Republic of Ireland,
which was provided by the Department of Social Protection. Of 16,136 mothers selected from the
sampling frame, 11,134 agreed to take part in the study, a response rate of 69%. Fieldwork was carried
out over 7 months, extending from September 2008, to the end of April 2009. Children were selected
so as to be 9-months-old at the time of the interview; consequently, eligible children were all those
born between 1 December 2007 and 30 June 2008 [28].
The sampling frame for the study was the Child Benefit Register for the Republic of Ireland. The
sample was selected on a systematic basis, pre-stratifying by marital status, county of residence,
nationality and number of children (where child is defined as <16 years of age) in the household, using
a random start and constant sampling fraction. The completed sample was statistically grossed or
reweighted on the basis of external population estimates to ensure that it was wholly representative of
all children aged one year or less in Ireland [28].
Interviewers measured and recorded both parents’ height and weight. A medically approved
mechanical SECA 761 weighing scale was used for the adults’ weight and a Leicester measuring stick
for their height. All stages of the Growing Up in Ireland project were subject to rigorous ethical review
by a Research Ethics Committee convened by the Department of Children and Youth Affairs of the
Irish Government. This included a review of all instrumentation, recruitment, consent and
implementation protocols adopted at all stages of the study [28].
 58

2.3. Statistical Analysis and Dependent Variable

The Statistical Package for the Social Sciences statistical software package version 19.0
(SPSS, Inc., Chicago, IL, USA) and STATA 13 (StataCorp LP, College Station, TX, USA) were used
for the statistical analysis. Several independent variables considered as risk factors for early
complimentary feeding were selected from the database. These included demographic factors, such as
maternal age, maternal education, socioeconomic status and parity, and biological factors, such as
maternal BMI, mode of delivery and infant’s health. In order to study the variations in complimentary
feeding by ethnic group, as well as the influence of breastfeeding in its timing, the model was also
adjusted for ethnicity, length of stay in the ROI, breastfeeding initiation and duration of exclusive
breastfeeding. Data was analysed using cross-WDEXODWLRQV DQG WKH Ȥ2 statistical test, as well as
multivariate binary logistic regression. Independent variables were included in the multivariate
analysis if they were significant in the bivariate analysis.
Mothers were asked to report their ethnic or cultural background. The following options were
provided; Irish, Irish traveller, any other white background, African, any other black background,
Chinese, any other Asian background, and other, including mixed background. A recoded binary
variable was constructed with two categories: Irish ethnic background and non-Irish ethnic
background. This recoded variable was then combined with a variable that asked non-Irish mothers
how long they had been residing in the ROI.
The dependent variable “early complimentary feeding” was constructed from a question in the
database that asked mothers to indicate when they started to give their infants solid foods at least twice
a day for several weeks. Solid foods were defined as baby cereals, pureed fruits, etc., and not milk or
drinks. The dependent variable used is therefore the age at which complimentary feeding was
established rather than the child’s age when solid foods were first introduced. Following ESPGHAN’s
guidelines, a binary dependent variable was created with two categories <17 weeks for early
complimentary feeding and • ZHHNV IRU WKH DFFHSWDEOH LQWURGXFWLRQ RI FRPSOLPHQWDU\
feeding [5,28]. Statistical significance was taken as a p-value of < 0.05. The weights were on for all
statistical analysis.

2.4. Definition of Covariates

Socio-economic status (SES) was assessed using three different indicators: household class,
equivalised household income quintiles and household type. Income is equivalised to take into account
household size and composition using the modified Organization for Economic Cooperation and
Development equivalence scale (first adult value, 1; second or higher adults, 0.5; children
aged < 14, 0.3). Primary and secondary caregivers were asked questions about their current occupation
to derive the variable household class. Where the respondent was economically inactive (retired or
unemployed) at the time of interview, previous employment was considered. The household class
classification adopted was that used by the Central Statistics Office (CSO): professional workers,
managerial and technical, non-manual, skilled-manual, semi-skilled, unskilled, all other gainfully
occupied and unknown and never worked at all. This variable was recoded to contain only five
categories: professional, managerial and technical workers; non-manual; skilled and semi-skilled
 59

manual; unskilled and all other gainfully occupied and unknown, and never worked at all. Household
type is a fourfold variable derived from whether the study child is living in a one or two parent family,
as well as the number of children (<18 years) living in the household. This resulted in a classification
as follows: one parent, one child; one parent, two or more children; two parents, one child;
two parents, two or more children [28].
Maternal education was coded as follows: no formal or primary education, secondary education and
third-level education. Maternal age was coded as follows: ” 4, 25–34 and • \HDUV ROG 0HDVXUHG
parent BMI was classified according to the World Health Organization (WHO) classifications as
underweight <18.5 kg/m2, normal weight •DQGNJP 2, overweight •NJP 2 and <30 kg/m2
and obese •NJP2 [28].
Mothers were asked about their infants’ overall health using the question: “In general, how would
you describe the baby’s current health” with response categories “very healthy, no problems”,
“healthy, but a few minor problems”, “sometimes quite ill” and “always unwell”. Children are defined
as having been breastfed if they consumed breast milk at any stage regardless of the amount of time
the baby was breastfed, including the colostrum in the first few days after birth. Exclusive
breastfeeding was defined as the infant receiving only breast milk without any additional food or drink,
regardless of the length of exclusive breastfeeding. The variable “duration of exclusive breastfeeding”
was constructed from the question “how old was the baby when he/she stopped being
exclusively breastfed” [28,29].

2.5. Missing Data

Some of the independent variables used in the analysis had a large percentage of missing cases:
maternal BMI (5.1%) and equivalised household annual income (7.8%). This would have resulted in a
proportion of the sample being lost from the analysis. In response, multiple imputation has been
carried out in STATA 13 using the variables maternal age, maternal education, father’s education,
household class, maternal employment status, current maternal smoking and migrant status to predict
the missing values in maternal BMI and equivalised household annual income.

3. Results

3.1. Characteristics of the Study Cohort

Table 1 shows the characteristics of mothers disaggregated by ethnic group. The primary caregiver
was defined as the person who spent the most time with the study infant. Irish mothers were younger
on average than non-Irish (31.7 years with SD 5.3 compared to 30.9 with SD 5.4). Of those mothers
with an Irish ethnic background, 49.5% initiated breastfeeding compared to 88.1% of those mothers
with a non-Irish ethnic background. The mean duration of any breastfeeding for those that breastfed at
all was 71.1 days (SD 66.4) for Irish mothers compared to 95.8 days (SD 69.2) for non-Irish. A higher
percentage of mothers with an Irish ethnic background (15.5% vs. 7.6%) introduced complimentary
foods early.
 60

3.2. Percentage of Infants’ Breastfed and the Introduction of Complimentary Feeding in the
<17 Weeks and •:HHNV&DWHJRULHV&ODVVLILHGE\(WKQLF*URXS

Figure 1 shows that the ethnic group with the highest percentage of mothers who breastfed their
infants was African or any other black background, with 92.5% of mothers initiating breastfeeding.
This group was followed by Chinese mothers, with 91.6% of breastfeeding initiation. Figure 2 shows
the percentage of mothers introducing complimentary foods in the <17 weeks • and
ZHHNV
categories classified by ethnic group. Fifteen-point-five percent of mothers with an Irish ethnic
background introduced complimentary foods early (<17 weeks). The group with the lowest percentage
of mothers introducing complimentary foods early (4.8%) was Chinese or any other Asian background.

Figure 1. Breastfeeding initiation by ethnic group.

100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Irish Any other white African or any Chinese or any Other (including
( n = 9,273) background other black other Asian mixed
( n = 1,204) background background background)
(n = 295) ( n = 273) (n = 52)

Yes No

Figure 2. Introduction of complimentary feeding by ethnic group.

Introduction of complimentary feeding

by ethnic group
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Irish Any other white African or any Chinese or any Other (including
( n = 9,273) background other black other Asian mixed
( n = 1,204) background background background)
(n = 295) ( n = 273) (n = 52)

< 17 weeks > = 17 weeks

 61

Table 1. Characteristics of whole sample, national and non-national primary caregivers.

Irish Non-Irish
Characteristic Primary Caregiver †
Sample (n *) Mean or % SD Sample (n *) Mean or % † SD
Mean age 31.7 5.3 30.9 5.4
Underweight (less than 18.5) 9275 2.5 1859 3.7
Normal weight (18.5–24.9) 50.5 51.1
BMI primary carer kg/m2
Overweight (25–29.9) 30.1 28.5
Obese (• 16.8 16.7
Yes 9273 49.5 1859 88.1
Having ever breastfed
No 50.5 11.9
Yes 4592 39.3 1634 66.5
Having exclusively breastfed ‡
No 10.2 21.4
Mean duration of any breastfeeding (days) ‡ 4019 71.1 66.4 1153 95.8 69.2
‡
Mean duration of exclusive breastfeeding (days) 3556 74.7 62 1158 95.9 63.2
<17 weeks 9151 15.5 1755 7.6
Introduction of complimentary feeding
•ZHHNV 84.5 92.4
† ‡
* n provided is number of primary caregivers who answered each question; Percentages provided are based on the total sample; mothers who reported not having ever
breastfed were filtered out.
 62

3.3. Predictors of Early Complimentary Feeding

Table 2 shows the adjusted model of significant factors that independently predicted early
complimentary feeding introduction for the whole sample. After adjustment, the significant factors
included the primary caregiver’s age, education, BMI, ethnicity and length of stay in the ROI,
household class, household type and exclusive breastfeeding duration.
Non-Irish mothers who had been living in the ROI < 6 years were 50.7% less likely to introduce
complimentary feeding early compared to Irish mothers (OR 0.493, 95% CI 0.371, 0.656). This
protective effect of ethnicity decreased with the length of stay in the ROI, with non-Irish mothers who
had been living in the ROI 11 to >20 years being 3.1% less likely to introduce complimentary feeding
early when compared to Irish mothers (OR 0.969, 95% CI 0.613, 1.532).
Those mothers who exclusively breastfed >90 days were 93.9% less likely to introduce early
complimentary feeding (OR 0.061, 95% CI 0.037, 0.101) when compared to those who did not
exclusively breastfed.

3.4. Effects of Acculturation on Breastfeeding Initiation and Early Complimentary Feeding

Figure 3 shows the effects of acculturation on breastfeeding initiation and early complimentary
feeding. Breastfeeding initiation in the non-Irish cohort reduced from 89.4 of mothers who had arrived
within the last year to five years ago to 67.5% for those who had arrived 11 to >20 years ago
(p < 0.001). The percentage of non-Irish mothers introducing complimentary foods early increased
from 6.6% for those who had arrived within the last year to five years ago to 16.0% for those who
arrived 11 to >20 years ago (p < 0.001).

Figure 3. The effect of acculturation on breastfeeding initiation and early complimentary feeding.

100
89.4
90 85.3

80
67.5
% Non-Irish mothers

70
60
50
40
30
16.0
20
6.6 7.9
10
0
Within the last year – 5 years 6 – 10 years ago 11 – >20 years ago
ago
Early Complementary Feeding Breastfeeding initiation
 63

Table 2. Characteristics of Irish and non-Irish primary caregivers and households in the <17 weeks and •ZHHNVFRPSOLPHQWDU\IHHGLQJ
categories and binary logistic regression of the significant factors associated with the introduction of complimentary feeding.
Complimentary Feeding Introduction *
Characteristics Total <17 Weeks •:HHNV Unadjusted † Adjusted ‡
n %Ȗ n %Ș n %Ș OR p§ OR 95% CI pį
Primary caregiver age
(years)
•35 3467 31.8 426 12.3 3041 87.7 0.467 0.606 0.494 0.744 <0.001
25–34 6119 56.1 823 13.4 5296 86.6 0.519 0.664 0.555 0.794 <0.001
$ $
”24 1321 12.1 304 23.0 1017 77.0 1.0 <0.001 1.0 <0.001
Maternal education
Third level education 5330 48.9 557 10.5 4773 89.5 0.534 0.820 0.602 1.117 0.208
Secondary level
5189 47.6 926 17.8 4263 82.2 0.994 1.031 0.773 1.374 0.835
education
No formal or primary
378 3.5 68 18 310 82 1.0$ <0.001 1.0 $ 0.004
education
Household class
Professional, managerial
5228 47.9 578 11.1 4650 88.9 1.0 $ 1.0 $ 0.001
and technical workers
Non-manual 1983 18.2 332 16.7 1651 83.3 1.620 1.179 1.001 1.387 0.048
Skilled and semi-skilled
2428 22.3 377 15.5 2051 84.5 1.478 1.111 0.941 1.311 0.214
manual
Unskilled and all other
gainfully occupied and 283 2.6 66 23.3 217 76.7 2.430 1.850 1.346 2.544 <0.001
unknown
Never worked at all 985 9.0 200 20.3 785 79.7 2.051 <0.001 0.972 0.752 1.258 0.830
 64

Table 2. Cont.
Household type
One parent one child
799 7.3 177 22.2 622 77.8 1.042 0.925 0.720 1.189 0.545
under 18 years
One parent two or more
831 7.6 178 21.4 653 78.6 1.0$ 1.0 $ <0.001
children under 18 years
Two parents one child
3536 32.4 385 10.9 3151 89.1 0.447 0.614 0.479 0.788 <0.001
under 18 years
Two parents two or more
5740 52.6 813 14.2 4927 85.8 0.604 <0.001 0.827 0.659 1.038 0.101
children under 18 years
BMI primary carer
Underweight
298 2.7 43 14.4 255 85.6 1.132 0.903 0.639 1.274 0.569
(less than 18.5 kg/m2)
Normal weight
5519 50.6 714 12.9 4805 87.1 1.0 $ 1.0 $ 0.001
(18.5–24.9 kg/m2)
Overweight
3264 29.9 456 14.0 2808 86 1.093 1.074 0.942 1.224 0.284
(25–29.9 kg/m2)
Obese (• kg/m2) 1825 16.7 340 18.6 1485 81.4 1.542 <0.001 1.336 1.151 1.551 <0.001
Primary carer ethnicity
& length of stay in the
ROI
Irish 9188 84.2 1424 15.5 7764 84.5 1.0 $ 1.0 $ 0.001
Non-Irish arrived within
964 8.8 62 6.4 902 93.6 0.374 0.493 0.371 0.656 <0.001
the last year-5 years ago
Non-Irish arrived
570 5.2 44 7.7 526 92.3 0.456 0.566 0.408 0.785 0.001
6–10 years ago
Non-Irish arrived
184 1.7 23 12.5 161 87.5 0.784 <0.001 0.969 0.613 1.532 0.893
11–>20 years ago
 65

Table 2. Cont.
Exclusive breastfeeding
duration
0–30 days 1579 14.5 241 15.3 1338 84.7 0.818 0.954 0.776 1.172 0.652
>30–60 days 770 7.1 101 13.1 669 86.9 0.687 0.864 0.665 1.122 0.272
>60–90 days 549 5.0 69 12.6 480 87.4 0.652 0.813 0.604 1.094 0.172
>90 days 1782 16.3 17 1.0 1765 99.0 0.044 0.061 0.037 0.101 <0.001
No exclusive
6227 57.1 1125 18.1 5102 81.9 1.0 $ <0.001 1.0 $ <0.001
breastfeeding
†
* Bivariate analysis using Ȥ2 statistical tests to compare the differences between primary caregivers, infants and households in the <17 weeks and
•ZHHN s groups.
Values are OR that were obtained from individual bivariate analysis of independent variables when compared to the dependent complimentary feeding variable <17 weeks
Ș
and • ZHHNV JURXSV Ȗ Total percentage. Percentage within each independent variable category who introduced complimentary feeding in the <17 weeks •
and
§
weeks groups. P-value resulting from unadjusted regression analysis of the independent variable with the complimentary feeding dependent variable <17 weeks and •
‡ į
weeks. Values are OR that were obtained from the final binary logistic regression model. P-values obtained from the adjusted binary logistic regression model. The
model was adjusted for maternal age, education, BMI, SES, parity, mode of delivery, breastfeeding initiation, duration of exclusive breastfeeding, and infant's health
status. 1.0 $ Denotes the reference group.
 66

4. Discussion

Breastfeeding is the most beneficial and nutritionally complete feeding method during infancy [30].
However, breastfeeding initiation rates in the ROI were the lowest compared to 14 European countries
in 2010 [21]. Despite modest increases in breastfeeding rates, as shown in the Perinatal Statistics
Report in 2012, these rates are still far from national targets and international averages [21,31–34].
In Table 1, it can be observed how a lower percentage of Irish mothers (49.5%) initiated
breastfeeding compared to their non-Irish counterparts (88.1%). These findings correlate with previous
studies in the ROI, which found similar percentages of breastfeeding initiation in the Irish and
non-Irish mothers [23,26,27]. The percentages between the two cohorts are nearer to each other when
mothers are asked about exclusive breastfeeding. However, the mean duration of any breastfeeding, as
well as exclusive breastfeeding is lower for the Irish cohort. Both cohorts are far from complying with
guidelines recommending six months of exclusive breastfeeding. However, a stronger predisposition
towards breastfeeding, possibly due to cultural differences, is observed in the non-Irish group.
Figure 1 also shows that breastfeeding initiation was higher in all other ethnic groups when
compared to the Irish cohort. The fact that any other white background has 86.5% breastfeeding
initiation concurs with the study findings from 2010 in which Ireland had the lowest breastfeeding
rates when compared to 14 European countries.
Differences in breastfeeding rates by ethnic background have been pointed out by other studies
internationally [35–39]. Acculturation plays a role in infant feeding practices; as shown in Figure 2,
the amount of non-national mothers initiating breastfeeding decreased the longer they had been living
in the ROI (p < 0.001). Moreover, the percentage of these mothers introducing complimentary foods
early also increased with a longer stay in the country (p < 0.001). This finding suggests the close
relationship between the early milk feeding method chosen and the introduction of complimentary
foods. Further exploration of the reasons behind these changes in infant feeding choices by non-Irish
mothers is needed. Factors, such as societal pressures, language barriers and the perception of
behaviours in the adopted culture as being modern, could potentially play a role in the
acculturation mechanisms.
The relationship between acculturation and milk feeding choices has been reported by different
studies in the United States (US) and Australia [40–44]. A study published in 2010 found that a group
of Chinese mothers living in Ireland had a less positive attitude and more misconceptions about
breastfeeding than a group of Chinese mothers living in Perth, Australia, suggesting a possible role of
‘acculturation’ and the mothers adapting themselves to the formula feeding culture of Ireland [45].
On the other hand, a 2013 study from Australia pointed out that Chinese mothers living in Perth had
higher breastfeeding initiation rates and a longer duration of breastfeeding than Chinese mothers in
Chengdu. Reported breastfeeding initiation rates in Australia are much higher than in the ROI [31,46].
These findings suggest that the culture of the adopted country may be an important influence on infant
feeding practices among migrants.
Lack of breastfeeding and the use of formula feeding have been related to early complimentary
feeding by many studies [14,15,17,18]. Formula feeding has been associated with impairment of
appetite self-regulatory mechanisms, leading to infants demanding the introduction of solids earlier,
with no subsequent reduction in milk intake during the complimentary feeding period. This
 67

interference with self-regulating mechanisms early in life could have long-term health consequences,
increasing the risk of being overweight and obesity later in life [7,8,13,47].
Several studies have linked early complimentary feeding to a higher risk of being overweight and
obesity during childhood and later in life [6,7,9]. An analysis of the same cohort at three years of age
found that those children who were introduced to complimentary feeding later had a lower prevalence
of being overweight or obesity [48]. Previous studies on complimentary feeding in the ROI have
shown poor compliance with current guidelines, with more than 70% of infants being introduced to
complimentary foods <17 weeks [15]. However, these studies did not explore ethnic variations in
complimentary feeding.
An important finding in this study is observed in Figure 2, which shows that a higher percentage of
Irish mothers (15.5%) introduced complimentary foods early when compared to the other ethnic
groups. It has to be noted that the prevalence of infants introduced early to complimentary foods in this
study is probably an underestimation, because mothers were asked for the child’s age at which point
solid foods had been regularly given. The group with the lowest percentage of mothers introducing
complimentary foods early were those of Chinese or any other Asian background (4.8%). Interestingly,
this was one of the ethnic groups with one of the highest breastfeeding rates, which suggests a close
relationship between early milk feeding and complimentary feeding.
The predictors of early complimentary feeding were studied for the whole sample. An important
finding is that belonging to a different ethnic background than Irish had a protective role against early
complimentary feeding, which was reduced with a longer length of stay in the ROI (Table 2). Figure 3
shows how the acculturation of non-Irish mothers resulted in a decrease in the breastfeeding rate,
which correlates with an increase in the percentage of mothers introducing early complimentary foods.
This finding highlights again the role played by acculturation and the adoption of formula milk in the
timing of complimentary feeding introduction.
The inclusion of the duration of exclusive breastfeeding in the adjusted model resulted in a loss of
the significance of breastfeeding initiation with little change in the rest of the significant predictors.
This result suggests the importance of exclusive breastfeeding and its potential role in the timing of
solids introduction and, ultimately, in the development of being overweight and obesity.
There is inconsistency in the results of studies on early complimentary feeding and the risk of
developing being overweight and obesity. Moreover, a longer duration of breastfeeding is associated
with the later introduction of complimentary foods. In the present study, a longer duration of exclusive
breastfeeding resulted in a decrease in the probability of early complimentary feeding. Therefore,
complimentary feeding could potentially be a confounder in the relationship between breastfeeding
and being overweight or obesity [49–52].
Another interesting finding was the fact that maternal BMI was a predictor of early complimentary
feeding. The relationship between being overweight, obesity and breastfeeding duration has been well
studied, suggesting that overweight and obese women are at higher risk of early cessation of
breastfeeding, due to biological and mechanical factors. [53–56].
 68

5. Strengths and Limitations

GUI is a large and nationally representative sample. The results of the study can be applied at a
population level, due to the application of the sampling weights. Parental BMI was measured by
trained professionals using validated techniques.
However, there are several limitations to the present study. It would have been desirable to collect
information on the first introduction of solids into the infant’s diet to allow comparability with other
studies on complimentary feeding. The results must also be interpreted with caution, as the information
was collected retrospectively, when the infant was nine months of age, increasing the possibility of
recall bias.
Maternal BMI was measured at the time of interview, which took place when the infant was nine
months old. Therefore, we assume that those mothers who were overweight or obese at that point in
time belonged to the same BMI category pre-pregnancy.

6. Conclusions

The results from this study suggest that, after adjusting for other maternal characteristics,
inappropriate infant feeding practices are more common among Irish mothers when compared to
non-Irish mothers residing in the ROI. Acculturation plays an important role in infant feeding practices
among non-Irish mothers. Therefore, cultural differences are an important factor in shaping patterns of
infant feeding in the ROI.
There is a strong association between breastfeeding and the early introduction of complimentary
feeding. The ROI continues to have one of the lowest breastfeeding rates in the world. Existing
policies to increase breastfeeding rates have been largely ineffective and with recent increases in the
breastfeeding rate explained by an increase in the proportion of non-Irish mothers residing in the ROI
and increasing maternal education and age, characteristics that are associated with a higher propensity
to breastfeed in Ireland. The immediate revision of current support, education and policies on infant
feeding practices would appear desirable to achieve the implementation of desirable infant feeding
practices in line with WHO and ESPGHAN recommendations.

Acknowledgments

This research received no specific grant from any funding agency in the public, commercial or
not-for-profit sectors.

Author Contributions

Patricia Dominguez Castro contributed towards data analysis and interpretation and led the writing.
Richard Layte helped in interpreting the results and provided critical feedback on the statistical
analysis of the data, as well as methods used to collect the same. He also provided feedback on drafts
of the paper. John Kearney helped in interpreting the results and provided feedback on drafts of the
paper. All authors approved the final version of the paper.
 69

Conflicts of Interest

The authors declare no conflict of interest.

References

1. Lanigan, J.; Singhal, A. Early nutrition and long-term health: A practical approach. Proc. Nutr. Soc.
2009, 68, 422–429.
2. Koletzko, B.; Brands, B.; Poston, L.; Godfrey, K.; Demmelmair, H. Early nutrition programming
of long-term health. Proc. Nutr. Soc. 2012, 71, 371–378.
3. Singhal, A.; Lucas, A. Early origins of cardiovascular disease: Is there a unifying hypothesis?
Lancet 2004, 363, 1642–1645.
4. Complementary Feeding: Report of the Global Consultation, and Summary of Guiding Principles
for Complementary Feeding of the Breastfed Child. Available online: http://www.who.int/
maternal_child_adolescent/documents/924154614X/en/ (accessed on the 9 February 2014).
5. Agostoni, C.; Decsi, T.; Fewtrell, M.; Goulet, O.; Kolacek, S.; Koletzko, B.; Michaelsen, K.F.;
Moreno, L.; Puntis, J.; Rigo, J.; et al. Complementary feeding: A commentary by the ESPGHAN
Committee on Nutrition. J. Pediatr. Gastroenterol. Nutr. 2008, 46, 99–110.
6. Wilson, A.C.; Forsyth, J.S.; Greene, S.A.; Irvine, L.; Hau, C.; Howie, P.W. Relation of infant diet
to childhood health: Seven year follow up of cohort of children in Dundee infant feeding study.
BMJ 1998, 316, doi:10.1136/bmj.316.7124.21.
7. Huh, S.Y.; Rifas-Shiman, S.L.; Taveras, E.M.; Oken, E.; Gillman, M.W. Timing of solid food
introduction and risk of obesity in preschool-aged children. Pediatrics 2011, 127, e544–e551.
8. Mihrshahi, S.; Battistutta, D.; Magarey, A.; Daniels, L.A. Determinants of rapid weight gain
during infancy: Baseline results from the NOURISH randomised controlled trial. BMC Pediatr.
2011, 11, doi:10.1186/1471-2431-11-99.
9. Schack-Nielsen, L.; Sorensen, T.; Mortensen, E.L.; Michaelsen, K.F. Late introduction of
complementary feeding, rather than duration of breastfeeding, may protect against adult
overweight. Am. J. Clin. Nutr. 2010, 91, 619–627.
10. Cooke, L. The importance of exposure for healthy eating in childhood: A review. J. Hum. Nutr.
Diet. 2007, 20, 294–301.
11. Nicklaus, S.; Boggio, V.; Chabanet, C.; Issanchou, S. A prospective study of food variety seeking
in childhood, adolescence and early adult life. Appetite 2005, 44, 289–297.
12. Nicklaus, S. Children’s acceptance of new foods at weaning. Role of practices of weaning and of
food sensory properties. Appetite 2011, 57, 812–815.
13. Grote, V.; Theurich, M.; Koletzko, B. Do complementary feeding practices predict the later risk
of obesity? Curr. Opin. Clin. Nutr. Metab. Care. 2012, 15, 293–297.
14. Schiess, S.; Grote, V.; Scaglioni, S.; Luque, V.; Martin, F.; Stolarczyk, A.; Vecchi, F.;
Koletzko, B. Introduction of complementary feeding in 5 European countries. J. Pediatr.
Gastroenterol. Nutr. 2010, 50, 92–98.
 70

15. Tarrant, R.C.; Younger, K.M.; Sheridan-Pereira, M.; White, M.J.; Kearney, J.M. Factors
associated with weaning practices in term infants: A prospective observational study in Ireland.
Br. J. Nutr. 2010, 104, 1544–1554.
16. Clayton, H.B.; Li, R.; Perrine, C.G.; Scanlon, K.S. Prevalence and reasons for introducing infants
early to solid foods: Variations by milk feeding type. Pediatrics 2013, 131, e1108–e1114.
17. Avery, M.; Duckett, L.; Dodgson, J.; Savik, K.; Henly, S.J. Factors associated with very early
weaning among primiparas intending to breastfeed. Matern. Child Health J. 1998, 2, 167–179.
18. Wijndaele, K.; Lakshman, R.; Landsbaugh, J.R.; Ong, K.K.; Ogilvie, D. Determinants of early
weaning and use of unmodified cow’s milk in infants: A systematic review. J. Am. Diet. Assoc.
2009, 109, 2017–2028.
19. Department of Health and Children. Policy Change in Breastfeeding Guidelines; Department of
Health and Children: Dublin, Ireland, 2003.
20. FSAI. Scientific Recommendations for a National Infant Feeding Policy, 2nd ed.; Food Safety
Authority of Ireland: Dublin, Ireland, 2011.
21. Euro-Peristat Project with SCPE and EUROCAT. European Perinatal Health Report. The Health
and Care of Pregnant Women and Babies in Europe in 2010. Available online:
www.europeristat.com (accessed on the 24 April 2014).
22. McSweeney, M.; Kevany, J. A national survey of infant feeding practices in Ireland 1981:
A preliminary report. Ir. Med. J. 1982, 75, 452–455.
23. Brick, A.; Nolan, A. Explaining the increase in breastfeeding at hospital discharge in Ireland,
2004–2010. Ir. J. Med. Sci. 2013, doi:10.1007/s11845-013-1012-0.
24. Department of Health and Children. A National Breastfeeding Policy for Ireland; Department of
Health and Children: Dublin, Ireland, 1994.
25. Department of Health and Children. Breastfeeding in Ireland: A Five Year Strategic Action Plan;
Department of Health and Children: Dublin, Ireland, 2005.
26. Tarrant, R.C.; Younger, K.M.; Sheridan-Pereira, M.; White, M.J.; Kearney, J.M. The prevalence
and determinants of breast-feeding initiation and duration in a sample of women in Ireland. Public
Health Nutr. 2010, 13, 760–770.
27. Begley, C.; Gallagher, L.; Clarke, M.; Carroll, M.; Millar, S. National Infant Feeding Survey
2008; Trinity College Dublin: Dublin, Ireland, 2009.
28. Quail, A.; Williams, J.; McCrory, C.; Murray, A.; Thornton, M. Summary Guide to Wave 1 of the
Infant Cohort (at 9 months) of Growing Up in Ireland. Available online: www.ucd.ie/issda/static/
documentation/esri/GUI-SummaryGuideInfants.pdf (accessed on the 10 February 2014).
29. Quail, A.; Williams, J.; McCrory, C.; Murray, A.; Thornton, M. Questionnaires for Wave 1 of the
Infant Cohort [at nine months] of Growing Up in Ireland. Available online:
http://www.ucd.ie/issda/static/documentation/esri/GUI-QuestionnairesInfants.pdf (accessed on 24
April 2014).
30. Kramer, M.S.; Kakuma, R. Optimal Duration of Exclusive Breastfeeding. Available online:
www.who.int/rhl/reviews/CD003517.pdf (accessed on 24 April 2014).
31. Economic and Social Research Institute. Perinatal Statistics Report; Economic and Social
Research Institute: Dublin, Ireland, 2012.
 71

32. Oakley, L.L.; Renfrew, M.J.; Kurinczuk, J.J.; Quigley, M.A. Factors associated with
breastfeeding in England: an analysis by primary care trust. BMJ Open 2013, 3, doi:10.1136/
bmjopen-2013-002765.
33. Parker, M.; Burnham, L.; Cook, J.; Sanchez, E.; Philipp, B.L.; Merewood, A. 10 Years after
baby-friendly designation: Breastfeeding rates continue to increase in a US neonatal intensive
care unit. J. Hum. Lact. 2013, 29, 354–358.
34. Norris, S.; Collin, S.M.; Ingram, J. Breastfeeding practices among the old order Mennonites in
Ontario, Canada: A multiple methods study. J. Hum. Lact. 2013, 29, 605–610.
35. Belanoff, C.M.; McManus, B.M.; Carle, A.C.; McCormick, M.C.; Subramanian, S.V.
Racial/ethnic variation in breastfeeding across the US: A multilevel analysis from the National
Survey of Children’s Health, 2007. Matern. Child Health J. 2012, 16, S14–S26.
36. Bulk-Bunschoten, A.M.; Pasker-de Jong, P.C.; van Wouwe, J.P.; de Groot, C.J. Ethnic variation
in infant-feeding practices in the Netherlands and weight gain at 4 months. J. Hum. Lact. 2008,
24, 42–49.
37. Ahluwalia, I.B.; Morrow, B.; D’Angelo, D.; Li, R. Maternity care practices and breastfeeding
experiences of women in different racial and ethnic groups: Pregnancy Risk Assessment and
Monitoring System [PRAMS]. Matern. Child Health J. 2012, 16, 1672–1678.
38. Sparks, P.J. Racial/ethnic differences in breastfeeding duration among WIC-eligible families.
Womens Health Issues 2011, 21, 374–382.
39. Kelly, Y.J.; Watt, R.G.; Nazroo, J.Y. Racial/ethnic differences in breastfeeding initiation and
continuation in the United Kingdom and comparison with findings in the United States. Pediatrics
2006, 118, e1428–e1435.
40. Ahluwalia, I.B.; D’Angelo, D.; Morrow, B.; McDonald, J.A. Association between acculturation
and breastfeeding among Hispanic women: Data from the Pregnancy Risk Assessment and
Monitoring System. J. Hum. Lact. 2012, 28, 167–173.
41. Chapman, D.J.; Perez-Escamilla, R. Acculturative type is associated with breastfeeding duration
among low-income Latinas. Matern. Child Nutr. 2013, 9, 188–198.
42. Haughton, J.; Gregorio, D.; Perez-Escamilla, R. Factors associated with breastfeeding duration
among Connecticut Special Supplemental Nutrition Program for Women, Infants, and Children
[WIC] participants. J. Hum. Lact. 2010, 26, 266–273.
43. Singh, G.K.; Kogan, M.D.; Dee, D.L. Nativity/immigrant status, race/ethnicity, and
socioeconomic determinants of breastfeeding initiation and duration in the United States, 2003.
Pediatrics 2007, 119, S38–S46.
44. Chen, S.; Binns, C.W.; Zhao, Y.; Maycock, B.; Liu, Y. Breastfeeding by Chinese mothers in
Australia and China: The healthy migrant effect. J. Hum. Lact. 2013, 29, 246–252.
45. Zhou, Q.; Younger, K.M.; Kearney, J.M. An exploration of the knowledge and attitudes towards
breastfeeding among a sample of Chinese mothers in Ireland. BMC Public Health 2010, 10,
doi:10.1186/1471-2458-10-722.
46. Department of Health Australian Government. Available online: http://www.health.gov.au/
breastfeeding (accessed on 24 April 2014).
 72

47. Mehta, U.J.; Siega-Riz, A.M.; Herring, A.H.; Adair, L.S.; Bentley, M.E. Pregravid body mass
index is associated with early introduction of complementary foods. J. Acad. Nutr. Diet. 2012,
112, 1374–1379.
48. Children’s Physical Growth from Birth to Age 3. Available online: http://www.growingup.ie/
index.php?id=61 (accessed on 24 April 2014).
49. Van Rossem, L.; Kiefte-de Jong, J.C.; Looman, C.W.; Jaddoe, V.W.; Hofman, A.;
Hokken-Koelega, A.C.; Mackenbach, J.P.; Moll, H.A.; Raat, H. Weight change before and after
the introduction of solids: Results from a longitudinal birth cohort. Br. J. Nutr. 2013, 109,
370–375.
50. Griffiths, L.J.; Smeeth, L.; Hawkins, S.S.; Cole, T.J.; Dezateux, C. Effects of infant feeding
practice on weight gain from birth to 3 years. Arch. Dis. Child. 2009, 94, 577–582.
51. Moorcroft, K.E.; Marshall, J.L.; McCormick, F.M. Association between timing of introducing
solid foods and obesity in infancy and childhood: A systematic review. Matern. Child Nutr. 2011,
7, 3–26.
52. Woo, J.G.; Guerrero, M.L.; Ruiz-Palacios, G.M.; Peng, Y.M.; Herbers, P.M.; Yao, W.;
Ortega, H.; Davidson, B.S.; McMahon, R.J.; Morrow, A.L. Specific infant feeding practices do
not consistently explain variation in anthropometry at age 1 year in urban United States, Mexico,
and China cohorts. J. Nutr. 2013, 143, 166–174.
53. Baker, J.L.; Michaelsen, K.F.; Sorensen, T.I.; Rasmussen, K.M. High prepregnant body mass
index is associated with early termination of full and any breastfeeding in Danish women.
Am. J. Clin. Nutr. 2007, 86, 404–411.
54. Hilson, J.A.; Rasmussen, K.M.; Kjolhede, C.L. High prepregnant body mass index is associated
with poor lactation outcomes among white, rural women independent of psychosocial and
demographic correlates. J. Hum. Lact. 2004, 20, 18–29.
55. Li, R.; Jewell, S.; Grummer-Strawn, L. Maternal obesity and breast-feeding practices.
Am. J. Clin. Nutr. 2003, 77, 931–936.
56. Neville, M.C.; Morton, J. Physiology and endocrine changes underlying human lactogenesis II.
J. Nutr. 2001, 131, 3005S–3008S.

© 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).
 73

Reprinted from Nutrients. Cite as: Khanal, V.; da Cruz, J.L.N.B.; Karkee, R.; Lee, A.H. Factors
Associated with Exclusive Breastfeeding in Timor-Leste: Findings from Demographic and Health
Survey 2009–2010. Nutrients 2014, 6, 1691-1700.

Factors Associated with Exclusive Breastfeeding in

Timor-Leste: Findings from Demographic and Health Survey
2009–2010
Vishnu Khanal 1, Jonia Lourenca Nunes Brites da Cruz 2, Rajendra Karkee 3 and
Andy H. Lee 4,*
1
Sanjeevani College of Medical Sciences, Butwal, Rupandehi, Nepal;
E-Mail: khanal.vishnu@gmail.com
2
National Hospital Guido Valadares, Ministry of Health, Dili, Timor Leste;
E-Mail: joniadacruz@yahoo.com
3
School of Public Health and Community Medicine, BP Koirala Institute of Health Sciences,
Dharan, Nepal; E-Mail: rkarkee@yahoo.com
4
School of Public Health, Curtin University, Perth, WA 6153, Australia

* Author to whom correspondence should be addressed; E-Mail: Andy.Lee@curtin.edu.au;

Tel.: +61-8-9266-4180; Fax: +61-8-9266-2958.

Received: 30 December 2013; in revised form: 15 April 2014 / Accepted: 16 April 2014 /
Published: 22 April 2014

Abstract: Exclusive breastfeeding is known to have nutritional and health benefits. This
study investigated factors associated with exclusive breastfeeding among infants aged five
months or less in Timor-Leste. The latest data from the national Demographic and Health
Survey 2009–2010 were analyzed by binary logistic regression. Of the 975 infants included
in the study, overall 49% (95% confidence interval 45.4% to 52.7%) were exclusively
breastfed. The exclusive breastfeeding prevalence declined with increasing infant age,
from 68.0% at less than one month to 24.9% at five months. Increasing infant age, mothers
with a paid occupation, who perceived their newborn as non-average size, and residence in
the capital city Dili, were associated with a lower likelihood of exclusive breastfeeding. On
the other hand, women who could decide health-related matters tended to breastfeed
exclusively, which was not the case for others whose decisions were made by someone
else. The results suggested the need of breastfeeding promotion programs to improve the
exclusive breastfeeding rate. Antenatal counseling, peer support network, and home visits
by health workers could be feasible options to promote exclusive breastfeeding given that
the majority of births occur at home.
 74

Keywords: breastfeeding; demographic and health survey; exclusive breastfeeding;

Timor-Leste

1. Introduction

Exclusive breastfeeding means only breastmilk is allowed with the exception of medicine, vitamin
syrup and oral rehydration solution [1,2]. It is known that exclusive breastfeeding for six months can
protect infants from short term illnesses such as gastroenteritis, respiratory infection and under
nutrition; and in the long term, against chronic diseases such as type 2 diabetes, hypertension and
obesity [3,4]. The economic benefits include cost savings from avoiding illness, workdays lost and the
purchase of infant formula [5]. Moreover, it has been projected that 11.6% of child deaths in 2011
could be attributable to sub-optimal breastfeeding [6]. A variety of factors have been reported to affect
the practice of exclusive breastfeeding, including maternal characteristics (education, occupation,
health condition, age), infant characteristics (sex, birth order, illness), and cultural practices (initiation
of breastfeeding, time of introduction of complementary feeds) [7,8]. The effects of these factors vary
according to cultural context and related socioeconomic conditions.
Timor-Leste is one of the youngest countries in Asia which gained independence since 2002 [9].
The country went through a long armed conflict during 1990s, leading to destruction of most of the
infrastructure [9], leaving thousands of its citizens being displaced from East to West Timor [10]. The
majority of health workforce returned to Indonesia after independence. Only a small number of health
professionals remained to re-establish the health system with assistance from the United Nations,
expatriate workers, and other international support [10]. The population of Timor-Leste is estimated to
be 1.07 million in 2010 with a growth rate of 2.4% [11]. The country has a high infant mortality
(45 per 1000 live births) and under five mortality rate (64 per 1000 live births). The proportion of
under five children suffering from under-nutrition is still high: stunting (53%), wasting (17%) and
under-weight (52%) [11]. Exclusive breastfeeding has been found positively associated with infant
stature [12] and protective against overweight and obesity in childhood [13].
Currently, information on infant nutrition in Timor-Leste is lacking. A previous survey conducted
in 2003 [14] reported that breastfeeding was almost universal (97.6%), but a much lower exclusive
breastfeeding rate of 30.7% for infants less than 6 months. Some health infrastructures have been
restored after independence in 2002. For instance, 5 hospitals, 69 health centers and 85 health posts
have been established between 2002 and 2005. A national referral hospital and community health
centers were also functioning by this time with the support from the government and international
community. The per capita expenditure on health was US $45 in 2006 (7.5% of GDP), higher than
many other Asian countries [9]. As part of the global measure Demographic and Health Surveys
(DHS) project, an updated Timor-Leste DHS was conducted during 2009–2010 [11]. The aim of the
present study was to investigate factors associated with exclusive breastfeeding among infants aged
five months or less based on the updated information from the national DHS, findings from which will
enable policy makers and public health researchers to develop interventions to improve exclusive
breastfeeding in the country.
 75

2. Methods

2.1. Survey Design

DHS are conducted every five years in more than 50 countries using a validated questionnaire. The
Timor-Leste DHS 2009–2010, conducted in two stages, was the second survey after the initial 2003
DHS. During the first stage, 455 enumeration areas were selected based on probability proportionate to
size: 116 urban and 339 rural areas. More rural samples were included because the majority of the
Timor-Leste population lives in rural areas. At the second stage, 27 households were selected
randomly from each enumeration area following a systematic sampling procedure.

2.2. Participants

The final survey included 11,463 households, comprising 9806 children under five years of age.
The present study focused on the subgroup of 975 infants (1) with a singleton birth; (2) who were aged
less than six months; (3) alive and living with the respondent; and (4) who were the youngest child in
the family; in order to avoid the selection of children from the same household and parents. The DHS
was approved by the ethics committee of Macro International Inc. and the Ministry of Health of
Timor-Leste. The data were de-identified and made available for public use [15].

2.3. Exclusive Breastfeeding

The operational definition of exclusive breastfeeding, as defined by the World Health Organization
(WHO) infant feeding guidelines [2], was adopted: “infants 0–5 months of age who are fed exclusively
with breastmilk”. Apart from breastmilk or wet nurse’s milk, no other fluid was allowed, with the
exception of oral rehydration solution, drops or syrups (vitamins, minerals and medicine). The binary
status of exclusive breastfeeding (coded as ‘1’ for yes and ‘0’ for no) was determined for each of the
975 selected infants. Previous published studies have used such definition to report a period prevalence
of exclusive breastfeeding based on 24-h recall [16,17], but it should not be treated as the rate of
exclusive breastfeeding for six months.

2.4. Independent Variables

Selection and categorization of independent variables in this study were based on literature
review [14,18]. Maternal age was recoded into three groups: 15–19, 20–34 and 35–49 years.
Frequency of antennal care (ANC) visit was categorized as: 0, 1–3 and •4. Mother’s perceived size of
newborn was coded as: small, average and large. Religion was originally recorded as: Roman Catholic,
Muslim, Protestant, Hindu and others. Because the vast majority of population follows Roman
Catholic, the other religions were grouped together. Maternal occupation was re-categorized as: no
paid work (housewives and household work), agriculture (self-employed or employee), professional,
clerical, sales and services, and manual work (skilled or unskilled). Education level was classified as:
no education, primary, secondary and higher. Decision making on health had three categories:
respondent alone, respondent with others (e.g., husband), others only (e.g., husband or someone
else) [19]. Place of delivery was regrouped as either health facility (national hospital, referral hospitals,
 76

community health centers, health post, SisCa post, private sectors, Marrie stops) or home (home of
respondent or others). Birth order referred to: first time birth, second or third, and fourth or above.
Residential location was defined as either rural or urban.

2.5. Statistical Analysis

Timor-Leste is divided into 13 administrative districts, which has been incorporated into the
analysis to examine geographical differences in the 24-h period prevalence of exclusive breastfeeding
among infants aged <6 months. Further age-wise disaggregated proportions of exclusive breastfeeding
were reported at age <1 month, 1, 2, 3, 4, and 5 months. Factors associated with exclusive
breastfeeding were screened by Chi-square tests and then assessed by backward stepwise logistic
regression [20], taking into account the apparent collinearity between independent variables. Complex
sample analysis was performed to estimate proportions, odds ratios and their 95% confidence
intervals (CI) [21].

3. Results

3.1. Exclusive Breastfeeding

As shown in Table 1, of the 975 infants aged ”5 months, overall 49% (95% CI 45.4% to 52.7%)
were exclusively breastfed in the 24 h preceding the 2009–2010 survey. The exclusive breastfeeding
prevalence appeared to decline with increasing infant age, from 68.0% at less than one month to 24.9%
at five months. In particular, a sharp decrease was observed between the 4th and 5th month postpartum
for the respondents.

Table 1. Prevalence of exclusive breastfeeding by infant age, Timor-Leste, 2009–2010

(n = 975).
Infant Age Prevalence
n
(months) (95% Confidence Interval)
<1 80 68.0 (55.4, 78.5)
1 177 67.6 (59.2, 75.0)
2 151 56.5 (46.5, 66.1)
3 183 48.5 (40.5, 56.6)
4 209 41.8 (34.3, 49.7)
5 175 24.9 (19.5, 31.2)
”5 975 49.0 (45.4, 52.7)

3.2. Sample Characteristics

The distribution of maternal age (years) was: 15–19 (6.8%), 20–34 (67.5%), 35–49 (25.7%).
The majority (78.6%) of participants resided in rural areas. About one-third (34.6%) of mothers and
just over a quarter (28.5%) of fathers received no education. The majority of women had no paid
employment (73.2%) and followed Roman Catholic as their religion (98.2%). More of them came
from poorer (45.4%) than middle class (39.6%) and rich households (15%). Most respondents did not
read newspaper (70.3%), listen to the radio (59.9%), or watch television (68.6%) at all.
 77

Slightly more than half (52.7%) of mothers had high parity (• 4), with only 17.4% being first time
motherhood. Only 52% of the women had paid four or more ANC visits. There were more male
(52.9%) then female (47.1%) infants. The majority of infants were born at home (74.2%) by vaginal
delivery (98.3%). Surprisingly, only about a quarter (26.6%) of women could make decision by
themselves with regard to health-related matters. Although 56% of mothers perceived their newborn as
average size, 18.5% believed they were small size. Most mothers (88.8%) initiated breastfeeding
within one hour of delivery. Only a small proportion (2.9%) of mothers continued to smoke at the time
of the survey.
A number of socio-demographic and health-related variables appeared to be associated with the
prevalence of exclusive breastfeeding according to Chi-square tests. These included residential
location (P = 0.014), ecological region (P < 0.001), maternal occupation (P < 0.001), wealth status
(P = 0.003), frequency of listening to the radio (P = 0.007), frequency of watching television
(P = 0.004), and decision making on health (P = 0.026).

3.3. Factors Affecting Exclusive Breastfeeding

Stepwise logistic regression analysis further confirmed that infant age, ecological region, maternal
occupation, perceived size of newborn, and decision making on health were significantly associated
with exclusive breastfeeding; results of which are shown in Table 2. In particular, increase in infant
age could reduce the likelihood of exclusive breastfeeding, consistent with the previous observation on
the decline in exclusive breastfeeding prevalence by advancing infant age in Table 1. When compared
to the capital city Dili, mothers from other regions were more likely to exclusively breastfeed their
infants. On the other hand, mothers who maintained employment seemed less likely to continue
exclusive breastfeeding than their non-working counterparts. Those mothers who perceived their
newborn as either large or small size were also less likely to exclusively breastfeed. Finally, mothers
who could decide health-related matters by themselves tended to exclusively breastfeed, which was not
the case for others whose decisions were made by someone else.

4. Discussion

This study found that half (49.0%, 95% CI 45.4% to 52.7%) of the infants aged five months or
below were exclusively breastfed at the time of the 2009–2010 DHS, which appeared to increase
substantially from the previously reported 24-h recall prevalence rate of 30.77% (95% CI 27.2% to
34.5%) in 2003 [14]. According to the report by UNICEF [22], the proportion of exclusively breastfed
children of 0–5 months during the period 2000–2007 was 43% in East Asia and Pacific, 44% in
South Asia, and 39% overall in developing countries. However, the differences in survey period
between countries should be taken into account. The apparent increase in exclusive breastfeeding
prevalence may be attributable to a number of changes in Timor-Leste since 2003. The country has
become more stable after the conflict, with social and health services being restored [9,11]. While it is
encouraging to note the improvement in exclusive breastfeeding practice, the rate is still much lower
than the recommended 90% by the WHO [23].
 78

Table 2. Factors associated with exclusive breastfeeding in Timor-Leste, 2009–2010 (n = 975).

Adjusted Odds Ratio
Factor n (%) EBF (%) P*
(95% Confidence Interval) *
Infant age (months) Mean 2.81 SD 1.58 0.67 (0.60, 0.71) <0.001
Ecological region <0.001
Dili (Capital) 86 (6.3) 30 (33.2) 1.00
Aileu 83 (8.5) 61 (73.6) 8.05 (3.43, 18.89)
Ainaro 73 (7.5) 49 (67.2) 5.10 (2.49, 10.43)
Baucau 65 (6.7) 36 (52.3) 2.83 (1.23, 6.51)
Bobonaro 78 (8.0) 30 (39.4) 1.23 (0.62, 2.44)
Cova Lima 61 (6.3) 21 (34.0) 1.06 (0.51, 2.21)
Ermera 102 (10.5) 52 (50.4) 2.21 (1.10, 4.44)
Liquica 84 (8.6) 53 (63.3) 3.75 (1.90, 7.38)
Lautem 72 (7.4) 35 (46.5) 1.93 (1.03, 3.64)
Manufahi 63 (6.5) 39 (61.6) 3.80 (1.75, 8.28)
Manatuto 81 (8.1) 34 (41.5) 1.44 (0.76, 2.74)
Oecussi 84 (8.6) 50(59.5) 3.98 (2.21, 7.15)
Viqueque 43 (4.4) 24 (55.3) 2.26 (2.21, 7.15)
Maternal occupation 0.003
No paid work 712 (73.2) 394 (52.8) 1.00
Agriculture 176 (18.1) 91 (46.1) 0.68 (0.46, 1.02)
Professional, clerical,
67 (6.9) 23 (24.3) 0.33 (0.18, 0.62)
sales, services
Manual work 18 (1.8) 5 (34.4) 0.67 (0.19, 2.31)
Perceived size of newborn 0.009
Average 536 (56.0) 301 (53.0) 1.00
Small 177 (18.5) 90 (45.6) 0.61 (0.39, 0.93)
Large 244 (25.5) 114 (43.2) 0.58 (0.38, 0.88)
Decision making on health 0.023
Others only 106 (11.1) 43 (35.1) 1.00
Respondent alone 254 (26.6) 143 (52.1) 2.02 (1.11, 3.67)
Respondent with others 595 (62.3) 318 (50.6) 1.63 (0.89, 2.98)
EBF: exclusive breastfeeding. * From backward stepwise logistic regression; variables excluded were: maternal age,
residential location, maternal education, paternal education, religion, sex of infant, wealth status, frequency of reading
newspaper/magazine, frequency of listening radio, frequency of watching television, birth order, frequency of antenatal
care visit, maternal tobacco smoking, method of delivery, place of delivery.

The prevalence of exclusive breastfeeding declined with increasing infant age, from 68.0% at less
than one month to 24.9% at five months. The inverse association between infant age and exclusive
breastfeeding practice was also observed in other Asian countries such as Bangladesh, China and
Nepal [18,24,25]. According to the local culture, it is common that Timorese infants are introduced
complementary foods at about the 4th month. The decision is usually made by the senior women of the
family such as the grandmother or grandmother-in-law.
Mothers residing in Dili were less likely to breastfeed exclusively when compared with mothers
from other regions. Such regional differences have been reported by previous studies in Timor-Leste
and other Asian countries [14,17]. Dili is the capital and economic center of the country, where infant
 79

formulas are readily accessible at supermarkets. Besides, the capital city citizens are more exposed
to advertisement of infant formula, consequently leading to the early cessation of exclusive
breastfeeding [26].
Moreover, women who maintained employment after giving birth were less likely to provide
exclusive breastfeeding to their infants than their non-working counterparts. Similarly, Chinese
mothers who had to return to their office job before six months were unlikely to breastfeed their infant
exclusively [24]. Another qualitative study from Bangladesh reported that caretakers introduced
formula, cow or buffalo milk when mothers attended work [27]. Working mothers in Timor-Leste are
entitled to less than three months of maternity leave. This short duration makes it difficult to continue
exclusive breastfeeding.
Newborns perceived to be non-average size by their mothers were less likely to be exclusively
breastfed. Experience in other countries has similarly shown that preterm and low birth weight infants
are breastfed for shorter duration [28]. Mothers may experience a number of barriers to breastfeed
smaller infants, for instance, poor sucking, infants being kept separately for intensive care, illness, and
lack of confidence [29], which may lead to the early introduction of complementary foods.
In this study, Timorese mothers who could decide health-related matters tended to continue
exclusive breastfeeding, when compared with those that relied on the advice from someone else. This
finding was consistent with the literature, which suggested that the ability of a woman to make
decision on utilization of services can lead to better maternal and child health outcomes [19,30].
Several issues should be considered when interpreting the results. This study utilized the dataset
from the latest national survey with a representative sample and a high response rate, while complex
sample analysis was performed to account for the sampling strategy and sample weight [21].
Therefore, the findings are generalizable to the entire country. However, the 24-h recall would
inevitably induce over-reporting of exclusive breastfeeding at six months [31] so that caution should
be taken [2]. The DHS data nonetheless remain the only available information to estimate exclusive
breastfeeding rate in many developing countries.
There is an immediate need of breastfeeding promotion programs in Timor-Leste. Given the high
infant and child mortality in the country [11], improving the practice of exclusive breastfeeding will
reduce such burden and partially overcome the problem of under-nutrition. Antenatal counseling on
breastfeeding and peer support network are recommended [32]. Because the majority of births occur at
home, home visits by health workers/volunteers would be an effective option to consider by healthcare
planners to further promote exclusive breastfeeding and to increase its duration.

5. Conclusions

Slightly less than half the infants in Timor-Leste were exclusively breastfed within 24-h preceding
the latest national survey. This represented a significant improvement in exclusive breastfeeding
practice since 2003 when the country restored peace. Mothers should be provided with continuous
support to sustain their initial high rate of exclusive breastfeeding for six months. It is desirable to
target mothers who are working, who perceive their newborns as non-average size and those residing
in the capital Dili for breastfeeding promotion programs. In addition, mothers must be involved in the
decision making process so that they can sustain breastfeeding exclusively.
 80

Acknowledgments

The authors would like to thank ICF International (the Measure DHS program) for permission to
use the dataset for this study.

Conflicts of Interest

The authors declare no conflict of interest.

References

1. Binns, C.W.; Fraser, M.L.; Lee, A.H.; Scott, J. Defining exclusive breastfeeding in Australia.
J. Paediatr. Child. Health 2009, 45, 174–180.
2. World Health Organization. Indicators for Assessing Infant and Young Child Feeding Practices
Part. 1. Definitions; World Health Organization: Geneva, Switzerland, 2008.
3. Kramer, M.S.; Kakuma, R. Optimal duration of exclusive breastfeeding. Cochrane Database
Syst. Rev. 2012, doi:10.1002/14651858.CD003517.pub2.
4. Ip, S.; Chung, M.; Raman, G.; Trikalinos, T.A.; Lau, J. A summary of the Agency for Healthcare
Research and Quality’s evidence report on breastfeeding in developed countries. Breastfeed. Med.
2009, 4, S17–S30.
5. Weimer, J. The Economic Benefits of Breastfeeding: A Review and Analysis; USDA Economic
Research Service: Washington, DC, USA, 2001.
6. Black, R.E.; Victora, C.G.; Walker, S.P.; Bhutta, Z.A.; Christian, P.; de Onis, M.; Ezzati, M.;
Grantham-McGregor, S.; Katz, J.; Martorell, R.; et al. Maternal and child undernutrition and
overweight in low-income and middle-income countries. Lancet 2013, 382, 427–451.
7. Duong, D.V.; Binns, C.W.; Lee, A.H. Breast-feeding initiation and exclusive breast-feeding in
rural Vietnam. Public Health Nutr. 2004, 7, 795–799.
8. Senarath, U.; Agho, K.; Akram, D.; Godakandage, S.; Hazir, T.; Jayawickrama, H.; Joshi, N.;
Kabir, I.; Khanam, M.; Patel, A.; et al. Comparisons of complementary feeding indicators and
associated factors in children aged 6–23 months across five South Asian countries. Matern. Child.
Nutr. 2012, 8, 89–106.
9. Saikia, U.; Hosgelen, M. Timor-Leste’s demographic destiny and its implications for the health
sector by 2020. J. Popul. Res. 2010, 27, 133–146.
10. Wild, K.J. Maternity waiting homes and the shaping of maternal health policy in Timor-Leste.
Ph.D. Thesis, Charles Darwin University, Darwin, Australia, 2009.
11. National Statistics Directorate (NSD); Ministry of Finance; ICF Macro. Timor-Leste
Demographic and Health Survey 2009–2010; NSD: Dili, Timor-Leste; ICF Macro: Calverton,
MD, USA, 2010.
12. Martin, R.M.; Smith, G.D.; Mangtani, P.; Frankel, S.; Gunnell, D. Association between breast
feeding and growth: the Boyd-Orr cohort study. Arch. Dis. Child. Fetal Neonatal Ed. 2002, 87,
F193–F201.
 81

13. Gillman, M.W.; Rifas-Shiman, S.L.; Camargo, C.A., Jr.; Berkey, C.S.; Frazier, A.L.;
Rockett, H.R.; Field, A.E.; Colditz, G.A. Risk of overweight among adolescents who were
breastfed as infants. JAMA 2001, 285, 2461–2467.
14. Senarath, U.; Dibley, M.; Agho, K. Breastfeeding practices and associated factors among children
under 24 months of age in Timor-Leste. Eur. J. Clin. Nutr. 2006, 61, 387–397.
15. Measure DHS: Demographic and Health Surveys. Available online: http://www.measuredhs.com/
what-we-do/survey/survey-display-356.cfm (accessed on 1 November 2013).
16. Agho, K.E.; Dibley, M.J.; Odiase, J.I.; Ogbonmwan, S.M. Determinants of exclusive
breastfeeding in Nigeria. BMC Pregnancy Childbirth 2011, 11, doi:10.1186/1471-2393-11-2.
17. Senarath, U.; Dibley, M.J.; Agho, K.E. Factors associated with nonexclusive breastfeeding in
5 east and southeast Asian countries: A multilevel analysis. J. Hum. Lact. 2010, 26, 248–257.
18. Khanal, V.; Sauer, K.; Zhao, Y. Exclusive breastfeeding practices in relation to social and health
determinants: A comparison of the 2006 and 2011 Nepal Demographic and Health Surveys.
BMC Public Health 2013, 13, 958.
19. Titaley, C.; Dibley, M.; Roberts, C. Factors associated with underutilization of antenatal care
services in Indonesia: Results of Indonesia Demographic and Health Survey 2002/2003 and 2007.
BMC Public Health 2010, 10, 485.
20. Sun, G.-W.; Shook, T.L.; Kay, G.L. Inappropriate use of bivariable analysis to screen risk factors
for use in multivariable analysis. J. Clin. Epidemiol. 1996, 49, 907–916.
21. West, B.T. Statistical and methodological issues in the analysis of complex sample survey data:
Practical guidance for trauma researchers. J. Trauma. Stress 2008, 21, 440–447.
22. Overview of Breastfeeding Patterns. Available online: http://www.childinfo.org/breastfeeding_
overview.html (accessed on 29 August 2012).
23. World Health Organization. Infant and Young Child Feeding: Model Chapter for Textbooks for
Medical Students and Allied Health Professionals; World Health Organization: Geneva,
Switzerland, 2009.
24. Xu, F.; Binns, C.; Zheng, S.; Wang, Y.; Zhao, Y.; Lee, A. Determinants of exclusive
breastfeeding duration in Xinjiang, PR China. Asia Pac. J. Clin. Nutr. 2007, 16, 316–321.
25. Mihrshahi, S.; Kabir, I.; Roy, S.; Agho, K.E.; Senarath, U.; Dibley, M.J. Determinants of infant
and young child feeding practices in Bangladesh: Secondary data analysis of Demographic and
Health Survey 2004. Food Nutr. Bull. 2010, 31, 295–313.
26. Declercq, E.; Labbok, M.H.; Sakala, C.; O’Hara, M. Hospital Practices and Women’s Likelihood
of Fulfilling Their Intention to Exclusively Breastfeed. Am. J. Public Health 2009, 99, 929–935.
27. Haider, R.; Rasheed, S.; Sanghvi, T.G.; Hassan, N.; Pachon, H.; Islam, S.; Jalal, C.S.
Breastfeeding in infancy: Identifying the program-relevant issues in Bangladesh. Int. Breastfeed. J.
2010, 5, 21.
28. Killersreiter, B.; Grimmer, I.; Bührer, C.; Dudenhausen, J.W.; Obladen, M. Early cessation of
breast milk feeding in very low birthweight infants. Infant Child. Dev. 2001, 60, 193–205.
29. Mulready-Ward, C.; Sackoff, J. Outcomes and Factors Associated with Breastfeeding for
<8 Weeks among Preterm Infants: Findings from 6 States and NYC, 2004–2007. Matern. Child
Health J. 2013, 17, 1648–1657.
 82

30. Shroff, M.; Griffiths, P.; Adair, L.; Suchindran, C.; Bentley, M. Maternal autonomy is inversely
related to child stunting in Andhra Pradesh, India. Matern. Child Nutr. 2009, 5, 64–74.
31. Binns, C.; Lee, A.; Sauer, K.; Hewitt, K. Reported Breastfeeding Rates in the Asia-Pacific
Region. Curr. Pediatr. Rev. 2012, 8, 339–345.
32. Lawrence, R.A. Promotion of Breastfeeding Intervention Trial (PROBIT) a randomized trial in
the Republic of Belarus. J. Pediatr. 2001, 139, 164–165.

© 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).
 83

Reprinted from Nutrients. Cite as: Dashti, M.; Scott, J.A.; Edwards, C.A.; Al-Sughayer, M. Predictors
of Breastfeeding Duration among Women in Kuwait: Results of a Prospective Cohort Study. Nutrients
2014, 6, 711-728.

Predictors of Breastfeeding Duration among Women in Kuwait:

Results of a Prospective Cohort Study
Manal Dashti 1,2, Jane A. Scott 3,*, Christine A. Edwards 1 and Mona Al-Sughayer 4
1
Human Nutrition, School of Medicine, College of Medical Veterinary and Life Sciences,
University of Glasgow, Glasgow G31 2ER, UK; E-Mails: manaldashti@yahoo.com (M.D.);
christine.edwards@glasgow.ac.uk (C.A.E.)
2
Nutrition Unit, Jaber Al-Ahmed Armed Forces Hospital, Ministry of Defense,
Sabhan 46004, Kuwait
3
School of Public Health, Curtin University, Perth 6102, Australia
4
Department of Biological Sciences, Faculty of Science, Kuwait University, Safat 13060, Kuwait;
E-Mail: mona.alsughayer@ku.edu.kw

* Author to whom correspondence should be addressed; E-Mail: jane.scott@curtin.edu.au;

Tel.: +61-8-9266-9050; Fax: +61-8-9266-2958.

Received: 20 January 2014; in revised form: 28 January 2014 / Accepted: 8 February 2014 /
Published: 20 February 2014

Abstract: The purposes of this paper are to report the prevalence of breastfeeding to
six months among women in Kuwait and to determine the factors that are associated with
the duration of breastfeeding. A cohort of 373 women recruited from maternity wards in
four hospitals in Kuwait city were followed from birth to 26 weeks postpartum. The
association of any and full breastfeeding duration and predictor variables were explored
using multivariate Cox’s proportional hazards models. At six months, 39% of all infants
were receiving some breast milk and only 2% of infants had been fully breastfed to
26 weeks. Women born in other Arab countries were less likely to discontinue
breastfeeding than women born in Kuwait. Other factors positively associated with
breastfeeding duration were level of maternal education, higher parity, infant being
demand fed in hospital and a preference for breastfeeding on the part of the infant’s father
and maternal grandmother. The introduction of a pacifier before four weeks of age and the
mother intending to return to work by six months were negatively associated with duration.
These findings present a number of opportunities for prolonging breastfeeding duration
in Kuwait.
 84

Keywords: breastfeeding; duration; determinants; Middle East

1. Introduction

Breastfeeding is an unequalled way to feed an infant. In addition to its unique nutritional properties,
human breast milk contains a wide-variety of immunoprotective factors that augment the immature
immune system of the infant [1]. Infants who are formula fed are at greater risk of infections common
to infancy including gastroenteritis, respiratory infection and otitis media [2]. The World Health
Organization (WHO) and the United Nations Children’s Fund (UNICEF) recommend that infants be
exclusively breastfed for the first six months of life with breastfeeding continuing for up to two years
of age or beyond [3]. The wide-spread practice of delayed initiation of breastfeeding and prelacteal
feeding [4–8], along with the early introduction of complementary feeding [9] in the Middle Eastern
region however, mean that very few infants in this region are exclusively breastfed from birth for
six months as recommended.
Breastfeeding practices are influenced by a complex mix of factors which are related to maternal
and family socio-demographic characteristics, biomedical factors, health-care practices, psychosocial
factors, social support, community attitudes, and public policy factors [10,11]. The direction of effect
of these factors is not consistent across all cultures. For instance, in industrialised countries, better
educated women are more likely to initiate breastfeeding and to breastfeed for longer than their less
educated counterparts, whereas in poorer countries the opposite tends to be the case. In common with
industrialised countries [10–12], amongst Middle Eastern women breastfeeding duration has been
positively associated with maternal age [13–16] and parity [14,17,18]. Whereas inconsistent
associations have been reported for level of maternal education with breastfeeding duration being
associated both negatively [16,19,20] and positively [17] with a higher level of maternal education.
Other factors reported to be negatively associated with duration of breastfeeding include maternal
employment [17,19–22], mode of delivery [21,23–25] and the use of infant formula while in
hospital [24–26].
Regular breastfeeding surveillance is essential to determine the extent to which national
breastfeeding targets are being met, the impact of breastfeeding promotion interventions and how
breastfeeding practices are changing over time. In addition, it is important to investigate the
determinants of infant feeding practices so that breastfeeding interventions can be targeted at the most
vulnerable population groups and address potentially modifiable risk factors which adversely affect
breastfeeding practices. Relatively few studies have investigated infant feeding practices in
Kuwait [27–29] and none have been longitudinal in nature. The reported mean duration of
breastfeeding appears to have declined from 6.4 months in 1988 [27] to 4.9 months in 1997 [28] and
there is a lack of more recent data to determine if this downward trend has continued. The aim of the
Kuwait Infant Feeding Study (KIFS) was to identify the incidence and prevalence of breastfeeding up
to 26 weeks postpartum among a population of women living in Kuwait and to identify the factors
associated with the initiation and duration of breastfeeding. The determinants of breastfeeding
initiation have been reported previously [30] and the purposes of this paper, therefore, are to report the
 85

prevalence of breastfeeding to six months and to determine the factors that are associated with the
duration of breastfeeding.

2. Experimental Section

2.1. Recruitment of Subjects

A prospective cohort study of infant feeding practices among women in Kuwait was conducted
between October 2007 and October 2008. The study methods have been described previously [30], but
briefly mothers were recruited from three major public hospitals and one private hospital located in
Kuwait City. Within 72 h of delivery, eligible mothers were visited and invited to participate in the
study by the researcher (MD) who provided a written and verbal description of the study. Women were
considered to be eligible for the study if they were able to read or understand Arabic or English and
had delivered a live, healthy, singleton of 36 weeks or more gestational age. Mothers whose infants
were admitted to the Special Care Nursery (SCN) for minor illnesses or observation were eligible
for recruitment.
The study was approved by the Medical Faculty Ethics Committee of the University of Glasgow,
(Application No. FM03906: Approved May 29, 2007) and by the Ministry of Health in Kuwait.
Participants provided signed informed consent and were advised that they could refuse to participate or
withdraw from the study at any time, without prejudicing their post-natal care or the care of their baby.

2.2. Data Collection

Mothers who agreed to join the study were interviewed face-to-face to complete a baseline
questionnaire prior to discharge from hospital. Women who declined to participate were asked to
provide some basic socio-demographic data to determine if the sample were representative of the
population of women giving birth at the participating hospitals. All participants were followed up by
telephone interview at 6, 12, 18 and 26 weeks postpartum. Data were collected using questionnaires
previously used in similarly designed studies of Australian women [31] and modified slightly to meet
the needs of this study population. Information on socio-demographic characteristics, maternal
lifestyle factors, infant characteristics, biomedical factors, hospital practices, psychosocial factors and
feeding practices were collected at baseline. Information on current feeding practices, changes to
feeding practices, breastfeeding experiences and introduction of a pacifier were collected during
follow-up interviews.

Infant Feeding Assessments

Breastfeeding terms used in this study were those defined by the World Health Organization [3].
An infant was considered to be exclusively breastfed when he or she had received only breast milk
with no other liquids (except for oral rehydration solutions (ORS), drops or syrups) or complementary
(solid) foods, and to be predominantly breastfed when he or she received breast milk as the main
source of nourishment, with certain liquids (water, water-based fluids, fruit juices, ritual fluids, ORS,
drops or syrups) but received no other liquids (including formula milk and non-human milks)
or complementary foods. Full breastfeeding was defined as either exclusive or predominant
 86

breastfeeding [32]. Any breastfeeding was defined as an infant who was receiving breast milk, with or
without formula, other milks, fluids or complementary foods. Duration of exclusive, predominant and
any breastfeeding was determined by using information about the age at which other types of milks,
liquids (e.g., water, fruit juice) and/or complementary foods were introduced in the first
six months of life. In this study, prelacteal feeding was defined as the act of giving any liquid or food
item (except breast milk) to a newborn within the first three days after birth [33].

2.3. Statistical Analysis

We explored the associations of breastfeeding duration and a variety of characteristics and practices
reported in the literature to be associated with the duration of breastfeeding using Cox’s proportional
hazards model. This model allows joint estimation of the effects of independent variables on the risk of
cessation of breastfeeding and can be used to analyze data that contain censored observations [34]. We
tested the role of: (1) socio-demographic factors (maternal age, education, country of birth,
employment plans for six month postpartum); (2) maternal lifestyle factors (pre-pregnancy BMI
calculated from self-reported weight and height, smoking during pregnancy); (3) infant factors
(gender, having spent time in the SCN); (4) biomedical factors (parity, delivery method, breastfeeding
problems at baseline, breastfeeding problems at six weeks postpartum, age at which pacifier was
introduced) (5) hospital practices (time to first breastfeed, composition of infant’s first feed, use of
prelacteal feeds, infant roomed-in, infant demand fed in hospital); and (6) psychosocial factors (when
decided on feeding method, whether pregnancy was planned, father’s feeding preference, maternal
grandmother’s feeding preference, paternal grandmother’s feeding preference). The association of
individual variables and the duration of any and full breastfeeding was first evaluated in a univariate
model. Any variable with a P-value of <0.100 was then included in a multivariate model which was
reduced using the backward stepwise procedure. The fitness of each model was assessed at every step
to avoid dropping non-significant variables that affected the model fitness. All variables in the final
model were variables for which, when excluded, the change in deviance compared with the
corresponding statistics on the relevant degrees of freedom was significant. Statistical analyses were
performed using the Statistical Package of Social Sciences version 19 (SPSS Inc., Chicago, IL, USA).

3. Results

A total of 439 women were invited to participate in the study and 373 mothers completed the
baseline questionnaire while in hospital, giving a response rate of 85%. There were no significant
differences between participants and those declining to participate (n = 66) with respect to age
(Ȥ2 4.413, P = 0.110), level of education (Ȥ2 2.455, P = 0.117) and chosen method of feeding at
discharge (Ȥ2 447, P = 0.800), suggesting that the sample was representative of the population from
which it was drawn [30]. In all, 80 women dropped out of the study prior to completing the final
follow-up interview at 26 weeks however, there were no differences in the age, level of education and
chosen feeding method of those who completed or withdrew from the study (data not reported). Data
for the duration of any and full breastfeeding for women who withdrew were censored in accordance
with the woman’s status at the time of last contact, allowing all participants to be included in the
survival analysis.
 87

Almost all women (92.5%) initiated breastfeeding, and at six months, just over one third of all
infants (39%) were receiving some breast milk and only 2% of infants had been fully breastfed to
26 weeks (Table 1). As not all women had ceased breastfeeding it was not possible to estimate mean
duration of breastfeeding, however the median duration of any breastfeeding was 13.9 weeks and the
mean duration of those women who had stopped breastfeeding before 26 weeks was six weeks. The
median duration of full breastfeeding was less than one week with 50% of infants having received
formula feeds within the first week post-partum.

Table 1. Prevalence a of full and any breastfeeding at selected ages.

Age (weeks) Any Breastfeeding (%) Full Breastfeeding (%)
4 66 31
8 56 26
12 53 22
26 39 2
a
Survival analysis using censored cases.

The majority of women who initiated breastfeeding (n = 345) were between 25 and 34 years of age
(64.3%), born in Kuwait (54.2%) and had completed 12 or more years of education (78%). Just over
one third of women had delivered by Caesarean section (36.5%). Delayed initiation of breastfeeding
and prelacteal feeding were the norm amongst this cohort with just over one half of women (52.8%)
initiating breastfeeding more than 24 h after giving birth and the majority of infants (88.7%) receiving
formula as either their first feed and/or at some time during the first three days of their hospital stay
(Table 2). Only one third of women who initiated breastfeeding left hospital fully breastfeeding their
infant, with the majority of women (59.4%) partially breastfeeding. A small number of women
(n = 29, 8.4%) who initiated breastfeeding left hospital exclusively formula feeding.

3.1. Univariate Analysis

Due to the small number of breastfed infants (11.3%) who had been exclusively breastfed whilst in
hospital we investigated the association of covariates with the duration of full and any breastfeeding
for the first six months of life. In the univariate analysis (Table 2), the duration of any and/or full
breastfeeding was associated with socio-demographic factors (maternal age, country of birth and level
of education), parity, prelacteal feeding, introduction of a pacifier, whether a mother had roomed-in for
24 h with her infant and psychosocial factors (partner’s and maternal grandmother’s support
for breastfeeding).

3.2. Multivariate Analysis

Table 3 shows the results of the multivariate analysis. After adjustment, maternal level of education
and country of birth were independently associated with both duration of full and any breastfeeding.
Mothers with 12 or more years of education were less likely to stop any (Adjusted Hazard Ratio
(AdjHR) = 0.68) and full (AdjHR = 0.74) breastfeeding during the six month follow-up period
compared with mothers with less than 12 years of education. Similarly, mothers born in other Arab
countries were less likely to stop any (AdjHR = 0.53) and full (AdjHR = 0.65) breastfeeding compared
 88

with women born in Kuwait. Mothers who did not intend to return to work within six months
post-partum (AdjHR = 0.76) and those who did not experience breastfeeding problems in hospital
(AdjHR = 0.80) were less likely to have stopped full breastfeeding. Conversely, women who did not
feed on demand while in hospital (AdjHR = 1.28) or whose partner preferred formula feeding or was
ambivalent as to how his child was fed (AdjHR = 1.33) were more likely to stop full breastfeeding.
Multiparous women (Adj HR = 0.63) were less likely to cease any breastfeeding while those women
who introduced a pacifier to their infant before four weeks (AdjHR = 1.66) or whose own mother
preferred formula feeding or was ambivalent as to how her grandchild was fed (AdjHR = 2.11) were
more likely to stop breastfeeding during the six month follow-up period.

3.3. Reasons for Discontinuing Breastfeeding

The reasons given by mothers for stopping breastfeeding are given in Table 4. The majority of
women (86.8%) indicated that they were concerned about the adequacy of their breast milk in terms of
either quantity or quality. Almost half (49.1%) indicated that their baby had either weaned them self,
preferred a bottle or were ready for solids. A notable proportion stopped breastfeeding because they
had returned to work or study. Only a small number of women cited mother-centered reasons like
inconvenience and dislike of breastfeeding. The reasons for cessation did not vary markedly according
to the infant age at which women stopped breastfeeding.

4. Discussion

While breastfeeding initiation is virtually universal amongst women living in Kuwait, targets for
breastfeeding duration are not being met, with no woman in this study exclusively breastfeeding to six
month of age and only 2% of women fully breastfeeding their infants to this age. As not all women had
ceased breastfeeding by 26 weeks, it was not possible to estimate mean duration of breastfeeding
however, the median duration of any breastfeeding was slightly more than three months and the mean
duration of those women who had stopped breastfeeding before 26 weeks was six weeks. This suggests
that the mean duration of any breastfeeding in Kuwait is likely to be less than the 4.9 months reported
in 1997 [28] and that breastfeeding duration is declining.
In this study, women born in other Arab countries were less likely to have discontinued any and full
breastfeeding than women born in Kuwait or other countries. These women are likely to be the wives
of Middle Eastern guest workers employed in the oil and construction industries and the infant feeding
practices of these women likely reflect those of their home country where children are breastfed for
longer than infants in Kuwait. For instance, contemporaneous studies have reported a mean duration of
breastfeeding of 8.6 months in the UAE [16] and 7.6 months in Bahrain [35], and a median duration of
12.4 months in Jordan [36]. A recent study in Kuwait reported that Kuwaiti mothers use bottle feeding
more than the non-Kuwaiti mothers [37] which is consistent with our findings. Al Fadli et al. [37]
proposed that such practice could be explained by the lifestyle changes that occurred in Kuwait due to
oil revenue and through using modern technology similar to what happened in Western countries in the
1960s and 1970s.
 89

Table 2. Characteristics of a cohort of mother-infant dyads who had breastfed (n = 345) and the unadjusted association with the risk of
discontinuing any or full breastfeeding during the six months’ follow-up (Crude Hazards Ratio and 95% Confidence Interval).
Any Breastfeeding Full Breastfeeding
Characteristic Na %
Crude HR 95% CI P value Crude HR 95% CI P value
Maternal Factors
Age (years) 0.018 0.380
<25 77 22.3 1.36 0.83–2.26 0.95 0.66–1.38
25–34 222 64.3 0.82 0.52–1.29 0.83 0.61–1.14
• 46 13.3 1.00 1.00
Mother’s country of birth <0.001 0.002
Kuwait & Gulf States 187 54.2 1.00 1.00
Other Arab countries b 119 34.5 0.44 0.31–0.63 0.66 0.52–0.83
Other non-Arab Countries 39 11.3 0.77 0.48–1.23 0.92 0.65–1.30
Years of education 0.002 0.090
<12 76 22.0 1.00 1.00
• 269 78.0 0.59 0.42–0.83 0.80 0.62–1.04
Employment intention for 6 months postpartum 0.156 0.055
Intend to be working 136 39.4 1.00 1.00
Do not intend to be working or don’t know 209 60.6 0.80 0.59–1.09 0.81 0.65–1.00
Smoked pregnancy 0.358 0.706
Yes 20 5.8 1.00 1.00
No 325 94.2 0.73 0.37–1.43 0.92 0.58–1.44
Pre-pregnancy Body Mass Index (kg/m2) 0.623 0.964
<24.99 149 43.2 1.00 1.00
25.00 to 29.99 124 35.9 1.02 0.72–1.45 1.03 0.81–1.30
• 72 20.9 1.21 0.81–1.82 0.99 0.75–1.31
 90

Table 2. Cont.
Infant factors
Gender 0.962 0.709
Male 176 51.0 1.00 1.00
Female 169 49.0 0.99 0.73–1.35 1.04 0.84–1.29
Spent time in Special Care Nursery 0.800 0.148
Yes 61 17.7 1.00 1.00
No 284 82.3 1.05 0.70–1.58 0.81 0.62–1.08
Biomedical factors
Parity 0.014 0.081
Primiparous 112 32.5 1.00 1.00
Multiparous 233 67.5 0.67 0.49–0.92 0.82 0.65–1.03
Delivery method 0.690 0.723
Vaginal 219 63.5 1.00 1.00
Caesarean Section 126 36.5 1.07 0.78–1.46 1.04 0.84–1.30
Breastfeeding problems in hospital 0.100 0.045
Yes 126 36.5 1.00 1.00
No 219 63.5 0.77 0.56–1.05 0.80 0.64–1.00
Breastfeeding problems at 6 weeks postpartum 0.165 0.402
Yes 153 44.3 1.00 1.00
No 192 55.7 1.25 0.91–1.70 1.10 0.89–1.36
Age pacifier introduced <0.001 0.244
<4 weeks 113 32.8 1.99 1.44–2.76 1.30 0.96–1.75
At or after 4 weeks 26 7.5 1.36 0.78–2.37 1.05 0.82–1.35
Not using a pacifier at 26 weeks 206 59.7 1.00 1.00
Hospital practices
Time to first breastfeed 0.158 0.424
Within 6 h of delivery 81 23.5 1.37 0.83–2.25 1.18 0.86–1.63
Between 6 and 24 h 68 19.7 1.50 0.99–2.26 1.18 0.91–1.54
More than 24 h after delivery 182 52.8 1.00 1.00
 91
Table 2. Cont.
Missing 14 4.1 - -
Infant’s first feed 0.162 0.336
Formula/other 277 80.3 1.00 1.00
Breast milk/colostrum 68 19.7 0.75 0.49–1.13 0.89 0.68–1.15
Prelacteal feed given 0.008 0.031
Yes 306 88.7 1.00 1.00
No 39 11.3 0.42 0.22–0.80 0.69 0.50–0.97
Infant roomed-in for 24 h/day 0.183 0.401
Yes 183 53.0 1.00 1.00
No 162 47.0 0.81 0.59–1.10 1.10 0.89–1.36
Infant fed on demand in hospital 0.042 0.064
Yes 245 71.0 1.00 1.00
No 100 29.0 1.41 1.01–1.95 1.25 0.99–1.58
Psycho-social factors
When decided how to feed infant 0.884 0.089
Before pregnant 262 75.9 1.00 1.00
After pregnant 83 24.1 0.97 0.66–1.43 1.24 0.97–1.59
Planned pregnancy 0.155 0.085
Yes 196 56.8 1.00 1.00
No 149 43.2 0.80 0.58–1.09 0.83 0.67–1.03
Father’s infant feeding preferences 0.008 0.007
Prefers breastfeeding 280 81.2 1.00 1.00
Prefers Bottle or ambivalent 65 18.8 1.64 1.14–2.36 1.46 1.11–1.92
Maternal grandmother’s infant feeding preference 0.004 0.013
Prefers breastfeeding 312 90.4 1.00 1.00
Prefers Bottle or ambivalent 33 9.6 2.06 1.26–3.36 1.59 1.10–2.28
Paternal grandmother’s infant feeding preference 0.448 0.400
Prefers breastfeeding 261 75.7 1.00 1.00
Prefers Bottle or ambivalent 84 24.3 1.15 0.81–1.63 1.11 0.87–1.42
a
N = Number of subjects; b.Other Arab countries included Algeria, Egypt, Iran, Iraq, Jordan, Lebanon, Morocco, Palestine, Qatar, Saudi Arabia, Syria, Yemen.
 92

Table 3. Factors independently associated with the risk for discontinuing any or full breastfeeding in a cohort of mother-infant dyads followed
to six months postpartum (n = 345).
Any Breastfeeding Full Breastfeeding
Characteristic a b
Adjusted HR 95% CI P Value Adjusted HR 95% CI P Value
Mother’s country of birth 0.003 0.001
Kuwait & Gulf States 1.00 1.00
Other Arab countries c 0.53 0.36–0.76 0.65 0.51–0.83
Other non-Arab Countries 0.75 0.46–1.22 0.93 0.65–1.33
Years of education 0.030 0.030
<12 1.00 1.00
• 0.68 0.47–0.96 0.74 0.56–0.97
Employment intention for 6 months postpartum 0.022
Intend to be working NS d 1.00
Do not intend to be working or don't know 0.76 0.60–0.96
Parity 0.005
Primiparous 1.00 NS
Multiparous 0.63 0.46–0.87
Age pacifier introduced 0.013
<4 weeks 1.66 1.18–2.33 NS
At or after 4 weeks 1.25 0.71–2.20
Not using a pacifier at 26 weeks 1.00
Prelacteal feed given 0.111
Yes 1.00 NS
No 0.59 0.31–1.13
Breastfeeding problems in hospital 0.046
Yes 1.00
No NS 0.80 0.64–0.99
Demand fed in hospital 0.040
Yes NS 1.00
No 1.28 1.01–1.62
 93

Table 3. Cont.
Father’s infant feeding preferences 0.045
Prefers breastfeeding NS 1.00
Prefers Bottle or ambivalent 1.33 1.01–1.77
Maternal grandmother’s infant feeding preference 0.005
Prefers breastfeeding 1.00 NS
Prefers Bottle or ambivalent 2.11 1.26–3.54
í/RJ/LNHOLKRRG 1681.143 í/RJ/LNHOLKRRG 
a
Adjusted for mother’s age, country of birth, level of education, parity, infant received prelacteal feed, age pacifier introduced, father’s and maternal grandmother’s infant
b
feeding preference; Adjusted for mother’s country of birth, level of education, employment intention for six months post-partum, parity, breastfeeding problems at
baseline, infant received prelacteal feed, infant demand fed in hospital, when infant feeding decision was made, whether pregnancy was planned, father’s and maternal
c
grandmother’s infant feeding preference; Other Arab countries included Algeria, Egypt, Iran, Iraq, Jordan, Lebanon, Morocco, Palestine, Qatar, Saudi Arabia,
d
Syria, Yemen. NS = non-significant.

Table 4. Reasons a for stopping breastfeeding.

Stopped < 4 weeks Stopped 4–12 weeks Stopped 13–26 weeks Total
Reason (N b = 51) (N = 68) (N = 40) (N = 159)
N % N % N % N %
Concerned about quantity and quality of breast milk 42 82.4 59 86.8 37 92.5 138 86.8
Baby weaned self, prefers bottle or ready for solids 24 47.1 31 45.6 23 57.5 78 49.1
Returned to work or study 24 47.1 24 35.3 19 47.5 67 42.1
Breastfeeding too difficult or requires too much motivation 4 7.8 8 11.8 2 5.0 14 8.8
Mother ill, stressed or too tired 3 5.8 4 5.9 3 7.5 10 6.3
Mother-centered reasons (breastfeeding inconvenient,
dislike breastfeeding, concern for effect on figure, 3 5.9 3 4.4 3 7.5 9 5.7
“done my bit”)
Breast-related problems (cracked or sore nipples,
1 2.0 1 1.5 2 5.0 4 2.5
engorgement, mastitis)
a
Women may have given more than one reason for stopping; b N = number of subjects.
 94

This study found no independent association with maternal age but, consistent with the findings of
studies of women in Western countries [11,12], breastfeeding duration was positively associated with
level of maternal education and parity, and negatively associated with maternal employment. Women
with 12 or more years of education were less likely to have discontinued any or full breastfeeding
compared with women with less than 12 years of education. Multiparous women were less likely to
discontinue any breastfeeding than primiparous women which is consistent among women from
Western [11,12] and other Middle-Eastern [14,18,22,23] countries. Previous breastfeeding success is a
strong predictor of breastfeeding duration [12] and in general women breastfeed for longer with each
successive pregnancy.
While there was no association with the duration for any breastfeeding, women who did not
intend to return to work within 26 weeks were less likely to have discontinued full breastfeeding than
those who planned to return to work. This suggests that women supplement breastfeeding with formula
feeding either on return to work or in preparation for a return to work. This negative association
between early return to work and breastfeeding duration has been reported in studies of women from
other Middle-Eastern countries [17,19,20,22] and is consistent with studies of women from Western
countries [31,38–40], and suggests that women everywhere have difficulty combining working with
exclusive breastfeeding.
Other factors negatively associated with the continuation of full breastfeeding were whether prior to
discharge the mother had experienced breastfeeding problems or her infant had been fed on demand,
both of which may be inter-related. Milk production is directly related to suckling frequency [41] and
there is evidence that fixed feeding schedules lead to insufficient milk supply and breastfeeding
problems [42]. These findings highlight the importance of unrestricted breastfeeding in the early post-
partum period to the successful establishment of breastfeeding. Hospital staff should encourage
demand feeding and support and encourage women to persevere when they are experiencing difficulty
establishing breastfeeding rather than resorting quickly to supplementing breast milk with formula.
This was the first study to investigate the association between pacifier use and breastfeeding
duration among Middle Eastern women. The incidence of pacifier use amongst breastfeeding women
in Kuwait (40%) was approximately half that reported for women in Australia [43,44] and the USA [45],
and introduction of a pacifier before four weeks of age was found to be negatively associated with any
breastfeeding duration which is consistent with the international literature [46]. While the mechanism
remains unclear, it has been suggested that the non-nutritive sucking on a pacifier reduces the
frequency of nutritive sucking from the breast, thereby leading to less stimulation of the breast and
consequently less milk production [43,45].
Finally, this study highlights the importance of social support for breastfeeding. Support can come
from a woman’s partner, family and friends, and the degree to which each of these groups influences a
woman’s decision to breastfeed varies according to the mother’s age, social class and cultural or ethnic
background [47]. In traditional societies, women rely more on the advice and support of their mother,
whereas in Western cultures they are more likely to identify their husband as their main source of
support [48]. This and other studies of Muslim women have highlighted the importance of grandmothers
both in providing practical support and as major influences on infant feeding decisions [8,49]. Advice
received from their mother and mother-in-law can have both a negative and positive affect on a
woman’s breastfeeding practices. For instance, on one hand, breastfeeding is promoted in the Quran
 95

(Al Baqara, 233) and by elders as the desired way to feed an infant, and the mean duration of
breastfeeding is longer in most Muslim countries than in Western countries. On the other hand, there is
a common perception amongst older women that the heavier the baby the healthier he or she is. There
is anecdotal evidence that Kuwaiti grandmothers often encourage topping up with formula to ensure
the baby is satiated and to stop hunger cries, which explains in this study the positive association with
any breastfeeding but not full breastfeeding.
This study also showed that a husband’s preference for breastfeeding over formula feeding was
positively associated with breastfeeding initiation [30] and longer duration of full-breastfeeding which
is consistent with Western studies [31,50,51]. To the best of our knowledge, no Middle-Eastern study
has investigated previously the association of paternal attitudes and breastfeeding duration. There is,
however, some evidence from Middle Eastern studies that support from a woman’s husband is
important for breastfeeding success and a study of women in Saudi Arabia found that mothers were
more likely to initiate breastfeeding if their partners supported breastfeeding [25]. A Turkish
intervention study reported the positive effects of an antenatal education program for fathers on their
reproductive health knowledge, attitudes and behaviours, and women whose husbands attended these
classes reported that their husbands became more supportive and communicative [52].
Women everywhere doubt the adequacy of their milk supply [53] and in this study more than eight
in 10 women gave this as one of their reasons for discontinuing breastfeeding. Perceived breast milk
insufficiency or insufficient milk syndrome (IMS) is frequently associated with the premature
introduction of complementary foods [9] and with the cessation of breastfeeding [14,16] in Middle
Eastern countries. It has been proposed [53] that IMS is increasing with “aspects of ‘modernization’:
urbanization, education, and female employment—factors that are repeatedly found to be inversely
associated with both the prevalence and duration of breastfeeding” (p42). It has been suggested that
‘insufficient milk’ is given as a socially acceptable reason for discontinuing breastfeeding when a
mother decides she no longer wishes to breastfeed [54] and that claims of IMS should not be taken
literally when they occur in cultural contexts that present the use of infant formula as an acceptable, if
not preferred, alternative [55].
As we have previously identified [30], there are a number of limitations to this study. Firstly, the
sample size is relatively small and this is reflected in the wide confidence intervals around some of the
adjusted hazard ratios reported. Secondly, we may have underestimated the rate of prelacteal feeding,
which in this study was defined as within the first three days after birth. The average length of
post-partum stay for Kuwaiti public hospitals is a maximum of two nights for uncomplicated deliveries
and five nights for a caesarean section. Therefore, it is possible that some mothers discharged within
48 h may have gone on to supplement breastfeeding with formula following discharge from hospital
and within this 72 h period. Given, however, that almost nine out of 10 infants received prelacteal
feeding in hospital, any underestimation of prelacteal feeding is likely to have had only a negligible
effect on the results. Finally, the number of women who delivered by caesarean section is three times
that of the national average. While every attempt was made to recruit mothers within 72 h and in most
cases 48 h, women who had undergone a caesarean section had a greater chance of being recruited
because of their longer hospital stay.
The major strength of this study is that it was the first prospective study of infant feeding practices
in Kuwait, all other previously reported studies being cross-sectional. Mother-infant dyads were
 96

followed from birth to 26 weeks with data being collected at five time points during this period, thus
minimizing the potential for maternal recall bias [56] as women were recalling events close to the time
at which they occurred. The findings of the study are consistent with those of other studies in the
region and, in most instances, studies of Western women, and can be used to inform infant feeding
policy, hospital practices and the design of breastfeeding promotion interventions.

5. Conclusions

The duration of breastfeeding amongst women in Kuwait, particularly Kuwaiti born women,
appears to be declining. Exclusive breastfeeding is virtually non-existent with almost nine in 10 infants
receiving prelacteal feeds within the first three days of birth. Full breastfeeding is also relatively
uncommon, and by four weeks, less than one third of infants were fully breastfed, while at six months,
only four in 10 infants were receiving some breast milk. This study identified a number of areas for
intervention. Hospitals should follow the 10 Steps for Successful Breastfeeding [57] and, in particular,
promote the early initiation of breastfeeding, encourage women to feed on demand and avoid the
unnecessary practice of prelacteal feeding. Women and health professionals need to be alerted to the
negative consequences of early pacifier use on breastfeeding duration. The role of family members
should not be underestimated in planning breastfeeding interventions. Community-based interventions
are needed to support women to breastfeed and to provide a supportive environment. Close family
members, especially husbands and maternal grandmothers, should be targeted in these interventions to
ensure higher rates of exclusive breastfeeding and prolonged duration.

Acknowledgments

We sincerely appreciate the assistance given by mothers in our study and the enthusiastic support
from the Kuwait Ministry of Health and hospital staff in all participating hospitals. MD was supported
by a PhD Scholarship from the Civil Services in Kuwait. We would like to acknowledge the statistical
support given by Rosie Meng of Curtin University.

Author Contributions

MD participated in the design of the study, collected the data, performed the statistical analysis and
co-wrote the first draft of the manuscript. JAS conceived of the study, developed the original
questionnaires on which the study instruments were based, assisted with statistical analysis and
co-wrote the first draft of the manuscript. CAE advised on the statistical analysis and commented on
drafts of the manuscript, and MAS provided assistance with the on-site coordination of the study and
commented on drafts of the manuscript. All authors read and approved the final manuscript.

Conflicts of Interest

The authors declare no conflict of interest.

 97

References

1. Labbok, M.; Clark, D.; Goldman, A. Breastfeeding: Maintaining an irreplaceable immunological

resource. Nat. Rev. Immunol. 2004, 4, 565–572.
2. Ip, S.; Chung, M.; Raman, G.; Chem, P.; Magula, N.; Devine, D.; Trikalinos, T.; Lau, J.
Breastfeeding and Maternal and Child Health Outcomes in Developed Countries; Agency for
Healthcare Research and Quality: Rockville, MD, USA, 2007.
3. WHO/UNICEF. Global Strategy for Infant and Young Child Feeding; World Health
Organization: Geneva, Switzerland, 2003.
4. El-Gilany, A.; Sarraf, B.; Al-Wehady, A. Factors associated with timely initiation of
breastfeeding in Al-Hassa province, Saudi Arabia. East Mediterr. Health J. 2012, 18, 250–254.
5. Sallam, S.A.; Babrs, G.M.; Sadek, R.R.; Mostafa, A.M. Knowledge, attitude, and practices
regarding early start of breastfeeding among pregnant, lactating women and healthcare workers in
El-Minia University Hospital. Breastfeed. Med. 2013, 8, 312–316.
6. Abdul Ameer, A.; Al-Hadi, A.; Abdulla, M. Knowledge, attitudes and practices of Iraqi mothers
and family child-caring women regarding breastfeeding. East Mediterr. Health J. 2008, 14,
1003–1014.
7. Batal, M.; Boulghaurjian, C. Breastfeeding initiation and duration in Lebanon: Are the hospitals
“mother friendly”? J. Pediatr. Nurs. 2005, 20, 53–59.
8. Ergenekon-Ozelci, P.; Elmaci, N.; Ertem, M.; Saka, G. Breastfeeding beliefs and practices among
migrant mothers in slums of Diyarbakir, Turkey, 2001. Eur. J. Public Health 2006, 16, 143–148.
9. Nasreddine, L.; Zeidan, M.N.; Naja, F.; Hwalla, N. Complementary feeding in the MENA region:
Practices and challenges. Nutr. Metab. Cardiovasc. Dis. 2012, 22, 793–798.
10. Yngve, A.; Sjostrom, M. Breastfeeding determinants and a suggested framework for action in
Europe. Public Health Nutr. 2001, 4, 729–739.
11. Scott, J.; Binns, C. Factors associated with the initiation and duration of breastfeeding: A review
of the literature. Breastfeed. Rev. 1998, 7, 5–16.
12. Dennis, C.L. Breastfeeding initiation and duration: A 1990–2000 literature review. J. Obstet.
Gynecol. Neonatal Nurs. 2002, 31, 12–32.
13. Khattab, M.S. Cross-sectional study of a child health care programme at one family practice
centre in Saudi Arabia. East Mediterr. Health J. 2000, 6, 246–259.
14. Batal, M.; Boulghourjian, C.; Abdallah, A.; Afifi, R. Breast-feeding and feeding practices of
infants in a developing country: A national survey in Lebanon. Public Health Nutr. 2006, 9,
313–319.
15. Osman, N.A.; El-Sabban, F.F. Infant-feeding practices in Al-Ain, United Arab Emirates.
East Mediterr. Health J. 1999, 5, 103–110.
16. Radwan, H. Patterns and determinants of breastfeeding and complementary feeding practices of
Emirati Mothers in the United Arab Emirates. BMC Public Health 2013, 13, 171.
17. Hajian-Tilaki, K.O. Factors associated with the pattern of breastfeeding in the north of Iran.
Ann. Hum. Biol. 2005, 32, 702–713.
 98

18. Al-Sahab, B.; Tamim, H.; Mumtaz, G.; Khawaja, M.; Khogali, M.; Afifi, R.; Nassif, Y.;
Yunis, K.A.; National Collaborative Perinatal Neonatal Network (NCPNN). Predictors of
breast-feeding in a developing country: Results of a prospective cohort study. Public Health Nutr.
2008, 11, 1350–1356.
19. Musaiger, A.O.; Abdulkhalek, N. Breastfeeding and weaning practices in Bahrain: The role of
mothers’ education. Nutr. Health 2000, 14, 257–263.
20. Sharief, N.M.; Margolis, S.; Townsend, T. Breastfeeding patterns in Fujairah, United Arab
Emirates. J. Trop. Pediatr. 2001, 47, 304–306.
21. Khassawneh, M.; Khader, Y.; Amarin, Z.; Alkafajei, A. Knowledge, attitude and practice of
breastfeeding in the north of Jordan: A cross-sectional study. Int. Breastfeed. J. 2006, 1, 17.
22. Al-Hreashy, F.A.; Tamim, H.M.; Al-Baz, N.; Al-Kharji, N.H.; Al-Amer, A.; Al-Ajmi, H.;
Eldemerdash, A.A. Patterns of breastfeeding practice during the first 6 months of life in
Saudi Arabia. Saudi Med. J. 2008, 29, 427–431.
23. Shiva, F.; Nasiri, M. A study of feeding patterns in young infants. J. Trop. Pediatr. 2003, 49,
89–92.
24. Shawky, S.; Abalkhail, B. Maternal factors associated with the duration of breast feeding in
Jeddah, Saudi Arabia. Pediatr. Perinat. Epidemiol. 2003, 17, 91–96.
25. Ogbeide, D.O.; Siddiqui, S.; Al Khalifa, I.M.; Karim, A. Breast feeding in a Saudi Arabian
community. Profile of parents and influencing factors. Saudi Med. J. 2004, 25, 580–584.
26. Alikasifoglu, M.; Erginoz, E.; Gur, E.T.; Baltas, Z.; Beker, B.; Arvas, A. Factors influencing the
duration of exclusive breastfeeding in a group of Turkish women. J. Hum. Lact. 2001, 17, 220–226.
27. Al Bustan, M.; Kohli, B.R. Socio-economic and demographic factors influencing breast-feeding
among Kuwaiti women. Genus 1988, 44, 265–278.
28. Al-Awadi, F.A.; Amine, E.K. Recent trends in infant feeding patterns and weaning practices in
Kuwait. East Mediterr. Health J. 1997, 3, 501–510.
29. Ministry of Health. Kuwait Family Health Survey 1996, Preliminary Report; Ministry of Health:
Riyadh, Kuwait, 1996.
30. Dashti, M.; Scott, J.A.; Edwards, C.A.; Al-Sughayer, M. Determinants of breastfeeding initiation
among mothers in Kuwait. Int. Breastfeed. J. 2010, 5, 7.
31. Scott, J.A.; Binns, C.W.; Oddy, W.H.; Graham, K.I. Predictors of breastfeeding duration:
Evidence from a cohort study. Pediatrics 2006, 17, e646–e655.
32. Labbok, M.; Krasovec, K. Toward consistency in breastfeeding definitions. Stud. Fam. Plan.
1990, 4, 226–230.
33. Kramer, M.; Kakuma, R. The Optimal Duration of Exclusive Breastfeeding: A Systematic Review;
World Health Organization: Geneva, Switzerland, 2002.
34. Norusis, M.J. SPSS for Windows. Advanced Statistics. Release 6.0; SPSS Inc.: Chicago, IL, USA,
1993; p. 577.
35. Al-Sairafi, M.; Al-Dallal, Z.; Moosa, K. Breastfeeding Patterns and Practices in the Kingdom of
Bahrain (Children Aged 0–24 Months); Ministry of Health: Juffair, Kingdom of Bahrain, 2002.
36. Mubaideen, M.; Al-Saraireh, R. Breast-feeding pattern in selected antenatal clinics in Jordon.
JRMS 2006, 13, 74–78.
 99

37. Al Fadli, H.; Al-Jasem, L.; Al-Jady, A.; Masoud, G.M. Factors underlying bottle-feeding practices
in Kuwait. Kuwait Med. J. 2006, 38, 118–121.
38. Ong, G.; Yap, M.; Li, F.L.; Choo, T.B. Impact of working status on breastfeeding in Singapore:
Evidence from the National Breastfeeding Survey 2001. Eur. J. Public Health 2005, 15, 424–430.
39. Hawkins, S.S.; Griffiths, L.J.; Dezateux, C.; Law, C.; Millennium Cohort Study Child Health
Group. The impact of maternal employment on breast-feeding duration in the UK Millennium
Cohort Study. Public Health Nutr. 2007, 10, 891–896.
40. Taveras, E.M.; Capra, A.M.; Braveman, P.A.; Jensvold, N.G.; Escobar, G.J.; Lieu, T.A. Clinician
support and psychosocial risk factors associated with breastfeeding discontinuation. Pediatrics
2003, 112, 108–115.
41. Klaus, M.H. The frequency of suckling. A neglected but essential ingredient of breast-feeding.
Obstet. Gynecol. Clin. N. Am. 1987, 14, 623–633.
42. Division of Child Health and Development. Evidence for the Ten Steps to Successful
Breastfeeding; World Health Organization: Geneva, Switzerland, 1998.
43. Binns, C.; Scott, J. Using pacifiers: What are breastfeeding mothers doing? Breastfeed. Rev. 2002,
10, 21–25.
44. Mauch, C.E.; Scott, J.A.; Magarey, A.M.; Daniels, L.A. Predictors of and reasons for pacifier use
in first-time mothers: An observational study. BMC Pediatr. 2012, 12, 7.
45. Howard, C.R.; Howard, F.M.; Lanphear, B.; Eberly, S.; deBlieck, E.A.; Oakes, D.;
Lawrence, R.A. Randomized trial of pacifier use and bottle-feeding or cupfeeding and their effect
on breastfeeding. Pediatrics 2003, 111, 511–518.
46. Hewitt, K.; Binns, C.; Scott, J.; Forbes, D. Pacifier use, breastfeeding and sudden infant death
syndrome. Curr. Pediatr. Rev. 2012, 8, 285–291.
47. Matich, J.R.; Sims, L.S. A comparison of social support variables between women who intend to
breast or bottle feed. Soc. Sci. Med. 1992, 34, 919–927.
48. Bryant, C.A. The impact of kin, friend and neighbour networks on infant feeding practices.
Soc. Sci. Med. 1982, 16, 1757–1765.
49. Khadduri, R.; Marsh, D.R.; Rasmussen, B.; Bari, A.; Nazir, R.; Darmstadt, G.L. Household
knowledge and practices of newborn and maternal health in Haripur district, Pakistan. J. Perinatol.
2008, 28, 182–187.
50. Scott, J.; Shaker, I.; Reid, M. Parental attitudes toward breastfeeding: Their association with
feeding outcome at hospital discharge. Birth 2004, 31, 125–131.
51. Giugliani, E.R.J.; Caiaffa, W.T.; Vogelhut, J.; Witter, F.R.; Perman, J.A. Effect of breastfeeding
support from different sources on mothers’ decisions to breastfeed. J. Hum. Lact. 1994, 10, 157–161.
52. Molzan Turan, J.M.; Nalbant, H.; Bulut, A.; Sahip, Y. Including expectant fathers in antenatal
education programmes in Istanbul, Turkey. Reprod. Health Matters 2001, 9, 114–125.
53. Obermeyer, C.M.; Castle, S. Back to nature? Historical and cross-cultural perspectives on barriers
to optimal breastfeeding. Med. Anthropol. 1996, 17, 39–63.
54. Hitchcock, N.E.; Coy, J.F. Infant-feeding practices in Western Australia and Tasmania: A joint
survey, 1984–1985. Med. J. Aust. 1988, 148, 114–117.
55. Quandt, S.A. Patterns of variation in breast-feeding behaviours. Soc. Sci. Med. 1986, 23, 445–453.
 100

56. Haaga, J.G. Reliability of retrospective survey data on infant feeding. Demography 1988, 25,
307–315.
57. UNICEF, Baby Friendly Hospital Initiative. Available online: http://www.unicef.org/programme/
breastfeeding/baby.htm (accessed on 6 December 2005).

© 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).
 101

Reprinted from Nutrients. Cite as: Jonsdottir, O.H.; Thorsdottir, I.; Gunnlaugsson, G.; Fewtrell, M.S.;
Hibberd, P.L.; Kleinman, R.E. Exclusive Breastfeeding and Developmental and Behavioral Status in
Early Childhood. Nutrients 2013, 5, 4414-4428.

Exclusive Breastfeeding and Developmental and Behavioral

Status in Early Childhood
Olof H. Jonsdottir 1,*, Inga Thorsdottir 1, Geir Gunnlaugsson 2, Mary S. Fewtrell 3,
Patricia L. Hibberd 4 and Ronald E. Kleinman 5
1
Unit for Nutrition Research, Landspitali—The National University Hospital of Iceland and Faculty
of Food Science and Nutrition, School of Health Sciences, University of Iceland, Eiriksgata 29,
Reykjavik 101, Iceland; E-Mail: ingathor@hi.is
2
Directorate of Health and Reykjavik University, Reykjavik 101, Iceland; E-Mail: geirg@hr.is
3
Childhood Nutrition Research Centre, UCL Institute of Child Health, London WC1N 1EH, UK;
E-Mail: m.fewtrell@ucl.ac.uk
4
Division of Global Health, Massachusetts General Hospital for Children, Harvard Medical School,
Boston, MA 02114, USA; E-Mail: phibberd@partners.org
5
Department of Pediatrics, Massachusetts General Hospital for Children, Harvard Medical School,
Boston, MA 02114, USA; E-Mail: rkleinman@partners.org

* Author to whom correspondence should be addressed; E-Mail: ohj@hi.is; Tel.: +354-543-1452;

Fax: +354-543-4824.

Received: 27 August 2013; in revised form: 30 October 2013 / Accepted: 31 October 2013 /
Published: 11 November 2013

Abstract: Breastfeeding during infancy may have beneficial effects on various

developmental outcomes in childhood. In this study, exclusively breastfed infants were
randomly assigned to receive complementary foods from the age of 4 months in addition
to breast milk (CF, n = 60), or to exclusively breastfeed to 6 months (EBF, n = 59).
At 18 months and again at 30–35 months of age, the children were evaluated with the
Parent’s Evaluation of Developmental Status questionnaire (PEDS) and the Brigance
Screens-II. The parents completed the PEDS questionnaire at both time intervals and the
children underwent the Brigance Screens-II at 30–35 months. At 30–35 months, no significant
differences were seen in developmental scores from the Brigance screening test (p = 0.82).
However, at 30–35 months a smaller percentage of parents in group CF (2%) had concerns
about their children’s gross motor development compared to those in group EBF (19%;
p = 0.01), which remained significant when adjusted for differences in pre-randomization
characteristics (p = 0.03). No sustained effect of a longer duration of exclusive
 102

breastfeeding was seen on selected measures of developmental and behavioral status at

18 months, although at 30–35 months, a smaller percentage of parents of children
introduced to complementary foods at four months of age expressed concerns about their
gross motor development.

Keywords: early childhood; exclusive breastfeeding; complementary feeding;

developmental status; behavior; randomized trial

Clinical Trial Registration: ISRCTN41946519

1. Introduction

Breastfeeding may have beneficial effects on development in childhood, adolescence and even in
adulthood [1,2], although this has not been a consistent finding [3]. Furthermore, some studies indicate
that a longer duration of exclusive breastfeeding is important for this positive association with
developmental outcomes in childhood, especially for those born small for gestational age [4–6]. While
most studies have focused on cognitive development, less is known about the impact of breastfeeding
and the duration of exclusive breastfeeding on non-cognitive developmental and behavioral status in
childhood. Some studies indicate that breastfeeding in general, and also, a longer duration of breastfeeding
may be associated with decreased risk of behavioral problems and developmental delays in childhood [7–9];
however, findings on this subject are inconsistent. A large breastfeeding promotion intervention in
Belarus showed no relationship between prolonged breastfeeding or longer duration of exclusive
breastfeeding and childrens’ behavior at 6.5 years of age [10,11]. Other studies have shown that
increased duration and exclusivity of breastfeeding may have beneficial effects on language and motor
development in childhood [12–18].
There has been a longstanding debate about the optimal duration of exclusive breastfeeding;
whether infants should be exclusively breastfed for 4 or 6 months after birth [19]. The current
recommendations of the WHO are that infants should be exclusively breastfed for the first 6 months of
life [20] but until May 2001 the WHO recommended exclusive breastfeeding for 4–6 months of
age [21]. We have previously reported the results of a parallel group, masked, randomized controlled
trial of the effects of exclusive breastfeeding for 4 vs. 6 months on growth, body composition,
breast-milk intake and iron status of the infant [22,23]. We now report a secondary analysis from this
cohort of exclusive breastfeeding infants for 4 vs. 6 months on selected measures of development and
behavior in early childhood. We hypothesized that infants exclusively breastfed for 6 months would
have better outcomes in selected measures of developmental and behavioral status at 18 months and
30–35 months of age than those receiving complementary foods from 4 months in addition to
breast milk.
 103

2. Experimental Section

2.1. Study Design

As described previously [22,23], between November 2007 and November 2009, a total of
119 mother-infant pairs were recruited at seven health care centers in Iceland where 50% and 35% of
mothers exclusively breastfeed through 4 and 5 months of age, respectively [24]. A total of 656 infants
were assessed for eligibility in this randomized controlled trial. Eligibility criteria for the study were
singleton birth, gestational length • ZHHNV H[FOXVLYHO\ EUHDVWIHG LQIDQW FKDUDFWHUL]HG DV KHDOWK\
absence of congenital abnormalities or chronic health issues likely to affect growth, development or
iron status. Mothers of eligible infants were invited to participate in the study and infants who were
still exclusively breastfed and whose parents were willing to participate were enrolled in the study at
4 months of age. Eligible mother-infant pairs were randomly assigned to receive complementary foods
from the age of 4 months in addition to breast milk (CF), or to continue being exclusively breastfed to
the age of 6 months (EBF). Vitamin D supplements were recommended in both groups. Exclusive
breastfeeding was defined as breastfeeding with no additional liquid or solid foods other than vitamins
and medications [25]. The use of up to 10 feedings of formula or water during the first 6 months was
allowed to avoid having to exclude infants that in fact were otherwise exclusively breastfed.
The study was reviewed and approved by the Data Protection Authority and National Bioethical
Committee in Iceland and the Partners Health System IRB, Boston, MA, USA.

2.2. Selected Measures of Developmental and Behavioral Status

Children in the present study were assessed both at 18 months and 30–35 months of age, during
their routine health care visits at the health center, where developmental and behavioral status was
assessed with both the Parent’s Evaluation of Developmental Status (PEDS) questionnaire and
the Brigance Screens-II. The parents filled out the PEDS questionnaire at both visits, at 18 months and
30–35 months of age, and the children underwent the Brigance Screens-II at 30–35 months.
Both tests were administered by trained nurses at each health care center following prescribed
protocols [26–28]. PEDS questionnaire and Brigance Screens-II were both introduced in 2010 as part
of routine health care visits at health centers in Iceland.
The PEDS is designed to detect parental concerns about the developmental status and behaviors of
their child; it has been found to have very good reliability and has been validated for children from
birth to 8 years of age [29–33]. The PEDS questionnaire consists of 10 brief questions, two open-
ended about general cognitive function and other concerns and eight domain-specific items. For each
of the eight domain-specific questions the parents are asked if they have any concerns about the
development or behavior of their child and their response option is in a multiple-choice format (no,
yes, a little). Certain parental expressions of concern in response to certain of these questions are
predictive of developmental delay [26]. If parents express concern in response to >2 of these predictive
questions, then health center procedures require that the child be referred for further evaluation
(see Figure 1). It takes parents approximately 5 min to answer the questionnaire.
 104

Figure 1. Developmental screening tests used in the study, the Parent’s Evaluation of
Developmental Status (PEDS) questionnaire and the Brigance Screens-II.

Developmental screening tests used in

18 m the study
30-35 m 30-35 m

Parent-completed Direct elicitation of the

screening questionnaire children

PEDS: 10 questions Brigance: 11 components (100

2 open-ended points)
i General cognitive function i 1B Personal data response

i Other concerns i 2B Identifies body parts

8 domain-specific i 3B Gross motor skills

i Expression and articulation
i 4B Identifies objects
i Language comprehension
i 5B Repeats sentences
i Fine motor skills
i 6B Visual motor skills
i Gross motor skills
i 7B Number concepts
i Behavior
i 8B Builds tower with blocks
i Social-emotional
i 9BMatches colors
i Self-help skills
i 10B Picture vocabulary
i Preschool and school skills
i 11B Plural s and –ing

Predictive components: 2B, 4B, 5B,

10B, 11B (47 points)

Children at risk of
18 m developmental delay
30-35 m

•Predictive concerns Cut off value of total Cut off value of

•Predictive concerns i General cognitive score predictive components
function i <72 (30-32 m) i <34 (30-32 m)
i General cognitive

function i Expression and i <76 (33-35 m) i <35 (33-35 m)

articulation
i Expression and
i Language
articulation comprehension
i Language i Other concerns
comprehension
i Gross motor skills
i Other concerns (at 36 months)
 105

The Brigance Screens-II is administered by a trained nurse who observes the child and questions
his/her parents and the test is completed by the child itself. It has good reliability and has
been validated for measuring the developmental and behavioral status of toddlers and preschool
children [34–36]. The Brigance Screens-II for 30–35 month old children is valid for children from the
age of 29 months + 15 days to 35 months + 14 days old children. The Brigance Screens-II comprises
11 components and it takes children approximately 15–20 min to complete the test. The cut off points
for defining children at risk of developmental delay are <72 and <76 points of 100 points for children
aged 30–32 months and 33–35 months, respectively. As with the PEDS questionnaire, there are some
components of the Brigance Screens-II more predictive for developmental delay than others among all
the test components (see Figure 1). Cut off points for defining children at risk of developmental delay
are <34 and <35 points for children aged 30–32 and 33–35 months, respectively [37]. In the current
study we focused on assessment of gross motor skills (3B), fine motor skills (6B, 8B) and receptive or
expressive language (5B, 10B, 11B), since studies indicate that breastfeeding may influence these
factors [12–18].

2.3. Statistical Analysis

Data were analyzed with SPSS Windows statistical software package version 20.0 (SPSS Inc.,
Chicago, IL, USA) with a level of significance of p ” 0.05. Data were presented with means and
standard deviations (SD) for normally distributed variables and with median and interquartile range
(IQR) for variables with skewed distribution. Group comparisons were performed using independent-
samples t-test and Mann-Whitney U-test. Comparisons between categorical values were made using
the Chi-square tests of association or two-sided Fisher’s exact test. Regression analysis was performed
to adjust for any pre-randomization characteristics that were different between the two intervention
groups at baseline. Finally, we calculated the power to detect differences between the CF and EBF
groups based on proportions. To detect a significant difference between intervention groups in
developmental scores from the Brigance screening test at 30–35 months of age with a sample size of
66 and a power of 80%, the mean difference in developmental scores would have had to be
approximately 11.2 points, or approximately 5.4 points if excluding the three outliers (n = 63). Of the
100 mother-infant pairs who finished the breastfeeding intervention trial, a total of 95 children attended
routine care at 18 months and 82 at 30–35 months. Fifty-four parents answered the PEDS questionnaire
when their child was 18 months and 78 parents at the 30–35 months visit. These numbers are based on the
calculation of the sample size.

3. Results

3.1. Sample Size and Characteristics of Participants

Since both PEDS questionnaire and Brigance Screens-II were introduced in 2010 we have
41 missing data points from PEDS questionnaire at 18 months of age for those children in the study
born in 2007 and some who were born in 2008. Parents of 4 children who attended routine care at
30–35 months did not answer the PEDS questionnaire at that age. The Brigance Screens-II was
undertaken by 77 children at the age of 30–35 months, but 10 of them were too old (>35 months + 14 days)
 106

and 1 too young (<29 months + 15 days) when the Brigance Screens-II was performed and were
therefore excluded from the analysis. The PEDS questionnaire is for a wider age range but we chose to
use 30–35 months throughout the paper. The children that did not have developmental scores recorded
from the Brigance Screens-II (n = 23) were lost to follow-up for several reasons, such as the family had
moved abroad or failure to attend the routine health care visits at the health center.
Among those children with developmental scores from the Brigance Screens-II at 30–35 months of
age (n = 66), no differences between study groups were seen in baseline characteristics, except for
mode of delivery, where vaginal delivery was more common among children in the CF group (94% vs.
74% in the EBF group, p = 0.04) (see Table 1). No difference was seen in baseline characteristics
(same as seen in Table 1) among those who were followed-up (n = 82) and those who were lost to
follow-up (n = 18), except for parity, where those parents who were lost to follow-up had more children
(3.0 ± 1.0 children) than those who were followed-up (2.0 ± 2.0 children; p = 0.01).

Table 1. Baseline characteristics of participants with scores from Brigance Screens-II at

30–35 months of age in the two study groups: infants who received complementary foods in
addition to breast milk from 4 months (CF, n = 35) compared with infants who were exclusively
breastfed for 6 months (EBF, n = 31).
Variables Group CF Group EBF
Boys 17 (49%) * 13 (42%) *
Birth weight (g) 3687 (432) 3733 (526)
Length at birth (cm) 51.3 (1.8) 51.7 (1.9)
Head circumference at birth (cm) 35.8 (1.3) † 35.9 (1.4)
Gain in head circumference from
12.6 (1.2) ‡ 12.6 (1.7) §
birth–18 months (cm)
Age when Brigance Screens-II was
32.3 (1.6) 32.8 (1.6)
performed (months)
Gestational length (days) 280.5 (9.3) 280.8 (7.1)
Maternal age (years) 29.4 (4.4) 31.2 (4.8)
Maternal education Π22 (63%) * 16 (52%) *
Vaginal delivery 33 (94%) * 23 (74%) *
Parity 2.0 (2.0) ¶ 2.0 (1.0) ¶
Father’s education Œ 13 (38%) *‡ 14 (45%) *
†
Data are presented as mean (SD) unless otherwise indicated; * Data are presented as number (%); One
‡ § Œ
missing value, n = 34; Three missing values, n = 32; Five missing values, n = 26; Finished studies at
university level; ¶ Data are presented as median (IQR).

3.2. Developmental and Behavioral Status

Table 2 shows the developmental and behavioral status measures in the two study groups at
18 months (PEDS questionnaire) and at 30–35 months (PEDS questionnaire and Brigance
Screens-II). At 18 months, a significantly smaller percentage of parents had concerns about any of the
domains of PEDS on their children’s developmental and behavioral status in the CF group compared
with those in the EBF group (17% in the CF group vs. 44% in the EBF group; p = 0.03). A logistic
regression was done to test the impact of the intervention by group, and when adjusted for mode of
 107

delivery, the difference in parents’ concerns between groups at 18 months was not statistically
significant (p = 0.08). No difference was seen between groups in the number of concerns regarding
gross or fine motor skills or receptive and expressive language. At 18 months, parents most often
expressed concerns about their children’s expression and articulation of the eight domain-specific
questions; 10% and 20% in the CF and EBF group, respectively (p = 0.45). No significant differences
were seen even when those questions with greater predictive value for developmental delay were
compared among groups at 18 months (0% in the EBF group vs. 3% in the CF group; p = 1.0).
At 30–35 months of age no significant differences were seen between study groups in number of
parents with concerns about any of the domains of PEDS (42% in the EBF group vs. 33% in the CF
group; p = 0.45). A smaller proportion of parents of children in the CF group (2%) had concerns about
their gross motor development compared with parents of those in the EBF group (19%; p = 0.01).
When adjusted for mode of delivery the difference was still significant (p = 0.03). No difference was
seen between groups in number of concerns regarding fine motor skills or receptive and expressive
language. At 30–35 months, parents most often expressed their concerns about their children’s
expression and articulation of the eight domain-specific questions; 19% and 28% in the CF and EBF
group, respectively (p = 0.36). Use of the cut off of•SUHGLFWLYHFRQFHUQVIRU3('6TXHVWLRQQDLUHDW
30–35 months showed that 19% of parents in the EBF group were above the cut off value compared
with 5% of the parents in the CF group, although the difference was not significant (p = 0.07).

Table 2. Selected measures of developmental and behavioral status for children at

18 months and at 30–35 months of age in the two intervention groups: infants who received
complementary foods in addition to breast milk from 4 months (CF) compared with infants who
were exclusively breastfed for 6 months (EBF).
Variables Group CF Group EBF P-value
PEDS questionnaire
Parents with concerns according to PEDS at
5 (17%) *; n = 29 11 (44%) *; n = 25 0.03
18 months
Parents with concerns according to PEDS at
14 (33%) *; n = 42 15 (42%) *; n = 36 0.45
30–35 months
Brigance Screens-II n = 35 n = 31
Total score at 30–35 months 86.0 (12.5) † 86.5 (12.5) † 0.82
Total score above cut off value ‡ 2 (6%) * 4 (13%) * 0.41
Score of predictive factors combined above
7 (20%) * 3 (10%) *Π0.32
cut off value §
Components of the Brigance Screens-II
Gross motor skills 6.0 (6.0) † 6.0 (4.5) †Œ 0.44
Fine motor skills 19.0 (3.0) † 19.0 (3.0) †Œ 0.89
Expressive and receptive language 40.5 (8.0) † 42.0 (9.5) †Œ 0.81
Data are presented as mean (SD) unless otherwise indicated; * Data are presented as number (%); † Data are
‡
presented as median (IQR); Cut off values for defining risk of developmental delay were <72 and <76
points from the total score from the Brigance Screens-II for children aged 30–32 months and 33–35 months,
§
respectively; Cut off values for defining risk of developmental delay were <34 and <35 points from the
predictive components of the Brigance Screens-II combined for children aged 30–32 months and 33–35
months, respectively; ΠTwo missing values, n = 29.
 108

There was no significant difference between study groups at 30–35 months by the Brigance
Screens-II (p = 0.82). Neither was there a significant difference between the groups in the number of
children below the cut off value defining developmental delays for total score from the Brigance
Screens-II (p = 0.41) or number of children above the cut off value defining developmental delays
from predictive components of the Brigance screening test combined (p = 0.32). Furthermore, there
was no significant difference between groups in fine or gross motor skills or receptive and expressive
language according to the Brigance Screens-II at 30–35 months. Excluding three outliers found in the
EBF group in the Brigance screening test did not change the mean values for the study groups or the
lack of significance (86.1 ± 7.8 points in the CF group vs. 88.0 ± 7.4 points in the EBF group; p =
0.33).

4. Discussion

In this study of well-nourished children at 30–35 months of age, a smaller proportion of parents in
the CF group expressed their concerns about their children’s gross motor development on the PEDS
questionnaire, a difference that remained significant when adjusted for differences in pre-
randomization characteristics. However, there were no significant intergroup differences in
developmental total scores or in fine and gross motor skills or receptive and expressive language
according to the Brigance Screens-II at 30–35 months. No difference was seen in the percentage of
parents with concerns about their children’s developmental and behavioral status at the age of 18
months.
Results from the PEDS questionnaire are based on a small number of categorical variables.
Outcomes from the Brigance Screens-II, however, are based on continuous variables and therefore this
test is more responsive to detecting minor developmental disabilities. The Brigance Screens-II is a
comprehensive, reliable and valid screening tool of developmental status that is completed by the child
itself [36,38]. Similar general developmental screening tools that are directly administered to the child and
are used in primary care settings are the Battelle Developmental Inventory Screening Tool Test II, the
Bayley Infant Neurodevelopmental Screener and the Denver-II Developmental Screening Test, which are
all comparable to the Brigance Screens-II in sensitivity and specificity [38,39]. Per health center protocols
in the Icelandic healthcare system, children identified at risk for developmental delay or behavioral
problems according to the Brigance Screens-II or the PEDS questionnaire are referred for further
evaluation, diagnosis and then developmental intervention, if appropriate. Early detection of developmental
delay and appropriate intervention has been shown to be effective in improving developmental outcomes in
childhood [40].
Although the PEDS is solely based on parental perception of their children’s developmental and
behavioral status, a positive correlation has been shown between the results of the PEDS questionnaire
and the Brigance Screens-II [28]. The PEDS questionnaire is a valid and reliable developmental
screening tool [38] and the value of parents’ concerns in the detection of developmental delay has been
well studied [31,41]. Comparable commonly used parent-completed screening questionnaires in
primary care settings comparable are the Ages & Stages Questionnaires, the Child Development
Review-Parent Questionnaire and the Infant Development Inventory [38,39]. These tests are not
perfectly concordant, but are widely used and are considered appropriate for developmental evaluation
 109

in primary care settings [42,43]. It should be noted that although some parental concerns are predictive
over time, the PEDS questionnaire does not always capture longitudinal changes in developmental
status since parents may have fewer concerns after their child begins a developmental intervention,
even though developmental delays may still be present. In 2006, the American Academy of Pediatrics
recommended systematic developmental screening in primary care using a validated screening tool for
children aged 9, 18 and 30 months, but no specific guidance was provided for the specific screening
tools that should be used [38]. Health care centers in Iceland chose to use the PEDS questionnaire and
the Brigance screening test because of their good reliability and validity and well-established
sensitivity and specificity and because they are useable for a wide range of ages in childhood [28].
To our knowledge this is the first secondary analysis of a randomized controlled trial conducted in a
resource rich country to examine the effects of exclusive breastfeeding for 4 vs. 6 months on selected
measures of developmental and behavioral status in early childhood. The World Health Organization
(WHO) recommended exclusive breastfeeding for 4–6 months of life until the year 2001 when the
recommendation was changed to breastfeed exclusively for the first 6 months of life in an effort to
lower the risk of adverse health outcomes for infants during the first 6 months, particularly in resource
constrained countries [44]. Developmental status is influenced by a number of genetic and environmental
factors that cause cumulative risk effects of development delays that are generally not addressed in
observational studies. This is one possible explanation for the inconsistent findings among such studies
[13,45,46]. Studies investigating the relationship between breastfeeding and developmental status often
compare formula fed infants to breastfed infants [46–49], but less is known about the impact of
exclusive breastfeeding compared to partial breastfeeding.
There is strong evidence that nutrition early in life can have long-term effects on health and
development later in life [50–52]. It has been suggested that the concentration of long chain
polyunsaturated fatty acids in breast milk may explain the enhanced cognitive outcomes reported in
some studies comparing breastfed and formula fed infants [53–56] and therefore the effect of duration
of exclusive breastfeeding on developmental and behavioral status would also be a relevant factor in
these outcomes. Since infants exclusively breastfed for 6 months in the present study had significantly
higher breast milk intakes at 5.5–6 months of age [22], we hypothesized they would have better
developmental and behavioral status in early childhood. However, no intergroup differences in measures of
developmental and behavioral status outcomes were observed among those that completed the Brigance
Screens-II at 30–35 months of age. The parental impressions from the PEDS questionnaire administered
when the children were 30–35 months of age were thus not substantiated by the more objective and reliable
Brigance Screens-II at the same age. The reason for no difference in these developmental and behavioral
measures might be because both study groups consumed a significant amount of breast milk. While the
infants in the EBF group consumed significantly more breast milk than those in the CF group (83 g/day,
measured using the stable isotope technique) [22], the amount consumed by the CF group was consistent
with the recommendations of the WHO [57]. The mothers who participated in our study were generally
well-educated and well-supported, and we cannot generalize our findings to other populations. It is possible
that in less well-educated or supported mothers, the introduction of a small amount of CF might result in a
greater decrease in breast milk production with more impact on health outcomes, including development.
The strength of the present study lies in the fact that this is the only analysis of later developmental
and behavioral data from a randomized controlled trial of 4 vs. 6 months of exclusive breastfeeding
 110

and it therefore has a methodological advantage over previously published observational studies.
Furthermore, approximately 78% of the cohort was follow-up until the age of 30–35 months. The main
limitation of the present study is that data were collected in routine health care visits at the health
center. We recognize that this secondary data analysis may have been underpowered to detect small
effects on developmental and behavioral outcomes that may be biologically relevant, but the sample
size was adequate to exclude large effects on developmental and behavioral outcomes in the two
groups [58].
In addition to breast milk per se, other factors that influence infant development may have played a
role in the outcomes we observed in this randomized trial. Infants with depleted iron stores, iron
deficiency or iron deficiency anemia can have lower developmental scores in childhood [59,60],
however, we have previously reported that no differences was seen in the prevalence of iron deficiency
with or without anemia between both groups at 6 months of age [23]. Mothers who choose to
breastfeed may differ from those who never breastfeed in many ways that can influence an infant’s
development, including socio-economic status and nurturing qualities. However, mothers in both study
groups exclusively breastfed for the first 4 months of their infant’s life all were from a similar
socioeconomic background and thereafter all of them continued breastfeeding partially or exclusively
until 6 months of age or beyond, minimizing the impact of these other influential developmental
factors. It is possible that the mothers participating in the study might differ from other mothers in the
population.
In conclusion, the present study showed no sustained effect of a longer duration of exclusive
breastfeeding on selected measures of developmental and behavioral status at 18 months of age
although at 30–35 months, a smaller percentage of parents of infants introduced to complementary
foods at 4 months of age expressed concerns about their children’s gross motor development. Further
investigation is needed in a larger randomized controlled trial using the same or other measures of
developmental and behavioral status to extend and confirm these findings.

5. Conclusions

Breastfeeding during infancy may have beneficial effects on various developmental outcomes in
childhood. The association between breastfeeding and developmental status is based on observational
studies that are subject to bias by confounding factors. In this randomized controlled trial, no sustained
difference were seen on selected measures of development and behavior in early childhood between
those receiving complementary foods in addition to breast milk from 4 months or those exclusively
breastfed for 6 months. Further investigation is needed in a larger randomized controlled trial using the
same or other measures of developmental and behavioral status to extend and confirm these findings.

Acknowledgments

The authors greatly acknowledge all participants in the study and the directors and staff at
the participating health centers for support to conduct the study. In particular, the nurses Sesselja
Gudmundsdottir, Dagny Gudmundsdottir (HC Hvammur), Audur Egilsdottir (HC Hamraborg) and
Kristin J. Vigfusdottir (HC Grafarvogur) and the nurses and the lactation consultants Jona Margret
Jonsdottir (Centre for Preventive Child Health Services), Alma Maria Rognvaldsdottir (HC Fjordur),
 111

Elin Sigurbjornsdottir (HC Akranes), Gudrun Gudmundsdottir (HC Sudurnes) and Ingibjorg
Eiriksdottir (HC Grafarvogur) and for recruiting mother-infant pairs and their great amount of
assistance in data collection. Furthermore we gratefully acknowledge Frances Page Glascoe for
constructive comments on earlier drafts of the paper. This study was supported by Mead Johnson and
the Eimskip Fund of the University of Iceland. The sponsors of the study had no role in study design,
data collection, data analysis, data interpretation, preparation of the report or the decision to submit for
publication. The authors have no financial relationship relevant to this article to disclose.

Conflicts of Interest

Mary S. Fewtrell has received research funding from and undertaken advisory work for companies
manufacturing infant foods and feeding products within the past 3 years, all the other authors declare
that they have no conflicts of interests.

References

1. Horta, B.L.; Bahl, R.; Martinés, J.C.; Victora, C.G. Evidence on the Long-Term Effects of
Breastfeeding: Systematic Reviews and Meta-Analyses; World Health Organization: Geneva,
Switzerland, 2007; pp. 11–39.
2. Schack-Nielsen, L.; Michaelsen, K.F. Advances in our understanding of the biology of human
milk and its effects on the offspring. J. Nutr. 2007, 137, 503S–510S.
3. Ip, S.; Chung, M.; Raman, G.; Chew, P.; Magula, N.; DeVine, D.; Trikalinos, T.; Lau, J.
Breastfeeding and maternal and infant health outcomes in developed countries. Evid. Rep.
Technol. Assess. 2007, 153, 1–186.
4. Kramer, M.S.; Aboud, F.; Mironova, E.; Vanilovich, I.; Platt, R.W.; Matush, L.; Igumnov, S.;
Fombonne, E.; Bogdanovich, N.; Ducruet, T.; et al. Breastfeeding and child cognitive
development: New evidence from a large randomized trial. Arch. Gen. Psychiatry 2008, 65,
578–584.
5. Jedrychowski, W.; Perera, F.; Jankowski, J.; Butscher, M.; Mroz, E.; Flak, E.; Kaim, I.;
Lisowska-Miszczyk, I.; Skarupa, A.; Sowa, A. Effect of exclusive breastfeeding on the
development of children’s cognitive function in the Krakow prospective birth cohort study. Eur.
J. Pediatr. 2012, 171, 151–158.
6. Rao, M.R.; Hediger, M.L.; Levine, R.J.; Naficy, A.B.; Vik, T. Effect of breastfeeding on cognitive
development of infants born small for gestational age. Acta Paediatr. 2002, 91, 267–274.
7. Oddy, W.H.; Kendall, G.E.; Li, J.; Jacoby, P.; Robinson, M.; de Klerk, N.H.; Silburn, S.R.;
Zubrick, S.R.; Landau, L.I.; Stanley, F.J. The long-term effects of breastfeeding on child and
adolescent mental health: A pregnancy cohort study followed for 14 years. J. Pediatr. 2010, 156,
568–574.
8. Halpern, R.; Giugliani, E.R.; Victora, C.G.; Barros, F.C.; Horta, B.L. Risk factors for suspicion of
developmental delays at 12 months of age. J. Pediatr. (Rio. J.) 2000, 76, 421–428.
9. Paterson, J.; Iusitini, L.; Gao, W. Child developmental assessment at two-years of age: Data from
the Pacific Islands Families Study. Pac. Health Dialog. 2011, 17, 51–63.
 112

10. Kramer, M.S.; Fombonne, E.; Igumnov, S.; Vanilovich, I.; Matush, L.; Mironova, E.; Bogdanovich,
N.; Tremblay, R.E.; Chalmers, B.; Zhang, X.; et al. Effects of prolonged and exclusive
breastfeeding on child behavior and maternal adjustment: Evidence from a large, randomized
trial. Pediatrics 2008, 121, e435–e440.
11. Kramer, M.S.; Kakuma, R. Optimal duration of exclusive breastfeeding. Cochrane Database Syst.
Rev. 2012, 8, CD003517.
12. Chiu, W.C.; Liao, H.F.; Chang, P.J.; Chen, P.C.; Chen, Y.C. Duration of breast feeding and risk
of developmental delay in Taiwanese children: A nationwide birth cohort study. Paediatr.
Perinat. Epidemiol. 2011, 25, 519–527.
13. Whitehouse, A.J.; Robinson, M.; Li, J.; Oddy, W.H. Duration of breast feeding and language
ability in middle childhood. Paediatr. Perinat. Epidemiol. 2011, 25, 44–52.
14. Quinn, P.J.; O’Callaghan, M.; Williams, G.M.; Najman, J.M.; Andersen, M.J.; Bor, W. The effect
of breastfeeding on child development at 5 years: A cohort study. J. Paediatr. Child Health 2001,
37, 465–469.
15. Tozzi, A.E.; Bisiacchi, P.; Tarantino, V.; Chiarotti, F.; D’Elia, L.; de Mei, B.; Romano, M.;
Gesualdo, F.; Salmaso, S. Effect of duration of breastfeeding on neuropsychological development
at 10 to 12 years of age in a cohort of healthy children. Dev. Med. Child Neurol. 2012, 54,
843–848.
16. Dee, D.L.; Li, R.; Lee, L.C.; Grummer-Strawn, L.M. Associations between breastfeeding
practices and young children’s language and motor skill development. Pediatrics 2007, 119,
S92–S98.
17. Vestergaard, M.; Obel, C.; Henriksen, T.B.; Sørensen, H.T.; Skajaa, E.; Ostergaard, J. Duration of
breastfeeding and developmental milestones during the latter half of infancy. Acta Paediatr. 1999,
88, 1327–1332.
18. Thorsdottir, I.; Gunnarsdottir, I.; Kvaran, M.A.; Gretarsson, S.J. Maternal body mass index,
duration of exclusive breastfeeding and children’s developmental status at the age of 6 years. Eur.
J. Clin. Nutr. 2005, 59, 426–431.
19. Fewtrell, M.S.; Morgan, J.B.; Duggan, C.; Gunnlaugsson, G.; Hibberd, P.L.; Lucas, A.;
Kleinman, R.E. Optimal duration of exclusive breastfeeding: What is the evidence to support
current recommendations? Am. J. Clin. Nutr. 2007, 85, 635S–638S.
20. World Health Organization (WHO). Infant and Young Child Nutrition. Global Strategy on Infant
and Young Child Feeding. Report by the Secretariat. In Proceedings of the 55th World Health
Assembly A55/15 Provisional Agenda Item 13.10, Geneva, Switzerland, 16 April 2002.
21. World Health Organization (WHO). Nutrition: Information and attitudes among health personnel
about early infant-feeding practices. WHO Wkly. Epidem. Rec. 1995, 70, 117–120.
22. Wells, J.C.; Jonsdottir, O.H.; Hibberd, P.L.; Fewtrell, M.S.; Thorsdottir, I.; Eaton, S.; Lucas, A.;
Gunnlaugsson, G.; Kleinman, R.E. Randomized controlled trial of 4 compared with 6 mo of
exclusive breastfeeding in Iceland: Differences in breast-milk intake by stable-isotope probe. Am.
J. Clin. Nutr. 2012, 96, 73–79.
23. Jonsdottir, O.H.; Thorsdottir, I.; Hibberd, P.L.; Fewtrell, M.S.; Wells, J.C.; Palsson, G.I.; Lucas,
A.; Gunnlaugsson, G.; Kleinman, R.E. Timing of the introduction of complementary foods in
infancy: A randomized controlled trial. Pediatrics 2012, 130, 1038–1045.
 113

24. Directorate of Health. Brjóstagjöf og næring ungbarna á Íslandi sem fædd eru 2004–2008
(Breastfeeding and Infant Nutrition among Infants Born 2004–2008 in Iceland), 2012. Available
online: http://www.landlaeknir.is/servlet/file/store93/item16573/version4/brjostagjof_og_naering_
2004-2008_juni.2012.pdf (accessed on 13 August 2013).
25. WHO. Indicators for Assessing Breastfeeding Practices: Report of an Informal Meeting; Division
of Child and Adolescent Health, WHO/CDD/SER/91.14.ed., World Health Organization: Geneva,
Switzerland, 1991; p. 14.
26. Glascoe, F.P. Technical Report for the BRIGANCE SCREENS; Educational Testing Institute and
Directorate of Health: Reykjavik, Iceland, 2008.
27. Glascoe, F.P. PEDS, Mat Foreldra á Þroska Barna (Parental Perception on Their Child’s
Developmental Outcomes). Educational Testing Institute and Directorate of Health: Reykjavik,
Iceland, 2009.
28. Gunnlaugsson, G.; Saemundsson, E. Að Finna Frávik í Þroska og Hegðun 5 Ára Barna (Detection
of Deviation in Development and Behaviour of Five-Year Old Children). In Ungir Íslendingar í
Ljósi Vísindanna; Children’s Ombudsman and University of Iceland: Reykjavik, Iceland, 2005;
pp. 237–245.
29. Glascoe, F.P. Using parents’ concerns to detect and address developmental and behavioral
problems. J. Soc. Pediatr. Nurs. 1999, 4, 24–35.
30. Glascoe, F.P. Detecting and addressing developmental and behavioral problems in primary care.
Pediatr. Nurs. 2000, 26, 251–257.
31. Glascoe, F.P. Evidence-based approach to developmental and behavioural surveillance using
parents’ concerns. Child. Care Health Dev. 2000, 26, 137–149.
32. Rydz, D.; Shevell, M.I.; Majnemer, A.; Oskoui, M. Developmental screening. J. Child Neurol.
2005, 20, 4–21.
33. Brothers, K.B.; Glascoe, F.P.; Robertshaw, N.S. PEDS: Developmental milestones—An accurate
brief tool for surveillance and screening. Clin. Pediatr. (Phila.) 2008, 47, 271–279.
34. Glascoe, F.P. Screening for developmental and behavioral problems. Ment. Retard. Dev. Disabil.
Res. Rev. 2005, 11, 173–179.
35. Glascoe, F.P. Are overreferrals on developmental screening tests really a problem? Arch. Pediatr.
Adolesc. Med. 2001, 155, 54–59.
36. Glascoe, F.P. The brigance infant and toddler screen: Standardization and validation. J. Dev.
Behav. Pediatr. 2002, 23, 145–150.
37. Directorate of Health and Primary Health Care Organisation of Greater Reykjavik Area. Ung- Og
Smábarnavernd. Leiðbeiningar um Heilsuvernd Barna 0–5 Ára (Preventive Child Health Services:
Guidelines for Preventive Health Services for Children 0–5 Years of Age). Gudmundsdóttir, S.,
Aradóttir, A.B., Gunnlaugsson, G., Eds.; The Organisations: Reykjavík, Iceland, 2010. Available
online: http://www.landlaeknir.is/servlet/file/store93/item21268/2ungbarnavernd_leidbeiningar_
12.05.13.pdf (accessed on 13 August 2013).
 114

38. American Academy of Pediatrics, Council on Children With Disabilities; Section on

Developmental Behavioral Pediatrics; Bright Futures Steering Committee; Medical Home
Initiatives for Children With Special Needs Project Advisory Committee. Identifying infants and
young children with developmental disorders in the medical home: An algorithm for
developmental surveillance and screening. Pediatrics 2006, 118, 405–420.
39. Glascoe, F.P.; Squires, J. Issues with the new developmental screening and surveillance policy
statement. Pediatrics 2007, 119, 861–862, discussion 862–863.
40. Anderson, L.M.; Shinn, C.; Fullilove, M.T.; Scrimshaw, S.C.; Fielding, J.E.; Normand, J.;
Carande-Kulis, V.G. The effectiveness of early childhood development programs: A systematic
review. Am. J. Prev. Med. 2003, 24, 32–46.
41. Glascoe, F.P. Parents’ evaluation of developmental status: How well do parents’ concerns identify
children with behavioral and emotional problems? Clin. Pediatr. (Phila.) 2003, 42, 133–138.
42. Sices, L.; Stancin, T.; Kirchner, L.; Bauchner, H. PEDS and ASQ developmental screening tests
may not identify the same children. Pediatrics 2009, 124, e640–e647.
43. Limbos, M.M.; Joyce, D.P. Comparison of the ASQ and PEDS in screening for developmental
delay in children presenting for primary care. J. Dev. Behav. Pediatr. 2011, 32, 499–511.
44. Kramer, M.S.; Chalmers, B.; Hodnett, E.D.; Sevkovskaya, Z.; Dzikovich, I.; Shapiro, S.;
Collet, J.P.; Vanilovich, I.; Mezen, I.; Ducruet, T.; et al. Promotion of Breastfeeding Intervention
Trial (PROBIT): A Randomized Trial in the Republic of Belarus. J. Am. Med. Assoc. 2001, 285,
413–420.
45. Zhou, S.J.; Baghurst, P.; Gibson, R.A.; Makrides, M. Home environment, not duration of breast-
feeding, predicts intelligence quotient of children at four years. Nutrition 2007, 23, 236–241.
46. Quigley, M.A.; Hockley, C.; Carson, C.; Kelly, Y.; Renfrew, M.J.; Sacker, A. Breastfeeding is
associated with improved child cognitive development: A population-based cohort study.
J. Pediatr. 2012, 160, 25–32.
47. Anderson, J.W.; Johnstone, B.M.; Remley, D.T. Breast-feeding and cognitive development: A
meta-analysis. Am. J. Clin. Nutr. 1999, 70, 525–535.
48. Fewtrell, M.S.; Morley, R.; Abbott, R.A.; Singhal, A.; Isaacs, E.B.; Stephenson, T.; MacFadyen, U.;
Lucas, A. Double-blind, randomized trial of long-chain polyunsaturated fatty acid
supplementation in formula fed to preterm infants. Pediatrics 2002, 110, 73–82.
49. Slykerman, R.F.; Thompson, J.M.; Becroft, D.M.; Robinson, E.; Pryor, J.E.; Clark, P.M.;
Wild, C.J.; Mitchell, E.A. Breastfeeding and intelligence of preschool children. Acta Paediatr.
2005, 94, 832–837.
50. Koletzko, B.; Brands, B.; Poston, L.; Godfrey, K.; Demmelmair, H. Early nutrition programming
of long-term health. Proc. Nutr. Soc. 2012, 71, 371–378.
51. Turck, D. Later effects of breastfeeding practice: The evidence. Nestle Nutr. Workshop Ser.
Pediatr. Program. 2007, 60, 31–39, discussion 39–42.
52. Wu, T.C.; Chen, P.H. Health consequences of nutrition in childhood and early infancy.
Pediatr. Neonatol. 2009, 50, 135–142.
53. Helland, I.B.; Smith, L.; Saarem, K.; Saugstad, O.D.; Drevon, C.A. Maternal supplementation
with very-long-chain n-3 fatty acids during pregnancy and lactation augments children’s IQ at
4 years of age. Pediatrics 2003, 111, e39–e44.
 115

54. Mitmesser, S.H.; Jensen, C.L. Roles of long-chain polyunsaturated fatty acids in the term infant:
Developmental benefits. Neonatal Netw. 2007, 26, 229–234.
55. Koletzko, B.; Rodriguez-Palmero, M. Polyunsaturated fatty acids in human milk and their role in
early infant development. J. Mammary Gland Biol. Neoplasia 1999, 4, 269–284.
56. Lanting, C.I.; Fidler, V.; Huisman, M.; Touwen, B.C.; Boersma, E.R. Neurological differences
between 9-year-old children fed breast-milk or formula-milk as babies. Lancet 1994, 344, 1319–1322.
57. Butte, N.F.; Lopez-Alarcon, M.G.; Garza, C. Nutrient Adequacy of Exclusive Breastfeeding for
the Term Infant during the First Six Months of Life; World Health Organisation: Geneva,
Switzerland, 2002; pp. 8–14.
58. Fewtrell, M.S.; Kennedy, K.; Singhal, A.; Martin, R.M.; Ness, A.; Hadders-Algra, M.; Koletzko, B.;
Lucas, A. How much loss to follow-up is acceptable in long-term randomised trials and
prospective studies? Arch. Dis. Child. 2008, 93, 458–461.
59. Iannotti, L.L.; Tielsch, J.M.; Black, M.M.; Black, R.E. Iron supplementation in early childhood:
Health benefits and risks. Am. J. Clin. Nutr. 2006, 84, 1261–1276.
60. Gunnarsson, B.S.; Thorsdottir, I.; Palsson, G.; Gretarsson, S.J. Iron status at 1 and 6 years versus
developmental scores at 6 years in a well-nourished affluent population. Acta Paediatr. 2007, 96,
391–395.

© 2013 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).
116
 117

Reprinted from Nutrients. Cite as: Liu, J.; Leung, P.; Yang, A. Breastfeeding and Active Bonding
Protects against Children’s Internalizing Behavior Problems. Nutrients 2014, 6, 76-89.

Breastfeeding and Active Bonding Protects against Children’s

Internalizing Behavior Problems
Jianghong Liu 1,*, Patrick Leung 2 and Amy Yang 3
1
Department of Family and Community Health, School of Nursing, University of Pennsylvania,
Philadelphia, PA 19104, USA
2
Department of Psychology, Chinese University of Hong Kong, Hong Kong, China;
E-Mail: pleung@cuhk.edu.hk
3
Department of Biostatistics and Epidemiology, University of Pennsylvania, Philadelphia,
PA 19104, USA; E-Mail: yyang@mail.med.upenn.edu

* Author to whom correspondence should be addressed; E-Mail: jhliu@nursing.upenn.edu;

Tel.: +1-215-898-8293.

Received: 4 October 2013; in revised form: 4 December 2013 / Accepted: 10 December 2013 /
Published: 24 December 2013

Abstract: Breastfeeding is associated with numerous health benefits to offspring and

mothers and may improve maternal-infant bonding. Ample evidence suggests breastfeeding
can improve child neurodevelopment, but more research is needed to establish whether
breastfeeding is linked to the development of child psychopathology. This paper aims to
explore the effects of both breastfeeding and mother-child interactions on child behavioral
outcomes at a later age. Children from the China Jintan Child Cohort Study (N = 1267), at
age six years old were assessed, along with their parents. Children who were breastfed
exclusively for a period of time in the presence of active bonding were compared to those
who were breastfed in the absence of active bonding as well as to children who were not
exclusively breastfed, with or without active bonding. Results from ANOVA and GLM,
using SPSS20, indicate that children who were breastfed and whose mothers actively
engaged with them displayed the lowest risk of internalizing problems (mean = 10.01,
SD = 7.21), while those who were neither exclusively breastfed nor exposed to active
bonding had the least protection against later internalizing problems (mean = 12.79,
SD = 8.14). The effect of breastfeeding on internalizing pathology likely represents a
biosocial and holistic effect of physiological, and nutritive, and maternal-infant
bonding benefits.
 118

Keywords: breastfeeding; bonding; attachment; internalizing behaviors

1. Introduction

Breastfeeding is associated with a wide range of positive health outcomes in children and mothers.
For example, a systematic review and meta-analysis of approximately 400 studies found that breastfeeding
was related to a reduced risk of acute ear infections, respiratory tract infections, asthma, obesity, type 1
and 2 diabetes mellitus, and childhood leukemia [1]. The production of prolactin and oxytocin during
breastfeeding is associated with lower levels of maternal stress and enhanced bonding [2].
Furthermore, early cessation of breastfeeding or not breastfeeding at all has been linked to an increased
risk of maternal postpartum depression [1].
One particular outcome of interest is that of cognitive development in breastfed children. Breast
milk is rich in vital nutrients, such as essential fatty acids, vitamins, minerals, and amino acids, that are
associated with improved cognitive functioning [3], language development [4], and overall neurological
development [2,5]. In addition, breastfeeding has been associated with improved mother-infant
bonding [6,7]. For instance, early feeding interactions between mother and infant may result in more
positive feeding experiences and produce greater maternal sensitivity and responsiveness to infant
needs [8].
Although previous studies have indicated a wealth of nutritional, physiological, and cognitive
benefits to children from breastfeeding, little has been done on emotional development and regulation.
It is known that childhood internalizing disorders, including depression and anxiety, can affect up to
about 20% of children and adolescents [9]. They also increase the risk of future psychopathology in
adulthood [10]. The identification of readily modifiable factors, such as breastfeeding, that may protect
against childhood internalizing behaviors is therefore important. Studies have not yet found a
relationship between breastfeeding and behavioral outcomes during early childhood [11,12]. However,
there are limited studies conducted in older age groups. Oddy et al. found that breastfeeding may be
associated with adverse mental health outcomes in early adolescents while Kwok et al. found
inconsistent associations [13,14]. However, few studies have tested the long term effect of both
breastfeeding as well as the mother-infant interaction during feeding on child behavioral outcomes.
The purpose of this study is to examine whether breastfeeding was related to fewer internalizing
disorders later in childhood in a large, community-based sample of Chinese children and parents, and
to understand whether breastfeeding and active bonding (i.e., verbal interactions during feeding) were
associated with the reduced risk of internalizing behaviors. Finally, this study will assess whether there
was any breastfeeding duration (or dosage) effect on internalizing behaviors.

2. Experimental Section

2.1. Participants and Sample Design

The current study was part of a larger population-based community cohort study of 1656 Chinese
children (55.5% boys, 44.5% girls) initially recruited from four preschools in the town of Jintan,
 119

located in the southeastern coastal region of Mainland China. Briefly, all children and parents taking
part in the original cohort study were invited to participate for assessment of children’s behaviors
while the children were in the final few months of their senior year in preschool (Spring 2005 to
Spring 2007). At that point, some children dropped out of the study because of changing schools or
because bio-markers were not obtained originally; therefore, only 1419 children in the original sample
were followed up in the later waves. Detailed sampling and research procedures of this larger cohort
study have been described elsewhere [15,16].
We excluded the cases where the children are over six years old (there are quite a few because it
was late in the school year) for the purpose of this analysis because we are using Child Behavior
Checklist (CBCL) 1.5 to 5 to measure behavioral outcome. We performed a comparison on measures
such as mother’s age when the children were born, neighborhood problems, gender, parent’s education
and occupation prestige, as well as whether parents are separated or divorced, and found there is no
difference except the older children tend to be boys and have parents who are less educated. We
acknowledge that Chinese tradition prefers to hold boys back in the earlier years of education because
of the cultural belief that boys develop and mature later than girls, and that more educated parents
would want to place their children in school as early as possible.
The analysis sample consists of 1267 complete data. The mean age of the analysis sample was
66.6 months (SD = 5, range = 50.0–71.9). This is close to the common kindergarten age in the US.

2.2. Measures

2.2.1. Internalizing Behavior Problems

We used the Internalizing Behavior scale from the Child Behavior Checklist (CBCL)/1.5-5 as the
measure for the dependent variable. The factor structures of CBCL/1.5-5 have been validated in our
previous study [17]. The internalizing behavior scale consists of 36 items out of 99 items in total,
including Emotionally reactive, Anxious/Depressed, Withdrawn and Somatic Complaints. Items are
rated on a 3-point scale (0 = not true, 1 = sometimes true, or 2 = often true) [18]. In this study, we
utilized both the full Internalizing Behavior scale and the four syndromes as dependent variables.
Alpha is 0.87 for the scale in our sample. We used raw scores on all behavioral assessments for the
analysis, as recommended by Achenbach [19].

2.2.2. Breastfeeding

Mothers completed a retrospective questionnaire that asked whether they breastfed (78.3%), used
formula (5.6%) or both (16%). In this study, we define exclusive breastfeeding as exclusive for a
period of time with a minimum of one month. Non-exclusive breastfeeding is defined to include
formula only or mixed methods. Mothers were also asked to report breastfeeding duration in months
(mean = 8.79, range 0–24). Duration was categorized into three levels: less or equal to 7 months
(25%), between 7 and 10 months (51%) and greater or equal to 10 months (24%).
 120

2.2.3. Maternal Interaction

Mothers were asked a follow-up question for (breast) feeding: “Did you talk to the child while
(breast) feeding in the first two years” These responses were: 1 = Never, 2 = Sometimes, 3 = Always.
In our study, not all mothers were always talking to the child while breastfeeding. We combined the
two measured into one and named it “Feeding types and bonding”, reflecting the beneficial effects of
both nutritional and active communication. In our sample, we identified four groups. 34.8% of the
mothers who always use breast milk and always talk to the child while feeding were classified into
Breastfed and Active bonding (group 1); 43.5% who never talk to the child even when they used breast
milk were assigned into Breastfed and no Active Bonding (group 2); 9% of the mothers were in the
group of Non-breastfed and Active Bonding (group 3), and the final 12.7% in the Non-breastfed and
no Active Bonding (group 4).

2.2.4. Social Adversity

Parents were asked to fill in a sociodemographic questionnaire at the same time they completed the
CBCL when children were six years old. A number of researchers have demonstrated the utility of
combining of several individual measurements of psychosocial risk factors in studying child behavior
problems [20,21]. The adversity index was created along lines similar to those described by Rutter and
Moffitt [20,21]. A total adversity score was derived based on 11 variables. This score was created by
adding 1 point (for 9 of the 11 indicators) or 2 points (for 2 indicators) for 11 adversity indicators:
mother’s low education (below middle school, 8.8%), father’s low education (below middle school,
5.4%), mother’s low occupational status (3-point scale: 0 = professional or skilled work, 22.2%;
1 = un-skilled worker, 38.9%; and 2 = no occupation, 31.8%), father’s low occupational status (3-point
scale: 0 = professional or skilled work, 26.8%; 1 = unskilled worker, 60.3%; and 2 = no job, 3.8%),
mother’s poor health status (2.1%), father’s poor health status (3.8%), obstetric complication (bleeding,
hypertension, diabetes, Caesarian section, difficult birth, low birth weight, difficulty breathing,
35.6%), divorce (3.3%), absence of biological mother (4.2%), house size below 70 m2 (8.4%), and
poor neighborhood (overcrowded neighborhood, noise pollution, damp, 35.6%). Details on these
indicators are given in Liu et al. [22]. The adversity score ranged from 0 to 13 (M = 3.51, SD = 2.04).

2.3. Statistical Methods

To test whether there is any social adversity or internalizing behavior score differences between
exclusive and non-exclusive breastfeeding groups, independent t tests were employed. Differences in
proportions for gender were tested using Ȥ2 tests. The same tests were conducted again to test
differences between active bonding and non-active bonding group. Cohen’s d was calculated to show
the size of the effect between two comparison groups.
To test for the combined effects of feeding and bonding types on internalizing behaviors, a series of
analysis of variance (ANOVA) were performed. “Feeding and bonding types” was a new variable
(four groups defined in the method section) created from the two independent binary variables.
It was the grouping variable in ANOVA, and the dependent variables were emotionally reactive,
 121

Anxious/Depressed, Somatic Complaints, Withdrawn and total internalizing problems respectively.

Omega squared w2 was computed as an index of effect size with several groups.
To test for the effect of social adversity as a potential confound, general linear models were fitted
with adversity entered as covariate. The effect moderator of sex was assessed by entering the measure
as a factor alongside feeding and bonding type. Gender × breastfeeding and bonding types interaction
term was included in the model as well. The interaction effect of breastfeeding types and bonding
types was tested in the models besides including these two variables independently in the models. To
test for a dose-response relationship between breastfeeding duration and internalizing behavior, the
grouping variable of duration took on three levels (” PRQWKV  –10 months and• PRQWKV  IRU
univariate ANOVAs. All results were considered significant if P < 0.05 using a two-tailed test.
Statistical analyses were conducted by using SPSS version 20.0 (IBM SPSS Statistics).

3. Results

3.1. Single Effect of Breastfeeding or Active Bonding

Of the 1267 participants with completed data, 77% exclusively breastfed their babies and 43.4%
always talk to their babies while feeding. Mean (SD) scores, effect sizes (Cohen’s d) and results of
specific t test comparisons for the two pairs of comparison groups are given in Table 1. The children
who were exclusively breastfed had significantly lower mean scores for somatic complaints and total
internalizing problems. Active bonded children had significant lower scores across emotionally
reactive, anxious/depressed, withdrawn syndromes and total internalizing problems.

3.2. Combined Effect of Breastfeeding and Active Bonding

ANOVA results showed significant group differences on mean behavior scores for total
internalizing (F(3,1264) = 5.21; P = 0.001), anxious/depression (F(3,1263) = 2.779; P = 0.04), somatic
complains (F(3,1264) = 3.20; P = 0.023) and withdrawn (F(3,1264) = 6.75; P < 0.001). Emotionally
reactive showed borderline significance (F(3,1261) = 2.38; P = 0.068). Mean (SD) scores, effect sizes
(omega squared w2) and P values from ANOVA for the four comparison groups are given in Table 2.
Results revealed identical trend across all dependent variables, with group 1 displaying the lowest
scores, followed by group 3. Group 4 had the highest scores.

3.3. Potential Confounds

Demographic and social variables are known to influence breastfeeding and children’s behavior.
Our study showed that boys were more likely be fed by formula or a mixed method rather than pure
breast milk (Ȥ2 = 2.17; P = 0.031). Social adversity scores were significantly lower for those whom
actively bonded with their baby while feeding (t = 5.18; P < 0.001), indicating they had a better
socioeconomic background and health status (Table 1). Consequently, it is possible that social
adversity and gender could account for the main effects of breastfeeding and bonding. We also tested
the correlations between potential confounders and dependent variables. In our sample, Spearman
correlation indicates social adversity was positively correlated to each syndrome and total internalizing
problems (P < 0.001 for all), although no significant correlation was detected between gender
 122

and dependent variables. As gender and social adversity were correlated to either the breastfeeding and
bonding types or the internalizing problems, we entered these two constructs in to a series of general
linear models. The main effects remained significant for Anxious/Depressed, Somatic Complaints,
Withdrawn and total internalizing (Table 3) after controlling for adversity (F(1,1252) • 
P ”IRUDOO DQGJHQGHU F(1,1252) ” 1.56; P • 0.212 for all).

3.4. Effect Moderator

There was no interaction between breastfeeding and bonding grouping and gender
(F(1,1252) ” 2.1; P • 0.097 for all), indicating that this measure did not moderate the effects of
breastfeeding and bonding.

3.5. Dose Response Relationship

Univariate ANOVAs (with three groups: 0–7 months, 7–10 months and 10 months duration levers)
were conducted on each syndrome and total internalizing scores. Results showed significant main
effect for breastfeeding duration on anxious/depression (F(2,1170) = 3.28; P = 0.038), somatic
complains (F(2,1171) = 3.25; P = 0.039) and total internalizing (F(2,1171) = 2.99; P = 0.051). No
significant dose-response effect was detected for emotionally reactive (F(2,1170) = 1.21; P = 0.298)
and withdrawn (F(2,1171) = 0.910; P = 0.403). The mean scores for each syndrome and total
internalizing problems were plotted against breastfeeding duration levels in Figure 1a,b.

4. Discussion

Three key findings emerged from this study. First, compared to children whose mothers breastfed
them, children who were not breastfed showed an increased number of internalizing behavioral
problems, particularly anxious/depressed and somatic symptoms. Second, the group of children whose
mothers both breastfed and actively interacted with their infants had the least likelihood of displaying
internalizing problems. Children who were not breastfed but whose mothers still engaged in active
interactions displayed the next-lowest risk, while being neither breastfed nor exposed to active bonding
had the smallest effect on internalizing behaviors. Finally, a duration effect (dosage effect) appeared
such that breastfeeding for 10 months or longer had the strongest impact on reducing anxious/depressed
and somatic symptoms in children.
Breastfeeding confers a strong biological benefit to infants and their development [23]. From a
nutritive standpoint, breast milk contains docosahexaenoic acid (DHA) omega-3 fats, the consumption
of which, along with eicosapentaenoic acid (EPA) fats, may reduce the risk for affective disorders,
including major depression and bipolar disorders, particularly among women [24,25]. However, the
overall literature on DHA and depression remains mixed [26,27]. What is known, however, is that
DHA plays a vital role in neural development, neurotransmitter transmission, and genetic expression,
making it highly relevant to child neurodevelopment as well as developmental disorders, such as
attention-deficit/hyperactivity disorder and motor deficits [28].
 123

Table 1. Descriptives of the Breastfeeding and internalizing behavior Outcome (N = 1267).

Breastfeeding type Active bonding
Variables Exclusive Non-exclusive Effect Size P Always 43.4% Never/Sometimes Effect Size P
77% 23% (Cohen’s d) 56.6% (Cohen’s d)
Emotionally Reactive 2.63 (2.39) 2.85 (2.38) í 0.168 2.52 (2.30) 2.82 (2.43) í 0.031
Anxious/Depressed 3.22 (2.28) 3.43 (2.30) í 0.175 3.13 (2.17) 3.44 (2.36) í 0.02
Somatic Complaints 2.64 (2.47) 3.01 (2.34) í 0.025 2.60 (2.33) 2.86 (2.54) í 0.076
Withdrawn 2.18 (2.34) 2.43 (2.40) í 0.11 1.89 (2.19) 2.52 (2.44) í <0.001
Total Internalizing 10.66 (7.64) 11.7 (7.53) í 0.04 10.14 (7.05) 11.63 (7.94) í 0.001
Boy % 53.3 60.4 \ 0.031 57.3 53.1 \ 0.141
Social Adversity 3.53 (2.03) 3.48 (2.11) 0.0241 0.701 3.17 (1.83) 3.77 (2.15) í <0.001
Note: significant results were highlighted in bold; \ not available.

Table 2. Combined Effect of Breastfeeding and Active Bonding on Behavior Problems from ANOVA.
Group 1 Group 2 Group 3 Group 4
Variables Exclusive Exclusive Non-exclusive Non-exclusive Effect Size (omega P
Breastfeeding with Breastfeeding without Breastfeeding with Breastfeeding without squared w2)
Active Bonding 34.8% Active Bonding 43.5% Active Bonding 9% Active Bonding 12.7%
Emotionally Reactive 2.48 (2.30) 2.73 (2.45) 2.62 (2.22) 3.07 (2.50) 0.003 0.068
Anxious/Depressed 3.11 (2.17) 3.30 (2.34) 3.15 (2.15) 3.71 (2.41) 0.004 0.04
Somatic Complaints 2.56 (2.39) 2.70 (2.53) 2.72 (2.14) 3.27 (2.50) 0.005 0.023
Withdrawn 1.87 (2.20) 2.40 (2.41) 2.13 (2.22) 2.73 (2.53) 0.013 <0.001
Total Internalizing 10.01 (7.21) 11.12 (7.90) 10.61 (6.66) 12.79 (8.14) 0.01 0.001
Note: significant results were highlighted in bold.
 124

Table 3. General linear model statistics of breastfeeding type and bonding against
internalizing problems, controlling for gender and social adversity.
Dependent Variables Type III Sum of Squares df Mean square F Significance
Emotionally Reactive 34.4 3, 1255 11 2.045 0.106
Anxious/Depressed 38.12 3, 1257 12.71 2.471 0.06
Somatic Complaints 54.6 3, 1258 18.2 3.108 0.026
Withdrawn 90.07 3, 1258 32.36 6.016 <0.001
Total Internalizing 774.39 3, 1258 258.13 4.573 0.003
Note: significant results (P<0.05) were highlighted in bold.

Figure 1. Dose-response Relationship between duration of breastfeeding and internalizing

behavior. (a) Breastfeeding duration and internalizing syndromes; (b) Breastfeeding
duration and total internalizing problems.

(a)

(b)
 125

It may be that the biological benefit offered to breastfed infants plays a role in healthy cognitive
maturation which in turn lowers their risk for psychopathology [29]. Feldman and Eidelman report that
breastfeeding is associated with improved motor and social skills [5], but other authors have not found
an impact on emotional regulation and behavioral disruption, indicating the need for further research
on breastfeeding and child psychodevelopment [30]. Interestingly, a recent study examined effects of
breastfeeding on mental health outcomes among children at age 14 years and found that breastfeeding
at age six months was associated with a lower rate of child psychopathology, including social and
attention difficulties and aggression [31]. However, more longitudinal data is needed to better
understand the potential long-term benefits of breastfeeding to child mental health. Whether the
nutritional, physiological and cognitive benefits from breastfeeding directly enhance mental health via
a biological route in our brain may still require further exploration. However, speculatively, the
reported nutritional, physiological and cognitive benefits can confer a lot of advantages to a child to
negotiate with the challenges of growing up. For example, a healthier, energetic physical body and a
faster cognitive growth can help a child to cope with the arduous demands of modern-day schooling,
particularly in mainland China, the tradition of which has long emphasized education as the avenue for
upward socio-economic migration. A child who excels at school will also be well liked and accepted
by parents, relatives, teachers and peers. School success and social popularity are both known key
precursors to mental health [32]. Thus, it is likely that there may be both a direct biological route and
an indirect psychosocial route from breastfeeding leading to positive mental health or fewer
internalizing problems. These dual routes should both be further examined in future study.
Breastfeeding also provides a biological benefit to the mother by reducing blood pressure and
pain [33]. Furthermore, the release of hormones oxytocin and prolactin not only confer analgesic and
relaxation benefits, but they also appear to play a key role in mother-infant bonding [11,33], which has
been shown to reduce emotional and behavior problems in children [6,7].
The attachment aspect of breastfeeding underscores the need to consider its potential mental health
benefits. Psychologically, breastfeeding may enhance the mother-infant bonding process via active
talking, eye contact, and skin-to-skin touch. This may help mothers form stronger attachments to
offspring and improve maternal sensitivity [33], reduce postpartum depression [34], and ward off other
negative mood states like maternal stress [35]. This may indirectly benefit a child’s mental health, as
the literature detailing the impact of maternal depression on increasing the risk of future child and
adolescent psychopathology is compelling [36,37].
Infants may similarly derive a mental health benefit from being breastfed, including development
of more secure attachments and reduced negative temperament [38]. Several authors have
documented analgesic properties of breast milk, along with reductions in salivary cortisol, due to milk
odor and skin-to-skin contact [39,40]; these are hypothesized to help alleviate child distress and
strengthen bonding.
Taken together, these findings underlie a biopsychological aspect of breastfeeding wherein the
physiological benefits of breastfeeding (e.g., pain reduction, stress reduction, healthy cognitive
development) coupled with improved pair bonding and mother-infant attachments may provide
protective effects against the formation of child internalizing behaviors The biopsychosocial interaction
may also provide indirect benefits that operate through mediating or moderating variables [41]. For
example, secure parent-child attachment may improve child sleep quality [42], and reduced sleep
 126

problems in children has been linked to better emotional and behavioral functioning [43]. In addition,
breastfeeding is ultimately a holistic process and there are several aspects that facilitate the process,
including how the mother responds to the infant, the physical and social environment around the
mother and baby, and the level nutrients in the breast milk. These factors, such as the genetic and
environmental influences of nutrient intake (e.g., breast milk) should be considered [44]. Consideration
of such in larger samples will require further study before definitive conclusions can be drawn about
intervening variables on breastfeeding and internalizing conditions.
In this study, breastfeeding for a longer duration (at least 10 months) had the greatest effect on
reducing internalizing symptoms. This is consistent with other authors who report that longer duration
of breastfeeding was associated with greater protection against child mental health problems at age
five years [45]. Another recent longitudinal cohort study from Oddy and colleagues followed breastfed
infants to 14 years of age and found that breastfeeding for six months or less independently predicted
greater externalizing and internalizing problems in childhood and adolescence, compared to infants
who were breastfed for 6 months duration or longer [14].
Some important limitations on the present study’s findings exist. First, the use of retrospective data
may involve recall bias. However, in the current literature, it is not rare for studies examining
breastfeeding practices to use maternal recall data after much longer periods. For example, a study by
Promislow looked at maternal recall of breastfeeding duration of elderly US women from 34 to
50 years ago [46]. Nevertheless, future prospective designs should be considered. As previously
mentioned, we also did not include holistic measures in our assessment of breastfeeding as we do not
have data available due to the retrospective nature of this study. Instead, we included confounding
factors, such as demographic and social background, in our analysis of breastfeeding. Another
limitation of the present study is that it does not take into consideration the exact duration of exclusive
breastfeeding, which is particularly important given the fact that breastfeeding practices have
decreased in recent years, especially among urban and well-educated mothers [47]. Future studies
should test if there is a duration-dependent relationship between breastfeeding and internalizing
behavior in children. Additionally, as active bonding was one of the key predictors, future studies
should employ validated, empirical-based measures on this construct. However, despite the use of such
“crude” measures, they are able to produce consistent results, indicating evidently the benefits of
breastfeeding on the children’s mental health. Furthermore, future studies should stratify by region,
given that breastfeeding practices differ by location.

5. Conclusions

These results indicate that children who were breastfed and exposed to active bonding during
feeding displayed the lowest risks of internalizing behavior problems at age six years. Increased
duration of breastfeeding (•PRQWKV FRXOGDOVRKHOSORZHULQWHrnalizing problems in children (i.e., a
dosage effect). These findings were independent of several socio-demographic/family characteristics,
as well as gender. It is possible that both nutrients (e.g., fatty acids) and maternal bonding interactively
work to promote optimal neurodevelopment in early childhood, subsequently protecting children from
internalizing disorders, such as depression, anxiety, and somatic complaints [48]. The promotion of
 127

active bonding practices during feeding (whether breastfeeding or formula feeding) may help reduce
later internalizing behaviors in children by enhancing attachment between the mother and infant.

Acknowledgement

This study was supported by funding from the National Institute of Environmental Health
Sciences/NIH/NIEHS 1-K02-ES-019878 and 1-R01-ES-018858. The funders had no role in study
design, data collection and analysis, decision to publish, or preparation of the manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

References

1. Ip, S.; Chung, M.; Raman, G.; Chew, P.; Magula, N.; Devine, D.; Trikalinos, T.; Lau, J.
Breastfeeding and maternal and infant health outcomes in developed countries. Evid. Rep.
Technol. Assess. 2007, 153, 1–186.
2. Uauy, R.; de Andraca, I. Human milk and breast feeding for optimal mental development. J. Nutr.
1995, 125, 2278S–2280S.
3. Quinn, P.J.; O’Callaghan, M.; Williams, G.M.; Najman, J.M.; Andersen, M.J.; Bor, W. The effect
of breastfeeding on child development at 5 years: A cohort study. J. Paediatr. Child Health 2001,
37, 465–469.
4. Whitehouse, A.J.; Robinson, M.; Li, J.; Oddy, W.H. Duration of breastfeeding and language
ability in middle childhood. Paediatr. Perinat. Epidemiol. 2011, 25, 44–52.
5. Feldman, R.; Eidelman, A.I. Direct and indirect effects of breast milk on the neurobehavioral and
cognitive development of premature infants. Dev. Psychobiol. 2003, 43, 109–119.
6. Zetterstrom, R. Breastfeeding and infant-mother interaction. Acta Paediatr. Suppl. 1999, 88, 1–6.
7. Fergusson, D.M.; Woodward, L.J. Breast feeding and later psychosocial adjustment. Paediatr.
Perinat. Epidemiol. 1999, 13, 144–157.
8. DeWitt, S.J.; Sparks, J.W.; Swank, P.B.; Smith, K.; Denson, S.E.; Landry, S.H. Physical growth
of low birthweight infants in the first year of life: Impact of maternal behaviors. Early Hum. Dev.
1997, 47, 19–34.
9. Costello, E.J.; Mustillo, S.; Erkanli, A.; Keeler, G.; Angold, A. Prevalence and development of
psychiatric disorders in childhood and adolescence. Arch. Gen. Psychiatry 2003, 60, 837–844.
10. Kovacs, M.; Devlin, B. Internalizing disorders in childhood. J. Child Psychol. Psychiatry 1998,
39, 47–63.
11. Kramer, M.S.; Fombonne, E.; Igumnov, S.; Vanilovich, I.; Matush, L.; Mironova, E.;
Bogdanovich, N.; Tremblay, R.E.; Chalmers, B.; Zhang, X.; et al. Effects of prolonged and
exclusive breastfeeding on child behavior and maternal adjustment: Evidence from a large,
randomized trial. Pediatrics 2008, 121, e435–e440.
12. Papp, L.M. Longitudinal associations between breastfeeding and observed mother-child
interaction qualities in early childhood. Child Care Health Dev. 2013, doi:10.1111/cch.12106.
 128

13. Kwok, M.K.; Leung, G.M.; Schooling, C.M. Breast feeding and early adolescent behaviour,
self-esteem and depression: Hong Kong’s “Children of 1997” birth cohort. Arch. Dis. Child.
2013, 98, 887–894.
14. Oddy, W.H.; Kendall, G.E.; Li, J.; Jacoby, P.; Robinson, M.; de Klerk, N.H.; Silburn, S.R.;
Zubrick, S.R.; Landau, L.I.; Stanley, F.J. The long-term effects of breastfeeding on child and
adolescent mental health: A pregnancy cohort study followed for 14 years. J. Pediatr. 2010, 156,
568–574.
15. Liu, J.; McCauley, L.A.; Zhao, Y.; Zhang, H.; Pinto-Martin, J. Cohort profile: The China Jintan
child cohort study. Int. J. Epidemiol. 2010, 39, 668–674.
16. Liu, J.; McCauley, L.; Leung, P.; Wang, B.; Needleman, H.; Pinto-Martin, J. Community-based
participatory research (CBPR) approach to study children’s health in China: Experiences and
reflections. Int. J. Nurs. Stud. 2011, 48, 904–913.
17. Liu, J.; Cheng, H.; Leung, P.W. The application of the preschool child behavior checklist and the
caregiver-teacher report form to mainland Chinese children: Syndrome structure, gender
differences, country effects, and inter-informant agreement. J. Abnorm. Child Psychol. 2011, 39,
251–264.
18. Achenbach, T.; Rescorla, L. Manual for the ASEBA Preschool Forms & Profiles; University of
Vermont Research Center for Children, Youth & Families: Burlington, VT, USA, 2000.
19. Achenbach, T.M. Manual for the Child Behavior Checklist/4–18 and 1991 Profile; University of
Vermont, Department of Psychiatry: Burlington, VT, USA, 1991.
20. Rutter, M. Family, Area and School Influences in the Genesis of Conduct Disorders: Aggression
and Anti-Social Behaviour in Child-Hood and Adolescence; Hersov, L.A., Berger, M., Shaffer,
D., Eds.; Pergamon: Oxford, UK, 1978; pp. 95–113.
21. Moffit, T.E. Juvenile delinquency and attention deficit disorder: Boys’ developmental trajectories
from age 3 to age 15. Child Dev. 1990, 61, 893–910.
22. Liu, J.; Portnoy, J.; Raine, A. Association between a marker for prenatal testosterone exposure
and externalizing behavior problems in children. Dev. Psychopathol. 2012, 24, 771–782.
23. Rautava, S.; Walker, W.A. Academy of breastfeeding medicine founder’s lecture 2008:
Breastfeeding—An extrauterine link between mother and child. Breastfeed Med. 2009, 4, 3–10.
24. McNamara, R.K.; Jandacek, R.; Rider, T.; Tso, P.; Dwivedi, Y.; Pandey, G.N. Selective deficits
in erythrocyte docosahexaenoic acid composition in adult patients with bipolar disorder and major
depressive disorder. J. Affect. Disord. 2010, 126, 303–311.
25. Colangelo, L.A.; He, K.; Whooley, M.A.; Daviglus, ML.; Liu, K. Higher dietary intake of
long-chain omega-3 polyunsaturated fatty acids are inversely associated with depressive
symptoms in women. Nutrition 2009, 25, 1011–1019.
26. Makrides, M.; Gibson, R.A.; McPhee, A.J.; Yelland, L.; Quinlivan, J.; Ryan, P. Effect of DHA
supplementation during pregnancy on maternal depression and neurodevelopment of young
children: A randomized controlled trial. JAMA 2010, 304, 1675–1683.
27. Murakami, K.; Miyake, Y.; Sasaki, S.; Tanaka, K.; Arakawa, M. Fish and n-3 polyunsaturated
fatty acid intake and depressive symptoms: Ryukyus child health study. Pediatrics 2010, 126,
e623–e630.
 129

28. Kennedy, D.O.; Jackson, P.A.; Elliot, J.M.; Scholey, A.B.; Robertson, B.C.; Greer, J.; Tiplady,
B.; Buchanan, T.; Haskell, C.F. Cognitive and mood effects of 8 weeks’ supplementation with
400 mg or 1000 mg of the omega-3 essential fatty acid docosahexaenoic aicid (DHA) in healthy
children aged 10–12 years. Nutr. Neurosci. 2009, 12, 48–56.
29. Kramer, M.S.; Aboud, F.; Mironova, E.; Vanilovich, I.; Platt, R.W.; Matush, L.; Igumnov, S.;
Fombonne, E.; Bogdanovich, N.; Ducruet, T.; et al. Breastfeeding and child cognitive
development: A new evidence from a large randomized trial. Arch. Gen. Psychiatry 2008, 65,
578–584.
30. Hayatbakhsh, M.R.; O’Callaghan, M.J.; Bor, W.; Williams, G.M.; Najman, J.M. Association of
breastfeeding and adolescents’ psychopathology: A large prospective study. Breastfeed Med.
2012, 7, 480–486.
31. Rutter, M.; Rutter, M. Developing Minds: Challenge and Continuity Across the Life Span;
Basic Books: New York, NY, USA, 1993.
32. Gribble, K.D. Mental health, attachment and breastfeeding: Implications for adopted children and
their mothers. Int. Breastfeed J. 2006, 1, 5.
33. Britton, J.R.; Britton, H.L.; Gronwaldt, V. Breastfeeding, sensitivity, and attachment. Pediatrics
2006, 118, e1436–e1443.
34. Field, T.; Diego, M.; Hernandez-Reif, M.; Figueirdo, B.; Ezell, S.; Siblalingappa, V. Depressed
mothers and infants are more relaxed during breastfeeding versus bottlefeeding interactions: Brief
report. Infant Behav. Dev. 2010, 33, 241–244.
35. Groër, M.W. Differences between exclusive breastfeeders, formula-feeders, and controls: A study
of stress, mood, and endocrine variables. Biol. Res. Nurs. 2005, 7, 106–117.
36. Rahman, A.; Harrington, R.; Bunn, J. Can maternal depression increase infant risk of illness and
growth impairment in developing countries? Child Care Health Dev. 2002, 28, 51–56.
37. Hay, D.F.; Pawlby, S.; Waters, C.S.; Sharp, D. Antepartum and postpartum exposure to maternal
depression: Different effects on different adolescent outcomes. J. Child Psychol. Psychiatry 2008,
49, 1079–1088.
38. Niegel, S.; Ystrom, E.; Hagtvet, K.A.; Vollrath, M.E. Difficult temperament, breastfeeding, and
their mutual prospective effects: The Norwegian Mother and Child Cohort Study. JDBP 2008, 29,
458–462.
39. Nishitani, S.; Miyamura, T.; Tagawa, M.; Sumi, M.; Takase, R.; Doi, H.; Moriuchi, H.; Shinohara, K.
The calming effect of a maternal breast milk odor on the human newborn infant. Neurosci. Res.
2009, 63, 66–71.
40. Gray, L.; Watt, L.; Blass, E.M. Skin-to-skin contact is analgesic in healthy newborns. Pediatrics
2000, 105, e14.
41. Liu, J. Early health risk factors for violence: Conceptualization, review of the evidence, and
implications. Aggress. Violent Behav. 2011, 16, 63–73.
42. Vaughn, B.E.; El-Sheikh, M.; Shin, N.; Elmore-Staton, L.; Krzysik, L.; Monteiro, L. Attachment
representations, sleep quality and adaptive functioning in preschool age children. Attach. Hum.
Dev. 2011, 13, 525–540.
43. Liu, J.; Zhou, G.; Wang, J.; Ai, Y.; Pinto-Martin, J.; Liu, X. Sleep problems, fatigue, and
cognitive performance in Chinese kindergarten children. J. Pediatr. 2012, 161, 520–525.e2.
 130

44. Liu, J.; Tuvblad, C.; Raine, A.; Baker, L. Genetic and environmental influences on nutrient intake.
Genes Nutr. 2013, 8, 241–252.
45. Robinson, M.; Oddy, W.; Li, J.; Kendall, G.E.; de Klerk, N.H.; Silburn, S.R.; Zubrick, S.R.;
Newnham, J.P.; Stanley, F.J.; Mattes, E. Pre- and postnatal influences on preschool mental health:
A large-scale cohort study. J. Child Psychol. Psychiatry 2008, 49, 1118–1128.
46. Promislow, J.; Gladen, B.; Sandler, D. Maternal recall of breastfeeding duration by elderly
women. Am. J. Epidemiol. 2005, 161, 289–296.
47. Liu, J.; Shi, Z.; Spatz, D.; Loh, R.; Sun, G.; Grisso, J. Social and demographic determinants for
breastfeeding in a rural, suburban and city area of South East China. Contemp. Nurse 2013, 45,
234–243.
48. Liu, J.; Chen, X.; Lewis, G. Childhood internalizing behavior: Analysis and implications.
J. Psychiatr. Ment. Health Nurs. 2011, 18, 884–894.

© 2013 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).
 131

3. Infants

Reprinted from Nutrients. Cite as: Thoene, M.; Hanson, C.; Lyden, E.; Dugick, L.; Ruybal, L.;
Anderson-Berry, A. Comparison of the Effect of Two Human Milk Fortifiers on Clinical Outcomes in
Premature Infants. Nutrients 2014, 6, 261-275.

Comparison of the Effect of Two Human Milk Fortifiers on

Clinical Outcomes in Premature Infants
Melissa Thoene 1, Corrine Hanson 2, Elizabeth Lyden 3, Laura Dugick 1, Leslie Ruybal 4 and
Ann Anderson-Berry 4,*
1
Newborn Intensive Care Unit, Nebraska Medical Center, Omaha, NE 68198, USA;
E-Mails: mthoene@nebraskamed.com (M.T.); ldugick@nebraskamed.com (L.D.)
2
School of Allied Health Professionals, University of Nebraska Medical Center,
984045 Nebraska Medical Center, Omaha, NE 68198-4045, USA; E-Mail: ckhanson@unmc.edu
3
College of Public Health, University of Nebraska Medical Center, 984375 Nebraska Medical
Center, Omaha, NE 68198-4375, USA; E-Mail: elyden@unmc.edu
4
Department of Pediatrics, University of Nebraska Medical Center, Omaha, NE 68198-2185, USA;
E-Mail: leslie.ruybal@unmc.edu

* Author to whom correspondence should be addressed; E-Mail: alanders@unmc.edu;

Tel.: +1-402-559-6750.

Received: 26 October 2013; in revised form: 17 December 2013 / Accepted: 20 December 2013 /
Published: 3 January 2014

Abstract: The use of human milk fortifiers (HMF) helps to meet the high nutritional
requirements of the human milk-fed premature infant. Previously available powdered
products have not met the protein requirements of the preterm infant population and many
neonatologists add powder protein modulars to help meet protein needs. The use of
powdered products is discouraged in neonatal intensive care units (NICU) due to concern
for invasive infection. The use of a commercially available acidified liquid product with
higher protein content was implemented to address these two concerns. During the course
of this implementation, poor growth and clinically significant acidosis of infants on
Acidified Liquid HMF (ALHMF) was observed. The purpose of this study was to quantify
those observations by comparing infant outcomes between groups receiving the ALHMF
vs. infants receiving powdered HMF (PHMF). A retrospective chart review compared
outcomes of human milk-fed premature infants <2000 g receiving the ALHMF (n = 23)
and the PHMF (n = 46). Infant growth, enteral feeding tolerance and provision, and
 132

incidence of necrotizing enterocolitis (NEC), metabolic acidosis, and diaper dermatitis

were compared between the two groups. No infants were excluded from this study based
on acuity. Use of ALHMF resulted in a higher incidence of metabolic acidosis (p = 0.002).
Growth while on HMF as measured in both g/kg/day (10.59 vs. 15.37, p < 0.0001) and in
g/day (23.66 vs. 31.27, p = 0.0001) was slower in the ALHMF group, on increased mean
cal/kg/day (128.7 vs. 117.3, p = 0.13) with nearly twice as many infants on the ALHMF
requiring increased fortification of enteral feedings beyond 24 cal/ounce to promote
adequate growth (48% vs. 26%, p = 0.10). Although we were not powered to study NEC as
a primary outcome, NEC was significantly increased in the ALHMF group. (13% vs. 0%,
p = 0.03). Use of a LHMF in an unrestricted NICU population resulted in an increase in
clinical complications within a high-acuity NICU, including metabolic acidosis and poor
growth. Although further research is needed to assess outcomes among infants with a
variety of clinical acuities, gestational ages, and weights to confirm these findings, based
on this experience, caution is urged to avoid potential risks.

Keywords: prematurity; human milk; fortifier; infant feeding; growth; acidosis

1. Introduction

Infants born prematurely have increased nutrient needs compared to those born at term [1–3].
Nutrition-related goals for premature infants aim to mimic fetal nutrient accretion and growth
in utero [4], yet many develop extrauterine growth restriction (EUGR) [5].
Despite the availability of customized, nutrient-dense enteral formulas, the American Academy of
Pediatrics strongly supports the use of human milk for premature infants [6]. However, unfortified
human milk remains inadequate to meet the high nutrient requirements of premature infants [1,4,7].
Provision of unfortified human milk has subsequently been linked to suboptimal growth (development
of EUGR or growth < 15 g/kg/day), reduced bone density leading to osteopenia of prematurity and a
clinical diagnosis of rickets, and the secondary consequences of each [1,4].
The use of commercial human milk fortifiers (HMF) allows for a more optimal provision of
essential nutrients to meet premature infant requirements [1,4,7]. Macronutrient recommendations
for low birth weight premature infants vary, but consensus goal ranges suggest enteral intake
of 110–120 cal/kg/day and 3.4–4.4 g protein/kg/day [1]. Protein is specifically emphasized, as early
and higher provisions promote more desirable growth and clinical outcomes [8,9]. The use of HMF
has been shown to be both safe and effective in improving growth and nutrition status of premature
infants compared to unfortified human milk [7,10,11]. In recent years the use of HMF with additional
powdered protein modular has been presented as a method of supplying the preterm infant with the
recommended amount of enteral protein to provide improved linear growth and neurodevelopmental
outcomes [12,13].
Human milk fortifiers have primarily been available in powder form, although the United States
Food and Drug Administration discourages the use of powdered forms in the neonatal intensive care
units (NICU) secondary to contamination risk [14]. They additionally advise that “alternatives to
 133

powdered forms should be chosen when possible” [14]. To comply with this recommendation and
achieve improved protein intake, The Nebraska Medical Center (TNMC) NICU changed standard
human milk fortification practices with a powdered product when an acidified liquid HMF (ALHMF)
with improved protein delivery became available. However, in the four months following our initial
use of the ALHMF, clinical observations of infants receiving the ALHMF suggested an increased
feeding intolerance, increased incidence of metabolic acidosis, poor growth, and a need for higher
caloric densities of enteral feedings to promote adequate growth. Due to our concern for patient
outcomes, use of the ALHMF was discontinued. The purpose of this study is to objectively quantify
these clinical observations by comparing outcomes of infants receiving the ALHMF to those receiving
the originally-used PHMF. Our study also looked to identify potential risk factors for the development
of the observed clinical complications, as previous research evaluating the ALHMF also documented
changes in pH and CO2 when compared to a powder HMF (PHMF) [15].

2. Patients and Methods

2.1. Participants and Study Design

The institutional review board at the University of Nebraska Medical Center (Omaha, NE, USA)
approved this study. Data was retrospectively collected from inpatient electronic medical records of all
infants admitted to the NICU, between October 2009 and July 2011, if they met the following
inclusion criteria: birth weight (BW) < 2000 g, received enteral feedings as fortified maternal breast
milk during NICU stay, and remained in the NICU • 14 days. Exclusion criteria included infants with
congenital abnormalities or conditions that significantly inhibited growth, such as Trisomy 13. No
infants were excluded based on clinical acuity. After extensive chart review, 69 infants were eligible
for the study.

2.2. Comparison and Use of Human Milk Fortifiers

Maternal breast milk (MBM) was fortified according to manufacturer directions. Ingredient and
estimated nutrient compositions of fortified preterm human milk were obtained from online nutritional
references [16,17]. Table 1 provides a composition comparison for key nutrients and ingredients.
 134

Table 1. Comparison of ingrediants and key nutrients using powder and liquid HMF.
24-Calorie-Per-Ounce Fortified Premature Human Milk [16,17]
Per 100 mL Powder HMF Liquid HMF
Protein (g) 2.35 3.2
Iron (mg) 0.46 1.85
Calcium (mg) 138 141
Phosphorus (mg) 78 78
Vitamin D (IU) 119 200
pH - 4.7
Primary Fortifier nonfat milk, whey protein water, whey protein isolate hydrolysate
Macronutrient Ingredients concentrate, corn syrup solids, (milk), medium chain triglycerides
medium-chain triglycerides (MCT oil), vegetable oil
(MCT oil) (soy and high oleic sunflower oils)
-: Information not avaliable.

Enteral feedings are initiated in this NICU within the first one to three days of life with Human
Milk (MBM as available or donor milk form the Milk Bank of Austin) at 20 mL/kg/day, trophic
feedings are continued for three to five days at the discretion of the attending neonatologist, and
then feedings are advanced daily by 20 mL/kg/day with human milk fortification beginning at
80–100 mL/kg/day enteral volume. A protein modular is utilized to improve protein intake to
approximately 4 g/kg/day enteral protein once caloric density is 24 kcal/oz. While using the ALHMF,
no additional protein modular was utilized. There were no other nutrition differences during the two time
periods. Nutrition is managed closely per unit protocol and is very consistent from provider to provider.
According to unit policy, infants receiving the PHMF also received supplementation with a protein
modular to provide approximately 4 g protein/kg/day when fed at goal volumes.
Sole use of the ALHMF was initiated on April 1, 2011. Infants receiving the PHMF before this date
of fortification change were included in the control group (PHMF, n = 46). Infants receiving the
ALHMF following this date were included in the study group (ALHMF, n = 23). Infants transitioned
from the PHMF to the LHMF on the date of fortification change were excluded.

2.3. Data Collection

Four investigators familiar with the electronic medical record and NICU terminology obtained all
data in a consistent predetermined manner. Collected information was reviewed for accuracy and
corrected if an electronic error occurred. All available information on each infant was included in the
analysis and is displayed in the tables.

2.4. Demographics

Demographic information was collected for all infants including gender, gestational age at birth and
discharge, and day of life (DOL) at discharge. Additional clinical outcomes were collected including
the presence of bronchopulmonary dysplasia (BPD), retinopathy of prematurity (ROP), intraventricular
hemorrhage (IVH), necrotizing enterocolitis (NEC), diaper dermatitis, and death. Treatment
 135

requirements were coded similarly if an infant required: oxygen at 36 weeks estimated gestational age
(EGA), ROP procedure, IVH shunt, Avastin treatment, Dexamethasone use, and Bicitra use. ROP
stage, IVH grade, and number of days of Dexamethasone use were included if available.

2.5. Anthropometrics

Infants were weighed daily on a gram scale, and head circumference and length (centimeters)
were recorded weekly by nursing staff using a measuring tape. Fenton growth curve percentile
rankings [18,19] were electronically plotted for each recorded anthropometric measurement. Weight, head
circumference, and length measurements with associated Fenton percentile rankings were taken for
infants at birth and at 36 weeks EGA, if available.

2.6. Nutrition

Enteral feeding data collected included day of life (DOL) enteral feedings were initiated, DOL full
enteral feedings were reached (with a discontinuation of parenteral nutrition support), and the number
of times enteral feedings were held (not secondary to preparation for a procedure). Maximum caloric
density and number of days on enteral feedings >24 cal/ounce were collected for infants requiring
caloric densities higher than the standard 24 cal/ounce to promote adequate growth.
Daily average provision of calories and protein (g) per kg body weight were calculated for infants
in each group if they received •50% of enteral feedings as fortified MBM during NICU stay. These
averages were taken when fortified enteral feedings reached a minimum of 140 mL/kg/day until either
daily intake was consistently less than this amount, the infant was changed to unfortified MBM, or the
infant received greater than 50% infant formula. Growth and nutrition was evaluated for the groups
comparing only growth during the period where the infant received •50% of enteral feedings as
fortified MBM. An electronic medical system (Intuacare®: Omaha, NE, USA) contained protein
references for breast milk and specified enteral formulas and caloric density. Nursing staff documented
daily intake of breast milk or specified enteral formulas, thus, daily calorie and protein provision per
kg of body weight were electronically calculated using the daily-recorded weight. The electronic
medical system also calculated the percentages of MBM vs. infant formula received according to
nursing documentation.

2.7. Laboratory Measurements

Maximum creatinine, maximum blood urea nitrogen (BUN) level, maximum base deficit value,
maximum calcium level, and lowest carbon dioxide (CO2) lab values were collected, if available, after
DOL 14 and DOL 30 for all infants. Values were not collected before DOL 14 to eliminate those
reflective of parenteral nutrition support and unfortified enteral feedings. Phosphorus and pH were not
consistently nor routinely obtained in this patient population and were therefore not collected in this
retrospective study.

2.8. Data Analysis

 136

Descriptive statistics were displayed for all variables by type of milk (powder vs. liquid) given. The
Wilcoxon rank sum test was used to compare continuous data between the milks groups. Associations
of categorical variables were assessed with the Fisher’s exact test. A p-value ” 0.05 was considered
statistically significant. To assess the difference in growth patterns between infants given powder and
infants given liquid, a mixed effects model was used. We included random slopes and intercepts for
each subject to capture individual growth pattern as well as fixed effects for group and day and a group
day interaction term. A significant interaction of day and group indicates differing growth patterns
based on group. Growth Velocity (GV) was calculated using Equation 1 [20].
GV = [1000 × ln(Wn/W1)]/(Dn í D1) (1)

3. Results

There were 46 infants in the PHMF group (21 males, 25 females) and 23 infants in the ALHMF
group (13 males, 10 females) (p = 0.45). Additional baseline characteristics were not statistically
significant between the two groups, as shown in Table 2. Enteral feeding data, growth and analyzed lab
values are displayed in Table 3. Clinical outcomes are displayed in Table 4. ROP stage, IVH grade, and
number of days of Dexamethasone use were not statistically significant and are not included in Table 4.

Table 2. Baseline characteristics of the subjects.

PHMF ALHMF
Variable p-value
n Mean SD (±) n Mean SD (±)
CGA at Birth 46 29.5 3.0 23 30.3 2.5 0.21
Birth Weight (g) 46 1293.7 407.5 23 1437.3 375.6 0.13
Birth Weight Percentile 46 31.4 24.7 23 36 26.5 0.82
Weight at 36 Weeks CGA (g) # 44 2245.9 450.72 18 2071.2 367.4 0.17
Weight Percentile at 36 Weeks CGA # 44 18.6 24.4 18 10.3 13.8 0.22
HC at Birth (cm) 46 27.2 3.4 22 27.9 2.1 0.19
HC Percentile at Birth 46 29.9 23.1 22 33.6 26.3 0.7
HC at 36 Weeks CGA (cm) # 42 32.5 2.6 19 31.9 1.5 0.37
HC Percentile at 36 Weeks CGA # 42 38.8 30.7 19 31.4 24.6 0.5
Length at Birth (cm) 46 38.6 3.9 21 40.4 2.8 0.07
Length Percentile at Birth 46 31.4 24.6 22 32.8 21.9 0.68
Length at 36 Weeks CGA (cm) # 42 44.2 3.3 19 43.5 4.6 0.44
Length Percentile at 36 Weeks CGA # 42 17.3 22.3 19 21.3 28.1 0.93
#
Growth at these time points represents nutrition delivery throughout hospitalization not just breast milk with
PHMF and ALHMF.

Table 3. Enteral feeding, growth and laboratory data.

Variable PHMF ALHMF p-Value
N Median N Median
Average Daily Provision of Protein per kg Weight 42 3.9 18 4.3 0.0014
CO2 Minimum after DOL 14 33 23 16 18.5 0.002
CO2 Minimum after DOL 30 23 25 8 20 0.002
Growth Velocity (g/kg/day) while on HMF 46 15.37 21 10.59 <0.0001
Growth (g/day, while on HMF) 46 31.27 21 23.66 0.0001
DOL Enteral Feedings Started 46 3.0 22 1.1 0.12
 137

Calcium Maximum 34 10.3 16 10.45 0.17

BUN Maximum after DOL 14 33 18 16 20 0.28
BUN Maximum after DOL 30 23 18 8 16 0.91
Creatinine Maximum 46 0.92 22 0.9 0.52

Table 4. Clinical outcomes.

PHMF LHMF
Variable p-Value
n (%) n (%)
NEC 0 (0%) 3 (13%) 0.03
ROP 16 (35%) 3 (13%) 0.09
ROP Procedure 3 (7%) 2 (9%) 1.00
IVH (any) 18 (39%) 4 (17%) 0.10
Dexamethasone Treatment 9 (20%) 1 (5%) 0.15
Bicitra Treatment 0 (0%) 1 (5%) 0.31
Death 0 (0%) 1 (4%) 0.33
Diaper Dermatitis 5 (11%) 4 (18%) 0.46
BPD 9 (20%) 3 (14%) 0.74
3.1. Safety and Clinical Outcomes

Mean lowest CO2 lab values (collected while infants were enterally feeding and not acutely ill)
were significantly lower in the ALHMF group compared to the PHMF group after both DOL 14
(18.5 vs. 23 mmol/L, p = 0.002) and DOL 30 (20 vs. 25 mmol/L, p = 0.002). Lowest CO2 lab values
after DOL 14 are displayed comparatively in Figure 1. Lowest values after DOL 30 are displayed
similarly in Figure 2. Maximum BUN and creatinine levels were similar between the two fortifier
groups and were not statistically significant. All other analyzed lab values were not statistically
different. All laboratory data in this retrospective study was obtained for clinical purposes regardless
of the fortifier group.
Incidence of NEC (a variable we were not powered to evaluate) was significantly higher in the
ALHMF group compared to the PHMF group (13% vs. 0%, p = 0.03).

Figure 1. CO2 levels between groups after Day of Life 14. The lowest CO2 levels after
DOL 14 were collected from metabolic panels. The mean level in the powder group
was 23, the mean level in the liquid group was 18.5. Laboratory clinical reference range
22–32 mmol/L. The difference is statistically significant (p = 0.002).
 138

Figure 2. CO2 levels between groups after Day of Life 30. The lowest CO2 levels after
DOL 30 were collected from metabolic panels. The mean level in the powder group
was 25, the mean level in the liquid group was 20. Laboratory Clinical reference range
22–32 mmol/L. The difference is statistically significant (p = 0.002).

3.2. Enteral Nutrition and Growth

Growth was significantly different between the two groups as measured in g/kg/day and is
described in Figure 3. Infant growth as measured in g/day from birth to 36 weeks EGA was 23.7 in
the PHMF group and 18.8 in the LHMF group (p = 0.057). There were no statistically significant
differences in the length of time to full feedings or the number of times that feedings were held that
could account for the difference in growth rates between the two groups.

Figure 3. The growth pattern of infants receiving powder differs from the growth pattern
of infants receiving liquid on fortified feed days. The plot shows the growth pattern for
each infant and the fitted line by group. Based on the plot, infants on powder grow at a
faster rate than infants receiving liquid. Evaluation of growth in gm/kg/day for the days
infants were fed fortified breast milk, based on the mixed effects model, shows a
significant interaction between day and group (p = 0.0022). Truncating the analysis at 45
days did not attenuate the results.
Growth on Fortified Feed Days
4000 4000
3800 3800
3600 3600
3400 3400
3200 3200
Predicted and Observed Weight

Predicted and Observed Weight

3000 3000
2800 2800
2600 2600
2400 2400
2200 2200
2000 2000
1800 1800
1600 1600
1400 1400
1200 1200
1000 1000
800 800
600 600

0 10 20 30 40 50 60 70 80 90

day

Group Powder Liquid

 139

Daily average protein/kg/day provision was higher in the ALHMF group compared to the PHMF
group (4.3 vs. 3.9 g, p = 0.0014). Mean enteral calorie provisions in the ALHMF group were higher
than in the PHMF group, 117.3 kcal/kg/day in the PHMF group as compared to 128.7 kcal/kg/day for
infants in the ALHMF group (p = 0.057). A higher proportion of infants in the ALHMF group required
increased caloric density of feedings >24 cal/ounce as compared to infants in the PHMF group,
(48% vs. 26%, p = 0.10). While this did not reach a statistical difference, clinically this was notable.

4. Discussion

To our knowledge, we are the first study to date to report our clinical findings of increased
complications with the use of ALHMF in a Level IIIc clinical setting. In our retrospective analysis of
acidosis, growth, and clinical outcomes in NICU infants fed with human milk fortified with LHMF
and PHMF we found significant acidosis and poor growth in the infants receiving LHMF. These
findings were very consistent with our clinical impressions during our clinical use of the LHMF.
We were also surprised to see increased NEC in the ALHMF group. Although we were not powered
as a primary outcome to evaluate NEC, we strongly encourage cautious further evaluation of the
product in the clinical setting with regards to this serious outcome.
A key difference in the ALHMF as compared to the PHMF is the acidification process required for
sterilization. This difference is likely to explain the increased complications seen in the ALHMF
group. The preterm infant’s inability to buffer this acid load likely led to an increase in clinical
complications including acidosis, poor growth, and, possibly, NEC.

4.1. Acidosis

There was a higher incidence of clinically significant metabolic acidosis in the ALHMF group, with
one infant requiring treatment with Bicitra. No infants in the PHMF group required Bicitra treatment,
even with twice as many patients in this group. Premature infants are susceptible to metabolic
acidosis [21] and renal tubular acidosis. However, these imbalances of acid base status should begin to
normalize after the first weeks of extrauterine life [21]. Considering similar baseline characteristics,
we hypothesize additional enteral acid load was a potential contributor to this increased incidence of
metabolic acidosis in the LHMF group.
Premature infants are at risk for developing metabolic acidosis secondary to immature
metabolic processes, a lower renal capacity to adequately excrete acid, and higher urinary losses of
bicarbonate [2,4,22]. Quantity of protein may affect metabolic processes; however the median daily
average protein provisions for each fortifier group were within the currently recommended ranges [1].
No clearly defined amount for maximum protein provision exists, however, it is suggested that intakes
greater than 6 g/kg/day are poorly tolerated [2]. Maximum daily average protein provisions for both
groups were below this level. Another reference states that protein provisions greater than 5 g/kg/day
may cause azotemia [1], but each group had intakes below this value, and maximum BUN and
creatinine levels were not different in the two fortifier groups. Having increased protein intake in
the PHMF group as well as the ALHMF group helps to illuminate that increased protein content in
the ALHMF was not likely the cause of the adverse outcomes.
 140

We question if the acidification sterilization process of the ALHMF may contribute to this acidosis
in some fragile premature infants. Our patient population included in this study was not limited by
respiratory acuity, as was the population in the Moya et al. paper [23]. We hypothesize that our more
inclusive population of both healthy infants and more fragile infants who may have less respiratory
stability decreases their capability to buffer the acid load provided in the ALHMF resulting in clinical
acidosis in some cases requiring medical therapy. It may be unwise in a fragile preterm infant
population to minimize the clinical significance of the metabolic acidosis noted in the ALHMF groups
in our study, the study conducted by Moya et al. [24], who reported that infants fed the ALHMF had
significantly lower pH (at day six), bicarbonate (at day six and 14), and CO2 (at day 14 and 28), and
significantly higher chloride (at day 14 and 28). Additionally, in an abstract evaluating 100 infants, 50
fed with ALHMF and 50 fed with PHMF, published by Cibulskis et al., from Saint Louis University at
the 2013 AAP_NCE, similar metabolic acidosis is described in this patient population (54% ALHMF
vs. 10% PHMF, p = 0.0001) [25]. As reported in their abstract, this grouptreated the acidosis as if it
were clinically significant, discontinuing ALHMF on 21/50 patients due to a clinical diagnosis of
acidosis [25].
4.2. Enteral Nutrition and Growth

Infants in the PHMF group received a mean daily calorie intake of 117.3 kcal/kg/day as compared
to infants in the ALHMF group who received a mean calorie intake of 128.7 kcal/kg/day. Infants in the
ALHMF group also received a median of 0.4 g protein/kg/day more than the infants in the PHMF
group. Despite higher protein and calorie provisions in the LHMF group, growth during the HMF
period was slower between the two groups as evaluated by several methods: in a mixed effects model
evaluated in g/kg/day (p = 0.002), in g/day (p = 0.0001), and by growth velocity in g/kg/day
(p < 0.0001). Noted also, is that ALHMF infants experienced an additional decrease of 10 growth
curve percentiles for weight from birth to 36 weeks EGA when compared to infants in the PHMF
group (growth at 36 weeks is representative of nutrition delivery that is not limited to the period
evaluated on PHMF and ALHMF). As Dexamethasone use inhibits growth in premature infants [4],
we further note that fewer infants in the ALHMF group (5%) compared to the PHMF group (20%)
required this drug for clinical treatment (p = 0.15).
Maintaining appropriate growth in this patient population was a high priority, so infants with
suboptimal growth were fed increased caloric density feedings above 24 cal/oz. Though not
statistically significant, a higher proportion of infants in the ALHMF group (48%) required caloric
densities greater than 24 cal/ounce when compared to the PHMF group (26%). Had those 48% of
infants in the ALHMF group not been prescribed increased caloric densities due to clinical
observations of poor growth differences in growth throughout the hospitalization would likely have
been larger between the PHMF groups and ALHMF groups. The statistical significance in infant
growth as noted in g/kg/day is seen in spite of the high priority our unit takes in maintaining optimal
growth and the subsequent aggressive adjustment of caloric density to achieve desired results. This
was ultimately the reason 26% of infants receiving ALHMF were transitioned to receive the PHMF
once the ALHMF use was discontinued in the NICU.
Not only are these growth effects consistent with the findings of Moya et al., they raise further
questions [24]. Moya et al., reported no significant differences in rate of weight gain or head
 141

circumference growth between infants fed this same ALHMF and infants fed a PHMF, even though
the human milk fortified with ALHMF contained 23% more protein (3.2 vs. 2.6 g protein/100 mL
fortified preterm human milk) [24]. In our study, even though we compensated for the difference
in protein content so that protein intake was similar, there was still poorer growth in infants
fed ALHMF.
At least part of the inability of the additional protein to improve growth may be due to the acidosis
noted above. It is well-known that infants with metabolic acidosis hyperventilate, as the expiration of
CO2 drives the elimination of H+ ions through the bicarbonate buffering system. What is less well
recognized is that protein catabolism can also be utilized to decrease acidity by the elimination of H+
ions through the urinary excretion of NH4+. Acidosis reduces protein synthesis in rats [26] and leads to
protein catabolism in humans [27].

4.3. NEC

No infants in PHMF group developed NEC compared to 13% in the ALHMF group. Reasons for
these occurrences remain unclear, as similar prevention strategies were followed for each group.
Previous implementation of aggressive nutrition practices in our unit demonstrated improved feeding
tolerance and clinical outcomes, with no increased incidence of NEC [12]. These nutrition practices
remained unchanged during the study period, and no additional clinical practices were implemented
concurrent with the change in human milk fortification. Slow rate of enteral feeding advancement
remained consistent between both groups, as evidenced by no statistically significant differences in
length to full enteral feedings. No changes in brand or caloric density of premature infant formula were
made, and infant formula was utilized equally in both groups if no MBM was available. As no
additional practice changes were implemented during this study period, we can neither confirm nor
exclude use of the ALHMF as a contributor to these occurrences of NEC. Although this study was not
powered to detect NEC based on historical incidence in our unit with rates over the last five years
ranging from 2% to 5% from our Vermont Oxford Network data, one should consider that significant
differences with small sample sizes may either reflect coincidental effects due to sample size, or may be
due to a real difference that is unexpectedly large.

4.4. Metabolic Acidosis

Literature suggests that premature infant formulas contain a high renal acid load, though human
milk contains less [2,21]. Research has additionally demonstrated that the composition of infant
formulas may affect the urinary pH and nutrient excretion of premature infants [21,22]. It is further
proposed that high renal acid loads contribute to maximum renal acid stimulation (urine pH < 5.4) [28] in
premature infants with immature renal function. Previous research studies have demonstrated that
infants with metabolic acidosis or maximum renal acid stimulation exhibit decreased growth [28,29].
This may also result in an increase in urinary sodium excretion [24,29] and a decrease in nitrogen
assimilation [30]. Blood sampling for acid-base indicators may not be significantly abnormal in
the presence of maximum renal acid excretion [22,28]. However, CO2 values may trend low [28],
which was clearly observed among infants in the ALHMF group (p = 0.002).
 142

4.5. Summary

Our results showing increased acidosis in the ALHMF group raise further concerns with use of
the ALHMF, as infants with metabolic acidosis may experience altered nutrient metabolism [28,31]
and decreased bone mineralization [32], leading to poor growth and osteopenia of prematurity.
Poor growth in the ALHMF group may also be attributed to changes to the nutrient content of the
milk caused by acidification as described by Erickson et al. [23]. This group reported significant
changes in acidified breast milk, including decreased total protein content, lipase activity, and free
fatty acids [23]. The nutritional changes in the composition of acidified breast milk documented by
Erickson in vitro may have led to the in vivo growth deficiencies noted in our ALHMF population [23].

5. Strengths and Limitations

5.1. Strengths

This study is the first to quantify results of use of ALHMF in a Level IIIc NICU setting. We are
uniquely situated to evaluate outcomes of our use of ALHMF in our patient population for several
important reasons. First, we initiated utilization of this product on all infants at one time. There was no
possibility of crossover product use to decrease the validity of the data. Additionally, we used this
product on all infants who would be eligible to receive fortified human milk, as would be expected in
a clinical NICU practice. This makes our data very relevant and applicable to clinical NICU settings.
Second, our clinical management of nutrition in this patient population has been published and
remains very successful with excellent growth and low baseline rates of NEC. Not only do we manage
nutrition care of this population very closely, but we also have a defined protocol in place so that
infants (except for fortification method) receive the same nutrition interventions over time regardless
of which group, PHMF or ALHMF, they received.
Additionally, our nutrition management with additional protein added to the PHMF group makes
the comparison of the two groups more relevant by giving them a more similar nutrient intake at
baseline than a comparison of ALHMF and PHMF alone which compares a large difference in
delivered protein.
Finally, we have a very detailed nutrition documentation medical record system, Intuacare. This
system allows for easy retrieval of detailed nutrition information including daily percentages of breast
milk, daily caloric intake, and daily protein intake in g/kg/day. This allows for minimal reporting error
in a retrospective study, such as this, and provides an excellent historical representation of each
infant’s delivered nutrition.

5.2. Limitations

This retrospective review of a clinical trial of a commercially available acidified liquid human milk
fortifier has several limitations including the retrospective nature of the study, and a modest sample
size, which limits the power of some data points. These limitations were partially reduced by our
reliance on electronic documentation for data collection and analysis. All medical documentation
remains variable between individuals and we cannot quantify unrecorded data, but the system utilized
 143

allowed for complete assessment of all recorded data on each research subject. As with any study
evaluating growth, head circumference and length measurements are also variable as length boards
were not used and measuring tape placement may vary between nursing staff. Some subjects were
discharged prior to 36 weeks EGA, therefore, anthropometric measurements at 36 weeks EGA were
not available. Likewise, lab values were also unavailable for these infants and could not be included in
data analysis.
Alterations in human milk composition are continuous, so calculated nutrient compositions of
fortified human milk may only serve as general estimations for our nutrient comparisons. Standard
NICU nutrition practices are followed as consistently as possible, however feeding advancement may
remain variable according to infant clinical status. Furthermore, the proportion of feedings as human
milk or formula remained variable among each infant. In an ideal study, all enrolled infants would
receive human milk only.
Though the incidence of NEC was statistically significant, it was not powered as a primary outcome
for this study. We also suspect that diaper dermatitis was under-recorded during this study period, as
our clinical experience suggests that diaper dermatitis is infrequently documented in the electronic
medical record even when infants experience more serious medical complications. However, perceived
worsening skin breakdown in our unit while using the ALHMF prompted development of a unit list of
infants with diaper dermatitis. Unfortunately, not all of the infants recorded on the unit list had diaper
dermatitis electronically documented as a medical problem. As there was no way to quantify these
cases, these select infants were not coded positively for diaper dermatitis in this study. Therefore, our
data analysis remains limited.

6. Conclusions

Use of the new ALHMF resulted in an increase in clinical complications and a decrease in growth
as measured in both g/day and g/kg/day. To our knowledge, this is one of the first studies assessing use
of the new ALHMF within a high acuity NICU without excluding infants with significant respiratory
disease or low five-min APGAR scores. Further research with the ALHMF is needed to compare
infant tolerance and outcomes among infants with a variety of gestational ages, weights, and increased
clinical acuity.

Acknowledgments

The authors would like to thank Sarah Kennedy, Ashley Schlaepfer, and Allison Fischer for their
contribution to data collection.

Conflicts of Interest

The authors declare no conflict of interest.

References

1. American Academy of Pediatrics. Pediatric Nutrition Handbook, 6th ed.; American Academy of
Pediatrics: Washington, DC, USA, 2009.
 144

2. Groh-Wargo, S.; Thompson, M.; Cox, J.H. Nutritional Care for High-Risk Newborns, 3rd ed.;
Hartline, J.V., Ed.; Precept Press: Chicago, IL, USA, 2000.
3. Sullivan, S.; Schanler, R.J.; Kim, J.H.; Patel, A.L.; Trawöger, R.; Kiechl-Kohlendorfer, U.;
Chan, G.M.; Blanco, C.L.; Abrams, S.; Cotton, C.M.; et al. An exclusively human milk-based diet
is associated with a lower rate of necrotizing enterocolitis than a diet of human milk and bovine
milk-based products. J. Pediatr. 2010, 156, 562–567.
4. Tsang, R.C. Nutrition of the Premature Infant: Scientific Basis and Practical Guidelines, 2nd ed.;
Uauy, R., Koletzko, B., Zlotkin, S.H., Eds.; Digital Education Publishing: Cincinnati, OH,
USA, 2005.
5. Clark, R.H.; Thomas, P.; Peabody, J. Extrauterine growth restriction remains a serious problem in
prematurely born neonates. Pediatrics 2003, 111, 986–990.
6. Section on Breastfeeding. Breastfeeding and the use of human milk. Pediatrics 2012, 129,
e827–e841.
7. Hawthorne, K.M.; Abrams, S.A. Safety and efficacy of human milk fortification for
very-low-birthweight infants. Nutr. Rev. 2004, 62, 482–485.
8. Valentine, C.J.; Fernandez, S.; Rogers, L.K.; Gulati, P.; Hayes, J.; Lore, P.; Puthoff, T.; Dumm, M.;
Jones, A.; Collins, K.; et al. Early amino-acid administration improves premature infant weight.
J. Perinatol. 2009, 29, 428–432.
9. Stephens, B.E.; Walden, R.V.; Gargus, R.A.; Tucker, R.; Mckinley, L.; Mance, M.; Nye, J.;
Vohr, B.R. First-week protein and energy intakes are associated with 18-month developmental
outcomes in extremely low birth weight infants. Pediatrics 2009, 123, 1337–1343.
10. Maggio, L.; Costa, S.; Gallini, F. Human milk fortifiers in very low birth weight infants. Early
Hum. Dev. 2009, 85, S59–S61.
11. Kuschel, C.A.; Harding, J.E. Multicomponent fortified human milk for promoting growth in
premature infants. Cochrane Database Syst. Rev. 2004, 1, CD000343.
12. Hanson, C.; Sundermeier, J.; Dugick, L.; Lyden, E.; Anderson-Berry, A.L. Implementation,
process, and outcomes of nutrition best practices for infants < 1500 g. Nutr. Clin. Pract. 2011, 26,
614–624.
13. Ehrenkranz, R.A.; Dusick, A.M.; Vohr, B.R.; Wright, L.L.; Wrage, L.A.; Poole, W.K. Growth in
the neonatal intensive care unit influences neurodevelopmental and growth outcomes of extremely
low birth weight infants. Pediatrics 2006, 117, 1253–1261.
14. Himelright, I.; Harris, E.; Lorch, V.; Anderson, M.; Jones, T.; Craig, A.; Kuehnert, M.; Forster, T.;
Arduino, M.; Jensen, B.; et al. Enterobacter sakazakii infections associated with the use of
powdered infant formula—Tennessee, 2001. MMWR 2002, 51, 297–300.
15. Berseth, C.L.; Mitmesser, S.H.; Stolz, S.; Harris, F.M.; Sisk, P. Growth and safety of premature
infants fed a new liquid human milk fortifier. FASEB J. 2011, 25, 589.3.
16. Enfamil. Enfamil Human Milk Fortifier Acidified Liquid Website. Available online:
http://www.enfamil.com/app/iwp/enf10/content.do?dm=enf&id=/Consumer_Home3/FeedingSolu
tions/EnfamilHumanMilkFortifier2&iwpst=B2C&ls=0&csred=1&r=3508452874 (accessed on
6 March 2012).
17. Similac Human Milk Fortifier Website. Available online: http://abbottnutrition.com/products/
similac-human-milk-fortifier (accessed on 6 March 2012).
 145

18. Fenton, T.R. A new growth chart for premature babies: Babson and Benda’s chart updated with
recent data and a new format. BMC Pediatr. 2003, 3, 13.
19. Fenton, T.R.; Sauve, R.S. Using the LMS method to calculate z-scores for the Fenton premature
infant growth chart. Eur. J. Clin. Nutr. 2007, 61, 1380–1385.
20. Patel, A.; Engstrom, J.; Meier, P.; Kimura, R. Accuracy of methods for calculating postnatal
growth velocity for extremely low birth weight infants. Pediatrics 2005, 116, 1466–1473.
21. Kalhoff, H.; Manz, F.; Kiwull, P.; Kiwull-Schone, H. Food composition and acid-base balance:
Alimentary acid load and clinical implications in neonates. Open Nutr. J. 2008, 2, 5–8.
22. Manz, F.; Kalhoff, H.; Remer, T. Renal acid excretion in early infancy. Pediatr. Nephrol. 1997,
11, 231–243.
23. Erickson, T.; Gill, G.; Chan, G.M. The effects of acidification on human milk’s cellular and
nutritional content. J. Perinatol. 2013, 33, 371–373.
24. Moya, F.; Sisk, P.M.; Walsh, K.R.; Berseth, C.L. A new liquid human milk fortifier and linear
growth in preterm infants. Pediatrics 2012, 130, 928–935.
25. Cibulskis, C.C.; Bowles, M. Metabolic Acidosis Associated with Liquid Acidified Human Milk
Fortifier. In Proceedings of the AAP Section on Perinatal Pediatrics Scientific and Educational
Program Abstract and Poster Presentations, Orlando, FL, USA, 25–29 October 2013.
26. Caso, G.; Garlick, B.A.; Casella, G.A.; Sasvary, D.; Garlick, P.J. Acute metabolic acidosis
inhibits muscle protein synthesis in rats. Am. J. Physiol. Endocrinol. Metab. 2004, 287, E90–E96.
27. Reaich, D.; Channon, S.M.; Scrimgeour, C.M.; Goodship, T.H.J. Ammonium chloride-induced
acidosis increases protein breakdown and amino acid oxidation in humans Am. J. Physiol. 1992,
263, E735–E739.
28. Kalhoff, H.; Manz, F. Nutrition, acid-base status and growth in early childhood. Eur. J. Nutr.
2001, 40, 221–230.
29. Kalhoff, H.; Manz, F.; Diekmann, L.; Kunz, C.; Stock, G.J.; Weisser, F. Decreased growth rate of
low-birth-weight infants with prolonged maximum renal acid stimulation. Acta Paediatr. 1993,
82, 522–527.
30. Kalhoff, H.; Rascher, W.; Diekmann, L.; Stock, G.J.; Manz, F. Urinary excretion of aldosterone,
arginine vasopressin and cortisol in premature infants with maximum renal acid stimulation.
Acta Paediatr. 1995, 84, 490–494.
31. Kalhoff, H.; Diekmann, L.; Rudloff, S.; Manz, F. Renal excretion of calcium and phosphorus in
premature infants with incipient late metabolic acidosis. J. Pediatr. Gastroenterol. Nutr. 2001, 33,
565–569.
32. Rochow, N.; Jochum, F.; Redlich, A.; Korinekova, Z.; Linnemann, K.; Weitmann, K.; Boehm, G.;
Müler, H.; Kalhoff, H.; Topp, H.; et al. Fortification of breast milk in VLBW infants: Metabolic
acidosis is linked to the composition of fortifiers and alters weight gain and bone mineralization.
Clin. Nutr. 2011, 30, 99–105.

© 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).
146
 147

Reprinted from Nutrients. Cite as: Thorisdottir, B.; Gunnarsdottir, I.; Steingrimsdottir, L.; Palsson, G.I.;
Thorsdottir, I. Vitamin D Intake and Status in 12-Month-Old Infants at 63–66° N. Nutrients 2014, 6,
1182-1193.

Vitamin D Intake and Status in 12-Month-Old Infants

at 63–66° N
Birna Thorisdottir 1,*, Ingibjorg Gunnarsdottir 1, Laufey Steingrimsdottir 1, Gestur I. Palsson 2
and Inga Thorsdottir 1
1
Unit for Nutrition Research, Landspitali University Hospital & Faculty of Food Science and
Nutrition, School of Health Sciences, University of Iceland, Eiriksgata 29, IS-101 Reykjavik,
Iceland; E-Mails: ingigun@landspitali.is (I.G.); laufey@hi.is (L.S.); ingathor@hi.is (I.T.)
2
Children’s Hospital, Landspitali University Hospital, Hringbraut, IS-101 Reykjavik, Iceland;
E-Mail: gesturip@landspitali.is

* Author to whom correspondence should be addressed; E-Mail: bth50@hi.is;

Tel.: +354-5438-410; Fax: +354-5521-331.

Received: 7 February 2014; in revised form: 1 March 2014 / Accepted: 10 March 2014 /
Published: 21 March 2014

Abstract: The objective was to assess the vitamin D status in healthy 12-month-old infants
in relation to quantity and sources of dietary vitamin D, breastfeeding and seasons.
Subjects were 76 12-month-old infants. Serum levels of 25-hydroxyvitamin D (25(OH)D)
• 50 nmol/L were considered indicative of vitamin D sufficiency and 25(OH)D < 27.5 nmol/L
as being indicative of increased risk for rickets. Additionally, 25(OH)D > 125 nmol/L was
considered possibly adversely high. Total vitamin D at 9–12 months (eight data collection
days) included intake from diet and supplements. The mean ± SD of vitamin D intake was
8.8 ± 5.2 ȝg/day and serum 25(OH)D 98.1 ± 32.2 nmol/L (range 39.3–165.5). Ninety-two
percent of infants were vitamin D sufficient and none at increased risk for rickets.
The 26% infants using fortified products and supplements never/irregularly or in small
amounts had lower 25(OH)D (76.8 ± 27.1 nmol/L) than the 22% using fortified products
(100.0 ± 31.4 nmol/L), 18% using supplements (104.6 ± 37.0 nmol/L) and 33% using both
(110.3 ± 26.6 nmol/L). Five of six infants with 25(OH)D < 50 nmol/L had no intake of
supplements or fortified products from 0 to 12 months. Supplement use increased the odds
of 25(OH)D > 125 nmol/L. Breastfeeding and season did not affect vitamin D status.
The majority of infants were vitamin D sufficient. Our findings highlight the need for
vitamin D supplements or fortified products all year round, regardless of breastfeeding.
 148

Keywords: 25-hydroxyvitamin D; vitamin D; infant; dietary supplements; fortified foods

1. Introduction

Vitamin D is a key nutrient for children’s well-being and growth, is essential for bone health [1]
and may contribute to other health benefits [2]. Infant need for vitamin D can be met by synthesis in
the skin when exposed to appropriate ultraviolet B wavelengths and by sufficient vitamin D intake,
either from breast milk or other dietary sources [3]. At latitudes higher than ~50° N, little or no
cutaneous vitamin D synthesis is possible during winter months [4]. Vitamin D from breast milk alone
is unlikely to meet the needs of infants during complementary feeding [5]. Few common foods are
naturally rich in vitamin D [6]. Vitamin D intake from supplements or fortified foods or beverages is
therefore important in northern latitudes [3].
The recommended intake (RI) of vitamin D is 10 ȝg (400 IU) for infants and children from
six months of age according to Nordic nutrition recommendations [7]. This is in accordance with
the average intake specified by the Institute of Medicine (IOM) for infants from birth to 12 months of
age [8]. To ensure that the RI is met, parents are advised to give their infants a daily supplement of
10 ȝg D3 from the age of 1–2 weeks. Fortification schemes differ between countries [3]. In Iceland,
population-based studies on infants [9,10] and pre-school children [11] have shown that less than
two-thirds of young children use vitamin D supplements regularly. Several cases of rickets in the past
years have given cause for concern on the vitamin D status of Icelandic infants [12]. In 2003, a follow-up
formula intended for infants from 6 to 24 months, fortified with 1.2 ȝg D3 per 100 mL, was
introduced [13], and infant porridges and breakfast cereals fortified with vitamin D are available [14].
Population-based studies on both vitamin D intake and status during the complementary feeding period
in the Nordic countries are lacking [3]. The vitamin D status of Icelandic infants is unknown and has
never been studied.
The objectives of this study were to assess vitamin D status measured as serum levels of
25-hydroxyvitamin D (25(OH)D) in healthy 12-month-old infants and to consider it in relation to
quantity and sources of dietary vitamin D, breastfeeding and seasons.

2. Experimental Section

2.1. Subjects

Study subjects were 76 infants with data on dietary intake in infancy and quantitative analysis of
serum 25(OH)D levels at 12 months. They were a subsample of participants in a longitudinal cohort
study on diet, growth and health outcomes of infants born in the year 2005. In the original study, 250
healthy Icelandic infants born at term were randomly selected from the whole country (63–66° N).
Blood samples were obtained with the primary aim of analyzing the iron status and blood lipids of the
infants [13]. Analysis of serum 25(OH)D was only possible for those subjects with sufficient amounts
of blood available, resulting in the subsample of 76. Anthropometrical variables and dietary intake in
infancy, e.g., duration of breastfeeding, intake of vitamin D, formula and cod liver oil, sociodemographic
 149

factors (parents’ age and education) and parental BMI of the children included in the current analysis
did not differ from those of the children in the original study. More detailed information on the original
cohort is published elsewhere [13,15]. Informed written consent from the parents was obtained, and
all individual information was processed with strict confidentiality. The study was approved by
the Icelandic Bioethics Committee, the Icelandic Data Protection Authority and the Local Ethical
Committee at Landspitali University Hospital.

2.2. Dietary Assessment

A flowchart of the process of the study, relevant to the present analysis, is presented in Figure 1.

Figure 1. Flowchart on the progress of the study.

Dietary data from 0 to 4 months of age were collected by dietary history, including questions on
breastfeeding, infant formula-feeding, other food items and supplements. At 5–8 and 10–11 months,
24-h recalls using common household measures, such as cups and spoons, were made. At 9 and
12 months, weighed food records were kept for 3 consecutive days (72-h). All food and fluids were
weighed on accurate scales (Philips type HR 2385, Szekesfehervar, Hungary) with 1-g precision. The
amount of breast milk consumed was estimated by weighing the breastfed infants in the same clothes
before and after each breastfeeding session on baby scales (Tanita model 1583, Tokyo, Japan, or Sega
model 336, Hamburg, Germany) with 10-g precision. An average consumption of food and nutrients
was calculated using the Icelandic food composition database [14]. The total intake of vitamin D
included intake from the diet, breast milk and supplements. For the analysis presented here, the main
emphasis was on dietary intake at 9–12 months, because we believe that it may influence serum
25(OH)D concentration at 12 months [3,16]. We divided infants into four groups based on regular
intake of significant amounts of the main vitamin D sources at 9–12 months. The “fortified” group
included infants getting on average • ȝg of vitamin D per day from fortified products; the
“supplement” group included those getting on average • ȝg of vitamin D per day from
supplements; the “combined” group included those fulfilling both conditions; and the “no or irregular”
group included infants fulfilling neither conditions. Fortified products included infant formula, infant
porridges and breakfast cereals, and the cut-off at 2.4 ȝg of vitamin D was applied, because it
corresponds to consumption of•P/ of fortified formula, the most commonly consumed product
in this category. Supplements included cod liver oil and liquid vitamin A and D supplements (vitamin
 150

AD drops), and the cut-off at 5 ȝg of vitamin D was applied, because it corresponds to the
recommended dose on at least half of the data collection days. We also considered whether or not
infants were still partially breastfed at 12 months of age.

2.3. Blood Sampling and Biochemical Analyses

At 12 months of age, blood samples were collected in the morning in the fasting state. The samples
were centrifuged within 6 h of data collection. Separated serum samples were then storedí at ƒ&
until being analyzed. The quantitative analyses of serum 25(OH)D levels were conducted by the Roche
Diagnostics Vitamin D total assay (Roche Diagnostics, Mannheim, Germany), with a measuring range
of 7.5–175 nmol/L and a precision of 0.1 nmol/L. In accordance with a recent Nordic systematic
literature review (SLR), serum 25(OH)D •  QPRO/  QJP/  ZDV FRQVLGHUHG LQGLFDWLY e of a
sufficient vitamin D status, and serum 25(OH)D < 27.5 nmol/L (11 ng/mL) indicates increased risk for
rickets [3]. Additionally, serum 25(OH)D > 125 nmol/L (50 ng/mL) was considered as possibly
adversely high, as suggested by the IOM [8]. Infants were classified according to season when blood
samples were collected; winter/spring (January 2006–April 2006 and November 2006–December
2006) and summer/autumn (May 2006–October 2006).

2.4. Statistical Analyses

Statistical analyses were performed with SAS (Enterprise Guide 4.3; SAS Institute Inc., Cary, NC,
USA). Linear regression analysis was used to examine the relation between vitamin D intake and
serum 25(OH)D. Descriptive statistics were used to describe vitamin D intake and serum 25(OH)D
concentrations, presented as the means ± SD. For comparison between groups, an independent,
two-sample t-test with equal variances and a one-way ANOVA with equal variance were used.
Logistic regression was used to examine the risk of having serum 25(OH)D above 125 nmol/L among
infants using supplements or not. The results were presented as odds ratios (OR), with its 95% CI.
Spearman’s correlation analysis was used to assess correlations between 25(OH)D and breastfeeding,
SUHVHQWHG DV WKH FRUUHODWLRQ FRHIILFLHQW ȡ  DQG WKH p-value for correlation. A two-sided test with
a p-value < 0.05 was considered statistically significant.

3. Results

At the age of 12 months, the mean ± SD serum 25(OH)D concentration was 98.1 ± 32.2 nmol/L
(39.3 ± 12.9 ng/mL) and ranged from 39.3 to 165.5 nmol/L (15.7 to 66.3 ng/mL). Seventy infants
(92%) were considered vitamin D sufficient and none at increased risk for rickets. Eighteen infants
(24%) were considered to have a possibly adversely high 25(OH)D concentration.
Vitamin D intake at 9–12 months predicted 25(OH)D levels at 12 months (Figure 2). The
mean ± SD intake of vitamin D was 8.8 ± 5.2 ȝg, and 57% of the infants were below the RI of 10 ȝg.
Those infants had significantly lower mean ± SD 25(OH)D than infants above the RI (87.1 ± 31.1 vs.
111.8 ± 29.0 nmol/L, p = 0.001).
Supplements provided 56% of total vitamin D at 9–12 months. Another 38% came from fortified
products; thereof, 24% from formulas, 13% from infant porridges and 1% from breakfast cereals.
 151

Among natural sources of vitamin D were meat and fish (3%) and cow’s milk (1%). Breast milk
provided <1% of vitamin D. As presented in Table 1, infants in the “combined” group had a higher
vitamin D intake than infants in the “supplement” group (p < 0.001), who, in turn, had a higher vitamin
D intake than infants in the “fortified” group (p = 0.013). Mean serum 25(OH)D in these three groups
was, however, not significantly different (p > 0.05). Infants not using fortified products or supplements
regularly in significant amounts at 9–12 months (“no or irregular” group) had significantly lower
vitamin D intake than all the other groups (p < 0.001) and lower serum 25(OH)D (p < 0.001).

Figure 2. The linear regression line for serum 25-hydroxyvitamin D (25(OH)D) at

12 months in relation to vitamin D intake from diet and supplements at 9–12 months. The
dashed horizontal line at 50 nmol/L is the cut-off line applied for a sufficient vitamin D
status, and the dashed vertical line at 10 ȝg indicates the Nordic recommended intake (RI).

Table 1. Variables potentially associated with vitamin D intake at 9–12 months and serum
25(OH)D at 12 months.
Vitamin D Intake 25(OH)D
Variables n (%)
ȝJGD\ (nmol/L)
All 76 (100) 8.8 ± 5.2 98.1 ± 32.2
Boys 39 (51) 8.6 ± 5.7 96.6 ± 34.3
Girls 37 (49) 8.9 ± 4.6 99.7 ± 30.3
Vitamin D sources at 9–12 months a
“No or irregular” 20 (26) 2.5 ± 1.9 76.8 ± 27.1
“Fortified” 17 (22) 6.5 ± 2.2 100.0 ± 31.4
“Supplement” 14 (18) 8.8 ± 2.7 104.6 ± 37.0
“Combined” 25 (33) 14.3 ± 3.0 110.3 ± 26.6
Partially breastfed at 12 months
No 62 (82) 8.7 ± 5.0 97.7 ± 32.7
Yes 14 (18) 9.1 ± 6.0 101.9 ± 31.5
Season of blood sample collection
Winter/Spring 33 (43) 8.1 ± 4.9 94.4 ± 31.6
Summer/Autumn 43 (57) 9.2 ± 5.4 101.0 ± 32.8
a
Abbreviation: 25(OH)D, 25-hydroxyvitamin D. Mean ± SD. Infants were divided into groups based on the regular
intake of significant amounts of the main vitamin D sources at 9–12 months. “No or irregular”: neither fortified products
nor supplements; “Fortified”: fortified products; “Supplement”: supplements; “Combined”: both fortified products
and supplements.
 152

Five out of six infants with serum 25(OH)D below 50 nmol/L belonged to the “no or irregular”
group, i.e., they did not use fortified products or supplements on a regular basis or in significant
amounts at 9–12 months. Their intake of vitamin D was below 5 ȝg at 9–12 months. Further, they did
not use fortified products or supplements at all from birth to nine months of age. The sixth infant with
25(OH)D below 50 nmol/L was categorized in the “supplement” group, but only got half of the
recommended amount of supplements daily. Of the 18 infants with 25(OH)D levels above 125 nmol/L,
one belonged to the “no or irregular” group (5% of infants in that group), three to the “fortified” group
(18%), six to the “supplement” group (43%) and eight to the “combined” group (32%). Infants using
supplements (i.e., classified in the “supplement” or “combined” groups) were 4.6 times more likely
(95% CI = 1.4, 15.8) to have 25(OH)D above 125 nmol/L than infants not using supplements
(i.e., classified in the “no or irregular” or “fortified” groups). Infants in the “combined” group were not
more likely to have 25(OH)D above 125 mol/L than infants in the “supplement” group (OR (95% CI)
= 0.6 (0.2, 2.4)).
The duration of exclusive breastfeeding ranged from 0 to 6 months, with a median (25th,
75th percentiles) of four (1, 5) months. The total duration of breastfeeding ranged from 0 to 12 months,
with a median (25th, 75th percentiles) of eight (6, 10) months. There was no correlation between the
GXUDWLRQ RI H[FOXVLYH EUHDVWIHHGLQJ DQG  2+ ' ȡí
 p = 0.895) or the total duration of
EUHDVWIHHGLQJDQG 2+ ' ȡ p = 0.502). Among children partially breastfed at 12 months of
age, breast milk intake in the age period of 9–12 months ranged from 10 mL to 750 mL per day.
No difference was observed in vitamin D intake or 25(OH)D according to breastfeeding at 12 months
(p = 0.923 and 0.674, respectively), season of blood sample collection (p = 0.385 and p = 0.379,
respectively) or sex (p = 0.859 and p = 0.678, respectively).

4. Discussion

This study provides the first information on vitamin D status in Icelandic infants. Based on
thresholds proposed in a recent Nordic SLR [3], 92% of the infants were considered vitamin D
sufficient and none at increased risk for rickets. Consensus has not been reached on the optimal
25(OH)D concentration in infants, and uniformity is lacking in the description of sufficient and
deficient ranges for 25(OH)D levels. Using cut-off values proposed by IOM [8] or the Pediatric
Endocrine Society [16] does not change our results of 92% of infants being classified as vitamin D
sufficient, and according to those cut-offs, no infants are classified as vitamin D deficient. According
to a European consensus statement, vitamin D deficiency occurs commonly among healthy European
infants not adhering to recommendations for vitamin D supplementation [17]. However, studies on
healthy infants from Denmark [18], Norway [19] and Finland [20] have previously reported a high
proportion of vitamin D sufficiency amongst nine-month-olds, 12-month-olds and 14-month-olds,
respectively. Those studies were not population-based, and in the Danish and Finnish studies, selection
bias resulted in an unusually high frequency of infants using vitamin D supplements (97% and 100%,
respectively). The Nordic countries have a well-established newborn and infant healthcare. According
to protocols for the newborn and infant healthcare in Iceland [21], mothers are asked about their
infants’ vitamin D supplement use and encouraged to follow the recommendations on vitamin D
supplements at every visit, which are scheduled at least nine times during the first year of the infant.
 153

This may explain the relatively low proportion of vitamin D deficiency among infants in Iceland and,
more broadly, the Nordic countries. To our knowledge, this is the first study on infant vitamin D status
in the Nordic countries in a sample that is representative of the general infant population. Therefore,
we believe it is an important contribution to the literature on the vitamin D status of healthy infants
during complementary feeding in northern latitudes.
The relatively high 25(OH)D levels may, at least partly, be explained by 75% of the infants
regularly using vitamin D supplements and/or fortified foods or drinks in significant amounts. The
commonly used follow-up formula, fortified with vitamin D, was introduced in 2003. Before that, it
was common that regular cow’s milk gradually replaced breast milk in the age range of 5–12 months [9].
The main vitamin D source for Icelandic infants has, therefore, historically, been vitamin D drops or
cod liver oil, and even though studies on infants and children have shown a little less than two-thirds
of children complying with supplement use, the remaining one-third has been seen as a reason for
concern. Studies on Icelandic infants and young children have never before assessed how frequently
vitamin D fortified products are consumed or how they contribute to vitamin D status. We do not have
any data on the vitamin D status of infants and young children previous to the introduction of the
fortified follow-up formula.
The wide range of serum 25(OH)D concentrations observed in the study should be considered when
interpreting the results. Transferring the 8% of infants in our sample with serum 25(OH)D below
50 nmol/L to the whole infant population in Iceland (around 4600 12-month-olds annually from 2005
to 2012) [22] suggests that about 275 infants every year would be vitamin D insufficient, with the
possibility of some being at risk for vitamin D deficiency. Our study, showing that infants with an
insufficient vitamin D status did not use fortified products or supplements at all from birth to nine
months of age, in addition to a very low vitamin D intake from nine to 12 months of age, could be
considered in newborn and infant healthcare in Iceland to identify, at an early age, children with
undesirable diet habits that may increase the risk of vitamin D insufficiency or deficiency. Infants
using supplements with or without concurrent use of fortified foods or drinks were more likely than
infants not using supplements to have 25(OH)D concentrations that may be considered as possibly
adversely high [8]. Correct dosing of supplements is important, as well as caution when combining the
use of supplements and fortified foods or drinks. However, no infant exceeded the 25 ȝg vitamin D
intake, which is considered the tolerable upper intake level by the European Food Safety Authority [6],
and other estimations of the high end for safe concentration levels of 25(OH)D are higher than the
125 nmol/L estimated by IOM [16,23].
Since the time of this study, parents have been encouraged to give their infants vitamin D drops
instead of vitamin AD drops. The vitamin D content in the two products is the same, and other infant
guidelines remain unchanged. Therefore, we believe that the findings of this study are transferable to
Icelandic infants born today. Iceland is among the few countries that includes cod liver oil intake or
other vitamin D supplements in the population-based dietary guidelines for children and adults of all
ages [24]. A recent study from Denmark showed that parents’ perceived relevance of nutritional
guidelines declined from the early to later phases of complementary feeding [25], and a Finnish study
showed decreased use of supplements as children grew older [26]. Icelandic studies have also shown
low vitamin D intakes among children [27], adolescents [28] and adults [29], and results from a
follow-up of the infants participating in the current analysis reveal that only 27% used supplements
 154

at six-years of age [30]. While the vitamin D status in our study is considered sufficient for the
majority of infants, studies on Icelandic children and adults have shown lower 25(OH)D
concentrations than presented here [31–33]. This study, showing the importance of supplements and/or
fortified products on vitamin D status, is therefore of importance for public health policy.
We did not find differences in 25(OH)D levels between months when cutaneous synthesis is
expected to be very low or totally absent at northern latitudes (November to April) and months when
the quantity and quality of UV radiation might be sufficient for cutaneous synthesis (May to October).
Icelandic parents are advised to keep their infants out of direct sunlight, and summer temperatures in
Iceland usually require long sleeves and a hat for infants. In case infants get in contact with sun, the
use of sunscreen is advised [21]. We propose that cutaneous synthesis of vitamin D does not contribute
significantly to 25(OH)D in Icelandic infants and that the use of supplements and/or fortified foods
and drinks is therefore essential all year round in order to maintain a sufficient vitamin D status.
Seasons have, however, been shown to affect 25(OH)D in older children and adults [31,33]. No
difference was seen in 25(OH)D levels between infants breastfed or not breastfed at 12 months, which
may be explained by the emphasis put on supplement use regardless of feeding mode.
The main strengths of our study lie in the assessment of both vitamin D intake and status in
a population-based infant sample and the longitudinal design of the study. Although we are aware of
the possibility of altered dietary behavior on data collection days, the use of eight data collection days
from 9 to 12 months of age in the current analysis strengthens our confidence that we have reliably
estimated food and nutrient intake that may affect vitamin D status at 12 months. The dietary
information from 0 to 8 months of age gives practical information. The study also has some
limitations. Blood samples were obtained with the primary aim of analyzing the iron status and blood
lipids of the infants [13]. Analysis of serum 25(OH)D was only possible for those subjects with
sufficient amounts of blood available, resulting in a small sample size. Analyses on parameters that
have been used to complement 25(OH)D levels and/or used as blood safety measurements in other
studies [34,35], such as parathyroid hormone, serum calcium, alkaline phosphatase and C-reactive
protein, were not performed. There is a possibility that the method used for quantitative analyses of
serum 25(OH)D may overestimate the 25(OH)D concentration in infants, due to the possible presence
of C-3 epimers [36–38]. As all subjects were of Icelandic origin and healthy, transferring the results to
high-risk groups of vitamin D deficiency, such as infants of non-western immigrants residing in
northern latitudes and infants with chronic illnesses, should not be advised [7,39,40].

5. Conclusions

In conclusion, the majority of infants were vitamin D sufficient. Our findings highlight the need for
vitamin D supplements or fortified products all year round, regardless of breastfeeding in infant
populations with little or no sun exposure.

Acknowledgments

The authors are most grateful to the staff at the healthcare centers, Children’s Hospital and
laboratories at Landspitali University Hospital, for their cooperation, the nutritionists, students and
other healthcare professionals participating in data collection and last, but not least, the participating
 155

children and the families. This study was supported by Sumargjöf—The Icelandic Children’s Welfare
Society, the Doctoral Grants of the University of Iceland Research Fund and the Icelandic Research
Fund of the Icelandic Centre for Research.

Author Contributions

B.T. participated in data collection, analyzed the data and drafted the paper. I.G. and L.S. conceived
of and designed the study. G.P. performed the blood sampling. I.T. supervised data collection and
conceived of and designed the study. All authors contributed in writing and editing the manuscript and
approved the final version of the paper as submitted.

Conflicts of Interest

The authors declare no conflict of interest.

References

1. Turner, A.G.; Anderson, P.H.; Morris, H.A. Vitamin D and bone health. Scand. J. Clin. Lab.
Investig. 2012, 243, 65–72.
2. Moreno, L.A.; Valtuena, J.; Perez-Lopez, F.; Gonzalez-Gross, M. Health effects related to low
vitamin D concentrations: Beyond bone metabolism. Ann. Nutr. Metab. 2011, 59, 22–27.
3. Lamberg-Allardt, C.; Brustad, M.; Meyer, H.E.; Steingrimsdottir, L. Vitamin D—A systematic
literature review for the 5th edition of the Nordic Nutrition Recommendations. Food Nutr. Res.
2013, 57, doi:10.3402/fnr.v57i0.22671.
4. Engelsen, O.; Brustad, M.; Aksnes, L.; Lund, E. Daily duration of vitamin D synthesis in human
skin with relation to latitude, total ozone, altitude, ground cover, aerosols and cloud thickness.
Photochem. Photobiol. 2005, 81, 1287–1290.
5. Thiele, D.K.; Senti, J.L.; Anderson, C.M. Maternal vitamin D supplementation to meet the needs
of the breastfed infant: A systematic review. J. Hum. Lact. 2013, 29, 163–170.
6. EFSA Panel on Dietetic Products Nutrition and Allergies (NDA). Scientific opinion on the
tolerable upper intake level of vitamin D. EFSA J. 2012, 10, doi:10.2903/j.efsa.2012.2813.
7. Nordic Council of Ministers. Nordic Nutrition Recommendations 2012—Integrating Nutrition and
Physical Activity; Nordic Council of Ministers: Copenhagen, Denmark, 2014.
8. Institute of Medicine. Dietary Reference Intakes for Calcium and Vitamin D; The National
Academies Press: Washington, DC, USA, 2011.
9. Thorsdottir, I.; Atladottir, H.; Palsson, G. Infant Nutrition in Iceland 1995–2000; Unit for
Nutrition Research: Reykjavik, Iceland, 2000.
10. Thorsdottir, I.; Thorisdottir, A.V.; Palsson, G. Infant Nutrition in Iceland 2005–2007; Unit for
Nutrition Research: Reykjavik, Iceland, 2008.
11. Gunnarsdottir, I.; Eysteinsdottir, T.; Thorsdottir, I. Diet for Preschool Children age 3 and 5 Years
old in Iceland 2007; Unit for Nutrition Research and Public Health Institue of Iceland: Reykjavik,
Iceland, 2008.
 156

12. Kristinsdottir, H.; Jonasdottir, S.; Bjornsson, S.; Ludvigsson, P. Rickets in a child. Laeknabladid
2011, 97, 477–480.
13. Thorisdottir, A.V.; Thorsdottir, I.; Palsson, G.I. Nutrition and iron status of 1-year olds following
a revision in infant dietary recommendations. Anemia 2011, 2011, doi:10.1155/2011/986303.
14. Matis Ltd.—Icelandic Food and Biotech R & D. Food Composition Tables. ISGEM (The
Icelandic Food Composition Database); Matis Ltd.—Icelandic Food and Biotech R & D:
Reykjavik, Iceland, 2008. Available online: http://www.matis.is/ISGEM/en/ (accessed on 7
February 2014).
15. Thorisdottir, A.V.; Gunnarsdottir, I.; Thorsdottir, I. Revised infant dietary recommendations:
The impact of maternal education and other parental factors on adherence rates in Iceland.
Acta Paediatr. 2013, 102, 143–148.
16. Misra, M.; Pacaud, D.; Petryk, A.; Collett-Solberg, P.F.; Kappy, M. Vitamin D deficiency in
children and its management: Review of current knowledge and recommendations. Pediatrics
2008, 122, 398–417.
17. Braegger, C.; Campoy, C.; Colomb, V.; Decsi, T.; Domellof, M.; Fewtrell, M.; Hojsak, I.;
Mihatsch, W.; Molgaard, C.; Shamir, R.; et al. Vitamin D in the healthy European paediatric
population. J. Pediatr. Gastroenterol. Nutr. 2013, 56, 692–701.
18. Ostergard, M.; Arnberg, K.; Michaelsen, K.F.; Madsen, A.L.; Krarup, H.; Trolle, E.; Molgaard, C.
Vitamin D status in infants: Relation to nutrition and season. Eur. J. Clin. Nutr. 2011, 65,
657–660.
19. Holvik, K.; Brunvand, L.; Brustad, M.; Meyer, H.E. Vitamin D Status in the Norwegian
Population. In Solar Radiation and Human Health; Bjertness, E., Ed.; The Norwegian Academy
of Science and Letters: Oslo, Norway, 2008; pp. 216–228.
20. Viljakainen, H.T.; Korhonen, T.; Hytinantti, T.; Laitinen, E.K.; Andersson, S.; Makitie, O.;
Lamberg-Allardt, C. Maternal vitamin D status affects bone growth in early childhood—A
prospective cohort study. Osteoporos. Int. 2011, 22, 883–891.
21. Directorate of Health and Centre of Development. Infant and Children Care. Guidelines on
Health Protection of Children 0–5 Years, 3rd ed.; Directorate of Health: Reykjavik, Iceland, 2013.
22. Statistics Iceland. Population by Municipality, Age and Sex 1 December 1997–2013.
In Statistics, Population (Online); Statistics Iceland: Reykjavik, Iceland, 2013. Available online:
http://www.statice.is/ (accessed on 12 March 2014).
23. Vieth, R. Implications for 25-hydroxyvitamin D testing of public health policies about the benefits
and risks of vitamin D fortification and supplementation. Scand. J. Clin. Lab. Investig. Suppl.
2012, 243, 144–153.
24. The Public Health Institute of Iceland. Recommendations on Diet and Nutrients for Adults and
Children from Two Years of Age; Booklet: Reykjavik, Iceland, 2006.
25. Nielsen, A.; Michaelsen, K.F.; Holm, L. Parental concerns about complementary feeding:
Differences according to interviews with mothers with children of 7 and 13 months of age. Eur. J.
Clin. Nutr. 2013, 67, 1157–1162.
26. Kyttala, P.; Erkkola, M.; Kronberg-Kippila, C.; Tapanainen, H.; Veijola, R.; Simell, O.; Knip, M.;
Virtanen, S.M. Food consumption and nutrient intake in Finnish 1–6-year-old children.
Public Health Nutr. 2010, 13, 947–956.
 157

27. Kristjansdottir, A.G.; Thorsdottir, I. Adherence to food-based dietary guidelines and evaluation of
nutrient intake in 7-year-old children. Public Health Nutr. 2009, 12, 1999–2008.
28. Thorsdottir, I.; Gunnarsdottir, I. The Diet of Icelandic 9- and 15-Year-Old Children and Adolescents;
Unit for Nutrition Research and Public Health Institute of Iceland: Reykjavik, Iceland, 2006.
29. Thorgeirsdottir, H.; Valgeirsdottir, H.; Gunnarsdottir, I.; Gisladottir, E.; Gunnarsdottir, B.E.;
Thorsdottir, I.; Stefansdottir, J.; Steingrimsdottir, L. The Diet of Icelanders—Results from a
National Nutrition Survey 2010–2011; Directorate of Health, the Icelandic Food and Veterinary
Authority, Unit for Nutrition Research: Reykjavik, Iceland, 2011.
30. Gunnarsdottir, I.; Helgadottir, H.; Thorisdottir, B.; Thorsdottir, I. Diet of six-year-old Icelandic
children—National dietary survey 2011–2012. Laeknabladid 2013, 99, 17–23.
31. Steingrimsdottir, L.; Gunnarsson, O.; Indridason, O.S.; Franzson, L.; Sigurdsson, G. Relationship
between serum parathyroid hormone levels, vitamin D sufficiency, and calcium intake. JAMA
2005, 294, 2336–2341.
32. Gunnarsson, O.; Indriethason, O.S.; Franzson, L.; Halldorsdottir, E.; Sigurethsson, G. Vitamin-D
homeostasis amongst adult Icelandic population. Laeknabladid 2004, 90, 29–36.
33. Bjarnadottir, A.; Kristjansdottir, A.G.; Hrafnkelsson, H.; Johannsson, E.; Magnusson, K.T.;
Thorsdottir, I. Insufficient autumn vitamin-D intake and low vitamin-D status in 7-year-old
Icelandic children. Public Health Nutr. 2014, doi:10.1017/S1368980013003558.
34. Hower, J.; Knoll, A.; Ritzenthaler, K.L.; Steiner, C.; Berwind, R. Vitamin D fortification of
growing up milk prevents decrease of serum 25-hydroxyvitamin D concentrations during winter:
A clinical intervention study in Germany. Eur J. Pediatr. 2013, 172, 1597–1605.
35. Dawodu, A.; Akinbi, H. Vitamin D nutrition in pregnancy: Current opinion. Int. J. Women’s
Health 2013, 5, 333–343.
36. Janssen, M.J.; Wielders, J.P.; Bekker, C.C.; Boesten, L.S.; Buijs, M.M.; Heijboer, A.C.;
van der Horst, F.A.; Loupatty, F.J.; van den Ouweland, J.M. Multicenter comparison study of
current methods to measure 25-hydroxyvitamin D in serum. Steroids 2012, 77, 1366–1372.
37. Singh, R.J.; Taylor, R.L.; Reddy, G.S.; Grebe, S.K. C-3 epimers can account for a significant
proportion of total circulating 25-hydroxyvitamin D in infants, complicating accurate measurement
and interpretation of vitamin D status. J. Clin. Endocrinol. Metab. 2006, 91, 3055–3061.
38. Stepman, H.C.; Vanderroost, A.; Stockl, D.; Thienpont, L.M. Full-scan mass spectral evidence for
3-epi-25-hydroxyvitamin D(3) in serum of infants and adults. Clin. Chem. Lab. Med. 2011, 49,
253–256.
39. Madar, A.A.; Stene, L.C.; Meyer, H.E. Vitamin D status among immigrant mothers from Pakistan,
Turkey and Somalia and their infants attending child health clinics in Norway. Br. J. Nutr. 2009,
101, 1052–1058.
40. Holmlund-Suila, E.; Koskivirta, P.; Metso, T.; Andersson, S.; Makitie, O.; Viljakainen, H.T.
Vitamin D deficiency in children with a chronic illness-seasonal and age-related variations in
serum 25-hydroxy Vitamin D concentrations. PLoS One 2013, 8, e60856.

© 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).
158
 159

4. Children and Adolescents

Reprinted from Nutrients. Cite as: Gubbels, J.S.; Raaijmakers, L.G.M.; Gerards, S.M.P.L.; Kremers, S.P.J.
Dietary Intake by Dutch 1- to 3-Year-Old Children at Childcare and at Home. Nutrients 2014, 6, 304-318.

Dietary Intake by Dutch 1- to 3-Year-Old Children at Childcare

and at Home
Jessica S. Gubbels, Lieke G. M. Raaijmakers *, Sanne M. P. L. Gerards and Stef P. J. Kremers

Department of Health Promotion, NUTRIM School for Nutrition, Toxicology and Metabolism,
Maastricht University Medical Centre, PO Box 616, Maastricht 6200 MD, The Netherlands;
E-Mails: jessica.gubbels@maastrichtuniversity.nl (J.S.G.);
sanne.gerards@maastrichtuniversity.nl (S.M.P.L.G.); s.kremers@maastrichtuniversity.nl (S.P.J.K.)

* Author to whom correspondence should be addressed;

E-Mail: lieke.raaijmakers@maastrichtuniversity.nl; Tel.: +31-43-388-2401; Fax: +31-43-367-1032.

Received: 24 October 2013; in revised form: 10 December 2013 / Accepted: 24 December 2013 /
Published: 8 January 2014

Abstract: The goal of the current study was to assess dietary intake in a large sample
(N = 1016) of Dutch toddlers (1–3 years old), both at childcare and at home. Dietary intake
during two weekdays was recorded using an observation format applied by childcare staff
for intake at childcare, and partially pre-coded dietary journals filled out by parents for
intake at home. Children’s intake of energy, macronutrients and energy balance-related
food groups (fruit, vegetables, sweet snacks, savoury snacks) were compared with Dutch
dietary guidelines. In addition, differences between the dietary intake by various subgroups
(based on gender, age, childcare attendance, socio-economic status of childcare centre)
were explored using multilevel regression analyses, adjusting for nesting of children within
centres. Energy intake was high relative to dietary guidelines, and children consumed more
or less equal amounts of energy at home and at childcare. Dietary fibre, fruit and vegetable
and snack intakes were low. Intake at childcare mainly consisted of carbohydrates, while
intake at home contained more proteins and fat. The findings imply various opportunities
for childcare centres to improve children’s dietary intake, such as providing fruit and
vegetables at snacking moments. In addition, the findings underline the importance of
assessing dietary intake over a whole day, both at childcare and at home, to allow intake to
be compared with dietary guidelines.
 160

Keywords: childcare; day-care; dietary intake; dietary journal; nutrition; observation;

overweight; parent; toddler

1. Introduction

Worldwide, at least 42 million children under the age of five were overweight in 2010, and these
numbers are expected to continue to increase [1]. Childhood overweight is a major risk factor for
several chronic conditions such as cardiovascular diseases and type 2 diabetes mellitus [2]. Moreover,
childhood overweight is known to track into adulthood, in that overweight children often remain
overweight or obese during later life [3]. Dietary intake plays a crucial role in the development of
overweight [4]. Dietary habits are often established at a young age [5] and maintained throughout
life [6–8], indicating the urgency of increasing our understanding of the origin and development of
dietary habits in young children.
In Europe, over half of the toddlers (below primary school age) attend some form of childcare or
educational facilities [9]. It has been recommended that a child in full-time childcare (i.e., 8 h or
more per day) should consume one half to two-thirds of his or her daily dietary intake at
childcare [10], indicating the importance of childcare for the development of children’s dietary habits.
Childcare use has been found to be associated with an increased overweight risk throughout childhood
(e.g., [11–13]). Furthermore, various studies have shown that children attending childcare often do not
meet dietary intake recommendations: they may consume excess energy [14] and excessive amounts of
total fat [14,15], saturated fat [15,16] and sweets [14]. In addition, they are not consuming sufficient
amounts of fruit [16], vegetables [14,16,17] and dietary fibre [18]. However, several of these studies
were limited to dietary intake at childcare [16,19,20], ignoring the intake at home. As such, they have
to rely on the estimated proportion of the dietary intake that takes place at childcare. Since dietary
intake at home is not known, these studies assume the composition of the meals to be stable throughout
the day and do not take into account possible compensation behaviour at home. The studies that have
taken account of both dietary intake at home and intake at childcare [14,15,17,18,21–23] mostly had
small sample sizes (N < 200) [14,17,18,21,22], assessed dietary intake at childcare through the
parents [18], or only examined specific meals instead of dietary intake during a whole day [18].
In addition, the majority of the studies examining dietary intake have been conducted in the United
States [14–18,20,21], with only a few from Europe [19,22,23].
The current study aimed to assess dietary intake in terms of energy, macronutrients and the food
groups of fruit, vegetables, sweet snacks and savoury snacks, both at childcare and at home, in a large
sample (N = 1016) of Dutch toddlers (aged 1–3 years). In addition, it explored the dietary intake in
various subgroups (according to gender, age, childcare attendance and socio-economic status (SES) of
the childcare centre’s neighbourhood).

2. Methods

2.1. Respondents and Procedure

 161

Ethical approval for this study was not required according to Dutch law, since the current research
did not involve invasive procedures, and thus did not fall under the Dutch Medical Research Involving
Humans Act (Wet Medisch-Wetenschappelijk Onderzoek met Mensen) [24]. All childcare centres in
the Netherlands were approached to participate in the study from March 2011 onwards. Several
strategies were used to recruit childcare centres. A direct mailing of letters was sent to addresses
acquired by purchasing commercially available addresses. In addition, a digital mailing was sent, and
childcare centers were recruited at conferences and through appointments at the head offices of the
childcare organizations to which the centres belonged. If the head office was interested, the
recruitment was continued at the individual centres. All childcare centres were allowed to participate.
Sometimes a centre decided not to participate citing reasons such as that it would be too much effort,
the centre had been closed down, the parent committee did not agree or management had changed.
The 112 childcare centres that responded before August 2013 were included in the study. Data
collection started as soon as a childcare centre consented to participate. All parents of the children
aged 1 to 3 years from these centres were invited to participate. A total of 2788 children participated.
All parents of participating children provided informed consent. Children’s dietary intake was
recorded on two entire weekdays, randomly chosen during one week, both at home using food diaries
and at childcare using observations.

2.2. Assessment of Dietary Intake

In the Netherlands, children attending childcare usually consume their breakfast at home.
Subsequently, they consume a morning snack, lunch and afternoon snack at childcare, and their dinner
again at home.

2.2.1. Dietary Intake at Childcare

Staff at the childcare centre was instructed by a dietician to record the dietary intake of each of the
participating children on a poster. The poster was a partially pre-coded dietary record, providing a list
of the most common products that might consumed at each different eating moment. For instance, it
showed a list of sweet snacks, beverages and fruits commonly consumed at snacking moments in the
Netherlands. In addition, it provided space at each eating moment to record any other products
consumed which were not on the standard list. There was a separate column on the poster for each
participating child, where their intake could be recorded. Childcare staff was asked to specify the type
of product (e.g., whether the milk product consumed was milk, chocolate milk, butter milk or yoghurt
drink), the unit (e.g., whether it was a cup or a bottle), and the amount (i.e., number of units).
The first eating moment (the snacking moment of the first observation day) was recorded together
with the dietician, at which point the childcare staff received detailed instructions from the dietician on
how to record the dietary intake. If the childcare staff were still uncertain about any aspects, these
would also be explained by the dietician. During the rest of that day, and on the second observation
day, the childcare staff filled in the poster for all eating moments at the centre (i.e., morning snack,
lunch and afternoon snack).
An additional questionnaire was filled in by the childcare staff together with the dietician to record
further information regarding the meals and foods offered at the centre, such as the standard portion
 162

size used for certain products (e.g., how many mL were in the cups used) and the type and brand of
particular products (e.g., whether regular or low-fat margarine was used and what brand).

2.2.2. Dietary Intake at Home

Parents were also asked to record their child’s dietary intake at different eating moments at home
during the two measurement days (i.e., breakfast, dinner including dessert, and anything consumed
after dinner, including anything consumed during the night). The questionnaire consisted of a partially
pre-coded food journal, providing a list of common products that might be consumed at each different
eating moment. For instance, the food journal listed the bread and bread products, butters or
margarines, sandwich toppings, fruit, porridges and beverages that are often consumed at breakfast in
the Netherlands. In addition there was space to record any other products consumed at each eating
moment that were not on the standard list. Parents were asked to specify the type of product (e.g.,
whether bread was white or brown), the unit (e.g., whether it was a slice of bread or a roll), the brand,
and the amount (i.e., number of units).

2.3. Assessment of Background Characteristics

Children’s age (rounded off to whole months), gender and the number of days they attended
childcare were asked for in the parental questionnaire. The socio-economic status (SES) score of the
population catered for by each childcare centre was derived from the centre’s postal code. These SES
scores are standardized scores per neighbourhood, reflecting educational level, income, and work
status of the residents of that neighbourhood [25]. The SES scores were recoded into low, medium and
high, based on tertile cuts of all scores in the Netherlands [25].

2.4. Processing of Dietary Intake Data

Only respondents for whom complete dietary intake data were available (for both measurement
days, both at childcare and at home) were retained in the analyses. Of the 2788 children participating
in the entire study, 1016 (43.7%) provided complete dietary intake data for both measurement days,
both at home and at childcare. Of the 1773 children without complete data, the majority (75.0%) had
complete data at childcare, but data at home was only available for 1 day or no days at all.
Furthermore, 24.0% had only attended childcare on the day, so data for two complete days at childcare
could not be provided, and 1.0% only provided data for intake at home, but not for childcare. The
1016 children with complete data were included in the final analyses.
The observed and reported dietary intake data of the children were entered by the dieticians in the
FoodFigures Program [26] separately for each of the six eating moments (breakfast, morning snack,
lunch, afternoon snack, dinner, evening snack). The amounts consumed as reported by the childcare
staff and parents were recalculated by this program into weights and volumes using the procedures on
measures and weights of the Dutch nutrient database [27] where necessary (e.g., using a standardized
weight for a slice of bread). Amounts of half or a quarter of a portion were also recalculated by the
program. As the focus of the current paper was on dietary intake, the average intake per day of the
following nutrients was calculated by the program, based on the Dutch nutrient database [27]: energy
 163

(in kcal), proteins (in energy percent (en%)), carbohydrates (en%), total fat (en%), saturated fat (en%),
unsaturated fat (en%) and fibre (grams (g)). In addition, intake from the following energy balance-related
food groups was calculated: fruit (g), vegetables (g), sweet snacks (g; including sweets (e.g., jelly
candy, liquorice, marshmallows), chocolate, cookies (e.g., butter cookies) and pastry (e.g., cake, pie))
and savoury snacks (g; including salty snacks (e.g., potato chips) and fried snacks (e.g., fried meats)).

2.5. Data Analyses

All analyses were conducted using SPSS 20.0 [28]. p-Values < 0.05 were considered statistically
significant. Independent t-tests and chi-square tests were conducted to compare the background
characteristics (children’s age, gender, childcare attendance, and the childcare centre’s SES) of the
children in the final sample with those of children who had incomplete data and were thus excluded.
Descriptive statistics were used to explore the children’s background characteristics and total
dietary intake. In addition, children’s total dietary intake was compared with the dietary guidelines for
toddlers from the Netherlands Nutrition Centre (Voedingscentrum; see Table 1) [29]. The overall
dietary guidelines applied in the current study were specific Dutch guidelines [29]. The dietary
guidelines for toddlers from the Netherlands Nutrition Centre refer to the guidelines for a healthy food
choice of the Netherlands Nutrition Centre for a balanced dietary intake for children of one year and
older. These guidelines are therefore used as a benchmark source for nutrient and food group analyses.
Next, children’s dietary intakes at home and at childcare were analysed separately, as well as their
intakes at different eating moments (breakfast, morning snack, lunch, afternoon snack, dinner and
evening snack).

Table 1. Total dietary intake by toddlers, compared to national guidelines (N = 1016).

Percentage of
Actual Total Dietary
Dietary Intake Dietary Guideline a Children Meeting the
Intake Mean (SD)
Guideline
Energy (kcal) 1285.1 (238.2) 1200 37.5% b
Carbohydrates (en%) 55.7 (5.2) • 98.1%
Proteins (en%) 14.3 (2.1) ” 99.2%
Fat
Total (en%) 30.0 (4.8) 25–40 83.2% c
Saturated (en%) 10.7 (1.8) ” 98.6%
Unsaturated (en%) 16.6 (3.7) - -
Dietary fibre (g) 12.5 (2.7) • 17.2%
Fruit (g) 124.1 (61.9) • 27.5%
69.3% d
Vegetables (g) 64.7 (36.5) •–100
17.0% e
Sweet snacks f (g) 13.5 (12.0) - -
Savoury snacks g (g) 0.7 (4.0) - -
Notes: en% = energy percent, g = grams, mL = millilitres, N = number of children, SD = standard deviation;
a
Nutritional guidelines from the Netherlands Nutrition Centre ([29]). No specific guidelines are available for
unsaturated fats and snacks; b
10% deviation from the guideline allowed (i.e., 1080–1320 kcal). Below the
c
guideline: 18.8%; above the guideline: 43.7%; below the guideline: 14.9%; above the guideline: 2.0%;
 164

d
N (%) meeting the guideline of 50 g/day; e N (%) meeting the guideline of 100 g/day; f Including sweets,
chocolate, cookies and pastry; g Including salty snacks and fried snacks.

In addition, the intakes by subgroups of children based on gender (boys vs. girls), age (2, 3 or
4 years old), childcare attendance (up to 2 days vs. 3 or more days a week) and the childcare centre’s
SES (low, medium or high) were explored. Multi-level linear regression analyses were conducted to
examine the associations between these background variables and the dietary intake variables,
corrected for the nesting of children within childcare centres and the background variables.

3. Results

Of the 1016 children, 54.8% (N = 554) were male. The average age was 2 years and 1 month
(SD = 10 months), with 313 1-year-olds (34.1%), 330 2-year-olds (36.0%) and 274 3-year-olds
(29.9%). On average, the children went to childcare for 2.4 days per week (SD = 0.6). A total of 24.5%
of the children attended childcare centres that were located in low-SES neighbourhoods, 28.2% in
medium-SES neighbourhoods, and 47.3% in high-SES neighbourhoods.
Children included in the final sample attended childcare for slightly more days a week than children
with incomplete dietary intake data (2.4 vs. 2.1, p < 0.001). There were no other significant differences
in background characteristics (age, gender, childcare SES) between children who did or did not
drop out.

3.1. Dietary Intake and Guideline Compliance

Table 1 lists the total dietary intake of the toddlers in the current study, as well as the number and
percentage of children complying with the dietary guidelines of the Netherlands Nutrition
Centre [29]. About one third of the children (37.5%) complied with the guidelines regarding energy
intake. A smaller group (18.8%) consumed less than the recommended amount of energy, while most
children (43.7%) consumed more energy than recommended.
The vast majority of the children met the guidelines with regard to macronutrients (carbohydrates,
proteins and fat). However, only 17.2% of the children consumed sufficient dietary fibre. This was
also reflected in the low percentages of children consuming sufficient fruit and vegetables. Snack
intake (both sweet and savoury) was generally low.

3.2. Dietary Intake at Childcare and at Home

Children consumed more or less equal amounts of energy at home and at childcare. However, while
their intake at childcare mainly consisted of carbohydrates, a relatively larger proportion of the intake
at home consisted of proteins and fat (see Table 2). Furthermore, children consumed most of their fruit
at childcare, and most of their vegetables at home. Sweet snacks were mostly eaten at childcare.

3.3. Dietary Intake at Different Eating Moments

Table 3 shows the children’s dietary intake at different eating moments, as well as the intake as
a proportion of total daily intake (from different meals). The percentage of children consuming food
 165

at each of the eating moments was very high (98.6%–99.8%), except for evening snacks, which were
consumed by only 62.8% of the children.
The main energy sources were lunch and dinner, together accounting for over half (57.4%) of the
total energy intake. The snacking moments were very high in carbohydrates, while the main
meals contained relatively larger proportions of proteins and fat. The main sources of dietary fibre
were the main meals. Most fruit was consumed during the morning snacking moment, while the
afternoon snacking moment often involved sweet snacks (e.g., cookies, sweets, pastry). The evening
snacking moment involved a relatively large proportion of saturated fat compared to the other
snacking moments.

Table 2. Dietary intake by toddlers at childcare and at home (N = 1016).

Dietary Intake Mean (SD)
Dietary Intake
At Childcare At Home
Energy (kcal) 652.2 (169.9) 633.5 (171.9)
Carbohydrates (en%) 62.4 (6.6) 49.1 (8.4)
Proteins (en%) 12.0 (2.4) 16.8 (3.7)
Fat
Total (en%) 25.6 (5.6) 34.1 (7.8)
Saturated (en%) 9.6 (2.3) 11.7 (2.8)
Unsaturated (en%) 13.4 (3.9) 19.5 (6.1)
Dietary fibre (g) 6.6 (1.8) 5.9 (2.1)
Fruit (g) 98.7 (46.0) 25.5 (41.4)
Vegetables (g) 10.9 (21.6) 53.8 (33.4)
Sweet snacks a (g) 10.3 (9.4) 3.3 (7.9)
Savoury snacks b (g) 0.2 (1.5) 0.5 (3.7)
Notes: en% = energy percent, g = grams, mL = millilitres; a Including sweets, chocolate, cookies and pastry;
b
Including salty snacks and fried snacks.

3.4. Dietary Intake in Subgroups

Overall, there were few differences in total dietary intake between boys and girls. Boys consumed
significantly more energy than girls (1304.8 vs. 1264.6 kcal, p < 0.01), as well as more dietary fibre
(12.7 vs. 12.2 g, p < 0.02). There were no significant differences between boys and girls in intake
specifically at childcare.
Intake of energy was significantly higher among older children, both specifically at childcare and
for the day as a whole (see Table 4). Dietary fibre intake also increased with age, mainly at childcare.
Sweet snacks intake increased with age, although at childcare, this increase was only significant
between 2 and 3 years, while the increase in overall sweet snacks intake was only significant between
1 and 2 years. Total savoury snack intake increased between the ages of 1 and 2 years.
Children who attended childcare for 3 or more days a week had a higher total vegetables
consumption (73.3 vs. 62.0 g, p < 0.02), and consumed more savoury snacks (1.1 vs. 0.6 g, p < 0.04;
results not tabulated) than children attending childcare for 2 days or less. Childcare attendance was not
significantly related to specific dietary intake at childcare.
Nutrients 2014, 6 166

Table 3. Dietary intake by toddlers at different eating moments.

Dietary Intake Mean (SD)
Breakfast Morning Snack Lunch Afternoon Snack Dinner Evening Snack
at Home at Childcare at Childcare at Childcare at Home at Home
Dietary Intake
(N = 1006) (N = 1010) (N = 1014) (N = 1002) (N = 1013) (N = 638)
Mean % of total Mean % of Total Mean % of Total Mean % of Total Mean % of Total Mean % of Total
(SD) intake (SD) Intake (SD) Intake (SD) Intake (SD) Intake (SD) Intake
237.7 18.1 126.1 9.6 378.5 28.8 147.6 11.2 358.8 27.2 66.8 5.1
Energy (kcal)
(87.1) (51.8) (122.9) (67.9) (120.0) (63.5)
54.9 14.5 85.4 22.6 49.6 13.2 81.2 21.5 43.0 11.4 63.6 16.8
Carbohydrates (en%)
(10.6) (10.8) (7.7) (12.8) (11.6) (23.1)
15.0 19.6 5.2 6.8 16.2 21.3 6.4 8.4 19.1 25.1 14.3 18.8
Proteins (en%)
(4.6) (3.5) (3.9) (4.6) (5.4) (12.4)
Fat
30.0 20.5 9.5 6.5 34.5 23.5 12.5 8.5 37.9 25.9 22.1 15.1
Total (en%)
(10.0) (8.5) (7.1) (10.2) (11.1) (16.1)
11.5 20.8 3.1 5.6 12.9 23.3 5.0 9.1 12.1 21.9 10.7 19.3
Saturated (en%)
(4.2) (4.6) (3.1) (5.4) (3.6) (8.9)
15.7 20.8 4.1 5.4 18.3 24.2 5.5 7.3 22.7 30.0 9.3 12.3
Unsaturated (en%)
(7.8) (4.5) (5.2) (6.2) (9.0) (11.1)
2.6 20.5 1.5 11.8 3.9 30.7 1.2 9.5 3.0 23.6 0.5 3.9
Dietary fibre (g)
(1.2) (0.9) (1.4) (0.9) (1.3) (0.9)
7.5 5.9 70.5 55.9 1.2 1.0 30.7 24.3 12.8 10.2 3.4 2.7
Fruit (g)
(22.6) (51.7) (8.6) (45.0) (27.6) (14.2)
0.3 0.5 0.1 0.2 5.1 7.8 2.8 4.3 56.4 87.0 0.1 0.2
Vegetables (g)
(3.0) (1.0) (15.8) (10.2) (32.5) (2.0)
1.2 8.7 3.6 26.1 0.2 1.4 7.3 52.9 0.8 5.8 0.7 5.1
Sweet snacks a (g)
(4.3) (6.4) (1.3) (8.2) (4.1) (3.0)
0.0 0.0 0.0 0.0 0.0 0.0 0.3 42.9 0.4 57.1 0.0 0.0
Savoury snacks b (g)
(0.5) (0.5) (0.0) (1.6) (3.7) (0.4)
Notes: en% = energy percent, g = grams, mL = millilitres; a Including sweets, chocolate, cookies and pastry; b Including salty snacks and fried snacks.
Nutrients 2014, 6 167

Table 4. Dietary intake differences based on age.

Dietary Intake at Childcare Dietary Intake during a Whole Day
1-Year-Olds 2-year-olds a 3-Year-Olds 1-Year-Olds 2-Year-Olds a 3-Year-Olds
Dietary Intake b
Mean (SD) Mean (SD) Mean (SD) Significance Mean (SD) Mean (SD) Mean (SD) Significance b
N = 313 N = 330 N = 274 N = 313 N = 330 N = 274
Energy (kcal) 570.7 (150.1) 662.3 (158.2) 726.8 (170.9) *** 1165.9 (209.1) 1305.9 (210.1) 1400.4 (231.7) ***
Carbohydrates (en%) 62.5 (7.1) 63.0(5.8) 62.3 (6.3) 55.7 (5.5) 56.0 (5.1) 55.8 (4.8)
Proteins (en%) 12.1 (2.8) 11.8 (2.1) 12.0 (2.3) 14.5 (2.4) 14.1 (1.9) 14.2 (2.1)
Fat
Total (en%) 25.4 (5.8) 25.2 (5.2) 25.6 (5.3) 29.8 (5.1) 29.9 (4.7) 30.0 (4.4)
Saturated (en%) 9.6 (2.4) 9.4 (2.1) 9.5 (2.1) 10.7 (1.9) 10.6 (1.7) 10.6 (1.6)
Unsaturated (en%) 13.3 (4.0) 13.3 (3.8) 13.5 (3.7) 16.5 (3.9) 16.5 (3.5) 16.7 (3.5)
Dietary fibre (g) 6.1 (1.7) 6.6 (1.9) 7.2 (1.8) *** 12.3 (2.6) 12.3 (2.6) 13.0 (2.7) *c
Fruit (g) 99.8 (45.8) 98.4 (49.5) 98.9 (45.0) 122.6 (62.3) 124.3 (63.4) 127.8 (63.1)
Vegetables (g) 13.5 (23.7) 9.8 (20.5) 9.8 (20.7) 68.5 (36.3) 61.9 (36.1) 64.6 (38.2)
Sweet snacks d (g) 8.8 (8.3) 9.9 (8.6) 12.1 (11.4) *e 11.3 (10.6) 13.6 (12.2) 16.0 (13.5) *c
Savoury snacks f (g) 0.1 (1.3) 0.2 (1.5) 0.5 (1.9) 0.3 (1.9) 1.1 (5.8) 0.9 (3.5) *c
b
Notes: en% = energy percent, g = grams, mL = millilitres; a Reference category; Adjusted significance from multivariate multi-level regression analyses, adjusted for gender, childcare
c
attendance and socioeconomic status score of the childcare centre; Only significant for the 1-year-olds; d Including sweets, chocolate, cookies and pastry; e Only significant for the 3-year-olds;
* p < 0.05, ** p < 0.01, *** p < 0.001; f Including salty snacks and fried snacks.
 168

There were no differences in overall intake between childcare centres with different SES. With
regard to the specific intake at childcare, children at high-SES childcare centres consumed
significantly less fruit (93.0 g) than children at medium- and low-SES centres (106.2 g and 101.2 g,
respectively, p < 0.04). On the other hand, they consumed significantly more vegetables at childcare
(14.9 g) compared to children from medium- and low-SES centres (7.2 g and 7.3 g, respectively,
p < 0.01). Children at low-SES childcare centres consumed significantly lower amounts of energy
(619.0 kcal) than those at medium- and high-SES centres (679.8 and 652.9 kcal, respectively,
p < 0.04). Finally, children at low-SES centres consumed significantly less savoury snacks (0.4 g
compared to 0.1 g and 0.3 g in medium- and high-SES centres, respectively, p < 0.05; results
not tabulated).

4. Discussion

The current study assessed dietary intake at childcare and at home in a large sample (N = 1016) of
Dutch toddlers (1–3 years) who attended childcare. Energy intake was high relative to dietary
guidelines, while dietary fibre, fruit and vegetable intakes were low. Snack intake (both sweet and
savoury) was low. In 2005 and 2006, a national food consumption survey was conducted among
toddlers in the Netherlands including children who attended childcare as well as those who did not.
The dietary intake among the 2- to 3-year-olds (N = 788) from that survey was very similar to the
intake we found in the current sample, specifically as regards the intake of energy, all macronutrients,
dietary fibre and fruit (differences all <5%) [30]. This indicates that the overall dietary intake by
children attending childcare does not seem to be very different from that by children not using
childcare. Compared to the national survey, however, children in the current sample appeared to
consume far less snacks (13.5 g of sweet snacks vs. 47 g in the national survey; and <1 g of savoury
snacks vs. 3 g in the national survey) and more vegetables (64.7 g compared to 40 g in the national
survey) [30].
It is unclear why there were such considerable differences with regard to vegetable and snack
intakes, but not with regard to any other dietary intake measures. Perhaps the fact that the current study
included 1-year-olds can partly explain these differences, especially with regard to snacks, because the
1-year-olds in the current sample consumed significantly less snacks than the older children.
Furthermore, it should be noted that the current study did not include days on which the children did
not attend childcare. It is possible that the children from the current sample had different intake
patterns during a full day at home. However, a previous study by Ziegler and colleagues [18], which
compared lunch and snacking moments at childcare with the corresponding eating moments during a
full day at home, found only slight differences in intake between these two locations, which were only
significant for the afternoon snacking moment: at home, children seemed to consume a bit more
protein and fat in the afternoon. Furthermore, Ziegler et al. [18] reported more frequent consumption
of salty snacks in the afternoon at home than at childcare. A study by Lehtisalo [23] that compared the
dietary intake of children cared for at childcare with that of children cared for at home found lower
vegetable consumption and higher sweet pastry consumption by the children cared for at home. These
findings are in line with the deviating vegetable and snack consumption in the current sample
compared to the national survey [30]. A final explanation for the differences between the current study
 169

and the national survey may regard the fact that the current study did not include weekend days.
Several studies have shown that children’s dietary intake is generally less healthy on weekend days
(e.g., [22,23,31]), possibly explaining the lower snack consumption and higher vegetable intake in the
current study.
In line with previous research among young children (e.g., [30,32]), the children in the current
sample skipped very few meals and snacking moments: 98.6%–99.8% of the children consumed food
at each of the eating moments (except for an evening snack, consumed by 62.4%). As regards the
quality of children’s diets, the macronutrient content of their diets seemed to be very good, with 83.2%
to 99.2% of the children meeting the guidelines for carbohydrates, proteins, total fat and saturated fat.
Studies from the US found excess consumption of total and saturated fat in childcare [14–16], perhaps
reflecting a cultural difference between the US and the Netherlands. However, in line with previous
US studies [14,16–18], many children in the current sample did not consume sufficient dietary fibre,
vegetables and fruit. Furthermore, almost half of the children consumed excess amounts of energy
(i.e., >1320 kcal), which is in line with previous research [14]. About equal amounts of energy were
consumed at home and at childcare. Energy intake at the different eating moments in the current study
was comparable to that found in US studies [18,20].
There were few differences in dietary intake between subgroups, both at childcare and in total.
In line with previous research [19], boys consumed more energy than girls. In addition, boys consumed
more dietary fibre. Concerns about children’s diet seemed to change with age: while younger children
were more likely to consume insufficient dietary fibre, the older children often consumed more snacks
and energy. These differences were visible specifically at childcare as well as during a whole day, with
the exception of savoury snack intake, whose increase was only significant as regards intake during a
whole day. With regard to childcare attendance, children who attended three or more days a week
consumed more vegetables and savoury snacks, though not at childcare, indicating that this increased
vegetable and snack intake took place at home. Furthermore, we found older children to consume more
energy, dietary fibre, sweets and snacks. Despite the fact that The Netherlands Nutrition Centre
recommends the same intake for children aged 1–4 in their guidelines [29], our results show that
children within this age group have different needs. Children in the age of 1 may for example still be
nursed which influences their dietary intake.
Although the overall consumption of snacks seemed to be low (13.5 g of sweet snacks and less than
one gram of savoury snacks per day on average), the majority of the sweet snacks were consumed at
childcare, especially during the afternoon snacking moment. Fruit was consumed especially during the
morning snack at childcare, and to a lesser extent in the afternoon. This indicates an opportunity for
childcare centres to improve children’s fruit consumption (which was too low for almost three quarters
of the children), and at the same time even further lower snack consumption, by replacing the
afternoon sweet snacks with fruit. Fruit consumption seemed to be especially low in high-SES
childcare centres, while intake at home was not significantly different between childcare centres with
different SES. Previous studies have repeatedly shown that children from low-SES families often
consume less fruit (e.g., [33,34]). However, children from high-SES childcare centres in the current
study also consumed more vegetables. It seems that high-SES childcare centres place more emphasis
on vegetable intake, and less on fruit intake.
 170

Various previous studies examining dietary intake at childcare have used the guidelines of the
American Dietetic Association (that a child who spends a full day at childcare [i.e., 8 h or more]
should consume one half to two-thirds of his or her daily dietary intake at childcare [10]) to convert
daily dietary intake guidelines into estimated guidelines for intake at childcare (e.g., [16,17,19]).
However, such conversion into childcare-specific guidelines ignores the fact that the composition of
meals and other eating moments is not stable throughout a day (e.g., the composition of a typical lunch
is different from the composition of a typical dinner), as the current study shows. It therefore makes no
sense to apply the same guidelines at home and at childcare. This underlines the importance of studies
assessing dietary intake during a total day, both at home and at childcare, enabling comparison with
daily intake guidelines.
The current study had several limitations. There was a relatively high percentage of incomplete
cases (56.3%), although the final sample included in the study can still be considered very large (over
1000 children from over 100 different childcare centres) compared to previous studies. This large
sample size also provided sufficient statistical power to correct the analyses for the multi-level
structure of the data. Nonetheless, there was some selective drop-out with regard to longer childcare
attendance. In addition, the data collection took place over a relatively long period (30 months), which
could have influenced the results. A strength of the current study was that dietary intake was assessed
both at childcare and at home, making it possible to compare the intake with dietary guidelines without
having to estimate the proportion of intake taking place at childcare. However, dietary intake at
childcare was observed and recorded by childcare staff, while dietary intake at home was self-reported
by parents, possibly introducing bias. Moreover, the dietary intake assessment methodologies used in
both settings (childcare and home) were not validated, and weekend days were not assessed in the
current study. With regards the software used to recalculate intake, participants were not asked
whether children consumed their entire plate/glass which may have biased the reported intake.
However, the ability of the program to register amounts of half or a quarter of a portion helped
accurate assessment of children’s intake. Moreover, the dietary intake in the current sample was very
similar to the intake in a previous national survey, indicating that the assessment methods in the
current study were probably sufficiently reliable.

5. Conclusions

In terms of energy balance, the main concern of children’s dietary intake in our study was their low
fibre, vegetable and fruit consumption, and their high energy intake, putting them at risk for
developing overweight. Childcare has a large potential to contribute to resolving these issues, for
instance by offering fruit as a snack twice instead of once a day, and by providing vegetables during
lunch and snacking moments. This could potentially also further lower snack consumption and thereby
lower energy intake, thus also reducing the overweight risk. Previous studies have shown that
childcare staff can have an important positive influence on children’s dietary intake at childcare, for
example by serving sufficient healthy foods [16,35], preferably using a family serving style (in which
the child can take healthy foods him/herself and can decide how much to take) [19,36,37] or indulgent
feeding style (giving and offering seconds for healthy foods) [37,38]; by being a positive role model
 171

and eating healthy foods together with the children [19]; and by talking about healthy foods with the
children [19].
As regards research, the current findings underline the importance of future research assessing
dietary intake during a total day, both at home and at childcare. This enables comparison with daily
guidelines, instead of having to convert these guidelines to improvised childcare-specific guidelines.
In addition, intake should be assessed during a full day at home as well, including weekend days, to be
able to compare intake at home and at childcare and to check for variability in children’s diets
across locations.

Acknowledgments

The data collection for this study was financially supported by Nutricia, as part of the Eet Compleet
Test. Nutricia had no influence on the analysis and reporting of this study. We are grateful to all
childcare staff, children and parents who participated in the study.

Conflicts of Interest

The authors declare no conflict of interest.

References

1. World Health Organization. Childhood Overweight and Obesity on the Rise, 2010. Available
online: http://www.who.int/dietphysicalactivity/childhood/en/ (accessed on 10 September 2013).
2. World Health Organization (WHO). Obesity: Preventing and Managing the Global Epidemic;
WHO: Geneva, Switzerland, 2004.
3. Singh, A.S.; Mulder, C.; Twisk, J.W.; van Mechelen, W.; Chinapaw, M.J. Tracking of childhood
overweight into adulthood: A systematic review of the literature. Obes. Rev. 2008, 9, 474–488.
4. Rennie, K.L.; Johnson, L.; Jebb, S.A. Behavioural determinants of obesity. Best Pract. Res. Clin.
Endocrinol. Metab. 2005, 19, 343–358.
5. Savage, J.S.; Fisher, J.O.; Birch, L.L. Parental influence on eating behavior: Conception to
adolescence. J. Law Med. Ethics 2007, 35, 22–34.
6. Kelder, S.H.; Perry, C.L.; Klepp, K.I.; Lytle, L.L. Longitudinal tracking of adolescent smoking,
physical activity, and food choice behaviors. Am. J. Public Health 1994, 84, 1121–1126.
7. Singer, M.R.; Moore, L.L.; Garrahie, E.J.; Ellison, R.C. The tracking of nutrient intake in young
children: The Framingham Children’s Study. Am. J. Public Health 1995, 85, 1673–1677.
8. Te Velde, S.J.; Twisk, J.W.; Brug, J. Tracking of fruit and vegetable consumption from
adolescence into adulthood and its longitudinal association with overweight. Br. J. Nutr. 2007, 98,
431–438.
9. Blackburn, P. Childcare Services in the EU—What Future? European Foundation for the
Improvement of Living and Working Conditions: Dublin, Ireland, 2006.
10. Benjamin Neelon, S.E.; Briley, M.E.; American Dietetic Association. Position of the American
Dietetic Association: Benchmarks for nutrition in child care. J. Am. Diet. Assoc. 2011, 111,
607–615.
 172

11. Gubbels, J.S.; Kremers, S.P.; Stafleu, A.; Dagnelie, P.C.; De Vries, N.K.; van Buuren, S.;
Thijs, C. Child-care use and the association with body mass index and overweight in children
from 7 months to 2 years of age. Int. J. Obes. 2010, 34, 1480–1486.
12. Benjamin, S.E.; Rifas-Shiman, S.L.; Taveras, E.M.; Haines, J.; Finkelstein, J.; Kleinman, K.;
Gillman, M.W. Early child care and adiposity at ages 1 and 3 years. Pediatrics 2009, 124,
555–562.
13. Geoffroy, M.C.; Power, C.; Touchette, E.; Dubois, L.; Boivin, M.; Séguin, J.R.; Tremblay, R.E.;
Côté, S.M. Childcare and overweight or obesity over 10 years of follow-up. J. Pediatr. 2013, 162,
753–758.e1.
14. Briley, M.E.; Jastrow, S.; Vickers, J.; Roberts-Gray, C. Dietary intake at child-care centers and
away: Are parents and care providers working as partners or at cross-purposes? J. Am. Diet.
Assoc. 1999, 99, 950–954.
15. Bollella, M.C.; Spark, A.; Boccia, L.A.; Nicklas, T.A.; Pittman, B.P.; Williams, C.L.
Nutrient intake of Head Start children: Home vs. school. J. Am. Coll. Nutr. 1999, 18, 108–114.
16. Ball, S.C.; Benjamin, S.E.; Ward, D.S. Dietary intakes in North Carolina child-care centers: Are
children meeting current recommendations? J. Am. Diet. Assoc. 2008, 108, 718–721.
17. Padget, A.; Briley, M.E. Dietary intakes at child-care centers in central Texas fail to meet Food
Guide Pyramid Recommendations. J. Am. Diet. Assoc. 2005, 105, 790–793.
18. Ziegler, P.; Briefel, R.; Ponza, M.; Novak, T.; Hendricks, K. Nutrient intakes and food patterns of
toddlers’ lunches and snacks: Influence of location. J. Am. Diet. Assoc. 2006, 106, S124–S134.
19. Gubbels, J.S.; Kremers, S.P.; Stafleu, A.; Dagnelie, P.C.; De Vries, N.K.; Thijs, C. Child-care
environment and dietary intake of 2- and 3-year-old children. J. Hum. Nutr. Diet. 2010, 23,
97–101.
20. Nicklas, T.A.; Liu, Y.; Stuff, J.E.; Fisher, J.O.; Mendoza, J.A.; O’Neil, C.E. Characterizing lunch
meals served and consumed by pre-school children in Head Start. Public Health Nutr. 2013, 16,
2169–2177.
21. Bruening, K.S.; Gilbride, J.A.; Passannante, M.R.; McClowry, S. Dietary intake and health
outcomes among young children attending 2 urban day-care centers. J. Am. Diet. Assoc. 1999, 99,
1529–1535.
22. Sepp, H.; Lennernas, M.; Pettersson, R.; Abrahamsson, L. Children’s nutrient intake at preschool
and at home. Acta Paediatr. 2001, 90, 483–491.
23. Lehtisalo, J.; Erkkola, M.; Tapanainen, H.; Kronberg-Kippila, C.; Veijola, R.; Knip, M.;
Virtanen, S.M. Food consumption and nutrient intake in day care and at home in 3-year-old
Finnish children. Public Health Nutr. 2010, 13, 957–964.
24. Central Committee on Research Involving Human Subjects. Available online: http://www.ccmo.
nl/nl/uw-onderzoek-wmo-plichtig-of-niet (accessed on 20 December 2013).
25. Sociaal en Cultureel Planbureau. Statusscores van Wijken in Nederland, 2010. Available online:
http://www.scp.nl/content.jsp?objectid=default:20133 (accessed on 10 September 2013).
26. Foodfigures. Dé Voedingscalculator voor Health Professionals! Available online:
http://www.foodfigures.nl/ (accessed on 6 January 2014).
 173

27. Donders-Engelen, M.; van der Heijden, L.; Hulshof, K. Maten, Gewichten en Codenummers 2003
[CD-ROM]; Wageningen University: Wageningen, The Netherlands, 2003. Available online:
http://library.wur.nl/WebQuery/clc/1844390 (accessed on 10 September 2013).
28. IBM SPSS Statistics for Windows, version 20.0; IBM Corporation: Armonk, NY, USA.
29. Voeding van Zuigelingen en Peuters. Uitgangspunten voor de Voedingsadvisering voor Kinderen
van 0–4 Jaar. Available online: http://www.voedingscentrum.nl/Assets/Uploads/Documents/
Voedingsvoorlichters/Zuigelingen%20en%20peuters.pdf (accessed on 10 September 2013).
30. Ocké, M.C.; van Rossum, C.T.M.; Fransen, H.P.; Buurma, E.J.M.; de Boer, E.J.;
Brants, H.A.M.; Niekerk, E.M.; van der Laan, J.D.; Drijvers, J.J.M.M.; Ghameshlou, Z.
Dutch National Food Consumption Survey—Young Children 2005/2006; National Institue for
Public Health and the Environment: Bilthoven, The Netherlands, 2008; Available online:
http://www.rivm.nl/dsresource?objectid=rivmp:13711&type=org&disposition=inline&ns_nc=1
(accessed on 10 September 2013).
31. Rothausen, B.W.; Matthiessen, J.; Hoppe, C.; Brockhoff, P.B.; Andersen, L.F.; Tetens, I.
Differences in Danish children’s diet quality on weekdays v. weekend days. Public Health Nutr.
2012, 15, 1653–1660.
32. Gubbels, J.S.; Kremers, S.P.; Stafleu, A.; Goldbohm, R.A.; De Vries, N.K.; Thijs, C.
Clustering of energy balance-related behaviors in 5-year-old children: Lifestyle patterns and their
longitudinal association with weight status development in early childhood. Int. J. Behav. Nutr.
Phys. Act. 2012, 9, 77.
33. Hilsen, M.; van Stralen, M.M.; Klepp, K.I.; Bere, E. Changes in 10–12 year old’s fruit and
vegetable intake in Norway from 2001 to 2008 in relation to gender and socioeconomic status—A
comparison of two cross-sectional groups. Int. J. Behav. Nutr. Phys. Act. 2011, 8, 108.
34. Rodenburg, G.; Oenema, A.; Kremers, S.P.; van de Mheen, D. Parental and child fruit
consumption in the context of general parenting, parental education and ethnic background.
Appetite 2012, 58, 364–372.
35. Zuercher, J.L.; Kranz, S. Toddlers and preschoolers consume more dietary fiber when high-fiber
lunch items are served. Child Obes. 2012, 8, 71–75.
36. Johnson, S.L.; Birch, L.L. Parents’ and children’s adiposity and eating style. Pediatrics 1994, 94,
653–661.
37. Fisher, J.O.; Rolls, B.J.; Birch, L.L. Children’s bite size and intake of an entree are greater with
large portions than with age-appropriate or self-selected portions. Am. J. Clin. Nutr. 2003, 77,
1164–1170.
38. Hughes, S.O.; Patrick, H.; Power, T.G.; Fisher, J.O.; Anderson, C.B.; Nicklas, T.A. The impact of
child care providers’ feeding on children’s food consumption. J. Dev. Behav. Pediatr. 2007, 28,
100–107.

© 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).
174
 175

Reprinted from Nutrients. Cite as: Chen, S.; Binns, C.W.; Maycock, B.; Liu, Y.; Zhang, Y. Prevalence
of Dietary Supplement Use in Healthy Pre-School Chinese Children in Australia and China. Nutrients
2014, 6, 815-828.

Prevalence of Dietary Supplement Use in Healthy Pre-School

Chinese Children in Australia and China
Shu Chen 1, Colin W. Binns 2,*, Bruce Maycock 3, Yi Liu 4 and Yuexiao Zhang 5
1
School of Public Health, Curtin University, GPO Box U1907 Perth, Western Australia 6845,
Australia; E-Mail: shu.chen2@postgrad.curtin.edu.au
2
School of Public Health, Curtin University, GPO Box U1907 Perth, Western Australia 6845,
Australia
3
School of Public Health, Curtin University, GPO Box U1907 Perth, Western Australia 6845,
Australia; E-Mail: B.Maycock@curtin.edu.au
4
School of Public Health, Sichuan University, No.17 Section 3 Renmin South Road Chengdu,
Chengdu 610041, China; E-Mail: hxliuyi@163.com
5
Wuhan Maternal and Child Medical Centre, 100 Xianggang Rd, Wuhan 430016, China;
E-Mail: zhangyuexiao86@sina.com

* Author to whom correspondence should be addressed; E-Mail: C.Binns@curtin.edu.au;

Tel.: +61-8-9266-2952; Fax: +61-8-9266-2958.

Received: 13 December 2013; in revised form: 10 February 2014 / Accepted: 11 February 2014 /
Published: 21 February 2014

Abstract: There is a growing use of dietary supplements in many countries including

China. This study aimed to document the prevalence of dietary supplements use and
characteristics of Chinese pre-school children using dietary supplements in Australia and
China. A survey was carried out in Perth, Western Australia of 237 mothers with children
under five years old and 2079 in Chengdu and Wuhan, China. A total of 22.6% and 32.4%
of the Chinese children were taking dietary supplements in Australia and China,
respectively. In China, the most commonly used dietary supplements were calcium
(58.5%) and zinc (40.4%), while in Australia, the most frequently used types were
multi-vitamins/minerals (46.2%) and fish oil (42.3%). In Australia, “not working”, “never
breastfeed”, “higher education level of the mother” and “older age of the child” were
associated with dietary supplement use in children. In China, being unwell and “having
higher household income” were significantly related to dietary supplement usage. Because
of the unknown effects of many supplements on growth and development and the potential
for adverse drug interactions, parents should exercise caution when giving their infants or
 176

young children dietary supplements. Wherever possible it is preferable to achieve nutrient

intakes from a varied diet rather than from supplements.

Keywords: dietary supplements; Chinese; calcium; zinc; migrants; child; nutrition

1. Introduction

Infant nutrition is important for short term and long term health. A balanced variety of nutritious
foods are emphasized in the guidelines of the Australian and Chinese governments and other
professional organizations as the best source of nutrition for healthy children [1–3]. However,
the Chinese diet has been reported to be low in calcium, riboflavin, Vitamin A, and zinc [4,5].
A national survey in 2004 found that the average calcium intake among the city and suburban
populations was 430 mg per day, well below the recommended intake [6]. The iron intake appears to
be adequate in amount, but its bioavailability is very low and consequently the prevalence of
iron deficiency and iron deficiency anemia was 43.7% and 7.8%, respectively, among children aged
1–3 years in 2001 [4,7].
The consumption of fortified foods and/or supplements can help some children meet their
nutritional needs [8]. Examples of recommended use of supplements include the American Academy
of Pediatrics’ recommendation for oral Vitamin D supplementation for exclusively breastfed infants
and, under certain conditions, for specific older infants and toddlers [9]. However, other countries,
such as Australia, have different climatic conditions and do not recommend universal use of Vitamin D,
and the excessive intakes of single nutrients may have the potential for adverse effects [10–12].
Dietary supplements enriched with vitamins, minerals, and other substances are increasingly
consumed worldwide. The North America and the Asia Pacific regions are the dominant markets for
vitamins and dietary supplements [13]. The prevalence of supplement use varies in different ethnic
groups for a diversity of dietary and cultural reasons and economic conditions. Most published studies
on the use of supplements in children have been conducted in the US and only a small number of
studies have been conducted in Asian countries. It is reported that approximately 49% of the U.S.
population take dietary supplements and the prevalence of supplement use was 35% among children
aged 1–13 years [14,15]. In South Korea, approximately 34% of Korean children and adolescents were
taking dietary supplements in a national survey in 2007–2009 [16]. A survey of urban Japanese found
that 20.4% of children and adolescents between 3 and 17 years were using supplements, or had used
them in the past year [17]. A cross-sectional survey carried out in Zhejiang Province, PR China in
1999 reported a prevalence of 18% of vitamin supplements and 31% of other nutritional supplements
in adolescents [18]. A recent study from Taiwan reported that 34.9% of the infants had been given
a dietary supplement before six months [19].
Australians have a high prevalence of taking dietary supplements. A representative population
survey conducted in 2004 in South Australia reported the use of vitamin supplements by 39.2%
respondents and mineral supplements by 13.6% of the population [20]. No recent data is available on
the use of supplements by infants or young children in Australia.
 177

Until recently, there have been no reported studies of dietary supplementation among Chinese
young children in mainland China or overseas. The aim of this study was to document the prevalence
of use of dietary supplements in these populations. A survey was carried out of Chinese mothers living
in Perth, Australia and Chengdu and Wuhan, PR China.

2. Methods

This data was collected from October 2010 to October 2011 in Perth, Western Australia and from
September to December 2011 in Chengdu and Wuhan, China. Participants in Perth were mothers who
have at least one pre-school child under 5 years old. They were recruited from the Perth Chinese
community, including Chinese schools and community organizations. Mothers interested in taking part
in this study received an information sheet containing project details and were asked to sign the
consent form. A total of 248 questionnaires were distributed in Perth and 237 mothers agreed to
participate (response rate of 95.6%) and 230 mothers completed the dietary supplementation section of
the questionnaire. The response rate to the dietary questionnaire was 95.6%. Participants in China were
recruited from four kindergartens in four districts of Wuhan and 14 kindergartens in seven districts of
Chengdu. Both private and public kindergartens were included. A total of 2400 questionnaires were
distributed to mothers by kindergarten teachers and 2079 were returned, a response rate of 86.6%.
The dietary supplementation questionnaire was completed by 1464 mothers in China with a response
rate to the dietary questionnaire of 70.4%. The study was approved by the Curtin University Human
Research Ethics Committee (approval number: HR 96/2010) and the local education authorities in China.
Demographic and dietary supplement use was collected using a validated and reliable questionnaire
previously used in Chinese population studies [21]. Pre-coded questions were used to classify income
into three groups using categories based on local annual household income surveys [22,23]. A Dietary
Supplement Questionnaire is used to collect information on the participants’ use of medicine, vitamins,
minerals, herbals, and other supplements during the past two weeks. Detailed information about type,
consumption frequency, and amount taken was collected for each reported dietary supplement use.
Child’s health status was collected using a translated version of the Australian National Health Survey
Questionnaire [24].
Body mass index (BMI) was defined as weight (kg)/height (m)2. The 2012 revised international
child cut-offs developed by the International Obesity Task Force (IOTF) were used to classify
thinness, overweightness and obesity in children in this study [25]. They are based on BMI data from
six countries, corresponding to the body mass index (BMI) cut-offs at 18 years, which are BMI 25
(overweight), 30 (obesity) and 18.5 (underweight) [25].
All statistical analyses were performed using the IBM Statistical Package for Social Sciences
(SPSS) Version 20.0. Independent samples’ t-test was used to compare means between groups.
Mann-Whitney U test was applied to compare the average age of children from two countries.
Chi-VTXDUH Ȥ2) test was used to compare basic characteristics of mothers and children in Australia and
China. A multiple binary logistic regression model was used to evaluate the association between
mother and child’s characteristics and the use of dietary supplements. A backward elimination
procedure was applied to obtain final models. p values <0.05 were considered statistically significant.
 178

3. Results

A total of 230 Chinese mothers living in Perth, Australia and 1156 mothers living in Chengdu,
Sichuan Province and 308 mothers living in Wuhan, Hubei Province, PR China completed the
supplements questionnaire. The distribution analysis shows there were no differences between mothers
who completed the supplements questionnaire and mothers who did not in age, education attainment,
marital status, working status, family income status, breastfeeding initiation and duration. There was
also no difference in education attainment, marital status, family income status, breastfeeding initiation
and duration, between mothers in Chengdu and Wuhan. The only statistically significant difference
between mothers in Wuhan and Chengdu was the average age (31.0 years in Chengdu and 30.8 years
in Wuhan, p < 0.001). Because the difference is so small in Wuhan and Chengdu mothers, their data
were pooled for further analysis.
The average age of Chinese mothers in Australia was older than mothers in China (33.8 ± 4.9 years
compared to 31.0 ± 4.1 years, p < 0.001). The mothers in Australia also had higher education levels.
The median age of the “index child” in the China study population (median age = 3.7 years, the
interquartile range = 1.1 years) was older than in Perth (median age = 1.6 years, the interquartile
range = 1.9 years, p < 0.001). More Perth Chinese children were underweight (22.7%) and fewer
overweight and obese (8.0%) than children in China (11.6% underweight and 17.0% overweight and
obese, p = 0.003) (Table 1).

Table 1. Characteristics of Chinese mothers and their children completing dietary

questionnaires in Australia and China.
Australia (n * = 230) China (n * = 1464) 2-sided
Characteristic
n (%) n (%) p-value
Mothers Age (years) <0.001
” 68 (30.1) 604 (53.3)
>30 158 (69.9) 530 (46.7)
Marital status 0.116
Married 229 (99.6) 1151 (98.1)
Divorced/single/widow 1 (0.4) 22 (1.9)
Educational attainment <0.001
High school diploma/TAFE
57 (24.8) 661 (57.1)
certificate/diploma or less
University degree or higher 173 (75.2) 496 (42.9)
Working status <0.001
Working 105 (45.7) 968 (83.1)
Not employed 125 (54.3) 197 (16.9)
Household income 0.086
Low income 108 (49.5) 572 (55.9)
High income 110 (50.5) 451 (44.1)
Mother’s birth place
Mainland China 187 (81.3)
Other Asian countries 43 (18.7)
 179

Table 1. Cont.
Duration in Australia (years)
<5 126 (53.1)
5–10 73 (32.3)
>10 33 (14.6)
Age of the child (years) <0.001
0–1 62 (27.0) 15 (1.0)
1–2 81 (35.2) 24 (1.7)
2–3 38 (16.5) 268 (18.6)
3–4 30 (13.0) 638 (442)
4–5 19 (8.3) 497 (34.5)
Gender of the child 0.737
Boy 122 (53.0) 782 (54.2)
Girl 108 (47.0) 660 (45.8)
Weight status of the child (aged 2–4 years old) 0.003
Underweight 20 (22.7) 147 (11.6)
Normal 61 (69.3) 905 (71.4)
Overweight/obesity 7 (8.0) 216 (17.0)
Ever breastfed
Yes 217 (94.3) 1210 (85.2)
No 13 (5.7) 211 (14.8)
Regular exercises 0.002
Yes 117 (60.0) 861 (70.9)
No 78 (40.0) 353 (29.1)
Illness during the past 4 weeks <0.001
Yes 85 (37.3) 790 (55.4)
No 143 (62.7) 636 (44.6)
Dietary supplement use by child 0.002
Yes 52 (22.6) 475 (32.4)
No 178 (77.4) 989 (67.6)
* The missing values vary for each variable in both countries.

A total of 22.6% of the Chinese children living in Perth were taking dietary supplements, including
multi-vitamins/minerals, fish oil, protein, probiotics, colostrum, calcium, zinc and Vitamin AD (or cod
liver oil) and Chinese herbs (Table 1). In Chengdu and Wuhan, China, 32.4% of young children were
having dietary supplements, including multivitamins/minerals, calcium, zinc, iron, magnesium, fish
oil, probiotics, Vitamin A and/or Vitamin D, Chinese herbs or other botanicals (Table 1). Compared to
Chinese Australians, Chinese parents living in China were more likely to give their children dietary
VXSSOHPHQWV Ȥ2 = 9.2, df = 1, p = 0.002). However, in children aged over 12 months, there is no
statistical difference in the prevalence of dietary supplements between Australia (28.6%) and China
(32.7%, p = 0.284). A higher percentage of children over three years old living in Australia were
taking dietary supplements (40.8%) compared to children living in China (31.5%).
In China, the use of calcium supplements was very common among supplement users (58.5%).
About half of the Chinese children taking calcium supplements were also taking Vitamin D
(n = 140, including the use of multi-vitamins). In Australia, only four children were given specific
 180

calcium supplements. The most common forms of supplemental calcium used in Chinese children up
to five years old are gluconate (51.8%) and carbonate (37.5%). The dosage of calcium supplements
ranged from 54–725 mg/day (Table 2). The average intake for calcium carbonate users (307.4 mg/day)
is higher than gluconate calcium users (81 mg/day). When calculating the average intake, the intakes
from multi-vitamins/minerals were also summed up if they were reported.

Table 2. Main dietary supplements used by Chinese children in Australia and China.
Australia China

%
Supplement supplement Average intake * Intake range % supplement Average intake * Intake range
n n
users (n = (mg/day) (mg/day) users (n = 475) (mg/day) (mg/day)
52)
Calcium 4 9.6 105 (n = 5) 75–200 278 58.5 131.4 (n = 264) 54–725
Zinc 1 1.9 3.1 (n = 12) 1–7.5 192 40.4 4.4 (n = 166) 1.62–8.6
Multi-vitamins/
24 46.2 NA NA 94 19.8 NA NA
minerals
Vitamin A 4 7.7 1026 ** (n = 7) 582.5–1617 ** 83 17.5 1695 ** (n = 71) 600–2800 **
Vitamin D 4 7.7 177 ** (n = 5) 85–200 ** 91 19.2 568 ** (n = 75) 80–780 **
Vitamin C 10 19.2 62.1 (n = 12) 20–125 33 6.9 61.4 (n = 23) 30–200
Fish oil 22 42.3 859.6 (n = 13) 300–1000 4 0.8 NA NA
Probiotics 2 3.9 NA NA 22 4.6 NA NA
Herbs 4 7.7 NA NA 51 10.7 NA NA
* When calculated the average intake, the intakes from multi-vitamins/minerals were also summed if they were reported; ** IU/day,
IU: international unit; NA: not available.

The prevalence of the use of zinc supplementation was also high in China. Nearly half of
supplements users were using zinc supplements (40.4%). Almost all the zinc supplements were in
the form of gluconate (93.2%) and the average intake of zinc was 4.4 mg/day (n = 166, range from
2.15–8.6 mg) (Table 2).
In Australia, the types most frequently used by supplement users were multi-vitamins/minerals
(46.2%) and fish oil (42.3%). The average intake of fish oil was 859.6 mg per day (n = 13) with the
range from 300 to 1000 mg per day (Table 2).
Chinese herbal supplements were used by children in both countries, especially in China, where
10.7% of supplements users were taking herb supplements (Table 2). Some herbal supplements were
used for “better appetite” and some were believed to be beneficial to the immune system or to bring an
improvement of health or well-being. In this study, traditional Chinese medicines including cinnabar,
as arum, isatis root, kaladana, mangnolia officinalis, scaphium scaphigerum, coltsfoot, coptis chinensis
and realgar were included as ingredients in children’s dietary supplements or medicines for
(preventing) coughs or colds. Excluding dietary supplements, 7.6% of children in China reported
taking medicine during the last two weeks and 82.9% (n = 92, 6.3% of all the samples) were taking
herbal products for medical reasons, such as cough or upper respiratory tract infection. In China, a
total of 16.1% of supplements users (8.6% of the total sample) were using herbal products as dietary
supplements or medicine and 7.7% of supplement users (2.2% of the total sample) in Australia
reported taking herbal products.
 181

In 4–5 year old children in Australia, nearly half (47.4%) were taking at least one dietary
supplement 7DEOH ,Q$XVWUDOLDROGHUFKLOGUHQ Ȥ2 = 19.22, df = 4, p = 0.001), children who were
QHYHU EUHDVWIHG Ȥ2 = 4.32, df = 1, p < 0.05) and children who did regular physical exercises in
pre-VFKRRORUDWKRPH Ȥ2 = 10.88, df = 2, p = 0.001) were more likely to take dietary supplements than
other children. Mothers who had migrated from other Asian regions (including Hong Kong) were more
OLNHO\WRJLYHWKHLUFKLOGUHQGLHWDU\VXSSOHPHQWVWKDQPRWKHUVIURPPDLQODQG&KLQD Ȥ2 = 4.47, df = 1,
p < 0.05) (Table 3).

Table 3. Dietary supplement use by children: demographic variables.

Children used dietary supplements Children used dietary
in Australia supplements in China
n (%) 2-sided p-value n (%) 2-sided p-value
Mothers Age (years) 0.201 0.551
<30 12 (17.6) 206 (34.4)
• 40 (25.5) 171 (32.8)
Education of the mother 0.283 0.942
<University 10 (17.5) 217 (33.1)
•8QLYHUVLW\ 42 (24.4) 163 (33.3)
Working status 0.690 0.645
Working 25 (23.8) 321 (33.2)
Not employed 27 (21.6) 62 (31.5)
Household income 0.692 <0.001
Low 26 (24.1) 161 (28.1)
High 24 (21.8) 186 (41.2)
Mother’s birth place 0.034
Mainland China 37 (19.9)
Other Asian regions 15 (34.9)
Duration in Australia 0.160
” 21 (17.6)
5–10 21 (28.8)
>10 9 (27.3)
Gender of the child 0.868 0.201
Male 28 (23.1) 267 (34.2)
Female 24 (22.2) 204 (31.1)
Child’s age (year) 0.001 0.427
<1 year 4 (6.6) 6 (40.0)
1–2 20 (24.7) 8 (33.3)
2–3 8 (21.1) 100 (37.5)
3–4 11 (36.7) 203 (31.8)
4–5 9 (47.4) 155 (31.2)
Infant feeding 0.038 0.272
Ever breastfed 46 (21.3) 402 (33.2)
Never breastfed 6 (46.2) 62 (29.4)
Child’s BMI 0.406 0.596
Underweight 4 (20.0) 45 (31.7)
Normal 20 (33.9) 310 (34.3)
Overweight or obesity 3 (42.9) 64 (31.1)
 182

Table 3. Cont.
Regular exercises 0.001 0.042
Yes 37 (31.6) 306 (35.5)
No 9 (11.5) 104 (29.5)
Illness during the past 4 weeks 0.208 0.008
Yes 27 (26.7) 354 (34.8)
No 25 (19.7) 113 (27.6)

In China, the prevalence of dietary supplement use was higher in children who had been sick during
WKH SDVW IRXU ZHHNV Ȥ2 = 6.97, df = 1, p <   DQG FKLOGUHQ ZKR KDG UHJXODU H[HUFLVH Ȥ2 = 4.13,
df = 1, p < 0.05) than in their counterparts. Higher household income was significantly related to
WKHXVHRIFKLOGVXSSOHPHQWV Ȥ2 = 19.29, df = 1, p < 0.001) (Table 3).
Mother’s age, education level, working status, household income, the child’s age, BMI, regular
exercise, and “illness during the last month” were entered into a binary logistic regression model using
backward elimination. After controlling for those potential confounding variables, the results of the
binary logistic regression analysis showed that Chinese Australian mothers with higher education
levels (OR = 2.51, 95% CI 1.19–5.27), older children (OR = 3.11, 95% CI 1.42–6.83), who were not
employed (OR = 3.83, 95% CI 1.09–13.44), and never breastfed their children (OR = 6.75, 95%
CI 1.29–35.31) were more likely to give their child dietary supplements. In China, higher household
income (OR = 1.53, 95% CI 1.13–2.08) and “having illness during the past month” (OR = 1.44, 95%
CI 1.05–1.97) were associated with dietary supplement use in children (Table 4).

Table 4. Odds ratios of factors for dietary supplement use in Chinese children in Australia
and China.
China Australia
OR 95% CI OR 95% CI
Household income NS
Low 1
High 1.53 1.13–2.08
Education of the mother NS
<University 1
•8QLYHUVLW\ 2.51 1.19–5.27
Working status NS
Working 1
Not employed 3.83 1.09–13.4
Child age (year) 3.11 1.42–6.83
Breastfed NS
Yes 1
Never 6.75 1.29–35.31
Illness during the past 4 weeks
Yes 1
No 1.44 1.05–1.97
NS: not significant.
 183

4. Discussion

With the increasing prevalence of chronic disease throughout the world and increasing interest in
complementary medicine, dietary supplements have become more widely used in children [26,27]. Many
varieties of dietary supplements are now marketed in China and Australia, including single-ingredient products
and various combinations of vitamins, minerals, botanicals, and other constituents. Their use in healthy
children is addressed towards non-clinical deficiencies, the achievement of optimal status of nutrition
and health [17,28].
This study investigated the prevalence of dietary supplement use in Chinese children in mainland
China and in Australia. This is the first report, to our knowledge, on the use of dietary supplements in
young Chinese children under the age of five years. In this study, one fifth of Chinese children in Perth
and one third of children in Chengdu and Wuhan were taking at least one nutritional supplement with
no gender differences. The prevalence of dietary supplement use in young children in China was
similar to that of the US (35%) and South Korea (34%), but higher than Japan (20.4%) [14–17].
However, the comparison populations in these reports generally were older children. The lower
prevalence of dietary supplement use in Chinese immigrant children in Australia than children in
China may be due to the age difference of the subjects. In Australia, most children were under three
years old. It was found that older children in Australia were more likely to take dietary supplements.
The types of supplements commonly used in Chinese children in China and in Australia were quite
different. In China, calcium and zinc supplements were most commonly used, with many children
taking both. Although 58.5% of supplements users were taking calcium supplementation, the average
intake was still only 131 mg per day, which is about 20% of the Adequate Intake set for calcium for
Chinese children in this age group [3]. It is less than half of the calcium that can be provided from one
serve (250 mL) of milk; besides milk can provide other nutrients like protein to support child growth
[29]. A meta-analysis on randomized, controlled trials reported little effectiveness of calcium
supplementation on bone density in healthy children, either in childhood or later life [30]. The calcium
dose was of 300–1200 mg per day in 19 studies included in the meta-analysis, which was much higher
than the average calcium intake from supplements in this study (131 mg in China and 105 mg in
Australia). Since the level of intake of calcium supplements in China is very low, it is not possible
that intake from supplements would be likely to have a positive effect on bone mineral density in
Chinese children.
It has been reported in many studies that Chinese children have a low daily zinc intake [31,32].
This may be due to the higher reference value used to define the adequate daily intake in those studies.
The Recommended Nutrient Intakes (RNIs) for zinc for 1–7 years old Chinese children range from
9–13.5 mg/day, which are higher than in Japan (5–7 mg/day), USA (3–5 mg/day) and in Australia
(3–4 mg/day) [3,29,33,34]. The recommended intake for Chinese children is even higher than the upper
level of zinc intakes for those age groups in Australia and New Zealand, which is 7 mg/day for 1–3 years
and 12 mg/day for 4–8 years [33]. The 2002 China National Nutrition and Health Survey found
that the median intake of zinc in 2–8 year old Chinese children ranged from 5.1 to 7.1 mg/day
(the interquartile range: 3.9–9.3 mg/day), which already met the RNIs for this age group in Japan,
USA and Australia [35]. However, the adequacy of zinc intake depends not only on the amount, but
also its bioavailability. People consuming a diet that provides marginal zinc intake may not absorb
 184

an adequate amount of zinc if they are also consuming foods high in phytate together with high
calcium [36]. The average population phytate intake of people in China (1186 mg/day) is relatively
high compared to their western counterparts, but Chinese diets are low in calcium, reducing the
possibility of low zinc availability [35]. The elevation of calcium intake by increasing consumption of
milk is not affected by the inhibitory effect of phytate because animal sources of protein appear to
promote zinc release from its phytate complex and also provides intrinsic zinc in a highly available
form [36]. For young children from this study, their calcium intakes from calcium supplements were
low and because of their young age, they still rely on milk products as their main calcium source.
Considering the amount of zinc intake from their diet, they may not need to take zinc supplements.
Together with the amount of zinc from supplements (ranging from 2.15 to 8.6 mg/day), it is a concern
that some children might have reached the upper level of intakes for their age. Adverse events
associated with chronic intake of supplemental zinc may include suppression of immune response,
decrease in high density lipoprotein cholesterol and reduced copper status [33].
In Australia, the most popular supplements were multi-vitamins/minerals, which is consistent with
previous studies in children and adolescents [16,17,37]. Fish oil supplements (42.3%) were almost as
popular as multi-vitamins and minerals (46.2%). Another large sample size, cross-sectional study
(n = 266,848) undertaken in New South Wales, Australia also reported a high prevalence of fish oil
supplement use in healthy elderly people [38]. Few children were on calcium supplements in Australia.
This might be due to higher consumption of milk and milk products in Australia than in China.
Commercial advertising may also influence the choice of dietary supplement.
The types of dietary supplements used by young children living in China were distinct from those in
Australia. This may be due to the different regulations about supplements that apply to both countries.
Promotion and advertising of supplements is different in both countries. There are many reports in
the literature that suggest that unnecessary or reckless use of dietary supplements can lead to problems.
More studies related to the clinical effectiveness and/or safety of dietary supplements in infants and
children are required, especially over the longer term. In the case of Chinese children in China,
the intakes of calcium and zinc deserve special considerations in relation to development of dietary
supplement regulations. Further studies on fish oil supplements in young children in Australia are also
required to add to our knowledge of its health effects.
Herbal products are widely used both in China and by Chinese Australians. Most herbal traditional
products not only have plant-derived materials or preparations, but may also include animal products
(including scorpions, cicadas and centipedes) and mineral compounds (including cinnabar and
realgar) [39]. There is a public perception that these products are inherently safe, however, the
therapeutic basis of many ingredients is still not clear. Some traditional ingredients can be toxic when
used for inappropriate indications, or prepared inappropriately, or used in excessive dosages, or for a
prolonged duration [39–42]. It is known that some Chinese medicines can have nephrotoxicity or
hepatotoxicity effects and some cause increased risk of bleeding [43–46]. There is a need to increase
the awareness of toxic effects of some herbal products in the public and health care professions.
There are several limitations that need to be considered when interpreting the results of the present
study. First, our results may not be representative of all Chinese children in China or in Australia
because of the location of the sample and the number of subjects. Secondly, the age distribution of
the subjects from two countries in this study was slightly different and this may have a small influence
 185

on the results. Nevertheless, we believe our present study to be important for understanding
the present status of supplement use in Chinese pre-school children, and in monitoring future trends
of supplement use.

5. Conclusions

It is important for pre-school children to meet their energy and nutrient needs for growth and
development. Consuming a healthy diet is important to achieve adequate nutrient intakes. Dietary
supplements only need to be considered when individuals are not able to obtain an adequate nutrient
status from their diet alone. A large number of healthy Chinese children both in China and in Australia
use dietary supplements, which for most may not be medically indicated.
Calcium and zinc are the two most popular dietary supplements in young children in China, while
multi-vitamin/minerals and fish oil are the most frequently used in Australian Chinese children.
However, the supplements used in China contain relatively low amounts of calcium and the same
amount could easily be obtained from milk and other dairy products. For some other nutrients such as
zinc, the potential over-nutrient of taking supplements should be of concern. There is also a need to
increase the awareness of toxic effects of some herbal products in the public and health care professions.
There are many reports in the literature that suggest that unnecessary use of dietary supplements can
lead to problems. Parents should exercise caution when giving their infants or young children dietary
supplements and be aware of the potential toxicity of inappropriate use or excessive dosages. Before
providing dietary supplements, parents should seek advice from appropriate health professionals. For
all infants and young children, wherever possible, it is preferable to achieve nutrient intakes from a
varied diet rather than from supplements.

Conflicts of Interest

The authors declare no conflict of interest.

Acknowledgments

The authors gratefully acknowledge the assistance of the mothers who agreed to be interviewed
and the support of kindergarten teachers in Chengdu and Wuhan. This study was funded by Curtin
University. No competing financial interests exist.

References

1. American Dietetic Association. Position of the American Dietetic Association: Fortification and
nutritional supplements. J. Am. Diet. Assoc. 2005, 105, 1300–1311.
2. National Health and Medical Research Council. Dietary Guidelines for Children and Adolescents
in Australia Incorporating the Infant Feeding Guidelines for Health Workers; NHMRC, Ed.;
NHMRC: Canberra, Australia, 2003.
3. Chinese Nutrition Society. Dietary Guidelines of China; Chinese Nutrition Society, Ed.; The
Tibetan People’s Publishing House: Lhasa, China, 2010.
 186

4. Ge, K.Y. The Dietary and Nutritional Status of Chinese Population. In 1992 National Nutrition
Survey; People’s Medical Publishing House: Beijing, China, 1999.
5. Ge, K. The transition of Chinese Dietary Guidelines and Food Guide Pagoda. Asia Pac. J. Clin.
Nutr. 2011, 20, 439–446.
6. Zhai, F.; Wang, H.; Du, S.; He, Y.; Wang, Z.; Ge, K.; Popkin, B.M. Prospective study on nutrition
transition in China. Nutr. Rev. 2009, 67 (Suppl. 1), 56–61.
7. Zhu, Y.P.; Liao, Q.K. Prevalence of iron deficiency in children aged 7 months to 7 years in China.
Zhonghua Er Ke Za Zhi 2004, 42, 886–891.
8. Marra, M.V.; Boyar, A.P. Position of the American Dietetic Association: Nutrient
supplementation. J. Am. Diet Assoc. 2009, 109, 2073–2085.
9. Wagner, C.L.; Greer, F.R. Prevention of rickets and Vitamin D deficiency in infants, children, and
adolescents. Pediatrics 2008, 122, 1142–1152.
10. National Health and Medical Research Council; Ministry of Health. Nutrient Reference Values for
Australia and New Zealand: Including Recommended Dietary Intakes; Ministry of Health:
Canberra, Australia, 2006.
11. Institute of Medicine. Dietary Reference Intakes For Calcium and Vitamin D 2011; Institute of
Medicine, Ed.; National Academies Press: Washington, DC, USA, 2011.
12. Morris, J.S.; Crane, S.B. Selenium toxicity from a misformulated dietary supplement, adverse
health effects, and the temporal response in the nail biologic monitor. Nutrients 2013, 5, 1024–1057.
13. Crossley, A. Review of the global dietary supplement market. Nutr. Bus. J. 2007, 12, 18–21.
14. Bailey, R.L. Dietary supplement use in the united states, 2003–2006. J. Nutr. 2011, 141, 261–266.
15. Bailey, R.L. Total folate and folic acid intakes from foods and dietary supplements of us children
aged 1–13 y. Am. J. Clin. Nutr. 2010, 92, 353–358.
16. Yoon, J.Y.; Park, H.A.; Kang, J.H.; Kim, K.W.; Hur, Y.I.; Park, J.J.; Lee, R.; Lee, H.H.
Prevalence of dietary supplement use in korean children and adolescents: Insights from Korea
National Health and Nutrition Examination Survey 2007–2009. J. Korean Med. Sci. 2012, 27,
512–517.
17. Mori, N. Prevalence and characterization of supplement use among healthy children and
adolescents in an urban Japanese city. Health 2011, 3, 135–140.
18. Hesketh, T.; Ding, Q.J.; Tomkins, A.M. Disparities in economic development in Eastern China:
Impact on nutritional status of adolescents. Public Health Nutr. 2002, 5, 313–318.
19. Chuang, C.H.; Yang, S.H.; Chang, P.J.; Chen, P.C.; Chen, Y.C. Dietary supplement intake by
6-month-old Taiwanese infants. J. Pediatr. Gastroenterol. Nutr. 2012, 54, 71–76.
20. MacLennan, A. The continuing use of complementary and alternative medicine in South
Australia: Costs and beliefs in 2004. Med. J. Aust. 2006, 184, 27–31.
21. Li, L.; Zhang, M.; Binns, C.W. Chinese mothers’ knowledge and attitudes about breastfeeding in
Perth, Western Australia. Breastfeed. Rev. 2003, 11, 13–19.
22. Sichuan Bureau of Statistics. The Average Salary Income of Workers and Staff Members in
Sichuan in 2011; Sichuan Bureau of Statistics: Chengdu, China, 2012.
23. Australian Bureau of Statistics Measures of Australia’s progress; ABS: Canberra, Australia,
2010.
 187

24. Australian Bureau of Statistics National Health Survey 2001: Data Reference Package,
2001. Available online: http://www.abs.gov.au/AUSSTATS/abs@.nsf/allprimarymainfeatures/
2EAF21C0E6710A54CA257107007819F8?opendocument (accessed on 20 October 2013)
25. Cole, T.J.; Lobstein, T. Extended international (iotf) body mass index cut-offs for thinness,
overweight and obesity. Pediatr. Obes. 2012, 7, 284–294.
26. Mullie, P. Cultural, socioeconomic and nutritional determinants of functional food consumption
patterns. Eur. J. Clin. Nutr. 2009, 63, 1290–1296.
27. Crawford, N. A cross-sectional survey of complementary and alternative medicine use by children
and adolescents attending the university hospital of wales. BMC Complement. Altern. Med. 2006,
6, 16.
28. Woodside, J.V. Micronutrients: Dietary intake v. Supplement use. Proc. Nutr. Soc. 2005, 64,
543–553.
29. U.S. Department of Agriculture; U.S. Department of Health and Human Services. Dietary
Guidelines for Americans, 2010, 7th ed.; U.S. Government Printing Office: Washington, DC,
USA, 2010.
30. Winzenberg, T.; Shaw, K.; Fryer, J.; Jones, G. Effects of calcium supplementation on bone
density in healthy children: Meta-analysis of randomised controlled trials. BMJ 2006, 333, 775.
31. Chen, X.C.; Yin, T.A.; He, J.S.; Ma, Q.Y.; Han, Z.M.; Li, L.X. Low levels of zinc in hair and
blood, pica, anorexia, and poor growth in Chinese preschool children. Am. J. Clin. Nutr. 1985, 42,
694–700.
32. Penland, J.G.; Sandstead, H.H.; Alcock, N.W.; Dayal, H.H.; Chen, X.C.; Li, J.S.; Zhao, F.;
Yang, J.J. A preliminary report: Effects of zinc and micronutrient repletion on growth and
neuropsychological function of urban Chinese children. J. Am. Coll. Nutr. 1997, 16, 268–272.
33. Department of Health and Ageing; National Health and Medical Research Council. Nutrient
Reference Values for Australia and New Zealand: Including Recommended Dietary Intakes;
Commonwealth of Australia: Canberra, Australia, 2005.
34. National Institute of Health and Nutrition. Dietary Reference Intakes for Japanese 2010. The
Summary Report from the Scientific Committee of “Dietary Reference Intakes For Japanese”;
NIHN: Tokyo, Japan, 2010.
35. Ma, G.; Li, Y.; Jin, Y.; Du, S.; Kok, F.J.; Yang, X. Assessment of intake inadequacy and food
sources of zinc of people in china. Public Health Nutr. 2007, 10, 848–854.
36. World Health Organization. Trace Elements in Human Nutrition and Health; WHO: Geneva,
Switzerland, 1996.
37. Dwyer, J.; Dwyer, R.L.; Nahin, G.T.; Rogers, P.M.; Barnes, P.M.; Jacques, C.T.; Sempos, R.
Bailey, Prevalence and predictors of children’s dietary supplement use: The 2007 National Health
Interview Survey. Am. J. Clin. Nutr. 2013, 97, 1331–1337.
38. Adams, J. Q-3 fatty acid supplement use in the 45 and up study cohort. BMJ Open 2013,
3, e002292.
39. Phua, D.H.; Zosel, A.; Heard, K. Dietary supplements and herbal medicine toxicities-when to
anticipate them and how to manage them. Int. J. Emerg. Med. 2009, 2, 69–76.
40. Levy, S. The risk-benefit profile of commonly used herbal therapies. Ann. Internal Med. 2002,
137, 1008.
 188

41. Haddad, P.S.; Azar, G.A.; Groom, S.; Boivin, M. Natural health products, modulation of immune
function and prevention of chronic diseases. Evid. Based Complement. Altern. Med. 2005, 2,
513–520.
42. Cupp, M.J. Herbal remedies: Adverse effects and drug interactions. Am. Fam. Physicians 1999,
59, 1239–1245.
43. Ulbricht, C.; Basch, E.; Weissner, W.; Hackman, D. An evidence-based systematic review of herb
and supplement interactions by the natural standard research collaboration. Expert Opin. Drug
Saf. 2006, 5, 719–728.
44. Liu, J. Mercury in traditional medicines: Is cinnabar toxicologically similar to common
mercurials? Exp. Biol. Med. 2008, 233, 810–817.
45. Hintelmann, H. High precision isotope ratio measurements of mercury isotopes in cinnabar ores
using multi-collector inductively coupled plasma mass spectrometry. Analyst 2003, 128, 635–639.
46. Nortier, J.L.; Vanherweghem, J.L. Renal interstitial fibrosis and urothelial carcinoma associated
with the use of a chinese herb (Aristolochia fangchi). Toxicology 2002, 181–182, 577–580.

© 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).
 189

Reprinted from Nutrients. Cite as: Nasreddine, L.; Naja, F.; Akl, C.; Chamieh, M.C.; Karam, S.; Sibai, A.;
Hwalla, N. Dietary, Lifestyle and Socio-Economic Correlates of Overweight, Obesity and Central
Adiposity in Lebanese Children and Adolescents. Nutrients 2014, 6, 1038-1062.

Dietary, Lifestyle and Socio-Economic Correlates of

Overweight, Obesity and Central Adiposity in Lebanese
Children and Adolescents
Lara Nasreddine 1,†,‡, Farah Naja 1,†,‡, Christelle Akl l, Marie Claire Chamieh 1, Sabine Karam 1,
Abla-Mehio Sibai 2,‡,* and Nahla Hwalla 1,‡,*
1
Department of Nutrition and Food Sciences, Faculty of Agricultural and Food Sciences,
American University of Beirut, P.O. Box 11-0236, Riad El Solh, Beirut 1107-2020, Lebanon;
E-Mails: ln10@aub.edu.lb (L.N.); fn14@aub.edu.lb (F.N.); cristell@gmail.com (C.A.);
mc31@aub.edu.lb (M.C.C.); sek04@aub.edu.lb (S.K.)
2
Department of Epidemiology and Population Health, Faculty of Health Sciences,
American University of Beirut, P.O. Box 11-0236, Riad El Solh, Beirut 1107-2020, Lebanon
†
These authors contributed equally to this work.
‡
These authors are members of the Public Health and Nutrition (PHAN) Research Group and of the
Vascular Medicine Program (VMP) at the American University of Beirut, Beirut, Lebanon.

* Authors to whom correspondence should be addressed; E-Mails: ansibai@aub.edu.lb (A.-M.S.);

nahla@aub.edu.lb (N.H.); Tel.: 961-1-340-460 (A.-M.S.); +961-1-343-002 (N.H.);
Fax: +961-1-744-470 (A.-M.S.); +961-1-744-460 (N.H.).

Received: 24 December 2013; in revised form: 11 February 2014 / Accepted: 17 February 2014 /
Published: 10 March 2014

Abstract: The Eastern Mediterranean region is characterized by one of the highest

burdens of paediatric obesity worldwide. This study aims at examining dietary, lifestyle,
and socio-economic correlates of overweight, obesity, and abdominal adiposity amongst
children and adolescents in Lebanon, a country of the Eastern Mediterranean basin.
A nationally representative cross-sectional survey was conducted on 6–19-year-old
subjects (n = 868). Socio-demographic, lifestyle, dietary, and anthropometric data (weight,
height, waist circumference) were collected. Overweight and obesity were defined based
on BMI z-scores. Elevated waist circumference (WC) and elevated waist to height ratio
(WHtR) were used as indices of abdominal obesity. Of the study sample, 34.8% were
overweight, 13.2% were obese, 14.0% had elevated WC, and 21.3% had elevated WHtR.
 190

Multivariate logistic regression analyses showed that male gender, maternal employment,
residence in the capital Beirut, sedentarity, and higher consumption of fast food and sugar
sweetened beverages were associated with increased risk of obesity, overweight, and
abdominal adiposity, while regular breakfast consumption, higher intakes of milk/dairies
and added fats/oils were amongst the factors associated with decreased risk. The study’s
findings call for culture-specific intervention strategies for the promotion of physical
activity, healthy lifestyle, and dietary practices amongst Lebanese children and adolescents.

Keywords: paediatric; obesity; abdominal adiposity; prevalence; correlates; diet; Lebanon;

Eastern Mediterranean region

1. Introduction

The Eastern Mediterranean region is characterized by one of the highest burdens of overweight and
obesity worldwide [1]. Of more concern is the high level of childhood obesity in countries of the region,
with approximately 10% of school-aged children being obese, an estimate that is projected to follow an
escalating secular trend [2]. Paediatric obesity is associated with both immediate and longer-term risks to
health [3]. Among the immediate risks are metabolic abnormalities including increased blood cholesterol,
triglycerides and glucose levels, insulin resistance, metabolic syndrome, and hypertension [3–5].
Childhood obesity is also a strong risk factor for adult obesity and its consequences including type 2
diabetes, cardiovascular diseases (CVDs), and certain types of cancer, in addition to psychological
disturbances, such as low self-esteem and depression [6,7].
Obesity-related comorbidities were found to be more closely associated with abdominal adiposity
and visceral fat depots than with the amount of total body fat [8]. Consequently, the use of body
fat distribution indices has been increasingly recommended, and particularly the use of waist
circumference (WC) and waist to height ratio (WHtR). These simple and non-invasive indices were
shown to correlate with visceral fat in children and to predict risk for obesity-related comorbidities
beyond that predicted by Body Mass Index (BMI) alone [8–14]. Being a relatively age-independent
measure, the use of WHtR for assessing central fatness in children has been recommended in
paediatric primary care practice, as well as epidemiological studies [14–16]. In a cohort of almost
1500 Caucasian children aged 5 to 15 years, both WC and WtHR were able to identify children with
the highest metabolic and cardiovascular risks among those who were overweight [13]. An extensive
review by Huxley et al. (2010) concluded that measures of abdominal obesity including WC and
WHtR, may be better than BMI in predicting CVD risk, although combining BMI with these measures
may improve their discriminatory capability [17].
The high disease burden of childhood obesity highlights the need for rigorous investigations of its
determinants, context-specific patterns and associated factors. Most of the studies investigating obesity
correlates in youth have been conducted in high-income countries and, as such, findings may not be
applicable to low and middle-income countries. Among the latter, the Middle-East has been largely
under-represented, although the region has one of the highest rates of childhood obesity [2]. The
present study aims at examining the prevalence and correlates of overweight, obesity and abdominal
 191

adiposity in a nationally representative sample of children and adolescents, aged six years and above,
in Lebanon. Gaining greater insight into factors that are associated with paediatric obesity could
catalyze the development of effective interventions and policies aiming at curbing the obesity
epidemic in Lebanon, orient further studies, and assist policy makers in implementing successful,
culture specific childhood obesity prevention strategies in the region.

2. Materials and Methods

2.1. Study Design and Subjects

Data for the present study is drawn from a national cross-sectional survey that was conducted
in 2009, in Lebanon, on subjects aged six years and above. The study sample was based on the
sampling frame provided by the National Survey of Household Living Conditions, which was
conducted by the Ministry of Social Affairs/Central Administration of Statistics in collaboration with
United Nations Development Programme (UNDP) and which covered primary residences across the
Lebanese territory [18]. Sample size calculation for the study was performed based on previously
estimated prevalence rates for the main outcome of interest [19]. As such, a minimum of 751
participants were needed to estimate a prevalence of obesity of 4.8% in children and adolescents [19],
allowing a power of 80% and a margin of error of 1.5% at 95% confidence interval (CI). Recruitment
efforts targeted a sample with an age, sex and district distribution proportionate to that of the Lebanese
population [18].
Lebanon is divided into six administrative regions referred to as “governorates”, which cover the
totality of the country. Except for the governorate of Beirut, which is considered purely urban, the
other governorates are essentially composed of rural regions inter-mixed with urban cities. In this
study, the sample was drawn from randomly selected households, based on stratified cluster sampling:
the strata were the Lebanese governorates, the clusters were selected further at the level of districts,
urban and rural areas, and the housing units constituted the primary sampling units in the different
districts of Lebanon. One adult from each household and one child/adolescent from every other
household were selected from the household roster. Field-work was carried out between May 2008 and
August 2009. The final sample consisted of 3636 subjects, including 939 children and adolescents
aged 6 years and above [20]. Refusal rate at the household level was estimated at 10.7%, with the main
reasons for refusal to participate in the survey being lack of time or disinterest in the study. The design
and conduct of the survey was approved by the Institutional Review Board of the American University
of Beirut, and informed consent from adults/parents and informed assent from children and adolescents
were obtained prior to enrolment in the studies.
Socio-demographic and lifestyle data were collected from study participants using a multi-component
questionnaire that was developed for the purpose of this study. Data collection was performed by
trained nutritionists in the household setting through face to face interviews which lasted for
approximately one hour. Quality control measures including training, pre-testing of the study
instruments, equipment, and data collection procedure and field monitoring of data collection, were
applied. Household and parental data were collected from the adult participant (mother or father) using
a multicomponent questionnaire covering information on demographic, socioeconomic and lifestyle
 192

characteristics, in addition to medical history and health seeking behavior. Data pertinent to the
child/adolescent were collected using a child-specific questionnaire which enquired about sex, age,
medical history, meal pattern, eating habits, dietary intake, physical activity, and sedentary time. For
children aged less than 11 years old, data was obtained by proxy (typically the mother), while the
interview was conducted directly with subjects aged 11 years and above.

2.2. Anthropometric Measurements

Anthropometric measurements were taken using standardized protocols [21] and calibrated
equipment. Height and body weight were measured according to standard procedures, using a portable
stadiometer (Holtain, Crymych, UK) and a Secacalibrated electronic weighing scale (Hamburg,
Germany), respectively. Subjects were weighed to the nearest 0.1 kg in light indoor clothing and with
bare feet or stockings. Height was measured without shoes and recorded to the nearest 0.5 cm.
A calibrated plastic measuring tape was used to measure waist circumference at the level of the
umbilicus to the nearest 0.1 cm, with the subject standing and after normal expiration. Anthropometric
measurements were taken and recorded by trained nutritionists who were working in teams of two, the
examiner and the recorder. All measurements were taken twice and the average of the 2 values
was adopted.

2.3. Definitions of Overweight and Obesity

Body mass index (BMI) was calculated as the ratio of weight (kilograms) to the square of height
(meters). Overweight and obesity were defined based on sex and age specific +1 and +2 BMI
z-scores, respectively, according to the WHO new growth standards [22]. The WHO AnthroPlus
software (WHO, Geneva, Switzerland) was used to calculate BMI z-score for each specific age and
sex. To allow for comparisons with studies conducted in other countries, prevalence rates of
overweight and obesity were also determined using the International Obesity Taskforce (IOTF) [23]
and the US Centers for Disease Control and Prevention (CDC) 2000 criteria [24].
Elevated WC was defined based on the International Diabetes Federation (IDF) criteria [25], which
recommend the use of:
- Adult cut-off values for subjects aged 16–19 years (WC > 94 cm for males and >80 cm for females).
- Cut-off value of WC• 90th percentile for sex and age (or adult cut-offs if lower) for subjects
aged 6 to 15 years old. As national WC percentiles are lacking in Lebanon, the WC percentiles for
children and adolescents as developed by Fernandez et al. (2004) were used [26].
The WHtR index for abdominal obesity was calculated by dividing WC by height, both measured in
centimetres [13]. The suggested cut-off point of• ZDV XVHG WR LGHQWLI\ FKLOGUHQ ZLWK HOHYDWHG
WHtR [13,14].

2.4. Dietary Intake and Physical Activity Assessment

Dietary intake was assessed using the multiple pass 24-h recall approach. Interviewers followed the
5 steps of the USDA multiple pass 24-h recall, which included (1) the quick list; (2) the forgotten
foods list; (3) time and occasion at which foods were consumed; (4) the detail cycle; and (5) the final
 193

probe review [27]. To assist subjects in assessing the portion/amount of food consumed, quantification
tools, such as household measures and graduated food models, were used. 24-h recall data were converted
to energy and nutrient intake using the Nutritionist IV software through a hand-coding procedure
(N-squared Computing Nutritionist IV. Silverton, OR: N-squared Computing; 1995). The Nutritionist
IV food database was expanded by adding analyses of traditional Lebanese foods and recipes.
Information on the weekly frequency of physical activity outside the school setting was assessed by
means of a questionnaire that was developed for this study. Examples of activities proposed by the
questionnaire included moderate intensity activities such as playground activities, brisk walking,
dancing, bicycle riding, as well as higher intensity activities, such as ball games, jumping rope, active
games involving running and chasing, and swimming. Based on the weekly frequency, individuals
were classified into three levels of physical activity: Low (Never); Moderate (1–2 times/week) and
High (>2 times/week).

2.5. Statistical Analysis

Descriptive statistics were performed and expressed as means and standard error (SE) for
continuous variables (dietary variables) or as number of subjects and percentages for nominal variables
(demographic, socio-economic, physical activity, meal pattern, and lifestyle variables). Crowding
index was calculated as the total number of co-residents per household divided by the total number of
rooms, excluding the kitchen and bathrooms. Prevalence of overweight (including obesity), obesity,
elevated WC and elevated WHtR, expressed as percentage with 95% confidence interval (CI), were
computed by gender and age groups (6–11-year-old children and 12–19-year-old adolescents).
Independent t-test and chi-squared test were used to evaluate the differences between continuous and
categorical variables, respectively.
Multivariate logistic regression analysis was carried out to examine the association of overweight,
obesity, elevated WC and elevated WHtR as the dependent variables with baseline socio-demographic,
lifestyle, and dietary characteristics as covariates. The associations between dependent and
independent variables were analyzed according to two age groups: Children (6–11 years) and
adolescents (12–19 years). All statistical calculations were carried out using the Statistical Analysis
Package for Social Sciences, version 18.0 (SPSS Inc., Chicago, IL, USA). Statistical significance was
defined as p-value <0.05.

3. Results

3.1. Study Sample

For the purpose of this study, subjects for whom dietary data were missing or incomplete were
excluded (out of 939 subjects, 71 were excluded). Accordingly, the final study sample consisted
of 868 subjects (439 boys and 429 girls), with a mean age of 13.06 years (±3.91) and a median age
of 12.85 years. Of the study participants, 42.6% were 6–11 years old and 57.4% were 12–19 years old.
The male to female ratio was of 1.02 with 50.6% boys and 49.4% females.
The proportion of parents who had attained high school level education and above was of 38.6% for
fathers and 46.3% for mothers with no significant differences between age groups (Table 1).
 194

A significantly higher proportion of working mothers was reported amongst 12–19-year-old

adolescents (29.6%) compared to children (19.4%). Parental obesity (mother or father) was reported
amongst 31% of the study population with no significant differences between age groups. The majority
of subjects (81.7%) had a crowding index •SHUVRQVURRP 7DEOH 
The proportions of subjects reporting a daily consumption of breakfast was significantly higher
amongst children compared to adolescents (86.4% vs. 69.5% in adolescents) while a significantly
higher proportion of adolescents reported eating outside home more than once per week (58.4% in
adolescents compared to 44.4% in children) (Table 1). Similarly, sedentary time was significantly
higher amongst 12–19-year-old adolescents compared to children (10.09 ± 2.94 vs. 8.72 ± 2.77 h/day)
while the proportion of subjects reporting high physical activity was significantly higher in children
compared to adolescents (63.9% vs. 33.5%).
Mean weight (35.01 ± 12.55 in children vs. 60.77 ± 15.61 kg in adolescents), mean height
(135.63 ± 12.26 vs.164.17 ± 10.07 cm), mean BMI (18.53 ± 3.99 vs. 22.32 ± 4.37 kg/m2) and mean
WC (63.77 ± 10.75 vs. 74.93 ± 10.97 cm) were all significantly higher in 12–19-year-old adolescents
compared to children aged 6–11 years (Table 1).

Table 1. Socio-demographic, lifestyle and anthropometric characteristics of the study sample

by age group, Lebanon (n = 868).
Age Group (years) Total (1)
Variables p-Value (2)
6–11 (n = 370) 12–19 (n = 498) (n = 868)
Socio-Demographic characteristics n (%)
Gender
Male 191 (51.6) 248 (49.8) 439 (50.6)
0.595
Female 179 (48.4) 250 (50.2) 429 (49.4)
Governorates
Capital (Beirut) 26 (7.0) 36 (7.2) 62 (7.1)
0.909
Other governorates 344 (93.0) 462 (92.8) 806 (92.9)
Father’s Education
Primary or less 105 (28.8) 156 (31.7) 261 (30.5)
Intermediate 112 (30.8) 153 (31.1) 265 (31.0) 0.571
High school and above 147 (40.4) 183 (37.2) 330 (38.6)
Mother’s Education
Primary or less 76 (23.0) 112 (26.3) 188 (24.9)
Intermediate 97 (29.4) 121 (28.4) 218 (28.8) 0.587
High school and above 157 (47.6) 193 (45.3) 350 (46.3)
Mother’s working status
Not working 291 (80.6) 350 (70.4) 641 (74.7)
0.001
Working 70 (19.4) 147 (29.6) 217 (25.3)
Parental Obesity (3)
No 218 (71.5) 171 (66.0) 389 (69.0)
0.163
Yes 87 (28.5) 88 (34.0) 175 (31.0)
Crowding Index
<1 person/room 66 (18.0) 92 (18.5) 158 (18.3)
0.858
•SHUVRQURRP 300 (82.0) 405 (81.5) 705 (81.7)
 195

Table 1. Cont.
Lifestyle characteristics n (%)
Breakfast consumption
(per week)
Never 11 (3.0) 36 (7.2) 47 (5.4)
Sometimes 39 (10.6) 116 (23.3) 155 (17.9) <0.001
Daily 319 (86.4) 346 (69.5) 665 (76.7)
Frequency of eating
outside home (per week)
”WLPH 205 (55.6) 207 (41.6) 412 (47.5)
<0.001
>1 time 164 (44.4) 291 (58.4) 455 (52.5)
Physical Activity (4)
Low 80 (21.7) 193 (41.2) 273 (32.6)
Moderate 53 (14.4) 119 (25.4) 172 (20.5) <0.001
High 235 (63.9) 157 (33.5) 392 (46.8)
Sedentary time (h/day)
8.72 ± 2.77 10.09 ± 2.94 9.51 ± 2.95 <0.001
Mean ± SD
Anthropometric characteristics
Weight (kg)
Mean ± SD 35.01 ± 12.55 60.77 ± 15.61 49.81 ± 19.21
10th percentile 22.15 42.67 25.88
<0.001
50th percentile 32.45 59.00 48.75
90th percentile 50.81 80.81 74.52
Height (cm)
Mean ± SD 135.63 ± 12.26 164.17 ± 10.07 152.02 ± 17.93
10th percentile 119.75 152.00 126.00
<0.001
50th percentile 136.00 164.00 154.00
90th percentile 152.00 177.50 174.05
2
BMI (kg/m )
Mean ± SD 18.53 ± 3.99 22.32 ± 4.37 20.71 ± 4.61
10th percentile 14.65 17.47 15.60 <0.001
50th percentile 17.58 21.58 20.08
90th percentile 23.56 28.17 26.80
WC (cm)
Mean ± SD 63.77 ± 10.75 74.93 ± 10.97 70.18 ± 12.19
10th percentile 52.75 62.73 56.00
<0.001
50th percentile 61.50 73.20 69.00
90th percentile 78.50 91.00 86.05
WHtR
Mean ± SD 0.47 ± 0.06 0.46 ± 0.06 0.46 ± 0.06
10th percentile 0.41 0.39 0.40
0.002
50th percentile 0.46 0.44 0.45
90th percentile 0.55 0.54 0.55
(1) (2)
Lack of corresponding sum of frequencies with total sample size is due to missing data; Differences between age groups were
(3)
examined using t-test and chi-square test for continuous and categorical variables, respectively; Total number of parents with
(4)
anthropometric data was equal to 564; The three categories of physical activity (Low, Moderate, High) refer to the frequency of
physical activity outside the school setting (Never; 1–2 times/week; >2 times/week, respectively).
 196

3.2. Prevalence of Overweight, Obesity and Abdominal Obesity

Taking both genders, 40.2% of 6–11-year-olds and 30.8% of 12–19-year-olds were found to be
overweight (BMI z score >+1), while 17.1% and 10.3% were found to be obese (BMI z score >+2),
respectively (Table 2). Gender-based differences were noted amongst 12–19-year-olds, with the
prevalence of overweight (37.9% in boys vs. 23.7% in girls) and obesity (16.1% in boys vs. 4.4% in
girls) being significantly higher in boys compared to girls. Similar gender-based differentials were
noted in the total sample of 6–19-year-old subjects.
Based on WC as an indicator of central fatness, abdominal obesity was observed in 13.8% of
6–11-year-olds and 14.1% of 12–19-year-olds, with no significant differences between genders
(Table 2). Elevated WHtR was observed in 22% and 20.9% of children and adolescents, respectively,
with gender-based differences being observed amongst 12–19-year-old subjects (26.2% in boys vs.
15.6% in girls; p < 0.05). Similar gender-based differentials were noted in the prevalence of elevated
WHtR in the total sample of 6–19-year-old subjects (Table 2).
As shown in Figure 1, the prevalence of overweight, obesity, elevated WHtR and elevated WC
amongst boys reached the highest rates at the age of 10–13 years (47%, 23%, 32% and 18%,
respectively), while declining afterwards (Figure 1). Amongst girls, the prevalence of overweight and
obesity was the highest at 6–9 years (38% and 14%, respectively) and followed a consistent declining
trajectory with age. The prevalence of elevated WC reached its highest in girls aged between 14 and
17 years (16%) and declined afterwards (Figure 1).

Table 2. Prevalence of overweight, obesity and abdominal adiposity amongst Lebanese

children and adolescents (n = 868) by gender and by age group.
Age Groups (years) Total
Variables 6–11 12–19 6–19
n % (95% CI) n % (95% CI) n % (95% CI)
Male
Overweight (1) 81 42.4 (36–50) 94 37.9 a (32–44) 175 39.9 a (35–44)
(1) b
Obesity 39 20.4 (15–27) 40 16.1 (12–21) 79 18.0 b (15–22)
Elevated WC (2) 30 15.7 (11–22) 32 12.9 (9–18) 62 14.1 (11–18)
Elevated WHtR (3) 46 24.1 (19–31) 65 26.2 c (21–32) 111 25.3 c (21–30)
Female
Overweight (1) 67 37.9 (31–45) 59 23.7 a (19–29) 126 29.6 a (25–34)
(1) b b
Obesity 24 13.6 (9–19) 11 4.4 (2–8) 35 8.2 (6–11)
Elevated WC (2) 21 11.8 (8–17) 38 15.2 (11–20) 59 13.8 (11–17)
Elevated WHtR (3) 35 19.7 (14–26) 39 15.6 c (12–21) 74 17.3 c (14–21)
Both Genders
Overweight (1) 148 40.2 (35–45) 153 30.8 (27–35) 301 34.8 (32–38)
(1)
Obesity 63 17.1 (14–21) 51 10.3 (8–13) 114 13.2 (11–16)
Elevated WC (2) 51 13.8 (10–17) 70 14.1 (11–17) 121 14.0 (11–16)
Elevated WHtR (3) 81 22.0 (18–27) 104 20.9 (17–24) 185 21.3 (18–24)
(1)
WC: Waist Circumference; WHtR: Waist to Height ratio; Overweight and obesity defined based on sex and age specific +1 and +2
BMI z-scores, respectively [22]; (2) For subjects aged 6–15 years, abdominal obesity: WC > 90th percentile [26] or adult cut-off value if
lower [25]; For subjects aged 16–19 years, abdominal obesity: WC > 94 cm for males and >80 cm for females [25]; (3) Elevated WHtR
a,b,c
defined as WHtR > 0.5 [13]; Within each age group, values with the same superscripts are significantly different by gender at
p < 0.05 (Using Chi-square test).
 197

Figure 1. Prevalence of overweight, obesity, elevated WC and elevated WHtR amongst Lebanese children and adolescents (n = 868) by age
and gender (* Significant difference by gender p < 0.05).
 198

3.3. Dietary Intake

As shown in Table 3, average energy intake (2255.85 vs. 1736.48 kcal/day) and percent
contribution of fast food (17.27% vs. 11.35%) and legumes and nuts (3.28% vs. 2.05%) to daily energy
intake were significantly higher among 12–19-year-old adolescents compared to 6–11-year-old
children. On the other hand, the percent contribution of milk and dairies (8.90% vs. 6.47%) and breads
and cereals (36.92% vs. 32.67%) were significantly higher in 6–11-year-old children compared to
adolescents. No significant differences in macronutrient intake were observed between age groups.

Table 3. Energy, macronutrient and food group intake amongst Lebanese children and
adolescents according to age group (n = 868).
Age Group (years) Total
Dietary Variables 6–11 12–19 6–19
(n = 370) (n = 498) (n = 868)
Energy (kcal ± SE ) 1736.48 ± 36.12 a 2255.85 ± 51.60 a 2033.70 ± 34.47
Mean % Daily Energy Intake ± SE
Carbohydrates 52.01 ± 0.51 51.05 ± 0.48 51.46 ± 0.35
Protein 13.11 ± 0.18 13.53 ± 0.23 13.35 ± 0.15
Fat 35.86 ± 0.47 36.24 ± 0.43 36.058 ± 0.32
b b
Breads and Cereals 36.92 ±0.89 32.67 ±0.77 34.49 ±0.58
Milk and Dairies 8.90 ±0.52 c 6.47 ±0.37 c 7.51 ±0.31
Meat and Equivalent 10.22 ±0.57 10.15 ±0.57 10.18 ±0.40
d d
Legumes and Nuts 2.05 ±0.34 3.28 ±0.38 2.76 ±0.27
Fruits and Vegetables 5.35 ±0.33 5.42 ±0.33 5.39 ±0.24
Added Fats and Oils 7.58 ±0.47 8.42 ±0.42 8.06 ±0.31
e e
Fast Food 11.35 ±0.69 17.27 ±0.92 14.74 ±0.61
Sugar and Sweets 10.81 ±0.63 9.65 ±0.60 10.15 ±0.44
Sugar Sweetened Beverages 6.52 ±0.45 6.45 ±0.32 6.48 ±0.26
a,b,c,d,e
Values with the same superscripts are significantly different by age group at p < 0.05 (Using t-test).

3.4. Factors Associated with Overweigh, Obesity and Abdominal Obesity

Amongst 6–11-year-old children, results of the multivariate regression analysis showed that, as
compared to subjects living in the capital Beirut, those residing in other governorates had significantly
lower odds of being overweight (OR = 0.32; 95% CI: 0.1–0.98), obese (OR = 0.21; 95% CI:
0.06–0.71), and of having elevated WC (OR = 0.16; 95% CI: 0.04–0.56) (Table 4). Higher maternal
education was associated with significantly higher odds of overweight (OR = 2.45; 95% CI:
1.13–5.31), while higher paternal education was associated with lower odds of obesity in this age
group (OR = 0.32; 95% CI: 0.11–0.91). Maternal employment was shown to be associated with
significantly higher odds of obesity (OR = 2.6; 95% CI: 1.18–5.70) and elevated WHtR (OR = 2.27;
95% CI: 1.19–4.33). In contrast, daily breakfast consumption was associated with significantly lower
odds of overweight (OR = 0.2; 95% CI: 0.05–0.84) and obesity (OR = 0.07; 95% CI: 0.01–0.30);
higher intakes of milk & dairies were associated with lower odds of elevated WC (OR = 0.35; 95% CI:
0.13–0.91), and higher intakes of added fats/oils were associated with lower odds of obesity
 199

(OR = 0.30; 95% CI: 0.12–0.73), elevated WHtR (OR = 0.42; 95% CI: 0.20–0.88), and elevated WC
(OR = 0.32; 95% CI: 0.12–0.88). High consumption of fast food was associated with a threefold
increase in the risk of overweight in this age group (OR = 3.24; 95% CI: 1.21–8.69) (Table 4).

Table 4. Associations of socio-demographic, lifestyle and dietary factors with overweight,

obesity, elevated waist to height ratio (WHtR), and elevated waist circumference (WC) in
Lebanese 6–11-year-old children (n = 868).
6–11 Years
(n = 370)
Variables
Overweight 1 Obesity 1 Elevated WHtR Elevated WC
Odds Ratio [95% CI]
Socio-Demographic Factors
Age (years) 1.17 [1.00–1.36] 1.17 [0.91–1.37] 1.14 [0.97–1.34] 1.22 [0.96–1.54]
Sex
Female 1.00 1.00
1.00 1.92 [0.94–3.90] 1.00 1.47 [0.84–2.56]
Male 1.17 [0.69–1.97] 2.01 [0.90–4.48]
Place of Residence
Beirut (Capital) 1.00 1.00 1.00 1.00
Other Governorates 0.32 [0.10–0.98] 0.21 [0.06–0.71] 0.39 [0.14–1.05] 0.16 [0.04–0.56]
Father’s Education 2
Low 1.00 1.00 1.00 1.00
Medium 1.14 [0.57–2.29] 0.83 [0.35–1.96] 0.98 [0.46–2.07] 1.12 [0.43–2.92]
High 0.52 [0.24–1.11] 0.32 [0.11–0.91] 0.73 [0.33–1.62] 0.46 [0.14–1.48]
Mother’s Education 2
Low 1.00 1.00 1.00 1.00
Medium 1.98 [0.90–4.33] 1.59 [0.60–4.18] 0.66 [0.30–1.47] 0.68 [0.23–1.95]
High 2.45 [1.13–5.31] 1.36 [0.52–3.59] 0.68 [0.31–1.45] 0.80 [0.29–2.20]
Mother’s Working Status
Not Working 1.00 1.00 1.00 1.00
Working 1.06 [0.55–2.03] 2.60 [1.18–5.70] 2.27 [1.19–4.33] 1.47 [0.60–3.63]
Crowding Index
<1 person/room 1.00 1.00 1.00 1.00
•SHUVRQURRP 1.19 [0.59–2.40] 1.04 [0.40–2.70] 0.64 [0.31–1.31] 0.53 [0.20–1.40]
Parental Obesity
No 1.00 1.00 1.00 1.00
Yes 1.72 [0.97–3.04] 2.67 [1.34–5.31] 2.10 [1.18–3.72] 2.46 [1.15–5.23]
Lifestyle and Dietary Factors
Physical Activity 3
Low 1.00 1.00 1.00 1.00
Medium 1.60 [0.91–2.81 1.78 [0.90–3.52] 0.99 [0.52–1.86] 1.43 [0.71–2.89]
High 0.86 [0.56–1.33] 0.75 [0.42–1.33] 0.70 [0.43–1.14 0.68 [0.38–1.22]
Sedentary Time (h/day) 1.02 [0.93–1.12] 1.10 [0.97–1.25] 1.05 [0.93–1.19] 1.08 [0.94–1.25]
Daily Breakfast
Consumption
No 1.00 1.00 1.00 1.00
Yes 0.20 [0.05–0.84] 0.07 [0.01–0.30] 0.32 [0.07–1.32] 0.25 [0.05–1.25]
 200

Table 4. Cont.
Frequency of Eating Out
”WLPHZHHN 1.00 1.00 1.00 1.00
>1 time/week 1.01 [0.62–1.65] 0.96 [0.48–1.93] 1.18 [0.65–2.14] 1.75 [0.82–3.70]
2
Total Daily Energy Intake (kcal)
Low 1.00 1.00 1.00 1.00
Medium 1.21 [0.70–2.10] 1.19 [0.54–2.65] 1.43 [0.73–2.78] 1.13 [0.48–2.66]
High 1.44 [0.71–2.90] 1.10 [0.52–3.74] 1.13 [0.47–2.71] 1.54 [0.54–4.43]
Bread and Cereals 4
Low 1.00 1.00 1.00 1.00
Medium 1.19 [0.67–2.12] 0.76 [0.32–1.79] 0.72 [0.35–1.46] 0.51 [0.20–1.32]
High 1.01 [0.53–1.86] 1.36 [0.59–3.14] 0.97 [0.46–2.01] 1.13 [0.46–2.73]
Milk and Dairies 4
Low 1.00 1.00 1.00 1.00
Medium 0.92 [0.50–1.69] 0.99 [0.43–2.30] 1.07 [0.52–2.22] 0.53 [0.21–1.32]
High 1.16 [0.62–2.16] 0.64 [0.26–1.56] 0.66 [0.30–1.43] 0.35 [0.13–0.91]
Meat and Equivalent 4
Low 1.00 1.00 1.00 1.00
Medium 0.86 [0.48–1.53] 1.19 [0.54–2.61] 1.24 [0.63–2.44] 1.38 [0.56–3.18]
High 0.71 [0.34–1.47] 0.64 [0.21–1.94] 1.32 [0.53–3.29] 1.26 [0.38–4.17]
Legumes and Nuts 4
Low 1.00 1.00 1.00 1.00
Medium 1.08 [0.66–1.76] 0.65 [0.34–1.23] 0.74 [0.41–1.33] 0.86 [0.42–1.74]
High 0.90 [0.51–1.58] 0.68 [0.33–1.42] 0.93 [0.49–1.78] 1.17 [0.55–2.51]
4
Fruits and Vegetables
Low 1.00 1.00 1.00 1.00
Medium 1.55 [0.86–2.79] 2.24 [0.99–5.08] 1.08 [0.49–2.37] 2.01 [0.83–4.86]
High 0.91 [0.50–1.67] 1.11 [0.45–2.69] 0.99 [0.48–2.03] 1.11 [0.42–2.90]
4
Added Fats and Oils
Low 1.00 1.00 1.00 1.00
Medium 0.73 [0.40–1.32] 0.36 [0.17–0.86] 0.42 [0.20–0.88] 0.79 [0.32–1.93]
High 0.64 [0.34–1.18] 0.30 [0.12–0.73] 0.53 [0.26–1.08] 0.32 [0.12–0.88]
4
Fast Foods
Low 1.00 1.00 1.00 1.00
Medium 2.14 [0.91–5.02] 2.41 [0.70–8.23] 1.62 [0.51–5.10] 2.86 [0.64–12.64]
High 3.24 [1.21–8.69] 1.50 [0.43–5.23] 2.46 [0.79–7.67] 1.50 [0.37–6.01]
4
Sugar and Sweets
Low 1.00 1.00 1.00 1.00
Medium 1.04 [0.57–1.90] 0.86 [0.37–1.92] 1.03 [0.50–2.14] 1.01 [0.39–2.59]
High 1.12 [0.61–2.06] 0.76 [0.33–1.75] 1.17 [0.57–2.43 1.59 [0.65–3.88]
4
Sugar Sweetened Beverages
Low 1.00 1.00 1.00 1.00
Medium 1.13 [0.67–1.90] 0.66 [0.34–1.28] 0.53 [0.22–1.28] 0.57 [0.22–1.17]
High 1.32 [0.79–2.22] 0.59 [0.29–1.17] 0.81 [0.45–1.45] 0.54 [0.20–1.13]
1
Overweight and obesity defined based on sex and age specific +1 and +2 BMI z-scores, respectively [22]; 2 Low, medium and high
education levels refer to primary or less, intermediate or high school and above, respectively; 3 The three categories of physical activity
(Low, Moderate, High) refer to the frequency of physical activity outside the school setting (Never, 1–2 times/week; >2 times/week);
4
Food groups’ intake based on percent contribution to daily energy intake. Low, medium, and high refer to first, second, and third
tertiles, respectively.
 201

Amongst 12–19-year-old adolescents, male gender was associated with significantly higher odds
of obesity (OR = 5.18; 95% CI: 1.76–15.28) and elevated WHtR (OR = 1.82; 95% CI: 1.12–2.97)
(Table 5). Similar to findings amongst 6–11-year-old children, significantly lower odds of overweight
were observed amongst adolescents residing in other governorates as compared to those living in the
capital Beirut (OR = 0.40; 95% CI: 0.19–0.83). Parental obesity was associated with approximately
a 3-fold increase in the odds of overweight (OR = 3.01; 95% CI: 1.61–5.63), obesity (OR = 2.93;
95% CI: 1.09–7.86), and elevated WHtR (OR = 2.87; 95% CI: 1.55–5.30). A borderline significant
association between high physical activity and lower odds of overweight (OR = 0.62; 95% CI:
0.33–1.05) and central fatness as assessed by WHtR (OR = 0.53; 95% CI: 0.26–1.09) was also
observed. Sedentary time was significantly positively associated with all adiposity indicators amongst
12–19-year-old adolescents with higher odds of overweight (OR = 1.12; 95% CI: 1.03–1.21), obesity
(OR = 1.2; 95% CI: 1.06–1.35), elevated WHtR (OR = 1.27; 95% CI: 1.13–1.43), and elevated WC
(OR = 1.10; 95% CI: 1.01–1.22) being observed. Higher intakes of milk and dairies were associated
with significantly lower odds of overweight (OR = 0.56; 95% CI: 0.32–0.98) in this age group.
In contrast, higher intakes of sugar sweetened beverages were associated with significantly higher odds
of overweight (OR = 2.49; 95% CI: 1.5–4.12) and elevated WHtR (OR = 1.77; 95% CI: 1.02–3.07).
A borderline significant association was found between the consumption of fruits and vegetables and
lower odds of elevated WC (OR = 0.46; 95% CI: 0.21–1.00) (Table 5).

Table 5. Associations of socio-demographic, lifestyle and dietary factors with overweight,

obesity, elevated WHtR and elevated WC in Lebanese 12–19-year-old adolescents.
12–19 Years
(n = 498)
Variables
Overweight 1 Obesity 1 Elevated WHtR Elevated WC
Odds ratio [95%CI]
Socio-Demographic Factors
Age (years) 0.99 [0.83–1.18] 0.91 [0.69–1.20] 0.99 [089–1.10] 0.95 [0.77–1.17]
Sex
Female 1.00 1.00 1.00 1.00
Male 1.68 [0.92–3.07] 5.18 [1.76–15.28] 1.82 [1.12–2.97] 0.75 [0.35–1.58]
Place of Residence
Beirut (Capital) 1.00 1.00 1.00 1.00
Other Governorates 0.40 [0.19–0.83] 0.54 [0.14–2.06] 1.18 [0.47–2.96] 1.00 [0.29–3.44]
Father’s Education 2
Low 1.00 1.00 1.00 1.00
Medium 1.53 [0.67–3.50] 1.26 [0.30–5.22] 1.70 [0.87–3.31] 1.14 [0.38–3.42]
High 1.47 [0.63–3.42] 2.20 [0.56–8.64] 1.83 [0.90–3.72] 2.13 [0.75–6.08]
Mother’s Education 2
Low 1.00 1.00 1.00 1.00
Medium 0.84 [0.33–2.09] 1.33 [0.26–6.86] 0.67 [0.33–1.36] 0.48 [0.15–1.55]
High 1.30 [0.54–3.12] 2.12 [0.50–8.85] 0.73 [0.36–1.45] 0.73 [0.25–2.13]
 202

Table 5. Cont.
Mother’s Working
Status
Not Working 1.00 1.00 1.00 1.00
Working 1.50 [0.69–3.27] 0.47 [0.10–2.07] 1.10 [0.60–2.02] 1.11 [0.43–2.86]
Crowding Index
<1 person/room 1.00 1.00 1.00 1.00
•SHUVRQURRP 1.74 [0.69–4.40] 0.96 [0.25–3.65] 0.74 [0.39–1.41] 0.73 [0.28–1.95]
Parental Obesity
No 1.00 1.00 1.00 1.00
Yes 3.01 [1.61–5.63] 2.93 [1.09–7.86] 2.87 [1.55–5.30] 1.74 [0.81–3.72]
Lifestyle and Dietary Factors
3
Physical Activity
Low 1.00 1.00 1.00 1.00
Medium 0.79 [0.44–1.39] 0.89 [0.37–2.12] 0.75 [0.39–1.43] 0.77 [0.35–1.69]
High 0.62 [0.33–1.05] 0.43 [0.13–1.33] 0.53 [0.26–1.09] 0.53 [0.21–1.31]
Sedentary time
1.12 [1.03–1.21] 1.20 [1.06–1.35] 1.27 [1.13–1.43] 1.10 [1.01–1.22]
(h/day)
Daily Breakfast
Consumption
No 1.00 1.00 1.00 1.00
Yes 0.62 [0.28–1.40] 1.11 [0.28–4.41] 0.58 [0.24–1.37] 0.72 [0.27–1.88]
Frequency of
Eating Out
”WLPHZHHN 1.00 1.00 1.00 1.00
>1 time/week 0.74 [0.47–1.17] 0.87 [0.42–1.79] 0.87 [0.52–1.45] 1.09 [0.60–1.97]
Total Daily Energy
Intake (Kcal) 2
Low 1.00 1.00 1.00 1.00
Medium 1.05 [0.56–1.96] 0.62 [0.23–1.69] 0.67 [0.33–1.36] 0.62 [0.28–1.39]
High 0.80 [0.42–1.54] 0.85 [0.31–2.31] 0.75 [0.37–1.52] 0.91 [0.41–2.04]
4
Bread and Cereals
Low 1.00 1.00 1.00 1.00
Medium 1.18 [0.69–2.02] 1.71 [0.75–3.92] 1.21 [0.66–2.23] 1.21 [0.61–2.39]
High 0.57 [0.32–1.02] 1.07 [0.43–2.65] 0.79 [0.41–1.50] 0.76 [0.36–1.62]
4
Milk and Dairies
Low 1.00 1.00 1.00 1.00
Medium 1.04 [0.61–1.78] 0.79 [0.36–1.75] 1.26 [0.69–2.32] 1.39 [0.70–2.75]
High 0.56 [0.32–0.98] 0.50 [0.21–1.20] 1.17 [0.64–2.15] 1.04 [0.51–2.12]
4
Meat and equivalent
Low 1.00 1.00 1.00 1.00
Medium 0.78 [0.44–1.39] 0.89 [0.35–2.27] 0.94 [0.49–1.79] 1.38 [0.64–2.95]
High 1.12 [0.63–1.99] 1.01 [0.39–2.57] 1.01 [0.52–1.93] 1.45 [0.66–3.14]
 203

Table 5. Cont.
Legumes and Nuts 4
Low 1.00 1.00 1.00 1.00
Medium 0.66 [0.39–1.12] 1.59 [0.76–3.31] 1.34 [0.76–2.46] 1.28 [0.64–2.54]
High 0.97 [0.61–1.53] 0.81 [0.38–1.75] 1.33 [0.77–2.30] 1.25 [0.66–2.36]
4
Fruits and Vegetables
Low 1.00 1.00 1.00 1.00
Medium 0.87 [0.50–1.54] 0.73 [0.30–1.73] 0.62 [0.29–1.33] 0.46 [0.21–1.00]
High 0.69 [0.39–1.21] 0.87 [0.36–2.06] 0.61 [0.30–1.24] 0.60 [0.29–1.21]
4
Added Fats and Oils
Low 1.00 1.00 1.00 1.00
Medium 1.39 [0.79–2.45] 0.67 [0.28–1.57] 0.87 [0.46–1.65] 0.96 [0.45–2.03]
High 1.19 [0.67–2.13] 0.92 [0.39–2.18] 1.31 [0.70–2.45] 1.63 [0.79–3.35]
4
Fast Foods
Low 1.00 1.00 1.00 1.00
Medium 0.92 [0.31–2.73] 0.89 [0.25–1.31] 1.07 [0.44–2.59] 1.64 [0.59–4.57]
High 1.40 [0.53–3.66] 1.34 [0.45–4.03] 1.46 [0.67–3.16] 2.05 [0.83–5.08]
4
Sugar and Sweets
Low 1.00 1.00 1.00 1.00
Medium 0.72 [0.41–1.27] 0.82 [0.35–1.88] 0.82 [0.44–1.53] 0.50 [0.25–1.17]
High 0.82 [0.47–1.42] 0.75 [0.32–1.73] 0.82 [0.44–1.53] 0.72 [0.39–1.31]
Sugar Sweetened
4
Beverages
Low 1.00 1.00 1.00 1.00
Medium 2.49 [1.50–4.12] 1.74 [0.78–3.88] 1.77 [1.02–3.07] 1.62 [0.83–3.13]
High 1.36 [0.80–2.29] 1.69 [0.76–3.77] 0.94 [0.52–1.70] 1.19 [0.60–2.37]
1
Overweight and obesity defined based on sex and age specific +1 and +2 BMI z-scores, respectively [22];
2
Low, medium and high education levels refer to primary or less, intermediate or high school and above, respectively;
3
The three categories of physical activity (Low, Moderate, High) refer to the frequency of physical activity outside the
4
school setting (Never, 1–2 times/week; >2 times/week); Food groups’ intake based on percent contribution to daily
energy intake. Low, medium and high refer to first, second, and third tertiles, respectively.

4. Discussion

Based on a nationally representative survey, this paper reports on the prevalence of overweight,
obesity, and abdominal adiposity in Lebanese children and adolescents and provides evidence linking
specific dietary, lifestyle, and socioeconomic factors to increased risk of adiposity in this population
group. Recognizing that the development of successful obesity prevention strategies should rely on
evidence-based public health approaches, the results of this paper could represent a stepping stone for
the formulation of effective interventions and policies aiming at curbing the epidemic of obesity in
Lebanese youth.
The findings of the present study indicate high prevalence rates of overweight and obesity amongst
Lebanese children and adolescents. Using the WHO 2007 BMI criteria, it was found that more than
third of 6–19-year-old children and adolescents (34.8%) are overweight (BMI z score >+1), with about
one in seven (13.2%) being obese (BMI z score >+2). To allow comparison with findings reported
 204

from selected countries in the region and worldwide, data were re-analyzed according to IOTF and
CDC criteria. Based on the IOTF criteria, current prevalence rates of obesity amongst children and
adolescents in Lebanon (9.6%) are comparable to those reported from Bahrain (11.3%) [28] and Syria
(11.1%) [29], higher than those observed in Qatar (6.3%) [30], while being lower than those reported
from the UAE (13.7%) [31]. Based on the CDC 2000 definition, the prevalence of obesity in Lebanese
youth (12.6%) appears lower than that reported from the US (18.7%) [32], while being considerably
higher than estimates reported from Iran (1.8%) [33] and Saudi-Arabia (5.7%) [34]. When the results
of the present study are compared to those provided by the first national survey conducted in 1997 in
Lebanon [19,20], an approximate two-fold increase in the prevalence of obesity in 6–19-year-old
Lebanese children is noted (7.3% in 1997 vs. 13.2% in 2009, based on WHO 2007 criteria). As such, the
observed annual increase (+6.7%) in the prevalence of child obesity in Lebanon exceeds the estimated
annual increase in the EMRO region (+5.6%), as determined by Wang and Lobstein (2006) [2]. The
prevalence of abdominal adiposity in Lebanese youth has followed a parallel increasing trend between
1997 and 2009, with elevated WC rates increasing from 8.5 to 14% and elevated WHtR rates
increasing from 19.1 to 21.3% amongst 6–19-year-old children. Current prevalence rates of abdominal
obesity as assessed by WHtR (21.3%) were found to exceed those reported from several other
countries including Germany (10.7% in boys and 8% in girls) [35] and Pakistan (16.5%) [16] while
being lower than those reported from Italy (29.5%) [36]. The prevalence rates of elevated WC (14%)
were similar to those reported from Pakistan (13%) [16] and Germany (17.3%) [37], but were lower
than estimates reported from Italy (29%) [36].
Gender differentials in the prevalence of obesity and central fatness were noted in 12–19-year-old
adolescents, with the odds of obesity being five times higher in boys compared to girls. Adolescent
boys were also approximately two times more likely to be abdominally obese compared to girls based
on the WHtR indicator. The higher prevalence of obesity amongst boys in this age group, is in line
with previous reports from other countries in the region such as Syria, Qatar, Saudi-Arabia, and Greece
[29,30,38,39], and with previous studies conducted in Lebanon [19,40]. This may possibly be resulting
from stronger cultural and social pressure on adolescent girls to maintain an acceptable body image in
this age group [19]. Gender differentials may also be explained by differences in dietary patterns and
food choices. In this study, adolescent boys had a significantly higher intake of fast food, sugar
sweetened beverages, and breads and cereals, while having significantly lower intakes of fruits and
vegetables compared to girls (data not shown).
Our finding of a positive significant association between paediatric adiposity and parental obesity
corroborates those reported from other studies and underscores the importance of genetic factors in the
aetiology of body fatness [29,41]. However, strong evidence also suggests that childhood obesity is
linked to socio-economic development, changes in environmental factors, such as living and school
environments, diet, and physical activity patterns [2]. In the present study, specific dietary habits and
food choices were associated with the risk of adiposity in the study sample. In 6–11-year-old children,
and in line with several previous studies [42–44], regular breakfast consumption was associated with a
significantly lower risk of overweight and obesity. Although mechanisms linking breakfast consumption
to lower body weight are unclear, several possible explanations may exist [42]. Skipping breakfast
may lead to excess hunger, rebound overeating [42], and consumption of larger portion sizes [45] and
higher amounts of discretionary calories at subsequent meals [42]. Breakfast consumption may also be
 205

associated with the selection of more healthful food choices [46], more regular eating habits and
increased frequency of eating meals, which is suggested to reduce the efficiency of utilization of
metabolizable energy and promote diet-induced thermogenesis and energy expenditure [42,47].
In agreement with previous reports [48,49], fast food consumption was associated with a threefold
increase in the risk of overweight amongst 6–11-year-old children. Fast food’s poor nutritional quality
[50,51] and higher content of fat and saturated fat [52] underline their potential role as contributors to
childhood adiposity and weight gain. Previous studies have shown that compared with non-consumers,
children who consume fast food were found to have higher total energy, total fat, and saturated fat
intakes [53] and higher obesity risk, while having lower intakes of fiber, milk, fruit, vegetables and
fiber [48,49,53,54]. Contrary to the observed association between fast food and adiposity in the study
sample, the intake of milk and dairy products was found to be associated with lower odds of abdominal
adiposity in this age group (6–11-year-old children). In agreement with our findings, several
observational studies have illustrated inverse associations between dairy intake and adiposity in
children, while suggesting a role for dairy protein in the regulation of body weight [55,56]. Other
studies have found that dietary calcium intake, especially from dairy products, may have a protective
effect against overweight and obesity [57,58]. Based on a retrospective analysis of several studies,
Heaney et al. (2002) proposed that a daily increase of 300 mg of calcium, or approximately 1 dairy
serving, was associated with a yearly reduction of approximately 1 kg of body fat in children [59]. It is
hypothesized that the relationship between calcium and body weight may be mediated by the lower
intracellular calcium levels resulting from high calcium intakes, which reduce lipogenesis while
increasing lipolysis and decreasing adiposity [60]. Surprisingly, the intake of “added fats and oils” was
found to be associated with a protective effect against obesity and abdominal adiposity in 6–11-year-
old children. When looking at the types of fats and oils included in this food group, olive oil was found
to contribute 78% of added fats and oils, on a weight basis. Monounsaturated fats (MUFAs) and olive
oil, which represent one of the distinctive properties of the Mediterranean diet, was suggested to reduce
the risk of obesity in childhood [61]. In a one–year longitudinal study conducted on 13–166-month-old
children, the risk of weight gain was significantly lower in children who consumed olive oil compared
to those who did not [61]. MUFAs may act on the regulation of appetite, on the intestinal absorption of
fat, on the lipolytic activity of the adipocyte and on thermogenesis, among other functions and
therefore may contribute to the regulation of body weight [61–64].
Amongst 12–19-year-old adolescents, and similarly to the findings documented in 6–11-year-old
children, higher intakes of milk and dairy products were associated with lower odds of adiposity.
In addition, a positive association was documented between higher consumption of sugar-sweetened
beverages and a higher risk of overweight and elevated WHtR amongst adolescents. This is in agreement
with findings reported from large cross-sectional studies and several well-powered prospective cohort
studies [65], which document a positive association between greater intakes of sugar sweetened
beverages and obesity in children. A recent meta-analysis of cohort studies found that a higher intake
of sugar-sweetened beverages among children was associated with 55% (95% CI 32%–82%) higher
risk of being overweight or obese compared to lower intakes [66]. The high added sugar content, low
satiety and the resulting incomplete compensation of energy at subsequent meals are likely mechanism
by which sugar-sweetened beverages may lead to weight gain [67].
 206

Through combined effects on energy balance, physical activity and sedentary time were suggested
as two important and distinct modulators of obesity risk in children and adolescents [68]. In the present
study, a borderline significant association was documented between high physical activity and lower
odds of overweight and central fatness in adolescents. Similarly, sedentary time was associated with
significantly higher odds of overweight, obesity and abdominal adiposity (elevated WC and WHtR) in
the same age group. It is suggested that adolescents usually become more interested in screen-time
activities such as computer games or watching TV than their younger peers, and, hence, are more
prone to engage in sedentary behaviors [69]. When compared to the findings of the previous national
survey conducted in 1997 in Lebanon [19], sedentary behavior among Lebanese children and
adolescents (defined as •KVHGHQWDU\WLPHSHUGD\ ZDVIRXQGWRLQFUHDVHIURPLQWR
60.5% in 2009, a finding that may mirror the increased reliance of youth on TV and telecommunication
technology. Similarly, regression analyses showed that the risk of overweight/obesity and abdominal
obesity was higher in children and adolescents living in the capital Beirut as compared to their
counterparts residing in other governorates. Beirut, as a city, is characterized by a complete lack of
safe greens and public spaces, such as gardens, parks, playgrounds and sports fields which may have
direct repercussions on the lifestyle of children and adolescents such as decreased physical activity,
increased screen time and television watching and consequently sedentary behavior [6]. In a European
sample of 766 children, aged 10 to 12 years, engagement in more moderate to vigorous physical
activity and spending less sedentary time were associated with a more favorable weight status in the
study sample [68].
The results of this study document significant associations between certain parental socioeconomic
characteristics and adiposity amongst 6–11-year-old children, but not amongst adolescents. An inverse
association between fathers’ education level and child obesity was documented. This finding is in
disagreement with that reported from several developing countries [29,70] where a positive association
between paediatric obesity and higher parental education was documented. However, our findings are
in agreement with those reported from developed countries [71–73]. A study conducted in Italy among
8- to 9-year-old children showed that the prevalence of paediatric obesity was inversely related to the
educational level of fathers, thus highlighting the role of paternal education in modulating the family’s
lifestyle, economic and cultural resources, all of which may bear ramifications on nutritional and
behavioral choices and therefore obesity risk in childhood [73]. In contrast, and in agreement with
findings reported from various developing countries [29,70], higher maternal education was found to
be associated with significantly higher odds of overweight amongst 6–11-year-old children. This
finding may be a reflection of the association between maternal employment and adiposity in children
as the likelihood of employment of the mother increases as her education level increases. In the
6–11-year-old study sample, children with working mothers were found to carry more than a two-fold
increase in the risk of obesity and abdominal obesity (elevated WHtR) compared to their counterparts.
Maternal employment may in fact be one of the modulators of the family environment, which can have
a direct influence on children’s lifestyles, physical activity, and eating habits [74]. A recent
longitudinal study in the UK showed that children with working mothers were more likely to be
overweight or obese than those of non-working mothers, and children’s likelihood of being overweight
or obese increased with the mother’s working time [72].
 207

The results of this study should be considered in light of the following limitations. The use of
cut-offs that are not population-specific may jeopardize the sensitivity and specificity of the indices
used to assess overweight, obesity and abdominal adiposity. Another limitation of concern is the fact
that children aged above 11 years reported themselves on their dietary intake. Children’s recall of food
intake may be associated with under-reporting (missing foods), over-reporting [75], as well as
incorrect identification of foods due to their lower knowledge of foods and their preparation [76]. It is
also important to note that, in our study, dietary information was based on the collection of one 24-h
recall, which may not be representative of dietary intakes at the individual level. However, despite its
well-known limitations such as reliance on memory and day-to-day variation, the 24-h recall may
provide accurate estimates of energy intake at the population level [76]. In the present study, dietary
information was collected by the multiple pass 24-h recall approach, which was shown to provide
accurate estimates of dietary intake in children [77]. In addition, the recalls were taken by research
nutritionists who went through extensive training prior to data collection in order to minimize
interviewer errors. Similarly, inter-observer measurement error in anthropometric assessment was
minimized by extensive training and follow up to maintain quality of measurement among all research
nutritionists. It is important to note that the physical activity questionnaire that was used in this
study was not validated. However, the questionnaire was reviewed by a panel of experts including
a nutritionist, a physical activity educator and an epidemiologist, and was based on tools used in
similar studies.

5. Conclusions

This study has documented high prevalence rates of overweight, obesity and adnominal adiposity
amongst Lebanese children and adolescents. More importantly, the study’s findings pinpointed
towards specific socioeconomic, dietary, and lifestyle factors that may increase the risk of adiposity in
Lebanese youth. The documented high prevalence of child adiposity raises questions about its
implications for psychosocial development and disease burden in the country, given the association of
paediatric adiposity with metabolic syndrome, insulin resistance, hypertension, glucose intolerance,
and dyslipidaemia [3,4]. With those below 20 years of age, making up close to 50% of the Lebanese
population [78], these estimates do not bode well for the health and well-being of the population.
Childhood obesity is related to growing up in an obesogenic environment, in which changes in
physical activity and diet appear as the main drivers. In countries undergoing the nutrition transition
such as Lebanon, children and adolescents represent the age group that suffers the most from adoption
of western lifestyle characterized by long hours of television viewing, computer games, and heavy
reliance on fast food, all of which are key factors affecting nutritional habits and obesity levels [79].
In the present study, adiposity in children was positively associated with sedentarity, irregular
breakfast consumption, and higher intakes of fast food and sugar-sweetened beverages while the
consumption of milk/dairies and olive oil were associated with a lowered risk. Parental socioeconomic
characteristics, including education level and maternal employment, were documented as risk factors
for adiposity in 6–11-year-old children, but not in adolescents. This highlights the importance of the
home environment in modulating the child’s lifestyle and dietary habits and hence obesity risk early in
life. Taken together, these findings call for community-based intervention programs that involve
 208

multisectoral partnerships and that are responsive to the sociocultural norms of the population. The
prevention of paediatric overweight and obesity requires systems-level approaches and environmental
support across all sectors of society to achieve sustained dietary and physical-activity behavior change
[80]. Based on the results of this study, physical intervention strategies should in particular target
adolescents who were shown to have higher levels of sedentarity and to be less likely to engage in
physical activity compared to their younger peers. Family-focused interventions and behavioral
strategies are needed to instil healthy lifestyle and dietary habits early in life. School-based
interventions should integrate behavioral and environmental approaches that focus on dietary intake
and physical activity using a systems-level approach [80]. Policy and environmental interventions are
recommended as sustainable ways to support healthful lifestyles for children and families and to ensure
that all youth have the opportunity to achieve and maintain a weight that is optimal for health [80].

Acknowledgments

The study was funded by the Training Programs in Epidemiology and Public Health Interventions
Network (TEPHINET in the US), the World Health Organization (WHO)-Lebanon, and the Lebanese
National Council for Scientific Research through its support of the Associated Research Unit on
Under-nutrition and Obesity in Lebanon.

Authors’ Contributions

LN wrote the paper and contributed to data analysis/interpretation; NH and AMS designed and
supervised the study and critically reviewed the paper; MCC coordinated field work; FN participated
in data analysis and critically reviewed the paper; SK participated in data analysis; CA conducted data
analysis and contributed to data interpretation; LN and FN contributed equally to this manuscript. All
authors read and approved the final manuscript.

Conflicts of Interest

The authors declare no conflict of interest.

References

1. Sibai, A.M.; Nasreddine, L.; Mokdad, A.H.; Adra, N.; Tabet, M.; Hwalla, N. Nutrition transition
and cardiovascular disease risk factors in the MENA countries: Reviewing the evidence.
Ann. Nutr. Metab. 2010, 57, 193–203.
2. Wang, Y.; Lobstein, T. Worldwide trends in childhood overweight and obesity. Int. J. Pediatr.
Obes. 2006, 1, 11–25.
3. Black, R.E.; Victora, C.G.; Walker, S.P.; Bhutta, Z.A.; Christian, P.; de Onis, M.; Ezzati, M.;
Grantham-McGregor, S.; Katz, J.; Martorell, R.; et al. Maternal and child undernutrition and
overweight in low-income and middle-income countries. Lancet 2013, 382, 427–451.
4. Weiss, R.; Dziura, J.; Burgert, T.S. Obesity and the metabolic syndrome in children and
adolescents. N. Engl. J. Med. 2004, 350, 2362–2374.
 209

5. Nasreddine, L.; Ouaijan, K.; Mansour, M.; Adra, N.; Sinno, D.; Hwalla, N. Metabolic syndrome
and insulin resistance in obese pre-pubertal children in Lebanon: A primary health concern.
Ann. Nutr. Metab. 2010, 57, 135–142.
6. Ebbeling, C.B.; Pawlak, D.B.; Ludwig, D.S. Childhood obesity: Public-health crisis, common
sense cure. Lancet 2002, 360, 473–482.
7. Krebs, N.F.; Jacobson, M.S.; American Academy of Pediatrics Committee on Nutrition.
Prevention of pediatric overweight and obesity. Pediatrics 2003, 112, 424–430.
8. Blüher, S.; Molz, E.; Wiegand, S.; Otto, K.P.; Sergeyev, E.; Tuschy, S.; L’Allemand-Jander, D.;
Kiess, W.; Holl, R.W.; Adiposity Patients Registry Initiative and German Competence Net
Obesity. Body mass index, waist circumference, and waist-to-height ratio as predictors of
cardiometabolic risk in childhood obesity depending on pubertal development. J. Clin.
Endocrinol. Metab. 2013, 98, 3384–3393.
9. Hirschler, V.; Molinari, C.; Maccallini, G.; Aranda, C.; Oestreicher, K. Comparison of different
anthropometric indices for identifying dyslipidemia in school children. Clin. Biochem. 2011, 44,
659–664.
10. Janssen, I.; Katzmarzyk, P.T.; Ross, R. Waist circumference and not body mass index explains
obesity related health risk. Am. J. Clin. Nutr. 2004, 79, 379–384.
11. Janssen, I.; Katzmarzyk, P.T.; Srinivasan, S.R.; Chen, W.; Malina, R.M.; Bouchard, C.;
Berenson, G.S. Combined influence of body mass index and waist circumference on coronary
artery disease risk factors among children and adolescents. Pediatrics 2005, 115, 1623–1630.
12. Lee, S.; Bacha, F.; Gungor, N.; Arslanian, S.A. Waist circumference is an independent predictor
of insulin resistance in black and white youths. J. Pediatr. 2006, 148, 188–194.
13. Maffeis, C.; Banzato, C.; Talamini, G.; Obesity Study Group of the Italian Society of Pediatric
Endocrinology and Diabetology. Waist to-height ratio, a useful index to identify high metabolic
risk in overweight children. J. Pediatr. 2008, 152, 207–213.
14. McCarthy, H.D.; Ashwell, M. A study of central fatness using waist-to-height ratios in UK
children and adolescents over two decades supports the simple message—“Keep your waist
circumference to less than half your height”. Int. J. Obes. 2006, 30, 988–992.
15. Freedman, D.S.; Kahn, H.S.; Mei, Z.; Grummer-Strawn, L.M.; Dietz, W.H.; Srinivasan, S.R.;
Berenson, G.S. Relation of body mass index and waist-to-height ratio to cardiovascular disease
risk factors in children and adolescents: The Bogalusa Heart Study. Am. J. Clin. Nutr. 2007, 86,
33–40.
16. Mushtaq, M.U.; Gull, S.; Abdullah, H.M.; Shahid, U.; Shad, M.A.; Akram, J. Waist circumference,
waist-hip ratio and waist-height ratio percentiles and central obesity among Pakistani children
aged five to twelve years. BMC Pediatr. 2011, 11, 105.
17. Huxley, R.; Mendis, S.; Zheleznyakov, E.; Reddy, S.; Chan, J. Body mass index, waist
circumference and waist: Hip ratio as predictors of cardiovascular risk—A review of the literature.
Eur. J. Clin. Nutr. 2010, 64, 16–22.
18. Lebanese Republic Ministry of Social Affairs; Central Administration for Statistics and the United
Nations Development Programme (UNDP). Living Conditions of Households—The National
Survey of Household Living Conditions; MOSA: Beirut, Lebanon, 2004.
 210

19. Sibai, A.M.; Hwalla, N.; Adra, N.; Rahal, B. Prevalence and covariates of obesity in Lebanon:
Findings from the first epidemiological study. Obes. Res. 2003, 11, 1353–1361.
20. Nasreddine, L.; Naja, F.; Chamieh, M.C.; Adra, N.; Sibai, A.M.; Hwalla, N. Trends in overweight
and obesity in Lebanon: Evidence from two national cross-sectional surveys (1997 and 2009).
BMC Public Health 2012, 12, 798.
21. Lee, R.; Nieman, D. Nutritional Assessment, 4th ed.; McGraw-Hill Companies: Boston, NY, USA,
2007.
22. De Onis, M.; Onyango, A.W.; Borghi, E.; Siyam, A.; Nishida, C.; Siekmann, J. Development of a
WHO growth reference for school-aged children and adolescents. Bull. World Health Organ.
2007, 85, 660–667.
23. Cole, T.J.; Bellizzi, M.C.; Flegal, K.M.; Dietz, W.H. Establishing a standard definition for child
overweight and obesity worldwide: International survey. BMJ 2000, 320, 1240–1243.
24. Kuczmarski, R.J.; Ogden, C.L.; Guo, S.S.; Grummer-Strawn, L.M.; Flegal, K.M.; Mei, Z.; Wei, R.;
Curtin, L.R.; Roche, A.F.; Johnson, C.L. 2000 CDC growth charts for the United States: Methods
and development. Vital Health Stat. 2002, 246, 1–190.
25. Zimmet, P.; Alberti, K.G.; Kaufman, F. The metabolic syndrome in children and adolescents—An
IDF consensus report. Pediatr. Diabetes 2007, 8, 299–306.
26. Fernandez, J.R.; Redden, D.T.; Pietrobelli, A.; Allison, D.B. Waist circumference
percentiles in nationally representative samples of African-American, European-American, and
Mexican-American children and adolescents. J. Pediatr. 2004, 145, 439–444.
27. Moshfegh, A.J.; Borrud, L.; Perloff, B.; LaComb, R. Improved method for the 24-h dietary recall
for use in national surveys. FASEB J. 1999, 13, A603.
28. Gharib, M.; Rasheed, P. Obesity among Bahrani children and adolescents: Prevalence and
associated factors. JBMS 2008, 20, 114–123.
29. Nasreddine, L.; Sibai, A.M.; Mrayati, M.; Adra, N.; Hwalla, N. Adolescent obesity in Syria:
Prevalence and associated factors. Child Care Health Dev. 2010, 36, 404–413.
30. Bener, A. Prevalence of obesity, overweight, and underweight in Qatari adolescents. Food Nutr.
Bull. 2006, 27, 39–45.
31. Malik, M.; Bakir, A. Prevalence of overweight and obesity among children in the United Arab
Emirates. Obes. Rev. 2006, 8, 15–20.
32. Ogden, C.L.; Carroll, M.D.; Curtin, L.R.; Lamb, M.M.; Flegal, K.M. Prevalence of high body
mass index in US children and adolescents, 2007–2008. JAMA 2010, 303, 242–249.
33. Taheri, F.; Kazemi, T. Prevalence of overweight and obesity in 7 to 18 year-old children in
Birjand/Iran. Iran. J. Pediatr. 2009, 19, 135–140.
34. El Mouzan, M.I.; Foster, P.J.; Herbish, A.S.; Al Salloum, A.A.; Al Omer, A.A.; Qurachi, M.M.;
Kecojevic, T. Prevalence of overweight and obesity in Saudi children and adolescents.
Ann. Saudi Med. 2010, 30, 203–208.
35. Kromeyer-Hauschild, K.; Neuhauser, H.; Schaffrath Rosario, A.; Schienkiewitz, A.
Abdominal obesity in German adolescents defined by waist-to-height ratio and its association to
elevated blood pressure: The KiGGS Study. Obes. Facts 2013, 6, 165–175.
36. Papalia, T.; Greco, R.; Lofaro, D.; Mollica, A.; Roberti, R.; Bonofiglio, R. Anthropometric
measures can better predict high blood pressure in adolescents. J. Nephrol. 2013, 26, 899–905.
 211

37. Gröber-Grätz, D.; Widhalm, K.; de Zwaan, M.; Reinehr, T.; Bluher, S.; Schwab, K.O.; Wiegand, S.;
Holl, R.W. Body mass index or waist circumference: Which is the better predictor for
hypertension and dyslipidemia in overweight/obese children and adolescents? Association of
cardiovascular risk related to body mass index or waist circumference. Horm. Res. Paediatr. 2013,
80, 170–178.
38. Al-Almaie, S.M. Prevalence of obesity and overweight among Saudi adolescents in Eastern Saudi
Arabia. Saudi Med. J. 2006, 26, 607–611.
39. Hassapidou, M.; Fotiadou, E.; Maglara, E.; Papadopoulou, S.K. Energy intake, diet composition,
energy expenditure, and body fatness of adolescents in Northern Greece. Obesity 2006, 14,
855–862.
40. Hwalla, N.; Sibai, A.M.; Adra, N. Adolescent Obesity and Physical Activity. World Rev. Nutr.
Diet. 2005, 94, 42–50.
41. Oliveira, A.M.; Oliveira, A.C.; Almeida, M.S.; Oliveira, N.; Adan, L. Influence of the family
nucleus on obesity in children from northeastern Brazil: A cross-sectional study. BMC Public
Health 2007, 7, 235.
42. Deshmukh-Taskar, P.R.; Radcliffe, J.D.; Liu, Y.; Nicklas, T.A. Do breakfast skipping and
breakfast type affect energy intake, nutrient intake, nutrient adequacy, and diet quality in young
adults? NHANES 1999–2002. J. Am. Coll. Nutr. 2010, 29, 407–418.
43. Szajewska, H.; Ruszczynski, M. Systematic review demonstrating that breakfast consumption
influences body weight outcomes in children and adolescents in Europe. Crit. Rev. Food Sci. Nutr.
2010, 50, 113–119.
44. Maddah, M.; Nikooyeh, B. Factors associated with overweight in children in Rasht, Iran: Gender,
maternal education, skipping breakfast and parental obesity. Public Health Nutr. 2009, 13, 196–200.
45. Lioret, S.; Volatier, J.L.; Lafay, L.; Touvier, M.; Maire, B. Is food portion size a risk factor of
childhood overweight? Eur. J. Clin. Nutr. 2009, 63, 382–391.
46. Utter, J.; Scragg, R.; Mhurchu, C.N.; Schaaf, D. At-home breakfast consumption among
New Zealand children: Associations with body mass index and related nutrition behaviors. J. Am.
Diet. Assoc. 2007, 107, 570–576.
47. Drummond, S.; Crombie, N.; Kirk, T. A critique of the effects of snacking on body weight status.
Eur. J. Clin. Nutr. 1996, 50, 779–783.
48. Niemeier, H.M.; Raynor, H.A.; Lloyd-Richardson, E.E.; Rogers, M.L.; Wing, R.R. Fast food
consumption and breakfast skipping: Predictors of weight gain from adolescence to adulthood in a
nationally representative sample. J. Adolesc. Health 2006, 39, 842–849.
49. Thompson, O.M.; Ballew, C.; Resnicow, K.; Must, A.; Bandini, L.G.; Cyr, H.; Dietz, W.H.
Food purchased away from home as a predictor of change in BMI z-score among girls. Int. J.
Obes. Relat. Metab. Disord. 2004, 28, 282–289.
50. O’Donnell, S.I.; Hoerr, S.L.; Mendoza, J.A.; Tsuei Goh, E. Nutrient quality of fast food kids
meals. Am. J. Clin. Nutr. 2008, 88, 1388–1395.
51. Poti, J.M.; Slining, M.M.; Popkin, B.M. Where are kids getting their empty calories?
Stores, schools, and fast-food restaurants each played an important role in empty calorie intake
among US children during 2009–2010. J. Acad. Nutr. Diet. 2013, doi:10.1016/j.jand.2013.08.012.
 212

52. Poti, J.M.; Duffey, K.J.; Popkin, B.M. The association of fast food consumption with poor dietary
outcomes and obesity among children: Is it the fast food or the remainder of diet? Am. J. Clin. Nutr.
2013, doi:10.3945/ajcn.113.071928.
53. Powell, L.M.; Nguyen, B.T. Fast-food and full-service restaurant consumption among children
and adolescents: Effect on energy, beverage, and nutrient intake. JAMA Pediatr. 2013, 167,
14–20.
54. Paeratakul, S.; Ferdinand, D.P.; Champagne, C.M.; Ryan, D.H.; Bray, G.A. Fast-food
consumption among US adults and children: Dietary and nutrient intake profile. J. Am. Diet.
Assoc. 2003, 103, 1332–1338.
55. Anderson, G.H.; Moore, S.E. Dietary proteins in the regulation of food intake and body weight in
humans. J. Nutr. 2004, 134, 974S–979S.
56. Barba, G.; Troiano, E.; Russo, P.; Venezia, A.; Siani, A. Inverse association between body mass
and frequency of milk consumption in children. Br. J. Nutr. 2005, 93, 15–19.
57. Carruth, B.R.; Skinner, J.D. The role of dietary calcium and other nutrients in moderating body fat
in preschool children. Int. J. Obes. Relat. Metab. Disord. 2001, 25, 559–566.
58. Skinner, J.D.; Bounds, W.; Carruth, B.R.; Ziegler, P. Longitudinal calcium intake is negatively
related to children’s body fat indexes. J. Am. Diet. Assoc. 2003, 103, 1626–1631.
59. Heaney, R.P.; Davies, K.M.; Barger-Lux, M.J. Calcium and weight: Clinical studies. J. Am. Coll.
Nutr. 2002, 21, 152S–155S.
60. Zemel, M.B. Nutritional and endocrine modulation of intracellular calcium: Implications in
obesity, insulin resistance and hypertension. Mol. Cell. Biochem. 1998, 188, 129–136.
61. Haro-Mora, J.J.; García-Escobar, E.; Porras, N.; Alcázar, D.; Gaztambide, J.; Ruíz-Órpez, A.;
García-Serrano, S.; Rubio-Martín, E.; García-Fuentes, E.; López-Siguero, J.P.; et al. Children
whose diet contained olive oil had a lower likelihood of increasing their body mass index z-score
over 1 year. Eur. J. Endocrinol. 2011, 165, 435–439.
62. Rodriguez, V.M.; Portillo, M.P.; Pico, C.; Macarulla, M.T.; Palou, A. Olive oil feeding up-regulates
uncoupling protein genes in rat brown adipose tissue and skeletal muscle. Am. J. Clin. Nutr. 2002,
75, 213–220.
63. Soriguer, F.; Moreno, F.; Rojo-Martinez, G.; Cardona, F.; Tinahones, F.; Gomez-Zumaquero, J.M.;
Garcia Fuentes, E.; Morcillo, S. Redistribution of abdominal fat after a period of food restriction
in rats is related to the type of dietary fat. Br. J. Nutr. 2003, 89, 115–122.
64. Soriguer, F.; Moreno, F.; Rojo-Martinez, G.; Garcia-Fuentes, E.; Tinahones, F.;
Gomez-Zumaquero, J.M.; Cuesta Munoz, A.L.; Cardona, F.; Morcillo, S. Monounsaturated n-9
fatty acids and adipocyte lipolysis in rats. Br. J. Nutr. 2003, 90, 1015–1022.
65. Malik, V.S.; Pan, A.; Willett, W.C.; Hu, F.B. Sugar-sweetened beverages and weight gain in
children and adults: A systematic review and meta-analysis. Am. J. Clin. Nutr. 2013, 98,
1084–1102.
66. Te Morenga, L.; Mallard, S.; Mann, J. Dietary sugars and body weight: Systematic review and
meta-analyses of randomised controlled trials and cohort studies. BMJ 2012, 346, e7492.
67. Malik, V.S.; Schulze, M.B.; Hu, F.B. Intake of sugar-sweetened beverages and weight gain:
A systematic review. Am. J. Clin. Nutr. 2006, 84, 274–288.
 213

68. De Bourdeaudhuij, I.; Verloigne, M.; Maes, L.; Van Lippevelde, W.; Chinapaw, M.J.;
Te Velde, S.J.; Manios, Y.; Androutsos, O.; Kovacs, E.; Dössegger, A.; et al. Associations of
physical activity and sedentary time with weight and weight status among 10- to 12-year-old boys
and girls in Europe: A cluster analysis within the ENERGY project. Pediatr. Obes. 2013, 8, 367–375.
69. Sisson, S.B.; Church, T.S.; Martin, C.K.; Tudor-Locke, C.; Smith, S.R.; Bouchard, C.;
Earnest, C.P.; Rankinen, T.; Newton, R.L.; Katzmarzyk, P.T. Profiles of sedentary behavior in
children and adolescents: The US national health and nutrition examination survey, 2001–2006.
Int. J. Pediatr. Obes. 2009, 4, 353–359.
70. Mushtaq, M.U.; Gull, S.; Shahid, U.; Shafique, M.M.; Abdullah, H.M.; Shad, M.A.; Siddiqui, A.M.
Family-based factors associated with overweight and obesity among Pakistani primary school
children. BMC Pediatr. 2011, 11, 114.
71. Apfelbacher, C.J.; Loerbroks, A.; Cairns, J.; Behrendt, H.; Ring, J.; Krämer, U. Predictors of
overweight and obesity in five to seven-year-old children in Germany: Results from cross-sectional
studies. BMC Public Health 2008, 8, 171.
72. Hawkins, S.S.; Cole, T.J.; Law, C.; The Millennium Cohort Study Child Health Group. An
ecological systems approach to examining risk factors for early childhood overweight: Findings
from the UK Millennium Cohort Study. J. Epidemiol. Community Health 2009, 63, 147–155.
73. Lazzeri, G.; Pammolli, A.; Pilato, V.; Giacchi, M.V. Relationship between 8/9-year-old school
children BMI, parents’ BMI and educational level: A cross sectional survey. Nutr. J. 2011, 10, 76.
74. Watanabe, E.; Lee, J.S.; Kawakubo, K. Associations of maternal employment and three-generation
families with pre-school children’s overweight and obesity in Japan. Int. J. Obes. 2011, 35,
945–952.
75. Crawford, P.B.; Obarzanek, E.; Morrison, J.; Sabry, Z.I. Comparative advantage of 3-day
food records over 24-h recall and 5-day food frequency validated by observation of 9- and
10-year-old girls. J. Am. Diet. Assoc. 1994, 94, 626–630.
76. Livingstone, M.B.; Robson, P.J. Measurement of dietary intake in children. Proc. Nutr. Soc. 2000,
59, 279–293.
77. Burrows, T.L.; Martin, R.J.; Collins, C.E. A systematic review of the validity of dietary
assessment methods in children when compared with the method of doubly labeled water. J. Am.
Diet. Assoc. 2010, 110, 1501–1510.
78. United Nations; Division for Public Administration and Development Management (DPADM);
Department of Economic and Social Affairs (DESA). Republic of Lebanon, Public Administration,
Country Profile; United Nations: New York, NY, USA, 2004.
79. Poskitt, E.M. Countries in transition: Underweight to obesity non-stop? Ann. Trop. Paediatr.
2009, 29, 1–11.
80. Hoelscher, D.M.; Kirk, S.; Ritchie, L.; Cunningham-Sabo, L.; Academy Positions Committee.
Position of the Academy of Nutrition and Dietetics: Interventions for the prevention and treatment
of pediatric overweight and obesity. J. Acad. Nutr. Diet. 2013, 113, 1375–1394.
© 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).
214
 215

Reprinted from Nutrients. Cite as: Gallagher, C.M.; Black, L.J.; Oddy, W.H. Micronutrient Intakes
from Food and Supplements in Australian Adolescents. Nutrients 2014, 6, 342-354.

Micronutrient Intakes from Food and Supplements in

Australian Adolescents
Caroline M. Gallagher, Lucinda J. Black and Wendy H. Oddy *

Telethon Institute for Child Health Research, The University of Western Australia, Perth,
Western Australia 6008, Australia; E-Mails: cgallagher@ichr.uwa.edu.au (C.M.G.);
lblack@ichr.uwa.edu.au (L.J.B.)

* Author to whom correspondence should be addressed; E-Mail: wendyo@ichr.uwa.edu.au;

Tel.: +61-8-9489-7879; Fax: +61-8-9489-7700.

Received: 28 November 2013; in revised form: 17 December 2013 / Accepted: 8 January 2014 /
Published: 14 January 2014

Abstract: Objective: Low micronutrient intakes in adolescents are frequently reported. We

assessed micronutrient intakes in adolescents to determine whether supplement use
optimises intakes. Methods: Dietary intake was assessed using a food frequency
questionnaire in 17 year old participating in the Western Australian Pregnancy Cohort
(Raine) Study (n = 991). We calculated median daily micronutrient intakes in supplement
users and non-users (from food sources only and from food and supplements), along with
the percentage of adolescents meeting the Estimated Average Requirements (EAR) or
Adequate Intake (AI) where appropriate. Results: Intakes of calcium, magnesium, folate
and vitamins D and E from food only were low. Although supplements significantly
increased micronutrient intakes in supplement users, more than half of supplement users
failed to meet the EAR or AI for some key micronutrients. Compared with non-users,
supplement users had higher micronutrient intakes from food sources with the exception of
vitamins D and B12 and were more likely to achieve the EAR or AI for many
micronutrients from food only. Conclusions: Intakes of some key micronutrients were low
in this population, even among supplement users. Those facing the greatest risk of
micronutrient deficiencies were less likely to use supplements.

Keywords: adolescents; food intake; micronutrients; dietary supplements; Raine Study

 216

1. Introduction

Low intakes of micronutrients, including calcium, folate, magnesium and potassium, have been
previously reported in Australian adolescents [1]. Similarly, adolescent diets in Europe and the United
States (US) have been associated with low intakes of calcium, vitamin D, iron, folate and zinc [2–5].
Assessing micronutrient status in adolescents is important due to the contribution of micronutrients to
disease prevention [6]. Herbison and colleagues reported that low intake of B vitamins was associated
with poor mental health and behaviour in adolescents [7]; calcium and magnesium may play a
protective role in type 2 diabetes [8]; and young adults with higher magnesium have a lower risk of
developing the metabolic syndrome [9]. Furthermore, adequate calcium and vitamin D levels are
essential during adolescence, when approximately 40% of total bone mass is accumulated [10,11].
In order to reliably assess micronutrient intakes, the contribution of nutritional supplements to
intake must be taken into account [12]. Nutritional supplement use is increasing in many countries and
is popular worldwide in adolescents [5,12–16]. The EPIC study in the United Kingdom (UK) found
that the contribution of nutritional supplements to nutrient intakes can be substantial, and
miscalculation of nutrient intakes can occur if supplement use is not considered [17]. Therefore, we
aimed to assess micronutrient intakes in 17 year old adolescents in Western Australia and to determine
whether supplement use optimises micronutrient intakes.

2. Experimental Section

2.1. Participants

Participants were from the Western Australian Pregnancy Cohort (Raine) Study, which has been
described previously [18]. In brief, 2900 pregnant women were recruited through the public antenatal
clinic at King Edward Memorial Hospital and nearby private clinics in Perth, Western Australia
between May 1981 and November 1991. A total of 2868 children were available for follow-up. The
King Edward Memorial Hospital and Princess Margaret Hospital Ethic Committees approved the study
protocol. The participant and/or their primary caregiver provided written consent for their participation
in the study. In order to increase participation at each follow-up, participants were sent regular
newsletters, Christmas cards, birthday cards and received regular updates and results of the study.
Participants for the 17 year follow-up were contacted by a research assistant over the telephone. A total
of 2168 adolescents were eligible for follow-up at 17 years between July 2006 and June 2009.
Of these, 1754 individuals participated and 1009 provided dietary intake data.

2.2. Dietary Intake

Dietary intake at the 17 year follow-up was assessed using a self-reported semi-quantitative food
frequency questionnaire (FFQ) developed by the Commonwealth Scientific and Industrial Research
Organisation (CSIRO) in Adelaide, Australia [19]. This FFQ has been validated for reliability against a
3-day food record in the same cohort [20] and also in adults [21]. From the FFQ we collected
information on 212 foods, mixed dishes and beverages, including beverages and snacks popular among
adolescents. An overall estimate of the adolescents’ usual dietary intakes in the past year was
 217

established using the portion size in standard household measures, and the number of times the food
was eaten per day, per week or per month. Participants were asked to record any additional items that
were consumed regularly but were not included in the FFQ. All questionnaires were checked by
a research nurse and queries were clarified with the adolescent. Seasonal differences were accounted
for by asking how often foods were eaten in summer and winter. Food intake data were entered into a
database and verified by CSIRO. Estimated daily micronutrient intakes were provided by CSIRO using
nutrient composition derived from four sources: the Australian nutrient database (NUTTAB95) [22];
the British Food Composition Tables [23]; the US Department of Agriculture food tables [24]; and
manufacturers’ data. Questionnaires were excluded if the daily energy intake reported was implausible
(<3000 or >20,000 kJ per day). Micronutrient intakes were calculated for thiamin, riboflavin, niacin,
pantothenic acid, pyridoxine, vitamin B12, folate, beta-carotene, vitamins A, C, D, E, calcium, iron,
potassium, magnesium, zinc, phosphorus and copper.

2.3. Supplement Use

Participants were asked to record any supplements they used over the last twelve months, including
brand, name of product, dose and frequency of use. Composition data for supplements were obtained
from the product label or directly from the manufacturer. If the frequency of use was less than daily,
the nutrient intake was calculated to reflect daily intake over the last twelve months. When there were
insufficient data regarding the brand, name, dose or frequency of use, a standardised default was used
based on the most common supplement of that type recorded by the participants. Micronutrient intake
from supplements was added to the intake from food sources to give a total daily micronutrient intake
for supplement users.

2.4. Demographic Characteristics

Height was measured using a Holtain Stadiometer to the nearest 0.1 cm. Weight was measured
using a Wedderburn Digital Chair Scale to the nearest 100 g. Body Mass Index (BMI) was calculated
as weight in kilograms divided by height in metres squared. Underweight, normal weight, overweight
and obesity were defined according to age- and sex-specific BMI cut-offs [25,26]. Physical activity
was assessed using a self-reported questionnaire, based on exercise outside of school hours per week,
where exercise was defined as activity causing breathlessness or sweating (• WLPHV SHU ZHHN
1–3 times per week and <once per week). Television/computer viewing was assessed by the amount of
hours watching television or using the computer per day (<2 h per day, 2–4 h per day, >4 h per day).
Family income was defined as the gross income before tax and was determined as AUD (per year)
<$35,000, $35,001–$70,000 or >$70,001 (average gross salary in 2009 was AUD $63,612 [27]).
Maternal education level was indexed by whether the mother had completed 12 years of education or
not by the time the child was 8 years old.

2.5. Statistical Analysis

Chi-square tests were applied to identify differences in sex, BMI category, maternal education,
family income, screen use and physical activity between supplement users and non-users. Median daily
 218

micronutrient intakes in males and female supplement users and non-users were calculated from food
sources alone and from food and supplements. The percentage of males and female supplement users
and non-users meeting the Estimated Average Requirement (EAR) or Adequate Intake (AI) was
calculated. The EAR is defined as the daily nutrient level estimated to meet the requirements of half
the healthy individuals in a particular life stage and gender [28]. Where evidence was insufficient or
too conflicting to establish an EAR, an Adequate Intake (AI) was set. The AI is defined as the average
daily nutrient intake level based on observed or experimentally-determined approximations or
estimates of nutrient intake by a group (or groups) of apparently healthy people that are assumed to be
adequate [27].
Since most micronutrient intakes were non-parametrically distributed, Mann Whitney-U tests were
applied to investigate differences in male and female intakes from food sources in supplement users
and non-users. We applied Wilcoxon signed rank tests to identify differences in micronutrient intakes
between male and female supplement users, from food sources only and from food sources plus
supplements. Chi-square tests were used to determine differences in the percentage of adolescents
achieving the EAR from food sources between supplement users and non-users. We used Statistical
Package for Social Science for Windows Rel.20.0.0 (Chicago: SPSSS Inc., Illinois, IL, USA) and
defined statistical significance as p < 0.05.

3. Results

3.1. Characteristics

At the 17 year follow-up, 1009 participants provided dietary intake data; however, 18 of these were
excluded due to implausible energy intakes (<3000 or >20,000 kJ per day) [29]. Ultimately, dietary
intake data were available for 238 supplement users (24%) and 753 non-users. Supplement users were
more likely to be physically active than non-users (p < 0.05) (Table 1). No other significant differences
between supplement users and non-users were observed.

Table 1. Characteristics of 17 year old adolescents providing dietary intake data in the
Western Australian Pregnancy Cohort (Raine) Study.
Total Population Supplement Users Non-Users p value
(n = 991) (n = 238) (n = 753)
n (%) n (%) n (%)
Sex 0.061
Male 454 (45.8) 96 (40.3) 358 (47.5)
Female 537 (54.2) 142 (59.7) 395 (52.5)
BMI category 0.165
Underweight 65 (7.7) 13 (6.2) 52 (8.3)
Healthy weight 606 (72.1) 158 (75.2) 448 (71.1)
 219

Table 1. Cont.
Overweight 108 (12.9) 30 (14.3) 78 (12.4)
Obese 61 (7.3) 9 (4.3) 52 (8.3)
Computer and/or television use 0.273
<2 h per day 33 (3.6) 11 (4.6) 22 (2.9)
2–4 h per day 486 (53.1) 124 (52.1) 362 (48.1)
>4 h per day 397 (43.3) 89 (39.8) 308 (40.9)
Physical activity
Once per week or less 186 (19.7) 42 (18.4) 144 (20.2) 0.013 *
1–3 times per week 512 (54.4) 110 (48.2) 402 (53.4)
4+ times per week 244 (25.9) 76 (33.3) 168 (23.5)
Maternal Education 0.082
<12 years of 551 (55.7) 122 (51.3) 429 (57.1)
education
>12 years of 438 (44.3) 116 (48.7) 322 (42.9)
education
Annual Family income 1 0.214
<$35,000 118 (13.0) 34 (15.5) 84 (12.3)
>$35,001–$70,000 228 (25.2) 47 (21.4) 181 (26.4)
>$70,001 559 (61.8) 139 (63.2) 420 (61.3)
1
Average gross salary in Australia in 2009 was $63, 612 [26]; * Significant at p < 0.05.

3.2. Types of Supplements Consumed

The most common supplement was a multivitamin, used by 42% of male supplement users and
33% of female supplement users. This was followed by vitamin C, consumed by 39% of male
supplement users and 29% of female supplement users. Only two female supplement users consumed a
folate supplement. Protein was only taken by male supplement users. No participants were taking a
dedicated vitamin D supplement (Table 2).

Table 2. Nutritional supplements consumed by 17 year old adolescents in the Western

Australian Pregnancy Cohort (Raine) Study.
Males Females
Supplement Type
(n = 96) (n = 142)
n (%)
Vitamin C 37 (39.0) 41 (29.0)
Vitamin B/B complex 4 (4.2) 8 (5.6)
Folate 0 (0.0) 2 (1.4)
Iron 4 (4.2) 35 (24.6)
Calcium 1 (1) 7 (5.0)
Magnesium 5 (5.2) 7 (5.0)
Zinc 8 (8.3) 12 (8.5)
Multivitamin/mineral 40 (42.0) 47 (33.0)
Fish/Cod liver oil 25 (26.0) 37 (26.0)
Primrose/Starflower Oil 4 (4.2) 4 (2.8)
 220

Table 2. Cont.
Probiotics 2 (2.1) 4 (2.8)
Protein 1 7 (7.3) 0 (0.0)
Other 2 9 (9.4) 12 (8.5)
1
Protein powder, micronized creatine monohydrate, muscle building supplement, lipo 6, mixed amino acids,
2
complete protein; Fibre, cranberry, phytelle, garlic and horseradish, echinacea, garlic, glucosamine,
chondroitin, spirulina, olive oil extract, butter menthol, l-lysine.

3.3. Median Daily Micronutrient Intakes from Food and Supplements

When micronutrient intakes from food only were compared between users and non-users, male
supplement users had significantly higher intakes than male non-users (p < 0.05) for all nutrients with
the exception of vitamin D (Table 3). Female supplement users had significantly higher intakes of
magnesium, potassium, vitamin A, beta-carotene, pantothenic acid, pyridoxine, folate and vitamin C
(p < 0.05) from food sources than female non-users. In supplement users, all micronutrient intakes
were significantly higher (p < 0.05) from food and supplements compared with food only.

3.4. Adequacy of Intakes

Fewer than 50% of females (both supplement non-users and supplement users) met the EAR for
calcium, magnesium, folate or the AI for vitamins D and E (Table 4). Fewer than 50% of male
non-users met the EAR for magnesium, potassium, pantothenic acid, folate or the AI for vitamins D
and E. From food sources only, in male supplement users, fewer than 50% met the EAR for folate or
the AI for vitamins D and E. When the contribution of supplements was accounted for, there was an
approximate 20% increase in the number of males and females meeting the EAR for folate and those
meeting the AI for vitamins D and E. However, more than 50% of males still failed to meet the AI
requirements for vitamin D and more than 50% of females failed to meet the requirements for folate,
vitamins D and E as well as calcium and magnesium.

4. Discussion

We found that adolescents in Western Australia had intakes below recommendations for calcium,
folate and vitamins D and E. Females also had low intakes of magnesium. Low micronutrient intakes
were also found in other studies carried out in Australia amongst 16–18 years old, where females
reported low intakes of calcium, folate, and magnesium, and males reported low intakes of folate,
calcium and potassium [1]. Similarly, in the UK, adolescent males aged 11–18 years had low intakes
of magnesium, potassium, zinc, folate, iron and vitamin D, while females of the same age had low
intakes of calcium, magnesium, potassium, folate, iron and vitamin D [30]. In 11–14 year old
adolescents in Spain, more than 50% of males had intakes below the recommended nutrient intake
(RNI), for magnesium, calcium, folate and vitamins A, B6, D and E, while more than 50% of females
had intakes below the RNI for magnesium, calcium, folate and vitamins A, B6, D and E [31]. In the
US around 90% of adolescent girls have inadequate intakes of calcium, magnesium, potassium and
vitamins D and E, and many do not meet the recommendations for, zinc, phosphorous and vitamins A,
B6, B12 and C [32].
 221

Table 3. Median daily micronutrient intakes from food and supplements in male and female supplement non-users (n = 753) and users
(n = 238) in 17 year old adolescents in the Western Australian Pregnancy Cohort (Raine) Study.
Supplement Non-Users (n = 753) Supplement Users (n = 238) Supplement Users (n = 238)
Food Sources Food Sources Food and Supplements
Nutrient
Males (n = 358) Females (n = 395) Males (n =96) Females (n = 142) Males (n = 96) Females (n = 142)
Median Median Median
Calcium (mg) 1089.5 842.3 1395.3 * 851.4 1402.1 880.9
Iron (mg) 13.8 10.8 16.4 * 11.1 19.4 14.8
Zinc (mg) 12.8 10 14.6 * 10.1 18.8 11.7
Magnesium (mg) 309.1 248.1 397.8 * 265.7 * 428.2 287.8
Potassium (mg) 3421.8 2890 4232.9 * 3167.7 * 4243.5 3180.1
Phosphorous (mg) 1654.4 1252.9 2064.5 * 1293.8 2064.5 1293.8
Copper (mg) 1.9 1.5 2.2 * 1.7 2.2 1.7
Vitamin A (μg) 1003.1 886.7 1218.9 * 985.2 * 1447.6 1136.7
Beta-carotene (μg) 3229.5 3397.3 3805.4 * 3733.6 * 4113.1 4494.5
Thiamin (mg) 1.8 1.3 2.1 * 1.3 2.7 1.6
Riboflavin (mg) 2.3 1.8 2.9 * 1.9 3.8 2.3
Niacin (mg) 38.6 29.6 44.5 * 29.7 54.3 35.3
Pantothenic acid (mg) 5.2 4.2 6.3 * 4.6 * 8.7 5.4
Pyridoxine (mg) 1.6 1.4 2.0 * 1.5 * 2.9 2.4
B12 (μg) 4.6 3.4 5.1 * 3.4 7.4 4.3
Folate (μg) 252.4 204 313.1 * 224.4 * 399.9 280.8
Vitamin C (mg) 146.4 128.3 178.4 * 142.9 * 335 253.3
Vitamin D (μg) 1.7 1.3 1.9 1.2 3 1.5
Vitamin E (mg) 7.1 5.8 8.6 * 5.9 11.5 6.9
* Significant difference in micronutrient intakes from food sources between supplement users and non-users (p < 0.05).
 222

Table 4. Number and percentage [n (%)] of adolescents meeting the EAR or AI [27] in male and female supplement non-users (n = 753) and
users (n = 238) in 17 year old adolescents in the Western Australian Pregnancy Cohort (Raine) Study.
Supplement Non-Users (n = 753) Supplement Users (n = 238) Supplement Users (n = 238)
EAR/AI
Food Sources Food Sources Food and Supplements
Males Females Males Females Males Females
Nutrient Males Females (n = 358) (n = 395) (n = 96) (n = 142) (n = 96) (n = 142)
n (%) n (%) n (%)
Calcium (mg) 1 1050 1050 185 (51.7) 116 (29.4) 70 (72.9) * 50 (35.2) 73 (76.0) 56 (39.4)
Iron (mg) 1 8 8 336 (93.9) 312 (79.0) 92 (95.8) 114 (80.3) 94 (97.9) 128 (90.1)
Zinc (mg) 1 11 6 248 (69.3) 354 (89.6) 79 (82.3) * 126 (88.7) 95 (99.1) 128 (90.1)
Magnesium (mg) 1 340 300 153 (42.7) 125 (31.6) 59 (61.5) * 58 (40.8) * 75 (78.1) 67 (47.2)
Potassium (mg) 1 3600 2600 163 (45.5) 239 (60.5) 63 (65.6) * 101 (71.1) * 85 (88.5) 101 (71.1)
Phosphorous (mg) 1 1055 1055 308 (86.0) 266 (67.3) 90 (93.8) * 102 (71.8) 90 (93.8) 102 (71.8)
Copper (mg) 1 1.5 1.1 260 (72.6) 327 (82.8) 82 (85.4) * 126 (88.7) 94 (97.9) 128 (90.1)
Vitamin A (μg) 1 630 485 296 (82.7) 356 (90.1) 89 (92.7) * 132 (93.0) 95 (99.0) 134 (94.4)
Beta-carotene (μg) 1 n/a n/a
Thiamin (mg) 1 1.1 0.9 332 (92.7) 318 (80.5) 90 (93.8) 125 (88.0) * 94 (97.9) 130 (91.50
Riboflavin (mg) 1 1.1 0.9 339 (94.7) 363 (91.9) 93 (96.9) 137 (96.5) 96 (100) 137 (96.5)
Niacin (mg) 1 12 11 358 (100) 391 (99.0) 95 (99.0) 139 (97.9) 96 (100) 139 (97.9)
Pantothenic acid (mg) 1 6 4 133 (37.2) 214 (54.2) 54 (56.3) * 93 (65.5) * 94 (97.9) 100 (70.4)
Pyridoxine (mg) 1 1.1 1 304 (84.9) 306 (77.5) 90 (93.8) * 122 (85.9) * 95 (99.0) 129 (90.8)
B12 (μg) 1 2 2 345 (96.4) 332 (84.1) 92 (95.8) 114 (80.3) 95 (99.0) 118 (83.1)
Folate (μg) 1 330 330 45 (12.5) 26 (6.5) 94 (26.3) * 46 (32.4) * 63 (65.6) 61 (43.0)
Vitamin C (mg) 1 28 28 349 (97.5) 388(98.2) 96 (100) 141 (99.3) 96 (100) 141 (99.3)
Vitamin D (μg) 2 51 51 17 (4.7) 5 (1.3) 11 (11.5) * 3 (2.1) 33 (34.4) 38 (26.8)
Vitamin E (mg) 2 10 1 81 90 (25.1) 96 (24.3) 56 (58.3) * 33 (23.2) 56 (58.3) 57 (40.1)
EAR, Estimated Average Requirement; AI, Adequate Intake [27]. 1 EAR; 2 AI; * Significant difference in the percentage of adolescents meeting the EAR or AI from food
sources only between supplement users and non-users (p < 0.05).
 223

Approximately 24% of adolescents in this cohort consumed supplements. The prevalence of

supplement use varies in other countries from 20% to 26% in adolescents in the US [14,33] and
11%–45% in European countries [5,13,14]. In our adolescent cohort, supplement use significantly
increased the intakes of all nutrients in both males and females. However, even among supplement
users, intakes of some micronutrients, including calcium, folate, vitamins D and E, remained lower
than the recommendations. The low intakes of calcium, folate and vitamin D raises concern since
inadequate calcium intake may lead to decreased bone-mineral density and increased risk of
developing osteoporosis [16], low folate intake in females may lead to neural tube defects in the
baby [34], and low vitamin D levels may affect both skeletal and non-skeletal health [35].
Vitamin D occurs naturally in very few foods, many of which are consumed episodically and
contain relatively small amounts of vitamin D [36]. The low dietary supply of vitamin D makes it
unrealistic that individuals would achieve the recommended intake of vitamin D. Indeed, the major
source of vitamin D is exposure to sunlight and dietary sources of vitamin D are not necessary in the
presence of adequate sunlight exposure. The AI for vitamin D assumes no, or minimal, sunlight
exposure and individuals who do not meet the AI for vitamin D may have sufficient vitamin D status
based on exposure to sunlight. Furthermore, care should be taken when estimating inadequacy based
on an AI: although the AI can be used as a goal for individual intake, there is limited certainty about
the value [27].
Particular characteristics appear to be associated with supplement use. We found that supplement
users were more likely to be physically active than non-users. This finding is supported by
Reaves et al. 2006 in the CATCH study, where 9–18 year old supplement users were more physically
active than non-users and 47% of supplement users participated in team sports compared to 40% of
non-users [14]. Adolescent supplement users in the NHANES were more physically active and were
more likely to engage in less than 2 h of television/video or computer use per day than non-users [15].
In Finland, leisure time physical activity was the strongest factor associated with supplement use in
12–18 year old adolescents [13]. Grm and colleagues [5] showed that 12 and 17 year old adolescent
supplement users were more likely to be members of sports clubs than non-users.
Higher supplement use in physically active adolescents may be due to the perceived benefits of
dietary supplements on performance [37]. Indeed, protein supplements were more commonly
consumed by males in our study and in others [5,13,38], which may be due to the perceived benefits of
protein in increasing or improving sports performance [33]. In general, there appears to be a
discrepancy in the type of supplements consumed and the micronutrients that are deficient in the diet.
For example, almost all adolescents in this study achieved the recommended intake for vitamin C
through food alone, even though vitamin C was the most commonly consumed supplement. In contrast
calcium intakes were low in females but calcium supplements were only consumed by 3% of
supplement users. It has been suggested that manufacturers develop a multivitamin/mineral that
responds to the need for micronutrients that may be inadequate in the diet [39].
Supplement users in this study had higher micronutrient intakes from food sources alone compared
with non-users. Therefore, supplements were less likely to be consumed by those who would benefit
most from them. This pattern was also observed in the National Health and Nutrition Examination
Survey (NHANES), where supplement non-users were more likely to have inadequate intakes of
calcium, magnesium, phosphorous, vitamins A and C from food sources, than supplement users [40].
 224

Reaves et al. 2006 in the Child and Adolescent Trial for Cardiovascular Health (CATCH) study found
that adolescents who followed a healthier dietary pattern were also supplement users and those facing
the greatest risk of micronutrient deficiencies were less likely to use supplements [15]. This limits the
effectiveness of supplementation as a public health strategy to increase micronutrient intakes.
One strategy to improve micronutrient inadequacies is to promote a balanced diet rich in nutrient
dense foods. Another alternative to supplementation is fortification of foods with micronutrients
known to be low in the population. For example, in order to increase folate intakes at a population
level, all wheat flour for making bread (excluding organic) in Australia is fortified with 2–3 mg of
folic acid per kilogram of wheat [41]. In the US and Canada, foods are widely fortified with vitamin D
in order to increase intakes and reduce to the risk of vitamin D deficiency [42].

5. Conclusions

Finding a balance between inadequate and excessive nutrient intakes is paramount to ensuring
healthy development in adolescents. Along with increasing the consumption of nutrient-dense foods,
supplement use may help to correct micronutrient imbalances. However, our results suggest that those
who use supplements have higher micronutrient intakes from food sources and are less likely to require
supplements than non-users. Furthermore, the type of supplements used by adolescents may not match
the micronutrient deficiencies in the diet. Professional advice should be sought for correcting
micronutrient imbalances using food and/or supplements.

Acknowledgments

We are grateful to the Raine Study team and to all the Raine Study participants and their families
who took part in this study. Data collection at the 17 year follow-up was funded by the National Health
and Medical Research Council (program grant ID 353514 and project grant ID 403981). We thank the
Telstra Research Foundation, the West Australian Health Promotion Foundation, the Australian Rotary
Health Research Fund, the National Heart Foundation of Australia/Beyond Blue and the National
Health and Medical Research Council (project grant ID 634445; project grant ID 1022134; program
grant ID 003209) for their provision of further funding for investigator and data support. We
appreciate core funding support from the University of Western Australia (UWA), the Raine Medical
Research Foundation, the Telethon Institute for Child Health Research, the UWA Faculty of Medicine,
Dentistry and Health Sciences, the Women and Infants Research Foundation and Curtin University.

Conflicts of Interest

The authors declare no conflict of interest.

References

1. Australian Bureau of Statistics. National Nutrition Survey: Nutrient Intakes and Physical
Measurements Australia; ABS: Canberra, Australia, 1995.
2. Lytle, L.A. Nutritional issues for adolescents. J. Am. Diet. Assoc. 2002, 102, 8S–12S.
 225

3. O’Connor, M.P.; O’Brien, N.M.; O’Sullivan, K.; O’Connor, T.P. Micronutrient intakes and
inadequacies in school-going Irish adolescents. Proc. Nutr. Soc. 2009, 67, doi:10.1017/S002
9665108008926.
4. Rolland-Cachera, M.F.; Bellisle, F.; Deheeger, M. Nutritional status and food intake in
adolescents living in Western Europe. Eur. J. Clin. Nutr. 2000, 54, 41S–46S.
5. Sterlinko Grm, H.; Stubelj Ars, M.; Besednjak-Kocijancic, L.; Golja, P. Nutritional supplement
use among Slovenian adolescents. Public Health Nutr. 2012, 15, 587–593.
6. Kawade, R. Zinc status and its association with the health of adolescents: A review of studies in
India. Glob. Health Action 2012, 5, 7353–7362.
7. Herbison, C.E.; Hickling, S.; Allen, K.L.; O’Sullivan, T.A.; Robinson, M.; Bremner, A.P.;
Huang, R.C.; Beilin, L.J.; Mori, T.A.; Oddy, W.H. Low intake of B-vitamins is associated with
poor adolescent mental health and behaviour. Prev. Med. 2012, 55, 634–638.
8. Villegas, R.; Gao, Y.T.; Dai, Q.; Yang, G.; Cai, H.; Li, H.; Zheng, W.; Shu, X.O. Dietary calcium
and magnesium intakes and the risk of type 2 diabetes: The Shanghai Women’s Health Study.
Am. J. Clin. Nutr. 2009, 89, 1059–1067.
9. He, K.; Liu, K.; Daviglus, M.L.; Morris, S.J.; Loria, C.M.; van Horn, L.; Jacobs, D.R., Jr.;
Savage, P.J. Magnesium intake and incidence of metabolic syndrome among young adults.
Circulation 2006, 113, 1675–1682.
10. Baroncelli, G.I.; Bertelloni, S.; Sordini, F.; Saggese, G. Osteoporosis in children and adolescents:
Etiology and management. Paediatr. Drugs 2005, 7, 295–323.
11. Cranney, A.; Horsley, T.; O’Donnell, S.; Weiler, H.; Puil, L.; Ooi, D.; Atkinson, S.; Ward, L.;
Moher, D.; Hanley, D.; et al. Effectiveness and safety of vitamin D in relation to bone health.
Evid. Rep. Technol. Assess. 2007, 158, 1–235.
12. Picciano, M.F.; Dwyer, J.T.; Radimer, K.L.; Wilson, D.H.; Fisher, K.D.; Thomas, P.R.;
Yetley, E.A.; Moshfegh, A.J.; Levy, P.S.; Nielsen, S.J.; et al. Dietary supplement use among
infants, children, and adolescents in the United States, 1999–2002. Arch. Pediatr. Adolesc. Med.
2007, 161, 978–985.
13. Mattila, V.M.; Parkkari, J.; Laaksos, L.; Pihlajamaki, H.; Rimpela, A. Use of dietary supplements
and anabolic-androgenic steroids among Finnish adolescents in 1991–2005. Eur. J. Public Health
2010, 20, 306–311.
14. Reaves, L.; Steffen, L.M.; Dwyer, J.T.; Webber, L.S.; Lytle, L.A.; Feldman, H.A.; Hoelscher, D.M.;
Zive, M.M.; Osganian, S.K. Vitamin supplement intake is related to dietary intake and physical
activity: The Child and Adolescent Trial for Cardiovascular Health (CATCH). J. Am. Diet. Assoc.
2006, 106, 2018–2023.
15. Shaikh, U.; Byrd, R.S.; Auinger, P. Vitamin and mineral supplement use by children and
adolescents in the 1999–2004 National Health and Nutrition Examination Survey: Relationship
with nutrition, food security, physical activity, and health care access. Arch. Pediatr. Adolesc. Med.
2009, 163, 150–157.
16. Stang, J.; Story, M.T.; Harnack, L.; Neumark-Sztainer, D. Relationships between vitamin and
mineral supplement use, dietary intake, and dietary adequacy among adolescents. J. Am.
Diet. Assoc. 2000, 100, 905–910.
 226

17. Welch, A.A.; Mulligan, A.A.; Luben, R.N.; Bingham, S.A. Intakes of selected nutrients from
vitamin and mineral supplements by subjects in the EPIC study. Proc. Nutr. Soc. 1998, 57, 154A.
18. Newnham, J.P.; Evans, S.F.; Michael, C.A.; Stanley, F.J.; Landau, L.I. Effects of frequent
ultrasound during pregnancy: A randomised controlled trial. Lancet 1993, 342, 887–891.
19. Baghurst, K.I.; Record, S.J. A computerised dietary analysis system for use with diet diaries or
food frequency questionnaires. Community Health Stud. 1984, 8, 11–18.
20. Ambrosini, G.L.; de Klerk, N.H.; O’Sullivan, T.A.; Beilin, L.J.; Oddy, W.H. The reliability of
a food frequency questionnaire for use among adolescents. Eur. J. Clin. Nutr. 2009, 63, 1251–1259.
21. Ambrosini, G.L.; Mackerras, D.; de Klerk, N.H.; Musk, A.W. Comparison of an Australian
food-frequency questionnaire with diet records: Implications for nutrition surveillance.
Public Health Nutr. 2003, 6, 415–422.
22. Lewis, J.; Milligan, G.C.; Hunt, A. NUTTAB95 Nutrient Data Table for Use in Australia;
Australian Government Publising Service: Canberra, Australia, 1995.
23. Holland, B.W.A.; Uwin, I.D.; Buss, D.H.; Paul, A.A.; Southgate, D.A.T. McCance and
Widdowson’s the Composition of Foods; Royal Society of Chemistry: Cambridge, UK, 1993.
24. US departure of Agriculture. Available online: http://www.nal.usda.gov/fnic/foodcomp/search/
(accessed on 7 December 2011).
25. Cole, T.J.; Bellizzi, M.C.; Flegal, K.M.; Dietz, W.H. Establishing a standard definition for child
overweight and obesity worldwide: International survey. BMJ 2000, 320, 1240–1246.
26. Cole, T.J.; Flegal, K.M.; Nicholls, D.; Jackson, A.A. Body mass index cut offs to define thinness
in children and adolescents: International survey. BMJ 2007, 335, 194–201.
27. Australian Bureau of Statistics: Average Weekly Earnings. Available online:
http://www.abs.gov.au (accessed on 16 December 2013).
28. Australian Government; Department of Aging; National Heath and Medical Research Council.
Nutrient References Values for Australia and New Zealand Including Recommended Dietary
Intakes; Commonwealth of Australia: Canberra, Australia, 2005.
29. Rockett, H.R.; Breitenbach, M.; Frazier, A.L.; Witschi, J.; Wolf, A.M.; Field, A.E.; Colditz, G.A.
Validation of a youth/adolescent food frequency questionnaire. Prev. Med. 1997, 26, 808–816.
30. Whitton, C.; Nicholson, S.K.; Roberts, C.; Prynne, C.J.; Pot, G.K.; Olson, A.; Fitt, E.; Cole, D.;
Teucher, B.; Bates, B.; et al. National Diet and Nutrition Survey: UK food consumption and
nutrient intakes from the first year of the rolling programme and comparisons with previous
surveys. Br. J. Nutr. 2011, 106, 1899–1914.
31. Serra-Majem, L.; Ribas-Barba, L.; Perez-Rodrigo, C.; Bartrina, J.A. Nutrient adequacy in Spanish
children and adolescents. Br. J. Nutr. 2006, 96, 49S–57S.
32. Moore, L.L.; Singer, M.R.; Qureshi, M.M.; Bradlee, M.L.; Daniels, S.R. Food group intake and
micronutrient adequacy in adolescent girls. Nutrients 2012, 4, 1692–1708.
33. Alves, C.; Lima, R.V. Dietary supplement use by adolescents. J. Pediatr. 2009, 85, 287–294.
34. De-Regil, L.M.; Fernandez-Gaxiola, A.C.; Dowswell, T.; Pena-Rosas, J.P. Effects and safety of
periconceptional folate supplementation for preventing birth defects. Cochrane Database
Syst. Rev. 2010, 10, 1–141.
35. Nowson, C.A.; Margerison, C. Vitamin D intake and vitamin D status of Australians.
Med. J. Aust. 2002, 177, 149–152.
 227

36. Calvo, M.S.; Whiting, S.J.; Barton, C.N. Vitamin D fortification in the United States and Canada:
Current status and data needs. Am. J. Clin. Nutr. 2004, 80, 1710S–1716S.
37. O’Dea, J.A. Consumption of nutritional supplements among adolescents: Usage and perceived
benefits. Health Educ. Res. 2003, 18, 98–107.
38. Dorsch, K.D.; Bell, A. Dietary supplement use in adolescents. Curr. Opin. Pediatr. 2005, 17,
653–657.
39. Murphy, S.P.; White, K.K.; Park, S.Y.; Sharma, S. Multivitamin-multimineral supplements’ effect
on total nutrient inatake. Am. J. Clin. Nutr. 2007, 85, 280S–284S.
40. Bailey, R.L.; Fulgoni, V.L.; Keast, D.R.; Lentino, C.V.; Dwyer, J.T. Do dietary supplements
improve micronutrient sufficiency in children and adolescents? J. Pediatr. 2012, 161, 837–842.
41. Food Standards Australia New Zealand. Australian User Guide: Mandatory Folic Acid
Fortification. Implementing the Requirements of the Mandatory Fortification with Folic Acid
under Standard 2.1.1—Cereals and Cereal Products, February 2009. Available online:
http://www.foodstandards.gov.au/code/userguide/Documents/Mandatory%20Folic%20Acid%20F
ortification%20User%20Guide.pdf (accessed on 29 October 2013).
42. Nowson, C. Fortification of the Food Supply with Vitamin D: A Public Health Issue for
Australians. Available online: http://www.deakin.edu.au/health/cpan/VitaminDfortification_final.pdf
(accessed on 25 October 2013).

© 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).
228
 229

Reprinted from Nutrients. Cite as: Yakub, M.; Schulze, K.J.; Khatry, S.K.; Stewart, C.P.; Christian, P.;
West, K.P., Jr. High Plasma Homocysteine Increases Risk of Metabolic Syndrome in 6 to 8 Year Old
Children in Rural Nepal. Nutrients 2014, 6, 1649-1661.

High Plasma Homocysteine Increases Risk of Metabolic

Syndrome in 6 to 8 Year Old Children in Rural Nepal
Mohsin Yakub 1, Kerry J. Schulze 1,*, Subarna K. Khatry 2, Christine P. Stewart 3,
Parul Christian 1 and Keith P. West, Jr. 1
1
Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore,
MD 21205, USA; E-Mails: myakub1@jhu.edu (M.Y); pchristi@jhsph.edu (P.C);
kwest@jhsph.edu (K.P.W)
2
Nepal Nutrition Intervention Project-Sarlahi, Tripureswor, Kathmandu 45104, Nepal;
E-Mail: skk@mos.com.np
3
Program in International and Community Nutrition, University of California, Davis, CA 95616,
USA; E-Mail: cpstewart@ucdavis.edu

* Author to whom correspondence should be addresses; E-Mail: kschulz1@jhu.edu;

Tel.: +1-410-955-2794.

Received: 18 December 2013; in revised form: 21 March 2014 / Accepted: 2 April 2014 /
Published: 21 April 2014

Abstract: Little attention has been given to the association of plasma homocysteine (Hcy)
and metabolic syndrome (MetS) in children. We have evaluated the risk of MetS with
plasma Hcy in a cohort of 6 to 8 year old rural Nepalese children, born to mothers who had
participated in an antenatal micronutrient supplementation trial. We assessed Hcy in
plasma from a random selection of n = 1000 children and determined the relationship of
HOHYDWHG +F\ ! ȝPRO/  WR 0HW6 defined as the presence of any three of the
following: abdominal adiposity (waist circumference • WK SHUFHQWLOH RI WKH VWXG\
population), high plasma glucose•WK ( SHUFHQWLOH  KLJK V\VWROLF RU GLDVWROLF EORRG
pressure (•WK SHUFHQWLOH RI UHIHUHQFH SRSXODWLRQ  WULJO\FHULGH •  PPRO/ DQG KLJK
density lipoprotein < 0.9 mmol/L.) and its components. There was an increased risk of low
high-density lipoproteins (HDL), [odds ratios (OR) = 1.77, 95% confidence intervals
(CI) = 1.08–2.88; p = 0.020], high blood pressure [OR = 1.60, 95% CI = 1.10–2.46;
p = 0.015] and high body mass index (BMI) [OR = 1.98, 95% CI = 1.33–2.96; p = 0.001]
with elevated Hcy. We observed an increased risk of MetS (OR = 1.75, 95% CI = 1.06–
2.90; p = 0.029) with elevated Hcy in age and gender-adjusted logistic regression models.
 230

High plasma Hcy is associated with increased risk of MetS and may have implications for
chronic disease later in life.

Keywords: metabolic syndrome; homocysteine; Nepal

1. Introduction

Metabolic syndrome (MetS) is a complex disorder comprising abdominal adiposity, high-blood

pressure (BP), plasma glucose (PG), dyslipidemia [high-plasma triglycerides (TG) and/or low
concentrations of high-density lipoproteins (HDL)]. Insulin resistance and cardiovascular disease
(CVD) have gained attention as major manifestations of the syndrome. MetS has been considered an
illness of adulthood; however an increase in the prevalence of insulin resistance and MetS has been
reported among children recently [1,2]. According to a systematic review by Friend et al., the median
prevalence of MetS was 3.3% and 22% among Far East (India, South Korea and China) non-obese and
obese children, respectively [3]. Given global trends toward increased adiposity, obesity and diabetes,
deaths due to outcomes related to MetS, such as coronary heart disease (CHD) and type-2 diabetes, are
expected to rise across the age spectrum [4].
In low income societies, the incidence in MetS among children has also been associated with a
pattern of intrauterine conditions leading to low birth weight (<2500 g, LBW), rapid postnatal weight
gain and less frequent breastfeeding during early life [5,6]. Given a high prevalence of LBW, rural
Asian populations experiencing the nutrition transition to diets higher in fat and processed foods [7]
may, therefore, be at particular risk of developing MetS. In the terai of southern Nepal, longitudinal
studies have shown the incidence of LBW to be 43% [8] and the prevalence of MetS to be 11.7% at
6–8 years of age [9], with a lower risk among children whose mothers were provided antenatal folic
acid supplements [9]. The research described here examines homocysteine (Hcy) as a risk factor for
the MetS that was observed in this rural setting.
Homocysteine, an intermediary product of methionine metabolism, is elevated in folate and vitamin
B12 deficiencies [10]. Hcy has also been accepted widely as an independent risk factor for CVD such
as CHD and stroke [11,12]. Moreover, Hcy has been shown to be a thrombogenic and atherogenic
substrate that potentiates atherosclerotic phenomena that may lead to adverse cardiometabolic events [13].
Despite its link with CVD, whether Hcy is more broadly associated with MetS is not well
established. It is known that atherosclerosis begins in childhood and CVD can occur if children and
adolescents have earlier exposure to risk factors such as elevated Hcy and components of MetS.
However, to the best of our knowledge, no study has yet been done to assess the relationship of Hcy to
MetS among young children living in rural South Asia. Detection of risk factors for MetS early in life
in South Asian populations may suggest means to attenuate the progression of MetS and possible
development of type-2 diabetes and CVD later in life [14,15].
Our hypothesis in the present study was that MetS and its components such as excess adiposity,
hypertension, dyslipidemia and high glucose are associated with increased plasma Hcy concentrations
in children. We addressed this question by determining the prevalence of hyperhomocysteinemia and
evaluating its associations with aspects of MetS in a cohort of 6–8 year old children in rural Nepal.
 231

2. Experimental Section

Children born to mothers who participated in a 5-arm trial of antenatal (to 3 months postpartum)
micronutrient supplementation (1999–2001) in the rural District of Sarlahi, Nepal [8], were revisited
and consented in 2006–2008 at 6–8 years of age as part of a large cohort follow-up study [9].
Procedures and primary outcomes of the original trial [8,16] and follow-up assessment [9] have been
previously reported. Briefly, from 2006 to 2008, 3524 of 4130 children who were born to mothers
enrolled in the micronutrient supplementation trial [8,16] were revisited in their homes to reassess vital
status, height (measured by stadiometer; Harpenden, Croswell, UK), weight (measured by electronic
scale; Model 881, Seca, Cambridge, MD, USA) and BP, and a 10 mL venous blood sample was
collected in a heparinized tube in 3305 children, two-thirds of whom reported being fasted. For BP, the
mean of last three values out of four measurements collected at one minute intervals using an
automated oscillometric device (BpTRU™ BPM-300 Medical Devices Ltd., Coquitlam, BC, Canada)
was taken. Standard test kits (DCA 2000 analyzer MN, Bayer HealthCare LLC, Elkhart, IN, USA)
were used to estimate glycosylated hemoglobin (HbA1c) from whole blood. Plasma was separated in a
field laboratory and total cholesterol, HDL-cholesterol, TG and glucose concentrations were measured
using a Cholestech LDX analyzer (Cholestech Corp., Haywood, CA, USA). Low density lipoprotein
(LDL) cholesterol was estimated using Friedewald’s formula [17]. Frozen plasma was shipped to
Johns Hopkins University Bloomberg School of Public Health in liquid nitrogen for further analysis,
including plasma insulin, in fasted children only, by ultrasensitive sandwich immunoassay (ALPCO
Diagnostic, Salem, NH, USA). Data from these laboratory analyses have been reported previously [9].
Based on criteria of having multiple aliquots of plasma and complete data collected from the initial
trial and follow up activity, 2130 children were identified as eligible for further biomarker analysis,
forming the sampling frame for this analysis. This group of children was similar to the 1394 children
without sufficient plasma aliquots or complete data on a range of personal and household
characteristics [9]. From the 2130 children we randomly selected 1000 children (511 boys and 489 girls),
balanced across maternal supplementation groups (n = 200 children per group evenly distributed over
the duration of the field activity), for a subsequent study of micronutrient status. Hcy was measured in
this subset of samples by chemiluminescent immunoassay (Immulite 1000, Siemens Diagnostics, Los
Angeles, CA, USA), along with other measures of inflammation and micronutrient status [18]. Out of
the 1000 participating children, 30.8% children had cholesterol, 7.2% had HDL, 6.0% had TG, and
3.2% had glucose concentrations that were below the detectable limits for the Cholestech assay. For all
those who had values below the detectable limits we used the minimum detectable value of that
biomarker as an estimate of biomarker concentration, allowing us to retain complete data in our
analyses, but thereby overestimating the actual concentrations of those analytes for those children.
LDL was not calculated when other lipid data were out of range. Also, 324 children had not adhered to
fasting instructions, in whom insulin was not measured and therefore HOMA not determined. Data for
lipids and glucose were analyzed from both fasted and non-fasted participants and included in the final
analysis as we have reported previously [9].
Ethical clearance of the study was obtained from Institutional Review Boards at the Johns Hopkins
Bloomberg School of Public Health (protocol H.22.06.05.26.A2, 20 September 2006) and the Ethics
 232

Review Committee at Institute of Medicine at Katmandu, Nepal. The study was conducted according
to the principles of the Declaration of Helsinki.

2.1. Variable Definitions

The definition of hyperhomocysteinemia in healthy children and adolescents is not well defined in
WKHOLWHUDWXUH&XWRIIVKDYHYDULHGIURPWRȝPRO/DQGGLIIHUE\DJHDQGHWKQLFLW\>–22].
In adults, hyperhomocysteinemia is commonly defined as FRQFHQWUDWLRQVJUHDWHUWKDQȝPRO/>@
RUȝPRO/>@%HFDXVHWKHUHLVQRVSHFLILFFXW-off for defining high plasma homocysteine level
in children, population-specific 85th percentile values may guide the definition of hyperhomocysteinemia
for FKLOGUHQ >@ ,Q WKLV DQDO\VLV WKH WK SHUFHQWLOH RI +F\ ZDV  ȝPRO/ FORVH WR WKH ZLGHO\
accepted cutoff for hyperhomocysteinemia in adults, and used in previously published reports [22].
7KHUHIRUHZHFKRVHDFXWRIIRIȝPRO/WRLQGLFDWHhyperhomocysteinemia, despite lower cutoffs
sometimes used in children, because it better reflected the risk of higher homocysteine observed in
Asian populations [22] and because it is conventionally used. We have established and reported
previously the cutoffs used for other components of MetS and the definition of MetS [9,25–28], which
are summarized in Table 1.

Table 1. Variable definitions.

Variables Definitions
Hyperhomocysteinemia !ȝPRO/
Adiposity BMI: •WKSHUFHQWLOHRIWKHHQWLUHVWXG\SRSXODWLRQ REVHUYHGLQ
n = 150/1000 children)
Waist circumference: •WKSHUFHQWLOHRIWKHHQWLUHVWXG\SRSXODWLRQ REVHUYHG
in n = 153/1000 children)
Hypertension Systolic blood pressure or diastolic BP •WKpercentile of the U.S. reference
population adjusted for age, height and sex [25]
Dyslipidemia TG •PPRO/ (1)
HDL cholesterol < 0.9 mmol/L (2)
Insulin Resistance Homeostasis model assessment (HOMA) [to estimate insulin resistance]:
Product of FPI (mU/L) and PG (mmol/L) standard factor 22.5;
HOMA = (FPIxPG)/22.5 [28]
PG: (•WKSHUFHQWLOHRIWKHVWXG\SRSXODWLRQGHWHUPLQHGLQ
n = 150/1000 children)
MetS Presence of any three of the following constituents: elevated waist circumference,
high PG, high systolic or diastolic BP, high TG and low HDL [9]
(1)
As described by the NCEP criteria set for adults, because there is no separate recommendation for
(2)
children [26]; As described by the NCEP criteria for cholesterol in children and adolescents [27]; BMI,
basal metabolic rate; BP, blood pressure; TG, triglyceride; HDL, high density lipoprotein; PG, plasma
glucose; FPI, fasting plasma insulin; MetS, metabolic syndrome.

2.2. Statistical Analysis

Anthropometric measures, BP and biomarkers were examined by Hcy ( ”ȝPRO/

vs. !ȝPRO/
using independent sample t-tests. Values were expressed as mean ± SD.
 233

The association of hyperhomocysteinemia (>12.0 ȝPRO/ ZLWK0HW6DQGGLFKRWRPL]HGLQGLYLGXDO

components of MetS was expressed as odds ratios (OR) determined by separate logistic regressions for
each outcome variable, with adjustment for age and gender. We also estimated OR for the combined
risk of having multiple MetS components (such as low HDL and high BMI combined) against
hyperhomocysteinemia through multiple logistic regression. Since we had previously shown an effect
of maternal antenatal micronutrient supplementation, particularly with folic acid, on MetS in these
children [9], we also initially adjusted our regression models for maternal intervention groups and birth
weight, but this adjustment is not reported as these variables were not statistically significant. Likewise
we examined statistical models adjusted for various aspects of socioeconomic status (SES), including
ownership of televisions, radios, bicycles, livestock and use of electricity, as well as for seasonal
effects. However, adjustment for these variables is not reported as their influence was not statistically
significant. Finally, we explored models adjusted for fasting status and observed no effect on the OR;
thus, that adjustment was not included in our results. Risks were expressed as OR and associated 95%
confidence intervals (CI). All statistical analyses were done with IBM SPSS® (Statistical Package for
Social Sciences, IBM Corp., Armonk, NY, USA) software version 21 for Windows®.

3. Results

The prevalence of underweight, stunting and low BMI (below-2 Z-score) [29] in these children was
48.2%, 42.0% and 16.1% respectively. The prevalence of hyperhomocysteinemia was 18.4%. Among
the 1000 children, 827 (83%) had low HDL cholesterol (<0.9 mmol/L), 108 (11%) had high
triglyceride (• mmol/L), 190 (19%) had high blood pressure • ( 0th percentile of reference),
153 (15%) had high waist circumference•WK ( SHUFHQWLOH  $ WRWDO RI    FKLOGUHQ PHW WKH
criteria for MetS.
7DEOHLQGLFDWHVWKDWFKLOGUHQZLWKHOHYDWHG+F\ !ȝPRO/ KDGKLJKHU%0, p = 0.002), waist
circumference (p = 0.043), systolic (p = 0.071) and diastolic (p = 0.052) blood pressure, and
triglycerides (p = 0.016) and total lipid concentrations (p = 0.048). HDL concentrations were lower in
subjects with higher vs. lower Hcy levels (p = 0.031). In logistic regression analyses adjusted for age
and sex (Table 3), an elevated Hcy level was associated with an increased risk of low HDL cholesterol
[OR = 1.77 (95% CI = 1.08–2.88); p = 0.022], high BP [OR = 1.65 (95% CI = 1.10–2.46); p = 0.015]
and MetS [OR = 1.75 (95% CI = 1.06–2.90); p = 0.029]. No association was observed between
elevated Hcy and high TG, waist circumference or PG.
We detected an increased risk of high BMI (•WKSHUFHQWLOH [OR = 1.98 (95% CI = 1.33–2.96);
p = 0.001] with hyperhomocysteinemia. Moreover, we observed that the relative odds of having
combined low HDL cholesterol and high BP was 1.76 (95% CI = 1.14–2.70; p = 0.010). Similarly the
odds ratio for combined low HDL cholesterol and high BMI was 2.31 (95% CI = 1.51–3.53;
p < 0.001). We didn’t see additional risk for combined low HDL, high BP and high BMI with
hyperhomocysteinemia (p = 0.351) (Table 3).
 234

Table 2. Anthropometric measures, blood pressure, and plasma biochemical biomarkers of

study children by plasma homocysteine levels 1.
Total Homocysteine ”ȝPRO/ Homocysteine !ȝPRO/
Variables p-Value 2
n = 1000 n = 813 n = 184
Age 7.48 ± 0.65 7.47 ± 0.44 7.52 ± 0.40 0.24
Hcy 9.40 ± 3.50 8.06 ± 1.99 15.02 ± 2.97 <0.001
BMI 3 14.02 ± 1.04 13.97 ± 1.03 14.24 ± 1.05 0.002
3
Waist Circumference (cm) 51.40 ± 3.06 51.31 ± 3.13 51.81 ± 2.72 0.043
3
Systolic BP (mmHg) 95.2 ± 8.3 95.0 ± 8.1 96.2 ± 9.0 0.071
Diastolic BP 3 (mmHg) 63.8 ± 8.5 63.6 ± 8.1 64.9 ± 9.8 0.052
Total Cholesterol (mmol/L) 3.01 ± 0.48 3.01 ± 0.49 3.0 ± 0.45 0.760
TG (mmol/L) 2.55 ± 1.09 2.51 ± 1.06 2.73 ± 1.20 0.016
Total lipids (mmol/L) 5.56 ± 1.26 5.53 ± 1.22 5.73 ± 1.40 0.048
LDL (mmol/L) 1.92 ± 0.43 1.91 ± 0.45 1.92 ± 0.36 0.940
HDL (mmol/L) 0.71 ± 0.22 0.72 ± 0.23 0.68 ± 0.20 0.031
PG (mmol/L) 3.99 ± 1.06 4.00 ± 1.11 3.97 ± 0.77 0.783
FPI (pmol/L) 22.56 ± 23.76 22.62 ± 23.58 22.20 ± 25.02 0.847
HbA1c (%) 5.11 ± 0.27 5.12 ± 0.28 5.09 ± 0.25 0.167
1 2
Values are means±SD; p compares mean values in Homocysteine
”( ȝPRO/  vs. Homocysteine
3
! ȝPRO/  JURXSV XVLQJ LQGHSHQGHQW VDPSOH t-test; Data were missing for BMI (n = 1), waist circumference
(n = 3), systolic blood pressure (n = 8), diastolic blood pressure (n = 8), LDL (n = 367) and FPI (n = 324); BMI, body
mass index; TG, triglyceride; LDL, low density lipoprotein; HDL, high density lipoprotein; PG, plasma glucose; FPI,
fasting plasma insulin; HbA1c, glycosylate hemoglobin.

Table 3. Risk of MetS and its components in 6–8 years old children related to
hyperhomocysteinemia (n = 1000) 1.
Homocysteine !ȝPRO/
Outcome p-Value
OR (95% CI)
MetS 2 1.75 (1.06–2.90) 0.029
Low HDL (<0.9 mmol/L) 1.77 (1.08–2.88) 0.022
High TG (• mmol/L) 1.31 (0.80–2.13) 0.276
High systolic OR diastolic BP (•WKSHUFHQWLOH 1.65 (1.10–2.46) 0.015
High waist circumference (•WKSHUFHQWLOH 0.98 (0.63–1.53) 0.982
High PG (•WKSHUFHQWLOH 1.26 (0.83–1.94) 0.275
High BMI (•WKSHUFHQWLOH 1.98 (1.33–2.96) 0.001
High HOMA (•WKSHUFHQWLOH 1.26 (0.73–2.16) 0.401
High BP + High BMI 1.36 (0.54–0.346) 0.512
Low HDL + High BP + High BMI 1.63 (0.58–4.60) 0.351
Low HDL + High BP 1.76 (1.14–2.70) 0.010
Low HDL + High BMI 2.31 (1.51–3.53) <0.001
1
Values are OR (95% CI) based on separate logistic regression analyses for each set of outcomes, adjusted
for age and gender; 2 MetS defined as presence of any three of the following constituents: elevated waist
circumference, high PG, high systolic or diastolic BP, high TG and low HDL.
 235

4. Discussion

We observed an increased risk of MetS and its components, specifically dyslipidemia, BMI, and
hypertension, with elevated Hcy in young school-aged children in rural Nepal. The distributions of
these components of MetS were similar to those reported previously for the larger cohort of
3524 children from whom this sample was derived [9]. That study identified intrauterine exposure to
folate as a factor that ameliorated the risk of MetS in childhood; the current study expands on those
findings by identifying a more proximal potential risk factor for MetS.
Previous studies have revealed that individuals who meet the criteria of MetS are at increased
risk for developing CHD and diabetes mellitus [30], both of which are increasing in South Asia. It has
been reported that 20%–25% of South Asians have MetS, with the frequency expected to rise in
the future [31]. Nepal is not an exception, with CHD considered the leading cause of death in 2010
followed by stroke, hypertension and diabetes mellitus among other major causes of death [32].
However, data for these reports have been generated from urban areas with no study, so far, carried out
to establish the burden of CVD in rural Nepal. Shaik, et al. collected hospital admission data in one of
the tertiary hospitals which served the terai of Nepal and observed that around two patients out of ten
were admitted due to stroke [33]. The extent of CVD in the region surrounding Nepal is not different,
as it has been reported that 32% of all deaths in rural India are due to CVD [34].
Not limited to adults, South Asian children are also reported to have increased susceptibility to
MetS and CVD, as South Asian children living in the UK showed higher mean heart rate, elevated
mean triglyceride and fibrinogen levels, and higher mean fasting and post-glucose load insulin
concentrations compared to white children [35]. Moreover, evidence also suggests that Indian children
who are born small for gestational age have higher blood pressure, cholesterol levels (total cholesterol
and LDL-C), as well as increased adiposity at the age of eight years [36]. Our results are in line with
the above-mentioned studies, as we also observed high mean triglycerides and low mean HDL in these
children, who were nonetheless thin (prevalence of low BMI 16%), underweight and stunted (>40%
prevalence for both) [18]. Our data also suggest a role for elevated homocysteine with hypertension,
dyslipidemia, and adiposity.
Although Hcy concentrations have been investigated in children in Western populations [20,37], no
previous study has examined this risk factor in young South Asian children. The 95th percentile of Hcy
 ȝPRO/  PHDQ FRQFHQWUDWLRQ +F\ RI  ȝPRO/ DQG SUHYDOHQFH RI K\SHUKRPRF\VWHLQHPLD
(18.4%) observed in our study area reflect elevated levels of Hcy in rural Nepal compared to Western
populations [20,37]. A high burden of hyperhomocysteinemia in children could be a risk factor for
CVD in later life, as elevated Hcy (95th percentile) in Dutch children was associated with a subsequent
4-fold increased risk of ischemic cerebrovascular disease [38]. High levels of Hcy in children of rural
Nepal could be due to genetic polymorphisms, environmental exposures (e.g., high blood lead), physical
activity patterns, and diet, particularly one that may lead to low vitamin B12 concentrations [39].
Atherosclerosis is known to begin in childhood, as autopsy reports in children and young adults
with unexpected death have revealed positive associations between atherosclerotic lesions and risk
factors such as LDL-C, triglycerides, systolic and diastolic blood pressure, body mass index and
cigarette smoking, making it imperative to maintain healthy lipid profiles, blood pressure [40] and
plasma Hcy to minimize the burden of diseases like CVD and MetS. High levels of Hcy and high
 236

cholesterol are both associated with CVD risk. However, very few studies have reported the combined
effect of hyperhomocysteinemia and dyslipidemia on the risk of CVD [41]. Since we observed a high
prevalence of low HDL (83%) and considerable prevalence of hyperhomocysteinemia (18.5%) in these
children, a potential interaction between Hcy and HDL cholesterol seems clinically relevant. Although
exact mechanisms relating Hcy and cholesterol are not known, animal studies have revealed that
hypomethylation due to hyperhomocysteinemia could be accountable for lipid accumulation in
tissues [42]. Moreover, Hcy could also modulate activity of some inhibiting enzymes which play a role
in HDL-particle assembly [42].
The association we observed between Hcy and blood pressure also merits discussion. We observed
that an increase in Hcy was associated with rise in systolic and diastolic blood pressure, a finding
consistent with National Health and Nutrition Examination Survey (1994–1998) and other studies of
adult populations in developed countries [43,44]. Possible mechanisms to explain a link of Hcy with
increased blood pressure are Hcy-induced arteriolar constriction [45] vascular endothelial damage [46]
and decreased vasodilator responsiveness [47].
To the best of our knowledge no previous study has addressed the relationship of Hcy with MetS or
components of MetS in children. Those studies which have addressed the relationship of Hcy with
MetS in adults present contradictory results. Relationships we observed were similar to data published
by Kang et al. showing a positive association between Hcy and triglycerides, BMI and systolic and
diastolic blood pressure and a negative association with HDL in Korean adults [48]. Likewise, Hcy
was positively associated with waist circumference, BMI, blood pressure, LDL-C, and triglycerides,
but inversely associated with HDL in a Chinese sample of 1680 adults [49]. On the other hand, there
are published reports where authors did not find associations [50] or observed associations with few
components of MetS [51]. A lack of association of hyperhomocysteinemia with insulin resistance is
consistent with other reports [52,53].
This study had some limitations. Due to the cross-sectional design we cannot establish a temporal
or causal relationship between Hcy and MetS. The elevated Hcy and MetS could be results of common
pathways such as poor diet, environment, inadequate physical activity [54,55], and genetic
polymorphisms. Since no stringent range has been set to define hyperhomocysteinemia in children, our
definition of hyperhomocysteinemia may underestimate the prevalence of hyperhomocysteinemia as
WKHFXWRII!ȝPRO/LVJHQHUDOO\XVHGIRUDGXOWV0RUHRYHUZHsomewhat overestimated the actual
mean of total cholesterol, HDL, TG and plasma glucose by assigning values for these analytes
equivalent to the lowest detectable concentration of these analytes, although prevalence of abnormal
findings would not have been affected We believe that this study provides unique information about
the high prevalence of hyperhomocysteinemia and its role as a risk factor for MetS in a population of
otherwise undernourished children residing in rural Nepal. Our findings reveal important issues to be
considered further on the relationship of Hcy and MetS in a population which is known to have a high
burden of CVD in adulthood.

5. Conclusions

We conclude that hyperhomocysteinemia exerts risk towards development of MetS and that high
prevalence of hyperhomocysteinemia and MetS in this low income population suggest that Nepalese
 237

children are at a greater risk of developing CVD and diabetes in future. It is therefore necessary to
understand the causes of hyperhomocysteinemia, such as dietary habits in this population, in order to
attenuate the future development of diabetes and cardiovascular disease symptoms, both of which
constitute a growing concern among populations of South Asia.

Acknowledgments

This study was supported by Grants #GH614 and OPPGH5241 from the Bill & Melinda Gates
Foundation, Seattle, WA, USA. The original Nepal Nutrition Intervention Project- Sarlahi (NNIPS)-2
trial was supported through the Vitamin A for Health Cooperative Agreement (HRN-A-00-97-00015-00)
between Johns Hopkins University and the Office of Health, Infectious Diseases and Nutrition, United
States Agency for International Development (USAID), Washington DC, with additional support from
the Sight and Life Research Institute, Baltimore, MD, USA and Basel, Switzerland.

Conflicts of Interest

The authors declare no conflict of interest.

References

1. Alberti, K.G.; Eckel, R.H.; Grundy, S.M.; Zimmet, P.Z.; Cleeman, J.I.; Donato, K.A.;
Fruchart, J.C.; James, W.P.; Loria, C.M.; Smith, S.C., Jr. Harmonizing the metabolic syndrome:
A joint interim statement of the International Diabetes Federation Task Force on Epidemiology
and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World
Heart Federation; International Atherosclerosis Society; and International Association for the
Study of Obesity. Circulation 2009, 120, 1640–1645.
2. Ogden, C.L.; Carroll, M.D.; Curtin, L.R.; McDowell, M.A.; Tabak, C.J.; Flegal, K.M.
Prevalence of overweight and obesity in the United States, 1999–2004. JAMA 2006, 295,
1549–1555.
3. Friend, A.; Craig, L.; Turner, S. The Prevalence of Metabolic Syndrome in Children:
A Systematic Review of the Literature. Metab. Syndr. Relat. Disord. 2013, 11, 71–80.
4. Alberti, G.; Zimmet, P.; Shaw, J.; Bloomgarden, Z.; Kaufman, F.; Silink, M. Type 2 diabetes in
the young: The evolving epidemic: The international diabetes federation consensus workshop.
Diabetese Care 2004, 27, 1798–1811.
5. Eriksson, J.G.; Osmond, C.; Kajantie, E.; Forsen, T.J.; Barker, D.J. Patterns of growth among
children who later develop type 2 diabetes or its risk factors. Diabetologia 2006, 49, 2853–2858.
6. Khuc, K.; Blanco, E.; Burrows, R.; Reyes, M.; Castillo, M.; Lozoff, B.; Gahagan, S.
Adolescent metabolic syndrome risk is increased with higher infancy weight gain and decreased
with longer breast feeding. Int. J. Pediatr. 2012, 2012, 478610.
7. Popkin, B.M.; Adair, L.S.; Ng, S.W. Global nutrition transition and the pandemic of obesity in
developing countries. Nutr. Rev. 2012, 70, 3–21.
 238

8. Christian, P.; Khatry, S.K.; Katz, J.; Pradhan, E.K.; LeClerq, S.C.; Shrestha, S.R.; Adhikari, R.K.;
Sommer, A.; West, K.P., Jr. Effects of alternative maternal micronutrient supplements on low
birth weight in rural Nepal: Double blind randomised community trial. BMJ 2003, 326, 571.
9. Stewart, C.P.; Christian, P.; Schulze, K.J.; Leclerq, S.C.; West, K.P., Jr.; Khatry, S.K.
Antenatal micronutrient supplementation reduces metabolic syndrome in 6- to 8-year-old children
in rural Nepal. J. Nutr. 2009, 139, 1575–1581.
10. Castro, R.; Rivera, I.; Blom, H.J.; Jakobs, C.; Tavares de Almeida, I. Homocysteine metabolism,
hyperhomocysteinaemia and vascular disease: An overview. J. Inherit. Metab. Dis. 2006, 29,
3–20.
11. Selhub, J. Public health significance of elevated homocysteine. Food Nutr. Bull. 2008, 29,
S116–S125.
12. Wang, X.; Qin, X.; Demirtas, H.; Li, J.; Mao, G.; Huo, Y. Efficacy of folic acid supplementation
in stroke prevention: A meta-analysis. Lancet 2007, 369, 1876–1882.
13. Guilland, J.C.; Favier, A.; Potier de Courcy, G.; Galan, P.; Hercberg, S. Hyperhomocysteinemia:
An independent risk factor or a simple marker of vascular disease? I. Basic data. Pathol. Biol.
2003, 51, 101–110.
14. Ghaffar, A.; Reddy, K.S.; Singhi, M. Burden of non-communicable diseases in South Asia.
BMJ 2004, 328, 807–810.
15. Goyal, A.; Yusuf, S. The burden of cardiovascular disease in the Indian subcontinent. Indian J.
Med. Res. 2006, 124, 235–244.
16. Christian, P.; West, K.P.; Khatry, S.K.; Leclerq, S.C.; Pradhan, E.K.; Katz, J.; Shrestha, S.R.;
Sommer, A. Effects of maternal micronutrient supplementation on fetal loss and infant mortality:
A cluster-randomized trial in Nepal. Am. J. Clin. Nutr. 2003, 78, 1194–1202.
17. Friedewald, W.T.; Levy, R.I.; Fredrickson, D.S. Estimation of the concentration of low-density
lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin. Chem.
1972, 18, 499–402.
18. Schulze, K.J.; Christian, P.; Wu, L.S.F.; Stewart, C.; Arguello, M.; Cui, H.; Nanayakkara-Bind, A.;
Khatry, S.K.; LeClerq, S.; West, K.P., Jr. Micronutrient deficiencies are common in 6–8 year old
children of rural Nepal, with prevalence estimates modestly affected by inflammation. J. Nutr.
2014, in press.
19. Osganian, S.K.; Stampfer, M.J.; Spiegelman, D.; Rimm, E.; Cutler, J.A.; Feldman, H.A.;
Montgomery, D.H.; Webber, L.S.; Lytle, L.A.; Bausserman, L.; et al. Distribution of and factors
associated with serum homocysteine levels in children: Child and Adolescent Trial for
Cardiovascular Health. JAMA 1999, 281, 1189–1196.
20. Papandreou, D.; Mavromichalis, I.; Makedou, A.; Rousso, I.; Arvanitidou, M. Reference range of
total serum homocysteine level and dietary indexes in healthy Greek schoolchildren aged
6–15 years. Br. J. Nutr. 2006, 96, 719–724.
21. Vilaseca, M.A.; Moyano, D.; Ferrer, I. Artuch, R. Total homocysteine in pediatric patients.
Clin. Chem. 1997, 43, 690–692.
22. Chang, J.B.; Chu, N.F.; Shen, M.H.; Wu, D.M.; Liang, Y.H.; Shieh, S.M. Determinants and
distributions of plasma total homocysteine concentrations among school children in Taiwan.
Eur. J. Epidemiol. 2003, 18, 33–38.
 239

23. Kang, S.S.; Wong, P.W.; Malinow, M.R. Hyperhomocyst(e)inemia as a risk factor for occlusive
vascular disease. Annu. Rev. Nutr. 1992, 12, 279–298.
24. Yakub, M.; Iqbal, M.P.; Iqbal, R. Dietary patterns are associated with hyperhomocysteinemia in
an urban Pakistani population. J. Nutr. 2010, 140, 1261–1266.
25. National High Blood Pressure Education Program Working Group on High Blood Pressure in
Children and Adolescents. The fourth report on the diagnosis, evaluation, and treatment of high
blood pressure in children and adolescents. Pediatrics 2004, 114, 555–576.
26. National Cholesterol Education Program Expert Panel. Detection Evaluation and Treatment of
High Blood Cholesterol in Adults (Adults Treatment Panel III); National Heart Lung and Blood
Institute: Bathesda, MD, USA, 2001.
27. NCEP Expert Panel on Blood Cholesterol Levels in Children and Adolescents. National
Cholesterol Education Program (NCEP): Highlights of the report of the Expert Panel on Blood
Cholesterol Levels in Children and Adolescents. Pediatrics 1992, 89, 495–501.
28. Matthews, D.R.; Hosker, J.P.; Rudenski, A.S.; Naylor, B.A.; Treacher, D.F.; Turner, R.C.
+RPHRVWDVLVPRGHODVVHVVPHQW,QVXOLQUHVLVWDQFHDQGȕ-cell function from fasting plasma glucose
and insulin concentrations in man. Diabetologia 1985, 28, 412–419.
29. World Health Organization, Growth reference 5–19 years. Available online:
http://www.who.int/growthref/en/ (accessed on 19 November 2013).
30. Vega, G.L. Results of Expert Meetings: Obesity and Cardiovascular Disease. Obesity, the
metabolic syndrome, and cardiovascular disease. Am. Heart. J. 2001, 142, 1108–1116.
31. Misra, A.; Misra, R.; Wijesuriya, M.; Banerjee, D. The metabolic syndrome in South Asians:
Continuing escalation & possible solutions. Indian J. Med. Res. 2007, 125, 345–354.
32. Health Profile: Nepal. Available online: http://www.worldlifeexpectancy.com/country-health-
profile/Nepal (accessed on 03 December 2013).
33. Shaik, M.M.; Loo, K.W.; Gan, S.H. Burden of stroke in Nepal. Int. J. Stroke 2012, 7, 517–520.
34. Joshi, R.; Cardona, M.; Iyengar, S.; Sukumar, A.; Raju, C.R.; Raju, K.R.; Reddy, K.S.; Lopez, A.;
Neal, B. Chronic diseases now a leading cause of death in rural India—Mortality data from the
Andhra Pradesh Rural Health Initiative. Int. J. Epidemiol. 2006, 5, 1522–1529.
35. Whincup, P.H.; Gilg, J.A.; Papacosta, O.; Seymour, C.; Miller, G.J.; Alberti, K.G.; Cook, D.G.
Early evidence of ethnic differences in cardiovascular risk: Cross sectional comparison of British
South Asian and white children. BMJ 2002, 324, 635.
36. Bavdekar, A.; Yajnik, C.S.; Fall, C.H.; Bapat, S.; Pandit, A.N.; Deshpande, V.; Bhave, S.;
Kellingray, S.D.; Joglekar, C. Insulin resistance syndrome in 8-year-old Indian children: Small at
birth, big at 8 years, or both? Diabetes 1999, 48, 2422–2429.
37. Ganji, V.; Kafai, M.R. Population references for plasma total homocysteine concentrations for
U.S. children and adolescents in the post-folic acid fortification era. J. Nutr. 2005, 135,
2253–2256.
38. Van Beynum, I.M.; Smeitink, J.A.; den Heijer, M.; te Poele Pothoff, M.T.; Blom, H.J.
Hyperhomocysteinemia: A risk factor for ischemic stroke in children. Circulation 1999, 99,
2070–2072.
39. Schulze, K.J.; Yakub, M.; Christian, P.; Stewart, C.P.; West, K.P. Prevalence and Determinants of
Hyperhomocysteinemia in School Aged Children in Rural Nepal. FASEB J. 2013, 27, 1055.24.
 240

40. Kavey, R.E.; Daniels, S.R.; Lauer, R.M.; Atkins, D.L.; Hayman, L.L.; Taubert, K.
American Heart Association guidelines for primary prevention of atherosclerotic cardiovascular
disease beginning in childhood. Circulation 2003, 107, 1562–1566.
41. Herrmann, W.; Obeid, R.; Hubner, U.; Jouma, M.; Geisel, J. Homocysteine in relation to
C-reactive protein and low-density lipoprotein cholesterol in assessment of cardiovascular risk.
Cell. Mol. Biol. 2004, 50, 895–901.
42. Obeid, R.; Herrmann, W. Homocysteine and lipids: S-adenosyl methionine as a key intermediate.
FEBS Lett. 2009, 583, 1215–1225.
43. Lim, U.; Cassano, P.A.; Homocysteine and blood pressure in the Third National Health and
Nutrition Examination Survey, 1988–1994. Am. J. Epidemiol. 2002, 156, 1105–1113.
44. Refsum, H.; Nurk, E.; Smith, A.D.; Ueland, P.M.; Gjesdal, C.G.; Bjelland, I.; Tverdal, A.;
Tell, G.S.; Nygard, O.; Vollset, S.E. The Hordaland Homocysteine Study: A community-based
study of homocysteine, its determinants, and associations with disease. J. Nutr. 2006, 136,
1731S–1740S.
45. Vermeulen, E.G.; Niessen, H.W.; Bogels, M.; Stehouwer, C.D.; Rauwerda, J.A.;
van Hinsbergh, V.W. Decreased smooth muscle cell/extracellular matrix ratio of media of femoral
artery in patients with atherosclerosis and hyperhomocysteinemia. Arterioscler. Thromb. Vasc.
Biol. 2001, 21, 573–577.
46. Harker, L.A.; Harlan, J.M.; Ross, R. Effect of sulfinpyrazone on homocysteine-induced
endothelial injury and arteriosclerosis in baboons. Circ. Res. 1983, 53, 731–739.
47. Mujumdar, V.S.; Aru, G.M.; Tyagi, S.C. Induction of oxidative stress by homocyst(e)ine impairs
endothelial function. J. Cell. Biochem. 2001, 82, 491–500.
48. Kang, J.Y.; Park, I.K.; Lee, J.Y.; Sung, S.H.; Chang, Y.K.; Park, Y.K.; Choi, T.I. Use of serum
homocysteine to predict cardiovascular disease in Korean men with or without metabolic
syndrome. JKMS 2012, 27, 500–505.
49. Feng, S.Q.; Ye, P.; Luo, L.M.; Xiao, W.K.; Xu, R.Y.; Wu, H.M. Relationship between serum
homocysteine and metabolic syndrome: A cross-sectional study. Zhonghua Liu Xing Bing Xue Za
Zhi 2012, 33, 256–259.
50. Vaya, A.; Carmona, P.; Badia, N.; Perez, R.; Hernandez M.A.; Corella, D. Homocysteine levels
and the metabolic syndrome in a Mediterranean population: A case-control study.
Clin. Hemorheol. Microcirc. 2011, 47, 59–66.
51. Nabipour, I.; Ebrahimi, A.; Jafari, S.M.; Vahdat, K.; Assadi, M.; Movahed, A.; Moradhaseli, F.;
Obeidi, N.; Sanjdideh, Z. The metabolic syndrome is not associated with homocysteinemia: The
Persian Gulf Healthy Heart Study. J. Endocrinol. Investig. 2009, 32, 406–410.
52. Meigs, J.B.; Jacques, P.F.; Selhub, J.; Singer, D.E.; Nathan, D.M.; Rifai, N.; D’Agostino, R.B., Sr.;
Wilson, P.W. Fasting plasma homocysteine levels in the insulin resistance syndrome: The
Framingham offspring study. Diabetes Care 2001, 24, 1403–1410.
53. Tanrikulu-Kilic, F.; Bekpinar, S.; Unlucerci, Y.; Orhan, Y. Insulin resistance is not related to
plasma homocysteine concentration in healthy premenopausal women. Physiol. Res. 2006, 55,
285–290.
 241

54. Dankner, R.; Chetrit, A.; Dror, G.K.; Sela, B.A. Physical activity is inversely associated with total
homocysteine levels, independent of C677T MTHFR genotype and plasma B vitamins.
Age 2007, 29, 219–227.
55. Bankoski, A.; Harris, T.B.; McClain, J.J.; Brychta, R.J.; Caserotti, P.; Chen, K.Y.; Berrigan, D.;
Troiano, R.P.; Koster, A. Sedentary activity associated with metabolic syndrome independent of
physical activity. Diabetes Care 2011, 34, 497–503.

© 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).
242
 243

Reprinted from Nutrients. Cite as: Woo, H.D.; Kim, D.W.; Hong, Y.; Kim, Y.; Seo, J.; Choe, B.M.;
Park, J.H.; Kang, J.; Yoo, J.; Chueh, H.W.; Lee, J.H.; Kwak, M.J.; Kim, J. Dietary Patterns in Children
with Attention Deficit/Hyperactivity Disorder (ADHD). Nutrients 2014, 6, 1539-1553.

Dietary Patterns in Children with Attention

Deficit/Hyperactivity Disorder (ADHD)
Hae Dong Woo 1, Dong Woo Kim 1, Young-Seoub Hong 2,3, Yu-Mi Kim 2,3, Ju-Hee Seo 3,
Byeong Moo Choe 4, Jae Hong Park 4, Je-Wook Kang 5, Jae-Ho Yoo 6, Hee Won Chueh 6,
Jung Hyun Lee 7, Min Jung Kwak 8 and Jeongseon Kim 1,*
1
Molecular Epidemiology Branch, National Cancer Center, 323 Ilsan-ro, Ilsandong-gu, Goyang-si,
Gyeonggi-do 410-769, Korea; E-Mails: eastsea93@hanmail.net (H.D.W.);
ellebass@gmail.com (D.W.K.)
2
Department of Preventive Medicine, College of Medicine, Dong-A University,
Dong-A University Hospital, 26, Daesingongwon-ro, Seo-gu, Busan 602-715, Korea;
E-Mails: yshong@dau.ac.kr (Y.-S.H.); kimyumi@dau.ac.kr (Y.-M.K.)
3
Heavy Metal Exposure Environmental Health Center, Dong-A University, 32, Daesingongwon-ro,
Seo-gu, Busan 602-714, Korea; E-Mail: juhui978@dau.ac.kr
4
Department of Psychiatry, College of Medicine, Dong-A University, Dong-A University Hospital,
26 Daesingongwon-ro, Seo-gu, Busan 602-715, Korea; E-Mails: bmchoe@dau.ac.kr (B.M.C.);
pjhkorea@hanmail.net (J.H.P.)
5
Department of Child and Adolescent Psychiatry, College of Medicine, Inje University Busan Paik
Hospital, 75 Bokji-ro, Busanjin-gu, Busan 614-735, Korea; E-Mail: forevery99@hanmail.net
6
Department of Pediatrics, College of Medicine, Dong-A University, Dong-A University Hospital,
26 Daesingongwon-ro, Seo-gu, Busan 602-715, Korea; E-Mails: pedendo@dau.ac.kr (J.-H.Y.);
hwchueh@dau.ac.kr (H.W.C.)
7
Department of Pediatrics, Kosin University Gospel Hospital, 262, Gamcheon-ro, Seo-gu,
Busan 602-702, Korea; E-Mail: agasoa@hanmail.net
8
Department of Pediatrics, Pusan National University Hospital, Pusan National University School of
Medicine, 179, Gudeok-ro, Seo-gu, Busan 602-739, Korea; E-Mail: glory0123@hanmail.net

* Author to whom correspondence should be addressed; E-Mail: jskim@ncc.re.kr;

Tel.: +82-31-920-2570; Fax: +82-31-920-2579.

Received: 12 February 2014; in revised from: 17 March 2014 / Accepted: 28 March 2014 /
Published: 14 April 2014
 244

Abstract: The role of diet in the behavior of children has been controversial, but the
association of several nutritional factors with childhood behavioral disorders has been
continually suggested. We conducted a case-control study to identify dietary patterns
associated with attention deficit hyperactivity disorder (ADHD). The study included
192 elementary school students aged seven to 12 years. Three non-consecutive 24-h recall
(HR) interviews were employed to assess dietary intake, and 32 predefined food groups
were considered in a principal components analysis (PCA). PCA identified four major
dietary patterns: the “traditional” pattern, the “seaweed-egg” pattern, the “traditional-healthy”
pattern, and the “snack” pattern. The traditional-healthy pattern is characterized by a diet
low in fat and high in carbohydrates as well as high intakes of fatty acids and minerals.
The multivariate-adjusted odds ratio (OR) of ADHD for the highest tertile of the
traditional-healthy pattern in comparison with the lowest tertile was 0.31 (95% CI:
0.12–0.79). The score of the snack pattern was positively associated with the risk of
ADHD, but a significant association was observed only in the second tertile. A significant
association between ADHD and the dietary pattern score was not found for the other two
dietary patterns. In conclusion, the traditional-healthy dietary pattern was associated with
lower odds having ADHD.

Keywords: dietary pattern; attention deficit/hyperactivity disorder (ADHD); school-aged

children; Korean

1. Introduction

Attention deficit hyperactivity disorder (ADHD) is one of the most commonly diagnosed
neurobehavioral disorders in childhood, and it often lasts into adulthood [1]. ADHD prevalence rates
vary by age, gender, and ethnicity [2,3]. Boys are more likely to have ADHD than girls, and higher
rates of ADHD in younger age groups have been observed in studies of children and adolescents [4].
Worldwide, the overall prevalence of ADHD/hyperkinetic disorder (HD) was found to be 5.29% in a
pooled analysis [2]. The prevalence of ADHD is 8.7% in US children aged eight to 15 years [5] and
9.7% in Iranian school-aged children [6]. In Korea, the prevalence of ADHD is 7.6% in elementary
school children with a mean age of 9.4 years [7] and upper-grade elementary school children with a
mean age of 11.6 years [8]. The etiology of ADHD is multifactorial, and both genetic and
environmental factors may be involved in ADHD [9]. Family and twin studies have shown that genes
play an important role in the development of ADHD. Genome-wide association studies are
inconclusive, but candidate gene studies suggest the involvement of genes related to the receptors and
transporters of dopamine and serotonin [10,11]. Proposed ADHD environmental risk factors include
heavy metal and chemical exposures such as lead, mercury, organochlorine, organophosphates, and
phthalates, as well as nutritional and lifestyle/psychosocial factors [5].
The effect of diet and dietary supplements is unclear, but considerable evidence suggests that
dietary factors are associated with childhood behavioral disorders such as ADHD [12,13]. Low levels
of copper, iron, zinc, magnesium, and omega-3 fatty acids have been reported in children with ADHD,
 245

and sugar, artificial food colorings, and preservatives are associated with an increased risk of
ADHD [12,13]. Recently, the association between dietary pattern and ADHD has been examined
in several studies [6,20,21]. As nutrients are consumed in combination and because nutrients are highly
interrelated, the study of dietary patterns is useful to further understand the overall role of diet in
ADHD. Thus, the purpose of this study was to determine the association between various dietary
patterns and ADHD among Korean school-aged children.

2. Experimental Section

2.1. Study Population

We conducted a hospital-based case-control study using elementary school students who visited
several university hospitals in Busan, Korea, from April to September, 2013. ADHD cases were
recruited from two university hospitals (Dong-A and Inje University). ADHD was diagnosed by
psychiatrists based on the Diagnostic and Statistical Manual of Mental Disorders-Fourth Edition
(DSM-IV). Some children with ADHD have concurrent condition such as tic disorder (motor type),
anxiety disorder, oppositional defiant disorder, Tourette’s disorder, depression, and learning disability.
A total of 117 cases, which consented to participate in research, were recruited, and age- and
sex-matched controls were recruited from three university hospitals (Dong-A, Pusan, and Kosin
University). Controls who did not have severe chronic diseases, a history of ADHD diagnosis and any
related disease, such as mental disorder and tic disorder were recruited. Additional test using ADHD
Rating Scale (ARS) for controls was performed to exclude ADHD cases. After excluding
seven participants who did not complete the questionnaire, a total of 202 controls were recruited.
To exclude the seasonal variation in dietary intake, the dietary survey season was also matched in the
analysis. Frequency matching by grade (two years), sex, and season (three months) was conducted.
A total of 192 elementary school students aged seven to 12 years (96 students with ADHD and
96 healthy controls) were finally selected. Each participant and their legal guardian were provided with
an informed consent form according to the procedures approved by the Institutional Review Board of
the National Cancer Center.

2.2. Data Collection

The legal guardians of the participants were asked to complete a self-administered questionnaire,
which was used to gather information on demographics, lifestyle, and the medical histories of the
participants and their parents. A trained interviewer facilitated the 24-h recalls (24HR) interviews
face-to-face, and another two non-consecutive 24HR interviews were conducted by telephone between
April and September 2013. Individual food intake was calculated using CAN-PRO 4.0 (Computer
Aided Nutritional Analysis Program, The Korean Nutrition Society, Seoul, Korea). Mercury and lead
exposure from food was calculated using dietary consumption data and their concentrations in
118 core food items. Consumption of omega-3 fatty acids was estimated as the sum of
eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA).
 246

2.3. Statistical Analysis

Principal-components analysis (PROC FACTOR) was used to extract the participants’ dietary
patterns using 32 predefined food groups. We used a varimax rotation to enhance the interpretability of
the analyzed factors. We determined how many factors to retain after evaluating the eigenvalue, scree
test, and interpretability. The dietary patterns were named according to the factors with the highest
scores among the defined food groups for each dietary factor. Each dietary pattern’s factor score was
categorized by tertile for further analysis. Using a Student t-test for continuous variables and a
chi-square test for categorical variables, we compared the general characteristics between students with
ADHD and controls. The trend test was performed to analyze the associations between each of the
dietary patterns and ADHD using a generalized linear model with adjustments for total energy intake.
Odds ratios (ORs) and 95% confidence intervals (CIs) for ADHD were calculated across the tertiles of
dietary pattern scores using logistic regression models. The lowest tertile of each dietary pattern was
used as the reference. To assess the trend across the tertiles, we assigned median values to each tertile
of the dietary pattern scores as a continuous variable. We performed the statistical analysis using SAS
version 9.2 (SAS Institute Inc., Cary, NC, USA). All P values were two-WDLOHG Į  

3. Results

The general characteristics of the study population are presented in Table 1. The mean ages of the
controls and students with ADHD were 9.1 and 9.0 years, respectively. The total energy intake was higher
in controls than in students with ADHD (p = 0.008). Father’s educational background and occupation
significantly differed between ADHD students and controls (p < 0.001 and p = 0.001, respectively).

Table 1. General characteristics of study population 1.

Characteristics Controls (n = 96) Cases (n = 96) P
Age (year) 9.1 ± 1.8 9.0 ± 1.7
Sex, male (%) 65 (67.7) 65 (67.7)
Total energy intake (kcal) 2027.3 ± 381.7 1879.2 ± 380.7 0.008
Weight (kg) 34.3 ± 9.6 33.1 ± 10.4 0.427
Body mass index (cm/m2) 18.2 ± 2.9 17.5 ± 3.2 0.122
Gestation age (week) 39.0 ± 1.7 39.0 ± 1.7 0.787
Birth weight (kg) 3.3 ± 0.5 3.2 ± 0.5 0.099
Breastfeeding, yes (%) 80 (83.3) 70 (72.9) 0.081
Mother’s age (year) 38.6 ± 3.7 39.5 ± 4.1 0.133
Birth order 1.5 ± 0.7 1.4 ± 1.1 0.651
Father’s education, n (%)
<High school 22 (22.9) 45 (47.9) <0.001
College 53 (55.2) 41 (43.6)
Graduate school 21 (21.9) 8 (8.5)
Father’s occupation, n (%)
Professional 26 (27.1) 15 (15.6) 0.001
 247

Table 1. Cont.
Office/service worker 44 (45.8) 28 (29.2)
Manual worker 13 (13.5) 28 (29.2)
Other 13 (13.5) 25 (26.0)
Father’s smoking status, n (%)
Seldom or Never 37 (39.0) 24 (26.4) 0.188
Current 42 (44.2) 49 (53.9)
Former 16 (16.8) 18 (19.8)
1
All analyses were performed with the data matched for age, sex, and dietary survey season.

PCA identified four major dietary patterns among the 32 food groups, and the associated factor
loading scores with absolute values•DUHVKRZQLQ7DEOH7KH³WUDGLWLRQDO´GLHWDU\SDWWHUQZDV
characterized by high intakes of condiments, vegetables, tofu/soymilk, and mushrooms. The
“seaweed-egg” dietary pattern included high intakes of seaweeds, fats/oils, sweets, and eggs. The
“traditional-healthy” dietary pattern included high intakes of kimchi, grains, bonefish, and low intakes
of fast foods and beverages. The “snack” dietary pattern was characterized by high intakes of snacks
and processed meat and a low intake of noodles. Lean fish, other seafood, and yogurt were not listed
due to their low factor loadings in all examined dietary patterns. Each dietary pattern explained 8.0%,
6.0%, 5.6%, and 5.4% of the variation in food intake, respectively.

Table 2. Factor loadings for the four major dietary patterns derived from principal
components analysis with orthogonal rotation.
Foods/Food Groups Traditional Seaweed-Egg Traditional-Healthy Snack
Condiments 0.75
Vegetables 0.56 0.20
Tofu, Soymilk 0.53
Mushrooms 0.49
Salted fermented seafood 0.34
Fruits 0.32 í í
Seaweeds 0.69
Fats, Oils 0.29 0.68
Sweets 0.27 0.43 0.33
Egg 0.41
Potatoes 0.22 0.35
Processed fruit products 0.33
Legumes 0.29
Kimchi 0.58 í
Grains 0.23 0.56
Bonefish 0.28 0.52 0.26
Fatty fish 0.29 0.23 í
Snack 0.49
Processed meats 0.44
Bread 0.43
Milk 0.30
 248

Table 2. Cont.
Shellfish í
Beverages 0.22 í
Fast foods í
Rice cake í
Seeds 0.23
Dairy products í í
Meats í í
Noodles 0.24 í
Variance of explained (%) 8.0 6.0 5.6 5.4
Factor loadings with absolute values •ZHUHOLVWHGLQWKHWDEOHDPRQJIRRGJURXSs.

The distribution of characteristics by dietary pattern score tertiles is presented in Table 3. Increasing
scores in the traditional and traditional-healthy patterns were correlated with a decreased percent
energy from fat (P for trend = 0.001; P for trend <0.001, respectively), whereas the percent energy
from carbohydrate increased as the score of the traditional-healthy pattern increased
(P for trend <0.001). Fatty acids were significantly associated with dietary pattern scores. The
traditional pattern score was associated with a high intake of total fatty acids; the seaweed-egg and
traditional-healthy pattern scores were associated with high intakes of PUFAs and omega-3 fatty acids,
whereas the snack pattern score was negatively associated with the intakes of total fatty acids, PUFAs,
and MUFAs. Regarding mineral intake, calcium intake was positively associated with the scores of the
traditional, traditional-healthy, and snack patterns, and iron was positively associated with the scores
of the traditional and traditional-healthy patterns. Heavy metal exposure via food consumption was
also assessed, and mercury was positively associated with the traditional, traditional-healthy, and snack
patterns; lead was positively associated with the traditional and snack patterns.
The ORs and 95% CIs of ADHD were analyzed across the tertiles of dietary pattern scores
(Table 4). The OR (95% CI) in the highest tertiles of the traditional dietary pattern compared to those
in the lowest tertiles in crude model was 0.29 (0.13–0.64), but a significant association was not
observed in multivariate model 2 (OR: 0.76, 95% CI: 0.26–2.24). The seaweed-egg pattern was not
significantly associated with ADHD in any of the models. The snack pattern score was positively
associated with the risk of ADHD, but a significant association was observed only in the second tertile
in crude model and multivariate model 1. Students in the highest tertile of the traditional-healthy
pattern score had an increased risk of ADHD in the multivariate-adjusted models when compared with
those in the lowest tertile (OR (95% CI): 0.32 (0.13–0.82) in multivariate model 1; 0.31 (0.12–0.79) in
multivariate model 2).
 249

Table 3. Distribution of characteristics by the tertiles of dietary pattern scores.

Traditional Seaweed-Egg
Characteristics 1 2
T1 T2 T3 P Trend T1 T2 T3 P Trend
Age (year) 8.7 (1.8) 9.1 (1.7) 9.5 (1.6) 0.014 8.9 (1.6) 9.0 (1.8) 9.2 (1.8) 0.270
Sex, female (%) 32 (43.2) 19 (25.7) 11 (25.0) 0.023 28 (40.6) 22 (33.9) 12 (20.7) 0.018
BMI (kg/m2) 17.9 (3.3) 17.6 (3.0) 18.3 (2.7) 0.533 17.6 (2.5) 17.7 (3.1) 18.4 (3.6) 0.171
Education, •college (%) 41 (56.9) 51 (68.9) 31 (70.5) 0.109 41 (59.4) 40 (63.5) 42 (72.4) 0.132
Total energy intake (kcal) 1757.7 (353) 2005.6 (325) 2194.2 (383) <0.001 1898.9 (418) 1897.2 (362) 2080.8 (352) 0.008
Carbohydrate (g) 248.6 (44.7) 284.5 (49.5) 315.4 (59.2) 0.013 273.0 (61.6) 272.3 (51.4) 289.5 (53.9) 0.216
Carbohydrate (% energy) 56.6 (6.5) 56.5 (6.2) 57.0 (4.9) 0.057 57.4 (6.3) 57.1 (6.0) 55.3 (5.5) 0.173
Protein (g) 67.1 (22.3) 78.2 (14.4) 86.0 (15.2) 0.198 71.0 (18.7) 74.0 (21.6) 83.2 (15.3) 0.016
Protein (% energy) 15.0 (3.0) 15.6 (2.2) 15.6 (1.7) 0.075 14.9 (2.2) 15.4 (2.8) 16.0 (2.2) 0.005
Fat (g) 56.9 (19.1) 63.3 (18.4) 68.1 (18.4) <0.001 59.7 (20.6) 59.4 (18.2) 67.3 (17.4) 0.951
Fat (% energy) 28.3 (5.1) 27.9 (5.1) 27.4 (4.4) 0.001 27.7 (5.3) 27.6 (4.9) 28.7 (4.6) 0.792
Total fatty acids (g) 29.2 (12.8) 33.1 (14.2) 33.2 (12.6) 0.031 28.9 (13.9) 30.9 (12.7) 35.7 (12.7) 0.080
PUFAs (g) 6.9 (2.5) 8.7 (3.4) 8.5 (2.7) 0.798 6.4 (2.6) 7.8 (2.1) 10.1 (3.2) <0.001
MUFAs (g) 10.9 (5.7) 12.2 (5.9) 12.4 (5.3) 0.057 10.8 (5.9) 11.5 (5.4) 13.1 (5.5) 0.245
Omega-3fatty acids (g) 0.10 (0.26) 0.23 (0.48) 0.34 (0.62) 0.134 0.25 (0.57) 0.20 (0.38) 0.15 (0.38) 0.059
Calcium (mg) 491.5 (222) 587.7 (171) 730.0 (224) 0.002 565.1 (225) 558.1 (230) 632.8 (206) 0.755
Iron (mg) 10.4 (2.3) 13.5 (4.8) 16.6 (8.4) <0.001 12.2 (7.3) 12.8 (5.3) 14.1 (3.4) 0.447
Zinc (mg) 8.2 (2.1) 9.8 (2.0) 10.5 (1.9) 0.112 8.9 (2.2) 9.1 (2.3) 10.2 (1.9) 0.063
Mercury (μg/kg bw) 0.19 (0.05) 0.22 (0.06) 0.22 (0.07) 0.027 0.20 (0.06) 0.21 (0.06) 0.21 (0.06) 0.965
Lead (μg/kg bw) 0.43 (0.14) 0.50 (0.14) 0.53 (0.17) 0.022 0.48 (0.17) 0.47 (0.13) 0.49 (0.16) 0.954
 250

Table 3. Cont.
Traditional-Healthy Snack
Characteristics
T1 T2 T3 P Trend T1 T2 T3 P Trend
Age (year) 9.2 (1.7) 8.8 (1.8) 9.1 (1.7) 0.746 9.8 (1.7) 8.8 (1.7) 8.8 (1.7) 0.002
Sex, female (%) 31 (41.9) 18 (29.0) 13 (23.2) 0.022 13 (25.0) 34 (42.5) 15 (25.0) 0.906
BMI (kg/m2) 17.8 (3.1) 17.8 (3.0) 18.0 (3.3) 0.644 19.5 (3.7) 17.4 (2.3) 17.0 (2.8) <0.001
Education, •FROOHJH  47 (63.5) 41 (67.2) 35 (63.6) 0.956 32 (61.5) 52 (66.7) 39 (65.0) 0.719
Total energy intake (kcal) 1946.2 (411) 1893.1 (391) 2029.3 (343) 0.225 2088.5 (361) 1812.7 (362) 2023.5 (387) 0.354
Carbohydrate (g) 266.5 (56.5) 272.8 (52.1) 298.1 (55.8) <0.001 292.7 (58.3) 260.5 (53.4) 287.7 (52.8) 0.566
Carbohydrate (% energy) 54.6 (6.0) 57.7 (5.9) 58.3 (5.5) <0.001 55.6 (6.1) 57.2 (6.3) 56.9 (5.5) 0.326
Protein (g) 77.4 (23.4) 71.8 (17.3) 77.9 (15.2) 0.294 82.6 (15.6) 71.2 (21.4) 75.8 (18.1) 0.079
Protein (% energy) 15.7 (3.1) 15.1 (1.9) 15.2 (1.9) 0.293 15.7 (2.4) 15.5 (2.8) 14.8 (1.9) 0.047
Fat (g) 65.7 (20.5) 58.5 (19.2) 60.6 (16.3) <0.001 67.7 (18.0) 56.0 (17.9) 64.7 (19.8) 0.872
Fat (% energy) 29.7 (4.6) 27.2 (5.0) 26.5 (4.6) <0.001 28.7 (5.1) 27.3 (5.0) 28.2 (4.6) 0.843
Total fatty acids (g) 30.4 (12.9) 30.6 (12.1) 34.3 (15.1) 0.244 38.3 (16.9) 28.3 (10.0) 30.3 (11.9) 0.001
PUFAs (g) 7.6 (3.1) 7.7 (2.8) 8.8 (3.0) 0.046 9.1 (3.2) 7.5 (2.5) 7.7 (3.3) 0.018
MUFAs (g) 11.3 (5.5) 11.4 (5.2) 12.8 (6.4) 0.300 14.8 (7.5) 10.4 (4.0) 10.9 (4.7) <0.001
Omega-3fatty acids (g) 0.13 (0.34) 0.18 (0.36) 0.33 (0.63) 0.024 0.29 (0.66) 0.19 (0.38) 0.14 (0.31) 0.124
Calcium (mg) 553.3 (227) 551.3 (195) 658.0 (231) 0.016 555.1 (221) 541.2 (206) 663.5 (227) <0.001
Iron (mg) 12.0 (4.1) 12.8 (5.0) 14.6 (7.6) 0.026 13.8 (3.8) 12.0 (4.6) 13.7 (7.8) 0.669
Zinc (mg) 9.3 (2.5) 9.2 (2.1) 9.7 (2.0) 0.909 10.1 (2.1) 8.7 (2.1) 9.6 (2.3) 0.293
Mercury (μg/kg bw) 0.19 (0.05) 0.21 (0.06) 0.23 (0.06) 0.003 0.19 (0.06) 0.21 (0.06) 0.22 (0.06) 0.001
Lead (μg/kg bw) 0.45 (0.13) 0.49 (0.15) 0.50 (0.18) 0.125 0.44 (0.16) 0.47 (0.14) 0.51 (0.16) 0.006
1 2
Tertiles of dietary pattern scores; P trend was calculated using generalized linear models for continuous variables and using Mantel–Haenszel chi-squared tests for categorical variables; P
trend of nutrient and metal consumption was adjusted for total energy intake; PUFAs: Polyunsaturated fatty acids, MUFAs: Monounsaturated fatty acids.
 251

Table 4. Distribution of characteristics by the tertiles of dietary pattern scores 1.

N Multivariate Multivariate
Dietary Pattern Crude Model
Control/Case Model 1 2 Model 2 3
T1 4 32/42 1 1 1
T2 32/42 1.00 (0.52–1.92) 1.32 (0.61–2.84) 1.88 (0.80–4.42)
Traditional
T3 32/12 0.29 (0.13–0.64) 0.43 (0.18–1.04) 0.76 (0.26–2.24)
5
P trend 0.003 0.072 0.615
T1 32/37 1 1 1
T2 32/33 0.89 (0.45–1.76) 0.66 (0.30–1.44) 0.70 (0.31–1.55)
Seaweed-egg
T3 32/26 0.70 (0.35–1.42) 0.64 (0.29–1.41) 0.84 (0.36–1.94)
P trend 0.321 0.271 0.682
T1 32/42 1 1 1
Traditional- T2 32/30 0.71 (0.36–1.41) 0.60 (0.27–1.32) 0.57 (0.25–1.29)
healthy T3 32/24 0.57 (0.28–1.15) 0.32 (0.13–0.77) 0.31 (0.12–0.79)
P trend 0.113 0.011 0.014
T1 32/20 1 1 1
T2 32/48 2.40 (1.17–4.91) 2.93 (1.22–7.05) 2.34 (0.95–5.79)
Snack
T3 32/28 1.40 (0.66–2.98) 1.69 (0.70–4.07) 1.59 (0.65–3.91)
P trend 0.571 0.451 0.505
1 2
All analyses were performed with the data matched for age, sex, and dietary survey season; Adjusted for gestation age,
birth weight, mother’s age, birth order, father’s education, and father’s occupation; 3 Model 2 + additional adjustment for
total energy intake, omega-3 fatty acids, lead, and mercury consumption; 4 Tertiles of dietary pattern scores; 5 Tests for
trend were conducted by assigning the median value to each tertile of heavy metal intake as a continuous variable.

4. Discussion

The present study identified four dietary patterns. The traditional-healthy dietary pattern,
characterized by high intakes of kimchi, grains, and bonefish, and low intakes of fast foods and
beverages, was associated with lower odds having ADHD. Although the present study focused on
dietary factors, significant associations with ADHD were found in father’s education and occupation.
Socioeconomic status of children is generally related to household income, and parent’s educational
background and occupation. Children from lower socioeconomic status are more likely diagnosed
with ADHD than children from higher socioeconomic status in previous studies [14í@. Family
income [14,15], parent’s education [15–17] and occupation [15,16] were significantly associated with
ADHD. Education status of mother was highly correlated with that of fathers in this study, and
occupation of mother did not vary compared to that of father’s. Thus, fathers’ educational background
and occupation were used as surrogate of socioeconomic status. As those variables were high
associated with ADHD, we adjusted them for the analysis.
The role of diet in the behavior of children has been controversial, but associations between several
nutritional factors and child behavior such as ADHD have been continually suggested [12,13]. Food
additives, sugar, and aspartame are considered negative factors in the development of ADHD, and
thus, dietary intervention studies with special diets, including additive-free and sugar elimination diets,
have been conducted. A meta-analysis has reported that artificial food coloring is associated with
 252

childhood hyperactivity [18]. However, in a sugar elimination intervention study, there was no
evidence that refined sugar affected child behavior [19–24].
The role of polyunsaturated fatty acids (PUFAs), particularly omega-3 fatty acids, in relation to
neurodevelopmental disorders has been studied because omega-3 fatty acids play a critical role in brain
development and function [25]. Children with ADHD have lower levels of omega-3 fatty acids, and
the supplementation of omega-3 fatty acids can reduce the symptoms of ADHD in school-aged
children and adolescents [26,27]. However, there was no clear evidence of improvement in ADHD
symptoms with omega-3 supplementation in randomized controlled trials, but these findings could be
the result of methodological problems [28,29]. The association between dietary pattern score and fatty
acid intake was investigated in this study. The traditional, seaweed-egg, and traditional-healthy pattern
scores were negatively associated with ADHD, although only the traditional-healthy pattern had a
statistically significant association; moreover, they were positively associated with fatty acid intake.
By contrast, the snack pattern score showed a positive association with ADHD and was negatively
associated with the intake of total fatty acids, PUFAs, and MUFAs. However, additional adjustment
for omega-3 fatty acid intake did not change the statistically significant association between the
traditional-healthy dietary pattern and ADHD. Thus, the factors associated with the beneficial effects
of a healthy dietary pattern might be complex.
Regarding mineral intake, calcium was positively associated with the scores of the traditional,
traditional-healthy, and snack patterns, and iron was positively associated with the scores of the
traditional and traditional-healthy patterns. Zinc was not associated with any of the four pattern scores.
Iron deficiency may be associated with ADHD [30] because iron stores in the brain can influence
dopamine-dependent functions [31,32]. A case-control study in India reported that the serum ferritin
level was lower in children with ADHD [33], while another study found that ADHD symptoms in
children with low serum ferritin levels were alleviated following iron supplementation [34]. In a
19-year follow-up study, the iron status of Costa Rican children was found to be associated with
behavioral problems in adolescents [35]. The role of zinc nutrition in ADHD is not clear, but evidence
suggests that zinc is beneficial in the treatment of children with ADHD [36,37]. Zinc deficiency is
involved in dopamine transporter dysfunction [38], and intervention studies have found that zinc
supplementation can reduce ADHD symptoms in children with low zinc levels [39–41]. Both, low iron
and zinc levels have been associated with dopamine metabolism, and low levels of iron and zinc are
involved in impaired dopamine transmission in subjects with ADHD [42–45].
Heavy metal exposure via food consumption was also investigated. Mercury was positively
associated with the traditional, traditional-healthy, and snack patterns, and lead was positively
associated with the traditional and snack patterns. The association between lead exposure and ADHD
has been widely studied, and a meta-analysis has reported that lead exposure is positively associated
with ADHD symptoms [46]. In a study with school-aged children living in two Romanian cities near a
metal-processing plant, an association with ADHD was observed only for lead exposure, not
aluminum or mercury exposure [47]. An association between the blood mercury level and ADHD in
Chinese children in Hong Kong has been observed [48], but a significant association was not found in
a cross-sectional study of Romanian children [47] or in a Children’s Health and Environment Research
(CHEER) study that surveyed elementary schools in six South Korean cities [49,50]. A more clear
association with ADHD has been observed for lead exposure, even at low concentrations [49,50].
 253

Prenatal mercury exposure is associated with an increased risk of neurobehavioral disorders, and lead
exposure in childhood has been associated with ADHD [51]. In this study, lead and mercury
consumption was positively correlated with the traditional-healthy dietary pattern, but it did not alter
the beneficial effects of the traditional-healthy dietary pattern on ADHD.
Recently, associations between dietary patterns and ADHD have been examined in several
cross-sectional studies [6,20,21]. One study, which included a population-based cohort of adolescents,
reported that a Western-style dietary pattern, characterized by high intakes of fat, refined sugars, and
sodium and low intakes of fiber, folate, and omega-3 fatty acids, was associated with increased odds of
an ADHD diagnosis, whereas a healthy dietary pattern, with high intakes of fiber, folate, and omega-3
fatty acids, was not correlated with the diagnosis of ADHD [20]. In a study of adolescents in China,
three major dietary patterns were identified, and dietary patterns characterized by a high intake of
snacks or animal-derived foods were associated with higher odds for psychological symptoms [21].
In Iranian school-aged children, four major dietary patterns were identified. The higher scores of
the dietary patterns associated with a high intake of sweets and fast food were associated with greater
odds for having ADHD, but no significant association was observed for the healthy or Western dietary
patterns [6]. In this study, traditional-healthy dietary pattern was positively associated with dietary
factors, such as PUFAs and minerals that are known for beneficial effects on ADHD. Another
beneficial effect of the traditional-healthy dietary pattern might be associated with the low fast food
intake. Junk foods are generally high in fat, sugar, additives, artificial food colorings, and
preservatives, which may negatively affect ADHD symptoms [52]. Overall, the traditional-healthy
dietary pattern was associated with many dietary factors that affect childhood behavioral disorders,
such as ADHD.
The present study has several limitations. As this was a case-control study, it is possible that dietary
intake was affected by an individual’s health status and social background. Thus, causal inference
cannot be determined. Results could differ by ADHD types, but information about ADHD type was
not gathered for subgroup analysis due to small sample size. However, such pattern analyses are useful
to further understand the diet of ADHD children as a whole rather than classifying it by a single
nutrient or food group.

5. Conclusions

The traditional-healthy dietary pattern, which is characterized by high intakes of kimchi, grains, and
bonefish, and low consumption of fast foods and beverages, appears to be negatively associated with
ADHD in school-aged Korean children.

Acknowledgments

This research was supported by a grant from the Korea Food and Drug Administration
(13162MFDS892).
 254

Author Contributions

Conceived and designed the experiments: HDW, DWK, YSH, BMC, JHP, JWK, JHY, HWC, JHL,
MJK, YMK, JHS, JK. Contributed to the acquisition of data: BMC, JHP, JWK, JHY, HWC, JHL,
MJK, YMK, JHS. Analyzed the data: HDW, DWK, JK. Wrote the paper: HDW, JK.

Conflicts of Interest

The authors declare no conflict of interest.

References

1. Wilens, T.E.; Biederman, J.; Spencer, T.J. Attention deficit/hyperactivity disorder across the
lifespan. Annu. Rev. Med. 2002, 53, 113–131.
2. Polanczyk, G.; de Lima, M.; Horta, B.; Biederman, J.; Rohde, L. The worldwide prevalence of
ADHD: A systematic review and metaregression analysis. Am. J. Psychiatry 2007, 164, 942–948.
3. Cuffe, S.P.; Moore, C.G.; McKeown, R.E. Prevalence and correlates of ADHD symptoms in the
national health interview survey. J. Atten. Disord. 2005, 9, 392–401.
4. Boyle, C.A.; Boulet, S.; Schieve, L.A.; Cohen, R.A.; Blumberg, S.J.; Yeargin-Allsopp, M.;
Visser, S.; Kogan, M.D. Trends in the prevalence of developmental disabilities in US children,
1997–2008. Pediatrics 2011, 127, 1034–1042.
5. Froehlich, T.E.; Anixt, J.S.; Loe, I.M.; Chirdkiatgumchai, V.; Kuan, L.; Gilman, R.C. Update on
environmental risk factors for attention-deficit/hyperactivity disorder. Curr. Psychiatry Rep. 2011,
13, 333–344.
6. Azadbakht, L.; Esmaillzadeh, A. Dietary patterns and attention deficit hyperactivity disorder
among Iranian children. Nutrition 2012, 28, 242–249.
7. Kim, E.; Kwon, H.; Ha, M.; Lim, M.; Oh, S.; Kim, J.; Yoo, S.; Paik, K. Relationship among
attention-deficit hyperactivity disorder, dietary behaviours and obesity. Child Care Health Dev.
2014, doi:10.1111/cch.12129.
8. Jang, C.-B.; Kim, H.-Y. The relationship between attention deficit hyperactivity disorder, dietary
habit and caffeine intake in upper-grade elementary school children. Korean J. Nutr. 2012, 45,
522–530.
9. Biederman, J. Attention-deficit/hyperactivity disorder: A selective overview. Biol. Psychiatry
2005, 57, 1215–1220.
10. Bobb, A.J.; Castellanos, F.X.; Addington, A.M.; Rapoport, J.L. Molecular genetic studies of
ADHD: 1991 to 2004. Am. J. Med. Genet. 2006, 141, 551–565.
11. Faraone, S.V.; Khan, S.A. Candidate gene studies of attention-deficit/hyperactivity disorder.
J. Clin. Psychiatry 2005, 67, 13–20.
12. Millichap, J.G.; Yee, M.M. The diet factor in attention-deficit/hyperactivity disorder. Pediatrics
2012, 129, 330–337.
13. Cormier, E.; Elder, J.H. Diet and child behavior problems: Fact or fiction? Pediatr. Nurs. 2007,
33, 138–143.
 255

14. Saadi, H.R.; Shamsuddin, K.; Sutan, R.; Alshaham, S.A. Socio-maternal risk factors of ADHD
among Iraqi children: A case-control study. Open J. Prev. Med. 2013, 3, 251–257.
15. Al Hamed, J.; Taha, A.; Sabra, A.; Bella, H. Attention deficit hyperactivity disorder (ADHD): Is it
a health problem among male primary school children. Bahrain Med. Bull. 2008, 30, 1–9.
16. Kalff, A.; Kroes, M.; Vles, J.; Bosma, H.; Feron, F.; Hendriksen, J.; Steyaert, J.; van Zeben, T.;
Crolla, I.; Jolles, J. Factors affecting the relation between parental education as well as occupation
and problem behaviour in Dutch 5-to 6-year-old children. Soc. Psychiatry Psychiatr. Epidemiol.
2001, 36, 324–331.
17. Khamis, V. Family environment and parenting as predictors of attention-deficit and hyperactivity
among Palestinian children. J. Soc. Serv. Res. 2006, 32, 99–116.
18. Schab, D.W.; Trinh, N.H.T. Do artificial food colors promote hyperactivity in children with
hyperactive syndromes? A meta-analysis of double-blind placebo-controlled trials. J. Dev.
Behav. Pediatr. 2004, 25, 423–434.
19. Wolraich, M.L.; Wilson, D.B.; White, J.W. The effect of sugar on behavior or cognition in
children. JAMA 1995, 274, 1617–1621.
20. Howard, A.L.; Robinson, M.; Smith, G.J.; Ambrosini, G.L.; Piek, J.P.; Oddy, W.H. ADHD is
associated with a “Western” dietary pattern in adolescents. J. Atten. Disord. 2011, 15, 403–411.
21. Weng, T.T.; Hao, J.H.; Qian, Q.W.; Cao, H.; Fu, J.L.; Sun, Y.; Huang, L.; Tao, F.B. Is there any
relationship between dietary patterns and depression and anxiety in Chinese adolescents?
Public Health Nutr. 2012, 15, 673–682.
22. Sinha, S.R.; Kossoff, E.H. The ketogenic diet. Neurologist 2005, 11, 161–170.
23. Murphy, P.; Likhodii, S.S.; Hatamian, M.; Burnham, W.M. Effect of the ketogenic diet on the
activity level of Wistar rats. Pediatr. Res. 2005, 57, 353–357.
24. Murphy, P.; Burnham, W. The ketogenic diet causes a reversible decrease in activity level in
Long-Evans rats. Exp. Neurol. 2006, 201, 84–89.
25. Horrocks, L.A.; Farooqui, A.A. Docosahexaenoic acid in the diet: Its importance in maintenance
and restoration of neural membrane function. Prostaglandins Leukot. Essent. Fatty Acids 2004,
70, 361–372.
26. Kirby, A.; Woodward, A.; Jackson, S.; Wang, Y.; Crawford, M. Childrens’ learning and
behaviour and the association with cheek cell polyunsaturated fatty acid levels. Res. Dev. Disabil.
2010, 31, 731–742.
27. Colter, A.L.; Cutler, C.; Meckling, K.A. Fatty acid status and behavioural symptoms of attention
deficit hyperactivity disorder in adolescents: A case-control study. Nutr. J. 2008, 7, 79–85.
28. Raz, R.; Gabis, L. Essential fatty acids and attention/deficit-hyperactivity disorder: A systematic
review. Dev. Med. Child Neurol. 2009, 51, 580–592.
29. Richardson, A.J. Omega-3 fatty acids in ADHD and related neurodevelopmental disorders.
Int. Rev. Psychiatry 2006, 18, 155–172.
30. Konofal, E.; Lecendreux, M.; Arnulf, I.; Mouren, M.-C. Iron deficiency in children with
attention-deficit/hyperactivity disorder. Arch. Pediatr. Adolesc. Med. 2004, 158, 1113–1115.
31. Solanto, M.V. Dopamine dysfunction in AD/HD: Integrating clinical and basic neuroscience
research. Behav. Brain Res. 2002, 130, 65–71.
 256

32. Erikson, K.M.; Jones, B.C.; Hess, E.J.; Zhang, Q.; Beard, J.L. Iron deficiency decreases dopamine
D1 and D2 receptors in rat brain. Pharmacol. Biochem. Behav. 2001, 69, 409–418.
33. Juneja, M.; Jain, R.; Singh, V.; Mallika, V. Iron deficiency in Indian children with attention
deficit hyperactivity disorder. Indian Pediatr. 2010, 47, 955–958.
34. Konofal, E.; Lecendreux, M.; Deron, J.; Marchand, M.; Cortese, S.; Zaïm, M.; Mouren, M.C.;
Arnulf, I. Effects of iron supplementation on attention deficit hyperactivity disorder in children.
Pediatr. Neurol. 2008, 38, 20–26.
35. Corapci, F.; Calatroni, A.; Kaciroti, N.; Jimenez, E.; Lozoff, B. Longitudinal evaluation of
externalizing and internalizing behavior problems following iron deficiency in infancy. J. Pediatr.
Psychol. 2010, 35, 296–305.
36. Arnold, L.E.; DiSilvestro, R.A. Zinc in attention-deficit/hyperactivity disorder. J. Child Adolesc.
Psychopharmacol. 2005, 15, 619–627.
37. Sinn, N. Nutritional and dietary influences on attention deficit hyperactivity disorder. Nutr. Rev.
2008, 66, 558–568.
38. Lepping, P.; Huber, M. Role of Zinc in the Pathogenesis of Attention-Deficit Hyperactivity
Disorder. CNS Drugs 2010, 24, 721–728.
39. Bilici, M.; YÕOGÕUÕP ) .DQGLO 6 %HNDUR÷OX 0 <ÕOGÕUPÕú 6 'H÷HU 2 hOJHQ 0
<ÕOGÕUDQ$$NVX+'RXEOH-blind, placebo-controlled study of zinc sulfate in the treatment of
attention deficit hyperactivity disorder. Prog. Neuropsychopharmacol. Biol. Psychiatry 2004, 28,
181–190.
40. Akhondzadeh, S.; Mohammadi, M.-R.; Khademi, M. Zinc sulfate as an adjunct to methylphenidate
for the treatment of attention deficit hyperactivity disorder in children: A double blind and
randomized trial [ISRCTN64132371]. BMC Psychiatry 2004, 4, doi:10.1186/1471-244X-4-9.
41. Arnold, L.E.; DiSilvestro, R.A.; Bozzolo, D.; Bozzolo, H.; Crowl, L.; Fernandez, S.; Ramadan, Y.;
Thompson, S.; Mo, X.; Abdel-Rasoul, M. Zinc for attention-deficit/hyperactivity disorder:
Placebo-controlled double-blind pilot trial alone and combined with amphetamine. J. Child
Adolesc. Psychopharmacol. 2011, 21, 1–19.
42. Oner, O.; Oner, P.; Bozkurt, O.H.; 2GDEDV( .HVHU1.DUDGDJ+.Õ]ÕOJQ0(IIHFWVRI
zinc and ferritin levels on parent and teacher reported symptom scores in attention deficit
hyperactivity disorder. Child Psychiatry Hum. Dev. 2010, 41, 441–447.
43. Kozielec, T.; Starobrat-Hermelin, B. Assessment of magnesium levels in children with attention
deficit hyperactivity disorder (ADHD). Magnes. Res. 1997, 10, 143–148.
44. Starobrat-Hermelin, B.; Kozielec, T. The effects of magnesium physiological supplementation on
hyperactivity in children with attention deficit hyperactivity disorder (ADHD). Positive response
to magnesium oral loading test. Magnes. Res. 1997, 10, 149–156.
45. Mousain-Bosc, M.; Roche, M.; Rapin, J.; Bali, J.P. Magnesium VitB6 intake reduces central
nervous system hyperexcitability in children. J. Am. Coll. Nutr. 2004, 23, 545S–548S.
46. Goodlad, J.K.; Marcus, D.K.; Fulton, J.J. Lead and attention-deficit/hyperactivity disorder
(ADHD) symptoms: A meta-analysis. Clin. Psychol. Rev. 2013, 33, 417–425.
47. Nicolescu, R.; Petcu, C.; Cordeanu, A.; Fabritius, K.; Schlumpf, M.; Krebs, R.; Krämer, U.;
Winneke, G. Environmental exposure to lead, but not other neurotoxic metals, relates to core
 257

elements of ADHD in Romanian children: Performance and questionnaire data. Environ. Res.
2010, 110, 476–483.
48. Wong, V. Attention-deficit hyperactivity disorder and blood mercury level: A case-control study
in Chinese children. Neuropediatrics 2006, 37, 234–240.
49. Ha, M.; Kwon, H.J.; Lim, M.H.; Jee, Y.K.; Hong, Y.C.; Leem, J.H.; Sakong, J.; Bae, J.M.;
Hong, S.J.; Roh, Y.M. Low blood levels of lead and mercury and symptoms of attention deficit
hyperactivity in children: A report of the children’s health and environment research (CHEER).
Neurotoxicology 2009, 30, 31–36.
50. Kim, S.; Arora, M.; Fernandez, C.; Landero, J.; Caruso, J.; Chen, A. Lead, mercury, and cadmium
exposure and attention deficit hyperactivity disorder in children. Environ. Res. 2013, 126,
105–110.
51. Boucher, O.; Jacobson, S.W.; Plusquellec, P.; Dewailly, É.; Ayotte, P.; Forget-Dubois, N.;
Jacobson, J.L.; Muckle, G. Prenatal methylmercury, postnatal lead exposure, and evidence of attention
deficit/hyperactivity disorder among inuit children in Arctic Québec. Environ. Health Perspect.
2012, 120, 1456–1461.
52. Wiles, N.J.; Northstone, K.; Emmett, P.; Lewis, G. “Junk food” diet and childhood behavioural
problems: Results from the ALSPAC cohort. Eur. J. Clin. Nutr. 2007, 63, 491–498.

© 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).
258
 259

Reprinted from Nutrients. Cite as: Liu, J.; Hanlon, A.; Ma, C.; Zhao, S.R.; Cao, S.; Compher, C. Low
Blood Zinc, Iron, and Other Sociodemographic Factors Associated with Behavior Problems in
Preschoolers. Nutrients 2014, 6, 530-545.

Low Blood Zinc, Iron, and Other Sociodemographic Factors

Associated with Behavior Problems in Preschoolers
Jianghong Liu *, Alexandra Hanlon, Chenjuan Ma, Sophie R. Zhao, Siyuan Cao
and Charlene Compher

School of Nursing, University of Pennsylvania, 418 Curie Blvd., Philadelphia, PA 19104, USA;
E-Mails: alhanlon@nursing.upenn.edu (A.H.); chenjuan@nursing.upenn.edu (C.M.);
sophie.r.zhao@hotmail.com (S.R.Z.); caos@sas.upenn.edu (S.C.);
compherc@nursing.upenn.edu (C.C.)

* Author to whom correspondence should be addressed; E-Mail: jhliu@nursing.upenn.edu;

Tel.: +1-215-573-7492; Fax: +1-215-746-3374.

Received: 25 November 2013; in revised form: 13 January 2014 / Accepted: 20 January 2014 /
Published: 27 January 2014

Abstract: Previous research supports the link among malnutrition, cognitive dysfunction,
and behavioral outcomes; however, less research has focused on micronutrient
deficiencies. This study investigates whether micronutrient deficiencies, specifically blood
zinc and iron levels, will be associated with increased behavior problem scores, including
internalizing and externalizing behaviors. 1314 Children (55% boys and 45% girls) from
the Jintan Preschool Cohort in China participated in this study. Venous blood samples were
collected and analyzed for zinc and iron when the children were 3–5 years old. Behavior
problems were measured with the Child Behavior Checklist (CBCL), which was completed
by the parents when children were in their last months of preschool (mean age 5.6 years).
General linear multivariate modeling was used, with adjustment for important
sociodemographic variables. The results indicate that low zinc levels alone (p = 0.024) and
combined low zinc and iron levels (p = 0.022) are significantly associated with increased
reports of total behavior problems. We did not find an association between low iron and
behavior problems. With regards to sociodemographics, living in the suburbs is associated
with increased internalizing problems, while higher mother’s education and being female
were associated with decreased externalizing problems. This study suggests that
micronutrient deficiencies and sociodemographic facts are associated with behavior problems
in preschoolers.
 260

Keywords: internalizing; externalizing; total behavior; CBCL; child; micronutrient

deficiency; zinc and iron

1. Introduction

The link between early nutrition deficiency and behavior outcomes has been receiving increasing
attention [1–3]. At the prenatal level, Neugebauer, Hoek, and Susser [4] found that the male offspring
of nutritionally-deprived pregnant women had 2.5 times the normal rate of antisocial personality
disorder in adulthood. At the postnatal level, in a longitudinal study from the Mauritius birth
cohort [5], it was found that children with malnutrition (protein, zinc, iron and vitamin B deficiencies)
at age 3 years, compared to controls, have higher externalizing behavior problems (i.e., antisocial,
aggressive, and hyperactive behavior) at ages 8, 11, and 17 years [6]. In another more recent longitudinal
study, Galler et al. [3] found that children who were malnourished at an early age showed significantly
higher parent-reported levels of behavior problems, particularly aggression, and decreased executive
functioning at age 9–15 and again at 11–17, independent of baseline age, sex, household standard of
living, and maternal depressive symptoms. Finally, at the intervention level [7], a double-blind,
placebo-controlled randomized trial from England showed that supplementation of adult prisoners’ diet
with vitamins, minerals, and essential fatty acids significantly reduced antisocial and violent behavior
in prison. These findings have been recently replicated in young prisoners in the Netherlands [8].
This initial evidence supports the relationship between nutrition and behavioral problems; however,
more research is still needed.
While increasing studies have showed the association between overall nutritional status and child
behavior, few studies have specifically investigated blood zinc and iron status in relation to behavior.
In developing countries, low zinc and iron levels are common [9–11]. Indeed, more than 90% of
affected individuals live in developing countries, and approximately one-tenth of the worldwide
population suffers from iron deficiency [12]. Furthermore, few studies have been conducted in Asian
populations. In China, zinc and iron deficiency were previously very common, but over the past two to
three decades, reports on zinc and iron intake have been mixed due to socioeconomic reform and rapid
economic development that have taken place since 1979. The availability [13,14] and affordability [15]
of foods have increased dramatically during this time, and as a result of this increased food production and
access to food, the prevalence of malnutrition has decreased, while over-nutrition has increased [16,17].
Studies indicate that iron deficiency is less prevalent than zinc deficiency among Chinese children [18,19],
and although the prevalence of anemia has decreased in China, it still exists among children [16,20].
Taken together, this makes the consideration of zinc and iron intake in Chinese samples a relevant
issue for better understanding putative risk factors for behavioral outcomes.
Zinc and iron play important roles in children’s physical and behavioral health; however, there is a
relative lack of attention given to the effects of specific micronutrient (e.g., zinc and iron) deficiency
on behavior problems, including internalizing and externalizing disorders. Zinc is a component of
enzymes that affect growth in infancy and childhood, sexual maturation, neuromotor development, and
immunity. Mental function is improved by zinc’s promotion of normal brain development and
 261

physiology [21–23]. Iron similarly boosts mental functioning by serving as a co-enzyme involved in
the production and release of neurotransmitters [21,22] and by influencing cognitive function [24,25]
and behavioral disorders such as attention-deficit hyperactivity disorder [6,26].
There is now increasing evidence of the relationship between malnutrition and childhood behavior
problems [27–30], though more data are needed to address the impact of specific micronutrient
deficiencies on both internalizing and externalizing problems separately [2]. The importance of zinc
and iron in physiological development seems to warrant particular attention with regard to how these
micronutrients relate to behavioral outcomes. Childhood behavioral problems represent an important
sub-area of developmental psychopathology [31–34]. Thus, identifying early childhood behavior
problems—and, perhaps more importantly, their early risk factors, including nutritional and
sociodemographic factors—is important for understanding and preventing problem behaviors later in
life [35,36]. The purpose of this study is to assess the association of micronutrients controlled for
sociodemographic factors with behavior outcomes. We hypothesize that nutritional deficiencies,
specifically zinc and iron deficiencies, will be associated with increased behavior problems.

2. Experimental Section

2.1. Participants and Procedures

The current study was part of a population-based community preschool cohort study of
1656 Chinese children (55.5% boys, 44.5% girls) initially recruited between the Fall of 2004 and the
Spring of 2005 from four preschools in the city of Jintan, located in the southeastern coastal region
of Mainland China. In China, preschools are called kindergartens and enroll children from ages
3–6 years, after which children enter the elementary school system; to be consistent, we use preschool
to refer to our study sample. Detailed sampling and research procedures of this larger cohort study
have been described elsewhere [37,38]. Briefly, all children and parents taking part in the original
cohort study were invited to participate for assessment of children’s behaviors while the children were
in the final few months of their senior year in preschool (spring 2005 to spring 2007). At that point,
some children dropped out of the study because they changed schools or because data were not fully
available. Therefore, only 1385 children in the original sample were followed up in the later waves.
There was no statistically significant difference between those who dropped out of the study and those
who were retained [37,39].
In the last year of preschool, parents were asked to assess their children with the Chinese version of
the Child Behavior Checklist (CBCL/1.5–5). Since some of the children were beyond the age limit
of the CBCL/1.5–5, the current analysis only addressed the subset of the original sample that was
under age 6 to adhere to the age requirement of the measure. Our final data set for analysis was thus
comprised of 1314 preschoolers with a mean age of 66.6 months (SD = 5, range = 50–71), which is
close to the common kindergarten age in the US. Written informed consent was obtained from parents.
Institutional Review Board (IRB) approval was obtained from the University of Pennsylvania and the
ethical committee for research at Jintan Hospital in China.
 262

2.2. Measures

2.2.1. Micronutrient Deficiency

Blood specimens were collected in Fall 2004 and Spring 2005 by trained pediatric nurses using a
strict research protocol to avoid lead contamination. Approximately 0.5 mL of venous blood was
collected in a lead-free EDTA tube for zinc and iron analysis. Samples were frozen and shipped to the
Child Development Center, Nanjing Medical University, Nanjing, China, for analysis. Specimens
remained frozen at í °C until analysis. Blood concentrations of zinc and iron were determined by
atomic absorption spectrophotometry (BH model 5.100 manufactured by Beijing Bohu Innovative
Electronic Technology Corporation), with duplicate readings taken with an integration time of 2 s.
The reliability and validity of the analysis and the detailed analytic procedure have been described
previously [40]. Detailed information on blood sample data collection and analysis is given in [39].
/RZ]LQFOHYHOVZHUHGHILQHGE\FRQFHQWUDWLRQȝJG/DQGORZLURQE\FRQFHQWUDWLRQȝJG/
in blood, with cutoffs determined from the middle of the normal range. Combined low zinc and
iron were defined as currently low zinc and low iron concentrations, i.e., children in this category have
ERWK=QȝJG/DQG)HȝJG/

2.2.2. Behavior Problems at Ages 5–6

Childhood behavior problems were measured with the Chinese version of the Achenbach System of
Empirically Based Assessment (ASEBA) CBCL/1.5–5 [41]. The CBCL is a widely used scale for
assessing behavioral and emotional problems in children. In this study, parents were asked to answer
the 99 items of the CBCL instrument, which dealt with their children’s behavior within the past
12 months, and give a rating from a 3-point scale (0 = not true; 1 = sometimes true, or 2 = often true) [41].
Factor analysis performed on the CBCL/1.5–5 has revealed two broadband factors: Internalizing
behaviors and Externalizing behaviors [42]. Separately, factor analysis has produced four syndromes
for Internalizing behaviors: Emotionally Reactive, Anxious/Depressed, Somatic Complaints, and
Withdrawn; and two syndromes for Externalizing behaviors: Attention Problems and Aggressive
Behavior [41,43]. These factor structures have also been validated in our previous study [44]. The
internal reliabilities (coefficient alpha) for the scales in our study sample were as follows: Emotionally
Reactive (0.71), Anxious/Depressed (0.64), Somatic Complaints (0.58), and Withdrawn (0.73),
Attention Problems (0.64) and Aggression (0.87). The strategy we employ in this study is to use these
established scales as predictors of latent construct “Internalizing behavior” and “Externalizing
Behavior”. The sum of the items in the scales was used.

2.2.3. Sociodemographic Variables

Sociodemographic information was obtained from the questionnaire filled out by the parents, and
included information on gender, parental education, home living conditions, and the age of mother
when the child was born. These data were collected as control variables given their potential direct
effects on child behavioral problems. As discussed in our previous publication [39], we did not ask for
data on household income because it is often not the best indicator of socioeconomic status, therefore
 263

we used information on house size as a proxy for evaluating socioeconomic status. A descriptive
summary of these demographic variables is presented in Table 1.

Table 1. Baseline characteristics of study population (N = 1314).

Characteristic N %
Gender
Male 758 55
Female 614 45
Location
City 959 74
Suburb 188 14
Rural 153 12
Mother’s Education
Low (”9 years) 625 48
Medium (9–12 years) 411 32
High (>12 years) 264 20
Family Size
>3 persons/household 713 57
”3 persons/household 548 43
House Size (m2)
<100 561 45
•100 696 55

2.3. Statistical Analyses

Descriptive statistics, including frequencies and percentages, are used to characterize categorical
demographic subject characteristics (Table 1). Behavioral outcomes measured on a continuum are
described using means and standard deviations, and compared by low zinc and iron level groups.
Multivariate modeling of behavioral outcomes regressed on low zinc and iron group, with adjustment
for important sociodemographic variables (family size, gender, house size, mother’s education) is
accomplished using general linear modeling. The father’s education was not included in the analysis to
avoid multicolinearity because the variable was found to be highly correlated with the mother’s
education. Models also accounted for clustering at the school level. Levine’s tests are used to
test for homogeneity of variance across all cells. In an attempt to identify multicollinearity, bivariate
associations between independent variables are examined using chi-square tests. Statistical
significance was taken at the two-sided p < 0.05 level. All the analyses were performed using STATA
version 11.0 [45].

3. Results

Mean behavior problem scores for total behavior, and internalizing and externalizing behavior
problems are described in Tables 2–4 according to Zn level (Table 2), Fe level (Table 3), and to Zn and
Fe levels (Table 4).
 264

Table 2. Children’s behavior problems by Zn level (N = 1314).

Low Zn † Normal Zn
Variable
N Mean SD N Mean SD
Total Behavior Problems 502 21.32 17.64 812 18.8 16.61
Internalizing Behavior Problems 502 6.42 5.91 812 5.74 5.84
Externalizing Behavior Problems 502 9.22 9.69 812 7.95 8.43
†
'HILQHGDV]LQFȝJG/.

Table 3. Children’s behavior problems by Fe level (N = 1314).

Low Fe † Normal Fe
Variable
N Mean SD N Mean SD
Total Behavior Problems 307 21.1 17.65 1007 19.36 16.85
Internalizing Behavior Problems 307 6.57 5.94 1007 5.83 5.85
Externalizing Behavior Problems 307 8.78 9.65 1007 8.33 8.73
†
'HILQHGDVLURQȝJG/

Table 4. Children’s behavior problems by Zn and Fe level (N = 1314).

Low Zn and Fe † Others
Variable
N Mean SD N Mean SD
Total Behavior Problems 215 21.41 18.2 1099 19.44 16.8
Internalizing Behavior Problems 215 6.63 5.99 1099 5.88 5.85
Externalizing Behavior Problems 215 9.07 10.18 1099 8.31 8.69
†
'HILQHGDV]LQFȝJG/DQGLURQȝJG/.

3.1. Zinc Status and Children’s Behavior

Results from our analyses (Table 5) using the model of total behavior score regressed on deficient
zinc showed that low zinc status was significantly associated with higher total behavior problems
(p = 0.024) in children, along with living in the suburbs, whereas high mother’s education and being
female were associated with lower behavior problems. The model of externalization score regressed on
low zinc showed being female and high mother’s education are associated with significantly lower
externalizing behavior scores. The model of internalizing score regressed on low zinc showed living in
the suburbs to be positively associated with internalizing problems.

3.2. Iron Status and Children’s Behavior

We did not find significant association between low iron status and internalizing, externalizing, or
total behavior problems (Table 6). However, we observed that several sociodemographic indicators
were significantly associated with behavior problems. The analysis using the model of total behavior
score regressed on low iron indicated that being female and having high level of mother’s education
are associated with significantly lower (better) total behavior scores. Living in the suburbs was also
positively associated with total behavior problems. For the model of externalization score regressed on
low iron, being female and high mother’s education were also associated with significantly lower
 265

externalizing behavior scores. The model of internalizing score regressed on low iron showed living in
the suburbs to be positively associated with internalizing problems.

3.3. Zinc and Iron Status and Children’s Behavior

Results from our analyses (Table 7) using the model of total behavior score regressed on combined
low levels of zinc and iron, having low zinc and iron was significantly associated with higher total
behavior problems (p = 0.022), whereas high mother’s education and being female were associated
with reduced total behavior score. The model of externalization score regressed on combined low zinc
and iron showed that being female and high mother’s education are also associated with significantly
lower externalizing behavior scores. The model of internalizing score regressed on combined low zinc
and iron showed living in the suburbs to be positively associated with internalizing problems.

4. Discussion

In this community sample of Chinese pre-school children (N = 1314), with micronutrient levels
measured at ages 3–5 years and behavioral problems measured at mean age 5.6 years, we found an
association between micronutrient deficiency and total behavior problems. Firstly, we found that low
zinc concentration is positively correlated with total behavior problems. Secondly, we found that
combined low blood levels of zinc and iron is positively correlated with total behavior problems.
These effects remained significant after controlling for sociodemographic factors such as gender
and mother’s education. We did not find a significant association between low iron status and
behavior problems.
The finding of the association of zinc deficiency with child behavior problems is consistent with
previous findings [46,47]. Zinc is a component of enzymes that affect growth in infancy and
childhood, sexual maturation, neuro-motor development, and immunity. Specifically, zinc acts as the
integral enzymatic agent in metabolic processes of proteins, carbohydrates, and lipids [48] and is used
as a neurotransmitter or neuromodulator in the central nervous system [49]. Mental function is
improved by zinc’s promotion of normal brain development and physiology [22,23]. A recent study
reported a relationship between low zinc and greater levels of hyperactivity, anxiety, and conduct
problems [50]. Indeed, animal and human models have suggested a relationship between low serum
zinc and anxiety, fear-like behaviors, and depression—implicating the role of the dopaminergic and
serotonergic systems [51,52].
Nutrients 2014, 6 266

Table 5. The effect of Zinc on children’s behavior problems (N = 1314).

Total Problems (R2 = 0.08) Internalizing (R2 = 0.04) Externalizing (R2 = 0.11)
Variable
ȕ 6( 95% CI p ȕ 6( 95% CI p ȕ 6( 95% CI p
†
Low Zn 2.13 (0.33) 0.70–3.57 0.024 * 0.63 (0.31) í–1.97 0.180 1.04 (0.28) í–2.25 0.066
Gender:
Male - - - - - - - - -
Female í  í–í 0.049 * í  í–2.33 0.994 í  í–í 0.007 *
Location:
City - - - - - - - - -
Suburb 8.44 (1.47) 2.11–14.78 0.029 * 3.23 (0.32) 1.87–4.60 0.010 * 3.13 (0.91) í–7.03 0.075
Rural 1.96 (3.61) í–17.51 0.641 0.55 (0.83) í–4.13 0.576 0.96 (2.05) í–9.78 0.686
Mother’s Education:
Low (”\ears) - - - - - - - - -
Medium (9–12 years) í  í–3.41 0.209 í  í–3.04 0.595 í  í–0.15 0.060
High (>12 years) í  í–í 0.012 * í  í–1.57 0.301 í  í–í 0.042 *
Family size:
” - - - - - - - - -
>3 í  í–1.70 0.295 í  í–0.48 0.622 í  í–0.65 0.205
House Size:
<100 m2 - - - - - - - - -
•P2 í  í–2.83 0.354 í  í–0.55 0.687 í  í–1.09 0.191
†
'HILQHGDV]LQFȝJG/; * Statistically significant at two-sided p < 0.05 level.
Nutrients 2014, 6 267

Table 6. The effect of Iron on children’s behavior problems (N = 1314).

Total Problems (R2 = 0.07) Internalizing (R2 = 0.04) Externalizing (R2 = 0.10)
Variable
ȕ 6( 95% CI p ȕ 6( 95% CI p ȕ 6( 95% CI p
†
Low Fe 1.34 (0.79) í–4.74 0.233 0.59 (0.27) í–1.77 0.166 0.31 (0.39) í–1.98 0.510
Gender:
Male - - - - - - - - -
Female í  í–í 0.046 * 0.00 (0.53) í–2.28 0.999 í  í–í 0.007 *
Location:
City - - - - - - - - -
Suburb 8.61 (1.57) 1.84–15.38 0.032 * 3.27 (0.32) 1.89–4.66 0.009 * 3.22 (0.96) í–7.35 0.079
Rural 2.07 (3.75) í–18.22 0.636 0.58 (0.87) í–4.32 0.571 1.01 (2.12) í–10.14 0.680
Mother’s Education:
Low (”\ears) - - - - - - - - -
Medium (9–12 years) í  í–3.77 0.232 í  í–3.17 0.619 í  í–0.06 0.054
High (>12 years) í  í–í 0.011 * í  í–1.59 0.336 í  í–í 0.037 *
Family size:
” - - - - - - - - -
>3 í  í–1.40 0.224 í  í–0.41 0.467 í  í–0.49 0.150
House Size:
<100 m2 - - - - - - - - -
•P2 í  í–2.79 0.319 í  í–0.61 0.586 í  í–1.00 0.168
†
'HILQHGDVLURQȝJG/; * Statistically significant at two-sided p < 0.05 level.
Nutrients 2014, 6 268

Table 7. The effect of Zinc and Iron on children’s behavior problems (N = 1314).
Total Problems (R2 = 0.07) Internalizing (R2 = 0.04) Externalizing (R2 = 0.10)
Variable
ȕ 6( 95% CI p ȕ 6( 95% CI p ȕ 6( 95% CI p
†
Low Zn & Fe 1.53 (0.23) 0.53–2.53 0.022 * 0.69 (0.23) í–1.69 0.095 0.49 (0.23) í–1.49 0.174
Gender:
Male - - - - - - - - -
Female í  í–í 0.046 * í  í–2.30 0.991 í  í–í 0.007 *
Location:
City - - - - - - - - -
Suburb 8.62 (1.58) 1.83–15.42 0.032 * 3.28 (0.32) 1.90–4.66 0.009 * 3.22 (0.95) í–7.31 0.078
Rural 2.09 (3.77) í–18.31 0.634 0.59 (0.87) í–4.32 0.564 1.02 (2.12) í–10.16 0.678
Mother’s Education:
Low (”\ears) - - - - - - - - -
Medium (9–12 years) í  í–3.67 0.227 í  í–3.13 0.615 í  í–0.07 0.054
High (>12 years) í  í–í 0.012 * í  í–1.68 0.329 í  í–í 0.038 *
Family size:
” - - - - - - - - -
>3 í  í–1.31 0.209 í  í–0.39 0.424 í  í–0.45 0.140
House Size:
<100 m2 - - - - - - - - -
•P2 í  í–2.86 0.333 í  í–0.65 0.649 í  í–1.04 0.174
†
'HILQHGDV]LQFȝJG/DQGLURQȝJG/ * Statistically significant at two-sided p < 0.05 level.
 269

The association of combined zinc and iron with children’s behavior has been reported
previously in [53]. It has been postulated previously that zinc and iron are thought to interact with
one another for absorption. When levels of both zinc and iron are low, a more severe pattern of
nutritional compromise is suggested. In fact, our results showed that blood zinc was positively
correlated with blood iron (r = 0.31, p = 0.000), consistent with other published reports [54,55].
As previously discussed, zinc and iron both play important enzymatic roles in the dopamine
metabolism pathways [56–58], and it has been suggested that zinc and iron deficiencies together
can lead to additive effects in dopaminergic system alterations [50]. Interestingly, Oner et al. [50]
also reported that while combined low serum levels of zinc and iron are correlated with increased
hyperactivity, as a single nutrient effect, only zinc deficiency, not iron deficiency, was related to
conduct problems and anxiety. As concluded in the paper, Oner et al. [50] suggested that zinc and
iron deficiencies might be associated with different types of behavior problems. However, it is also
possible that the association of combine low zinc and iron with behavior problems is driven by the
effect of low zinc alone.
While not the focus of this current paper, the sociodemographic control variables also had
effects on our outcome constructs. It is worth noting that the children whose mothers had more than
12 years education exhibited decreased externalizing behavior problems in all of our analyses. As
the mothers are postulated to be the primary caregiver of the child, this observation points to the
fact that children of mothers who had more education tend to exhibit less behavior problems,
possibly as the result of better parenting practices in addition to better nutritional habits as results
of increased education concerning nutrition. Furthermore, being female is also associated with
decreased externalizing behavior and total behavior problems; such that girls in our sample tended
to have lower incidents of externalizing and total behavior problems than boys.
Additionally, living in the suburbs has consistently been shown to be significantly associated
with increase in total behavior problems and, even more strongly, in internalizing behavior
problems. This result might not be too surprising because there has been ample evidence in the
literature supporting that children of economically affluent families tend to develop more
internalizing behavior problems, such as anxiety, depression, and substance abuse [59,60].
Globally, an epidemiological study has found depression rates to be higher in developed countries
than in others [61]. The results from our data support the literature in that the suburban residents in
our sample have the highest parental occupation and parental education levels, above both the
urban and rural groups. This finding is not surprising given the suburban preschool is in a “new
development zone”, where up-and-coming young parents, generally of a better educational and
socio-economic background, are pursuing relatively better and higher-paid occupations.
Furthermore, parents who are in the transitional stage of social and economic rise in their lives and
careers are likely to have high expectation of their children. Additionally, the parents themselves
live very high-pressured lives from their own occupational demands, leading to the possibility that
they are more likely to suffer from internalizing problems themselves, such as anxiety and
depression. Previous studies have shown that parental symptoms of depression have been
associated with children’s problem behavior in clinical and community samples [62,63].
We postulate that this finding could be the result of a more stressful suburban parental lifestyle due
 270

to social and occupational stress, factors that may indirectly affect their children by not having
enough time to interact with them or through a decreased emphasis on food choice, which can
contribute to nutrition status. Nevertheless, future research could include stress and lifestyle factors
and their effects on malnutrition.
Limitations of this study should not be overlooked. First, these findings do not establish a causal
relationship between zinc, iron and behavioral disturbance. However, results from intervention
trials should be considered to elucidate whether a causal relationship truly exists. The nature of the
study design also required that behavior be assessed during the last year of preschool while blood
micronutrient levels were assessed when the children were ages 3–5 years. Consequently,
participants differed in the time between times of micronutrient and behavior assessment.
Secondly, nutrition was only assessed at a single point in time, making it difficult to generalize
findings and assess the role of sustained nutrition deficiency, or even nutrition deficiency during
the prenatal period. As a result, we were unable to separate the effects of chronic versus acute
nutrition, which may have different implications on behavior [64]. Thirdly, although we included
some sociodemographic factors in our analyses, other factors, such as income, should be
considered in future studies. In addition, potential confounders such as the effects of other nutrients
(e.g., vitamin D), food intake, and physical activity level should also be considered. Fourthly,
examination of an all-Chinese sample in this age range limits application to other cultures, as
cultural, ethnic, social, and age factors impact child rearing behaviors, including nutrition and
feeding. Finally, while this study only included the parent-report, it would be equally important to
consider other informants. Currently, the children in our cohort are at school age, and future studies
will include youth self-report of behavior to assess the relationship between micronutrient
deficiencies and behavior problems.

5. Conclusions

Few studies have specifically examined the role of zinc and iron status in relation to child
behavior. This sample of Chinese preschoolers suggests that low blood zinc is correlated to
increased total behavior problems and that, additionally, combined low blood zinc and iron levels
are also linked to increased total behavior problems early in childhood. Implications may include
more public awareness of the importance of micronutrients. While sociodemographic factors are
not easily modifiable, it is possible to encourage parents, children, and baby/child care
professionals to make healthy food choices, including foods rich in zinc and iron.

Acknowledgements

This research is supported by the National Institute of Environment Health Sciences (NIEHS,
R01-ES018858, 1K02-ES019878-01) US; Jintan City Government; Jintan Hospital, China.
 271

Conflicts of Interest

Ethical approval was obtained from the Institutional Review Board at the University of
Pennsylvania and Jintan Hospital. None of the authors declare any conflict of interest regarding the
data and materials presented in this paper.

References

1. Lozoff, B.; Georgieff, M.K. Iron deficiency and brain development. Sem. Pediatr. Neurol.
2006, 13, 158–165.
2. Corapci, F.; Calatroni, A.; Kaciroti, N.; Jimenez, E.; Lozoff, B. Longitudinal evaluation of
externalizing and internalizing behavior problems following iron deficiency in infancy.
J. Pediatr. Psychol. 2010, 35, 296–305.
3. Galler, J.R.; Bryce, C.P.; Waber, D.P.; Medford, G.; Eaglesfield, G.D.; Fitzmaurice, G. Early
malnutrition predicts parent reports of externalizing behaviors at ages 9–17. Nutr. Neurosci.
2011, 14, 138–144.
4. Neugebauer, R.; Hoek, H.W.; Susser, E. Prenatal exposure to wartime famine and
development of antisocial personality disorder in early adulthood. JAMA 1999, 282, 455–462.
5. Raine, A.; Liu, J.H.; Venables, P.H.; Mednick, S.A.; Dalais, C. Cohort profile: The mauritius
child health project. Int. J. Epidemiol. 2010, 39, 1441–1451.
6. Liu, J.; Raine, A.; Venables, P.; Mednick, S.A. Malnutrition at age 3 years predisposes to
externalizing behavior problems at ages 8, 11 and 17 years. Am. J. Psychiatry 2004, 161,
2005–2013.
7. Gesch, C.B.; Hammond, S.M.; Hampson, S.E.; Eves, A.; Crowder, M.J. Influence of
supplementary vitamins, minerals and essential fatty acids on the antisocial behaviour of
young adult prisoners. Randomised, placebo-controlled trial. Br. J. Psychiatry 2002, 181,
22–28.
8. Zaalberg, A.; Nijman, H.; Bulten, E.; Stroosma, L.; van der Staak, C. Effects of nutritional
supplements on aggression, rule breaking, and psychopathology among young adult prisoners.
Aggress. Behav. 2010, 36, 117–126.
9. Zimmermann, M.B.; Hurrell, R.F. Nutritional iron deficiency. Lancet 2007, 370, 511–520.
10. Diaz, J.R.; de las Caigas, A.; Rodriguez, R. Micronutrient deficiencies in developing and
affluent countries. Eur. J. Clin. Nutr. 2003, 57, S70–S72.
11. Bhutta, Z.A. Iron and zinc deficiency in children in developing countries. BMJ 2007, 334,
104–105.
12. Lopez, M.A.; Martos, F.C. Iron availability: An updated review. Int. J. Food Sci. Nutr. 2004,
55, 597–606.
13. Lin, J.Y.F. Rural reforms and agricultural growth in China. Am. Econ. Rev. 1992, 82, 34–51.
14. Zhang, J. China’s success in increasing per capita food production. J. Exp. Bot. 2011, 62,
3707–3711.
15. Chai, J.C.H. Consumption and living standards in China. China Quart. 1992, 131, 721–749.
 272

16. Li, L.M.; Rao, K.Q.; Kong, L.Z.; Yao, C.H.; Xiang, H.D.; Zhai, F.Y.; Ma, G.S.; Yang, X.G.;
Technical Working Group of China National Nutrition and Health Survey. A description on
the chinese national nutrition and health survey in 2002. Zhonghua Liu Xing Bing Xue Za Zhi
2005, 26, 478–484.
17. Zhai, F.; Wang, H.; Du, S.; He, Y.; Wang, Z.; Ge, K.; Popkin, B.M. Lifespan nutrition and
changing socio-economic conditions in China. Asia Pac. J. Clin. Nutr. 2007, 16, 374–382.
18. Qin, Y.; Melse-Boonstra, A.; Shi, Z.; Pan, X.; Yuan, B.; Dai, Y.; Zhao, J.; Zimmermann,
M.B.; Kok, F.J.; Zhou, M. Dietary intake of zinc in the population of Jiangsu Province, China.
Asia Pac. J. Clin. Nutr. 2009, 18, 193–199.
19. Ma, G.; Li, Y.; Jin, Y.; Du, S.; Kok, F.J.; Yang, X. Assessment of intake inadequacy and food
sources of zinc of people in China. Public Health Nutr. 2007, 10, 848–854.
20. Zhu, Y.P.; Liao, Q.K. Prevalence of iron deficiency in children aged 7 months to 7 years in
China. Zhonghua Er Ke Za Zhi 2004, 42, 886–891.
21. Liu, J.; Raine, A. The effect of childhood malnutrition on externalizing behavior. Curr. Opin.
Pediatr. 2006, 18, 565–570.
22. Singh, M. Role of micronutrients for physical growth and mental development. Indian J.
Pediatr. 2004, 71, 59–62.
23. Bitanihirwe, B.K.; Cunningham, M.G. Zinc: The brain’s dark horse. Synapse 2009, 63,
1029–1049.
24. Grantham-McGregor, S.; Ani, C. A review of studies on the effect of iron deficiency on
cognitive development in children. J. Nutr. 2001, 131, 649s–666s.
25. Liu, J.; Raine, A.; Venables, P.H.; Dalais, C.; Mednick, S.A. Malnutrition at age 3 years and
lower cognitive ability at age 11 years: Independence from psychosocial adversity. Arch.
Pediatr. Adolesc. Med. 2003, 157, 593–600.
26. Konofal, E.; Lecendreux, M.; Deron, J.; Marchand, M.; Cortese, S.; Zaim, M.; Mouren, M.C.;
Arnulf, I. Effects of iron supplementation on attention deficit hyperactivity disorder in
children. Pediatr. Neurol. 2008, 38, 20–26.
27. Fishbein, D.H.P.; Pease, S.E. Diet, nutrition, and aggression. J. Offender Rehabil. 1995, 21,
117–144.
28. Breakey, J. The role of diet and behaviour in childhood. J. Paediatr. Child Health 1997, 33,
190–194.
29. Werbach, M. Nutritional influences on aggressive behavior. J. Orthomol. Med. 1992, 7, 45–51.
30. Lozoff, B.; Beard, J.; Connor, J.; Felt, B.; Georgieff, M.; Schallert, T. Long-lasting neural and
behavioral effects of iron deficiency in infancy. Nutr. Rev. 2006, 64, S34–S43.
31. Raine, A. Annotation: The role of prefrontal deficits, low autonomic arousal, and early health
factors in the development of antisocial and aggressive behavior in children. J. Child Psychol.
Psychiatry 2002, 43, 417–434.
32. Liu, J.; Wuerker, A. Biosocial bases of aggressive and violent behavior—Implications for
nursing studies. Int. J. Nurs. Stud. 2005, 42, 229–241.
33. Farrington, D.P. Early predictors of adolescent aggression and adult violence. Violence Vict.
1989, 4, 79–100.
 273

34. Moffitt, T.E. Adolescence-limited and life-course-persistent antisocial behavior: A

developmental taxonomy. Psychol. Rev. 1993, 100, 674–701.
35. Hann, D.M. Taking Stock of Risk Factors for Child/Youth Externalizing Behavior Problems;
NIMH: Bethesda, MD, USA, 2000.
36. Liu, J. Early health risk factors for violence: Conceptualization, review of the evidence, and
implications. Aggress. Violent Behav. 2011, 16, 63–73.
37. Liu, J.; McCauley, L.A.; Zhao, Y.; Zhang, H.; Pinto-Martin, J. Cohort profile: The China
Jintan child cohort study. Int. J. Epidemiol. 2010, 39, 668–674.
38. Liu, J.; McCauley, L.; Leung, P.; Wang, B.; Needleman, H.; Pinto-Martin, J.; Jintan Cohort
Group. Community-based participatory research (CBPR) approach to study children’s health
in China: Experiences and reflections. Int. J. Nurs. Stud. 2011, 48, 904–913.
39. Liu, J.; Ai, Y-X.; Hanlon, A.; Shi, Z.; Dickerman, B.; Compher, C. Micronutrients deficiency
and associated sociodemographic factors in Chinese children. World J. Pediatr. 2011, 7, 217–
223.
40. Dong, S.F.; Zhu, Z.G.; Liu, W.M.; Zhao, W.X. Determination of Ca, Mg, Fe, Mn, Cu and
Zn in lixin pill by atomic absorption spectrophotometry. Spectrosc. Spect. Anal. 2001, 21,
391–392.
41. Achenbach, T.; Rescorla, L. Manual for the ASEBA Preschool Forms & Profiles; University
of Vermont, Research Center for Children, Youth & Families: Burlington, VT, USA, 2000.
42. Liu, J.; Portnoy, J.; Raine, A. Association between a marker for prenatal testosterone exposure
and externalizing behavior problems in children. Dev. Psychopathol. 2012, 24, 771–782.
43. Achenbach, T.M.; Rescorla, L.A. Manual for the ASEBA School-Age: Forms & Profiles;
University of Vermont, Research Center for Children, Youth & Families: Burlington, VT,
USA, 2001.
44. Liu, J.; Cheng, H.; Leung, P.W. The application of the preschool child behavior checklist and
the caregiver-teacher report form to mainland Chinese children: Syndrome structure, gender
differences, country effects, and inter-informant agreement. J. Abnorm. Child Psychol. 2011,
39, 251–264.
45. STATA, Stata User’s Guide Release 11; Stata Press: College Station, TX, USA, 2009.
46. Black, M.M. Zinc deficiency and child development. Am. J. Clin. Nutr. 1998, 68, 464S–469S.
47. Grantham-McGregor, S.M.; Ani, C.C. The role of micronutrients in psychomotor and
cognitive development. Br. Med. Bull. 1999, 55, 511–527.
48. Vallee, B.L.; Falchuk, K.H. The biochemical basis of zinc physiology. Physiol. Rev. 1993, 73,
79–118.
49. Frederickson, C.; Danscher, G. Zinc-containing neurons in hippocampus and related CNS
structures. Prog. Brain Res. 1990, 83, 71–84.
50. Oner, O.; Oner, P.; Bozkurt, O.H.; Odabas, E.; Keser, N.; Karadag, H.; Kizilgün, M. Effects of
zinc and ferritin levels on parent and teacher reported symptom scores in attention deficit
hyperactivity disorder. Child Psychiatry Hum. Dev. 2010, 41, 441–447.
51. Tassabehji, N.M.; Corniola, R.S.; Alshingiti, A.; Levenson, C.W. Zinc deficiency induces
depression-like symptoms in adult rats. Physiol. Behav. 2008, 95, 365–369.
 274

52. DiGirolamo, A.M.; Ramirez-Zea, M. Role of zinc in maternal and child mental health. Am. J.
Clin. Nutr. 2009, 89, 940s–945s.
53. Black, M.M.; Baqui, A.H.; Zaman, K.; Ake Persson, L.; El Arifeen, S.; Le, K.; McNary, S.W.;
Parveen, M.; Hamadani, J.D.; Black, R.E. Iron and zinc supplementation promote motor
development and exploratory behavior among Bangladeshi infants. Am. J. Clin. Nutr. 2004,
80, 903–910.
54. Shi, Z.; Lien, N.; Kumar, B.N.; Dalen, I.; Holmboe-Ottesen, G. The sociodemographic
correlates of nutritional status of school adolescents in Jiangsu Province, China. J. Adolesc.
Health 2005, 37, 313–322.
55. Shi, Z.; Hu, X.; Yuan, B.; Pan, X.; Dai, Y.; Holmboe-Ottesen, G. Association between dietary
patterns and anaemia in adults from Jiangsu Province in Eastern China. Br. J. Nutr. 2006, 96,
906–912.
56. Volkow, N.D.; Wang, G.J.; Newcorn, J.; Telang, F.; Solanto, M.V.; Fowler, J.S.; Logan, J.;
Ma, Y.; Schulz, K.; Pradhan, K.; et al. Depressed dopamine activity in caudate and
preliminary evidence of limbic involvement in adults with attention-deficit/hyperactivity
disorder. Arch. Gen. Psychiatry 2007, 64, 932–940.
57. Erikson, K.; Pinero, D.; Connor, J.; Beard, J. Iron status and distribution on iron in the brains
of the developing rats. J. Nutr. 1997, 127, 2030–2038.
58. Krause, J. SPECT and PET of the dopamine transporter in attention-deficit/hyperactivity
disorder. Expert Rev. Neurother. 2008, 8, 611–625.
59. Luthar, S.S. The culture of affluence: Psychological costs of material wealth. Child Dev. 2003,
74, 1581–1593.
60. Luthar, S.S.; Becker, B.E. Privileged but pressured? A study of affluent youth. Child Dev.
2002, 73, 1593–1610.
61. Buss, D.M. The evolution of happiness. Am. Psychol. 2000, 55, 15–23.
62. Downey, G.; Coyne, J.C. Children of depressed parents: An integrative review. Psychol. Bull.
1990, 108, 50–76.
63. Campbell, S.B. Hard-to-manage preschool boys: Externalizing behavior, social competence,
and family context at two-year followup. J. Abnorm. Child Psychol. 1994, 22, 147–166.
64. Galler, J.R.; Waber, D.; Harrison, R.; Ramsey, F.; Liu, J.; Raine, A.; Venables, P.H.;
Mednick, S.A. Behavioral effects of childhood malnutrition. Am. J. Psychiatry 2005, 162,
1760–1761.

© 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).
 275

5. Nutrition Assessment and Body Composition

Reprinted from Nutrients. Cite as: Grewal, N.K.; Mosdøl, A.; Aunan, M.B.; Monsen, C.; Torheim, L.E.
Development and Pilot Testing of 24-Hour Multiple-Pass Recall to Assess Dietary Intake of
Toddlers of Somali- and Iraqi-Born Mothers Living in Norway. Nutrients 2014, 6, 2333-2347.

Development and Pilot Testing of 24-Hour Multiple-Pass

Recall to Assess Dietary Intake of Toddlers of Somali- and
Iraqi-Born Mothers Living in Norway
Navnit Kaur Grewal 1,*, Annhild Mosdøl 2,3, Marte Bergsund Aunan 2, Carina Monsen 2
and Liv Elin Torheim 1,2
1
Fafo Institute for Applied International Studies, P.O. Box 2947 Tøyen, Oslo NO-0608, Norway;
E-Mail: liv.elin.torheim@hioa.no
2
Department of Health, Nutrition and Management, Faculty of Health Sciences,
Oslo and Akershus University College of Applied Sciences, P.O. Box 4 St. Olavs plass,
Oslo NO-0130, Norway; E-Mails: annhild.mosdol@kunnskapssenteret.no (A.M.);
marteaunan@hotmail.com (M.B.A.); carina_monsen86@hotmail.com (C.M.)
3
Department of Evidence Summaries, The Norwegian Knowledge Centre for Health Services,
P.O. Box 7004 St. Olavs plass, Oslo NO-0130, Norway

* Author to whom correspondence should be addressed; E-Mail: navnit.grewal@fafo.no;

Tel.: +47-970-781-15; Fax: +47-220-887-00.

Received: 7 February 2014; in revised form: 4 June 2014 / Accepted: 6 June 2014 /
Published: 19 June 2014

Abstract: The aim of this study was to develop, test, and evaluate a 24-h recall procedure
to assess the dietary intake of toddlers of Somali- and Iraqi-born mothers living in Norway.
A protocol for a 24-h multiple-pass recall procedure, registration forms, and visual tools
(a picture library for food identification and portion size estimation) was developed and
tested in 12 mothers from Somalia and Iraq with children aged 10–21 months. Five female
field workers were recruited and trained to conduct the interviews. Evaluation data for the
24-h recall procedure were collected from both the mothers and the field workers. Nutrient
intake was calculated using a Norwegian dietary calculation system. Each child’s estimated
energy intake was compared with its estimated energy requirement. Both the mothers and
the field workers found the method feasible and the visual tools useful. The estimated
energy intake corresponded well with the estimated energy requirement for most of the
children (within mean ± 2 SD, except for three). The pilot study identified the need for
 276

additional foods in the picture library and some crucial aspects in training and supervising
the field workers to reduce sources of error in the data collection.

Keywords: 24-h recall; dietary assessment; infants; toddlers; Somalia; Iraq;

immigrants; Norway

1. Introduction

Dietary assessment studies are important for the development of public nutrition policies and
interventions because they can identify population groups at risk of nutritional health problems and
describe their dietary habits [1]. In Europe, some immigrant groups have a higher risk of
developing nutrition-related diseases than the host population, in particular overweight/obesity and
diabetes mellitus type 2 [2–4]. This may be due to changes in dietary habits and physical activity
patterns influenced by a process of acculturation, urbanization and westernization [2,5]. Only
relatively few studies describe dietary habits in immigrant groups, but these indicate diets with
increased consumption of processed food after migration that replace healthy dietary components,
such as fruits, vegetables, nuts, and whole grains [2,3].
Dietary studies among immigrant groups are hampered by a lack of suitable cultural-sensitive
assessment methods and data collection procedures. Missing food composition data on ethnic foods
and the use of dietary assessment methods, which are not critically assessed for suitability in these
groups, may limit the reliability of dietary intake data among immigrants [2]. Furthermore, various
methodological aspects, such as sampling and recruitment, tools and method of administration,
among others, often require special attention [6]. In addition, dietary assessment among infants and
children in particular has several inherent challenges, and these might be amplified among immigrant
groups [1,7]. It is, therefore, important to exercise considerable caution when conducting dietary
studies in this study group in order to reduce possible errors and increase validity.
Previous Norwegian national dietary surveys among infants (“Spedkost”, aged six months and
12 months) and toddlers (“Småbarnskost”, aged 24 months) excluded children of mothers born
outside Scandinavia [8–10]. There are two main reasons for this exclusion: (1) the dietary
assessment method used in the study was a food frequency questionnaire (FFQ) and the food list
was not adapted to non-Scandinavian food habits and (2) the FFQ was only available in
Norwegian. Thus, the method was less suitable for population groups with atypical food habits,
poor Norwegian skills or lower literacy levels. The researchers expressed a need for separate
studies among children of immigrant parents using more appropriate methods [11].
The “InnBaKost” study was initiated to address the limited knowledge about dietary habits and
health among children in Norway with immigrant backgrounds. Children of Somali- and Iraqi-born
mothers were chosen because they are the two non-Western immigrant groups currently with the
highest number of births in Norway [12]. The aim of this research project was to collect
information about breastfeeding practices and feeding patterns among infants aged six months with
follow-up at 12 and 24 months to supplement the Spedkost and Småbarnskost surveys [8–10].
 277

A modified FFQ was used for the data collection at six months of age. However, a structured 24-h
recall method was considered more appropriate at ages 12 and 24 months. This because the FFQ
requires that variations in food habits must be known and included in the food list to develop
suitable FFQs [13,14]. The aim of the present study was to develop, pilot test and evaluate a
protocol for a 24-h recall procedure, with registration forms and visual tools, to assess the dietary
intake of toddlers of Somali- and Iraqi-born mothers living in Norway.

2. Methods

2.1. Subjects and Study Design

The pilot study was carried out January–June, 2013. Twelve Somali- and Iraqi-born mothers
with children aged 10–21 months living in Oslo and Akershus counties, Norway, were recruited
through several methods: the Norwegian National Population Register, open kindergartens and by
using the snowball method. Inclusion criteria were the mothers’ country of birth and the child being
approximately 12 months old, born in Norway and with no serious health problem or disease
requiring a special diet. The mothers received a bilingual information letter and provided written
consent. Respondents received a shop voucher after completing two recalls. The study was
approved by the Norwegian Regional Committees for Medical and Health Research Ethics.
To measure the dietary intake among toddlers, a structured 24-h recall method was used [15].
In the 24-h recall method, the mothers were interviewed twice, usually 1–2 weeks apart, by trained
field workers about the exact food and beverage intake of their child during the preceding 24 h. If
other caretakers were involved, the mother was asked to obtain information about the child’s food
consumption while under their care. A researcher (C.M., M.B.A., or N.K.G.) was present and
observed the interviews. In addition to the food and beverage intake of the child, information about
the performance of the interviews and methods was collected through an evaluation form.

2.2. The 24-H Recall Method

2.2.1. Picture Library for Food Identification

To help the mothers and the field workers identify the correct foods given to the child, a library
with pictures of food items commonly eaten by children in Norway was developed. A list of food
items to be included in the library was made based on knowledge about Norwegian children’s
dietary intake. In addition, food items identified as eaten by Somali and Iraqi children through an
informal qualitative prestudy and food items suggested by the field workers were also included.
The foods were photographed in supermarkets and independent shops owned by immigrants in
Oslo, with permission from the owners/managers. A Canon Ixus 860 IS digital camera was used,
and the pictures were edited in iPhoto on a MacBook Pro (Apple Inc., Cupertino, CA, USA).
The library contained pictures of a wide selection of industrially produced baby foods for
children aged 8–15 months, as well as other foods and beverages. Before the pilot study, the field
workers suggested adding Nido milk powder, different types of meat, cheese and biscuits. The
 278

“bread scale”, a Norwegian labeling scheme for fiber and wholemeal content of bread, was also
included in the library.
To ease retrieval of pictures during the interviews, the 336 unique pictures were categorized in
19 different folders on an iPad (Table 1). The folders contained between 4 and 50 pictures, with the
largest number of pictures in the folder “fruits and vegetables”. Because certain foods may be
categorized differently by different people, some pictures were placed in more than one folder. For
instance, smoothies appeared both in the folders “snacks” and “juice and nectar”. Thus, the final
library contained 405 pictures in 19 folders.

Table 1. Number of pictures in each folder of the picture library.

Food Folder Number of Pictures
Baby cereals 16
Snacks 33
Infant formula 16
Ready-made meals 10
Bread spreads 19
Dinner 39
Yoghurt and desserts 29
Oils and butter 13
Dairy products 13
Fruits and vegetables 50
Breads 22
Pasta, rice and beans 17
Supplements 8
Milk 38
Juice and nectar 46
Soda 4
Squash, lemonade, etc. 17
Meat 4
Biscuits 11
Total 405

2.2.2. Photographic Booklet and Measuring Equipment for Portion Size Estimation

This study used a photographic booklet for portion size estimation developed for the Spedkost
and Småbarnskost surveys [8–10]. It included 17 color photograph series of selected food items
representing different, usually four, portion sizes appropriate for toddlers ranging from small (A) to
large (D), with up to six different portion sizes for baby cereal. The mother used the booklet as a
tool to identify portion sizes eaten by her child. The field workers also brought a kitchen scale and
three measuring cups (in deciliters and milliliters) to weigh or measure foods or volume in
tableware from the respondents’ homes whenever possible. Frequently, both methods were used to
compare the results.

2.2.3. 24-H Recall Protocol and Registration Form

The protocol contained instructions on how the field workers should conduct the 24-h recalls
based on standard procedures for face-to-face 24-h recall in the literature [16]. A 24-h period was
 279

defined as starting at the time the child woke up the previous day until the time the child woke up
the day of interview. The field worker informed the mother about the recall procedure at the
beginning of each interview and recorded the answers in specially designed paper-based forms that
matched the three-stage, multiple-pass interviewing technique [16]. In the first pass, the mother
was asked to give a complete overview of all foods and beverages consumed during the 24-h
period. If the mother was still breastfeeding her child, each breastfeeding occurrence was
registered. In the second pass, a detailed description of each food and beverage consumed were
obtained. This included type of product, brand names, cooking methods, amounts, and food
leftovers. In this pass, the picture library was used to identify foods, and the photographic booklet
and measuring equipment to estimate amounts. The field workers used standardized probe
questions to collect specific details. In the last pass, the field workers summarized and reviewed
the information to ensure that all items were recorded correctly. This phase also included a
checklist of foods and beverages that are often forgotten, such as water, snacks, and supplements.
Representativeness of the day and food allergies/intolerances was also registered. A separate
questionnaire covered background information of the mother and child.

2.3. Training of Field Workers

Five female field workers were recruited to conduct the interviews, of which two spoke Somali
and three Arabic. All of them spoke fluent Norwegian and one of the Arabic-speaking field workers
also spoke Kurdish. Thus, the mothers could choose to speak either Norwegian or their own
language during the interviews.
The field workers received 1–2 weeks of training on how to conduct 24-h recalls according to
the protocol using the forms and tools. Practice took place in pairs and in plenary using different
languages. The training particularly emphasized how to ask follow-up questions to make sure all
food items were registered, to identify the correct food items and to estimate portion sizes as
accurately as possible.

2.4. Pilot Testing of the Procedures for 24-H Recall

The pilot study enabled a full appraisal of all aspects of the 24-h recall procedure. The field
worker and the observer recorded data and answered questions regarding the method after each
interview using an evaluation form (Table 2). The mothers were asked for their views on the
method, including the visual tools, after the second interview.
The dietary data obtained from the 24-h recalls were manually coded and entered by C.M.,
M.B.A., and N.K.G. in a software system (KBS, database AE-10) developed at the Department of
Nutrition, University of Oslo, Norway. The food database in KBS was mainly based on the official
Norwegian food composition table. Breast milk intake among the breastfed children was calculated
by multiplying the number of feeding events by an estimated breast milk intake per feed of
124 mL. This amount of breast milk per feed was derived from an estimated daily breast milk
intake of 497 mL among 12-month old children in developed countries [17] divided by the average
breastfeeding frequency in Norwegian 12-month old breastfed children of 4 times per day [9]
(497 mL/4 feeds = 124 mL/feed).
 280

As an objective measure of validity for this pilot test, each child’s estimated energy intake (EEI)
was compared with its estimated energy requirement (EER). According to the Nordic Nutrition
Recommendations, estimated average daily energy requirement for 12-month-old boys is 337 kJ/kg
and 333 kJ/kg for girls [18]. We did not measure the children’s weight, but recorded the weight
registered at clinic for the 12-month health check-up (Table 3). However, eight of the children were
interviewed at ages between 13 and 21 months and two were interviewed younger than 12 months
of age. One of the children younger than 12 months of age had his body weight measured at the 8-
month consultation at the child health center. Thus, for 9 of the 11 children with registered body
weight, at least one month had passed between the weighing and the 24-h recall. To adjust for this,
an estimate of monthly weight gain was calculated using the World Health Organization’s growth
standards for children between 0 and 24 months [19]. The estimated weight gain for boys
8–21 months of age varies by month and is highest from 8 to 9 months (3.25%) and decreases
gradually to 1.76% from 20 to 21 months. For girls, weight gain from 12 to 13 months was found to
be 2.64% and from 13 to 14 months average weight gain is 2.46%. Each child’s body weight was
thus calculated by adding the monthly estimated weight gain to its weight. For the child with no
records of body weight registered, the average weight of 11-month old girls was used. Using the
estimated body weight at the time of interview, EER for each child was calculated and compared
with the EEI calculated from the 24-h recalls, using the mean intake of the two recalls. Differences
between EER and EEI were tested with paired samples t-test. Bland-Altman plot [20] was used to
visualize the dispersion between EER and EEI. Linear regression analysis was applied to study
whether there was any relationship between the mean of the estimates EER and EEI and the
difference between the two estimates.

Table 2. Evaluation form for the pilot study.

Source of
Evaluation Topic Question Asked
Information
Time spent on picture library (iPad)?
Time spent by the field Time spent on photographic booklet?
worker Time spent on measuring equipment?
Observation by Other notes?
researchers Use of visual tools Which pictures were used most frequently or not at all?
Did the field workers ask the questions in the same way?
Standardisation of
Did they follow the protocol?
methods/field workers
Did they use the visual tools?
Clarity of questions Were any of the questions difficult to answer/unclear? If yes, which and why?
Did you miss pictures of any foods/beverages?
Missing pictures Are there some foods/beverages you give your child often, but not yesterday?
Questions to
Are there any other foods/beverages you know Somali/Iraqi children often eat/drink?
respondents
Did the portion sizes in the booklet match the portion sizes your child usually eats?
Portion sizes Was it easier to estimate the amount the child had eaten by using the booklet,
measuring equipment or by showing it on/in the plate/cup used?
Was the protocol easy to understand? If no, why not?
24-h recall protocol How did you experience the different passes during the interview? Was it easy to
distinguish these from each other?
Questions to
How did you experience using the picture library during the interview? Was it user
field workers Picture library
friendly? If no, why not?
How did you experience using the photographic booklet to estimate portion sizes?
Photographic booklet
Did you miss photos of any foods/beverages?
 281

How did you experience to estimate amounts using the measuring equipment?
Measuring equipment
Did you miss any equipment?
How did you experience using the form?
Registration form for
Was the order of items logical to you?
24-h recall
Was there enough space to write? If no, where did you want more space?
Table 3. Estimated energy requirements and energy intake among children (10–21 months)
with mothers from Iraq (ID 1–5) and Somalia (ID 6–12) living in Norway.
Estimated Estimated
Body Age at Percentage
Average Daily Body Weight EER at Time EEI at Time
Weight at Time of Differences
ID Sex Energy at Time of of Interview of Interview
12 Months Interview between EER
Requirements Interview (g) (kJ/day) b (kJ/day) c
(g) (months) a
and EEI (%) d
(kJ/kg)
1 F 333 8440 14 8855 2949 2843 í
2 M 337 11,600 13 11,864 3998 3486 í
3 M 337 11,083 21 13,268 4471 3649 í
e
4 F 333 8719 11 8719 2903 4415 41
5 M 337 8300 12 8300 2797 3102 10
6 F 333 10,000 13 10,246 3412 4043 17
7 M 337 8200 12 8200 2763 2764 0
8 F 333 9970 14 10,460 3483 3157 í
9 M 337 10,000 14 10,466 3527 2232 í
10 M 337 11,000 14 11,513 3880 4635 18
f
11 M 337 9890 10 10,509 3542 3556 0
12 F 333 9270 13 9498 3163 4636 38
Mean 335 9706 13 10,158 3407 3543 18 g
SD 2 1146 3 1544 527 773 15 g
a
Estimated body weight at time of interview calculated based on average growth rate from World Health
Organization’s growth standards [19] multiplied by number of months between time of weighing and
b
time of interview; Estimated body weight at time of interview multiplied with estimated average
c d
requirement per kilogram; Mean estimated energy intake of the two recalls; Calculated as percent
difference of mean. Difference between EER and EEI tested with paired samples t-test: p = 0.58; e Body
weight not registered. Average weight for girls at 11 months of age used as reference [19]; f Body weight
at 8 months of age; g Calculated using absolute values of percentage differences

3. Results

3.1. Subjects

A total of 28 Somali-born mothers were asked to participate in the pilot study, and 13 consented.
However, only seven of these showed up to the appointed interview. Of the fourteen 24-h recalls,
eight were conducted in Norwegian and six in Somali. Likewise, 48 Iraqi mothers were contacted,
seven consented, but only five showed up at the interview. Five of the 24-h recalls were conducted
in Kurdish, four in Arabic and one in Norwegian. Among the 12 participating mothers, mean age
was 31 (range 22–42) and the average number of years lived in Norway was 15 (range 3–24).
Three mothers had no education from Norway. Two of them had, however, completed a Norwegian
 282

course. Seven mothers had completed high school education and two had completed higher
education. Seven mothers had more than one child.

3.2. Results from the Evaluation Form

The mean (minimum-maximum) time spent on the total 24-h recall interviews was 47 (20–75)
min. There was a decreasing trend in the time spent on each interview conducted over time in both
the Somali and Iraqi groups. The repeat interviews were conducted by the same field worker,
except for three of the Somali mothers and one Iraqi mother where two different field workers
conducted the interviews.
The field workers spent an average of four minutes during each interview showing pictures on
the iPad. The visual tools were used during all interviews. Of the 19 folders in the picture library,
eight were used by the Somali mothers to identify foods. The most frequently used folders were
“baby cereal” (seven interviews) and “oils and butter” (four interviews). Among the Iraqi mothers,
pictures from 13 of the 19 folders were used. The most frequently used folders in this group were
“breads” (eight interviews) and “baby cereal” (four interviews). Although the mothers browsed
through all folders to identify foods given to the child, none of the Somali or Iraqi mothers used or
identified foods from the five folders “ready-made meals”, “fruits and vegetables”, “soda”,
“squash, lemonade, etc.” or “meat”.
Eleven of the 17 colored photograph series were used by the Somali mothers to estimate portion
sizes eaten by the child. The portion sizes of baby cereal and butter were the most frequently used
total in 10 and 9 of the 14 interviews, respectively. The Iraqi mothers used 10 of the 17 photograph
series of portion sizes to estimate foods consumed by the child. The most frequently used series
were the portion sizes for milk and butter, which were referred to in 9 and 5 of the 10 interviews,
respectively. The measuring equipment was used together with the photographic booklet in the first
interviews, but over time the interviewers favored the photographic booklet over actual
measurements. Reasons given for this shift were that measurements were time consuming and
difficult to use when the interviews were conducted outside the informants’ homes. When mothers
were asked to identify amount with both the photographic booklet and by measurements of actual
foods, these seemed to correspond well.
The protocol was mostly only used during the last pass, when the field workers were going
through the checklist of foods and beverages often forgotten. When the interviews were conducted
in Norwegian, the observers noted that the field workers consistently asked about added
foods/ingredients, brands and amounts consumed. It was sometimes difficult for the field workers
to write down recipes and cooking methods because of limited space on the forms. However, the
amount of food eaten by the child was usually asked about and written down clearly.
The mothers expressed that the picture library was a good tool to be reminded of and to identify
the type of foods given to the child. It was especially useful for remembering brand names. Among
pictures missing in the picture library, some mothers mentioned different types of rice, fruit purees,
bread spreads, breads, butter, baby cereals, and yoghurts, as well as Weetabix and prunes.
One of the topics that emerged repeatedly was how difficult it was to estimate portion sizes.
Six mothers mentioned the difficulties in estimating the amount of bread eaten by the child, without
 283

pictures of bread in the booklet. Other pictures of portion sizes mentioned as missing were lasagna,
spaghetti and pancakes. Five mothers expressed that pre-packed industrially produced foods were
easier to estimate. All mothers found that the portion size options in the photographic booklet
matched amounts the child usually ate. Ten mentioned that it was easier to show amounts of
foods eaten using the booklet, whereas illustrating amounts of beverages was easier using the
measuring equipment.
Most mothers said that the day of interview represented the typical foods given to the child,
only three recall days were considered non-representative for the typical foods given. Two foods
(bread and bulgur) were mentioned by two mothers as being typical foods given to the child, but
not during the days in question. Pancakes (anjera), Weetabix and juice were mentioned by two
Somali mothers as cultural relevant foods often given to children, while four Iraqi mothers
mentioned different types of staple foods and vegetables, such as okra.
All field workers found the protocol easy to understand. Finding pictures in the picture library
took time to begin with, but became easier after a few interviews. The photographic booklet was
judged as a good tool for estimating portion sizes. However, similarly to the mothers, they missed
portion size pictures of pasta and bread.
The registration form was described as clear and easy to understand, but the field workers
missed more space to write down recipes.

3.3. Results from the 24-H Recalls

Six mothers were still breastfeeding their children (four Iraqi and two Somali mothers). The
average breastfeeding frequency was two times per day (range one to four times per day). Foods
given to the children included bread, porridge, different fruit and vegetables, snacks and
supplements. The porridge was either industrial produced or home-made using oatmeal or bulgur.
A couple of mothers also added margarine, olive oil, salt and/or sugar when preparing the porridge.
Some of the foods registered that are not commonly given to Norwegian children were nan bread,
feta cheese, Turkish delight, bulgur porridge, pancakes (anjera), and seeds. None of the mothers
reported using ready-made dinners, as most of them mentioned that they did not trust the contents
and the ready-made dinners were not considered to be fresh. The home-made dinners often
constituted of different staple foods, vegetables, meat, and fish.

3.4. Results from Energy Intake Estimation

Table 3 presents the comparison of each child’s EER with its EEI calculated from the 24-h
recall. The percent difference between EER and EEI was in the range of ±0%–10% for five of the
children and in the range of ±11%–20% for four of the children, whereas for three of the children
the percentage of difference between the two estimates was ±38%, 41% or 45%, respectively.
The mean (SD) for EER and EEI was 3407 (527) kJ/day and 3543 (773) kJ/day, respectively, and
the difference was not significant, as tested with paired samples t-test. The Bland-Altman plot
(Figure 1) showed large individual variations in the differences between EER and EEI but no clear
pattern. A linear regression analysis testing the relationship between the mean of the estimates EER
 284

and EEI and the difference between the two estimates was not significant, p = 0.24. This indicates
that the difference between the two estimates is not related to the magnitude of the estimates.
Excluding child number 4, for whom there was no registered weight, did not change the results.

Figure 1. The difference between estimated energy requirement (EER) and estimated energy
intake (EEI), plotted against the mean of EER and EEI (n = 12). SD = Standard deviation.

Mean +2SD
EER-EEI (kJ/day)

Mean

Mean -2SD

Mean of EER and EEI (kJ/day)

4. Discussion

For the InnBaKost study, we developed a protocol for a 24-h recall procedure, including a
picture library to assist in identifying the correct foods eaten. In addition, a photographic booklet
was used for portion size estimation. Although the latter approach has become a common method
for portion size estimation [21,22], including dietary assessment in children [23,24], the use of a
picture library is a rather novel approach. The hypothesis was that the picture library would be a
useful tool to identify the correct food and brand, particularly for dietary assessment among
immigrant mothers with varying levels of language and literacy skills.
A review conducted by Burrows et al. (2010), indicates that weighed food records provide the
best dietary estimates for younger children aged 0.5 to 4 years, while 24-h multiple-pass recall that
uses parents as reporters is the most accurate method to estimate total energy intake in children
aged 4 to 11 years [25]. The weighed food record method requires both motivation and good
literacy skills and is often time-consuming. Thus, the method has been considered to be less
suitable for dietary assessment in immigrants, as the method has led to misreporting and dropout in
immigrant groups due to the burden and time consumption the method carries [26]. The face-to-
face FFQs and multiple-pass 24-h recalls are reported to be the two most frequently used methods
with immigrant populations in Europe [6]. The 24-h recall is more flexible because it can capture
all foods and beverages consumed the preceding day, with no assumptions about the food culture or
dependency on literacy levels. In addition, as seen from the few recalls in this pilot, some mothers
gave selected atypical foods to their children and mostly made home-made dinners, which may
 285

vary from the general Norwegian population in regards to composition and preparation method.
The 24-h recall has therefore been recommended as the most optimal method for many immigrant
groups and is considered to provide valid information among children [26,27]. In addition, the
interactive nature and the personal contact of the method may contribute to more reliable data
collection, although social desirability bias may cause some misreporting [28]. The multiple-pass
technique is considered to give the most exact estimates, and limit misreporting, because the probing
questions encourage the respondent to remember more of the foods consumed [16]. The respondent
burden is usually small compared to weighed records [29].
The protocol was used sparingly during the interviews, because the field workers expressed that
they already knew the content in the protocol and that it was difficult to focus on the protocol while
registering the child’s food consumption. Thus, it was recommended that important guidelines from
the protocol could be included in the 24-h recall registration form instead. The decreasing time
spent on the second interview with each mother was mostly due to the mother being more prepared
and that the background information was already collected. Another reason for the decline in time
spent may have been that the field workers became more familiar with the method and navigated
the picture library and photographic booklet more easily. The measuring equipment was initially
used together with the photographic booklet to see how well both measurements corresponded, but
both the field workers and mothers expressed that it was too time-consuming.
Both the mothers and field workers reported the picture library to be a good tool to identify
foods given to the child. It was mostly used when the interviews were conducted outside the
respondents’ homes because the mothers could show foods available in the home. The use of a
picture library similar to this has not been described by many; however, the use of photo images
has been reported to be useful as a memory aid for respondents during 24-h recalls [30,31]. The
picture library seemed to strengthen the mothers’ ability to report the correct food and reduce
misunderstandings. However, the pilot study revealed many desired additions to both the picture
library and booklet.
Portion size estimation is one of the main challenges in dietary assessment studies. Estimating
amounts eaten other than direct weighing may contribute to a source of error, both among children
and adults [7,22,24,32]. The photographic booklet was considered to be a good tool for estimating
portion sizes among the field workers and the mothers, as has also been reported in several other
studies [21,24,30,33]. A study by Lillegaard et al. (2005) showed that children and adolescents
could accurately estimate portion sizes of pre-weighed foods by viewing photographs,
approximately 60% of the comparisons were made correctly [24]. The estimations were more
accurate when the served portions had the exact appearance as the food portrayed in the
photographic booklet [24]. Thus, the arguments can be made that more picture series in our
photographic booklet may be favorable rather than using pictures of similar foods. The studies
further emphasize the importance of validation studies to test the applicability of photographs for
estimating current portions and actual consumption [21,22], especially among immigrant
groups [13]. This was not done in this pilot study, but should be considered in the future.
Assessing children’s food intake accurately can be difficult for a number of reasons. Infants and
toddlers cannot account for their food intake, but parents are seen as reliable sources when
 286

affirming their children’s consumption of food [25,34]. Efforts should be made to assess foods
eaten outside the home or with other caretakers; for instance, at the kindergarten or with family
members. A possible challenge may be that the level of reporting and motivation may vary for each
caretaker [7]. In the pilot, one of the Somali fathers was on paternity leave and was in charge of the
child’s diet at the time; therefore, he was interviewed together with the mother. Among the Iraqi
mothers, only one mother reported that her child had spent much of the day with a nanny.
Although, this did not apply for many of the mothers in the pilot, it should be taken into
consideration for larger studies and dietary assessment of somewhat older children. Potential
solutions may be to ask the mothers prior to the interview if the child has other caretakers and if it
may be possible to include them to obtain information about their child’s food consumption during
their supervision.
In regard to EEI, it seemed to correspond well with the EER for most of the children (within ± 2 SD
of the average of the two estimates) except for three. The comparison of EEI and EER has some
weaknesses and can only give an indication of whether the method is suitable for capturing habitual
energy intake on a group basis. First, each child’s EER might not reflect the true energy
requirement of the child because an energy requirement is highly variable between children of the
same age and weight [35]. There is also intraindividual variation in energy requirement for
children, depending on their physical activity level and growth rate [35]. Second, the energy intake
measure was simply averaged over the two days without adjustment for intraindividual variation
over time. Thus, it may not be representative of habitual energy intake [36]. Although the sample
size was small, it was encouraging that there was no consistent over- or underreporting of EEI
compared to EER.
Recruitment of study participants in itself was challenging and time consuming in this pilot
study, as it was difficult to come in contact with the target group. This was mostly due to wrong
contact numbers registered on several mothers when tried to reach by phone. Some reasons for
refusals were that they were not interested, skeptical, or had to consult their partner. It was
necessary to seek the mothers through several methods and many did not show up to appointed
interviews. The use of bilingual field workers was an advantage and enabled the recruitment of
mothers who did not speak Norwegian. Challenges related to recruitment when conducting dietary
studies with immigrants have previously been reported [6,26]. Most studies conducted with a
European immigrant population group have also used nonprobability sampling methods, such as
the convenience sampling method [6]. The need for extra effort in recruiting participants has been
described, such as using bilingual field workers, involving key leaders and including places of
worship and media, to overcome cultural barriers and ensure representativeness [6]. Although the
convenience sampling method may lead to the inclusion of highly motivated participants, there
seemed to be variations in the background characteristics of the mothers included in the pilot.
Based on the pilot study presented, some suggestions were made for improving the 24-h
multiple-pass recall method. Observations of the interviews showed that the field workers were not
actively using the protocol, and a possible solution is to incorporate the protocol into the
registration form. Other important suggestion were to include more pictures in the library and supply
the photographic booklet with portion sizes of bread in particular, but also of foods such as lasagna,
 287

pancakes and other portion sizes of meat, fish, fruits and vegetables. Furthermore, a more thorough
training and follow-up of the field workers would be required to increase the quality of the
data collection.

5. Conclusions

Experiences from the current study indicate that the 24-h multiple-pass recall method with
inclusion of visual tools is appropriate method for assessing dietary intake among toddlers of
Somali- and Iraqi-born mothers living in Norway. The picture library and photographic booklet
were considered to have an added value to the method to aid to identify and describe foods and
beverages consumed. However, for the method to be applicable, there is a need for thorough
training and follow-up of the field workers during data collection and an update of the picture
library and photographic booklet to capture foods, which were not included.

Acknowledgments

The “InnBaKost” project is financed by the Norwegian Research Council. The authors thank the
field workers who assisted with the data collection and the mothers who participated in the study.

Author Contributions

N.K.G., A.M. and L.E.T. designed the research; all authors collaborated in refining the
methodology. M.B.A. and C.M. developed the tools and collected the data. N.K.G. wrote the paper
and had responsibility for the final content. All authors contributed substantially to the
development of the manuscript. All authors read, edited, and approved the final manuscript.

Conflicts of Interest

The authors declare no conflict of interest.

References

1. Gibson, R.S. Principles of Nutritional Assessment, 2nd ed.; Oxford University Press: Oxford,
UK, 2005.
2. Gilbert, P.A.; Khokhar, S. Changing dietary habits of ethnic groups in Europe and
implications for health. Nutr. Rev. 2008, 66, 203–215.
3. Holmboe-Ottesen, G.; Wandel, M. Changes in dietary habits after migration and consequences
for health: A focus on South Asians in Europe. Food Nutr. Res. 2012, 56, doi:10.3402/
fnr.v56i0.18891.
4. Landman, J.; Cruickshank, J.K. A review of ethnicity, health and nutrition-related diseases in
relation to migration in the United Kingdom. Public Health Nutr. 2001, 4, 647–657.
5. Kleiser, C.; Mensink, G.B.; Neuhauser, H.; Schenk, L.; Kurth, B.M. Food intake of young
people with a migration background living in Germany. Public Health Nutr. 2010, 13, 324–330.
 288

6. Ngo, J.; Gurinovic, M.; Frost-Anderson, L.; Serra-Majem, L. How dietary intake methodology
is adapted for use in European immigrant population groups—A review. Br. J. Nutr. 2009,
101, S86–S94.
7. Livingstone, M.B.; Robson, P.J.; Wallace, J.M. Issues in dietary intake assessment of children
and adolescents. Br. J. Nutr. 2004, 92, S213–S222.
8. Øverby, N.C.; Kristiansen, A.L.; Andersen, L.F. Spedkost—6 Måneder. Landsomfattende
Kostholdsundersøkelse Blant 6 Måneder Gamle Barn; Helsedirektoratet, Mattilsynet and
Universitetet i Oslo: Oslo, Norway, 2008.
9. Øverby, N.C.; Kristiansen, A.L.; Andersen, L.F.; Lande, B. Spedkost—12 Måneder.
Landsomfattende Kostholdsundersøkelse Blant 12 Måneder Gamle Barn; Helsedirektoratet,
Mattilsynet and Universitetet i Oslo: Oslo, Norway, 2009.
10. Kristiansen, A.L.; Andersen, L.F.; Lande, B. Småbarnskost—2 år. Landsomfattende
Kostholdesundersøkelse Blant 2 år Gamle Barn; Helsedirektoratet, Mattilsynet and
Universitetet i Oslo: Oslo, Norway, 2009.
11. Lande, B.; Andersen, L.F.; Baerug, A.; Trygg, K.; Lund-Larsen, K.; Bjørneboe, G.-E.
Development of a method for use in a national representative dietary survey among Norwegian
infants and children—Spedkost/Småbarnskost. Nor. J. Epidemiol. 2000, 10, 43–50.
12. Statistics Norway. Population 1 January 2012 and 2013 and Changes in 2012, by Immigration
Category and Country Background. Absolute Numbers. Private communication, 2013.
13. Khokhar, S.; Ashkanani, F.; Garduno-Diaz, S.D.; Husain, W. Application of ethnic food
composition data for understanding the diet and nutrition of South Asians in the UK.
Food Chem. 2013, 140, 436–442.
14. Abu-Saad, K.; Shahar, D.R.; Vardi, H.; Fraser, D. Importance of ethnic foods as predictors of
and contributors to nutrient intake levels in a minority population. Eur. J. Clin. Nutr. 2010, 64,
S88–S94.
15. Thompson, F.E.; Subar, A.F. Dietary assessment methodology. In Nutrition in the Prevention
and Treatment of Disease; Coulston, A.M., Boushey, C.J., Eds.; Academic Press: San Diego,
CA, USA, 2008; pp. 5–6.
16. Rutishauser, I.H. Dietary intake measurements. Public Health Nutr. 2005, 8, 1100–1107.
17. Butte, N.F.; Lopez-Alarcon, M.G.; Garza, C. Nutrient Adequacy of Exclusive Breastfeeding
for the Term Infant during the First Six Months of Life; WHO: Geneva, Switzerland, 2002.
18. Nordic Council of Ministers. Nordic Nutrition Recommendations 2012 Part 1: Summary,
Principles and Use; Nordic Council of Ministers: Copenhagen, Norway, 2013.
19. World Health Organization. The WHO Child Growth Standards, 2013. Available online:
http://www.who.int/childgrowth/standards/en/ (accessed on 1 August 2013).
20. Bland, J.M.; Altman, D.G. Statistical methods for assessing agreement between two methods
of clinical measurement. Lancet 1986, 1, 307–310.
21. Ovaskainen, M.L.; Paturi, M.; Reinivuo, H.; Hannila, M.L.; Sinkko, H.; Lehtisalo, J.;
Pynnönen-Polari, O.; Männistö, S. Accuracy in the estimation of food servings against the
portions in food photographs. Eur. J. Clin. Nutr. 2008, 62, 674–681.
 289

22. De Keyzer, W.; Huybrechts, I.; de Maeyer, M.; Ocké, M.; Slimani, N.; van’t Peer, P.; de Henauw,
S. Food photographs in nutritional surveillance: Errors in portion size estimation using
drawings of bread and photographs of margarine and beverages consumption. Br. J. Nutr.
2011, 105, 1073–1083.
23. Trolle, E.; Vandevijvere, S.; Ruprich, J.; Ege, M.; Dofková, M.; de Boer, E.; Ocké, M.
Validation of a food quantification picture book targeting children of 0–10 years of age for
pan-European and national dietary surveys. Br. J. Nutr. 2013, 110, 2298–2308.
24. Lillegaard, I.T.; Overby, N.C.; Andersen, L.F. Can children and adolescents use photographs
of food to estimate portion sizes? Eur. J. Clin. Nutr. 2005, 59, 611–617.
25. Burrows, T.L.; Martin, R.J.; Collins, C.E. A systematic review of the validity of dietary
assessment methods in children when compared with the method of doubly labeled water. J.
Am. Diet. Assoc. 2010, 110, 1501–1510.
26. Garduno-Diaz, S.D.; Husain, W.; Ashkanani, F.; Kokhar, S. Meeting challenges related to the
dietary assessment of ethnic minority populations. J. Hum. Nutr. Diet. 2013, doi:10.1111/
jhn.12153.
27. Donin, A.S.; Nightingale, C.M.; Owen, C.G.; Rudnicka, A.R.; McNamara, M.C.; Prynne, C.J.;
Stephen, A.M.; Whicup, P.H. Nutritional composition of the diets of South Asian, black
African-Caribbean and white European children in the United Kingdom: The Child Heart and
Health Study in England (CHASE). Br. J. Nutr. 2010, 104, 276–285.
28. Satia-Abouta, J.; Patterson, R.E.; Neuhouser, M.L.; Elder, J. Dietary acculturation:
Applications to nutrition research and dietetics. J. Am. Diet. Assoc. 2002, 102, 1105–1118.
29. Biro, G.; Hulshof, K.F.; Ovesen, L.; Amorim Cruz, J.A.; EFCOSUM Group. Selection of
methodology to assess food intake. Eur. J. Clin. Nutr. 2002, 56, S25–S32.
30. Lazarte, C.E.; Encinas, M.E.; Alegre, C.; Granfeldt, Y. Validation of digital photographs, as a
tool in 24-h recall, for the improvement of dietary assessment among rural populations in
developing countries. Nutr. J. 2012, 11, 61.
31. Arab, L.; Estrin, D.; Kim, D.H.; Burke, J.; Goldman, J. Feasibility testing of an automated
image-capture method to aid dietary recall. Eur. J. Clin. Nutr. 2011, 65, 1156–1162.
32. Frobisher, C.; Maxwell, S.M. The estimation of food portion sizes: A comparison between
using descriptions of portion sizes and a photographic food atlas by children and adults.
J. Hum. Nutr. Diet. 2003, 16, 181–188.
33. Huybregts, L.; Roberfroid, D.; Lachat, C.; van Camp, J.; Kolsteren, P. Validity of photographs
for food portion estimation in a rural West African setting. Public Health Nutr. 2008, 11, 581–587.
34. Klesges, R.C.; Klesges, L.M.; Brown, G.; Frank, G.C. Validation of the 24-hour dietary recall
in preschool children. J. Am. Diet. Assoc. 1987, 87, 1383–1385.
35. Wiskin, A.E.; Davies, J.H.; Wootton, S.A.; Beattie, R.M. Energy expenditure, nutrition and
growth. Arch. Dis. Child. 2011, 96, 567–572.
36. Dodd, K.W.; Guenther, P.M.; Freedman, L.S.; Subar, A.F.; Kipnis, V.; Midthune, D.; Tooze,
J.A.; Krebs-Smith, S.M. Statistical methods for estimating usual intake of nutrients and foods:
A review of the theory. J. Am. Diet. Assoc. 2006, 106, 1640–1650.
 290

© 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).
 291

Reprinted from Nutrients. Cite as: Moltu, S.J.; Sachse, D.; Blakstad, E.W.; Strømmen, K.; Nakstad, B.;
Almaas, A.N.; Westerberg, A.C.; Rønnestad, A.; Brække, K.; Veierød, M.B.; Iversen, P.O.;
Rise, F.; Berg, J.P.; Drevon, C.A. Urinary Metabolite Profiles in Premature Infants Show Early
Postnatal Metabolic Adaptation and Maturation. Nutrients 2014, 6, 1913-1930.

Urinary Metabolite Profiles in Premature Infants Show Early

Postnatal Metabolic Adaptation and Maturation
Sissel J. Moltu 1,2,†, Daniel Sachse 3,4,5,†,*, Elin W. Blakstad 6,7, Kenneth Strømmen 2,8,
Britt Nakstad 6,7, Astrid N. Almaas 6,7, Ane C. Westerberg 2,9, Arild Rønnestad 8,
Kristin Brække 8, Marit B. Veierød 2,10, Per O. Iversen 2,11, Frode Rise 5, Jens P. Berg 3,4 and
Christian A. Drevon 2
1
Department of Pediatrics, Oslo University Hospital, P.O. Box 4950 Nydalen, Oslo 0424,
Norway; E-Mail: sissel.moltu@medisin.uio.no
2
Department of Nutrition, University of Oslo, P.O. Box 1046 Blindern, Oslo 0317, Norway;
E-Mails: kestro@ous-hf.no (K.S.); a.c.westerberg@medisin.uio.no (A.C.W.);
m.b.veierod@medisin.uio.no (M.B.V.); p.o.iversen@medisin.uio.no (P.O.I.);
c.a.drevon@medisin.uio.no (C.A.D.)
3
Department of Medical Biochemistry, University of Oslo, P.O.Box 4950 Nydalen, Oslo 0424,
Norway; E-Mail: j.p.berg@medisin.uio.no
4
Department of Medical Biochemistry, Oslo University Hospital, P.O. Box 4956 Nydalen,
Oslo 0424, Norway
5
Department of Chemistry, University of Oslo, P.O. Box 1033 Blindern, Oslo 0315, Norway;
E-Mail: frode.rise@kjemi.uio.no
6
Department of Pediatric and Adolescent Medicine, Akershus University Hospital, Lørenskog
1478, Norway; E-Mails: e.w.blakstad@medisin.uio.no (E.W.B.); britt.nakstad@medisin.uio.no
(B.N.); a.n.almaas@gmail.com (A.N.A.)
7
Institute of Clinical Medicine, University of Oslo, P.O. Box 1171 Blindern, Oslo 0318, Norway
8
Department of Neonatal Intensive Care, Oslo University Hospital, P.O. Box 4950 Nydalen,
Oslo 0424, Norway; E-Mails: aronnest@ous-hf.no (A.R.); uxkbre@ous-hf.no (K.B.)
9
Atlantis Medical University College, P.O. Box 4290 Nydalen, Oslo 0402, Norway
10
Department of Biostatistics, University of Oslo, P.O. Box 1122 Blindern, Oslo 0317, Norway
11
Department of Hematology, Oslo University Hospital, P.O. Box 4950 Nydalen, Oslo 0424,
Norway
†
These authors contributed equally to this work.

* Author to whom correspondence should be addressed; E-Mail: daniel.sachse@medisin.uio.no;

Tel.: +47-230-27251; Fax: +47-221-58796.
 292

Received: 30 January 2014; in revised form: 14 April 2014 / Accepted: 30 April 2014 /
Published: 12 May 2014

Abstract: Objectives: Early nutrition influences metabolic programming and long-term

health. We explored the urinary metabolite profiles of 48 premature infants (birth weight
< 1500 g) randomized to an enhanced or a standard diet during neonatal hospitalization.
Methods: Metabolomics using nuclear magnetic resonance spectroscopy (NMR) was
conducted on urine samples obtained during the first week of life and thereafter
fortnightly. Results: The intervention group received significantly higher amounts of
energy, protein, lipids, vitamin A, arachidonic acid and docosahexaenoic acid as
compared to the control group. Enhanced nutrition did not appear to affect the urine
profiles to an extent exceeding individual variation. However, in all infants the
glucogenic amino acids glycine, threonine, hydroxyproline and tyrosine increased
substantially during the early postnatal period, along with metabolites of the
tricarboxylic acid cycle (succinate, oxoglutarate, fumarate and citrate). The metabolite
changes correlated with postmenstrual age. Moreover, we observed elevated threonine
and glycine levels in first-week urine samples of the small for gestational age (SGA;
birth weight < 10th percentile for gestational age) as compared to the appropriate for
gestational age infants. Conclusion: This first nutri-metabolomics study in premature
infants demonstrates that the physiological adaptation during the fetal-postnatal
transition as well as maturation influences metabolism during the breastfeeding period.
Elevated glycine and threonine levels were found in the first week urine samples of the
SGA infants and emerged as potential biomarkers of an altered metabolic phenotype.

Keywords: prematurity; very low birth weight; pediatric nutrition; intervention study;
metabolomics; urine; nuclear magnetic resonance spectroscopy; glycine; threonine

1. Introduction

Despite improved perinatal medical care and increased focus on enhanced nutritional support to
premature infants, pre- and postnatal growth-restriction still occurs in 60%–100% of infants with
a very low birth weight (<1500 g) [1,2]. Premature infants with growth restriction are at risk of
impaired cognitive function and adverse metabolic and cardiovascular outcomes later in life [3,4].
Metabolic changes occurring in utero, during birth and the postnatal weaning period, seem to be
of particular importance for future health [5–7]. Nutritional alterations during these periods are
associated with a predisposition to obesity, cardiovascular diseases and associated co-morbidities
later in life [5,8]. However, the time frame for these programming effects on long-term disease risk
is controversial. Present evidence favors proactive nutritional support in premature infants to
promote growth similar to the intrauterine growth rate and to support cognitive development [1,3,4].
 293

This is in contrast with the potentially advantageous effects of relative undernutrition and slower
growth on long-term cardiovascular health [8].
Recently, we published results from a randomized, controlled trial comparing the effect of enhanced
nutritional supply (intervention) as opposed to a standard (control) diet, on postnatal growth in
premature infants with a birth weight < 1500 g [9,10]. The infants in the intervention group, with
median nutrient supplies in the upper range of current recommendations (Table 1) [9,11,12],
exhibited postnatal growth along their birth percentiles for both weight and head circumference,
whereas the control infants fell off their expected growth trajectories from birth to 36 weeks
postmenstrual age (PMA). However, a preplanned safety analysis after the enrolment of 50 infants
revealed an increased occurrence of late onset septicemia without increased mortality in the
intervention group, and it was decided to halt further recruitment [10].

Table 1. Daily nutrient supply up to four weeks after birth.

Intervention (n = 23) Control (n = 21) p
Human milk, mL/kg/day 133 (110–139) 134 (124–141) 0.37
Energy, kcal/kg/day 139 (128–145) 126 (121–128) <0.001
Protein, g/kg/day 4.0 (3.9–4.2) 3.2 (3.1–3.3) <0.001
Lipids, g/kg/day 7.3 (6.4–7.6) 5.9 (5.6–6.1) <0.001
Carbohydrates, g/kg/day 14.4 (13.4–14.8) 14.7 (14.3–15.1) 0.12
Arachidonic acid, mg/kg/day 68 (57–73) 24 (23–25) <0.001
Docosahexaenoic acid, mg/kg/day 87 (81–91) 36 (34–38) <0.001
Vitamin A, ȝg/kg/day 1300 (1105–1442) 252 (238–257) <0.001
Detailed records of actual nutrient supply was available for 44 infants [9]. Data are presented as medians
(interquartile ranges) and compared using the Mann-Whitney U test.

To assess the metabolic status of these premature infants and to explore potentially different
responses to the two diets, we used state-of-the-art nuclear magnetic resonance (NMR)-based
metabolomics to analyze urine samples obtained during the postnatal period. Metabolomics is
recognized as a powerful top-down systems biology approach that explores the genetic-
environment-health interaction [13,14]. The approach is to obtain broad snapshots of the
metabolism by detecting and quantifying hundreds of small-molecular substances (molecular mass
< 1000 Da) in tissues or body fluids, and then link them to disease or development states using
multivariate statistical methods such as principal component analysis (PCA) and partial least
squares (PLS) regression that handle and integrate large datasets [15,16]. Metabolomics of
biofluids is thought to be a promising new tool in neonatology, especially in premature infants, due
to its comprehensive and usually non-invasive nature [17]. Metabolomic analysis of urine from the
neonatal period may be used to understand metabolic processes linked to early nutrition. It may
also be used to identify biomarkers for diagnosis, prognosis and risk prediction of different
diseases [5,13,17–19].
The main objectives of the present study were to analyze urine samples from the premature
infants of the previous trial in relation to the two different nutritional exposures and to assess the
infants’ postnatal metabolic maturation. Secondary objectives were to explore potential differences
related to age, sex, infections as well as pre- and postnatal growth.
 294

2. Materials and Methods

2.1. Study Design and Population

The study was part of an open, randomized, controlled clinical trial [9,10], approved by
the Regional Committee for Medical and Health Research Ethics and in accordance with the
principles of the Helsinki Declaration. Fifty premature infants with birth weight < 1500 g were
recruited from the neonatal intensive care units at Oslo University Hospital and Akershus
University Hospital, Norway, from 17 August to 21 December 2010; 24 in the intervention group
and 26 in the control group. Exclusion criteria were congenital malformations, chromosomal
abnormalities, critical illnesses with short life expectancy and clinical syndromes known to affect
growth and development. Morbidities were registered according to routine clinical practice and
standard definitions [20–23]. Infants were classified as small for gestational age (SGA) if their
birth weight was below the 10th percentile of a reference population [24], or as appropriate for
gestational age (AGA) otherwise. Growth velocity was calculated by the exponential equation
described by Patel et al. [25].
Two infants in the control group died during the first week of life, leaving 48 infants for the
analysis [10]. Demographic and clinical characteristics are presented in Table 2. The significantly
higher occurrence of septicemia and electrolyte deviations observed in the intervention group have
been reported previously [10].

Table 2. Baseline characteristics and clinical outcomes.

Intervention (n = 24) Control (n = 24) p
Gestational age (weeks), mean (range) 28.1 (25.0–33.6) 28.5 (24.0–32.6) 0.60
Birth weight (g), mean (range) 940 (460–1311) 1083 (571–1414) 0.03
Small for gestational age, n (%) 11/24 (46%) 5/24 (21%) 0.12
Sex, boys, n (%) 15/24 (63%) 15/24 (63%) 1.00
Cesarean section, n (%) 16/24 (67%) 19/24 (79%) 0.52
APGAR-score, 5-min, mean (± SD) 7.33 (± 1.7) 7.54 (± 1.7) 0.68
Prenatal steroid treatment, n (%) 22/24 (92%) 24/24 (100%) 0.49
Late onset septicemia, n (%) 15/24 (63%) 7/24 (29%) 0.04
NEC, n (%) 1/24 (4%) 2/24 (8%) 1.00
IVH, grade •3, n (%) 2/24 (9%) 2/24 (9%) 1.00
PVL, grade •3, n (%) 0/24 (0%) 1/24 (4%) 1.00
ROP (severe grade III/+disease), n (%) a 3/23 (13%) 2/23 (9%) 1.00
O2 at 36 weeks PMA, n (%) a 5/23 (22%) 6/23 (26%) 1.00
PDA surgical treatment, n (%) 4/24 (17%) 2/24 (8%) 0.67
Deaths before 36 weeks PMA, n (%) 1/24 (4%) 1/24 (4%) 1.00
Hypophosphatemia 1st week, n (%) b 17/22 (77%) 6/23 (26%) 0.001
Hypokalemia 1st week, n (%) 21/24 (88%) 11/24 (46%) 0.005
Student t-test or Fisher’s exact test was applied as appropriate. NEC: necrotizing enterocolitis;
IVH: intraventricular hemorrhage; PVL: periventricular leukomalacia; ROP: retinopathy of prematurity;
PDA: persistent ductus arteriosus; PMA: post-menstrual age; Hypophosphatemia < 1.4 mmol/L;
Hypokalemia < 3.5 mmol/L. a Two more infants died during hospitalization, leaving 46 infants in the
analyses of ROP and O2-dependency at 36 weeks PMA; b Only 45 infants had their phosphate
concentrations measured during the first week of life.
 295

2.2. Nutritional Intervention

The nutritional intervention was started on the first day of life, after informed consent was
obtained [9,10]. The intervention group started with 3.5 g/kg/day of amino acids and 2.0 g/kg/day
of intravenous lipids, whereas the control group started with 2.0 g/kg/day of amino acids and
0.5 g/kg/day of lipids. To improve the parenteral supply of the long chain polyunsaturated fatty
acids (PUFAs), the intervention group received a lipid emulsion containing marine omega 3 fatty
acids (SMOF®, Fresenius Kabi Norge AS, Oslo, Norway), whereas the control group received the
lipid emulsion used in our units (Clinoleic®, Baxter AS, Oslo, Norway). The supply of human milk
was increased equally in both groups, and standard fortification with 4.2 g Nutriprem® (Nutricia
Norge AS, Oslo, Norway)/100 mL human milk was initiated when the infants tolerated
110 mL/kg/day as enteral supply. In addition to standard fortification, the intervention group was
given 0.6 g Complete Amino Acid Mix® (Nutricia Norge AS, Oslo, Norway)/100 mL human milk,
60 mg/kg/day of docosahexaenoic acid (DHA; 22:6, n-3) as well as arachidonic acid (AA; 20:4,
n-6), and 1500 ȝg/kg/day vitamin A (Ås Laboratory, Ås, Norway). On average the energy supply
was approximately 10% higher and the protein supply 25% higher in the intervention group as
compared to the control group [9].

2.3. Sample Collection and Preparation

Urine samples were obtained from the infants during the first week of life and thereafter
approximately every other week until discharge (Figure 1). We collected 0.5–1.8 mL urine
non-invasively by the use of cotton pads, transferred them to Nunc® CryoTubes® (Nalge Nunc
International, Penfield, NY, USA) before they were stored at í80 °C. The urine samples had been
thawed once prior to the metabolic NMR profiling when they were acidified for electrolyte analysis
by mixing 400 ȝL of sample with 5 ȝL of 1 M HCl, resulting in a pH of approximately 3.
Metabolite profiling in the present study largely followed a protocol described earlier [26].
Briefly, 150 ȝL of distilled water and 50 ȝL of a buffer at pH 7.4 containing D2O and
trimethylsilylpropionate-d4 (TSP) were added to 350 ȝL of the samples, which were then
centrifuged at 13,400× g for 5 min and transferred to 5 mm NMR tubes (Wilmad LabGlass,
Vineland, NJ, USA). One-dimensional, water-suppressed proton NMR spectra were acquired at
300.0 K on a Bruker AVI-600 spectrometer (Bruker Biospin GmbH, Rheinstetten, Germany)
equipped with a TCI cryoprobe and a BACS-60 automatic sample changer, under full automation
of D2O locking, tuning and matching, and gradient shimming using TopSpin 2.1pl6 and iconNMR.
Of each sample 32 scans and 4 dummy scans were collected into 64 k data points using the Bruker
“noesygppr1d.comp” sequence with a spectral width of 20.6 ppm, 2.65 s acquisition time and
a 25 Hz water presaturation during the 4 s relaxation delay. An exponential line broadening of
0.3 Hz was applied. The TSP signal achieved a full width at half maximum of less than 1 Hz after
apodization and acted as spectral and concentration reference. The spectra were phase-corrected,
a smooth baseline was removed, and the spectra were binned to a spectral width of 0.01 ppm.
Signals were assigned to known metabolites using a reference database [27] and the software
Chenomx NMR Suite 7.5 professional (Chenomx Inc., Edmonton AB, Canada). Two example
 296

spectra are shown in Figure 2. Pseudo-concentrations were extracted by integrating manually

defined spectral regions corresponding to both known and unknown substances, and arranged in
a table. Pseudo-concentrations are proportional to absolute concentrations and can be used as such
in the statistical analysis. Both the spectra and the table of metabolite pseudo-concentrations were
subsequently normalized to the total intensity of the respective spectra, and the metabolite table
was log-transformed.

Figure 1. (a) Available urine samples by infants’ age in days, one infant per line, one
sample per symbol. Grouped by intervention and control (red and gray lines,
respectively), color-coded by week of life. Age in days was imputed for eight samples
where only the week was recorded; (b) Available urine samples by infants’ week of
life. Bars divided by nutritional intervention vs. control (left half of bar; red and gray,
respectively) and further subdivided by small for gestational age (SGA) or appropriate
for gestational age (AGA) infants (right half of bar; SGA white, AGA black).
Week 1 3 5 7 >7

45
(a) (b)
SGA

23 5
Control

38 Total sum
40

18
17 4 Control SGA
AGA

Number of urine samples

13 Intervention AGA
30
30

13 2
11 26
24 6 6
9 9
22 11
21 10
20

20 9
Intervention

SGA

17 8
15 5

11 11 11
10

10
9
AGA

0 10 20 30 40 50 150 1 3 5 7 >7

Age in days Week after birth

Figure 2. Selected regions of two NMR spectra (black for week 1 and red for week 1) of
an SGA infant in the intervention group.
0.30
0.25
0.06

Betaine
Unknown singlet

0.20
Intensity (a.u.)

Intensity (a.u.)

Glycine

Threonine
0.04

Dimethylamine
0.15

Unknown broad singlet

Sugar doublets
4-Hydroxyphenylacetate

N,N-Dimethylglycine
Creatinine
Unknown multiplet
Unknown multiplet

Sucrose/Maltose
Formate

0.10

2-Oxoglutarate
0.02

Sarcosine

Succinate
Histidine

Tyrosine
Hippurate

Fumarate

0.05

Citrate
0.00
0.00

8.0 7.5 7.0 6.5 6.0 5.5 3.5 3.0 2.5 2.0 1.5 1.0
Chemical Shift (ppm) Chemical Shift (ppm)
 297

2.4. Statistical Analysis

We used Student t-test, Mann-Whitney U test or Fisher’s exact test to evaluate differences in
baseline characteristics, clinical outcomes and nutrient supplies between the two study groups [9,10].
Results are presented as frequencies (%) for categorical data, and as means (ranges or standard
deviations) or medians (interquartile ranges) for continuous data [9,10]. For the metabolomics
study, PCA was applied to mean-centered and unit-variance scaled spectra to explore the major
variations in the dataset [28,29]. By definition, a PCA score plot arranges samples based on the
similarity of their spectra, thus enabling the identification of natural groupings of and systematic
changes between samples. The corresponding loadings reveal which spectral regions, i.e., which
metabolites, contribute to the scores. Multivariate PLS regression was used to associate the
endpoints in our study to the urine spectra. Again, the spectral variables were mean-centered and
scaled to unit variance, and 7-fold cross-validation was applied to evaluate the quality of the
resulting statistical models by considering the diagnostic measures R2 and Q2 [30], describing the
endpoint variation captured in regression model, and the variation reproduced in cross-validation,
respectively. Whereas R2 and Q2 represent measures of the strength of a multivariate relationship
between profiles and endpoints, their ratio Q2/R2 is a measure of cross-validation reproducibility. In
the present study, Q2/R2 ratios above 0.5 were considered indicative of relevant associations, which
were then studied further.
Univariate response approaches were used on log-transformed data in the pseudo-concentration
table to expand the results from the multivariate analyses. A linear mixed model for repeated
measures (first-order autoregressive covariance structure) was used to study the impact of
the two diets on the metabolite pseudo concentrations over time (weeks 1, 3, 5 and 7), adjusted for
gestational age at birth and SGA status. Linear regression was used to quantify the relations
between the metabolite pseudo concentrations at week 1 with SGA status and PMA in weeks, and
also between metabolite levels, growth velocity and PMA. Results are reported as fold-change
ratios (FC) with respect to back-transformed metabolite levels; ratios below 1 are presented
as í1/ratio. Bonferroni correction for multiple testing was applied. The analyses were carried out
on a Windows PC using SPSS version 20 (SPSS Inc., Chicago, IL, USA) and R 2.12.1, 64-bit
(R Foundation, Vienna, Austria), with packages pls 2.2-0 and pcaMethods 1.32.0.

3. Results

The PCA score plot in Figure 3 presents the overall NMR spectroscopic relations between
all the available urine samples, with lines between consecutive samples from the same infant.
Urine samples from the first week of life occupy the lower right quadrant of the plot, and mostly
progress towards the middle left with increasing age of the infant. There was no obvious difference
in distribution or temporal development between the intervention and control group.
Several samples in the upper right quadrant deviated from this general trend and are marked as
outliers. The deviation was characterized by strong NMR signals, predominantly in the aromatic
region of the NMR spectrum, which could not be identified as known metabolites. Whereas the 10
infants with outlier samples had a somewhat lower gestational age than the others, there were no
 298

significant differences with respect to the nutritional intervention, SGA status, sex or infections or
any other of the clinical parameters (data not shown).
The PCA loadings (not shown) revealed that the first principal component (PC1, the x-axis)
of Figure 3 corresponded to increasing levels of citrate, betaine, glycine and hydroxyproline from
right to left, along with decreasing unidentified spectral signals at 0.57 and 5.50 ppm. The second
(PC2, the y-axis) and the third principal component (not shown) were dominated by the
unidentified signals of the outlier samples mentioned above.

Figure 3. PCA score plot of NMR spectra of all available urine samples, presented as
points marked with infant age in weeks and color-coded as earlier. PCA: Principal
component analysis, NMR: Nuclear magnetic resonance, PC: Principal component with
percent of explained total variation. Lines connect consecutive samples from one infant;
line color red for intervention, gray for control group. Outlier samples marked with
a dashed line in the upper right quadrant. Inset: Cumulative explained variation (black)
and cross-validation (red) of the first five PCs.
40
Explained Variation (%)

5
3
30
15

1
5 7
20

3
3
5
10

3
5 5
10

3
0

PC1 PC2 PC3 PC4 PC5 3

7 7
3
3 77 9 3 3
PC2: 6.3%
5

9
5 75 73 3
35 1
5 7 93 37
11 53 5 7 1
5 3 3 3 3
3 7 5 5 31
7 5 7 1
93 9 5 3585 35 33 5 7 7 1
5 3 3 37 9
0

9 9
7 559 1 155 13
3 3 1
7 3 3 5 75 1
3 5 5 7 1
11 73 7 13 1 1
7 9 11 1
999 7 1 1
1 1 1 1 1
13 5 99 1
1
12 1
-5

15 19 1 1 1 11
15 31 1 11
1 1 1 11
20 11 1

11
-10

-15 -10 -5 0 5 10 15
PC1: 14.3%

The metabolites were studied in relation to changes over time in a linear mixed model for
repeated measures (Table 3). There was no significant interaction between intervention group and
time for any of the metabolites, thus interaction terms were not included in the final models. There
was no significant effect of the intervention, but for most of the metabolites, there was a significant
effect of time. The levels of amino acids and many other metabolites increased between weeks 1
and 3, whereas gluconate and two strong, unidentified signals at positions 0.57 and 5.50 ppm
 299

disappeared. The results were similar irrespective of whether the outliers were kept or removed in
the analyses (data not shown).

Table 3. Mixed model and change of mean metabolite levels between sampling weeks
1, 3, 5 and 7 (n = 48).
Weeks Weeks Weeks
Metabolite Weeks 1ĺ0L[HGModel
1ĺ 3ĺ 5ĺ
(n = 35 Pairs) (n = 28 Pairs) (n = 19 Pairs)
p Interaction p Diet p Time FC FC FC
Total Integral í1.2 í1.2 1.2
1-Methylnicotinamide 0.21 0.24 0.02 1.2 1.0 í1.1
2-Oxoglutarate 0.05 0.51 <0.001 4.1 1.4 1.2
4-Hydroxyphenylacetate 0.43 0.11 <0.001 5.1 í1.1 í1.3
Betaine 0.96 0.16 <0.001 1.5 1.2 1.3
Citrate 0.80 0.26 <0.001 5.5 1.9 1.5
Creatinine 0.79 0.52 0.01 1.0 1.1 1.2
Dimethylamine 0.87 0.39 0.003 1.1 1.1 1.0
Fumarate 0.06 0.29 <0.001 3.2 1.7 1.0
Gluconate - - - x - -
Glycine 0.62 0.05 <0.001 1.6 1.2 í1.1
Hipurate 0.56 0.10 0.31 í1.2 1.6 1.0
Histidine 0.70 0.31 <0.001 1.2 1.6 í1.2
myo-Inositol 0.60 0.28 0.08 1.3 1.0 í1.2
N,N-Dimethylglycine 0.23 0.62 <0.001 1.3 1.2 1.3
Succinate 0.42 0.67 <0.001 5.0 1.4 1.0
Sucrose/Maltose 0.12 0.51 <0.001 í8.4 1.1 í1.1
Sugar doublets,
0.30 0.08 0.03 1.0 í1.2 í1.2
5.23 ppm
Threonine 0.58 0.34 <0.001 2.0 1.0 í1.7
trans-4-Hydroxy-L-
0.26 0.16 <0.001 3.1 1.3 1.0
proline
Tyrosine 0.33 0.37 <0.001 3.7 1.0 í1.4
Unknown, 0.57 ppm - - - x - -
Unknown, 5.50 ppm - - - x - -
Unknown, 7.68 ppm 0.60 0.48 <0.001 1.2 1.1 1.1
Unknown, 7.76 ppm 0.43 0.9 <0.001 1.6 1.4 1.3
Linear mixed model for intervention (diet) and week of life (time) with adjustment for gestational age (GA) at birth
and small for gestational age (SGA) status was used. Statistical significance was assumed for p < 0.002. Increase or
decrease of log-transformed pseudo-concentrations, presented as fold-change (FC; ratios below 1 are presented as
í1/ratio). The FC is based on available paired urine samples from the same child at the respective weeks of age.
Metabolites marked “x” disappeared entirely; FC is therefore not applicable. Histidine is an uncertain assignment, based
on a narrow doublet at 7.9–8.0 ppm. Total integral of the urine spectra was determined early relative to the added
internal standard trimethylsilylpropionate-d4 (TSP) and then used to normalize all specified compounds.
 300

Multivariate PLS regression analyses were carried out between the NMR spectra and clinical
variables (Table 4). The nutritional intervention, presence of infections, and infants’ sex did not
influence the urine spectra, whereas SGA status did show an effect on the metabolite profiles:
The PLS model of the infants’ SGA status based on all spectra except those of the outliers reached
a Q2/R2 ratio of 0.40. This increased to 0.53 by focusing on spectra from the first week of life,
whereas spectra of the urine samples at 36 weeks PMA could not be linked to SGA status.
Inclusion of the outliers in the analysis did not change these results (data not shown). PLS
regression analysis also indicated that PMA as well as chronological age were associated with the
urine spectra.

Table 4. Partial least squares regression of variables to sample spectra.

Variable Samples a Ab Q2 R2 Q2/R2
Intervention all - - -
first - - -
36 weeks PMA - - -
SGA status all 3 34% 84% 0.40
first 1 27% 50% 0.53
36 weeks PMA - - -
Infections all - - -
first - - -
36 weeks PMA - - -
Age (since birth) all 2 41% 81% 0.51
36 weeks PMA - - -
Age (PMA) all 3 67% 89% 0.75
first 1 54% 70% 0.76
Sex all - - -
first - - -
36 weeks PMA - - -
R2: Endpoint variation contained in regression model. Q2: Variation reproduced in cross-validation.
Higher numbers, or at least high Q2/R2 ratios mean reliable models. a Samples: All except outliers, first
urine sampled from subject, and sample from 36 weeks PMA; b Number of PLS components resulting in
best (highest) Q2/R2 ratio; results with Q2/R2 ratio below 0.3 not shown.

Observations linking the first-week urine samples to SGA status as well as PMA were also
studied by linear regression analyses for selected metabolites (Table 5). In simple linear regression
analyses, SGA status was associated with increased levels of glycine, histidine and threonine
(8 × 10í4 ” p ” 0.003), as well as creatinine, succinate and trans-4-hydroxy-L-proline
(hydroxyproline) (0.016 ” p ”   30$ ZDV DVVRFLDWHG ZLWK D EURDGHU UDQJH RI PHWDEROLWHV
(3 × 10í ”p ”IRUDOOYDULDEOHVLQ7DEOH 7KH6*$LQIDQWVKDGDKLJKHUPHDQJHVWDWLRQDO
age at birth than AGA infants (29.9 vs. 27.5 weeks, p = 0.003). When adjusting for PMA in the
multiple linear regression analyses, there is an indication that SGA was associated with glycine
(p = 0.027) and threonine (p = 0.033), although not significant at the adjusted significance level.
 301

Table 5. Linear regression results for selected metabolites at week 1 with respect to
SGA status and PMA.
Metabolite a SGA Alone b PMA Alone c Mutually Adjusted d
(Week 1) FC p FC p FC (SGA) p (SGA) FC (PMA) p (PMA)
í4
2-Oxoglutarate 1.6 0.15 1.3 4 × 10 í 0.81 1.3 0.001
Betaine 1.4 0.057 1.1 9 × 10í5 1.0 0.95 1.1 5 × 10í4
Citrate 1.9 0.12 1.3 0.001 í 0.87 1.3 0.004
í7
Creatinine 1.8 0.018 1.2 3 × 10 1.0 0.86 1.2 6 × 10í6
Dimethylamine 1.5 0.036 1.2 2 × 10í4 1.1 0.69 1.1 0.001
Formate 1.6 0.083 1.2 0.005 1.1 0.66 1.1 0.022
í4 í5
Glycine 1.8 2 × 10 1.1 7 × 10 1.4 0.027 1.1 0.005
í4 í6
Histidine 2.0 8 × 10 1.2 9 × 10 1.4 0.091 1.2 6 × 10í4
myo-Inositol 1.5 0.062 1.1 0.001 1.4 0.73 1.1 0.006
N,N-Dimethylglycine 1.0 0.93 1.2 0.020 í 0.18 1.2 0.008
Succinate 2.1 0.039 1.3 3 × 10í4 1.2 0.63 1.3 0.003
Sugar doublets, 5.23 ppm í 0.11 í 0.002 í 0.66 í 0.012
Threonine 1.8 0.003 1.1 0.040 1.6 0.033 1.0 0.37
trans-4-Hydroxy-L-proline 1.9 0.016 1.3 5 × 10í7 1.1 0.81 1.3 1 × 10í5
Histidine is an uncertain assignment, based on a narrow doublet at 7.9–8.0 ppm. Significance assumed for p < 0.002.
a
Only selected contributions are shown; b Simple linear regression analyses of log-transformed pseudo-concentrations
and SGA status; results are presented as fold-change (FC), e.g., glycine levels were 1.8-fold higher in the SGA group
and 1.1-fold higher for each week’s difference in PMA at week 1 (ratios below 1 are presented as í1/ratio);
c
Corresponding analyses for PMA; FC is per one week difference in PMA; d Multiple linear regression including both
SGA status and PMA.

The previous considerations are summarized for the urinary metabolites glycine and threonine in
Figure 4. There were similar levels of glycine and threonine in the intervention and control group
(Figure 4a,b). Glycine and threonine levels appeared to differ between SGA and AGA children in
the first week of life, but not at later time points (Figure 4c,d). The same applied when the infants’
age was defined as PMA instead of chronological age (Figure 4e,f).
Finally, the metabolite pseudo-concentrations were examined with respect to growth velocity
from birth to four weeks of life [9]. In the initial linear regression models, first week glycine and
threonine levels and third week glycine and hydroxyproline levels correlated positively with
growth velocity. However, when adjusting for PMA in the models, these relations disappeared.
 302

Figure 4. Temporal development of glycine and threonine log-pseudo-concentrations

(means and 95% CIs) related to nutritional intervention, SGA status and age. (a)
Glycine levels by nutritional intervention (intervention red, control gray); (c) Glycine
levels by SGA status (SGA orange, AGA green) for samples from weeks 1, 3, 5 and 7;
(e) As above, but samples selected by PMA instead of weeks of life; (b, d, f)
Corresponding figures for threonine.

Glycine Threonine

4.0
Intervention Intervention
4.0

3.5
Log(Threonine/total)
3.0
Log(Glycine/total)
3.5

Control

2.5
Control
3.0

2.0
2.5

1.5

(a) (b)
2.0

1.0

first 3 5 7 first 3 5 7
Age (weeks) Age (weeks)
4.0

SGA SGA
4.0

3.5
Log(Threonine/total)
3.0
Log(Glycine/total)
3.5

2.5

AGA
3.0

AGA
2.0
2.5

1.5

(c) (d)
2.0

1.0

first 3 5 7 first 3 5 7
Age (weeks) Age (weeks)
4.0

SGA
SGA
4.0

3.5
Log(Threonine/total)
3.0
Log(Glycine/total)
3.5

2.5
3.0

AGA AGA
2.0
2.5

1.5

(e) (f)
2.0

1.0

<29.1 <31.3 <33.3 <35.1 >35.1 <29.1 <31.3 <33.3 <35.1 >35.1
Age (PMA) Age (PMA)

4. Discussion

Due to proposed risks of overfeeding, we investigated the impact of enhanced nutrition on the
urinary excretion profile during the first weeks of life in premature infants. We did not observe
significantly different metabolic trajectories between the intervention group receiving nutritional
support in the upper range of recent recommendations as compared to the infants on standard
 303

nutrient supply; neither in the first-week urine profiles nor in the change over time. Furthermore,
infants in the intervention group exhibited better overall growth [9]. Together, this suggests that
premature infants handle enhanced nutrient supply similarly to a standard diet because the urinary
profiles in the intervention group did not indicate an overload of the renal function as compared to
the controls.
Our study also revealed that all infants exhibited substantial changes in their urinary profiles
during the early postnatal period (Figure 3, Table 3), and these changes correlated with gestational
age at birth and with chronological age (Tables 3 and 5). The correlation between PMA and urinary
metabolite profiles has been reported previously [17,31,32], and may reflect the degree of organ
development and metabolic maturity. Between the first and the third week of life the glucogenic
amino acids glycine, threonine, hydroxyproline and tyrosine increased along with metabolites of
the tricarboxylic acid cycle like 2-oxoglutarate, citrate, fumarate and succinate. In most mammals,
the prenatal-postnatal transition is accompanied by important adaptations in carbohydrate
metabolism due to the abrupt change from the placental supply of nutrients to a cyclic supply of
nutrients via the breast milk [7]. In rodents this period is characterized by the appearance of
gluconeogenic enzymes to maintain glucose homeostasis in the newborn [33]. Thus, the metabolite
changes observed in our premature infants might reflect similar metabolic adaptations. The
presence of metabolites linked to the tricarboxylic acid cycle may be due to the high metabolic
turnover in premature infants. The tricarboxylic acid cycle is important in energy metabolism,
providing intermediates for the synthesis of glucose and some amino acids [34].
Hydroxyproline showed a threefold increase in concentration during the initial postnatal period.
Urinary hydroxyproline reflects collagen metabolism and is considered a marker of infant
growth [35,36]. Although we observed a positive correlation between third week hydroxyproline
levels and growth velocity, this correlation disappeared when PMA was introduced as a covariate,
suggesting that urinary hydroxyproline is closely related to PMA.
In parallel with the increase of the other metabolites during the first month of life, gluconate and
two unidentified metabolites with an NMR signals at 5.50 and 0.57 ppm disappeared. The latter
unidentified metabolite has also been observed in the urine of pregnant women [26,37]. Its
disappearance shortly after birth suggests that this substance was transferred from the mother to the
infant and may reflect a sulfate- or glucuronide-conjugate of pregnanediol or estrogen [26].
Most SGA infants have been exposed to a limited nutrient supply during fetal life, which may
cause irreversible metabolic changes (fetal programming). Subsequent catch-up growth, both in
early infancy and in childhood, is also associated with later obesity and cardiovascular disease
risk [4,5,8]. The so-called mismatch hypothesis proposes that an obesogenic childhood
environment increases later cardiovascular disease risk, whereas the postnatal programming or
postnatal growth acceleration hypothesis links rapid weight gain in early infancy to later
cardiovascular disease risk. In a recent review [3], growth during late infancy and childhood
appeared to be the major determinant of later metabolic and cardiovascular disease risk, and not the
early postnatal growth. It has also been shown that early postnatal growth has a significant impact
on later neurodevelopment [3]. Both these findings support the aggressive nutritional approach in
our intervention. We studied metabolic differences between SGA and AGA infants at birth
 304

and over time, and identified glycine and threonine as potential biomarkers of an altered
metabolic phenotype.
Glycine has been linked to nutrient restriction of pregnant baboons, where the fetal plasma
levels more than doubled compared to control fetuses [38]. A similar increase in fetal glycine levels
has been observed in human SGA fetuses [39,40]. Dessi et al., profiled newborn urines one and
four days after birth and reported that in addition to the glycine and threonine pathway, prenatal
growth restriction also affected metabolic pathways involving hydroxyproline, creatinine and
myo-inositol [19]. They interpreted these metabolites as potential early markers of the metabolic
syndrome. In line with their study, we found similar differences in the first-week urine profiles
between our SGA and AGA infants, although after adjusting for PMA, only threonine and glycine
remained independently elevated in the SGA infants. Glycine and threonine are glucogenic amino
acids, which may be converted to pyruvate during protein metabolism. Increased levels of plasma
glycine may be caused by reduced amino acid oxidation or reduced gluconeogenesis as a strategy
to conserve amino acids [38]. We observed that glycine and threonine were linked to SGA status in
the first urine sample, but we were unable to detect a persistent difference during the course of
time. Although our study did not exhibit similar results for all metabolites as compared to the study
by Dessi et al. [19], and the elevation of glycine and threonine levels were insignificant after
Bonferroni adjustment, it still highlights glycine, threonine and to some extent hydroxyproline
as important targets for future research.
Our study has several limitations. In our original intervention trial [10], we planned to recruit
240 infants. The early termination due to increased occurrence of septicemia in the intervention
group resulted in a reduced number of infants in our present study. In spite of the fact that the
intervention group had a significantly lower mean birth weight and a higher proportion of SGA
infants than the control group, we observed increased whole body growth [9], improved white
matter maturation and motion perception in the intervention group (manuscript submitted).
Moreover, we did not find any significant differences between the metabolic trajectories with
regard to the two different diets. The occurrence of septicemia and electrolyte deviations did not
seem to influence the urinary metabolite profiles, but the lack of significant differences must be
interpreted with caution in view of the relatively small study sample. Similar electrolyte
disturbances have been reported in other studies with early and enhanced nutrition to premature
infants during the first week of life [41–46]. Thus, a difference in metabolic profile would probably
occur during the early postnatal stay, and it raises the question as to whether we would have
been able to identify an effect with a more frequent first week monitoring of the urines in our
premature infants.
Multiple hypothesis testing was performed using Bonferroni correction. Although samples from
such vulnerable patients are challenging to come by, the randomized design of the current trial as
well as the strict adherence to the nutritional protocol reduced the number of confounding factors.
 305

5. Conclusions

The urinary metabolite profiles were unaltered by the enhanced postnatal nutrition, suggesting
that supply in the upper range of current recommendations did not overload renal function. Our
data show that both gestational age at birth, i.e., degree of maturation, and postnatal physiological
adaptations, may influence metabolism in premature infants during the neonatal period. Several of
the first-week urinary metabolites were associated to SGA status and postnatal growth and might
be markers for long-term health outcomes.

Author Contributions

D.S. acquired the NMR spectra of the urine samples. S.J.M. was responsible for the detailed
planning of the nutritional protocol, recruitment and treatment of infants, and the collection of
clinical parameters and endpoint data. D.S. and S.J.M. jointly carried out the statistical analysis and
drafted the manuscript. E.W.B., K.S., B.N. and A.N.A. contributed to the conception and the design
of the study, recruitment and treatment of infants, and the collection of clinical parameters and
endpoint data. A.C.W., A.R., K.B., M.B.V., P.O.I. and C.A.D. contributed to the conception and
the design of the study, and the interpretation of data, M.B.V., in addition to the statistical analysis.
F.R. contributed to the design of the study, the metabolic profiling, and the analysis and
interpretation of the results. J.P.B. planned the metabolic profiling and contributed to the statistical
analysis and the interpretation of the results. All authors revised the manuscript critically for
important intellectual comment.

Conflicts of Interest

The authors declare no conflict of interest.

References

1. Lapillonne, A.; O’Connor, D.L.; Wang, D.; Rigo, J. Nutritional recommendations for the
late-preterm infant and the preterm infant after hospital discharge. J. Pediatr. 2013, 162,
S90–S100.
2. Henriksen, C.; Westerberg, A.C.; Ronnestad, A.; Nakstad, B.; Veierod, M.B.; Drevon, C.A.;
Iversen, P.O. Growth and nutrient intake among very-low-birth-weight infants fed fortified
human milk during hospitalisation. Br. J. Nutr. 2009, 102, 1179–1186.
3. Lapillonne, A.; Griffin, I.J. Feeding preterm infants today for later metabolic and
cardiovascular outcomes. J. Pediatr. 2013, 162, S7–S16.
4. Lucas, A. Long-term programming effects of early nutrition—Implications for the preterm
infant. J. Perinatol. 2005, 25 (Suppl. 2), S2–S6.
5. Koletzko, B.; Brands, B.; Poston, L.; Godfrey, K.; Demmelmair, H. Early nutrition
programming of long-term health. Proc. Nutr. Soc. 2012, 71, 371–378.
6. Barker, D.J.; Osmond, C. Infant mortality, childhood nutrition, and ischaemic heart disease in
England and Wales. Lancet 1986, 1, 1077–1081.
 306

7. Patel, M.S.; Srinivasan, M. Metabolic programming as a consequence of the nutritional

environment during fetal and the immediate postnatal periods. In Neonatal Nutrition and
Metabolism, 2nd ed.; Thureen, P., Hay, W., Eds.; Cambridge University Press: Cambridge,
UK, 2006; pp. 76–90.
8. Singhal, A.; Lucas, A. Early origins of cardiovascular disease: Is there a unifying hypothesis?
Lancet 2004, 363, 1642–1645.
9. Moltu, S.J.; Blakstad, E.W.; Strommen, K.; Almaas, A.N.; Nakstad, B.; Ronnestad, A.;
Braekke, K.; Veierod, M.B.; Drevon, C.A.; Iversen, P.O.; et al. Enhanced feeding and
diminished postnatal growth failure in very-low-birth-weight infants. J. Pediatr. Gastroenterol.
Nutr. 2014, 58, 344–351.
10. Moltu, S.J.; Strommen, K.; Blakstad, E.W.; Almaas, A.N.; Westerberg, A.C.; Braekke, K.;
Ronnestad, A.; Nakstad, B.; Berg, J.P.; Veierod, M.B.; et al. Enhanced feeding in
very-low-birth-weight infants may cause electrolyte disturbances and septicemia—A
randomized, controlled trial. Clin. Nutr. 2013, 32, 207–212.
11. Koletzko, B.; Goulet, O.; Hunt, J.; Krohn, K.; Shamir, R. 1. Guidelines on paediatric
parenteral nutrition of the european society of paediatric gastroenterology, hepatology and
nutrition (ESPGHAN) and the european society for clinical nutrition and metabolism
(ESPEN), supported by the european society of paediatric research (ESPR). J. Pediatr.
Gastroenterol. Nutr. 2005, 41 (Suppl. 2), S1–S87.
12. Agostoni, C.; Buonocore, G.; Carnielli, V.P.; de, C.M.; Darmaun, D.; Decsi, T.; Domellof, M.;
Embleton, N.D.; Fusch, C.; Genzel-Boroviczeny, O.; et al. Enteral nutrient supply for preterm
infants: Commentary from the European Society of Paediatric Gastroenterology, Hepatology
and Nutrition Committee on Nutrition. J. Pediatr. Gastroenterol. Nutr. 2010, 50, 85–91.
13. Moco, S.; Collino, S.; Rezzi, S.; Martin, F.P. Metabolomics perspectives in pediatric research.
Pediatr. Res. 2013, 73, 570–576.
14. Emwas, A.H.; Salek, R.M.; Griffin, J.L.; Merzaban, J. NMR-based metabolomics in human
disease diagnosis: Applications, limitations, and recommendations. Metabolomics 2013, 9,
1048–1072.
15. Holmes, E.; Wilson, I.D.; Nicholson, J.K. Metabolic phenotyping in health and disease. Cell
2008, 134, 714–717.
16. Ellis, D.I.; Dunn, W.B.; Griffin, J.L.; Allwood, J.W.; Goodacre, R. Metabolic fingerprinting as
a diagnostic tool. Pharmacogenomics 2007, 8, 1243–1266.
17. Atzori, L.; Antonucci, R.; Barberini, L.; Griffin, J.L.; Fanos, V. Metabolomics: A new tool for
the neonatologist. J. Matern. Fetal Neonatal Med. 2009, 22 (Suppl. 3), 50–53.
18. Fanos, V.; Antonucci, R.; Barberini, L.; Noto, A.; Atzori, L. Clinical application of
metabolomics in neonatology. J. Matern. Fetal Neonatal Med. 2012, 25, 104–109.
19. Dessi, A.; Atzori, L.; Noto, A.; Visser, G.H.A.; Gazzolo, D.; Zanardo, V.; Barberini, L.;
Puddu, M.; Ottonello, G.; Atzei, A.; et al. Metabolomics in newborns with intrauterine growth
retardation (IUGR): Urine reveals markers of metabolic syndrome. J. Matern. Fetal Neonatal
Med. 2011, 24, 36–40.
 307

20. Bell, M.J.; Ternberg, J.L.; Feigin, R.D.; Keating, J.P.; Marshall, R.; Barton, L.; Brotherton, T.
Neonatal necrotizing enterocolitis. Therapeutic decisions based upon clinical staging. Ann.
Surg. 1978, 187, 1–7.
21. International Committee for the Classification of Retinopathy of Prematurity. The
international classification of retinopathy of prematurity revisited. Arch. Ophthalmol. 2005,
123, 991–999.
22. De Vries, L.S.; Eken, P.; Dubowitz, L.M.S. The spectrum of leukomalacia using cranial
ultrasound. Behav. Brain Res. 1992, 49, 1–6.
23. Papile, L.A.; Burstein, J.; Burstein, R.; Koffler, H. Incidence and evolution of subependymal
and intraventricular hemorrhage: A study of infants with birth weights less than 1500 gm. J.
Pediatr. 1978, 92, 529–534.
24. Skjaerven, R.; Gjessing, H.K.; Bakketeig, L.S. Birthweight by gestational age in Norway.
Acta Obstet. Gynecol. Scand. 2000, 79, 440–449.
25. Patel, A.L.; Engstrom, J.L.; Meier, P.P.; Kimura, R.E. Accuracy of methods for calculating
postnatal growth velocity for extremely low birth weight infants. Pediatrics 2005, 116, 1466–
1473.
26. Sachse, D.; Sletner, L.; Mørkrid, K.; Jenum, A.K.; Birkeland, K.I.; Rise, F.; Piehler, A.P.;
Berg, J.P. Metabolic changes in urine during and after pregnancy in a large, multiethnic
population-based cohort study of gestational diabetes. PLoS One 2012, 7, e52399.
27. Wishart, D.S.; Knox, C.; Guo, A.C.; Eisner, R.; Young, N.; Gautam, B.; Hau, D.D.;
Psychogios, N.; Dong, E.; Bouatra, S.; et al. HMDB: A knowledgebase for the human
metabolome. Nucleic Acids Res. 2009, 37, D603–D610.
28. Nicholson, J.K.; Lindon, J.C. Systems biology: Metabonomics. Nature 2008, 455, 1054–1056.
29. Schlotterbeck, G.; Ross, A.; Dieterle, F.; Senn, H. Metabolic profiling technologies for
biomarker discovery in biomedicine and drug development. Pharmacogenomics 2006, 7,
1055–1075.
30. Westerhuis, J.A.; Hoefsloot, H.C.J.; Smit, S.; Vis, D.J.; Smilde, A.K.; van Velzen, E.J.J.;
van Duijnhoven, J.P.M.; van Dorsten, F.A. Assessment of PLSDA cross validation.
Metabolomics 2008, 4, 81–89.
31. De Curtis, M.; Rigo, J. Nutrition and kidney in preterm infant. J. Matern. Fetal Neonatal Med.
2012, 25 (Suppl. 1), 55–59.
32. Foxall, P.J.; Bewley, S.; Neild, G.H.; Rodeck, C.H.; Nicholson, J.K. Analysis of fetal and
neonatal urine using proton nuclear magnetic resonance spectroscopy. Arch. Dis. Child. Fetal
Neonatal Ed. 1995, 73, F153–F157.
33. Greengard, O. Enzymic differentiation in mammalian tissues. Essays Biochem. 1971, 7, 159–205.
34. Harvey, R.A.; Ferrier, D.R. Tricarboxylic Acid Cycle. In Lippincott’s Illustrated Reviews:
Biochemistry, 5th ed.; Lippincott Williams & Wilkins: Philadelphia, PA, USA, 2010; pp. 109–
116.
35. Younoszai, M.K.; Andersen, D.W.; Filer, L.J., Jr.; Fomon, S.J. Urinary excretion of
endogenous hydroxyproline by normal male infants. Pediatr. Res. 1967, 1, 266–270.
 308

36. O’Sullivan, A.; He, X.; McNiven, E.M.; Hinde, K.; Haggarty, N.W.; Lonnerdal, B.;
Slupsky, C.M. Metabolomic phenotyping validates the infant rhesus monkey as a model of
human infant metabolism. J. Pediatr. Gastroenterol. Nutr. 2013, 56, 355–363.
37. Diaz, S.O.; Pinto, J.; Graca, G.; Duarte, I.F.; Barros, A.S.; Galhano, E.; Pita, C.; Almeida,
M.C.; Goodfellow, B.J.; Carreira, I.M.; et al. Metabolic biomarkers of prenatal disorders:
An exploratory NMR metabonomics study of second trimester maternal urine and blood
plasma. J. Proteome Res. 2011, 10, 3732–3742.
38. McDonald, T.J.; Wu, G.; Nijland, M.J.; Jenkins, S.L.; Nathanielsz, P.W.; Jansson, T. Effect of
30% nutrient restriction in the first half of gestation on maternal and fetal baboon serum amino
acid concentrations. Br. J. Nutr. 2013, 109, 1382–1388.
39. Cetin, I.; Corbetta, C.; Sereni, L.P.; Marconi, A.M.; Bozzetti, P.; Pardi, G.; Battaglia, F.C.
Umbilical amino acid concentrations in normal and growth-retarded fetuses sampled in utero
by cordocentesis. Am. J. Obstet. Gynecol. 1990, 162, 253–261.
40. Economides, D.L.; Nicolaides, K.H.; Gahl, W.A.; Bernardini, I.; Evans, M.I. Plasma amino
acids in appropriate- and small-for-gestational-age fetuses. Am. J. Obstet. Gynecol. 1989, 161,
1219–1227.
41. Bonsante, F.; Iacobelli, S.; Chantegret, C.; Martin, D.; Gouyon, J.B. The effect of parenteral
nitrogen and energy intake on electrolyte balance in the preterm infant. Eur. J. Clin. Nutr.
2011, 65, 1088–1093.
42. Bonsante, F.; Iacobelli, S.; Latorre, G.; Rigo, J.; de Felice, C.; Robillard, P.Y.; Gouyon, J.B.
Initial amino acid intake influences phosphorus and calcium homeostasis in preterm
infants—It is time to change the composition of the early parenteral nutrition. PLoS One 2013,
8, e72880.
43. Elstgeest, L.E.; Martens, S.E.; Lopriore, E.; Walther, F.J.; te Pas, A.B. Does parenteral
nutrition influence electrolyte and fluid balance in preterm infants in the first days after birth?
PLoS One 2010, 5, e9033.
44. Ichikawa, G.; Watabe, Y.; Suzumura, H.; Sairenchi, T.; Muto, T.; Arisaka, O.
Hypophosphatemia in small for gestational age extremely low birth weight infants receiving
parenteral nutrition in the first week after birth. J. Pediatr. Endocrinol. Metab. 2012, 25,
317–321.
45. Mizumoto, H.; Mikami, M.; Oda, H.; Hata, D. Refeeding syndrome in a small-for-dates
micro-preemie receiving early parenteral nutrition. Pediatr. Int. 2012, 54, 715–717.
46. Rigo, J.; Marlowe, M.L.; Bonnot, D.; Senterre, T.; Lapillonne, A.; Kermorvant-Duchemin, E.;
Hascoet, J.M.; Desandes, R.; Malfilatre, G.; Pladys, P.; et al. Benefits of a new pediatric
triple-chamber bag for parenteral nutrition in preterm infants. J. Pediatr. Gastroenterol. Nutr.
2012, 54, 210–217.

© 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).
 309

Reprinted from Nutrients. Cite as: Imai, C.M.; Gunnarsdottir, I.; Thorisdottir, B.; Halldorsson, T.I.;
Thorsdottir, I. Associations between Infant Feeding Practice Prior to Six Months and Body Mass
Index at Six Years of Age. Nutrients 2014, 6, 1608-1617.

Associations between Infant Feeding Practice Prior to Six

Months and Body Mass Index at Six Years of Age
Cindy Mari Imai 1,2,* Ingibjorg Gunnarsdottir 1,2, Birna Thorisdottir 1,2,
Thorhallur Ingi Halldorsson 1,2 and Inga Thorsdottir 1,2
1
Unit for Nutrition Research, Landspitali—National University Hospital of Iceland, Eiriksgata
29, Reykjavik 101, Iceland; E-Mails: ingigun@landspitali.is (I.G.); bth50@hi.is (B.T.);
tih@hi.is (T.I.H.); ingathor@hi.is (I.T.)
2
Faculty of Food Science and Nutrition, School of Health Sciences, University of Iceland,
Eiriksgata 29, Reykjavik 101, Iceland

* Author to whom correspondence should be addressed; E-Mail: cmi1@hi.is;

Tel: +354-543-8422; Fax: +354-552-1331.

Received: 16 January 2014; in revised form: 10 March 2014 / Accepted: 2 April 2014 /
Published: 17 April 2014

Abstract: Rapid growth during infancy is associated with increased risk of overweight
and obesity and differences in weight gain are at least partly explained by means of
infant feeding. The aim was to assess the associations between infant feeding practice
in early infancy and body mass index (BMI) at 6 years of age. Icelandic infants
(n = 154) were prospectively followed from birth to 12 months and again at age 6 years.
Birth weight and length were gathered from maternity wards, and healthcare centers
provided the measurements made during infancy up to 18 months of age. Information on
breastfeeding practices was documented 0–12 months and a 24-h dietary record was
collected at 5 months. Changes in infant weight gain were calculated from birth to
18 months. Linear regression analyses were performed to examine associations between
infant feeding practice at 5 months and body mass index (BMI) at 6 years. Infants who
were formula-fed at 5 months of age grew faster, particularly between 2 and 6 months,
compared to exclusively breastfed infants. At age 6 years, BMI was on average
1.1 kg/m2 (95% CI 0.2, 2.0) higher among infants who were formula fed and also
receiving solid foods at 5 months of age compared to those exclusively breastfed. In a
high-income country such as Iceland, early introduction of solid foods seems to further
 310

increase the risk of high childhood BMI among formula fed infants compared with
exclusively breastfed infants, although further studies with greater power are needed.

Keywords: MeSH terms; growth; infant; breastfeeding; weaning; overweight; child

1. Introduction

Childhood obesity is an ongoing public health concern [1]. Infant growth patterns are becoming
better understood and the timing and tempo of growth rate appears to be important variables in
predicting childhood obesity [2,3]. Rapid growth during the first year of life has been associated
with increased risk of overweight and obesity in children [4,5] and differences in weight gain
during early infancy are at least partly explained by means of infant feeding [6–8].
The World Health Organization (WHO) recommends exclusive breastfeeding for the first
6 months of life and its benefits are well accepted worldwide [6,9]. However, there is evidence that
most mothers in European countries begin to introduce complementary foods before 6 months of
age [10]. Studies have shown that formula fed children grow more rapidly than children who are
exclusively breastfed [11,12] and are at greater risk of childhood overweight or obesity [6] but less
is known about the long-term effects of early solid food introduction in early infancy. In a recent
randomized controlled trial, introduction of complementary feeding at 4 months of age did not
increase weight gain in infancy and did not appear to affect the risk of overweight or obesity at
18 months or 29–38 months of age compared to infants exclusively breastfed for 6 months [13–15].
The aim in this current analysis was to assess the associations between infant feeding practice in
early infancy and body mass index (BMI) at 6 years of age. Furthermore, we aimed to assess
whether the introduction of solid foods among partially breastfed or formula fed infants prior to
6 months of age was associated with childhood BMI.

2. Experimental Section

2.1. Study Design

The study population, recruitment and data collection have previously been described in detail
[16]. In brief, a random sample of 250 Icelandic infants born in 2005 was collected by Statistics
Iceland. The inclusion criteria were Icelandic parents, singleton birth, gestational length of 37–41
weeks, birth weight within the 10th and 90th percentiles, no birth defects or congenital long-term
diseases, and the mother had early and regular antenatal care. In this current analysis, eligible
subjects were those with complete data at birth and with a complete dietary record at 5 months, and
weight and height measurements at 6 years of age (n = 154). Our analysis are mainly based on the
feeding practice at the age of 5 months where we compare the growth in infancy between those
children who were exclusively breastfed at this time point to those who were either exclusively
formula fed or those who had started complementary feeding (defined here as introduction of solid
foods in addition to partial breastfeeding or formula feeding) at the age of 5 months. The reason
 311

why we chose to use the 5 month registration for our primary analysis is that this was the earliest
detailed food registration available in the present study and it gives information on variations in
duration of exclusive breastfeeding. Informed written consent was obtained from all parents.
The study was approved by the Icelandic Data Protection Authority (S5099/2011), Local Ethical
Committee at Landspitali-University Hospital (1104Ref.16 2011) and the Bioethics committee
(VSNb2011010008/037).

2.2. Infant and Childhood Growth Data Collection

Birth information on weight and length was gathered from the maternity wards. Healthcare
centers provided the measurements made during infancy at 1, 2, 3, 5, 6, 8, 10, 12 and 18 months of
age. As close to the child’s sixth birthday as possible (mean 73.4 ± 3.2 months) weight (Marel M
series 1100, Iceland; ±0.1 kg) and height (Ulmer stadiometer, Professor Heinze, Busse design,
Ulm, Germany; ±0.5 cm) were measured in a clinical examination at the Landspitali-University
Hospital. BMI was calculated as weight (kg)/height (m2). When the infant was 12 months of age,
the parent or caregiver was asked to complete a questionnaire regarding information on age,
education, and physical characteristics including self-reported height and weight of both parents.

2.3. Dietary Assessment

Information on breastfeeding was gathered monthly during the first 12 months. The parents or
caregivers completed a 24-h food record monthly from 5 to 8 months and 10 to 11 months using
common household measures such as cups and spoons. At 9 and 12 months of age, weighed food
records were kept for three consecutive days on accurate scales (PHILIPS HR 2385, Austria;
PHILIPS HR 2385, Hungary; ±1 g accuracy).
Average daily consumption of energy and the contribution of energy providing nutrients at the
age of 9 and 12 months were estimated using ICEFOOD, a software program used by the Icelandic
Nutrition Council. Special infant products, such as cereals and purées were added to the database
and nutrient losses due to food preparation were taken into account in the calculations.

2.4. Statistical Analysis

Mean and standard deviation (SD) or proportions (%) were used to describe infant and maternal
characteristics. Differences in infant and maternal characteristics were calculated using t-tests for
continuous variables and chi-squared tests for categorical variables. Linear regression analyses
were used to determine associations between infant feeding practice at 5 months of age and BMI at
6 years. All regression analyses were adjusted for sex, birth weight, and maternal education level
categorized as completion of elementary school, high school or vocational school, or university.
Information on mother’s education was missing for n = 12, thus a total of 142 subjects were
analyzed for the linear regression. Changes in growth were calculated from crude differences in
measurements at the two different time points. All statistical analyses were carried out using SPSS
version 20.0 (IBM Corp., Armonk, NY, USA). The level of significance was set at p < 0.05.
 312

3. Results

Table 1 shows the participants’ characteristics among infants at 5 months of age who were
exclusively breastfed (n = 62), exclusively formula fed (n = 12), started on solid foods with partial
breastfeeding (n = 57) or started on solid foods with formula feeding (n = 23).

Table 1. Participant characteristics and dietary variables by infant feeding practice at 5

months of age.
Infant Feeding Practice at 5 Months
Excusively Solid Foods with
Exclusively Solid Foods with
Formula Partial
Breastfed Formula Feeding
Variable Fed Breastfeeding
(n = 62) (n = 23)
(n = 12) (n = 57)
Mean ± SD Mean ± SD Mean ± SD Mean ± SD
Maternal characteristics
Age (years) 31.5 ± 4.5 30.4 ± 4.0 31.0 ± 4.8 30.1 ± 4.8
BMI (kg/m2) 23.7 ± 3.2 29.2 ± 9.1 24.5 ± 3.8 * 27.0 ± 6.5
University education [n (%)] 31 (50) 8 (67) 28 (49) 6 (26)
Infant characteristics
Girls [n (%)] 34 (55) 24 (55) 24 (42) 14 (61)
Exclusive breastfeeding (month) 5.0 ± 0.9 1.7 ± 1.7 * 3.1 ± 1.5 * 1.4 ± 1.5 *
Breastfeeding duration (month) 9.5 ± 1.9 3.9 ± 2.7 * 8.3 ± 2.4 * 2.8 ± 1.1 *
Food groups introduced at 5 months
Porridge [n (%)] - - 53 (93) 18 (78)
Fruits and vegetables [n (%)] - - 20 (35) 15 (65)
Dairy products [n (%)] - - 20 (35) 17 (74)
Legumes [n (%)] - - 1 (2) 3 (13)
Meat, organs [n (%)] - - 0 0
Eggs [n (%)] - - 0 0
Dietary composition at 9 months
Energy (kcal/kg) 101 ± 36 109 ± 27 101 ± 31 101 ± 48
Protein (g/kg) 3.0 ± 1.4 3.7 ± 1.4 3.2 ± 1.4 3.6 ± 1.7
Dietary composition at 12 months
Energy (kcal/kg) 89 ± 22 80 ± 16 86 ± 18 79 ± 16
Protein (g/kg) 3.1 ± 0.9 3.0 ± 0.8 3.6 ± 1.7 3.2 ± 0.9
* Significantly different from exclusively breastfed infants determined by t-tests for continuous variables and chi-
square tests for categorical variables.

Maternal characteristics were not markedly different although mothers who exclusively
breastfed their child at 5 months had lower BMI. The total duration of breastfeeding was shortest
among infants who were formula fed (both with and without solid food introduction) at the age of
5 months. Among infants who were provided solid foods at 5 months, the food items with the
highest frequency of consumption were porridge, dairy products, and fruits and vegetables. Based
on the 24-h dietary registration at 5 months, there were no children eating meat, fish, poultry, liver,
or eggs although it is possible they were exposed to these foods. Although no difference was
 313

observed in energy intake at 9 or 12 months based on infant feeding practice at the age of 5 months,
infants who were receiving formula feeds had higher protein intake per kilogram at 9 months.
There were no differences in birth weight between the different infant feeding groups (Table 2).
However, infants who were formula fed at 5 months grew faster after birth, particularly between
2 and 6 months of age compared to exclusively breastfed infants (Table 3).

Table 2. Anthropometrics from birth by infant feeding practice at 5 months of age.

Exclusively Exclusively Solid Foods with
Solid Foods with Partial
Growth Breastfed Formula Fed Formula Feeding
Breastfeeding (n = 57)
Variable (n = 62) (n = 12) (n = 23)
Mean ± SD Mean ± SD Mean ± SD Mean ± SD
Weight (kg)
At birth 3.8 ± 0.4 3.8 ± 0.4 3.7 ± 0.4 3.7 ± 0.3
2 months 5.6 ± 0.6 5.5 ± 0.7 5.8 ± 0.5 5.4 ± 0.4
6 months 7.8 ± 0.8 8.2 ± 1.0 8.2 ± 1.0 8.5 ± 0.8 *
2 months 9.6 ± 0.9 10.3 ± 1.4 10.1 ± 1.2 * 10.6 ± 1.0 *
18 months 11.3 ± 1.0 11.5 ± 1.2 11.6 ± 1.3 11.9 ± 1.2 *
Length (cm)
At birth 51.9 ± 1.5 52.0 ± 1.7 51.8 ± 1.6 51.7 ± 1.9
2 months 59.7 ± 1.5 59.4 ± 1.8 59.8 ± 2.0 58.9 ± 1.5
6 months 68.3 ± 2.0 69.4 ± 2.7 69.1 ± 2.3 * 68.4 ± 1.7
12 months 76.2 ± 2.5 77.4 ± 2.9 77.1 ± 2.7 77.6 ± 2.5 *
18 months 83.0 ± 2.5 83.8 ± 3.2 83.9 ± 2.7 83.1 ± 2.5
* Significantly different from exclusively breastfed infants determined by t-tests.

Table 3. Changes in weight from birth to 18 months by infant feeding practice at 5

PRQWKV FRPSDULQJ LQIDQWV RQ IRUPXOD RU VROLG IRRGV ǻ  &,  WR H[FOXVLYHO\
breastfed infants.
Exclusively Exclusively Solid Foods with Solid Foods with
Weight Variable Breastfed (referent) Formula Fed Partial Breastfeeding Formula Feeding
(n = 62) (n = 12) (n = 57) (n = 23)
a a a
Changes in weight (g) Mean ± SD ¨ &, ¨ &, ¨ &,
Birth to 2 months 1859 ± 500 í í 156 (í í í
2 to 6 months 2208 ± 480 512 (147, 877) * 161 (í 787 (472, 1102) *
6 to 12 months 1775 ± 550 172 (í 122 (í 283 (í
12 to 18 months 1649 ± 563 í í í í í í
Birth to 6 months 4081 ± 745 308 (í 360 (58, 661) * 755 (367, 1143) *
Birth to 12 months 5876 ± 884 510 (í 462 (92, 831) * 1019 (566, 1471) *
a
Mean difference in weight compared to exclusively breastfed infants (referent); * Significantly different
from exclusively breastfed infants determined by t-tests.
 314

Furthermore, the group that had been introduced to solid foods along with formula feeding at the
age of 5 months grew significantly faster compared to exclusively breastfed infants, while the
difference was smaller and non-significant for the group who had been provided solid foods and
were partially breastfed (Table 3). The addition of solid foods along with formula feeding at
5 months predicted greater BMI at 6 years, with BMI being on average 1.1 kg/m2 (95% CI 0.2, 2.0)
higher compared to exclusively breastfed infants at 5 months of age (Table 4).

Table 4. Associations between infant feeding practice at 5 months and BMI at 6 years
of age (n = 142).
Infant Feeding Practice at 5 Months ¨ &, a p*
Exclusively breastfed Referent -
Formula fed 0.3 (í 0.583
Solid foods with partial breastfeeding 0.5 (í 0.125
Solid foods with formula feeding 1.1 (0.2, 2.0) 0.020
a
Adjusted difference in BMI at 6 years with respect to exclusively breastfed infants; * Analyses adjusted
for sex, birth weight, and maternal education.

4. Discussion

In this current analysis, we found that infants who were provided solids foods in addition to
formula feeding at 5 months of age had faster growth during infancy and up to 12 months of age
compared to exclusively breastfed infants. Differences in weight were most pronounced between
the ages of 2 to 6 months. Furthermore, the addition of solid foods among formula fed infants prior
to the age of 6 months predicted greater BMI at 6 years.
The strength in this cohort lies in the longitudinal nature and the detailed information on infant
feeding practices. In this way, we are able to describe dietary intake in infancy in greater detail than
possible in many other studies in relation to BMI at 6 years of age. It is difficult to separate the
effects of exclusive breastfeeding (and total duration of breastfeeding) from the introduction of
solid foods at the age of 5 months. However, our findings are in line with existing evidence that
show exclusively breastfed infants grow slower compared to infants of the same age who receive
formula or solid foods [2,6] and that longer duration of breastfeeding may protect against later
obesity potentially due to slower weight gain in infancy [6].
There are several factors that contribute to variations in infant feeding practices. Predictors of
early introduction of solid foods include young maternal age, low maternal education and short
(<4 weeks) duration of breastfeeding [17]. There were no significant differences with maternal age,
and maternal education also did not appear to predict introduction of solid foods at 5 months of
age, although the proportion of mothers with higher education was greater among the infants who
were exclusively breastfed. We note that the mothers who started their infants on formula feeding
had a higher mean BMI compared to mothers of exclusively breastfed infants, although the
numbers are too few to detect a significant difference. There is evidence from epidemiological
studies that women who are overweight or obese are less likely to breastfeed compared to normal
weight women [18,19].
 315

A probable explanation for the observed difference in growth rate between breastfed and
formula fed infants is the relatively higher protein content of infant formula compared to
breastmilk. High protein intake may have a stimulating effect on insulin-like growth factor 1 which
can accelerate growth [20]. Furthermore, higher protein intake during infancy has been associated
with faster weight gain and greater adiposity [21] and there is evidence that this is associated with
higher BMI in childhood [4,7,22]. Another possible mechanism is that compared to breastmilk,
infant formula contains a higher amount of omega-6 fatty acids which may play a role in promoting
adipose tissue development and result in greater childhood adiposity [23].
Results from a randomized controlled trial performed in Iceland, showed no significant
differences in energy intake [15], growth [14], or risk of being overweight [13] between
those exclusively breastfed for 4 vs. 6 months. The reason may be a low amount of energy from
solid foods [14,15] mainly consisting of infant cereals (67%) and median protein intake of only
0.9 g/day [14]. In other studies, exposing infants to solid foods prior to the age of 4 months was
associated with being overweight or obese in early childhood [24,25], but it was suggested that the
risk associated with the timing of introduction of solid foods might be greater among children who
were no longer breastfed [24]. In the present study, we saw that partially breastfed infants who had
been introduced to solid foods at the age of 5 months did not grow as fast as the infants who were
formula fed and consuming solid foods.
Together these findings suggest that in addition to the timing of complementary foods, the type
of food, and the protein content of the infant formula introduced may influence infant growth [26].
In this cohort, 70% of infants who had been provided solid foods at 5 months were exclusively
breastfed at 2 months of age. Mothers may introduce solid foods earlier if they find their infant
appears hungry or worry that breast milk is inadequate for their infant’s needs [27]. However,
limited data exists on the effects of specific food groups introduced during the complementary
feeding period in relation to infant growth and childhood BMI. Further studies are needed in this
area to determine whether rapid infant growth leads to children demanding more feedings or
whether the introduction of formula feeding or solid foods is the main contributor.
Some limitations exist in this current analysis. The sample size is a possible limitation however,
the thorough data on dietary intake and growth variables provides valuable information and the
number is sufficient to analyze differences in infant feeding practices and its potential contribution
to childhood weight status. Information on the exact timing of solid food introduction would have
been useful as well as reasons affecting the duration of breastfeeding, particularly among the
infants who had been provided solid foods at 5 months to better understand the observed
association with BMI at 6 years. The aim of this present analysis was to determine the association
between differences in infant feeding practices and childhood BMI, and even more detailed data
related to diet and nutrient composition before the age of 5 months would have been beneficial.
Overall, it appears that earlier formula feeding contributes to growth in infancy and adding solid
foods along with formula feeding may influence growth in childhood compared to exclusively
breastfed infants.
 316

5. Conclusions

In this current analysis, we found that infants who were provided solid foods along with formula
feeding prior to 6 months of age had faster growth during infancy and greater BMI at 6 years of age
compared to exclusively breastfed infants. In a high-income country such as Iceland, early
introduction of solid foods seems to further increase the risk of high childhood BMI of formula fed
infants compared to exclusively breastfed infants. Furthermore, better breastfeeding promotion
strategies may help reduce the incidence of childhood overweight. Further studies with more
detailed information on infant feeding practices and statistical power are needed to determine the
appropriate composition of infant formula and solid foods especially if mothers are for some
reasons unable to breastfeed exclusively for 6 months.

Acknowledgments

The authors are grateful to the healthcare centers for their cooperation, the nutritionists, nurses,
and students who assisted in data collection. We are most thankful to the participating children and
families. This work was supported by the University of Iceland Research Fund, Landspitali
National University Hospital Research Fund and the American Scandinavian Foundation Thor
Thors Memorial Fund. The sponsors had no role in the design; in the collection, analysis and
interpretation of data; in the writing of the report; and in the decision to submit the article for
publication.

Author Contributions

C.M.I. assisted with the design of this current manuscript and performed the statistical analyses
and drafting and editing of the manuscript. I.G. assisted with the design, statistical analyses and
contributed in data collection, writing and editing the manuscript. B.T. assisted with statistical
analyses, did the data collection, and participated in writing and editing the manuscript. T.I.H.
participated in the statistical analyses and provided critical revision of the paper. I.T. was the
principal investigator and designer of the study and supervised data collection, data analyses and
writing and editing of the manuscript.

Conflicts of Interest

The authors report no conflict of interest.

References

1. Han, J.C.; Lawlor, D.A.; Kimm, S.Y. Childhood obesity. Lancet 2010, 375, 1737–1748.
2. Gunnarsdottir, I.; Schack-Nielsen, L.; Michaelsen, K.F.; Sorensen, T.I.; Thorsdottir, I.;
NordNet, G. Study, Infant weight gain, duration of exclusive breast-feeding and childhood
BMI-two similar follow-up cohorts. Public Health Nutr. 2010, 13, 201–207.
3. Lakshman, R.; Elks, C.E.; Ong, K.K. Childhood obesity. Circulation 2012, 126, 1770–1779.
 317

4. Gunther, A.L.; Remer, T.; Kroke, A.; Buyken, A.E. Early protein intake and later obesity risk:
Which protein sources at which time points throughout infancy and childhood are important
for body mass index and body fat percentage at 7 year of age? Am. J. Clin. Nutr. 2007, 86,
1765–1772.
5. Ong, K.K.; Loos, R.J. Rapid infancy weight gain and subsequent obesity: Systematic reviews
and hopeful suggestions. Acta Paediatr. 2006, 95, 904–908.
6. Hornell, A.; Lagstrom, H.; Lande, B.; Thorsdottir, I. Breastfeeding, introduction of other
foods and effects on health: A systematic literature review for the 5th Nordic Nutrition
Recommendations. Food Nutr. Res. 2013, 57, doi:10.3402/fnr.v57i0.20823.
7. Hornell, A.; Lagstrom, H.; Lande, B.; Thorsdottir, I. Protein intake from 0 to 18 years of age
and its relation to health: A systematic literature review for the 5th Nordic Nutrition
Recommendations. Food Nutr. Res. 2013, doi:10.3402/fnr.v57i0.21083.
8. Koletzko, B.; von Kries, R.; Monasterolo, R.C.; Subias, J.E.; Scaglioni, S.; Giovannini, M.; Beyer,
J.; Demmelmair, H.; Anton, B.; Gruszfeld, D.; et al. Infant feeding and later obesity risk.
Adv. Exp. Med. Biol. 2009, 646, 15–29.
9. World Health Organization. Infant and Young Child Nutrition; World Health Organization:
Geneva, Switzerland, 2002.
10. Caroli, M.; Mele, R.M.; Tomaselli, M.A.; Cammisa, M.; Longo, F.; Attolini, E. Complementary
feeding patterns in Europe with a special focus on Italy. Nutr. Metab. Cardiovasc. Dis. 2012,
22, 813–818.
11. Mihrshahi, S.; Battistutta, D.; Magarey, A.; Daniels, L.A. Determinants of rapid weight gain
during infancy: Baseline results from the NOURISH randomised controlled trial. BMC
Pediatr. 2011, 11, doi:10.1186/1471-2431-11-99.
12. Sloan, S.; Gildea, A.; Stewart, M.; Sneddon, H.; Iwaniec, D. Early weaning is related to
weight and rate of weight gain in infancy. Child Care Health Dev. 2008, 34, 59–64.
13. Jonsdottir, O.H.; Kleinman, R.E.; Wells, J.C.; Fewtrell, M.S.; Hibberd, P.L.; Gunnlaugsson,
G.; Thorsdottir, I. Exclusive breastfeeding for 4 vs. 6 months and growth in early childhood.
Acta Paediatr. 2014, 103, 105–111.
14. Jonsdottir, O.H.; Thorsdottir, I.; Hibberd, P.L.; Fewtrell, M.S.; Wells, J.C.; Palsson, G.I.; Lucas,
A.; Gunnlaugsson, G.; Kleinman, R.E. Timing of the introduction of complementary foods in
infancy: A randomized controlled trial. Pediatrics 2012, 130, 1038–1045.
15. Wells, J.C.; Jonsdottir, O.H.; Hibberd, P.L.; Fewtrell, M.S.; Thorsdottir, I.; Eaton, S.; Lucas,
A.; Gunnlaugsson, G.; Kleinman, R.E. Randomized controlled trial of 4 compared with 6 mo
of exclusive breastfeeding in Iceland: Differences in breast-milk intake by stable-isotope
probe. Am. J. Clin. Nutr. 2012, 96, 73–79.
16. Thorisdottir, A.V.; Thorsdottir, I.; Palsson, G.I. Nutrition and iron status of 1-year olds
following a revision in infant dietary recommendations. Anemia 2011, 2011, doi:10.1155/
2011/986303.
17. Scott, J.A.; Binns, C.W.; Graham, K.I.; Oddy, W.H. Predictors of the early introduction of
solid foods in infants: Results of a cohort study. BMC Pediatr. 2009, 9, doi:10.1186/
1471-2431-9-60.
 318

18. Amir, L.H.; Donath, S. A systematic review of maternal obesity and breastfeeding intention,
initiation and duration. BMC Pregnancy Childbirth 2007, 7, doi:10.1186/1471-2393-7-9.
19. Baker, J.L.; Michaelsen, K.F.; Sorensen, T.I.; Rasmussen, K.M. High prepregnant body mass
index is associated with early termination of full and any breastfeeding in Danish women. Am.
J. Clin. Nutr. 2007, 86, 404–411.
20. Schack-Nielsen, L.; Michaelsen, K.F. Advances in our understanding of the biology of human
milk and its effects on the offspring. J. Nutr. 2007, 137, 503S–510S.
21. Escribano, J.; Luque, V.; Ferre, N.; Mendez-Riera, G.; Koletzko, B.; Grote, V.; Demmelmair,
H.; Bluck, L.; Wright, A.; Closa-Monasterolo, R.; et al. Effect of protein intake and weight
gain velocity on body fat mass at 6 months of age: The EU childhood obesity programme.
Int. J. Obes. 2012, 36, 548–553.
22. Gunnarsdottir, I.; Thorsdottir, I. Relationship between growth and feeding in infancy and body
mass index at the age of 6 years. Int. J. Obes. Relat. Metab. Disord. 2003, 27, 1523–1527.
23. Ailhaud, G.; Massiera, F.; Weill, P.; Legrand, P.; Alessandri, J.M.; Guesnet, P. Temporal
changes in dietary fats: Role of n-6 polyunsaturated fatty acids in excessive adipose tissue
development and relationship to obesity. Prog. Lipid Res. 2006, 45, 203–236.
24. Huh, S.Y.; Rifas-Shiman, S.L.; Taveras, E.M.; Oken, E.; Gillman, M.W. Timing of solid food
introduction and risk of obesity in preschool-aged children. Pediatrics 2011, 127, e544–e551.
25. Abraham, E.C.; Godwin, J.; Sherriff, A.; Armstrong, J. Infant feeding in relation to eating
patterns in the second year of life and weight status in the fourth year. Public Health Nutr.
2012, 15, 1705–1714.
26. Koletzko, B.; von Kries, R.; Closa, R.; Escribano, J.; Scaglioni, S.; Giovannini, M.; Beyer, J.;
Demmelmair, H.; Gruszfeld, D.; Dobrzanska, A.; et al. Lower protein in infant formula is
associated with lower weight up to age 2 years: A randomized clinical trial. Am. J. Clin. Nutr.
2009, 89, 1836–1845.
27. Colin, W.B.; Scott, J.A. Breastfeeding: Reasons for starting, reasons for stopping and
problems along the way. Breastfeed Rev. 2002, 10, 13–19.

© 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).
 319

Reprinted from Nutrients. Cite as: Kagawa, M.; Wishart, C.; Hills, A.P. Influence of Posture and
Frequency Modes in Total Body Water Estimation Using Bioelectrical Impedance Spectroscopy in
Boys and Adult Males. Nutrients 2014, 6, 1886-1898.

Influence of Posture and Frequency Modes in Total Body

Water Estimation Using Bioelectrical Impedance
Spectroscopy in Boys and Adult Males
Masaharu Kagawa 1,2,3,4,*, Connie Wishart 4 and Andrew P. Hills 5
1
Institute of Nutrition Sciences, Kagawa Nutrition University, Saitama 350-0288, Japan
2
National Institute of Public Health, Saitama 351-0197, Japan
3
School of Public Health, Curtin University, Western Australia 6102, Australia
4
Institute of Health and Biomedical Innovation, Queensland University of Technology,
Queensland 4059, Australia; E-Mail: c.wishart@qut.edu.au
5
Mater Research Institute, the University of Queensland and Griffith Health Institute,
Griffith University, Queensland 4101, Australia; E-Mail: ahills@mmri.mater.org.au

* Author to whom correspondence should be addressed; E-Mail: mskagawa@eiyo.ac.jp;

Tel.: +81-49-281-7743; Fax: +81-49-284-3679.

Received: 21 February 2014; in revised form: 16 April 2014 / Accepted: 26 April 2014 /
Published: 5 May 2014

Abstract: The aim of the study was to examine differences in total body water (TBW)
measured using single-frequency (SF) and multi-frequency (MF) modes of bioelectrical
impedance spectroscopy (BIS) in children and adults measured in different postures
using the deuterium (2H) dilution technique as the reference. Twenty-three boys and
26 adult males underwent assessment of TBW using the dilution technique and BIS
measured in supine and standing positions using two frequencies of the SF mode
(50 kHz and 100 kHz) and the MF mode. While TBW estimated from the MF mode
was comparable, extra-cellular fluid (ECF) and intra-cellular fluid (ICF) values differed
significantly (p < 0.01) between the different postures in both groups. In addition, while
estimated TBW in adult males using the MF mode was significantly (p < 0.01) greater
than the result from the dilution technique, TBW estimated using the SF mode and
prediction equation was significantly (p < 0.01) lower in boys. Measurement posture
may not affect estimation of TBW in boys and adult males, however, body fluid shifts
may still occur. In addition, technical factors, including selection of prediction
equation, may be important when TBW is estimated from measured impedance.
 320

Keywords: body fluid; deuterium; dilution technique; impedance technique; prediction

equation; accuracy; technical error

1. Introduction

Body composition, including fat mass (FM) and fat-free mass (FFM), is an important variable in
the assessment of health status. Obesity has been defined as a state of excessive fat deposition [1,2]
and the assessment of body composition assists in identifying individuals with metabolic risks.
In addition, while body mass index (BMI: kg/m2) and other simple anthropometric indices have
been utilized as convenient screening tools for obesity, assessment of body composition reduces
misclassification of individuals at risk.
Body composition can be determined using a wide range of techniques. Each technique varies
not only in its accuracy and precision, but also in cost, portability, convenience, and requirements
for accredited operators. Bioelectrical impedance analysis (BIA) is one of the most commonly
utilized techniques as it is simple, portable and cost- and time-efficient. The technique assesses
differences in the electrical conductivity between tissues. Tissues that contain water and
electrolytes have higher conductivity compared to those with less body fluid. From the
measurement of electrical conductivity, resistance (R) and reactance (Xc) can be determined. These
components can be utilized to calculate impedance (Z) based on their association Z2 = R2 + Xc2 and
also a phase angle based on a ratio of Xc to R [3]. In addition, together with information on the
length (L) or height (Ht), a total volume of body water (TBW) can be determined [4,5].
Furthermore, while R has been used most frequently, R, Xc, and Z have been used to estimate
TBW, intra-cellular fluid (ICF) and extra-cellular fluid (ECF) as well as percentage body fat (%BF)
of individuals [4].
Existing BIA devices can be divided into single-frequency BIA (SFBIA), multi-frequency BIA
(MFBIA) and bioelectrical impedance spectroscopy (BIS). SFBIA devices generally use a
frequency of 50 kHz that passes through both ECF and ICF [4]. In comparison, MFBIA uses
multiple frequencies in the range of 1 to 1000 kHz and enables one to distinguish between ICF and
ECF. A previous study reported that a low frequency, generally below 20 kHz, is used to predict
ECF whereas a higher frequency (above 50 kHz) is used to estimate TBW in MFBIA [6]. As a
result, ICF can be determined from the difference of the two. Although it has been suggested that
MFBIA may overestimate %BF of lean individuals and underestimate that of obese individuals [7],
error in estimation of %BF may be minimized compared with SFBIA [8]. BIS is a more
sophisticated model that uses a wide range of frequencies and non-linear mathematical algorithm to
assess relationships between R and body fluid. This allows estimation of R extrapolated to zero
(R0) and infinite (R’) frequencies and development of empirically-derived prediction equations
[4,5,9]. Although both accuracy and precision of results may vary depending on the characteristics
of the study population [9], past studies have reported that BIS provides better estimation of ECF
than SFBIA [4,5,10,11] and also has an acceptable accuracy and precision using an animal model
[12]. In addition, a technique known as “segmental BIA” is available which determines information
 321

on total body composition through measurements of each segment (i.e., upper and lower limbs and
the trunk). A previous review has described a number of advantages and considerations [13], and
another study reported that segmental BIA can provide valid information on body composition
compared with the four-compartment model [14]. However, most studies have been undertaken on
adults and studies of children are relatively scarce. Consequently, little knowledge is available on
any differences in TBW estimation between adults and children and the effects of using
different frequencies.
In addition, wide variations in measurement posture are commonplace using the impedance
technique depending on the device used. Many hand-to-foot models measure in a supine position,
the posture recommended in the European Society for Parenteral and Enteral Nutrition (ESPEN)
guidelines [10]. However, modern segmental BIAs that are in a scale type are designed to measure
in a standing position [13]. A previous study on the influence of posture during measurements
reported 2%–5% changes in ECF and 1.8%–8.0% changes in ICF depending on body position [15].
Another study reported no significant differences in %BF using a hand-to-foot device but
a significant increase in %BF using a hand-to-hand device [8]. These results suggest that body
posture may influence estimation of body fluid status and therefore estimation of %BF using the
impedance technique. However, these studies were conducted on adults and no previous research
has reported differences between adults and children.
Therefore, the present study aimed to examine influences of frequency modes and measurement
posture in estimation of TBW in adult males and boys. The estimated TBW was compared with
the result obtained from the reference deuterium (2H) dilution technique.

2. Experimental Section

The study was approved by the Human Research Ethics Committee of Queensland University
of Technology and adhered to the principles of medical research established by the National Health
and Medical Research Council [16]. Boys aged below 15 years or adult males aged above 20 years
with no medical conditions or under medication were included in the study. Participants were
recruited through flyers. All participants were given an information package and all signed a
consent form prior to their participation. For participants below 18 years of age, parents or legal
guardians also signed the consent form. All participants were provided with a $20 gift voucher after
their full participation in the study. In total, 49 participants including 23 boys aged between 6 and
14 years and 26 adult males aged between 23 and 82 years, completed all assessments and were
included in the study. All participants were instructed to fast overnight and void their bladders in
the morning, prior to measurements being conducted. All assessments on children were conducted
by the primary investigator.

2.1. Anthropometry

All participants underwent measurements of stature using a stadiometer to the nearest 0.1 cm
and body weight using a digital weighing scale to the nearest 0.1 kg. Waist circumference was
measured at the narrowest point between the 10th rib and the iliac crest using a steel
 322

anthropometric tape. All measurements were conducted according to the protocol of the
International Society for the Advancement of Kinanthropometry (ISAK) [17]. All measurements
were taken by a level three anthropometrist accredited by ISAK with an intra-tester technical error
of measurements (TEM) of less than 1.0%, below recommended levels [18,19]. From these
measurements, BMI and waist-to-height ratio (WHtR) were calculated.

2.2. Bioelectrical Impedance Spectroscopy (BIS)

Body fluid status was determined using a BIS device (Imp SFB7, ImpediMed Ltd, Brisbane,
Australia) with functionality to switch between multi-frequency (MF) and single-frequency (SF)
modes. In the MF mode, TBW was estimated using a wide range of frequencies between 4 and
1000 kHz with 256 data points [20]. In the SF mode, two of the five fixed frequencies (i.e., 50 kHz
and 100 kHz) of the device that are commonly selected were used [20].
The device was calibrated before measurements of each participant. All participants rested
(lying on a bed) for at least five minutes prior to the measurement. Electrodes were placed on
the dorsal surface of the wrist and the ankle as well as at the base of the second or third
metacarpal-phalangeal joints of hand and foot after the skin was cleaned with an alcohol wipe. The
lead wires were attached to the appropriate electrodes and participants were instructed to abduct
their limbs from the trunk. After measurement in the supine position, participants were instructed to
stand up and stay in the same position for at least five minutes before measurement in the standing
position. The measurements were repeated in both positions after staying in the position at least
five minutes. Length of rest time was consistent with the protocol used in a previous study [21].
While the participant was in the same posture, triplicate measurements were conducted and
a median value determined. After completion of two sessions of triplicate measurements at each
measurement posture, an average of two median values was calculated for each variable.
From each measurement, TBW, ECF, ICF, FFM and FM were calculated for the MF mode
using built-in algorithms. For the SF mode, Xc and R were recorded for each frequency. Similar to
a previous study [22], all measurements were conducted with resistivity coefficients of 273.9 for
zero/extracellular (ȡe) and 937.2 for infinite/intracellular (ȡi), body density of 1.05 g/cm3 and body
proportion of 4.3. From the obtained FM and FFM results, %BF was calculated for the MF mode.
For the SF mode, Z was calculated from mean Xc and R using an equation¥ 5 2 + Xc2). Calculated
Z values were then utilized to estimate TBW of the study groups using age-, gender-, and
frequency-specific prediction equations. In the present study, estimation of TBW for boys was
conducted using an equation by Davies et al. [23] with Z determined from a frequency of 50 kHz (Z50):
TBW = 0.6 × (Ht2/Z50) í (1)
For males, equations by Deurenberg et al. [24] were used to estimate TBW using Z determined
from frequencies of 50 kHz (Z50) and 100 kHz (Z100):
TBW = 6.53 + 0.3674 × Ht2/Z50 + 0.17531 × body weight í × age + 2.83 (2)
TBW = 6.69 + 0.34573 × Ht2/Z100 + 0.17065 × body weight – 0.11 × age + 2.66 (3)
 323

2.3. Deuterium Dilution Technique

Deuterium (2H) dilution technique was the reference method used for TBW. Prior to the
assessment, a 10% deuterium oxide (D2O) solution was prepared by mixing 100 mL of 99.9% D2O
solution (Aldrich Chemistry, Sigma-Aldrich Pty Ltd, Sydney, Australia) and 900 mL of tap water.
The solution was sterilized at 120 °C for 10 min using an autoclave. After collecting a pre-dose
urine sample, the body weight of participants was measured using a weighing scale. The dose
amount of 10% D2O solution was calculated as 0.5 × body weight (kg) and weighed using a scale.
All participants consumed the weighed 10% D2O solution and re-weighed the drinking cup to
record the precise amount consumed. After consumption of the 10% D2O solution, participants
were instructed to collect a post-dose urine sample after five hours, to ensure that equilibration of
2
H within the body fluid pool was reached.
Both pre- and post-dose samples were analyzed using an isotope ratio mass spectrometry
(IRMS: Hydra 20-20, SerCon Mass Spectrometry, SerCon Limited, Cheshire, CW1 6YY UK). All
analyses were conducted at the laboratory at the Institute of Health and Biomedical Innovation
(IHBI) of Queensland University of Technology (QUT) in Brisbane, Australia. TBW was
calculated using the following equation:
TBW (kg) = ((W × A/a) × (ǻ''ǻ%:  × 1.041) (4)
where W = total weight of water added when making the dose dilution (g), A = weight of dose
taken by the participant (g), a ZHLJKWRIGRVHLQGLOXWHGGRVH J ǻ'' HQULFKPHQWRI 2H in the
diluted dose (ppm excess 2+ DQGǻ%: HQULFKPHQWRI 2H in body water (ppm excess 2H). The
value of 1.041 was based on an assumption that the dilution space or the volume of the distribution
of 2H is 1.041 times greater than TBW [25].
All statistical analyses were conducted using the PASW® Statistics package (version 18.0.0,
IBM, Chicago, IL, USA). Paired t-tests were conducted to compare results obtained from different
postures (i.e., supine and standing positions). In addition, TBW estimated from different frequency
modes were compared with the results from the dilution technique using repeated measures of
analysis of variance (ANOVA) and a Bonferroni post hoc test. Results were expressed as mean ±
standard error (SE). In addition, variability of estimated TBW for the study population from different
frequency modes were determined using correlation coefficients, limits of agreement (i.e., difference ±
1.96 × standard deviation), and the Bland and Altman plots [26] using the dilution technique as the
reference. All statistical tests used significant level of 0.05 unless otherwise stated.

3. Results

Physical characteristics of the participants were 9.8 ± 0.5 years, 144.1 ± 3.2 cm and 36.6 ± 2.4 kg
for boys and 36.9 ± 2.7 years, 174.3 ± 1.4 cm, and 76.5 ± 3.2 kg for adult males, respectively
(Table 1). WHtR, an index of abdominal fat accumulation with the cut-off point of 0.5, were
0.44 ± 0.01 in boys and 0.49 ± 0.01 in adult males, respectively.
 324

Table 1. Demographic characteristics of participants.

Boys (n = 23) Mean ± SE Males (n = 26) Mean ± SE
Age (years) 9.8 ± 0.5 36.9 ± 2.7
Stature (cm) 144.1 ± 3.2 174.3 ± 1.4
Body weight (kg) 36.6 ± 2.4 76.5 ± 3.2
BMI (kg/m2) 17.2 ± 0.5 25.0 ± 0.8
WHtR 0.44 ± 0.01 0.49 ± 0.01

Table 2 shows differences in results obtained from supine and standing positions. In both adult
males and boys, there were no significant differences in TBW using the MF mode in supine and
standing positions. There were no differences in estimated FFM, FM and %BF between
measurement postures using the MF mode. Adult males, however, showed a statistically significant
(p < 0.05) difference in TBW in independent measurement sessions in the supine position
(44.3 ± 1.5 L and 44.2 ± 1.5 L). ECF measured in the standing position was significantly (p < 0.01)
increased compared with the supine position in both groups (males: 19.2 ± 0.6 L vs. 18.9 ± 0.6 L,
boys: 10.0 ± 0.6 L vs. 9.7 ± 0.6 L). There was also a significant (p < 0.01) decrease in ICF in the
standing position in both groups. Using the SF mode, while no difference in Z50 was observed
between measurement postures in males, a significant decrease in Z50 was observed in the standing
position in boys (6“ȍLQWKHVXSLQHSRVLWLRQ vs. “ȍLQWKHVWDQGLQJSRVLWLRQ
p < 0.01). A similar result was observed using 100 kHz (p < 0.05 in males and p < 0.01 in boys).
In addition, adult males showed a significant difference in Z values in each supine measurement,
boys showed significantly different Z values from standing sessions (data not shown).

Table 2. Differences in variables between measurement postures in boys and adult males.
Boys (n = 23) Mean ± SE Adult Males (n = 26) Mean ± SE
Posture Supine Standing p-value Supine Standing p-value
TBW (L) 21.3 ± 1.4 21.2 ± 1.5 0.348 44.3 ± 1.5 44.2 ± 1.5 0.823
ECF (L) 9.7 ± 0.6 10.0 ± 0.6 <0.001 18.9 ± 0.6 19.2 ± 0.6 < 0.001
ICF (L) 11.6 ± 0.9 11.3 ± 0.8 0.005 25.4 ± 0.9 25.0 ± 0.9 0.005
Multi-Frequency
FFM (kg) 29.1 ± 2.0 29.0 ± 2.0 0.353 60.5 ± 2.0 60.4 ± 2.1 0.819
FM (kg) 7.4 ± 0.8 7.6 ± 0.8 0.353 16.0 ± 1.4 16.0 ± 1.5 0.824
%BF (%) 20.1 ± 1.3 20.5 ± 1.3 0.340 20.1 ± 1.2 20.1 ± 1.3 0.820
Z50 ȍ 665.3 ± 16.1 656.9 ± 16.3 <0.001 473.1 ± 8.9 472.2 ± 9.4 0.524
Single-Frequency
Z100 ȍ 635.0 ± 15.8 626.5 ± 16.2 <0.001 447.5 ± 8.6 444.3 ± 9.0 0.021

Using repeated measures ANOVA, the present study also examined the accuracy of the estimated
TBW from the impedance technique using the dilution technique as the reference (Table 3). In
boys, TBW values estimated from the MF mode were comparable to the results from the dilution
technique regardless of the measurement postures. However, in males, the MF mode provided a
significantly (p < 0.01) greater TBW in both postures compared to results from the dilution
technique. When TBW was estimated using the SF mode with a selected prediction equation, no
significant difference with the result from dilution technique was observed in males regardless of
their measurement posture and frequency used. However, TBW estimated from boys using a
 325

frequency of 50 kHz was significantly (p < 0.01) smaller compared to the result from the dilution
technique, regardless of their measurement posture.
The variability of TBW estimation using different frequency modes are shown in Table 4.
Compared with the dilution technique, results from both MF and SF modes showed high
correlation coefficients of 0.956 to 0.988. However, Bland and Altman plots for the supine position
showed that TBW of almost all adult males was overestimated when the MF mode was used
(Figure 1a) and calculated limits of agreement indicated an average of 2.4 L overestimation with a
wide variability of about 3.7 L (Table 4). In comparison, boys showed relatively accurate
estimation of TBW but some boys with a larger TBW were overestimated (Figure 1b). Calculated
limits of agreement indicated that TBW estimation using the MF mode for boys had an
underestimation of 0.38 L with variability of about 2 L. Similarly, TBW from the dilution
technique and the SF mode using 50 kHz was compared. In males, the Bland and Altman plot
indicated that individuals with relatively low TBW values (less than 40 L) were overestimated by
the SF mode whereas the opposite was true for those with relatively high TBW (greater than 40 L).
Limits of agreement indicated, on average, overestimated about 0.7 L with a variability of
approximately 4 L. In boys, TBW was underestimated in all participants with potentially greater
underestimation in individuals with a higher TBW (Figure 2). Limits of agreement showed about
2.7 L of underestimation with a variability of 2.4 L.

Table 3. Differences in TBW estimated from different techniques.

Males (n = 26) Mean ±
Boys (n = 23) Mean ± SE
SE
TBW(2H dilution) (L) 21.7 ± 1.4 41.9 ± 1.5
Supine
Device Mean ± SE p-value Mean ± SE p-value
Multi-Frequency (L) 21.3 ± 1.4 0.333 44.3 ± 1.5 <0.001
Single-Frequency Z50 (L) † 19.0 ± 1.3 <0.001 42.6 ± 1.4 0.566
Single-Frequency Z 100 (L) ‡ - NA 42.2 ± 1.3 1.000
Standing
Device Mean ± SE p-value Mean ± SE p-value
Multi-Frequency (L) 21.2 ± 1.5 0.250 44.2 ± 1.5 <0.001
Single-Frequency Z50 (L) † 19.3 ± 1.3 <0.001 42.7 ± 1.4 0.461
Single-Frequency Z 100 (L) ‡ - NA 42.4 ± 1.4 1.000
†
TBW was estimated using the estimation equation Deurenberg et al. [24] for males and Davies et al. [23] for boys;
‡
TBW was estimated using the estimation equation by Deurenberg et al. [24].

Table 4. Variability of TBW estimation using different frequencies compared with the
dilution technique.
Boys (n = 23) Males (n = 26)
Correlation coefficient 0.988 0.970
Multi-Frequency
Limits of Agreement 0.378 ± 2.14 (2.518, í í“ (1.29, í
Correlation coefficient 0.985 0.956
Single-Frequency Z50 †
Limits of Agreement 2.6524 ± 2.357 (5.009, 0.295) í“ í
†
TBW was estimated using the estimation equation Deurenberg et al. [24] for males and Davies et al. [23] for boys.
 326

Figure 1. Bland and Altman plots between TBW estimated from the dilution technique
and multi-frequency mode for (a) adult males and (b) boys.

(a)

(b)

Figure 2. Bland and Altman plots between TBW estimated from the dilution technique
and single-frequency mode (50 kHz) for (a) adult males and (b) boys.

(a)
 327

Figure 2. Cont.

(b)

4. Discussion

The present study investigated the influence of posture and frequency modes of impedance
technique in the estimation of TBW in adult males and boys. Results confirmed that measurement
posture had no significant influence on TBW estimation and therefore no influence on estimation
of body composition in this convenience sample. The results were consistent with earlier reports
using SF devices [21]. This suggests that influence of body posture during measurements using
impedance technique has a minimal impact on overall estimation of body composition. However,
the present study showed significant changes in ICF and ECF volumes depending on posture
during measurements in both adult males and boys. It has been suggested that a change in posture
will cause redistribution of ECF. The observed results, although smaller in magnitude, were
consistent with a previous study that reported change of ECF and ICF volumes by posture [15].
The result was inconsistent with another study that reported a redistribution of ECF only occurred
between body segments (i.e., the limbs and the trunk) with total ECF volume not altering as a
function of change in measurement posture [27]. The present findings of change in ECF and ICF
with no overall change in TBW may be due to a larger sample size compared to the previous study
that examined only 11 males.
In addition to a possible fluid shift as a result of change in measurement posture, the presence of
stray capacitance may also have influenced the observed outcomes. Weyer and colleagues [28]
suggested that of the two fundamental stray capacitances in the impedance technique, the one
formed between the human body and the ground may have considerable impact on the reading.
Technical factors such as stray capacitances as well as variables such as body proportions, body
density, and resistivity coefficients may be important in interpreting accuracy and quality of results.
Boys showed a significant decrease in Z50 measured in the standing position. Since the frequency
of 50 kHz can go through both ECF and ICF, a reduction in Z may indicate a reduction in TBW.
However, results from the MF mode did not show a change in TBW between postures. As no
 328

differences in Z50 were observed from males, this may suggest that, together with body fluid shift
some technical factors influenced measurements in boys.
While no specific pattern was observed from the MF mode, estimation of TBW in adult males
using the SF mode showed a different pattern depending on whether the participant had TBW
greater or lesser than 40 L. Compared to adult males, boys showed a better correlation and
agreement between the dilution technique and the impedance technique (both SF and MF modes).
A smaller difference in results from the dilution technique and both MF and SF modes and
narrower limits of agreement indicates the accuracy of the impedance technique. Observed
differences in TBW estimation may be associated with a number of technical factors, such as
resistivity coefficients, a body density, a body proportion factor and also prediction equations to
estimate TBW using measured Z for the SF mode. The current study used default values for
resistivity coefficients, a body density and a body proportion that were derived from healthy
Caucasian adults. Although other studies have adopted the same default settings in unhealthy
populations (e.g., obese) [22] or children [29,30], the estimation of TBW or Z values in the current
study, particularly in boys, may be the result of technical error. In addition, the estimated TBW
from a frequency of 50 kHz showed greater underestimation or noticeable pattern in both males
and boys. This may be explained by application of prediction equations to estimate TBW. In this
study, TBW for males was estimated by using the equation by Deurenberg et al. [24] and TBW of
boys were calculated using the equation by Davies et al. [23] that was derived from a small group
of children (n = 26). While the equation by Deurenberg et al. [24] was derived from 139 healthy
volunteers, the equation by Davies et al. [23] was derived from a group of children with particular
health conditions, including growth hormone deficiency, inflammatory bowel disease and diabetes.
In addition, while the equation was derived from both boys and girls, the equation does not include
gender as a variable. Although the age range was matched with the sample of the present study, it
may be possible that application of these equations may also affect accuracy and variability of the
results. These possibilities suggest the importance of considering the abovementioned technical
issues in differentiating biological influence such as fluid shift caused by a change in a
measurement posture and also to improve the accuracy of the results, particularly using the
SF mode.

5. Conclusions

In summary, the present study clarified that estimation of TBW using the MF mode of BIS
device is not affected by measurement posture regardless of participants’ maturational status or
body size. Accordingly, estimation of body composition, including %BF is not affected by change
in measurement posture. However, it should be noted that change in posture may be associated
with fluid shift within the body that may alter values for ECF, ICF and Z. In addition, it is
important to consider technical factors associated with measurements, including stray capacitances,
resistivity coefficients, body proportion factor and also selection of appropriate prediction
equations in order to differentiate the effect of measurement posture and technical error. As
information on appropriate resistivity coefficients and body proportion factors for children is very
scarce future research should consider explore appropriate values for this population. Similarly, as
 329

the current study was based on a relatively small sample size, future research should replicate the
study using a larger group in different age categories as well as including females to examine
gender differences.

Acknowledgments

The present study was funded by the Early Career Researchers Grant scheme of the Institute
of Health and Biomedical Innovation, Queensland University of Technology.

Author Contributions

M.K. contributed in study design, recruitment and data collection, sample analysis, data analysis
and preparation of the manuscript. C.W. contributed in sample analysis and preparation of the
manuscript. A.H. contributed in overall supervision and preparation of manuscript.

Conflicts of Interest

The authors declare no conflict of interest.

References

1. WHO. Obesity: Preventing and Managing the Global Epidemic; WHO Technical Report
Series 894; WHO: Geneva, Switzerland, 2000.
2. WHO/IASO/IOTF. Asia-Pacific Perspective: Redefining Obesity and Its Treatment; Health
Communications Australia Pty Ltd.: Melbourne, Australia, 2000.
3. Baumgartner, R.N.; Chumlea, W.C.; Roche, A.F. Bioelectric impedance phase angle and body
composition. Am. J. Clin. Nutr. 1988, 48, 16–23.
4. Kyle, U.G.; Bosaeus, I.; de Lorenzo, A.D.; Deurenberg, P.; Elia, M.; Gomez, J.M.; Heitmann,
B.L.; Kent-Smith, L.; Melchior, J.C.; Pirlich, M.; et al. ESPEN Guidelines, Bioelectrical
impedance analysis-part I: Review of principles and methods. Clin. Nutr. 2004, 23, 1226–
1243.
5. Jaffrin, M.Y.; Morel, H. Body fluid volumes measurements by impedance: A review of
bioimpedance spectroscopy (BIS) and bioimpedance analysis (BIA) methods. Med. Eng. Phys.
2008, 30, 1257–1269.
6. Ward, L.C.; Dyer, J.M.; Byrne, N.M.; Sharpe, K.K.; Hills, A.P. Validation of a three-frequency
bioimpedance spectroscopic method for body composition analysis. Nutrition 2007, 23,
657–664.
7. Sun, G.; French, C.R.; Martin, G.R.; Younghusband, B.; Green, R.C.; Xie, Y.-G.; Mathews, M.;
Barron, J.R.; Fitzpatrick, D.G.; Gulliver, W.; et al. Comparison of multifrequency bioelectrical
impedance analysis with dual-energy X-ray absorptiometry for assessment of percentage body fat
in a large, healthy population. Am. J. Clin. Nutr. 2005, 81, 74–78.
8. Demura, S.; Sato, S.; Kitabayashi, T. Percentage of total body fat as estimated by three
automatic bioelectrical impedance analyzers. J. Physiol. Anthropol. Appl. Hum. Sci. 2004, 23,
 330

93–99.
9. Lukaski, H.C. Evolution of bioimpedance: A circuitous journey from estimation of
physiological function to assessment of body composition and a return to clinical research.
Eur. J. Clin. Nutr. 2013, 67, S2–S9.
10. Kyle, U.G.; Bosaeus, I.; de Lorenzo, A.D.; Deurenberg, P.; Elia, M.; Manuel Gómez, J.;
Lilienthal Heitmann, B.; Kent-Smith, L.; Melchior, J.C.; Pirlich, M.; et al. ESPEN Guidelines,
Bioelectrical impedance analysis-part II: Utilization in clinical practice. Clin. Nutr. 2004, 23,
1430–1453.
11. Buchholz, A.C.; Bartok, C.; Schoeller, D.A. The validity of bioelectrical impedance models in
clinical populations. Nutr. Clin. Prac. 2004, 19, 433–446.
12. Smith, D.J.; Johnson, M.; Nagy, T. Precision and accuracy of bioimpedance spectroscopy for
determination of in vivo body composition in rats. Int. J. Body Compos. Res. 2009, 7, 21–26.
13. Ward, L.C. Segmental bioelectrical impedance analysis: An update. Curr. Opin. Clin. Nutr.
Metab. Care 2012, 15, 424–429.
14. Bosy-Westphal, A.; Schautz, B.; Later, W.; Kehayias, J.J.; Gallagher, D.; Muller, M.J.
What makes a BIA equation unique? Validity of eight-electrode multifrequency BIA to
estimate body composition in a healthy adult population. Eur. J. Clin. Nutr. 2013, 67,
S14–S21.
15. Scharfetter, H.; Monif, M.; László, Z.; Lambauer, T.; Hutten, H.; Hinghofer-Szalkay, H.
Effect of postural changes on the reliability of volume estimations from bioimpedance
spectroscopy data. Kidney Int. 1997, 51, 1078–1087.
16. National Health and Medical Research Council. National Statement on Ethical Conduct in
Research Involving Humans, 1st ed.; AusInfo.: Canberra, Australia, 1999.
17. Marfell-Jones, M.; Olds, T.; Stewart, A.; Carter, J.E.L. International Standards for
Anthropometric Assessment (Revised 2006), 2nd ed.; The International Society for the
Advancement of Kinanthropometry: Potchefstroom, South Africa, 2006.
18. Gore, C.; Norton, K.; Olds, T.; Whittingham, N.; Birchall, K.; Clough, M.; Dickerson, B.;
Downie, L. Accreditation in anthropometry: An Australian model. In Anthropometrica;
Norton, K., Olds, T., Eds.; University of New South Wales Press: Sydney, Australia, 1996;
pp. 395–411.
19. Wang, J.; Thornton, J.C.; Kolesnik, S.; Pierson, R.N., Jr.; Anthropometry in body composition:
An overview. Ann. N. Y. Acad. Sci. 2000, 904, 317–326.
20. ImpediMed. Imp SFB7 Instructions for Use; ImpediMed: Eight Mile Plains, Australia, 2007.
21. Demura, S.; Yamaji, S.; Goshi, F.; Nagasawa, Y. The influence of posture change on
measurements of relative body fat in the bioimpedance analysis method. J. Physiol. Anthropol.
Appl. Hum. Sci. 2001, 20, 29–35.
22. Moon, J.R.; Smith, A.E.; Tobkin, S.E.; Lockwood, C.M.; Kendall, K.L.; Graef, J.L.;
Roberts, M.D.; Dalbo, V.J.; Kerksick, C.M.; Cramer, J.T.; et al. Total body water changes
after an exercise intervention tracked using bioimpedance spectroscopy: A deuterium oxide
comparison. Clin. Nutr. 2009, 28, 516–525.
23. Davies, P.S.W.; Preece, M.A.; Hicks, C.J.; Halliday, D. The prediction of total body water
 331

using bioelectrical impedance in children and adolescents. Ann. Hum. Biol. 1988, 15, 237–240.
24. Deurenberg, P.; Tagliabue, A.; Schouten, F.J.M. Multi-frequency impedance for the prediction
of extracellular water and total body water. Br. J. Nutr. 1995, 73, 349–358.
25. International Atomic Energy Agency. Introduction to Body Composition Assessment Using the
Deuterium Dilution Technique with Analysis of Urine Samples by Isotope Ratio Mass
Spectrometry; IAEA: Vienna, Austria, 2010.
26. Bland, J.M.; Altman, D.G. Statistical methods for assessing agreement between two methods
of clinical measurement. Lancet 1986, 1, 307–310.
27. Zhu, F.; Schneditz, D.; Wang, E.; Levin, N.W. Dynamics of segmental extracellular volumes
during changes in body position by bioimpedance analysis. J. Appl. Physiol. 1998, 85,
497–504.

© 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).
 MDPI AG
Klybeckstrasse 64
4057 Basel, Switzerland
Tel. +41 61 683 77 34
Fax +41 61 302 89 18
http://www.mdpi.com/

Nutrients Editorial Office

E-mail: nutrients@mdpi.com
http://www.mdpi.com/journal/nutrients
0'3,‡%DVHO‡%HLMLQJ
,6%1
ZZZPGSLFRP