Optimal Duration of Exc Bfeeding Review Eng

8/8/2019 Optimal Duration of Exc Bfeeding Review Eng

1/52

DEPARTMENT OF NUTRITION FOR HEALTH AND DEVELOPMENTDEPARTMENT OF CHILD AND ADOLESCENT HEALTH AND DEVELOPMENT

WORLD HEALTH ORGANIZATION

W H O / N H D / 0 1 . 0 8W H O / F C H / C A H / 0 1 . 2 3O R I G I N A L : E N G L I S HD I S T R . : G E N E R A L

THE OPTIMALDURATION OF

EXCLUSIVEBREASTFEEDINGA SYSTEMATIC

REVIEW


2/52

For further information please contact:

Department of Nutrition for Healthand Development (NHD)

World Health Organization20 Avenue Appia1211 Geneva 27

SwitzerlandTel: +41 22 791 3320Fax: +41 22 791 4156

email: [email protected]: www.who.int/nut

Department of Child and AdolescentHealth and Development (CAH)

World Health Organization20 Avenue Appia

1211 Geneva 27Switzerland

Tel +41-22 791 3281Fax +41-22 791 4853

email: [email protected]: http://www.who.int/child-adolescent-health


3/52

DEPARTMENT OF NUTRITION FOR HEALTH AND DEVELOPMENTDEPARTMENT OF CHILD AND ADOLESCENT HEALTH AND DEVELOPMENT

WORLD HEALTH ORGANIZATION

THE OPTIMAL DURATIONOF EXCLUSIVE BREASTFEEDING

A SYSTEMATIC REVIEWMICHAEL S. KRAMER, MD

RITSUKO KAKUMA, MSc

W H O / N H D / 0 1 . 0 8W H O / F C H / C A H / 0 1 . 2 3

O R I G I N A L : E N G L I S HD I S T R . : G E N E R A L


4/52

World Health Organization, 2002

This document is not a formal publication of the World Health Organization (WHO), and all rights are reservedby the Organization. The document may, however, be freely reviewed, abstracted, reproduced and translated,

in part or in whole, but not for sale for use in conjunction with commercial purposes.

The views expressed in documents by named authors are solely the responsibility of those authors.

Designed by minimum graphicsPrinted in Switzerland


5/52

Contents

Abstract 1

Introduction 3

Search methods 5

Selection criteria for studies 5

Types of intervention/exposure 5

Types of outcome measures 5Search strategy for identification of studies 5

Selection of studies for inclusion 6

Review methods 9

Description of studies 9

Methodologic quality of included studies 9

Data collection 10

Data analysis 10

Results 14Comparison 1: Controlled trials of exclusive vs mixed breastfeeding for

46 months, developing countries 14

Comparison 2: Observational studies of exclusive vs mixed breastfeeding for37 months, developing countries 15

Comparison 3: Observational studies of exclusive vs mixed breastfeeding for37 months, developed countries 16

Discussion 19

Summary of findings 19

Implications for future research 19Conclusion 19

References 21

Annexes 25

Annex 1. Comparison 01: Exclusive vs mixed breastfeeding 46 months,developing countries, controlled trials 25

Annex 2. Comparison 02: Exclusive vs mixed breastfeeding 37 months,developing countries, observational studies 32

Annex 3. Comparison 03: Exclusive vs mixed breastfeeding 37 months,developed countries, observational studies 36

iii

R E F E R E N C E S


6/52

1

E X P E R T C O N S U L T A T I O N O N O P T I M A L D U R A T I O N O F E X C L U S I V E B R E A S T F E E D I N G

Background: The longstanding debate over the optimalduration of exclusive breastfeeding has centered on theso-called weanlings dilemma in developing countries:the choice between the known protective effect ofexclusive breastfeeding against infectious morbidityand the (theoretical) insufficiency of breast milk aloneto satisfy the infants energy and micronutrient require-

ments beyond 4 months of age. The debate over whetherto recommend exclusive breastfeeding for 46 monthsvs about 6 months has recently become more intense.

Objectives: The primary objective of this review wasto assess the effects on child health, growth, anddevelopment, and on maternal health, of exclusivebreastfeeding for 6 months vs exclusive breastfeedingfor 34 months with mixed breastfeeding (introductionof complementary liquid or solid foods with continuedbreastfeeding) thereafter through 6 months.

Search strategy: Two independent literature searches

were carried out, together comprising the followingdatabases: MEDLINE (as of 1966), Index Medicus (priorto 1966), CINAHL, HealthSTAR, BIOSIS, CABAbstracts, EMBASE-Medicine, EMBASE-Psychology,Econlit, Index Medicus for the WHO EasternMediterranean Region, African Index Medicus, Lilacs(Latin American and Carribean literature), EBMReviews-Best Evidence, the Cochrane Database ofSystematic Reviews, and the Cochrane ControlledTrials Register. No language restrictions were imposed.The two searches yielded a total of 2,668 unique

citations. Contacts with experts in the field yieldedadditional published and unpublished studies.

Selection criteria: We selected all internally-controlled clinical trials and observational studiescomparing child or maternal health outcomes withexclusive breastfeeding for 6 or more months vsexclusive breastfeeding for at least 34 months withcontinued mixed breastfeeding until at least 6 months.Studies were stratified according to study design(controlled trials vs observational studies), provenance(developing vs developed countries), and timing of

compared feeding groups (37 months vs later).

Data collection and analysis: Two reviewers indepen-dently assessed study quality (using a priori assessmentcriteria) and extracted data.

Main results: Sixteen independent studies meeting theselection criteria were identified by the literature search:7 from developing countries (2 of which were controlledtrials in Honduras) and 9 from developed countries (all

observational studies). The two trials did not receivehigh methodologic quality ratings but were nonethelesssuperior to any of the observational studies included inthis review. The observational studies were of variablequality; in addition, their nonexperimental designs werenot able to exclude potential sources of confoundingand selection bias. Definitions of exclusive breastfeedingvaried considerably across studies. Neither the trials northe observational studies suggest that infants whocontinue to be exclusively breastfed for 6 months showdeficits in weight or length gain, although larger sample

sizes would be required to rule out small increases inthe risk of undernutrition. The data are scarce withrespect to iron status, but at least in developing countrysettings where newborn iron stores may be suboptimal,suggest that exclusive breastfeeding without ironsupplementation through 6 months may compromisehematologic status. Based primarily on an observationalanalysis of a large randomized trial in Belarus, infantswho continue exclusive breastfeeding for 6 months ormore appear to have a significantly reduced risk of oneor more episodes of gastrointestinal infection. Nosignificant reduction in risk of atopic eczema, asthma,or other atopic outcomes has been demonstrated instudies from Finland, Australia, and Belarus. Data fromthe two Honduran trials suggest that exclusivebreastfeeding through 6 months is associated withdelayed resumption of menses and more rapidpostpartum weight loss in the mother.

Reviewers conclusions: We found no objectiveevidence of a weanlings dilemma. Infants who areexclusively breastfed for 6 months experience lessmorbidity from gastrointestinal infection than thosewho are mixed breastfed as of 3 or 4 months, and nodeficits have been demonstrated in growth amonginfants from either developing or developed countrieswho are exclusively breastfed for 6 months. Moreover,

Abstract


7/52

T H E O P T I M A L D U R A T I O N O F E X C L U S I V E B R E A S T F E E D I N G

2

the mothers of such infants have more prolongedlactational amenorrhea. Although infants should stillbe managed individually so that insufficient growth or

other adverse outcomes are not ignored and appropriateinterventions are provided, the available evidencedemonstrates no apparent risks in recommending, as

public health policy, exclusive breastfeeding for the first6 months of life in both developing and developedcountry settings. Large randomized trials are

recommended in both types of setting to rule out smalladverse effects on growth and to confirm the reportedhealth benefits of exclusive breastfeeding for 6 months.


8/52

3

I N T R O D U C T I O N

Introduction

The epidemiologic evidence is now overwhelming that,even in developed countries, breastfeeding protectsagainst gastrointestinal and (to a lesser extent)respiratory infection, and that the protective effect isenhanced with greater duration and exclusivity ofbreastfeeding.1317 (Greater duration and exclusivityis used in a general sense here; the references cited do

not pertain specifically to the subject of this review, i.e.,the optimal duration of exclusive breastfeeding.)Prolonged and exclusive breastfeeding has also beenassociated with a reduced risk of the sudden infant deathsyndrome (SIDS)18 and of atopic disease,1921 and somestudies even suggest acceleration of neurocognitivedevelopment2228 and protection against long-termchronic conditions and diseases like obesity,2931 type Idiabetes mellitus,32,33 Crohns disease,34 and lym-phoma.35,36 Maternal health benefits have also receivedconsiderable attention in developed countries, inclu-ding possible protection against breast cancer amongpremenopausal women,37-39 ovarian cancer,40 and osteo-porosis.4143

Although growth faltering is uncommon in developedcountries, a recent pooled analysis of U.S., Canadian,and European data sets undertaken by the WHOWorking Group on Infant Growth showed that infantsfrom developed countries who follow current WHOfeeding recommendations (to exclusively breastfeed for4 to 6 months of age and to continue breastfeeding withadequate complementary foods up to 2 years of age)show a deceleration in both weight and length gainrelative to the international WHO/NCHS growthreference from around 3 to 12 months, with partialcatch-up in the second year.44,45 More recent studies,including a Danish population-based cohort study,46 ananalysis based on the third U.S. National Health andNutrition Examination Survey,47 and the Euro-Growthstudy48 have also reported an association betweenprolonged and exclusive breastfeeding and slowergrowth during infancy. Unfortunately, the currentWHO/NCHS reference is based on the Fels Longitu-dinal Study, which was conducted many decades ago in

infants who were primarily bottle-fed. WHO hastherefore embarked on an ambitious study to establishnew growth standards for infants following currentfeeding recommendations.49,50 In developed country

The debate over the optimal duration of exclusivebreastfeeding has had a long history. Growth falteringis a commonly observed phenomenon in developingcountries after about 3 months of age.1,2 This growthfaltering has traditionally been attributed to threefactors: (1) the inadequacy of energy intake from breastmilk alone after 3 or 4 months; (2) the poor nutritional

quality (i.e., low energy and micronutrient content) ofthe complementary foods commonly introduced inmany developing countries; and (3) the adverse effectsof infection on energy intake and expenditure. Theinadequacy of breast milk for energy requirementsbeyond 3 or 4 months was initially based on calculationsmade by the Food and Agricultural Organization (FAO)and World Health Organization (WHO) in 1973.3 Morecareful studies since the 1980s47 and a later FAO/WHOreport,8 however, have shown that the earlier FAO/WHO figures substantially overestimate true energyrequirements in infancy.47

The belief that breast milk alone is nutritionallyinsufficient after 3 or 4 months, combined with the factthat complementary foods given in many developingcountries are both nutritionally inadequate andcontaminated, led to concern about the so-calledweanlings dilemma.9,10 Breastfeeding is a life-and-death issue in developing countries. A recent meta-analysis11 reported markedly reduced mortality(especially due to infectious disease) with breastfeedingeven into the second year of life. A recent study fromIndia reported an increased risk of postneonatal

mortality associated with exclusive breastfeeding >3months,12 but reverse causality (illness prior to deathpreventing the infants acceptance of complementaryfoods), selection bias (exclusion of infants who diedprior to each cross-sectional period), or uncontrolledconfounding might explain this result.

The weanlings dilemma and the risk of mortalityassociated with early introduction of complementaryfoods are concerns primarily in developing countries.In most developed countries, uncontaminated,nutritionally adequate complementary foods are readily

available, and growth faltering is relatively uncommon.With the resurgence of breastfeeding in developedcountries, however, recent attention has turned to theimportance of promoting its duration and exclusivity.


9/52


4

settings, it is not at all clear that the more rapid growthreported in infants who are formula-fed, or breastfedless exclusively and for a shorter duration, is an

advantage. Moreover, a recent, large randomized trialfrom Belarus has reported that breastfed infants bornand followed at sites randomized to a breastfeedingpromotion intervention (and who were breastfed moreexclusively and for a longer duration) actually grew morerapidly in the first 69 months than those born andfollowed at control (nonintervention) sites.51,52,53

In the last years, recommendations for the optimalduration of exclusive breastfeeding promoted by WHOand UNICEF started to differ. WHO had continued torecommend exclusive breastfeeding for 4 to 6 months,

with the introduction of complementary foods there-after,54 whereas UNICEF preferred the wording forabout 6 months.55 This led to concerns in the largerinfant nutrition and public health communities.56 TheAmerican Academy of Pediatrics position is unclear;in two different sections of their Pediatric NutritionHandbook,57 they alternatively recommend human milkas the exclusive nutrient source ... during the first 6months (p. 18) and to delay introduction of solid foodsuntil 4 to 6 months (p. 38).

Until recently, the only scientific evidence contributing

to this debate was based on observational studies, withwell-recognized sources of potential bias. Some of thesebiases tend to favor exclusively breastfed infants, whileothers favor those who receive earlier complementaryfeeding. Infants who continue to be exclusively breastfedtend to be those who remain healthy and on anacceptable growth trajectory; significant illness orgrowth faltering can lead to interruption of breastfeedingor supplementation with infant formula or solidfoods.58,59 Confounding by indication60 [i.e., the reason(indication) for the supplementation affects the out-come, rather than the supplementation itself] is another

important bias, and could operate in either direction.Poorly-growing infants (especially those in developingcountries) are likely to receive complementary feedingsearlier because of their slower growth. In developedcountries, however, rapidly-growing infants may requiremore energy than can be met by the increasingly spacedfeedings typical of such settings. This may result incrying and poor sleeping, supplementation with formulaand/or solid foods, reduced suckling, and a vicious cycleleading to earlier termination of breastfeeding. Reversecausality is another potential source of bias, particularly

with respect to infectious morbidity and neuromotordevelopment.61 Infants who develop a clinicallyimportant infection are likely to become anorectic andto reduce their breast milk intake, which can in turn

lead to reduction in milk production and eventermination of breastfeeding. This is particularly aproblem in cross-sectional studies, because the temporal

sequence of the early signs of infection and terminationof breastfeeding may not be adequately appreciated;infection may be blamed on the termination of breast-feeding, rather than the reverse. Advanced neuromotordevelopment may also lead to earlier induction of solidfoods, which could then receive credit for acceleratingmotor development.62 Finally, other unmeasured orpoorly measured confounding variables could also biasthe association between introduction of complementaryfoods and infant health outcomes.

Because of these well-recognized problems in

observational studies, two recent controlled clinicaltrials63,64 from Honduras have attracted considerableinterest. These trials allocated infants born to eithercontinue breastfeeding exclusively for 6 months or toreceive solid foods along with continued breastfeedingfrom 4 months onwards. The results showed nosignificant benefit for growth nor any disadvantage formorbidity with the earlier introduction of comple-mentary foods, but the small sample sizes and publishedanalyses based on compliance with allocation (i.e., noton intention-to-treat) have prevented universalacceptance of these results.65 In addition, the comple-

mentary foods used were those commonly found indeveloped countries, rather than in those traditionallyused in Honduras or other developing countries.

Most studies have reported effects in terms of groupdifferences in mean z-scores or in mean weight or lengthgain; few have provided data on the tails (extremes) ofthe distribution, e.g., anthropometric indices (z-scores


10/52

5

Search methods

Types of outcome measures

No infant or maternal health outcomes were excludedfrom consideration. The infant outcomes specificallysought (but not necessarily found) included growth[weight, length, and head circumference and z-scores(based on the WHO/NCHS reference) for weight-for-age (WAZ), length-for-age (LAZ), and weight-for-

length (WLZ)], infections, morbidity, mortality,micronutrient status, neuromotor and cognitivedevelopment, asthma, atopic eczema, other allergicdiseases, Type 1 diabetes, blood pressure, and subsequentadult chronic diseases such as coronary heart disease,hypertension, Type 2 diabetes, and inflammatory andautoimmune diseases. Maternal outcomes soughtincluded postpartum weight loss, duration of lactationalamenorrhea, and such chronic diseases as breast andovarian cancer and osteoporosis.

Search strategy for identificationof studies

In order to capture as many relevant studies as possible,two independent literature searches were conducted:one by staff at the Department of Nutrition of WHOand one by the authors. The search details are shownbelow.

The search by WHO was conducted between June andAugust 2000 in the following databases: MEDLINE(1966 to June 2000), Pre-MEDLINE (Index Medicus

previous to 1966), CINAHL (1982 to June 2000),HealthSTAR (1975 to August 2000), EBM Reviews-Best Evidence (1991 to July/August 2000), SocioFile(1974 to July 2000), Cochrane Database of SystematicReviews (Issue 2, 2000), CAB Abstracts (1973 to July2000), EMBASE-Psychology (1987 to 3rd Quarter,2000), Econlit (1969 to August 2000), Index Medicusfor the WHO Eastern Mediterranean (IMEMR),African Index Medicus (AIM), and Lilacs (LatinAmerican and Caribbean literature).

Where applicable, the medical subject heading (MeSH)

breast feeding, and otherwise the free language termsbreast-feeding, breast feeding, or breastfeedingcombined with exclusive or exclusively were used

S E A R C H M E T H O D S

Selection criteria for studies

We selected controlled clinical trials and observationalstudies, published in all languages, examining whetheror not exclusive breastfeeding (EBF) until 6 months ofage has an impact on growth, development, morbidity,and survival of healthy, term infants and their mothers.Studies of (or including) low-birth-weight (


11/52


6

in the search strategy. The search yielded 1,423 citations(MEDLINE 686, Pre-MEDLINE 15, CINAHL 25,HealthSTAR 1, EBM-Best Evidence 7, Socio File 2,

Cochrane Database of Systematic Reviews 8, CABAbstracts 680, EMBASE-Psychology 4, other databases0). Once duplicates were removed, 1,035 citationsremained; these were then assessed for eligibility.

The authors search was conducted on August 12, 2000in the following databases: MEDLINE (1966 to June2000), CINAHL (1982 to April 2000), HealthSTAR(1975 to August 2000), BIOSIS (1989 to 2000), CABAbstracts (1973 to June 2000), Cochrane Database ofSystematic Reviews (Issue 3, 2000), CochraneConrolled Trials Register (Issue 3, 2000), and EMBASE-

Medicine (1980 to present).The terms breast feeding, infant, and growth, asMeSH headings and text words, were combined in thesearch strategy. This search yielded a total of 2,496citations (MEDLINE yielded 1,079 citations, CINAHL75, HealthSTAR 2, BIOSIS 190, CAB 614, CochraneDatabase of Systematic Reviews 25, CochraneControlled Trials Register 122, and EMBASE 389).Once duplicates among the databases were removed,1,845 citations remained, 1,633 of which were differentfrom the 1,035 identified by the WHO search. Thus

2,668 unique citations were identified by the twosearches.

For both searches, every effort was made to identifyrelevant non-English language articles and abstracts.Given their own backgrounds, the reviewers themselveswere able to determine the eligibility of articles inFrench, Spanish, and Japanese. For publications in otherlanguages, two options were available. Many articles inlanguages other than English provided English abstracts.As such, all potentially relevant articles were obtainedand checked for availability of English abstracts. If such

abstracts were not available, or were available but didnot provide enough information to determine theireligibility, then assistance was requested from WHO todetermine their eligibility for inclusion. No article orabstract was excluded because of language ofpublication.

In addition to the studies found through the twoelectronic searches, reference lists of identified articleswere checked, and contacts with experts in the fieldwere made to identify other potentially relevantpublished or unpublished studies. Attempts were made

to contact the authors of all studies that qualified forinclusion in the review to obtain methodologic details,clarify inconsistencies, and obtain unpublished data.

Selection of studies for inclusion

Many studies were identified that either comparedoutcomes in infants receiving EBF vs MBF or

investigated the effects of age at introduction ofnonbreast-milk liquid or solid foods. The vast majorityof these studies were ineligible for inclusion, however,because they did not ensure EBF 3 months prior tointroducing these complementary foods in the MBFgroup and/or a comparison group with EBF 6 months.

We identified 32 unique citations (articles or abstracts)that met the selection criteria, comprising 16 separatestudies. Of the 16 included studies, 7 were carried outin developing countries and the other 9 in developedcountries.

Eight of the 32 total citations were found by bothsearches63,64,6772; 7 were identified only by the WHOsearch21,7378; 6 were found only by the authorssearch.45,62,7982 Eleven additional citations were locatedthrough contacts with experts and reference lists ofrelevant articles.7,44,51,53,8388 The selected studies arelisted below. They are generally referred to by the lastname of the first author of the earliest citation for eachstudy, along with the year of publication of that citation.Thirty-five references are listed; one85 of the 32 uniquecitations appears twice, and another7 3 times.

Developing countries

Adair 1993

a. Adair L, Popkin BM, Vanderslice J, Akin J, GuilkeyD, Black R, et al. Growth dynamics during the firsttwo years of life: a prospective study in thePhilippines. Eur J Clin Nutr 1993;47:4251.79

b. Brown K, Dewey K, Allen L. Complementary Feedingof Young Children in Developing Countries: A Review

of Current Scientific Knowledge. Geneva: WHO, 1998,pp. 3032.7

Brown 1991

a. Brown KH. The relationship between diarrhoealprevalence and growth of poor infants varies withtheir age and usual energy intake (abstract). FASEBJ 1991;5:A1079.86

b. Brown K, Dewey K, Allen L. Complementary Feedingof Young Children in Developing Countries: A Review

of Current Scientific Knowledge. Geneva: WHO, 1998,

pp. 3032.7


12/52

7

Castillo 1996

Castillo C, Atalah E, Riumallo J, Castro R. Breast-feeding and the nutritional status of nursing children

in Chile. Bull PAHO 1996;30:125133.68

Cohen 1994 (first Honduras trial)

a. Cohen RJ, Brown KH, Canahuati J, Rivera LL,Dewey KG. Effects of age of introduction ofcomplementary foods on infant breast milk intake,total energy intake, and growth: a randomisedintervention study in Honduras. Lancet 1994;344:288293.63

b. Cohen RJ, Brown KH, Canahuati J, Rivera LL,Dewey KG. Determinants of growth from birth to

12 months among breast-fed Honduran infants inrelation to age of introduction of complementaryfoods. Pediatrics 1995;96:504510.69

c. Dewey KG, Cohen RJ, Rivera LL, Canahuati J,Brown KH. Do exclusively breast-fed infants requireextra protein? Pediatr Res 1996;39:303307.72

d. Dewey KG, Cohen RJ, Rivera LL, Canahuati J,Brown KH. Effects of age at introduction ofcomplementary foods to breast-fed infants onduration of lactational amenorrhea in Honduranwomen.Am J Clin Nutr 1997;65:14031409.71

e. Dewey KG, Cohen RJ, Rivera LL, Brown KH. Effectsof age of introduction of complementary foods oniron status of breast-fed infants in Honduras. Am JClin Nutr 1998;67:878884.71

f. Dewey KG, Cohen RJ, Brown KH, Rivera LL. Effectsof exclusive breastfeeding for four versus six monthson maternal nutritional status and infant motordevelopment: results of two randomized trials inHonduras.J Nutr 2001;131:262267.85

Dewey 1999 (second Honduras trial)

a. Dewey KG, Cohen RJ, Rivera LL, Brown KH. Effectsof age of introduction of complementary foods onmicronutrient status of term, low-birthweight,breastfed infants in Honduras. FASEB J 1998;12:A648.70

b. Dewey KG, Cohen J, Brown KH, Rivera LL. Age ofintroduction of complementary foods and growthof term, low-birth-weight, breast-fed infants: arandomized intervention study in Honduras. Am JClin Nutr 1999;69:679686.64

c. Dewey KG, Cohen RJ, Brown KH, Rivera LL. Effectsof exclusive breastfeeding for four versus six monthson maternal nutritional status and infant motor

development: results of two randomized trials inHonduras.J Nutr 2001;131:262267.85

Huffman 1987

Huffman SL, Ford K, Allen HA, Streble P. Nutritionand fertility in Bangladesh: breastfeeding and postpartum amenorrhoea. Population Studies 1987;41:447462.84

Simondon 1997

Simondon KB, Simondon F. Age at introduction ofcomplementary food and physical growth from 2 to9 months in rural Senegal. Eur J Clin Nutr 1997;51:703707.80

Developed countries

keson 1996

a. keson PMK, Axelsson IE, Raiha NCR. Human milkand standard infant formula together with highquality supplementary foods is sufficient for normalgrowth during infancy. Pediatr Res 1996;39(Suppl):313A.81

b. keson PMK, Axelsson IE, Raiha NCR. Growth andnutrient intake in three- to twelve-month-old infants

fed human milk or formulas with varying proteinconcentrations. J Pediatr Gastroenterol Nutr 1998;26:18.67

c. keson PMK, Axelsson IE, Raiha NCR. Protein andamino acid metabolism in three-to twelve-month-old infants fed human milk or formulas with varyingprotein concentrations.J Pediatr Gastroenterol Nutr1998;26:297304.73

Duncan 1993

Duncan B, Ey J, Holberg CJ, Wright AL, MartinezFD, Taussig LM. Exclusive breast-feeding for at least4 months protects against otitis media. Pediatrics1993; 91:867872.74

Heinig 1993

Heinig MJ, Nommsen LA, Peerson JM, LonnerdalB, Dewey KG. Intake and growth of breast-fed andformula-fed infants in relation to the timing ofintroduction of complementary foods: theDARLING study.Acta Paediatr Scand 1993;82:9991006.62

Kajosaari 1983

a. Kajosaari M, Saarinen UM. Prophylaxis of atopicdisease by six months total solid food elimination.Evaluation of 135 exclusively breast-fed infants of

S E A R C H M E T H O D S


13/52


8

atopic families. Acta Paediatr Scand 1983;72:411414.75

b. Kajosaari M. Atopy prophylaxis in high-risk infants.

Prospective 5-year follow-up study of children withsix months exclusive breastfeeding and solid foodelimination.Adv Exp Med Biol 1991;453458.77

c. Kajosaari M. Atopy prevention in childhood: the roleof diet: prospective 5-year follow-up of high-riskinfants with six months exclusive breastfeeding andsolid food elimination. Pediatr Allerg Immunol1994;5:2628.76

Kramer 2000

a. Kramer MS, Chalmers B, Hodnett ED, et al.

Breastfeeding and infant growth: biology or biasPediatr Res 2000;47:151A.52

b. Kramer MS, Chalmers B, Hodnett ED, et al.Promotion of Breastfeeding Intervention Trial(PROBIT): a cluster-randomized trial in the Republicof Belarus. In: Koletzko, Michaelsen KF, Hernell O,editors. Short and Long Term Effects of Breast Feedingon Child Health. New York: Kluwer Academic/PlenumPublishers, 2000, pp. 327345. 53

c. Kramer MS, Chalmers B, Hodnett ED, et al.Promotion of breastfeeding intervention trial(PROBIT): a randomized trial in the Republic ofBelarus.JAMA 2001;285:413420.51

Oddy 1999

Oddy W, Holt P, Sly P, Read A, Landau L, Stanley F,et al. Association between breast feeding and asthmain 6 year old children: findings of a prospective birthcohort study. BMJ 1999;319:815819.21

Pisacane 1995

Pisacane A, de Vizla B, Valiante A, Vaccaro F, Russo

M, Grillo G, et al. Iron status in breast-fed infants.J Pediatr 1995;127:429431.78

WHO 1994

a. WHO Working Group on Infant Growth. An evaluationof infant growth. Doc WHO/NUT/94.8. Geneva:

World Health Organization, 1994.44

b. Dewey KG, Peerson JM, Brown KH, Krebs NF,Michaaelsen KF, Persson LA, Salmenpera L,Whitehead RG, Yeung DL and the World HealthOrganization Working Group on Infant Growth.Growth of breast-fed infants deviates from currentreference data: a pooled analysis of US, Canadian,and European data sets. Pediatrics 1995;96:495503.45

c. WHO Working Group on Infant Growth. Anevaluation of infant growth: the use andinterpretation of anthropometry in infants. BullWHO 1995;73:165174.88

d. Brown K, Dewey K, Allen L. Complementary feedingof young children in developing countries: a review of

current scientific knowledge. Geneva: WHO, 1998, pp.2829.7

WHO 1997

a. Frongillo EA Jr, de Onis M, Garza C, the WorldHealth Organization Task Force on Methods for theNatural Regulation of Fertility. Effects of timing ofcomplementary foods on postnatal growth. FASEB J1997;11:A574.82

b. WHO Working Group on the Growth ReferenceProtocol and the WHO Task Force on Methods forthe Natural Regulation of fertility. Growth of healthyinfants and the timing, type and frequency ofcomplementary foods. Am J Clin Nutr 2001. (Inpress).83


14/52

9

Review methods

1. Was the study described as randomized (this includesthe use of words such as randomly, random, andrandomization)?

a) not random or not mentioned (0)

b) random, described, and inappropriate (0)

c) random, not described (+1)

d) random, described, and appropriate (+2)

2. Was the study described as double-blind?

a) not double-blind (0)

b) double-blind, described, and not appropriate (0)

c) double-blind, not described (+1)

d) double-blind, described, and appropriate (+2)

3. Was there a description of withdrawals and dropouts?

Withdrawals (number and reasons) must be described

by group to get 1 pointObservational (cohort, case-control, and cross-sectionalstudies) were assessed for control for confounding, lossesto follow-up, and assessment of outcome as follows:

1. For growth and morbidity outcomes, control forconfounding by socioeconomic status, water supply,sanitation facilities, parental height and weight, birthweight, and weight and length at 3 months (or ageat which complementary feeding was introduced inthe MBF group):

A. Control for all (or almost all) pertinentconfounders

B. Partial control for some confounders

C. No control for confounding

2. Losses to follow-up:

A. Losses to follow-up were symmetrical and lessthan 15% in each group

B. Losses were 1525% and symmetrical

C. Losses were >25%, asymmetrical, or not reported

(and all cross-sectional studies)

R E V I E W M E T H O D S

Description of studies (see tables of includedstudies)

Methodologic quality of included studies(see tables of included studies)

Studies under consideration were evaluated for metho-

dologic quality and appropriateness for inclusionwithout consideration of their results. The criteria forquality assessment were developed a priori and arepresented below.

We used Cochrane criteria for assessing controlledclinical trials. As shown below, this method rates trialson three elements:

1. Adequacy of randomization and concealment:

A. Randomized and concealed appropriately

B. Randomized appropriately but concealmentunclear from the description

C. Not (or not reported as) randomized and/orinadequate concealment

2. Losses to follow-up and analysis:

A. Used intention-to-treat (ITT) analysis, withlosses to follow-up symmetrical and


15/52


10

3. Assessment of outcome (outcome-specific):

A. Blinding of observers or objective outcomes(e.g., measured weight)

B. Nonblinding of observers or measurements thatcould be affected by bias (including length, headcircumference, and self-reported outcomes)

Quality assessments of all of the eligible studies werecarried out independently by the two reviewers.Disagreements were resolved by consensus.

Data collection

Data were extracted independently by both reviewers,

with disagreements resolved by consensus. Attemptswere made to contact authors of all included studies toobtain additional data, resolve inconsistencies, andobtain additional methodologic details.

Data analysis

The studies were stratified according to study design(controlled trials vs observational studies), provenance

(developing vs developed country), and timing offeeding comparison [(37 months vs prolonged (>6months)]. One study,82,83 based on a pooled analysis of2 developed and 3 developing countries has beenincluded with developed country studies because of theselection criteria (literate, educated, urban mothers) andthe observed high infant growth rates. Analyses werecarried out using the Review Manager 4.1 software forpreparing Cochrane reviews. Effect measures arereported as the fixed-effect weighted mean difference(WMD) and its 95% confidence interval (CI) for

continuous outcomes and the fixed-effect pooledrelative risk (RR) and its 95% CI. For most continuousoutcomes, a positive WMD denotes a higher (morefavorable) value in the EBF group. All dichotomousoutcomes are formulated as adverse; thus an RR


16/52

11

Study Methods Participants Interventions Outcomes Notes

Cohen 1994 Design: controlled trial 141 Honduran EBF = BF wi th no other li quids Weight and length gain 46 1. Nonr andom(first Honduras Quality Assessment infants of low- or so lids unti l 6 months (n=50) and 612 months; WAZ, LAZ, allocat ion.

trial) (P,U) R andomization: C income fam ilies MBF = introduction of com pl e- and WLZ at 6 m onths; r ecei pt 2. Cl uster all ocationFollow-up: C with poor mentary solid food at 4 months of Fe supplements 69 months; by week of birth,Blinding: A for weight sanitation with either ad libitumn ur si ng he mog lo bin and fe rr it in at whi le ana ly se s do neand maternal postpartum (SF) or maintenance of baseline 6 months; % of days with at individual level.weight loss; B for length, nursing frequency (SF-M) fever, cough, nasal congestion, 3. Analysis not baseddevelopmental milestones, (n=91) nasal discharge, hoarseness, on intention-to-treat.and lactational amenorrhea and diarrhea; age first crawled, 4. SF-M and SF groupsJadad Scale age first sat from lying combined as MBFRandomization: 0/2 position, walking by 12 group.Double-blinding: 0/2 months;maternal postpartumWithdrawals: 1/1 weight loss 46 months;Total Jadad scale score: resumptionof menses by

1/5 6 monthsDewey 1999 Design: controlled trial 119 LBW EBF = BF with no other liquids Weight and length gain 46 1. Cluster randomized(second Honduras Quality Assessment Honduran term or so lids unt i l 6 months (n=59) and 612 months; WAZ, LAZ, by week of birth,trial) (P,U) Randomization: B infants MBF = introduction of comple- and WLZ at 6 months; plasma while analyses done

Follow-up: C mentary solid food at 4 months zinc concentration at 6 months; at individual level.Blinding: A for weight, with maintenance of baseline % of days with fever, cough, 2. Analysis not basedB for length nursing frequency (n=60) nasal congestion, nasal on intention-to-treat.Jadad Scale discharge, hoarseness, andRandomization: 1/2 diarrhea; age first crawled,Double blinding: 0/2 age first sat from lying position,Withdrawals: 1/1 walking by 12 months;Total Jadad scale score: maternal postpartum weight

2/5 loss 46 months; resumptionof menses by 6 months

Huf fman 1987 D esi gn : pros pec ti ve co hor t 1 ,018 Bang la - EB F = BF wi th no ot he r l iq ui ds Du ra ti on of la ct at ion al 1 . O ve r 95% of st udy(P,U) Quality Assessment deshi women or solids for 7 months amenorrhea women BF >16

Control for confounding: C with live births (n=647) months, so all MBFFollow-up: B MBF = EBF for 4 months with women assumed toBlinding: A introduction of liquid or solid continue BF 6

supplements before 7 months months.(n=371) 2. Multivariate (Cox)

regression controlledfor maternaleducation, parity,

religion, and weight,but reference groupEBF


17/52


12

Table of included studies: Developed Countries


keson 1996 (P) Desi gn: prospective cohor t 44 heal thy EBF = BF + 125 ml/day of formula for trations at 6 months length gain for theblood analyses, B for 6 months (n=18) two groups at 812length months: misprint?

Duncan 1993 Des ign: p ro spec tive cohort 279 heal thy U .S . EBF = EBF fo r 6 months Number of episodes of otitis(P,U) Quality Assessment infants (n=138) media (OM), one or more

Control for confounding: A MBF = EBF for 4 months with episodes of OM, and frequentFollow-up: B introduction of formula or OM in first 12 monthsBlinding: B solid foods between 4 and 6

months (n=141)Heini g 1993 D esi gn : p ros pec ti ve c ohor t 60 he al thy U .S . EB F = BF 120 ml/day of Mon thl y weigh t and len gth 1 . D ata on weigh t an d(P,U) Quality Assessment infants living in other milk or formula for gain at 69 and 912 length gain 46

Control for confounding: C Davis, CA 12 months and no sol ids mon ths ; t ot al sl eep in g time mon ths inc lu ded inFollow-up: C


18/52

13


Pisacane 1995 Design: pr ospective cohort 30 term , EBF = BF only without any Hemoglobi n and serum(P) Quality Assessment appropr ia te -for- o ther fluids or so lids for ferri tin concent ra tion s a t

Control for confounding: C gestational-age 7 months (n=9) 12 monthsFollow-up: C Italian infants MBF = EBF for 4-6 monthsBlinding: A recruited at 6 with other foods introduced

months and BF before 7 months (n=21)for first year oflife

WHO 1994 (P,U) Design: prospective cohort Pooled sample EBF = BF o the r li qui ds f or Mon thl y weigh t and l eng th Mul ti va ri at e c ont ro lQuality Assessment of healthy 6 months (n=200) gain 46 months for initial weight andControl for confoundi ng: C developed- MBF = BF other liquids for length, but data notFollow-up: C country infants 4 months with other milk presented.B li nd in g: A f or we ight , ( n=358) solids introduced betweenB for length 4 and 6 months (n=158)

WHO 1997 (P,U) Design: prospective cohort Pooled sample EBF = BF n oncal or ic l iq ui ds Mon thl y weigh t an d len gth 1 . Mul ti leve lQuality Assessment of mid- to for 6 months (n=179) gain 48 months regression used toControl for confounding: A high-SES MBF = BF caloric liquids control for maternalFollow-up: C infants from 2 or solids introduced at size and educationBlinding: A for weight, developed and 46 months (n=377) and infant size andB for length 3 developing growth


19/52


14

Results

noncompliance; the latter were restricted to 4 subjects(3 in the exclusive breastfeeding group, 1 in the mixedbreastfeeding group) in the first trial and 3 subjects (all3 in the exclusive breastfeeding group) in the secondtrial. Moreover, the investigators have provided(unpublished) data on weight and length gain on 5 ofthe 9 dropouts in the second Honduran trial (3 of the 9

moved away before 6 months), thereby substantiallyreducing the potential for selection bias in the analysisof that trial.

Most importantly, despite the above-notedmethodologic problems, these two trials are the onlystudies uncovered by our search that used anexperimental design to specifically address the 46months vs about 6 months controversy. Thus, at leastwith respect to bias due to known and unknownconfounding variables, these trials are methodologicallysuperior to any of the observational studies included in

this review despite their methodologic imperfections.Furthermore, the investigators made a considerableeffort to ensure compliance with the assigned allocationand to standardize the training of the observers whoperformed the anthropometric measurements, therebyreducing the random error (improving the precision)of these measurements. Finally, detailed comparisonsbetween trial participants and eligible nonparticipantsdemonstrated no differences that would detract fromthe external validity (generalizability) of the trialsfindings, at least for the specific type of setting wherethe study was conducted (an urban, low-incomepopulation in Honduras).

For all analyses, the two mixed breastfeeding groups (oneof which was intended to maintain frequency ofbreastfeeding) in the first trial were combined for thepurposes of this analysis. Monthly weight gain from 4to 6 months was nonsignificantly slightly higher amonginfants whose mothers were assigned to continuedexclusive breastfeeding [weighted mean difference(WMD) = +20.8 (95% CI -22.0 to +63.5) g/mo](Outcome 1). Thus the 95% CI is statisticallycompatible with a weight gain only 22 g/mo lower inthe EBF group, which represents approximately 5% ofthe mean and 15% of the SD for the monthly weightgain. Weight gain from 6 to 12 months (Outcome 2)

As discussed in the Review Methods, studies werestratified according to study design and provenance fromdeveloping vs developed countries. This resulted inthree separate strata for considering the results of thestudies located by the literature search: (1) controlledtrials of exclusive vs mixed breastfeeding for 46 monthsfrom developing countries, (2) observational studies of

exclusive vs mixed breastfeeding for 37 months fromdeveloping countries and (3) observational studies ofexclusive vs mixed breastfeeding for 37 months fromdeveloped countries.

In accordance with conventional terminology used inCochrane reviews, these strata are labeled below ascomparisons. Outcomes for each comparison arepresented sequentially.

Comparison 1: Controlled trials of

exclusive vs mixed breastfeeding for46 months, developing countries

Two studies were found in this category, both from thesame group of investigators and involving the same studysetting (Honduras). The first of these studies, Cohen1994, involved term infants unselected for birth weightbut included 29 infants (19.9%) weighing


20/52

15

was almost identical in the two groups [WMD = -2.6(-25.9 to +20.6) g/mo].

For length gain from 4 to 6 months, the WMD was +1.0

mm/mo (-0.4 to +2.4 mm/mo); the lower confidencelimit represents only 2% of the mean and 8% of the SDfor monthly length gain (Outcome 3). As with weightgain, length gain from 6 to 12 months (Outcome 4)was nearly identical in the two groups [WMD = -0.4(-1.0 to +0.2) mm/mo].

Weight-for-age, length-for-age, and weight-for-lengthz-scores at 6 months (Outcomes 57) were allnonsignificantly higher in the EBF group [WMD =+0.18 (-0.06 to +0.41), +0.11 (-0.11 to +0.33), and+0.09 (-0.13 to +0.31), respectively.

The impact of the small sample size of the two Hondurantrials is evident when examining the risk ofundernutrition, as represented by anthropometricz-scores


21/52


16

adjusted means for weight and length gain from 4 to 6months. Nonetheless, the inability of observationalstudies to control for subtle (and unknown) sources of

confounding and selection bias suggests the need forcautious interpretation. All three studies reported onmonthly weight gain from 4 to 6 months (Outcome 1).The WMD was -7.2 (-25.5 to +11.0) g/mo, a lowerconfidence limit compatible with a deficit of only 7%of the mean and


22/52

17

(Outcome 3); the WMD was -1.8 (-16.7 to +13.1) g/mo, which excludes a reduced length gain in the EBFgroup of 5% of the mean and 10% of the SD for the

Belarussian study.For length gain at 38 months (Outcome 4), the studiesagain show significant (P


23/52


18

109 g/L (95% CI for the difference = +4.0 to +12.0 g/L)] (Outcome 42), a nonsignificant reduction in anemia(hemoglobin


24/52


25/52


20

for 6 or more months had no observable deficits ingrowth, and their mothers were more likely to remainamenorrheic for 6 months postpartum. No benefits of

introducing complementary foods between 4 and 6months have been demonstrated, with the exceptionof improved iron status in one developing countrysetting (Honduras). Since the latter benefit can beachieved more effectively by medicinal ironsupplementation (e.g., vitamin drops), it does not appearto justify incurring the adverse effects of liquid or solidfood supplementation on infectious morbidity, andlactational amenorrhea.

Thus, with the caveat that individual infants must stillbe managed individually, so that insufficient growth orother adverse outcomes are not ignored and appropriate

interventions are provided, the available evidencedemonstrated no apparent risks in recommending, as ageneral policy, exclusive breastfeeding for the first 6months of life in both developing and developed countrysettings. Large, rigorous cluster-randomized trials shouldhelp resolve residual uncertainties about the possibleadvantages and disadvantages of such a policy.


26/52

21

References

14. Cunningham A, Jelliffe D, Jelliffe E. Breast-feedingand health in the 1980s: a global epidemiologicreview.J Pediatr 1991; 118:659666.

15. Beaudry M, Dufour R, Marcoux S. Relation betweeninfant feeding and infections during the first sixmonths of life.J Pediatr 1995; 126:191197.

16. Dewey K, Heinig M, Nommsen-Rivers L.Differences in morbidity between breast-fed andformula fed infants.J Pediatr 1995; 126:696702.

17. Raisler J, Alexander C, OCampo P. Breast-feedingand infant illness: a dose-response relationship?AmJ Public Health 1999; 89:2530.

18. Ford R, Taylor B, Mitchell E, Enright S, Stewart A,Becroft D, et al. Breastfeeding and the risk of suddeninfant death syndrome. Int J Epidemiol 1993; 22:885890.

19. Saarinen U, Backman A, Kajosaari M, Simes M.Prolonged breast-feeding as prophylaxis for atopicdisease. Lancet 1979; ii:163166.

20. Hide D, Guyer B. Clinical manifestations of allergyrelated to breast and cows milk feeding. Arch DisChild 1981; 56:172175.

21. Oddy W, Holt P, Sly P, Read A, Landau L, StanleyF, et al. Association between breast feeding andasthma in 6 year old children: findings of aprospective birth cohort study. BMJ 1999; 319:815819.

22. Lucas A, Morley R, Cole T, et al. Breast milk andsubsequent intelligence quotient in children bornpreterm. Lancet 1992; 339:261264.

23. Lanting C, Fidler V, Huisman M, et al. Neurologicaldifferences between 9-year old children fed breast-milk or formula-milk as babies. Lancet 1994;344:13191322.

24. Horwood L, Fergusson D. Breastfeeding and latercognitive and academic outcomes. Pediatrics 1998;101:e9.

25. Anderson J, Johnstone B, Remley D. Breast-feedingand cognitive development: a meta-analysis.Am JClin Nutr 1999; 70:525535.

R E F E R E N C E S

1. Waterlow J, Thomson A. Observations on theadequacy of breast-feeding. Lancet 1979; ii:238241.

2. Whitehead R, Paul A. Growth charts and theassessment of infant feeding practices in the westernworld and in developing countries. Early Hum Dev1984; 9:187207.

3. FAO/WHO. Energy and protein requirements.1973; Rome: FAO. 52: FAO Nutrition meetings.

4. Whitehead R, Paul A. Infant growth and humanmilk requirements. Lancet 1981; 161163.

5. Garza C, Butte N. Energy intakes of human milk-fed infants during the first year. J Pediatr 1990;117:S124S131.

6. Butte N. Energy requirements of infants. Eur J ClinNutr 1996; (suppl) 50:2436.

7. Brown K, Dewey K, Allen L. Complementary Feeding

of Young Children in Developing Countries: A Reviewof Current Scientific Knowledge. Geneva: WHO,1998.

8. World Health Organization. Energy and proteinrequirements. 1985; Report of a Joint FAO/WHO/UNU Expert Consultation ed., Geneva: TechnicalReport Series, 724.

9. Rowland M, Barrell R, Whitehead R. Bacterialcontamination in traditional Gambian weaningfoods. Lancet 1978; i:136138.

10. Rowland M. The weanlings dilemma: Are wemaking progress?Acta Paediatr Scand 1986; (suppl)323:3342.

11. WHO Collaborative Study Team on the Role ofBreastfeeding on the Prevention of Infant Mortality.Effect of breastfeeding on infant and child mortalitydue to infectious diseases in less developed countries:a pooled analysis. Lancet 2001; 355:451455.

12. Anandaiah R, Choe M.Are the WHO guidelines onbreastfeeding appropriate for India? National FamilyHealth Survey Subject Reports, Number 16 2000.

13. Howie P, Forsyth J, Ogston S, Clark A, du V FloreyC. Protective effect of breast feeding againstinfection. Br Med J 1990; 300:1116.


27/52


22

26. Vestergaard M, Obel C, Henriksen T, Srensen H,Skajaa E, stergaard J. Duration of breastfeedingand developmental milestones during the latter half

of infancy.Acta Paediatr 1999; 88:13271332.27. San Giovanni J, Berkey C, Dwyer J, Colditz G.

Dietary essential fatty acids, long chain polyun-saturated fatty acids, and visual resolution acuity inhealthy fullterm infants: a systematic review. EarlyHum Dev 2000; 57:165188.

28. San Giovanni J, Parra-Cabrera S, Colditz G, BerkeyC, Dwyer J. Meta-analysis of dietary essential fattyacids and long chain polyunsaturated fatty acids asthey relate to visual resolution acuity in healthypreterm infants. Pediatrics 2000; 105:12921298.

29. Kramer M. Do breast-feeding and delayedintroduction of solid foods protect against subse-quent obesity?J Pediatr 1981; 98:883887.

30. von Kries R, Koletzko B, Sauerwald T, von MutiusE, Barnert D, Grunert V, et al. Breast feeding andobesity: cross sectional study. BMJ 1999; 319:147150.

31. Gillman M, Rifas-Shiman S, Camargo C, BerkeyC, Frazier A, Rockett H, et al. Risk of overweightamong adolescents who were breastfed as infants.

JAMA 2001; 285:24612467.32. Mayer E, Hamman R, Gay E, et al. Reduced risk of

IDDM among breastfed children. Diabetes 1988;37:16251632.

33. Gerstein H. Cows-milk exposure and type-I diabetesmellitus: a critical review of the clinical literature.Diabetes Care 1994; 17:1319.

34. Koletzko S, Sherman P, Corey M, et al. Role ofinfant feeding practices in development of Crohnsdisease in childhood. Br Med J 1989; 298:1617

1618.35. Davis M, Savitz D, Graubard B. Infant feeding and

childhood cancer. Lancet 1988; 2:365368.

36. Davis M. Review of the evidence for an associationbetween infant feeding and childhood cancer. Int JCancer 1998; (suppl) 11:2933.

37. Chilvers C. Breastfeeding and risk of breast cancerin young women. Br Med J 1993; 307:1720.

38. Brinton L, Potischman N, Swanson C, et al.Breastfeeding and breast cancer risk. Cancer Causes

and Control 1995; 6:199208.

39. Enger S, Ross R, Henderson B, Bernstein L.Breastfeeding history pregnancy experience and riskof breast cancer. Br J Cancer 1997; 76:118123.

40. Rosenblatt K, Thomas D. Lactation and the risk ofepithelial ovarian cancer. Int J Epidemiol 1993;22:192197.

41. Alderman B, Weiss N, Daling J, et al. Reproductivehistory and postmenopausal risk of hip and forearmfracture.Am J Epidemiol 1986; 124:262267.

42. Melton L, Bryant S, Wahner H. Influence ofbreastfeeding and other reproductive factors in bonemass later in life. Osteoporosis Int 1993; 3:7683.

43. Cummings R, Klineberg R. Breastfeeding and other

reproductive factors and the risk of hip fractures inelderly women. Int J Epidemiol 1993; 22:684691.

44. WHO Working Group on Infant Growth. Anevaluation of infant growth. Geneva: World HealthOrganization, 1994.

45. Dewey K, Peerson J, Brown K, Krebs N, MichaelsenK, Persson L, et al. Growth of breast-fed infantsdeviates from current reference data: a pooledanalysis of US, Canadian, and European data sets.Pediatrics 1995; 96:495503.

46. Nielsen G, Thomsen B, Michaelsen K. Influenceof breastfeeding and complementary food on growthbetween 5 and 10 months. Acta Paediatr 1998;87:911917.

47. Hediger M, Overpeck M, Ruan W, Troendle J. Earlyinfant feeding and growth status of US-born infantsand children aged 471 mo: analyses from the thirdNational Health and Nutrition ExaminationSurvey, 19881994.Am J Clin Nutr 2000; 72:159167.

48. Haschke F, vant Hof M, Euro-growth study groups.

Euro-Growth references for breast-fed boys and girls:influence of breast-feeding and solids on growthuntil 36 months of age. J Pediatr Gastroenterol Nutr2000; 31:S60S71.

49. de Onis M, Garza C, Habicht J. Time for a newgrowth reference. Pediatrics 1997; 100:E8.

50. WHO Working Group on the Growth ReferenceProtocol and WHO Task Force on Methods for theNatural Regulation of Fertility. Growth patterns ofbreastfed infants in seven countries. Acta Paediatr

2000; 89:215222.51. Kramer M, Chalmers B, Hodnett E, Sevkovskaya

Z, Dzikovich I, Shapiro S, et al. Promotion of


28/52

23

R E F E R E N C E S

breastfeeding intervention trial (PROBIT): Arandomized trial in the Republic of Belarus.JAMA2001; 285:413420.

52. Kramer M, Chalmers B, Hodnett E. Breastfeedingand infant growth: biology or bias? Pediatr Res 2000;47:151A.

53. Kramer MS, Chalmers B, Hodnett ED, et al.Promotion of Breastfeeding Intervention Trial(PROBIT): a cluster-randomized trial in theRepublic of Belarus. In: Koletzko, Michaelsen KF,Hernell O, editors. Short and Long Term Effects ofBreast Feeding on Child Health. New York: KluwerAcademic/Plenum Publishers, 2000, pp. 327345.

54. World Health Organization. Nutrition: Informationand attitudes among health personnel about earlyinfant-feeding practices. WHO Wkly Epidem 1995;70:117120.

55. United Nations Childrens Fund. Facts for Life.Wallingford: P&LA, 1993.

56. Lutter C. Length of exclusive breastfeeding: linkingbiology and scientific evidence to a public healthrecommendation. J Nutr 2000; 130:13351338.

57. Committee on Nutrition, American Academy ofPediatrics. Pediatric Nutrition Handbook. 4th ed. ElkGrove Village, Illinois: American Academy ofPediatrics, 1998.

58. Hill A.A Short Textbook of Medical Statistics. London:Hodder & Stoughton, 1977, p.27.

59. Sauls H. Potential effect of demographic and othervariables in studies comparing morbidity of breast-fed and bottle-fed infants. Pediatrics 1979; 64:523527.

60. Miettinen O. The need for randomization in thestudy of intended effects. Stat Med 1983; 2:267271.

61. Bauchner H, Leventhal J, Shapiro E. Studies ofbreast-feeding and infections: How good is theevidence?JAMA 1986; 256:887892.

62. Heinig M, Nommsen L, Peerson J, Lonnerdal B,Dewey K. Intake and growth of breast-fed andformula-fed infants in relation to the timing of intro-duction of complementary foods: the DARLINGstudy.Acta Paediatr Scand 1993; 82:9991006.

63. Cohen R, Brown K, Canahuati J, Rivera L, DeweyK. Effects of age of introduction of complementary

foods on infant breast milk intake, total energyintake, and growth: a randomized intervention studyin Honduras. Lancet 1994; 344:288293.

64. Dewey K, Cohen R, Brown K, Rivera L. Age ofintroduction of complementary foods and growthof term, low-birthweight, breast-fed infants: a

randomized intervention study in Honduras.Am JClin Nutr 1999; 69:679686.

65. Frongillo EJr, Habicht J-P. Investigating theweanlings dilemma: lessons from Honduras.Nutrition Reviews 1997; 55:390395.

66. Indicators for Assessing Breast-Feeding Practices.Geneva, Switzerland: World Health Organization.WHO Document WHO/CDD/SER 1991; 91:14.

67. keson P, Axelsson I, Raiha N. Growth andnutrient intake in three-to-twelve-month-oldinfants fed human milk or formulas with varyingprotein concentrations. J Pediatr Gastroenterol Nutr1998; 26:18.

68. Castillo C, Atalah E, Riumallo J, Castro R. Breast-feeding and the nutritional status of nursing childrenin Chile. Bull PAHO 1996; 30:125133.

69. Cohen R, Brown K, Canahuati J, Rivera L, DeweyK. Determinants of growth from birth to 12 monthsamong breast-fed Honduran infants in relation toage of introduction of complementary foods.Pediatrics 1995; 96:504510.

70. Dewey K, Cohen R, Rivera L, Brown K. Effects ofage of introduction of complementary foods onmicronutrient status of term, low-birthweight,breastfed infants in Honduras. FASEB J 1998;12:A648.

71. Dewey K, Cohen R, Rivera L, Brown K. Effects ofage introduction of complementary foods on ironstatus of breast-fed infants in Honduras.Am J ClinNutr 1998; 67:878884.

72. Dewey K, Cohen R, Rivera L, Canahuati J, Brown

K. Do exclusively breast-fed infants require extraprotein? Pediatr Res 1996; 39:303307.

73. keson P, Axelsson I, Raiha N. Protein and aminoacid metabolism in three-to-twelve-month-oldinfants fed human milk or formulas with varyingprotein concentrations. J Pediatr Gastroenterol Nutr1998; 26:297304.

74. Duncan B, Ey J, Holberg C, Wright A, Martinez F,Taussig L. Exclusive breast-feeding for at least 4months protects against otitis media. Pediatrics 1993;91:867872.

75. Kajosaari M, Saarinen U. Prophylaxis of atopicdisease by six months total solid food elimination.Evaluation of 135 exclusively breast-fed infants of


29/52


24

atopic families.Acta Paediatr Scand 1983; 72:411414.

76. Kajosaari M. Atopy prevention in childhood: the

role of diet: a prospective 5-year follow-up of high-risk infants with six months exclusive breastfeedingand solid food elimination. Pediatr Allerg Immunol1994; 5:2628.

77. Kajosaari M. Atopy prophylaxis in high-risk infants.Prospective 5-year follow-up study of children withsix months exclusive breastfeeding and solid foodelimination.Adv Exp Med Biol 1991; 310:453458.

78. Pisacane A, de Vizia B, Valiante A, Vaccaro F, RussoM, Grillo G, et al. Iron status in breast-fed infants.J Pediatr 1995; 127:429431.

79. Adair L, Popkin B, Vanderslice J, Akin J, GuilkeyD, Black R, et al. Growth dynamics during the firsttwo years of life: a prospective study in tehPhillipines. Eur J Clin Nutr 1993; 47:4251.

80. Simondon K, Simondon F. Age at introduction ofcomplementary food and physical growth from 2 to9 months in rural Senegal. Eur J Clin Nutr 1997;51:703707.

81. keson P, Axelsson I, Raiha N. Human milk andstandard infant formula together with high qualitysupplementary foods is sufficient for normal growthduring infancy. Pediatr Res 1996; 39 (Suppl):313A.

82. Frongillo EJ, de Onis M, Garza C. The World HealthOrganization Task Force on Methods for the NaturalRegulation of Fertility. Effects of timing ofcomplementary foods on post-natal growth. FASEBJournal 1997; 11:A574.

83. WHO Working Group on the Growth ReferenceProtocol and the WHO Task Force on Methods forthe Natural Regulation of fertility. Growth of

healthy infants and the timing, type and frequencyof complementary foods.Am J Clin Nutr 2001. (Inpress).

84. Huffman S, Ford K, Allen H, Streble P. Nutritionand fertility in Bangladesh: breastfeeding and postpartum amenorrhoea. Population Studies 1987;41:447462.

85. Dewey K, Cohen R, Brown K, Rivera L. Effects ofexclusive breastfeeding for four versus six monthson maternal nutritional status and infant motordevelopment: results of two randomized trials in

Honduras. J Nutr 2001; 131:262267.86. Brown K. The relationship between diarrhoeal

prevalence and growth of poor infants varies withtheir age and usual energy intake (abstract). FASEBJ 1991; 5:A1079.

87. Dewey K, Cohen R, Rivera L, Canahuati J, BrownK. Effects of age at introduction of complementaryfoods to breast-fed infants on duration of lactationalamenorrhea in Honduran women.Am J Clin Nutr1997; 65:14031409.

88. WHO Working Group on Infant Growth. Anevaluation of infant growth: the use and interpre-tation of anthropometry in infants. Bull WHO 1995;73:165174.

89. Jadad A, Moore R, Carroll D, Jenkinson C,Reynolds J, Gavaghan D, et al. Assessing the qualityof reports of randomized clinical trials: is blindingnecessary? Controlled Clinical Trials 1996; 17:112.


2 4


30/52

25

A N N E X 1 . C O M P A R I S O N 0 1

ANNEX 1

Comparison 01: Exclusive vs mixed breastfeeding

46 months, developing countries, controlled trials


31/52


26


32/52

27



33/52


28


34/52

29



35/52


30


36/52

31



37/52


32

ANNEX 2


37 months, developing countries, observational studies


38/52

33



39/52


34


40/52

35



41/52


36

ANNEX 3


37 months, developed countries, observational studies


42/52

37



43/52


38


44/52

39



45/52


40


46/52

41



47/52


42


48/52

43



49/52


44


50/52

45



51/52


46


52/52


Optimal Duration of Exc Bfeeding Review Eng

Documents

Transcript of Optimal Duration of Exc Bfeeding Review Eng