3.2.pdf - OSSyR - Observatorio de Salud Sexual y Reproductiva

DOI: 10.1542/peds.2004-1441 2005;115;519-617 Pediatrics

Zulfiqar A. Bhutta, Gary L. Darmstadt, Babar S. Hasan and Rachel A. Haws Outcomes in Developing Countries: A Review of the Evidence

Community-Based Interventions for Improving Perinatal and Neonatal Health

http://www.pediatrics.org/cgi/content/full/115/2/S1/519located on the World Wide Web at:

The online version of this article, along with updated information and services, is

rights reserved. Print ISSN: 0031-4005. Online ISSN: 1098-4275. Grove Village, Illinois, 60007. Copyright © 2005 by the American Academy of Pediatrics. All and trademarked by the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elkpublication, it has been published continuously since 1948. PEDIATRICS is owned, published, PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly

by on April 13, 2010 www.pediatrics.orgDownloaded from

http://www.pediatrics.org/cgi/content/full/115/2/S1/519

http://www.pediatrics.org

Community-Based Interventions for Improving Perinatal andNeonatal Health Outcomes in Developing Countries:

A Review of the Evidence

Zulfiqar A. Bhutta, MBBS, FRCPCH, PhD*; Gary L. Darmstadt, MD, MS, FAAP‡§; Babar S. Hasan, MD*;and Rachel A. Haws, MHS§

ABSTRACT. Background. Infant and under-5 child-hood mortality rates in developing countries have de-clined significantly in the past 2 to 3 decades. However,2 critical indicators, maternal and newborn mortality,have hardly changed. World leaders at the United Na-tions Millennium Summit in September 2000 agreed on acritical goal to reduce deaths of children <5 years bytwo thirds, but this may be unattainable without halv-ing newborn deaths, which now comprise 40% of allunder-5 deaths. Greater emphasis on wide-scale imple-mentation of proven, cost-effective measures is requiredto save women’s and newborns’ lives. Approximately99% of neonatal deaths take place in developing coun-tries, mostly in homes and communities. A comprehen-sive review of the evidence base for impact of interven-tions on neonatal health and survival in developing-country communities has not been reported.

Objective. This review of community-based antena-tal, intrapartum, and postnatal intervention trials in de-veloping countries aimed to identify (1) key behaviorsand interventions for which the weight of evidence issufficient to recommend their inclusion in community-based neonatal care programs and (2) key gaps in knowl-edge and priority areas for future research and programlearning.

Methods. Available published and unpublished dataon the impact of community-based strategies and inter-ventions on perinatal and neonatal health status out-comes were reviewed. Evidence was summarized system-atically and categorized into 4 levels of evidence basedon study size, location, design, and reported impact, par-ticularly on perinatal or neonatal mortality. The evidencewas placed in the context of biological plausibility of theintervention; evidence from relevant developed-countrystudies; health care program experience in implementa-tion; and recommendations from the World Health Or-ganization and other leading agencies.

Results. A paucity of community-based data wasfound from developing-country studies on health statusimpact for many interventions currently being consid-ered for inclusion in neonatal health programs. However,review of the evidence and consideration of the broadercontext of knowledge, experience, and recommendationsregarding these interventions enabled us to categorizethem according to the strength of the evidence base andconfidence regarding their inclusion now in programs.This article identifies a package of priority interventionsto include in programs and formulates research prioritiesfor advancing the state of the art in neonatal health care.

Conclusions. This review emphasizes some new find-ings while recommending an integrated approach to safemotherhood and newborn health. The results of thisstudy provide a foundation for policies and programsrelated to maternal and newborn health and emphasizesthe importance of health systems research and evaluationof interventions. The review offers compelling supportfor using research to identify the most effective measuresto save newborn lives. It also may facilitate dialogue withpolicy makers about the importance of investing in neo-natal health. Pediatrics 2005;115:519–617.

ABBREVIATIONS. ARI, acute respiratory infection; CCS, case-control study; CHW, community health worker; CI, confidenceinterval; CKMC, community-based application of kangaroomother care; CQ, chloroquine; DBRCT, double-blind, randomized,controlled trial; DBRPCT, double-blind, randomized, placebo-controlled trial; EFA, essential fatty acid; EPI, Expanded Pro-gramme on Immunization; FHW, family health worker; Hb, he-moglobin; HBeAg, hepatitis B virus “e” antigen; HBsAg, hepatitisB surface antigen; HBV, hepatitis B virus; HDN, hemorrhagicdisease of the newborn; IM, intramuscular; IMR, infant mortalityrate; IPT, intermittent presumptive treatment; ITN, insecticide-treated bed net; IUGR, intrauterine growth restriction; IV, intra-venous; IVH, intraventricular hemorrhage; KMC, kangaroomother care; LBW, low birth weight; NIB, untreated bed net; NIH,National Institutes of Health; NMR, neonatal mortality rate; NTD,neural tube defect; OR, odds ratio; PCS, prospective cohort study;PMR, perinatal mortality rate; PROG, proguanil; PPROM, pretermpremature rupture of membranes; PROM, premature rupture ofmembranes; QT, quasi-experimental trial; RCS, retrospective co-hort study; RCT, randomized, controlled trial; RDA, recom-mended dietary allowance; RPCT, randomized, placebo-con-trolled trial; RPR, rapid plasma reagin; RR, relative risk; SEARCH,Society for Education, Action and Research in Community Health;SGA, small for gestational age; SP, sulfadoxine-pyrimethamine;STD, sexually transmitted disease; TBA, traditional birth atten-dant; TEWL, transepidermal water loss; TT, tetanus toxoid;UNICEF, United Nations Children’s Fund; UTI, urinary tract in-fection; VLBW, very low birth weight; WHO, World Health Or-ganization; WIC, Women, Infants, and Children SupplementalNutrition Program; VHW, village health worker.

From the *Department of Pediatrics and Child Health, Aga Khan Univer-sity, Karachi, Pakistan; ‡Saving Newborn Lives Initiative, Office of Health,Save the Children/USA, Washington, DC; and §Department of Interna-tional Health, Bloomberg School of Public Health, Johns Hopkins Univer-sity, Baltimore, MD.Accepted for publication Aug 10, 2004.doi:10.1542/peds.2004-1441No conflict of interest declared.Reprint requests to (Z.A.B.) Department of Pediatrics and Child Health, AgaKhan University, Stadium Road, Karachi 74800, Pakistan. E-mail:[email protected]; or Gary L. Darmstadt, MD, Department of Inter-national Health, E8153, Bloomberg School of Public Health, Johns HopkinsUniversity, 615 N Wolfe St, Baltimore, MD, 21205. E-mail: [email protected] (ISSN 0031 4005). Copyright © 2005 by the American Acad-emy of Pediatrics.

PEDIATRICS Vol. 115 No. 2 February 2005 519 by on April 13, 2010 www.pediatrics.orgDownloaded from


EXECUTIVE SUMMARY

Background

Although there has been considerable im-provement in child health globally, it is in-creasingly evident that important gaps and

disparities remain. In particular, it is apparent that adisproportionate burden of infant and under-5 child-hood mortality relates to deaths within the neonatalperiod, which frequently occur within the first fewdays after birth. Moreover, the vast majority of peri-natal and neonatal deaths occur in conditions ofsocioeconomic deprivation in developing countries.As the health of the newborn infant is inexorably tiedto the health of the mother, strategies to improve thehealth and care of women in low-resource commu-nities and countries are also expected to improveboth pregnancy and neonatal health outcomes. How-ever, although it is true that poverty, illiteracy, poorstatus and care of women, as well as dysfunctionalhealth systems are critical underlying factors thatadversely affect maternal and child health in manydeveloping countries, these factors are relatively dif-ficult to change in the short term. Moreover, in sub-Saharan Africa, the devastating epidemic of humanimmunodeficiency virus (HIV)/acquired immuno-deficiency syndrome (AIDS) threatens to reversemany of the gains achieved during decades of childhealth programming. This review does not attemptto evaluate the benefits of investing in social devel-opment, reducing inequity, and promoting economicgrowth among impoverished populations of devel-oping countries. Although these are important long-term goals, pragmatic reality in most developingcountries dictates the need for wide-scale implemen-tation now of evidence-based, cost-effective healthprograms and interventions to improve child healthoutcomes. Moreover, achievement of MillenniumDevelopment Goal 4 to reduce under-5 mortality bytwo thirds by the year 2015 is critically dependent ona substantial reduction in neonatal mortality over thenext decade.

The objectives of this review of community-basedantenatal, intrapartum and postnatal interventiontrials in developing countries were to (1) identify keybehaviors and interventions for which the weight ofevidence is sufficient to recommend their inclusionin community-based neonatal care programs and(2) identify key gaps in knowledge and priority areasfor future research and program learning.

MethodsCurrent practice of summarizing evidence for the

impact of interventions through meta-analyses ofrandomized, controlled trials (RCTs), although ofhigh scientific validity, has more limited relevancewhen applied to research in developing countries,because most studies meeting the rigorous criteriafor inclusion in such analyses were conducted indeveloped-country settings. In addition, the evi-dence base made up of interventions deriving fromeffectiveness trials in health system settings is scanty.In this review, we evaluated the available evidencein the global literature for the benefits and impact of

various community-based interventions during theantenatal, intrapartum, and postnatal periods onperinatal and neonatal health status outcomes. Theselection of interventions was based on biologicalplausibility and inclusion as a component in pro-grams of maternal and/or perinatal health care. Wedid not, however, review the evidence for impactof skilled birth attendants, because this is the sub-ject of other reviews.1–4 The scientific evidence avail-able from individual interventions or combinationsthereof was reviewed, and information from pro-grams and effectiveness trials that used packages ofinterventions was specifically solicited and analyzed.Although our principal focus was to seek informa-tion from community-based RCTs, we extended theevaluation to include studies with a variety of other,less rigorous designs. A few studies with a quasi-experimental trial (QT) design were included, espe-cially if they represented information from develop-ing countries and pertained to an intervention withlittle other evidence base. Community was definedas extending from the household to the peripheralhealth facility level; in general, studies at secondaryor tertiary referral-level health facilities were ex-cluded. However, where evidence for key interven-tions from community-based settings was sparse ornot available, information was included from facili-ty-based settings in developing countries and occa-sionally from the developed world. The evidencefrom developed-country studies, however, was usedprimarily to provide perspective and context for con-clusions drawn from developing-country data. Evi-dence from the Cochrane database of RCTs and theWorld Health Organization (WHO) ReproductiveHealth Library was also considered, and studiesfrom developing countries that were included in theCochrane Reference Library were specifically evalu-ated. We also complemented the review of the sci-entific evidence based on intervention trials in devel-oping countries with an assessment of public healthprograms and interventions currently in place; rec-ommendations from the WHO and other expert in-stitutions and individuals; and biological plausibilityand evidence from developed-country studies.

Sources for potentially eligible studies includedjournal articles, book chapters, technical reports, con-ference proceedings, and theses. The search for com-munity-based evidence encompassed all availableelectronic health and social science reference libraries(including indexed and nonindexed journals), andmanual reviews of Safe Motherhood and Child Sur-vival books and technical reports. Additional detailswere solicited directly from most agencies and insti-tutions involved in community-based care in devel-oping countries, especially Reproductive Health,Safe Motherhood, or Child Survival programs. Mostleading global public health researchers in the fieldof perinatal and maternal care were also individuallyapproached for information and unpublished mate-rial. It is important to underscore that, although wespecifically sought evidence from RCTs, we werecognizant of the danger of relying on RCTs as thesole source of evidence for interventions,5 especiallyin terms of consistency6 and external validity.7 This

520 COMMUNITY-BASED INTERVENTIONS FOR IMPROVING HEALTH OUTCOMES by on April 13, 2010 www.pediatrics.orgDownloaded from


TA

BL

E1.

Num

bers

ofC

omm

unit

y-B

ased

Tri

als

ofIn

terv

enti

ons

toIm

prov

ePe

rina

tal/

Neo

nata

lH

ealt

hin

Dev

elop

ing

Cou

ntri

es

Inte

rven

tion

Num

ber

ofD

evel

opin

g-C

ount

rySt

udie

sR

evie

wed

In-D

epth

Num

ber

ofC

omm

unity

-Bas

edSt

udie

sR

epor

ting

Prim

ary

Peri

nata

l/N

eona

talH

ealth

Stat

usO

utco

mes

*

Num

ber

ofC

omm

unity

-Bas

edR

CTs

Rep

ortin

gPr

imar

yPe

rina

tal/

Neo

nata

lHea

lthSt

atus

Out

com

es*

Num

ber

ofC

omm

unity

-Bas

edSt

udie

sR

epor

ting

Seco

ndar

yPe

rina

tal/

Neo

nata

lHea

lthSt

atus

Out

com

es†

Num

ber

ofC

omm

unity

-Bas

edR

CTs

Rep

ortin

gSe

cond

ary

Peri

nata

l/N

eona

talH

ealth

Stat

usO

utco

mes

†

Num

ber

ofH

ealth

Syst

emIn

terv

entio

nsor

Effe

ctiv

enes

sTr

ials

Mat

erna

lsc

hool

ing/

heal

thed

ucat

ion

00

00

00

Ant

enat

alca

repa

ckag

e(s)

31

12

21

Prot

ein

supp

lem

enta

tion

00

00

00

Bal

ance

dpr

otei

n-en

ergy

supp

lem

enta

tion

123

29

50

Ant

enat

alir

onsu

pple

men

tati

on5

11

43

0Pe

rico

ncep

tion

alfo

late

supp

lem

enta

tion

10

00

00

Ant

enat

alfo

late

supp

lem

enta

tion

10

01

01

Ant

enat

alio

din

esu

pple

men

tati

on3

21

20

1A

nten

atal

vita

min

Asu

pple

men

tati

on6

(1)

11

11

0A

nten

atal

zinc

supp

lem

enta

tion

60

04

40

Mul

tipl

em

icro

nutr

ient

supp

lem

enta

tion

5(3

)2

24

30

Mal

aria

chem

opro

phyl

axis

orIP

T16

65

1411

2M

alar

iapr

even

tion

usin

gIT

Ns

42

24

41

Dew

orm

ing

4(1

)2

02

00

Syph

ilis

scre

enin

gan

dtr

eatm

ent

51

02

00

Ant

ibio

tics

for

UT

Isan

dse

xual

lytr

ansm

itte

dd

isea

ses

(ST

Ds)

21

11

10

Ant

ibio

tics

for

asym

ptom

atic

bact

eriu

ria

00

00

00

Ant

ibio

tics

for

bact

eria

lva

gino

sis

10

00

00

Ant

ibio

tics

for

pret

erm

labo

r3

00

11

0A

ntib

ioti

csfo

rPP

RO

M4

00

00

0T

Tim

mun

izat

ion

85‡

3‡0

01‡

Cle

and

eliv

ery

prac

tice

s5

(3)‡

5(3

)§2

(2)§

00

1M

ater

nal

pneu

moc

occa

lim

mun

izat

ion

10

00

00

Prom

otio

nof

smok

ing

cess

atio

nin

preg

nanc

y�0

00

00

0M

ater

nal

care

pack

ages

21

10

00

Mat

erna

lva

gina

lan

dne

wbo

rnsk

inan

tise

psis

20

00

00

New

born

resu

scit

atio

n13

81

20

0D

elay

edum

bilic

alco

rdcl

ampi

ng1

00

00

0U

mbi

lical

cord

anti

seps

is4

4§3§

00

0Pr

even

tion

and

man

agem

ent

ofne

onat

alhy

poth

erm

ia4

00

31

0Pr

even

tion

and

man

agem

ent

ofhy

pogl

ycem

ia2

00

00

0B

reas

tfee

din

g13

30

33

0Pr

even

tion

ofop

htha

lmia

neon

ator

um2

00

00

0V

itam

inK

prop

hyla

xis

00

00

00

Hep

atit

isB

vacc

inat

ion

70

07

00

Neo

nata

lvi

tam

inA

supp

lem

enta

tion

32

20

00

KM

C9

10

00

0T

opic

alem

ollie

ntth

erap

y2

00

00

0H

yper

bilir

ubin

emia

scre

enin

g5

00

00

0T

BA

/C

HW

trai

ning

14(7

)12

(7)

4(1

)3

00

Pneu

mon

iaca

sem

anag

emen

t4

21

20

1N

eona

tal

care

pack

ages

10(4

)8

(4)

4(2

)2

10

Car

ein

peri

pher

alhe

alth

faci

litie

s9

30

00

1C

ost-

effe

ctiv

enes

sst

udie

s6

(2)

4(2

)2

(2)

10

0T

otal

s18

664

3174

4010

*Pr

imar

ype

rina

tal/

neon

atal

outc

omes

incl

uded

inth

isco

unt

wer

est

illbi

rth

rate

,PM

R,a

ndN

MR

.†

Seco

ndar

ype

rina

tal/

neon

atal

outc

omes

incl

uded

inth

isco

untw

ere

pret

erm

birt

hs,L

BW

and

VL

BW

birt

hs,b

irth

leng

th,h

ead

/ch

estc

ircu

mfe

renc

e,SG

A/

larg

e-fo

r-ge

stat

iona

l-ag

ebi

rths

,neo

nata

lw

eigh

tga

in,

dia

rrhe

alin

fect

ions

,re

spir

ator

yin

fect

ions

,co

ngen

ital

syph

ilis,

opht

halm

iane

onat

orum

,N

TD

s,H

IVin

fect

ion,

seps

is/

infe

ctio

ns,

anem

ia,

impe

tigo

,H

BV

carr

ier

rate

,fe

tal/

neon

atal

mal

aria

,hyp

othe

rmia

,hyp

ogly

cem

ia,a

ndbr

east

feed

ing

rate

s.‡

The

num

bers

ofst

udie

sin

pare

nthe

ses

are

repe

ated

inea

rlie

rta

bles

and

are

not

cons

ider

edin

the

tota

ls.

§T

etan

usin

fect

ion

rate

was

used

asa

prox

you

tcom

efo

rte

tanu

s-at

trib

utab

lene

onat

alca

sefa

talit

y.�

Eig

htsm

okin

g-ce

ssat

ion

stud

ies

wer

ere

view

edbu

tno

tco

unte

din

the

stud

yto

tal

beca

use

they

wer

eal

lfr

omd

evel

oped

coun

trie

s.

SUPPLEMENT 521 by on April 13, 2010 www.pediatrics.orgDownloaded from


is especially true for those interventions that must benested within health systems.8

The principal reviewers independently evaluatedthe data, and a common reporting matrix was usedin summarizing the findings. Studies were evaluatedfor size, setting, quality, and design, ie, either effi-cacy or effectiveness trials.9 The final categorizationand assessment of evidence for impact of the inter-ventions was made by mutual agreement and con-sensus. Emphasis was placed on assessment of im-pact on perinatal or neonatal primary health statusoutcomes. However, for some interventions forwhich data on primary health status outcomes werelacking, other indicators were considered.

The evidence from various interventions was cat-egorized as follows:

1. No evidence of benefit: These interventions hadbeen evaluated and found to have no demonstra-ble benefit either singly or in combination withother measures. In some cases, there was evidenceof an adverse effect of the intervention. Therefore,these interventions were not recommended forinclusion in neonatal health care strategies.

2. Uncertain evidence of benefit: This category in-cluded interventions for which there was someevidence of benefit, but contradictory evidence orissues such as study design, location, or size pre-cluded any firm conclusions. These interventionsmerited additional evaluation or research in de-veloping-country community settings using well-designed protocols.

3. Some evidence of benefit: These interventionshad some positive impact on perinatal and/orneonatal outcomes, but the evidence remainedpreliminary or the location of studies was notrepresentative of the developing world at large.Furthermore, the trial designs were mostly effi-cacy studies, and therefore their effectiveness, ifany, in large-scale programmatic interventions re-mained to be assessed. Inclusion of interventionsin this category in neonatal health programs wasconsidered optional, but a recommendation wasmade to evaluate the benefits whenever these in-terventions were implemented.

4. Clear evidence of benefit: This category of inter-ventions was of incontrovertible benefit to moth-ers and/or newborn infants, and thus it was rec-ommended that they be included in community-based intervention programs for maternal andneonatal care.

This report principally presents the initial analysisof the data based on quality and availability of theevidence. We do not report the projections of theimpact of these interventions, either singly or com-bined as packages, on the global or regional burdenof neonatal mortality. A preliminary exercise of thisnature on a limited number of maternal and neonatalinterventions was conducted by the Bellagio ChildSurvival Study Group,10 and a comprehensive anal-ysis is forthcoming in the Lancet Neonatal SurvivalSeries in March 2005.

ResultsWe found a paucity of data from community-

based settings in developing countries and a remark-able lack of large-scale effectiveness trials of a num-ber of key interventions, especially in relevant healthsystem settings. A total of 186 studies from develop-ing countries were identified for in-depth review, ofwhich only 64 were community-based studies re-porting primary perinatal/neonatal health statusoutcomes such as stillbirths and perinatal and/orneonatal mortality, and 74 were community-basedstudies reporting secondary perinatal/neonatalhealth outcomes such as low birth weight (LBW)and/or anthropometrics, preterm birth, breastfeed-ing rates, and morbidities (Table 1). Of these studiesreporting health outcomes, there were very fewRCTs: 31 community-based RCTs reported primaryneonatal health outcomes, and 40 reported second-ary neonatal health outcomes. Only 10 studies wereinterventions conducted in health system settings, oreffectiveness trials. Most interventions had beentested on relatively small numbers of individuals.There was also wide variation in the quality, size,location, design, and publication source of studies.This variability was considered while summarizingthe information, although we refrain from directcomment on the quality of the evidence in individualstudies, Table 2 (summarizing the strength of theevidence) represents a categorical ranking of inter-ventions based on review of individual studies. Inaddition, however, as noted above, the evidence wasplaced in the context of biological plausibility andknowledge from developed countries, experiencewith the intervention in the context of health pro-grams, and recommendations from the WHO andother leading maternal and child health agencies.

DiscussionAppropriate perinatal and neonatal care in any

given circumstance in developing countries requiresan integrated and holistic program of interventionsat various levels. Interventions must not only includehealth-related measures that have a direct bearing onperinatal and/or neonatal outcomes but severalother ancillary measures of equal importance. Thesemeasures include poverty alleviation; improved op-portunities for female education; and improvementof women’s social status, including empowermentand improvement of women’s decision-making abil-ity. Family size and short interpregnancy intervalsare also critical factors in perinatal health.11–13

Implications for ProgramsThis review of evidence from developing-country

community-based trials for impact of antenatal, in-trapartum, and postnatal interventions on perinataland neonatal outcomes highlights the paucity ofavailable information, particularly from RCTs. Cost-effectiveness data were found to be almost entirelyunavailable. The relative paucity of evidence for im-pact of interventions on neonatal mortality was alsoapparent in the recent analysis of the Bellagio ChildSurvival Study Group,10 which nevertheless in-



TA

BL

E2.

Sum

mar

yof

Impa

ctof

Ant

enat

al,I

ntra

part

um,a

ndPo

stna

talI

nter

vent

ions

onPe

rina

tala

ndN

eona

talH

ealth

Stat

usO

utco

mes

Inte

rven

tion

Evi

den

ceof

No

orN

egat

ive

Impa

ct(L

eave

out

ofPr

ogra

ms)

Unc

erta

inE

vid

ence

(Nee

dfo

rA

dd

itio

nal

Res

earc

hB

efor

eIn

clud

ing

inPr

ogra

ms)

Som

eE

vid

ence

(May

Incl

ude

inPr

ogra

ms,

but

Ad

dit

iona

lE

valu

atio

nIs

War

rant

ed)

Cle

arE

vid

ence

(Mer

its

Incl

usio

nin

Prog

ram

s)

Ant

enat

alin

terv

enti

ons

Mat

erna

lsc

hool

ing/

heal

thed

ucat

ion

XA

nten

atal

care

pack

age(

s)*

XN

utri

tion

-rel

ated

inte

rven

tion

sPr

otei

nsu

pple

men

tati

onX

Bal

ance

dpr

otei

n-en

ergy

supp

lem

enta

tion

XIr

onsu

pple

men

tati

onX

†Pe

rico

ncep

tion

alfo

late

supp

lem

enta

tion

XA

nten

atal

fola

tesu

pple

men

tati

onX

Iod

ine

supp

lem

enta

tion

XA

nten

atal

vita

min

Asu

pple

men

tati

onX

Zin

csu

pple

men

tati

onX

Mul

tipl

em

icro

nutr

ient

supp

lem

enta

tion

XIn

fect

ion-

rela

ted

inte

rven

tion

sM

alar

iaM

alar

iach

emop

roph

ylax

isor

IPT

X‡

ITN

sX

Dew

orm

ing

XU

TIs

and

repr

oduc

tive

trac

tin

fect

ions

Syph

ilis

scre

enin

gan

dtr

eatm

ent

X§

Ant

ibio

tics

for

asym

ptom

atic

bact

eriu

ria

XA

ntib

ioti

csfo

rba

cter

ial

vagi

nosi

sX

Ant

ibio

tics

for

pret

erm

labo

rX

Ant

ibio

tics

for

PPR

OM

X�

Tet

anus

TT

imm

uniz

atio

nX

Cle

and

eliv

ery

prac

tice

sX

Mat

erna

lpn

eum

ococ

cal

imm

uniz

atio

nX

¶O

ther

sPr

omot

ion

ofsm

okin

gce

ssat

ion

inpr

egna

ncy

XM

ater

nal

care

pack

ages

XIn

trap

artu

min

terv

enti

ons

Mat

erna

lva

gina

lan

dne

wbo

rnsk

inan

tise

psis

XPo

stna

tal

inte

rven

tion

sN

ewbo

rnre

susc

itat

ion

XD

elay

edum

bilic

alco

rdcl

ampi

ngX

¶U

mbi

lical

cord

anti

seps

isX

Hyp

othe

rmia

prev

enti

onan

dm

anag

emen

tX

Hyp

ogly

cem

iapr

even

tion

and

man

agem

ent

XB

reas

tfee

din

gX

Prev

enti

onof

opht

halm

iane

onat

orum

X¶

Vit

amin

Kpr

ophy

laxi

sX

Hep

atit

isB

vacc

inat

ion

X¶

Neo

nata

lvi

tam

inA

supp

lem

enta

tion

XK

MC

XT

opic

alem

ollie

ntth

erap

yX

Hyp

erbi

lirub

inem

iasc

reen

ing

X¶

TB

A/

CH

Wtr

aini

ngX

Pneu

mon

iaca

sem

anag

emen

tX

Neo

nata

lca

repa

ckag

esX

Car

ein

peri

pher

alhe

alth

faci

litie

sX

*Pr

iori

tyin

terv

enti

ons

incl

ude

TT

imm

uniz

atio

n,ir

on-f

olat

esu

pple

men

tati

on,d

etec

tion

and

trea

tmen

tof

pre-

ecla

mps

ia,a

nd,w

here

appr

opri

ate,

det

ecti

onan

dm

anag

emen

tof

syph

ilis

and

mal

aria

.†

Incl

usio

nin

prog

ram

sm

aybe

war

rant

edd

espi

tela

ckof

evid

ence

,whi

chis

due

inpa

rtto

ethi

cal

cons

trai

nts

inco

nduc

ting

RC

Ts

ofir

onsu

pple

men

tati

on.

‡In

clus

ion

isin

dic

ated

inen

dem

icar

eas

for

red

ucti

onof

mat

erna

lan

emia

and

para

site

mia

and

toim

prov

ebi

rth

wei

ght;

impa

cton

mor

talit

yne

eds

add

itio

nal

asse

ssm

ent.

§B

enef

its

ofd

iagn

osis

and

trea

tmen

tar

ecl

ear,

but

add

itio

nal

rese

arch

isne

eded

onco

st-e

ffec

tive

mea

nsof

prov

idin

gac

cess

ible

and

qual

ity

dia

gnos

isan

dtr

eatm

ent

atth

eco

mm

unit

yle

vel.

�A

ppro

pria

teat

refe

rral

-lev

elhe

alth

faci

litie

s.¶

Ass

essm

ent

base

don

heal

thin

dic

ator

sot

her

than

neon

atal

heal

thst

atus

outc

omes

.



cluded several neonatal interventions because oftheir proven impact on infant and child survival. Notwithstanding the above exercise, to broaden the rel-evance of the conclusions that can be drawn from theavailable data, we attempted to place the evidence inthe context of biological plausibility, data from stud-ies in developed countries, programmatic experi-ence, and recommendations by the WHO and otherleading child health agencies. In so doing, it is clearthat the evidence for benefit of a number of interven-tions (Table 2) warrants their broad programmaticimplementation (Fig 1). Interestingly, this group ofevidence-based interventions closely resembles thoseadvocated by the WHO14–16 and also identified re-cently through a strategic planning process at theinternational level and in multiple countries, led bythe Saving Newborn Lives Initiative of Save the Chil-dren/USA.17 Thus, there seems to be broad conver-gence of expert opinion and the evidence base re-garding priority interventions to advance perinataland neonatal health and survival at the communitylevel. Considering past experience of child healthprograms in implementation of various interventionsand current recommendations of the WHO and lead-ing child health agencies, a few additional interven-tions (marked with an asterisk) have been added toFig 1 despite the lack of rigorous, prospective scien-tific evidence for their impact. These interventionsinclude birth preparedness; recognition of and ap-

propriate response to danger signs in the antenatalperiod; skilled health care at delivery (evidence re-viewed elsewhere); recognition of and response tointrapartum danger signs; and early postnatal visita-tion for provision of anticipatory guidance and rec-ognition/management of maternal and newborn ill-ness. Many of these interventions have beenincluded in comprehensive packages of maternaland newborn interventions but have not been eval-uated per se for their specific contribution to the totalimpact of the package of care. Such evaluations mustnow be regarded as a priority, especially in healthsystem settings.

Effective interventions span maternal and neonatalcare, as anticipated when one considers that preg-nancy-related causes, delivery-related causes, andinfections each account for approximately one thirdof neonatal deaths.18 This need for a continuum ofcare serves to illustrate the importance of integratingmaternal and neonatal care while avoiding verticalprograms for either the mother or the newborn.Moreover, although we emphasize impact on peri-natal and neonatal outcomes (summarized in Tables4–42), this review has further illustrated the princi-ple that interventions in the antenatal and intrapar-tum periods frequently benefit both mother andnewborn simultaneously.19 Moreover, a coordinatedapproach to postpartum care for mother and new-born would similarly benefit both. Thus, one would

Fig 1. Summary of priority antepartum, intrapartum, and postnatal interventions for inclusion in programs of maternal and neonatalhealth care, based on assessment of available evidence for impact on perinatal and neonatal health status. * indicates essential elementsof the Saving Newborn Lives conceptual framework for advancing newborn health and survival, which either were not reviewed in thisreport or for which evidence is lacking (see “Methods”).



anticipate that an approach to maternal and neonatalhealth integrated within Safe Motherhood and ChildSurvival programs would not only foster continuityof care across the life cycle but also enhance thecost-effectiveness of packages of interventions.

Research GapsAlthough a number of interventions have been

shown to reduce perinatal and/or neonatal mortal-ity, there are fundamental gaps in our knowledge ofhow to most effectively improve perinatal and neo-natal outcomes in developing-country communities.We know that implementation of comprehensiveneonatal care programs can substantially reduceperinatal and/or neonatal mortality, but there is anurgent need to adapt and evaluate culturally andregionally appropriate packages of interventions in avariety of settings. Among these gaps in our knowl-edge is the critical issue of the determinants of familyand community practices and their influence onnewborn care and care-seeking behaviors. Betterdocumentation of how behavior-change interven-tions are implemented and evaluation of these meth-ods is needed to develop better tools for buildingindividual, household, and community capacity forappropriate self-care and care seeking. In the contextof the many parts of the developing world in whichgender inequity and female feticide are major issues,this need for effective behavior-change approachesextends to newborn care and outcomes.

Pivotal questions regarding implementation ofneonatal health care programs that demand addi-tional operational research include: Which cadre ofhealth workers in various settings can most effec-tively deliver the needed services for newborns at thecommunity level, and how can they be linked effec-tively with referral facilities to provide care for ma-ternal and neonatal illness? How will these workersbe trained and supervised in a sustainable manner atscale, and what are the most effective methods forpreservice and in-service training? What will be thescope of their service delivery (eg, with regard toclient age, breadth of services, and geographicreach)? Is a team of skilled birth attendants andnewborn care providers needed at the communitylevel to provide simultaneous care for the motherand newborn during the critical intrapartum period?

The Save the Children/USA conceptual frame-work for newborn care at the community level17 callsfor provision of both preventive and curative care,particularly for birth asphyxia and infections. How-ever, in many settings, provision of curative care forthese major causes of neonatal mortality is beyondthe capacity of current health care systems. Thus,critical unanswered questions are: Can effective im-plementation of a behavior-change communicationspackage at the domiciliary level, without active iden-tification and management of newborn illness, im-prove neonatal outcomes? What is the added benefitand cost-effectiveness of active identification andmanagement of neonatal illness, particularly seriousbacterial infections and intrapartum hypoxia/birthasphyxia? What are the most feasible and effectiveways to deliver life-saving newborn resuscitation

and antibiotic therapy in the community? How canbarriers to care seeking for newborn illness be over-come most effectively so that home-based care andcare seeking can be effectively linked with referral-level care at facilities? What is the impact and cost-effectiveness of postnatal visitation for promotion ofhealthful behaviors and recognition of neonatal ill-ness? Can the same worker address the postnatalneeds of both mothers and newborns? What is theoptimal timing and number of routine visits with ahealth care provider?

Skilled care during delivery is universally recog-nized as a major long-term priority for improving thecare of mothers and newborns, and plans for advanc-ing health system capabilities for providing this careare paramount. Based on a consideration of the factthat most births and neonatal deaths occur at homeduring the early neonatal period, due to birth as-phyxia and/or infections, and among LBW infants,the following emerge as major research gaps:

1. Understanding and improving household andcommunity practices and their determinants: Lo-cal formative research is needed to better under-stand local beliefs and practices and the reasonsbehind them so that effective behavior-changestrategies can be developed and evaluated.23 Thismust be followed by appropriate research to de-velop intervention strategies to improve care-seeking behaviors at the household and commu-nity levels.

2. Improving health systems’ capacity for providingessential preventive and special curative neonatalhealth care: As noted above, some of the mostchallenging questions in neonatal health care re-late to how to most effectively deliver services tonewborns in an integrated way within existingprograms for maternal and child health.20–22 Al-though difficult, determining the answers to thesequestions requires that many packages and com-binations of interventions be tested through effec-tiveness trials in health system settings.

3. Preventing and improving recognition and man-agement of birth asphyxia: Identification ofsustainable interventions for management of in-trapartum hypoxia/birth asphyxia is urgentlyneeded at the community level.24 Solutions mustallow for immediate response at the time of deliv-ery in a cost-effective manner and necessarily willrequire integration with skilled health care formothers at delivery16 and links with referral facil-ities.

4. Preventing and improving recognition and man-agement of infections: There is an urgent need toidentify how the burden and severity of maternalinfections relate to perinatal outcomes. These in-fections may range from subclinical intrauterineinfection and bacterial vaginosis to overt genitaltract infections that may lead to preterm labor.The true burden of bacterial neonatal infections incommunity settings is also unclear, because manyclinical bacterial infections may represent viralinfections. Narrowing this information gap is vi-tal; to devise optimal antibiotic treatment strate-



gies for neonatal infections,25 we need to know theagents of life-threatening infections in the commu-nity and their antibiotic susceptibility patterns.26

There is additional need for validated algorithmsfor accurate and rapid identification of infectedneonates by community health workers (CHWs)and caregivers. We also must advance antibiotic-treatment strategies for serious infections, whichmay include simplified antibiotic-delivery sys-tems and/or regimens. The potential develop-ment and evaluation of simplified oral treatmentregimens that include oral administration will bea major advance for public health programs.

5. Preventing and improving care for LBW infants:Given that the majority of newborns who die inmany developing countries are LBW, improvedstrategies for both prevention of and care for LBWinfants are urgently needed. These strategies in-clude interventions to reduce preterm births andthe incidence of intrauterine growth restriction(IUGR) or combinations thereof. Prevention maybe achieved by improved maternal nutrition anddetection and treatment of maternal infections.Improved postnatal care of LBW infants may beachieved in part by behavior-change communica-tions, topical emollient therapy, breastfeedingpromotion, and widespread implementation ofculturally adapted methods for skin-to-skin care(or variations thereof) with the mother and (whenindicated) other household members. The devel-opment, validation, and availability of low-costtechnology for the care of LBW infants in primaryand secondary health care facilities is an impor-tant adjunct to community-based managementstrategies.

6. Improving information on the magnitude andcauses of neonatal mortality: Lack of accurateglobal, regional, and country-specific data on themagnitude and causes of perinatal and neonatalmorbidity and mortality currently is limiting ad-vocacy and program planning in newborn health.Strengthening of information systems, includingbirth and death registration, and dissemination ofinformation at local levels about causes of new-born morbidity and mortality (and their determi-nants), are needed to guide resource allocationand program and research priorities. Moreover, asprograms incorporate newborn care, their impactmust be monitored and accurate data fed back tothose involved in health policy and program de-cision-making to enable them to use scarce re-sources more effectively. Integral to documentingand monitoring newborn health status is the needfor improved verbal autopsy instruments to en-able more accurate determination of causes ofperinatal and neonatal deaths in the communityand to assess the contribution of socioculturaland logistic factors. Perinatal audit may also be apowerful tool for identifying avoidable factors indeaths and mobilizing change in communities toimprove maternal and neonatal health care.

7. Cost-effectiveness analyses: Assessment of cost-effectiveness must be incorporated into neonatal

health research to guide selection of interventionsand stimulate investment in neonatal health.

8. Development of indicators and simple manage-ment tools for assessing and monitoring healthsystem performance for perinatal and newborncare at the national level: An important impedi-ment to wider implementation of neonatal healthprogramming is lack of inclusion of perinatal andneonatal health indicators among global indica-tors for measuring progress in child survival (eg,Millennium Development goals). Moreover, pro-grams too often fail to monitor adequately anddemonstrate the effectiveness of their programs.Tools for rapidly assessing the situation, prioritiz-ing program activities, and accurately monitoringand documenting program effectiveness are ur-gently needed.

A major factor currently limiting our ability toidentify effective interventions is the wide variationin study designs and indicators for assessing impactand the almost complete absence of cost-effective-ness data. In 2001, a group of neonatal health re-searchers met to discuss a common agenda andmethodologies for neonatal health research in devel-oping-country communities.27 Our review furtherhighlights the need, as recommended at that time,for dialogue among researchers, policy makers, pro-gram managers, and donors in the selection of re-search priorities, use of common (and, wheneverpossible, rigorous) study designs, and for sharing ofdata-collection instruments and research results.

ConclusionsA paucity of community-based data are available

from developing countries on health status impact ofmany interventions that are currently considered forinclusion in health programs for newborns. How-ever, a review of the evidence and consideration ofthe broader context of knowledge, experience, andrecommendations regarding these interventions en-abled us to categorize the interventions according tothe strength of the evidence base and confidence thatthe intervention could be implemented widely andwould improve perinatal and/or neonatal survival.As a result, a package of priority interventions forinclusion in programs was identified, and researchpriorities for advancing the state-of-the-art in neona-tal health care were formulated. Thus, this reviewcan serve as a guide for development of evidence-based maternal and newborn health care program-ming at the community level and for selection ofresearch to advance community-based neonatal care.It also may facilitate dialogue with policy makersabout the importance of investing in newborn health.

Clearly, there is ample evidence for benefit of sev-eral interventions, and, in many cases, operationalquestions of how to implement the intervention(s) inan affordable and acceptable manner at scale were ofoverriding concern. Thus, although there is greatneed for continued research on the cost-effectivenessof a number of interventions, it must not hamperimplementation now of many interventions ofknown impact at wider scale. However, it is impor-



tant that these intervention packages be structured asintegrated maternal and newborn care strategies thatcan be implemented in appropriate health systemsettings. Close communication between programmanagers as they gain experience with interventionimplementation, the researchers who can provideanswers to operational questions, and the donorswho fund the work will be critical to advancingmaternal and neonatal health care at the communitylevel.

INTRODUCTION AND BACKGROUNDInfant and under-5 childhood mortality rates in

developing countries have declined significantly inthe past 2 to 3 decades, whereas neonatal mortalityrates (NMRs) have remained relatively static.18,20,24,28

Neonatal mortality, amounting to an estimated 4million deaths worldwide each year, now comprisesnearly two thirds and two fifths of infant and un-der-5 childhood mortality, respectively, in develop-ing countries,29 and 98% of global neonatal mortalityoccurs in developing countries.30 An equal numberof deaths are thought to occur during the last tri-mester of pregnancy, although data precisely quan-tifying the burden of stillbirths are lacking. Unfortu-nately, most of the countries with high rates ofperinatal and neonatal deaths also have the lowestrates of vital registration of births and deaths.31,32

Moreover, the likelihood of missing live births ishighest for very low birth weight (VLBW) infants,33

and rates of LBW are highest in developing coun-tries, especially Asia.18 In addition, neonatal healthindicators are seldom included in Safe Motherhoodor Child Survival program evaluations, nor are theyamong the outcomes of interest of global agenciesand initiatives. Thus, current estimates of perinataland neonatal mortality, although startlingly high,may nevertheless underestimate the true burden.

It is now recognized that reducing perinatal andneonatal mortality is of paramount importance foradditional gains in child survival to be real-ized.20,26,32,34,35 Moreover, because the majority ofperinatal and neonatal deaths in developing coun-tries occur in the home, there is an urgent need toidentify solutions at the community level.18,20,22,26 Toachieve Millennium Development Goal 4 of halvingchild mortality by the year 2015, major advances inneonatal survival must be achieved through wide-scale implementation of cost-effective interventionsin the community.22

There is little debate that perinatal and neonatalmortality are profoundly affected by proximal fac-tors that influence maternal health such as socioeco-nomic deprivation, gender bias, illiteracy, and highfertility rates, and redress of these factors is critical toimproving maternal and neonatal health in develop-ing countries.4,36 However, these elements are rela-tively resistant to change in the short term.37–42

Moreover, as a consequence of such systematic ne-glect, a sense of fatalism and inevitability of adversefetal and neonatal outcomes sets in and further im-pedes care seeking.22,43,44 This in itself is a majorbarrier to improvement in perinatal and neonataloutcomes. The concept that all people possess equal

rights to health, education, and social services is akey factor in creating demand for better allocation ofhealth care resources for women and newborns. Thismust be coupled with greater participation of indi-viduals and communities in planning and meetingtheir own health care needs, particularly womenwithin traditional societies through empoweringthem to participate in decision-making processes.

Because the health of the mother and newborn areintimately entwined, they must be considered to-gether when planning strategies to improve perinataland neonatal outcomes. It is important to highlightthat the peak period of vulnerability for both themother and newborn is around pregnancy and child-birth. Thus, interventions must largely focus on ad-dressing joint outcomes. There is evidence, however,that this has not been widely adopted, that SafeMotherhood interventions have not adequately ad-dressed the newborn period, and that newborn in-terventions rarely focus on integration with existingmaternal care programs and services.

To redress the burden of perinatal and neonatalmortality, several factors are required: (1) politicalcommitment to newborn health at the global, re-gional, national, and local levels; (2) increased focuson the newborn within existing Safe Motherhoodand Child Survival programs; (3) efficient allocationof resources; (4) effective implementation of cost-effective interventions; and (5) clear documentationof impact.18 To aid in garnering political and pro-grammatic will and action to improve perinatal andneonatal health care and status, the magnitude of theproblem and evidence for effectiveness of interven-tions to prevent and manage adverse outcomes mustbe documented clearly. A recent analysis of the neo-natal burden of disease in south Asia and sub-Sa-haran Africa, in which approximately three fourthsof neonatal deaths occur, highlighted the dearth ofinformation available on neonatal outcomes in devel-oping countries, particularly at the community lev-el.28 Similarly, a recent meeting of neonatal healthresearchers highlighted the need for a review ofavailable evidence for impact of interventions onperinatal and neonatal health and survival.27

Neonatal health experts agree that improving neo-natal health and survival in developing countriesdepends in large measure on more effectively im-plementing what has already been shown towork.18,26,34,35 Moreover, a number of health inter-ventions for the mother and her newborn have beenproposed by the WHO and others as global prioritiesfor programmatic implementation.14,18,26,34,35,45,46 Al-though many advances in obstetric and neonatal careare costly and require technologies that are unavail-able in resource-poor countries, a substantial propor-tion of perinatal and neonatal morbidity and mortal-ity in developing countries could be preventedthrough appropriate adaptations and applications ofinexpensive, relatively simple methods to improveantenatal, obstetric, and neonatal care. The fact re-mains that improvements in care are often limitedmore by lack of adequate knowledge and its appro-priate application than by technologic barriers. Inother cases, however, additional research is needed



to devise, adapt, and evaluate sustainable solutions,particularly at the community level. Although re-views of the impact of certain antepartum, intrapar-tum, and postnatal interventions have been con-ducted, evidence for proven benefit, or lack thereof,of the many interventions that one might include ina neonatal health program at the community levelhas never been systematically evaluated and summa-rized. Major evidence gaps include lack of objectivedata on the methodologies of introducing interven-tions within health system settings and evaluatinghard outcomes through effectiveness-trial designs.The limitations of the strictly randomized-trial de-sign have been recognized in health systems researchand interventions.8

This review of community-based antenatal, intra-partum, and postnatal intervention trials in develop-ing countries was undertaken to (1) identify keybehaviors and interventions for which the weight ofevidence is sufficient to recommend their inclusionin community-based neonatal care programs and(2) identify key gaps in knowledge and priority areasfor future research and program learning. We didnot focus on long-term solutions of established andindisputable value in improving maternal and peri-natal outcomes, such as poverty reduction, genderequity, fertility regulation and control, and improvedhealth system performance. Rather, the focus of thisreview was on specific targeted interventions thatmay impact perinatal and neonatal health status out-comes, primarily perinatal and neonatal mortality.

METHODOLOGY USED FOR LITERATURESEARCH AND REVIEW

This review aimed to consider all available pub-lished and unpublished data on the impact of com-munity-based strategies and interventions on peri-natal and neonatal health status outcomes. Thecommunity was defined as extending from thehousehold to peripheral health facilities.

The search methodology included review of thefollowing sources of information:

1. All available electronic reference libraries of in-dexed medical journals and analytical reviews

2. Electronic reference libraries of nonindexed med-ical journals

3. Nonindexed journals not available in electroniclibraries

4. Pertinent books, monographs, and theses5. Project documents and reports

Electronic Reference SourcesThe following principal sources of electronic refer-

ence libraries were searched to access the availabledata on community-based intervention studies:Cochrane Reference Libraries, the WHO Reproduc-tive Health Libraries, Medline, PubMed, ExtraMed,Embase, and Popline. Several search strategies wereemployed using key words, combinations, and med-ical subject headings (MeSH) words including “com-munity-based care,” “community care,” “newborn orneonatal care,” “perinatal care,” “interventions,” “in-tervention strategies,” “perinatal or newborn care

programs,” “newborn survival,” “perinatal out-comes,” and “neonatal outcomes,” among others.

Manual Literature SearchA detailed examination of cross-references and

bibliographies of available data and publications wasperformed to identify additional sources of informa-tion. In particular, this search extended to reviewingthe gray literature in nonindexed and nonelectronicsources. The bibliographies of 37 recently publishedtextbooks or books with sections pertaining to com-munity-based maternal and/or newborn care werealso searched manually. Requests for informationwere sent to major development and aid agenciesincluding the World Bank, United Nations Chil-dren’s Fund (UNICEF), WHO, Department for Inter-national Development, United Nations DevelopmentProgramme, United States Agency for InternationalDevelopment, MotherCare, JHPIEGO, the WellcomeTrust, LINKAGES, John Snow Inc, National Instituteof Child Health and Human Development, NationalInstitutes of Health (NIH) Institute of Medicine,CARE, Save the Children/USA, and several othernongovernmental organizations. In particular, re-quests for information were made to regionally ac-tive development agencies and research councils. Inaddition, personal requests for information on com-munity-based perinatal and neonatal interventionswere made to leading public health scientists in thefield.

For in-depth review, we selected 186 studies fromdeveloping countries that directly related to the re-search question of health status impact of communi-ty-based perinatal and neonatal health care. Thesestudies were analyzed in detail and summarized inthe tables according to a standardized, prearrangedevaluation format as to their location, size, design,nature of intervention, and outcome. The informa-tion was categorized according to whether the targetgroup consisted of mothers, newborn infants, orboth.

The following categorization of interventions wasmade:

Maternal Interventions

1. Maternal schooling/health education2. Antenatal care packages3. Protein supplementation4. Balanced protein-energy supplementation5. Iron supplementation6. Folate supplementation7. Iodine supplementation8. Antenatal vitamin A supplementation9. Zinc supplementation

10. Multiple micronutrient supplementation11. Malaria chemoprophylaxis or intermittent pre-

sumptive treatment (IPT)12. Malaria protection using insecticide-treated bed

nets (ITNs)13. Deworming14. Syphilis screening and treatment15. Antibiotics for asymptomatic bacteriuria16. Antibiotics for bacterial vaginosis



17. Antibiotics for preterm labor18. Antibiotics for preterm premature rupture of

membranes (PPROM)19. Tetanus toxoid (TT) immunization and clean de-

livery20. Maternal pneumococcal immunization21. Promotion of smoking cessation during pregnancy

Composite InterventionsIn addition to the specific community-based inter-

ventions noted above, some studies evaluated pack-ages of maternal interventions in community set-tings:

22. Maternal care packages

Intrapartum Interventions

1. Maternal vaginal and newborn skin antisepsis

Postnatal Interventions

1. Newborn resuscitation2. Delayed umbilical cord clamping3. Umbilical cord antisepsis4. Hypothermia prevention and management5. Hypoglycemia prevention and management6. Breastfeeding7. Prevention and treatment of ophthalmia neona-

torum8. Vitamin K prophylaxis9. Hepatitis B vaccination

10. Neonatal vitamin A supplementation11. Kangaroo mother care (KMC)12. Topical emollient therapy13. Hyperbilirubinemia screening14. Traditional birth attendant (TBA)/CHW training15. Pneumonia case management

Composite InterventionsApart from community interventions focusing on

the aforementioned specific areas, some studies eval-uated packages of postnatal interventions or thefunctioning of hospitals in the community and inter-ventions performed within them, including use ofalternative methods of care to compensate for mea-ger resources and facilities:

16. Neonatal care packages17. Care in peripheral health facilities

ExclusionsSome interventions were excluded from this re-

view because other investigators were evaluating theevidence base for their impact. Interventions ex-cluded included the following:1. Roles of skilled birth attendants2. Family planning and birth spacing3. Safe Motherhood strategies such as prevention

and treatment for pre-eclampsia, pregnancy-in-duced hypertension, and antepartum hemor-rhage; newer strategies for prevention of pretermlabor (eg, magnesium, calcium, fish oil); emer-gency obstetric care; emergency transport serv-ices; communications strategies; community wait-ing homes; and use of fetal partograph

4. HIV prevention and mother-to-child transmissionreduction strategies

5. Maternal tuberculosis treatment

Synthesis of EvidenceThe principal reviewers independently evaluated

all the data, and a common reporting matrix wasused in summarizing the findings. Emphasis wasplaced on assessment of impact on perinatal or neo-natal primary health status outcomes. For some in-terventions, however, for which data on primaryhealth status outcomes were lacking, other indicatorswere considered.

The final categorization of the interventions wasdone by mutual agreement and consensus as follows:

1. No evidence of benefit: These interventions hadbeen evaluated and found to have no demonstra-ble benefit either singly or in combination withother measures. In some cases, there was evidenceof an adverse effect of the intervention. Therefore,these interventions were not recommended forinclusion in any neonatal health care strategy.

2. Uncertain evidence of benefit: This category in-cluded interventions for which there was someevidence of benefit, but contradictory evidence orissues such as study design, quality, location orsize precluded any firm conclusions. These inter-ventions merited additional evaluation or re-search using well-designed protocols and designs.

3. Some evidence of benefit: These interventions hadsome positive impact on perinatal or neonatal out-comes, but the evidence remained preliminary orthe location of the studies was not representativeof the developing world at large. Furthermore, thetrial designs were mostly efficacy studies; there-fore, their effectiveness, if any, in large-scale pro-grammatic interventions remained to be assessed.Their inclusion in intervention programs was con-sidered optional, but a recommendation wasmade to include an evaluation of benefits when-ever they were included.

4. Clear evidence of benefit: This category of inter-ventions was of incontrovertible benefit to moth-ers and/or newborn infants, and thus it was rec-ommended that they be included in community-based intervention programs for maternal andperinatal care.

When categorizing the evidence for impact of in-terventions, we considered a variety of factors in-cluding the study size, location, and rigor of design;consistency and magnitude of impact reported, par-ticularly on perinatal or neonatal mortality; biologi-cal plausibility of the intervention; evidence fromrelevant developed-country studies; experience withimplementing the intervention in health care pro-grams; and recommendations from the WHO andother leading agencies in maternal and child health.Thus, the evidence was put into a broader context toreach a composite assessment that was agreed on bythe principal investigators (Z.A.B. and G.L.D.).



REVIEW AND ANALYSIS OF AVAILABLE DATA

Antenatal Interventions

Maternal Schooling/Health Educationbackground. The relationship between maternal

antenatal education and perinatal and neonatal out-comes is well established. Caldwell and McDonald47

demonstrated the close relationship between mater-nal education level (ie, schooling) and infant mortal-ity from observations in Nigeria. Since then, thisassociation has been borne out through a number ofreviews.40,48–51 Victora et al,39 in a review of cause-specific infant mortality rates (IMRs) in Pelotas, Bra-zil, found a significant inverse relationship betweenmaternal schooling and deaths from perinatal condi-tions, particularly infectious diseases. The associa-tion between increased maternal education and de-creased infant mortality, particularly in reducingpostneonatal deaths, was further strengthened by areview of 34 cross-sectional World Fertility Surveysbetween 1974 and 1980,52 although the associationwas weakened when the data were corrected forsocioeconomic status53,54 (Table 3). The associationbetween maternal education and neonatal survivalwas also corroborated by analysis of data from 11Demographic and Health Surveys; in this review, theassociation held even after correction for socioeco-nomic status.38,55 More recently, the importance ofmaternal education in reducing birth weight–specificperinatal mortality in Nigeria was further stressed byHarrison,42 and the association between educationallevel of the maternal grandmother and utilization ofhealth services for prenatal care and delivery by askilled attendant was also demonstrated.56

There are surprisingly few intervention studies onthe specific impact of maternal health education onperinatal and newborn outcomes. Kramer57 evalu-ated the evidence, including that from an interven-tion trial in Greece,58 of the potential benefit of ma-ternal nutritional advice during pregnancy onseveral outcomes. Although an increase in energyand protein intake during pregnancy was notable,the overall impact on maternal, perinatal, and neo-natal outcomes was unclear.

Reasons for improved survival of neonates born tomore highly educated mothers is not clear, but theassociation is only partly explained by the economicadvantages and access to health care afforded byeducation. Potential links between maternal educa-tion and reduced perinatal and neonatal mortalityalso include appropriate birth spacing and health-seeking behavior, particularly for prenatal care.

There is strong evidence supporting the importanceof community- and hospital-based maternal educa-tion and support programs on breastfeeding prac-tices59–62; these programs are reviewed below (see“Breastfeeding”).

community-based evidence. Although there aredata available from developed countries on maternaleducational strategies specifically aimed to improveperinatal and neonatal outcomes,63–65 there are fewsystematic studies that have prospectively evaluatedtheir impact,57 particularly from developing coun-tries. Woods and Theron66,67 in South Africa demon-strated a significant improvement in cognitiveknowledge of midwives who participated in an ex-tended perinatal education program; however, im-pact on perinatal outcomes was not reported. In con-trast, providing postnatal maternal education inNepal through a limited didactic educational inter-action met with little success in improving knowl-edge and practices, except for family-planning prac-tices.68 Although the impact of the didactic form ofeducation was not found to be effective, the authorsof the latter study subsequently concluded that com-munity participation was a key to the success ofeducational strategies. These intervention strategiesinclude the development of intervention strategiesby community members themselves, based on theirunderstanding of barriers to care seeking for new-born care.69,70

conclusions. Maternal educational level isclearly associated with improved perinatal and neo-natal survival. Thus, building the capacity of moth-ers through basic education is a key long-term strat-egy to improve perinatal and neonatal health indeveloping-country communities. More work isneeded, however, to develop and test shorter-termmaternal educational strategies targeted toward im-proving pregnancy outcomes in developing coun-tries, particularly at the community level. The exactnature and content of the educational package, rolesof different cadres of health workers, and ways toconvey the messages at the community level mosteffectively may best be developed and evaluatedconsidering the principles of appropriate and partic-ipatory community-based research.71

Antenatal Care Packagesbackground. Antenatal care is well regarded as 1

of the 4 main pillars of Safe Motherhood by theWHO.14 Although the beneficial effects of antenatalcare for maternal health and outcomes are well rec-ognized and the practice is well established, therehave been few systematic studies of the impact of“standardized” antenatal care programs on perinataland neonatal outcomes.72,73 No intervention studiesare available that directly compared groups ofwomen who received antenatal care and those whodid not, thus limiting conclusions regarding the ex-tent to which antenatal care improves perinatal/neo-natal outcomes.

The benefits of antenatal care for maternal andnewborn outcomes, including assessment of themost effective components, were addressed in sys-

TABLE 3. Association of Maternal Education With RR ofNeonatal Death (52)

RRs Years of Schooling

0 1–3 4–6 �7

Neonatal mortalityUnadjusted 1.00 0.87 0.76 0.57Adjusted for socioeconomic factors 1.00 0.98 0.90 0.86

Postneonatal mortalityUnadjusted 1.00 0.82 0.73 0.45Adjusted for socioeconomic factors 1.00 0.98 0.90 0.78



tematic analyses by Bergsjo and Villar72,74 and Car-roli et al.75 Some of the major interventions intro-duced during antenatal care and their impact onpregnancy outcomes are detailed in Table 4. TT im-munization, iron-folate supplementation, detectionand management of pre-eclampsia, screening andtreatment for bacteriuria, and where appropriate,screening and treatment for syphilis and malaria arepriority activities. Although some studies have indi-cated that antenatal care alone may be insufficient forthe identification of pregnant women at risk of ob-stetric complications and emergencies,76–78 there isobservational evidence from a variety of geographicsettings that lack of antenatal care is associated withincreased risk for late fetal death.79,80 Although theevidence is somewhat mixed, the overall consensusis that quality antenatal care provided by a trainedattendant within a functional health system reducesthe risk of maternal mortality and adverse pregnancyoutcomes.81

Although there is some evidence that antenatalcare works, there is little consensus on critical relatedissues such as the minimum number of visits and themost cost-effective components of antenatal care. Inan evaluation of antenatal care models in the UnitedStates, McDuffie et al82 found comparable pregnancyoutcomes among women who had attended a mod-ified program of 2.7 fewer visits, on average, com-pared with the traditional program of 7 antenatalvisits. To further evaluate whether a reduced systemof 4 antenatal care visits was as effective as a pro-gram with more frequent visits, the WHO organizeda multicenter trial involving urban centers in SaudiArabia, Argentina, Cuba, and Thailand (Table 5).83,84

No impact was observed in this large trial on eitherpreterm birth or IUGR.84 Women who received in-formation about breastfeeding antenatally weremore likely to initiate breastfeeding after birth. Thoseassigned to the reduced-visit model had similar ma-ternal (ie, morbidity index, urinary tract infection[UTI], and anemia rates) and neonatal (ie, perinatalmortality rate [PMR], NMR, LBW rate) outcomes asthose who were given standard antenatal care, al-though women who had �4 antenatal visits weremore likely to feed their infants colostrum. The par-ticipants of the trials were generally satisfied withthe quality of care in the new, modified system ofantenatal care.85

A recent systematic review of randomized trialsalso yielded no strong evidence that the content,frequency, or timing of currently recommended an-tenatal care visits has an effect on reducing the inci-dence of IUGR or preterm delivery.75 This systematicreview of pooled data on alternative models of re-duced numbers of antenatal care visits showed sim-ilar odds of multiple health outcomes compared withthe standard care model, including pre-eclampsia(odds ratio [OR]: 0.91; 95% confidence interval [CI]:0.66–1.26), UTIs (OR: 0.93; CI: 0.79–1.1), maternalmortality (OR: 0.91; CI: 0.55–1.51), incidence of LBW(OR: 1.04; CI: 0.93–1.17), and overall perinatal mor-tality (OR: 1.06; CI: 0.82–1.36).T

AB

LE

4.C

ompo

nent

sof

Ant

enat

alC

are

Con

dit

ion/

Inte

rven

tion

Tes

t/T

reat

men

tE

xpec

ted

Impa

cton

Mat

erna

lan

dN

ewbo

rnO

utco

mes

Prev

enti

onof

anem

iaR

outi

nesu

pple

men

tati

onw

ith

iron

and

fola

ted

urin

gpr

egna

ncy

Prev

ents

orre

duc

esfa

llin

mat

erna

lH

b

Mal

aria

chem

opro

phyl

axis

Red

uces

perc

enta

geof

wom

enw

hobe

com

ean

emic

Tre

atm

ent

ofir

on-d

efic

ienc

yan

emia

Ora

lir

onan

dfo

late

Can

rais

em

ater

nal

Hb

by0.

4–0.

7g/

dL

per

wk

Det

ecti

onan

din

vest

igat

ion

ofhy

pert

ensi

ved

isea

seof

preg

nanc

yM

easu

rem

ent

ofbl

ood

pres

sure

wit

hsp

hygm

oman

omet

erus

ing

fift

hK

orot

koff

soun

dD

etec

tsm

ater

nal

hype

rten

sion

and

allo

ws

sele

ctio

nof

case

sat

high

erri

skof

mat

erna

lan

dpe

rina

tal

mor

bid

ity

Ana

lysi

sof

clea

nca

tch

urin

eD

etec

tsm

ater

nal

prot

einu

ria,

ind

icat

ive

ofpr

e-ec

lam

psia

inth

epr

esen

ceof

hype

rten

sion

Tre

atm

ent

ofse

vere

pre-

ecla

mps

iaT

rans

fer

tofi

rst

refe

rral

leve

lfo

rex

pert

care

Con

trol

ofm

ater

nal

dis

ease

;red

uces

mat

erna

lca

sefa

talit

ySc

reen

ing

for

infe

ctio

nSe

rolo

gica

lsc

reen

ing

and

trea

tmen

tfo

rsy

phili

sD

etec

tsas

ympt

omat

icm

ater

nal

dis

ease

;cou

pled

wit

hef

fect

ive

trea

tmen

t,co

ntac

ttr

acin

g,an

dfo

llow

-up,

red

uces

feta

llo

ssan

dm

ater

nal

and

infa

ntm

orbi

dit

yM

icro

biol

ogic

alsc

reen

ing

for

gono

rrhe

aD

etec

tsas

ympt

omat

icm

ater

nal

dis

ease

;cou

pled

wit

hef

fect

ive

trea

tmen

t,co

ntac

ttr

acin

g,an

dfo

llow

-up,

red

uces

feta

llo

ssan

dm

ater

nal

and

infa

ntm

orbi

dit

ySc

reen

ing

for

bact

eriu

ria

wit

hqu

anti

tati

vecu

ltur

eof

urin

eD

etec

tsas

ympt

omat

icd

isea

se;c

oupl

edw

ith

effe

ctiv

etr

eatm

ent,

prev

ents

pyel

onep

hrit

isan

dpr

eter

md

eliv

ery/

LB

WPr

even

tion

ofin

fect

ion

Tet

anus

imm

uniz

atio

nin

preg

nant

and

/or

child

-be

arin

g–ag

ew

omen

Prev

ents

mat

erna

lan

dne

onat

alte

tanu

s



TA

BL

E5.

Ant

enat

alC

are

Pack

ages

Sour

ceL

ocat

ion

and

Typ

eof

Tri

alIn

terv

enti

onM

ater

nal

Out

com

ePe

rina

tal/

Neo

nata

lO

utco

me

Vill

aret

al84

Mul

tice

nter

tria

lin

Saud

iA

rabi

a,A

rgen

tina

,Cub

a,an

dT

haila

nd;u

rban

cent

ers;

RC

T

Wom

en(n

�24

678)

rece

ived

eith

erst

and

ard

ante

nata

lca

re(n

�11

958)

ora

new

mod

elin

volv

ing

few

ervi

sits

inco

rpor

atin

gsc

ient

ific

ally

eval

uate

dan

dob

ject

ive-

orie

nted

acti

viti

es(n

�12

568)

;the

othe

rsw

ere

give

nth

eca

reap

prop

riat

eto

any

det

ecte

dco

ndit

ion

orri

skfa

ctor

.

The

mat

erna

lm

orbi

dit

yin

dex

(83)

reac

hed

ath

resh

old

valu

ein

14.5

%of

wom

enfo

rth

ene

wm

odel

and

16.5

%fo

rth

est

and

ard

mod

el;t

here

wer

eno

sign

ific

ant

dif

fere

nces

inU

TI

orm

ater

nal

anem

iara

tes

betw

een

the

2gr

oups

.

The

PMR

(2%

vs1.

7%),

NM

R(0

.4%

each

),an

dL

BW

rate

s(7

.2%

vs6.

7%)

wer

esi

mila

rin

both

mod

els.

Mun

janj

aet

al71

3Z

imba

bwe;

rura

lse

ttin

g;R

CT

Stud

ycl

inic

sha

dfe

wer

visi

tsbu

tm

ore

obje

ctiv

ely

stru

ctur

edvi

sits

and

few

erpr

oced

ures

(n�

1599

4w

omen

recr

uite

d:9

394

und

erth

ein

terv

enti

onpr

ogra

man

d61

38un

der

the

stan

dar

dpr

ogra

m).

The

rew

ere

6m

ater

nal

dea

ths

inst

udy

clin

ics

vers

us5

inco

ntro

lcl

inic

s.T

here

wer

e12

%fe

wer

prem

atur

ebi

rths

inth

est

udy

clin

ics

(OR

:0.8

6;C

I:0.

78–0

.96)

;no

effe

ctw

asse

enon

LB

W(O

R:0

.96;

CI:

0.85

–1.0

8)or

PMR

(OR

:1.2

;CI:

0.92

–1.6

;P�

.18)

.

Srin

ivas

anet

al71

4In

dia

;rur

alse

ttin

g;R

CT

Wom

enw

ere

rand

omiz

edto

rece

ive

1of

3pa

ckag

es:(

1)hi

gh-r

isk

pack

age,

inw

hich

trai

ned

mid

wiv

esid

enti

fied

high

-ris

kpr

egna

ncie

san

din

terv

ened

acco

rdin

gly;

(2)

Tam

ilN

adu

pack

age,

whi

chin

clud

edtr

aine

dm

idw

ives

but

not

high

-ris

kpr

egna

ncy

iden

tifi

cati

on;a

nd(3

)a

cont

rol

grou

pth

atre

ceiv

eda

basi

cpa

ckag

eof

gove

rnm

ent-

appr

oved

care

(n�

1145

preg

nant

wom

enin

the

stud

yar

ea;3

80w

ere

inth

ehi

gh-

risk

preg

nanc

ygr

oup,

320

wer

ein

the

Tam

ilN

adu

grou

p,an

d44

5w

ere

inth

eco

ntro

lgr

oup)

.

Hig

her

rate

sof

TT

imm

uniz

atio

nw

ere

seen

inth

ehi

gh-r

isk

pack

age

and

Tam

ilN

adu

grou

ps(9

8%an

d95

%,

resp

ecti

vely

)th

anin

the

cont

rol

grou

p(7

2%).

Sim

ilarl

y,hi

gher

rate

sof

iron

-fo

late

usag

ew

ere

obse

rved

inth

ehi

gh-

risk

pack

age

and

Tam

ilN

adu

grou

ps(7

5%an

d79

%,r

espe

ctiv

ely)

than

inth

eco

ntro

lgr

oup

(13%

);th

epr

opor

tion

ofw

omen

wit

hm

ater

nal

Hb

�8

g/d

Lat

34w

kw

as16

.6%

inth

eT

amil

Nad

use

ries

but

only

5.5%

inth

ehi

gh-r

isk

pack

age

seri

es.

The

prop

orti

onof

prev

enta

ble

neon

atal

mor

bid

ity

was

low

erin

the

high

-ris

kpa

ckag

ese

ries

than

inth

eT

amil

Nad

use

ries

by11

%(C

I:�

0.8–

22.8

;P�

.06)

.



conclusions. The benefits and importance of an-tenatal care in improving maternal health and preg-nancy outcomes are widely accepted, yet little directevidence of impact exists from intervention trials. Anantenatal care package that consists of fewer butqualitatively better and more goal-oriented visits isrecognized to be more cost-effective than the “con-ventional” antenatal care packages promoted previ-ously, which involved more frequent visits. How-ever, this evaluation was also not undertaken as aneffectiveness trial in health system settings. It is alsoimportant to point out that there are no studies eval-uating different community-based models of ante-natal care using primary heath care workers andCHWs.

The exact margin of improvement in neonatalmortality after antenatal care is unclear, and wecould not cite a specific figure based on objectiveevidence and controlled trials. Moreover, a con-trolled trial to determine the level of effect wouldnow be unethical. The exact contents of such a pack-age would need to be based on evidence of theefficacy of each individual component of the packageplus the cost-effectiveness and relative ease of imple-mentation by primary care workers. Based on theavailable evidence elaborated in this review, the an-tenatal care package should contain, at a minimum,TT immunization, iron-folate supplementation, andpromotion of clean delivery and exclusive breast-feeding. Based on health system capacity, the pack-age should also include supplementation with iodineand screening and treatment for bacteriuria, pre-eclampsia, and syphilis.

Nutrition Interventions in PregnancyMaternal malnutrition is widespread in develop-

ing countries and is an underlying factor in fetalmalnutrition and LBW as well as other adverse preg-nancy outcomes such as premature births, abruptioplacentae, and stillbirths.86,87 A large proportion(16%) of births in developing countries are LBW,which is a major underlying risk factor for morbidityand mortality in the perinatal and neonatal periodsand later in infancy.86,88 Poor maternal nutritionalstatus is associated with adverse birth outcomes,11,89

but the association with fetal mortality is less clear.Given the recognized association between maternalmalnutrition and LBW, there has been considerableinterest in nutritional interventions that may im-prove birth weight as well as other adverse preg-nancy outcomes.87 With the emerging evidence ofthe long-term implications of fetal malnutrition, nu-trition transition, and adverse metabolic outcomessuch as diabetes,90 it becomes even more imperativeto improve maternal and fetal nutrition in develop-ing countries. Two recent reviews evaluated the im-pact of nutrition interventions on prematurity91 andpregnancy outcomes92 and underscored the fact thatfew studies have addressed this problem in commu-nity settings in developing countries.

Evidence for impact of nutritional interventions onmaternal, perinatal, and neonatal outcomes has beenreviewed extensively, largely within the Cochranecollaboration using meta-analyses of RCTs. Avail-

able data have also been reviewed recently as part ofan evaluation of the evidence base for Safe Mother-hood strategies93 and a review of the efficacy andeffectiveness of nutrition interventions.94 Our evalu-ation of the evidence was drawn largely from thesesources, especially the individual community-basedstudies in developing countries within the Cochranereviews. In addition, we evaluated recent studiesthat have not yet been included in the Cochranereviews and others with a quasi-experimental designthat were not considered as part of the meta-analy-ses.

Protein Supplementationbackground. Benefits of unbalanced protein sup-

plementation in pregnancy were largely refuted re-cently in a meta-analysis of available evidence.95

Such interventions have been tried historically in avariety of malnourished and at-risk populations in-cluding poor communities in developed coun-tries.96,97 In 3 studies among Asian women in theUnited Kingdom and Chile, where the usual mater-nal energy intake was isocalorically replaced with10% to 11% protein,98–100 there was no effect onpregnancy outcomes, although there was a trendtoward reduced birth weight. Even higher levels ofprotein supplementation (�25% of energy) in rela-tively well-nourished populations failed to showany benefit on pregnancy outcomes and birthweight.101,102 Thus, protein supplementation alone isno longer viewed as a viable intervention duringpregnancy.103

conclusions. Based on a large body of evidence,pure or high levels of dietary protein supplementa-tion cannot be recommended as an antenatal inter-vention, nor is additional research warranted on thisintervention.

Balanced Protein-Energy Supplementationbackground. Balanced protein-energy supple-

ments, by definition, provide �25% of their totalenergy content in the form of protein. A systematicreview done by the Cochrane collaboration on theeffect of antenatal maternal balanced protein-energysupplementation95 concluded that this interventionsignificantly improved fetal growth and reduced therisk of fetal and neonatal death. The findings of thisreview, however, were largely influenced by 1 largetrial undertaken in The Gambia that indicated a sig-nificant reduction in perinatal mortality.104 How-ever, this efficacy study also included micronutrientsupplementation in addition to balanced protein-en-ergy intake. Excluding this single study drasticallyaltered the conclusions of this meta-analysis, leavingno demonstrable impact.

community-based evidence. A review of the lit-erature identified 19 studies, 12 of which were un-dertaken in community settings and discussed preg-nancy outcomes, thus fulfilling our criteria forselection. The details of these studies are given inTable 6. These trials were largely conducted in de-veloping countries and inner-city populations in in-dustrialized countries. Inconsistent results may havebeen related to the variability in the backgroundrates of maternal malnutrition in the different study



TABLE 6. Antenatal Balanced Protein-Energy Supplementation

Source Location and Typeof Trial

Intervention Maternal Outcome Perinatal/Neonatal Outcome

Ceesay etal104

The Gambia; ruralsetting; RCT

Treatment group consisted of pregnantwomen who received high–proteinenergy biscuits; the control groupconsisted of women who did notreceive the supplement (247 normalsingleton live births among 1460women; treatment group: 1010 livebirths; control group: 1037 livebirths).

Incidence of stillbirths decreasedby 53% in the interventiongroup (OR: 0.47; CI: 0.23–0.99;P � .05); a 46% decrease wasrecorded in the number ofdeaths occurring in the firstweek of life (OR: 0.54; CI:0.35–0.85; P � .01), but therewas no effect on mortalityrates after 7 d of life (P � .2);there was a significantincrease in birth weight ofinfants born to mothers in theintervention group (meanincrease: 136 g; P � .001); a39% decrease was seen inLBW rates among the infantsof the intervention group (P �.001).

Kusin etal109

East Java, Indonesia;rural setting; QT

Pregnant women (n � 542) wereassigned to low- and high-energysupplementation groups.

For women who consumed thesupplements for �45 d, therewas a significant difference inbirth weight (463 g) betweenthe 2 groups; a difference of421 g (P � .017) was seen ininfant weight at 12 mo of age.

Kardjati etal116

East Java, Indonesia;rural setting;double-blind,randomized,controlled trial(DBRCT)

Pregnant women (n � 741) in 3villages were assigned to a high-energy (n � 272) or low-energy (n �265) group; there were 204noncompliers excluded from thesegroups; among these, 6 additionalgroups were formed according tothe number of energy packetsconsumed during the trial.

No difference in weightgain was seen amongthe women of the 2groups comparedwith noncompliers.

There was no significant effectof the supplementation onbirth weight and LBW rateswhen compared to thebaseline period during whichno supplementation wasprovided.

Villar andRivera111

Guatemala; ruralsetting; QT

169 mothers and their offspring wereenrolled during 2 consecutivepregnancies and the interimlactation period; 2 types ofsupplements (atole, a high-proteinand high-calorie gruel, and fresco, ano-protein/low-calorie cold liquiddrink) were offered to pregnantand/or lactating women.

There was an impact of thesupplement on birth weight,with a linear trend forincreased (P � .05) birthweight from the highest (atole)to the lowest (fresco)supplemented groups.

Tontisirinet al110

Thailand; ruralsetting; RCT

Women (n � 43) in their thirdtrimester were divided into 3groups: groups 1 and 2 (14 in eachgroup) received 2 different levels ofsupplementation, whereas group 3(n � 15) acted as controls; group 2had the highest protein and caloriecontent in their supplement.

There was a significant increasein birth weight of infants bornto women in group 1(3089 � 308 g; P � .025) andgroup 2 (3104 � 259 g; P �.025) compared to the group 3controls (2853 � 247.9 g); theplacental weights also weresignificantly higher in the 2supplemented groups(630 � 669 and 616 � 68 g ingroups 1 and 2, respectively;P � .005) that were sup-plemented compared to thecontrol group (563 �509 g).

Girija etal117

India; urbanhospital setting;RCT

Pregnant women (n � 20) wererecruited for the study: 1 group (n �10) was supplemented in the lasttrimester with sesame cakes, jaggery(raw sugar), and oil; women in thecontrol group (n � 10) were notgiven the supplement.

There was no differencein weight gain ofwomen in thesupplementcompared with thecontrol group;women in thesupplemented group,however, had asignificantly greaterincrease in Hb levels.

There was no significantdifference in birth weight andlength at birth of infants bornto the 2 groups of women.



TABLE 6. Continued



Prentice etal112

The Gambia; ruralsetting; QT

Pregnant women who were, onaverage, in their 16th wk ofgestation were supplemented withpeanut-based biscuits and a vitamin-fortified tea drink (n � 652nonsupplemented women and 577supplemented women).

Infants born to thesupplemented mothers weresignificantly heavier (101 g)than those born tononsupplemented mothers (P� .05); this beneficial effect ofsupplementation was seenonly in the wet season, notduring the dry season; a67.5% decrease was observedin the proportion of LBWinfants in the supplementedgroup over the year (P � .01);this effect was greater in thewet season, during whichthere was a decrease of 83.3%in the proportion of LBWinfants (P � .01);supplementation also seemedto contribute to increasedhead circumference in infants.

McDonaldet al108

Taiwan; ruralsetting; DBRCT

Women (n � 212) were distributedrandomly to groups A (n � 107) andB (n � 105); pregnant women ingroup A were given high-proteinand -calorie liquid supplements, andthe supplement was provided togroup B was low in calories and hadno protein; the trial started after thebirth of the “first study infant,” andthe women were supplemented untilonset of lactation of the “secondstudy infant.”

There was no difference in theincidence of fetal deathsamong the “first” and“second” study infants of the2 groups; a mean increase inbirth weight of 161.4 g wasobserved in the second studyinfants in group A (P � .05);LBW rates decreased to 2.8%for the second infant in groupA (statistically insignificant),whereas it remained the samefor group B at 6.8%; there wasno effect of the supplementson prematurity rates.

Kielmannet al118

India; rural setting;RCT

3 groups of villages were randomizedinto nutrition care (NUT), nutritioncare plus medical care (NUT � MC),or medical care (MC); a fourthgroup (emergency care only) servedas the control group; nutrition carewas comprised of nutritioneducation, surveillance, and foodsupplementation (1675 J and 11 g ofprotein per day, fortified with ironand folate) through special feedingcenters; medical care includedimmunization, health education,early diagnosis, and treatmentthrough frequent surveillance,provided singly and combined asneeded; monitoring from birth toage 36 mo was done throughlongitudinal anthropometry,morbidity, and vital statistics.

There was a 41.3% decrease inPMR in the nutrition groupcompared to the controlgroup; this decrease washigher among the 3intervention groups (MED,NUT, and MED � NUTgroups) compared to controls;the decrease in NMR washighest in the nutritionintervention group (41%)compared to the controlgroup.

Moraet al245*

Colombia; urbansetting; QT

Women (n � 456) in their thirdtrimester and from a family havingat least 50% of childrenmalnourished were given dry milk,bread, and vegetable oil for thewhole family.

Antenatal weight gainsignificantlyincreased (meanincrease: 140 g/wk;P � .05) amongsupplementedwomen who gavebirth to male infants.

Birth weight was significantlygreater (95 g; P � .05) amongmale infants born tosupplemented compared tononsupplemented women;there was no significant effectof the supplement on LBWrates.

Moraet al105*

Colombia; urbansetting; RCT

456 pregnant women were recruitedfor the study; women in thesupplemented group (n � 226) weregiven a total of 230 g of dry skimmilk, enriched bread, vegetable oil,and vitamin mineral supplementduring the third trimester; controls(n � 230) were not supplemented.

There was a 46% decrease in theNMR and a 25% reduction inthe PMR. A 75% decrease wasdetected in the stillbirth rateamong the infants of womenin the supplemented group(on pooled sample ofstillbirths and neonatal deaths,�2 (1 df) � 4.1; P � .05).



settings and the relative size of the individual stud-ies. Of the trials included in this review, only 4reported preterm birth rates.97,101,105–107 Supplemen-tation was not associated with an increase in meangestational age (mean difference: �0.1 week; CI:�0.2 to �0.1 week) or a significant reduction inpreterm birth (OR: 0.83; CI: 0.65–1.06). Supplemen-tation generally resulted in increased birth weightand/or a reduction in the LBW rate.104,105,108–115

Overall, however, balanced energy-protein supple-mentation seems to have only a modest effect onmean birth weight (weighted mean difference: 25 g;CI: �4 to �55 g) but a more substantial effect onreducing IUGR (OR: 0.68; CI: 0.57–0.80). No evidencewas found that these effects were greater in under-nourished than in well-nourished women. However,the magnitude of the birth weight increase was sub-stantially larger (136 g) in the Gambian study,104 inwhich the supplement provided an additional 3780kJ per day, as compared with an 840- to 1050-kJ-per-day increase in most of the other trials. Although atrend toward increased weight gain of the supple-mented mothers was observed in 2 studies, the dif-ferences were nonsignificant.105,114 Moreover, otherstudies showed no impact on maternal weightgain.116,117 In the few studies that examined effectson the stillbirth104,105 and perinatal mortality104,105,118

rates, reductions were seen. The largest of thesestudies104 was undertaken in The Gambia, where, inan RCT, chronically undernourished pregnantwomen were provided a higher-energy supplement(3780 kJ), largely toward the last trimester, with littlemicronutrient content. Results from this Gambianstudy104 reported significant reductions in rates ofstillbirths (53% reduction), early neonatal deaths(46% reduction), and LBW (39% reduction).

In contrast to findings noted above regarding un-balanced protein supplementation alone in pregnan-cy,95 the impressive reduction in rates of stillbirths

and perinatal mortality from this large trial providedstrong evidence of the potential benefit of balancedprotein-energy supplementation during pregnancy.

conclusions. Although Kramer95 did not find adifferential effect of balanced protein-energy supple-ments according to the degree of maternal malnutri-tion, the weight of evidence is strong in favor ofimproving perinatal mortality and birth weightthrough balanced protein-energy supplementationof malnourished pregnant women. Most of the evi-dence, however, comes from strict efficacy trials con-ducted under intense supervision, and the overallresults are largely driven by a single trial from TheGambia. No effectiveness trials have been under-taken to evaluate the benefit of balanced-energy pro-tein supplementation at the community level nor ofusing home-available diets to provide these supple-ments. We believe that balanced protein-energy sup-plementation merits additional field evaluation indiverse geographic locations and may be cautiouslyincluded in intervention programs in malnourishedpopulations. However, if such a program is insti-tuted, data must be collected to evaluate the pro-gram’s benefit and cost-effectiveness. Ideally, thebenefit of improved protein-energy intake in preg-nancy may be achieved through dietary diversifica-tion strategies as well as targeted supplementationin at-risk populations, although the cost may besubstantial.

Iron Supplementationbackground. Global estimates by the WHO indi-

cate that 55% of all pregnant women living in devel-oping countries and 18% of those in developedcountries are anemic (hemoglobin [Hb] concentra-tion �11 g/dL).119 It is also recognized that anemiaunderlies some 8% to 15% of maternal deaths indeveloping countries.120–122 Although the exact con-tribution of maternal anemia to maternal mortality

TABLE 6. Continued



Lechtiget al115

Guatemala; ruralsetting; QT

Chronically malnourished women of 2villages (n � 405) received atole (ahigh-protein/-calorie gruel), whereaswomen of 2 villages were givenfresco (a low-calorie/no-protein cooldrink).

There was a significantcorrelation between caloricsupplementation duringpregnancy and birth weight(P � .01); the rate of LBWbirths was approximately halfin the high-calorie/protein-supplemented motherscompared to the low-calorie/protein-supplementedmothers (P � .05).

Qureshi etal114


3 groups of women (n � 126) wereformed and followed from 20 wkgestation to birth; women in group 1(n � 37) received an iron � folatesupplement, and group 2 (n � 39)received a dietary supplement highin protein and calories, as well asthe iron � folate supplement; group3 (n � 50) received none of theabove and served as the controlgroup.

Women in group 2gained an average of1 kg more duringpregnancy than thosein group 1.

Infants born to women in group1 were 0.27 kg heavier thanthe infants born to women inthe control group; the infantsborn to women in group 2were 0.81 kg heavier thanthose in the control group.

* Data are from the same trial.



may be unclear,123,124 it is also widely recognized asa major determinant of maternal morbidity in devel-oping countries. The majority of such cases of anemiaare related to iron deficiency, although malaria andhookworm infestation, as well as protein and othermicronutrient deficiencies, may play a role also.121

Iron-deficiency anemia is highly prevalent in de-veloping countries, affecting an estimated 2 billionpeople, including one fourth of the world’s womenand children.125,126 Thus, there has been much inter-est in interventions geared toward improving ironintake and status during pregnancy. Despite the ev-idence that gastrointestinal iron absorption increasesduring pregnancy, it is highly unlikely that sufficientamounts can be absorbed from the diet during thisperiod to compensate for the increased requirementof the body. Thus, supplementation with iron gener-ally is required, especially where diets may be defi-cient in iron and body stores of iron may be inade-quate to meet requirements.127

Although it is widely accepted that iron-deficiencyanemia poses an increased risk of complications inpregnancy and of maternal and perinatal mortali-ty,128 there is surprisingly little evidence to supportthis relationship. An association has been suggestedby some epidemiologic studies,129–131 but other stud-ies, in fact, have failed to demonstrate a relationshipbetween iron-deficiency anemia and adverse preg-nancy outcomes.132,133 Most of the evidence in theliterature has been largely derived from retrospec-tive studies and has not been controlled for ancillaryfactors such as overall nutrition, underlying health,and health service delivery.123 Several review-ers123,134–138 believe that anemia, which itself can bedue to a variety of factors, is just one of a multitudeof determinants of maternal and perinatal mortalityand that there is no conclusive evidence of a linkbetween maternal anemia and LBW or maternal orperinatal mortality. Moreover, there is little docu-mentation that health status can be improved bytreating anemia alone. There are no iron-supplemen-tation trials with maternal mortality as a measuredoutcome, and all intervention trials that used perina-tal mortality as the outcome involved nonanemicwomen,139 were poorly designed,140,141 or were toosmall to be conclusive about iron effects.142 How-ever, despite the fragmentary nature of the data onthe association between maternal anemia and mor-tality, one can infer that there is a steep rise in ma-ternal mortality with increasing severity of anemia,especially Hb levels �5 g/dL.143 Nevertheless, theassociation may not be causal.128 There also seems tobe a U-shaped relationship between maternal anemiaand birth weight, because both low and high mater-nal Hb values are associated with an increased risk ofLBW.135,144,145 Again, although no causal evidencehas been established overall to support or refute therelationship between iron-deficiency anemia andLBW, the evidence from developing countries, inwhich iron-deficiency anemia is common, shows thatmaternal iron deficiency is positively associated withLBW and poor obstetric outcome.136 There is alsoevidence suggesting that a relationship exists be-

tween maternal anemia in early pregnancy and in-creased risk of preterm birth.146–149

Meta-analyses of iron-supplementation trials, con-ducted under the auspices of the Cochrane collabo-ration,150–152 concluded that although iron supple-mentation significantly reduced the prevalence oflow Hb concentration (�105 g/L), it had no detect-able effect on any other substantive measures ofmaternal or perinatal outcomes. Although iron re-duces maternal anemia, there is no evidence that ironsupplements administered alone or with folate haveany effect on birth weight or fetal survival in devel-oped countries. There are insufficient data from de-veloping countries to draw conclusions; the fewRCTs are inconclusive because of small sample size,problems with compliance, and large losses to fol-low-up. Small trials in The Gambia, Nigeria, andIndia showed no significant effects on birthweight,140,153–155 although 1 large trial in Nigershowed a significant increase in birth length andApgar scores and a reduction in PMR.156

There are several other reports of iron therapy inpregnancy from developing countries that are rela-tively less stringent in their inclusion and exclusioncriteria.157–163 However, most of these studies werebased in health facilities within developing countriesand varied considerably in design. The overall effectsof iron therapy in these studies were largely mixedand did not suggest a particular effect. A review of 9RCTs of iron supplementation from health facilitiesin developing countries showed that the incrementin Hb level for given amounts of supplementationwas not greatest among those with the lowest Hblevels, suggesting that factors other than iron defi-ciency alone were operative.164

community-based evidence. Few community-based iron-supplementation trials have resulted inimprovement in either iron stores or measures ofmaternal and perinatal health.164–166 Among iron-supplementation trials conducted to date, 5 were indeveloping countries: 1 in a periurban setting andthe other 4 in rural settings (Table 7). All 5 studiesdemonstrated a convincing effect of iron supple-mentation on reducing rates of maternalanemia.141,154–156,167,168

Several studies have evaluated the benefit of com-bining iron with other micronutrients in an effort toaddress multiple deficiencies. In a randomized, pla-cebo-controlled trial (RPCT) in Indonesia, Suharno etal168 found a synergistic effect of vitamin A and ironsupplementation in reducing rates of maternal ane-mia, with two thirds of the response attributable tovitamin A. Another study in India141 demonstrated areduction in LBW rates and an increase in birthweight of infants born to women who were supple-mented with both iron and folate. Increased birthweight, however, was seen only in offspring ofwomen who began supplementation from 16 to 20weeks’ gestation but not in those supplemented afterweek 20 of gestation. Christian et al169 indicated thatboth iron-folate supplements and multiple micronu-trients reduced the prevalence of LBW comparablyamong pregnant women in rural Nepal (16% vs



TA

BL

E7.

Ant

enat

alIr

onSu

pple

men

tati

on

Sour

ceL

ocat

ion

and

Typ

eof

Tri

alIn

terv

enti

onM

ater

nal

Out

com

ePe

rina

tal/

Neo

nata

lO

utco

me

Prez

iosi

etal

156

Nig

er;p

eriu

rban

sett

ing;

DB

RC

TW

omen

(n�

197)

wer

egi

ven

100

mg/

del

emen

tal

iron

(n�

99)

orpl

aceb

o(n

�98

).

Supp

lem

enta

tion

had

asi

gnif

ican

tef

fect

ind

ecre

asin

gth

epr

eval

ence

ofan

emia

(a6%

dec

reas

ein

anem

iapr

eval

ence

was

obse

rved

atd

eliv

ery

com

pare

dto

6m

oof

gest

atio

n).

Of

8fe

tal

orne

onat

ald

eath

s,7

wer

ein

the

plac

ebo

grou

p,an

d1

was

inth

eir

ongr

oup;

Apg

arsc

ores

wer

esi

gnif

ican

tly

bett

erin

the

supp

lem

ente

dgr

oup

than

the

cont

rols

;the

supp

lem

ent

also

was

asso

ciat

edw

ith

incr

ease

dbi

rth

leng

thof

infa

nts

but

not

birt

hw

eigh

t(P

�.0

5).

Men

end

ezet

al15

4T

heG

ambi

a;ru

ral

sett

ing;

DB

RC

TM

ulti

grav

ida

wom

en(n

�55

0)w

ere

give

n60

mg/

del

emen

tal

iron

(n�

273)

orpl

aceb

o(n

�27

7);a

llw

omen

rece

ived

5m

gof

folic

acid

wee

kly.

Supp

lem

enta

tion

incr

ease

dH

ble

vel

by0.

5g/

dL

(P�

.01)

and

impr

oved

plas

ma

iron

leve

lby

2.7

�m

ol/

L(P

�.0

01)

inth

epr

egna

ntw

omen

.

Supp

lem

enta

tion

was

not

foun

dto

have

any

sign

ific

ant

infl

uenc

eon

birt

hw

eigh

tan

dpr

eval

ence

ofL

BW

exce

ptin

wom

enw

hoha

dta

ken

the

supp

lem

ent

for

�80

d;t

his

subg

roup

ofw

omen

del

iver

edsi

gnif

ican

tly

heav

ier

babi

es(m

ean

dif

fere

nce:

9.2

g;P

�.0

4)co

mpa

red

toth

eco

ntro

ls.

Atu

kora

laet

al16

7Sr

iL

anka

;rur

alse

ttin

g;pr

ospe

ctiv

eco

hort

stud

y(P

CS)

Preg

nant

wom

en(n

�19

5)w

ere

give

na

fort

ifie

dfo

odsu

pple

men

t(t

hrip

osha

),ir

onsu

pple

men

ts(6

0m

g/d

),an

d0.

25m

gof

fola

te.

Iron

supp

lem

enta

tion

impr

oved

mat

erna

lH

bpa

ram

eter

sir

resp

ecti

veof

how

man

yta

blet

sw

ere

take

n(P

�.0

1);s

uppl

emen

tati

onfo

r�

17w

kw

asm

ore

effe

ctiv

eth

an�

17w

k.

No

effe

ctof

the

supp

lem

enta

tion

onbi

rth

wei

ght

was

note

d.

Aga

rwal

etal

141

Ind

ia;r

ural

sett

ing;

RC

TPr

egna

ntw

omen

betw

een

16an

d24

wk

ofge

stat

ion

(n�

418)

wer

ese

lect

edfr

om6

subc

ente

rsof

aru

ral

bloc

kof

Var

anas

id

istr

ict;

wom

enfr

om3

subc

ente

rsw

hoch

ose

topa

rtic

ipat

ein

the

stud

ygr

oup

(n�

137)

rece

ived

60m

gof

elem

enta

lir

onco

mbi

ned

wit

h50

0�

gof

folic

acid

dai

lyfo

r10

0d

;w

omen

from

anot

her

3su

bcen

ters

who

wer

eun

supp

lem

ente

dse

rved

asco

ntro

ls(n

�12

3).

Com

pare

dto

base

line,

iron

-fol

ate

supp

lem

enta

tion

incr

ease

dH

ble

vel

(mea

nd

iffe

renc

e:1.

8g/

dL

;P�

.001

)as

wel

las

seru

mfe

rrit

inle

vel

(mea

nd

iffe

renc

e:20

.57

g/d

L;P

�.0

01)

inth

epr

egna

ntw

omen

.

Supp

lem

ente

dw

omen

del

iver

edsi

gnif

ican

tly

heav

ier

babi

esco

mpa

red

toco

ntro

ls(2

.88

�0.

41vs

2.59

�0.

34kg

;P�

.001

);th

ere

was

a46

%d

ecre

ase

inL

BW

rate

sin

the

supp

lem

ente

dgr

oup

com

pare

dto

cont

rols

;the

perc

enta

geof

LB

Wba

bies

was

20.4

%in

the

stud

ygr

oup

com

pare

dto

37.9

%in

the

cont

rol

grou

p(P

�.0

5);t

hem

ean

birt

hw

eigh

tof

infa

nts

born

tow

omen

give

nsu

pple

men

tsst

arti

ngbe

twee

n16

and

19w

kof

gest

atio

nw

asgr

eate

rth

anth

em

ean

birt

hw

eigh

tof

infa

nts

born

tow

omen

supp

lem

ente

daf

ter

the

20th

wee

kof

gest

atio

n(2

.95

�0.

31vs

2.86

�0.

43kg

).Su

harn

oet

al16

8W

est

Java

,Ind

ones

ia;

rura

lse

ttin

g;d

oubl

e-bl

ind

,ran

dom

ized

,pl

aceb

o-co

ntro

lled

tria

l(D

BR

PCT

)

Wom

en(n

�25

1)w

ith

Hb

leve

lsbe

twee

n80

and

109

g/L

wer

era

ndom

ized

to1

of4

grou

psre

ceiv

ing

vita

min

Aan

dpl

aceb

oir

on(n

�63

),ir

onan

dpl

aceb

ovi

tam

inA

(n�

63),

iron

and

vita

min

A(n

�63

),or

both

plac

ebos

(n�

62)

for

8w

k.

Max

imum

Hb

was

achi

eved

inw

omen

rece

ivin

gbo

thvi

tam

inA

and

iron

(12.

78g/

L;C

I:10

.86–

14.7

0);o

neth

ird

ofth

ere

spon

sew

asat

trib

utab

leto

vita

min

A,a

ndtw

oth

ird

sof

the

resp

onse

was

attr

ibut

able

toir

on;t

hepr

opor

tion

ofw

omen

who

beca

me

nona

nem

icw

as35

%in

the

vita

min

A-s

uppl

emen

ted

grou

p,68

%in

the

iron

-sup

plem

ente

dgr

oup,

97%

inth

egr

oup

rece

ivin

gvi

tam

inA

plus

iron

,an

d16

%in

the

plac

ebo

grou

p.



14%). A different study167 found a synergistic effectof antihelminthics with iron supplementation. How-ever, this particular study also indicated that womenwho were supplemented for �17 weeks did notshow any benefits, and there was no overall effect ofsupplementation on birth weight. One study154

showed an increase in birth weight but only in theoffspring of women who had taken the iron supple-ment for a period of �80 days.

It must be pointed out also that there is evidencethat administration of parenteral iron to iron-defi-cient women during pregnancy in malaria-endemicareas may be associated with an increase in the inci-dence of malarial infections and clinical episodes ofdisease.170,171 Thus, care must be exercised in admin-istering iron supplements alone (without chemopro-phylaxis) in malaria-endemic areas.

conclusions. Oral iron supplementation may im-prove maternal anemia, but there is no clear effect ofiron supplementation on maternal and perinatal orneonatal outcomes. However, the evidence for im-pact of iron supplementation on health outcomes isinconclusive, primarily due to a paucity of ade-quately designed and robust trials of iron supple-mentation in developing countries rather than ademonstrated lack of effect.123,136 In any case, itseems that initiation of therapy early in gestation isimportant.

Pending the results of well-designed trials thatmight more definitely delineate the role and mode ofiron supplementation in deficient populations, andgiven the apparent importance of anemia, especiallysevere anemia, as a risk factor for maternal mortalityand morbidity,128 it would seem prudent (pendingadditional research, including meta-analyses) to con-tinue with iron supplementation, concomitant withfolate administration (see “Folate Supplementa-tion”), during pregnancy and in at-risk populationsof women of reproductive age. In malaria-endemicareas, routine iron supplementation during preg-nancy must be accompanied by malaria chemo-prophylaxis (see “Malaria Chemoprophylaxis orIntermittent Therapy”). Current WHO/UNICEFguidelines recommend universal iron-folate pro-phylactic supplementation of young children andpregnant women in areas where anemia is highlyprevalent.172

Folate Supplementationbackground. Folate is critical for DNA synthesis,

and folate deficiency is associated with dysfunctionin rapidly dividing cells. Observational studies havesuggested that lower maternal serum folate levels areassociated with LBW and prematurity.173,174 A largeUS study suggests an association between highermaternal serum folate at 30 weeks’ gestation andlower risk of IUGR, higher birth weight, and higherApgar scores,175 although these results have not beencorroborated by data from South America.176,177

There is a large body of literature, mainly fromdeveloped countries, reporting observational studiesand RCTs of folic acid in pregnancy. These have beenreviewed by Ramakrishnan et al,145 de Onis et al,86

and Mahomed.151,152 Some observational studies178

have shown positive associations between maternalfolate status and birth weight, but the evidence isinconsistent. Twenty-one trials of folate supplemen-tation were included in the Cochrane review,151

which concluded that, despite a significant reductionin maternal anemia, there was only a small andnonsignificant effect on the incidence of LBW (OR:0.73; CI: 0.47–1.13). de Onis et al86 included 4 folatetrials in their review of nutritional interventions toprevent IUGR and found a significant reduction inLBW, but they commented on the poor quality ofmuch of the data. Although 2 studies from India179

and South Africa180 showed significant increases inbirth weight of offspring of malnourished women,this result has not been confirmed in subsequentstudies.

Folate deficiency in early pregnancy is an impor-tant factor underlying the occurrence of neural tubedefects (NTDs) such as spina bifida, encephalocele,and anencephaly. Worldwide, NTDs (particularlythe 2 major types, anencephaly and spina bifida) andencephalocele are estimated to affect �300 000 in-fants each year.181,182 A trial of periconceptional fo-late supplementation (400 �g of folic acid per daybefore and during the first 28 days after conception),conducted by the Medical Research Council183 in 33centers among developed countries, showed a reduc-tion of NTDs by 50% to 70%. Similar data werereported by Berry et al.184 Universal folic acid forti-fication of flour at 240 �g/100 g in consumed foodproducts has been shown to significantly reduceNTD-affected conceptions and births.185 Currently,the benefits of periconceptional folate supplementa-tion in terms of reducing the incidence of NTDs arewell established, and this intervention should bemade universally available.

Although there is little evidence of widespreadfolate deficiency, folic acid is commonly adminis-tered along with iron supplements during preg-nancy. The benefit of periconceptional folate admin-istration has been convincingly demonstrated in alarge cohort (n � 23 491) in the United States.186

Prevalence ratio estimates of NTDs among womenwho took multivitamins containing folic acid duringthe first 6 weeks of pregnancy were 0.27 and 0.29,respectively, depending on whether they had a fam-ily history of an NTD.

Two meta-analyses undertaken by the Cochranecollaboration evaluated the impact of folate supple-mentation during either pregnancy151 or the pericon-ceptional period.187 In the former review of 21 stud-ies, folate supplementation during pregnancy wasassociated with a reduction in the proportion ofwomen with low Hb levels in late pregnancy (OR:0.61; CI: 0.52–0.71). Apart from the small, nonsig-nificant reduction in the incidence of LBW notedabove (OR: 0.73; CI: 0.47–1.13), folate supplementa-tion seemed to have no measurable effect on anyother substantive measure of pregnancy outcomesuch as pregnancy-induced hypertension, placentalabruption, or preterm delivery. Although folate ad-ministration significantly improved maternal Hb lev-els and folate status, there was insufficient evidenceto evaluate whether folate supplementation during



pregnancy had any effect, either beneficial or harm-ful, on clinical outcomes for mother and infant.

The second meta-analysis of periconceptional fo-late supplementation trials187 principally included 4trials with 6425 women and concluded that pericon-ceptional folate supplementation significantly re-duced the incidence of NTDs (OR: 0.28; CI: 0.15–0.53). Of the 4 trials, 3 also evaluated differentmultivitamin combinations, including various micro-nutrients. Multivitamins or micronutrients alonewere not associated with prevention of NTDs anddid not produce any additive benefits when givenwith folate. There was no effect on rates of miscar-riage (OR: 1.12; CI: 0.98–1.29) or stillbirth (OR: 0.78;CI: 0.34–1.78). There was some evidence that folatesupplements increase the risk of multiple births (OR:1.40; CI: 0.93–2.11); although not statistically signifi-cant, this was a consistent finding in 3 studies. Noneof these studies reported birth weight as an outcome.A trial of periconceptional and first trimester folicacid and multivitamins conducted by the IndianCouncil of Medical Research showed no effect onabortions or stillbirths and a nonsignificant effect onLBW (12.5% [folic acid] vs 15.6% [placebo]).188

We identified an additional 10 studies in devel-oped countries that also evaluated the potential ben-efits of periconceptional folate administration; only 1of these studies189 presented data on neonatal out-comes other than congenital malformations. Thislarge study from Denmark189 showed a decrease inpreterm, small-for-gestational-age (SGA), and LBWrates in neonates born to mothers who were supple-mented with folate. Women who benefited werethose supplemented with folate in the periconcep-tional period, compared with women who eitherwere not supplemented or were supplemented dur-ing some other time of pregnancy. There was nodifference in the effect of a 1-mg daily dose com-pared with a 2.5-mg daily dose of folate. Other stud-ies revealed a significant reduction in the incidenceof primary and recurrent cases of NTDs among ne-onates born to mothers supplemented with folate.190,191

Furthermore, these studies showed a decrease in the

incidence of other congenital malformations (eg, limbdefects, cardiovascular problems, urinary tract anoma-lies, orofacial abnormalities).191–195

Given the recent interest in the prevalence of ma-ternal hyperhomocysteinemia among folate and B12-deficient populations and the potential relationshipwith adverse pregnancy outcomes,196,197 there is re-newed emphasis on ensuring adequate maternal fo-late and B12 status during pregnancy. The serumconcentrations of these nutrients have been shown tocorrelate with maternal pre-eclampsia,198 perinataloutcomes,199 and newborn measurements.200

community-based evidence. Two studies fromdeveloping countries provide evidence for efficacy offolate supplementation (Table 8). In South Africa,there was a significant decrease in the incidence ofLBW births among native African women supple-mented with folic acid during pregnancy.180 InChina, a large-scale intervention with optional folicacid supplementation given to women premaritallyrevealed a significant reduction in the incidence ofNTDs in infants born to women who had receivedthe folic acid supplements.184

conclusions. Based on the demonstrated impactof folic acid supplements given along with iron inimproving Hb levels in iron-deficient populations,folate merits inclusion in antenatal supplementationprograms in developing countries. Overall, the evi-dence also strongly supports the use of periconcep-tional folate supplementation to reduce NTDs, al-though the impact of folate supplementation onother pregnancy outcomes such as LBW and pretermbirths is not clear.

Several other aspects of folate use in pregnancyremain unanswered, such as the optimal dosage, therole of genetic polymorphisms at a population level,and the considerable logistic difficulties in providingfolic acid to susceptible women periconception-ally.201 Our review of the evidence on folate supple-mentation suggests that this intervention is impor-tant in areas of maternal malnutrition and endemicmicronutrient deficiencies, and it already is used

TABLE 8. Antenatal Folate Supplementation

Source Location and Type of Trial Intervention Perinatal/Neonatal Outcome

Berry etal184

China; population-based trial;QT

247 831 pregnant women were studied(31 960 from the northern provinceshaving high rates of NTDs and215 871 from the southern region ofChina with lower rates of NTDs); ofthe 31 960 women in the northernregion, 18 591 women used folicacid, whereas 13 369 did not; in thesouthern region, 111 551 used folicacid, whereas 104 320 did not.

Among the women who did not take anyfolic acid, the rates of NTDs amongoffspring were 6.5 of 1000 pregnanciesof at least 20 wk of gestation in thenorthern region and 0.8 of 1000pregnancies of at least 20 wk ofgestation in the southern region;among the fetuses or infants of thewomen who used any folic acid, therespective rates were 1.3 of 1000 (79%reduction; OR: 0.21; CI: 0.13–0.32) and0.7 of 1000 (16% reduction; OR: 0.84;CI: 0.61–1.14) pregnancies of at least 20wk gestation.

Baumslaget al180

South Africa; tertiary hospital;RCT

Women (n � 354) were randomized to1 of 2 groups; group 1 received 200mg of iron daily, group 2 received 5mg of folic acid daily � iron, andgroup 3 received 50 �g of vitaminB12 � iron and folate.

19 babies born to native African womenin group 1 had a birth weight �2270 g,compared with only 4 such babiesamong native African women ingroup 2.



routinely in maternal health programs. Better target-ing of this intervention, by ensuring that youngwomen receive folic acid along with iron, may re-duce rates of NTDs and other congenital malforma-tions and improve pregnancy outcomes.

Iodine Supplementationbackground. Iodine-dependent thyroid hor-

mones play a critical role in brain cell proliferation,synapse formation and microtubular assembly. Thebeneficial effect of iodine supplementation on en-demic cretinism and goiter has been well established,and deficiency disorders are now understood tomanifest across a spectrum ranging from subclinicaliodine deficiency to overt hypothyroidism and goi-ter. The global importance of endemic iodine defi-ciency in many developing countries is well recog-nized.202 It is typically associated with mountainousareas but may pose a problem anywhere that iodineis leached from the soil by heavy rainfall or flooding.It is estimated that approximately �1 billion peoplelive in high-risk areas, particularly China, India,Eastern Europe, and parts of Africa. Clinical effectsof iodine deficiency are particularly serious duringpregnancy, and can result in miscarriages, early in-fant death, LBW, and cognitive deficiencies.

Three trials involving 1551 women were includedin the Cochrane review of data on the impact ofmaternal iodine supplementation on pregnancy out-comes, including cretinism and mortality of off-spring.203 In these trials,204–207 which were con-ducted in areas of endemic iodine deficiency, the useof injectable iodized oil resulted in a significant re-duction in mortality in infancy and childhood (OR:0.71; CI: 0.56–0.90) and reduced risk of endemiccretinism by the age of 4 years (OR: 0.27; CI: 0.12–0.60). The trial from Zaire further revealed that ad-ministration of iodized-oil injections was effective inreducing infant mortality and improving neurologicoutcome even when given in midpregnancy.207–209

community-based evidence. Apart from the datafrom Zaire and Papua New Guinea, our search iden-tified 1 additional community-based trial of iodinesupplementation (Table 9). In rural China, adminis-tration of iodine in drinking water resulted in asignificant (65%) reduction in neonatal mortality (seeref 715). However, no data on neurologic develop-ment or cognitive function were available from thisintervention trial.

conclusions. Although limited data on the im-pact of iodine supplementation on pregnancy out-comes are available from developing-country com-munities, the potential benefits (ie, reduced neonatalmortality and neurologic deficits) are substantial.Moreover, these findings are in accord with a vastbody of literature from developed countries that alsoestablishes the crucial role of iodine in cognitivedevelopment. The evidence indicates the enormouspotential of public health approaches to this prob-lem, especially in countries with endemic iodine de-ficiency. Therefore, administration of adequateamounts of iodine during pregnancy to expectantmothers should be an important component ofhealth care interventions during pregnancy. Al-

though universal salt iodization is a well-establishedintervention to control endemic iodine deficiency,there is still a need to develop innovative means ofproviding iodine supplements to deficient populations.

Antenatal Vitamin A Supplementationbackground. Although there is conclusive evi-

dence for reduction of childhood mortality aftervitamin A supplementation,210 the correspondingevidence for impact of vitamin A supplementation inpregnancy on outcomes in the perinatal/neonatalperiod is less well established. In Nepalese women,night blindness in the third trimester was associatedwith a significantly increased risk of anemia, infec-tions, and pre-eclampsia as well as maternal mortal-ity.211 Similarly, maternal night blindness was asso-ciated with increased risk of mortality in earlyinfancy.212 Although there is evidence of an asso-ciation between serum vitamin A or carotenoidsand birth weight,213,214 it has not been foundconsistently.173,175,215–217

community-based evidence. Five trials per-formed in rural settings in developing countries pro-vided information on vitamin A supplementationrelevant to this review (Table 10). None of the trialsreported birth weight as an outcome. In an RPCT inIndonesia in which women were supplemented withvitamin A, iron, or both, Suharno et al168 found thatwomen given both supplements had maximal Hbincreases and one third of this response was attrib-utable to vitamin A supplementation, suggestingthat vitamin A may help reduce rates of maternalanemia and subsequent adverse birth outcomes.

In Nepal, a recent series of double-blind, cluster-randomized trials examined the impact of low-dosevitamin A or �-carotene supplementation on a num-ber of associated maternal and neonatal health out-comes. Daily supplementation of married women(n � 44 646) with vitamin A or �-carotene providedpreconceptionally, during pregnancy, and duringlactation resulted in a highly significant (44%) reduc-tion in maternal mortality and reduced the maternalmortality ratio by 40%.218 There was no impact, how-ever, on perinatal or neonatal mortality, nor wasthere any impact on rates of fetal loss (miscarriage,stillbirth, loss due to maternal death), preterm births,or early infant survival.219 However, the risk of earlyinfant mortality was reduced among night-blindwomen supplemented with vitamin A.212 Thesetrials in Nepal showed no broadly observed effect ofvitamin A supplementation on neonatal mortality ormortality in the first 6 months,219 although subanaly-sis suggests that there may have been a trend towardan effect.212

Some of the most important data on the potentialbenefits of high-dose antenatal vitamin A supple-mentation are from studies in HIV-infected women.Trials among HIV-positive mothers in Tanzania220

and South Africa221,222 showed no effect of vitamin Aalone on fetal growth or fetal loss. In an RCT, HIV-positive pregnant women supplemented with 5000units of vitamin A and 30 mg of �-carotene duringthe third trimester and 200 000 units of vitamin A atbirth had a reduced incidence of preterm births of



11%, compared with 17% among unsupplementedwomen.222 In Malawi, administration of 10 000 unitsof vitamin A in HIV-positive women lowered theincidence of LBW compared with placebo-treatedcontrol women.223,224

conclusions. Data on the impact of antenatal vi-tamin A supplementation are limited, and the bio-logical plausibility of the intervention has been ques-tioned.88 The impact seems to be robust for maternalmortality, but evidence of benefit for perinatal orneonatal outcomes is lacking. There is an urgentneed to undertake additional studies of antenatalvitamin A supplementation in areas of endemic orsubclinical vitamin A deficiency. These trials must bedesigned with sufficient power to determine impacton maternal mortality, pregnancy outcomes, andmortality in infancy. Similarly, the strong evidencefor efficacy of �-carotene in reducing maternal mor-tality suggests that other forms of supplementationalso merit consideration. However, all such trialsmust be designed as effectiveness trials, with suffi-cient information regarding appropriate process in-dicators to inform programs.

Notwithstanding the above, the data on reductionof maternal mortality are compelling. Thus, it makessense to ensure that pregnant women in areas ofendemic subclinical vitamin A deficiency and night

blindness receive at least the recommended dietaryallowance (RDA) of vitamin A during pregnancyfrom both dietary sources and supplements.

Zinc Supplementationbackground. Because zinc is a critical nutrient

involved in immunocompetence, growth, and devel-opment, there has been much interest in its role inpregnancy. King225 and Keen et al226 reviewed theevidence for an association between zinc deficiencyand adverse outcomes in pregnancy. They concludedthat, although evidence for the critical nature of zincin pregnancy is based largely on animal studies,there were data to suggest that zinc was indeed animportant nutrient in human pregnancy. A review of17 studies published from 1977 to 1994 also indicatedan association between maternal indicators of zincstatus and birth weight of offspring, although severalothers were inconclusive.175 A further in-depth re-view of 10 zinc-supplementation trials conducted upto 1996 revealed that birth weight increased afterzinc supplementation in 4 of the 10 trials.145 How-ever, the Cochrane review on maternal zinc supple-mentation227 revealed no significant differences inmaternal or neonatal outcomes after zinc supplemen-tation, although there was a small effect of supple-mentation on reduction of preterm delivery rates. It

TABLE 9. Antenatal Iodine Supplementation

Source Location andType of Trial


DeLonget al715

China; ruralsetting; RCT

Potassium iodate was addedto the rivers that were themain source of water forthe entire population ofintervention villages;control villages usedwater from rivers withoutany potassium iodate.

Neonatal mortality was reducedby 65% in the populationsupplemented with iodine(OR: 0.35; CI: 0.18–0.68).

Pharoahet al204,206

New Guinea;population-based; QT

The entire population(16 000 individuals in 27villages) was giveninjectable iodized oil(4 mL if �12 y of age);however, the follow-upcohort was considerablysmaller.

Definite cretinism was reducedby 100% in the neonates bornto the intervention group(0 of 274 in the interventiongroup vs 14 of 248 in thecontrol group; �2 (3 df) �33.87; P � .001); the 15-ycumulative survival of thechildren whose mothersreceived supplementaryiodine was significantlygreater than the survival ofcontrol children (P � .002).

Thillyet al208

Zaire; ruralsetting; PCS

Pregnant women (n � 109)around 28 wk of gestationattending antenatal clinicsin a severely iodinedeficient area in Zairewere allocated to either atreatment or controlgroup; the treatmentgroup received 1 iodizedoil injection at the time ofthe first antenatal clinicvisit, and the controlgroup received aninjection of iodine-freevitamins; there were 118newborns born to thesewomen during theintervention period.

Correlation between infant thyroidfunction and maternal thyroidfunction was noticed; treatmentwith iodized oil significantlyimproved mean thyroxineserum concentration (15.7 � 0.7vs 11.5 � 0.7 �g/dL inuntreated versus treatedmothers; P � .001) and loweredmean thyroid-stimulatinghormone (5.4 � 0.5 vs 8.7 � 1.2�U/mL; P � .001).

Birth weight increased by amean 101 g (�13.5 to215.5 g); there was also atrend toward reduced infantmortality (RR: 0.66; CI:0.44–1.03).



TA

BL

E10

.A

nten

atal

Vit

amin

ASu

pple

men

tati

on

Sour

ceL

ocat

ion

and

Typ

eof

Tri

alIn

terv

enti

onM

ater

nal

Out

com

ePe

rina

tal/

Neo

nata

lO

utco

me

Chr

istia

net

al*2

12N

epal

;rur

alse

ttin

g;D

BPR

CT

See

Kat

zet

al21

9(b

elow

)C

ompa

red

toof

fspr

ing

ofno

n–ni

ght-

blin

dw

omen

,ris

kof

infa

ntm

orta

lity

duri

ngth

efir

st6

mo

oflif

ew

asin

crea

sed

the

leas

tam

ong

offs

prin

gof

nigh

t-bl

ind

mot

hers

supp

lem

ente

dw

ithvi

tam

inA

(OR

:1.1

4;C

I:0.

67–1

.93)

,was

inte

rmed

iate

for

thos

ew

hose

nigh

t-bl

ind

mot

her

had

rece

ived

�-c

arot

ene

(OR

:1.5

0;C

I:0.

97–2

.33)

,and

was

high

est

for

thos

ebo

rnto

plac

ebo-

cont

roln

ight

-blin

dw

omen

(OR

:1.6

3;C

I:1.

09–2

.38)

.K

atz

etal

*219

Nep

al;r

ural

sett

ing;

RPC

TA

llm

arri

edw

omen

ofch

ild-b

eari

ngag

ein

each

ofth

e9

stud

yw

ards

(n�

4355

9)w

ere

assi

gned

rand

omly

to(1

)a

wee

kly

dose

of70

00�

gof

retin

oleq

uiva

lent

asre

tinyl

palm

itate

(vita

min

A),

(2)

42m

gal

l-tra

ns�

-ca

rote

ne(7

000

�g

ofre

tinol

equi

vale

nt),

or(3

)pl

aceb

o(n

�15

832

wom

enw

hoco

ntri

bute

dda

ta;d

urin

gth

est

udy

peri

od,1

737

3w

ere

preg

nant

,and

ther

ew

ere

1598

7liv

e-bo

rnin

fant

s).

Ther

ew

asno

effe

ctof

eith

ersu

pple

men

t(v

itam

inA

or�

-ca

rote

ne)

onfe

tall

oss,

neon

atal

mor

talit

y,or

prev

alen

ceof

pret

erm

birt

hs.

Cou

tsou

dis

etal

222

Sout

hA

fric

a;ur

ban

hosp

itals

ettin

g;D

BRPC

T

Preg

nant

HIV

-infe

cted

wom

en(n

�72

8)w

ere

rand

omiz

edto

rece

ive

eith

ervi

tam

inA

(n�

368)

orpl

aceb

o(n

�36

0);t

hevi

tam

inA

trea

tmen

tco

nsis

ted

ofa

daily

dose

of50

00IU

ofre

tinyl

palm

itate

and

30m

gof

�-c

arot

ene

duri

ngth

eth

ird

trim

este

rof

preg

nanc

yan

d20

000

0IU

ofre

tinyl

palm

itate

atde

liver

y.

Wom

enre

ceiv

ing

the

vita

min

Asu

pple

men

tw

ere

less

likel

yto

have

apr

eter

mde

liver

y(1

1.4%

inth

evi

tam

inA

grou

pan

d17

.4%

inth

epl

aceb

ogr

oup,

P�

.03)

.

HIV

-infe

ctio

nre

sults

wer

eav

aila

ble

on63

2ch

ildre

nw

how

ere

incl

uded

inth

eK

apla

n-M

eier

tran

smis

sion

anal

ysis

.The

rew

asno

diff

eren

cein

the

risk

ofH

IVin

fect

ion

by3

mo

ofag

ebe

twee

nth

evi

tam

inA

(20.

3%;C

I:15

.7–2

4.9)

and

plac

ebo

(22.

3%;C

I:17

.5–2

7.1)

grou

ps,n

orw

ere

ther

edi

ffer

ence

sin

feta

lmor

talit

yra

teor

IMR

betw

een

the

2gr

oups

;am

ong

the

80pr

eter

mde

liver

ies,

thos

eas

sign

edto

the

vita

min

Agr

oup

wer

e50

%le

sslik

ely

tobe

HIV

infe

cted

(17.

9%;C

I:3.

5–32

.2)

than

thos

eas

sign

edto

the

plac

ebo

grou

p(3

3.8%

;CI:

19.8

–47.

8).

Wes

tet

al*2

18N

epal

;rur

alse

ttin

g;D

BRPC

TSe

eK

atz

etal

(abo

ve)

(n�

4464

6m

arri

edw

omen

,of

who

m22

119

wer

epr

egna

nt).

Mat

erna

lmor

talit

yw

asre

duce

d44

%(O

R:0

.56;

CI:

0.37

–0.8

4;P

�.0

05)

amon

gw

omen

supp

lem

ente

dw

ithvi

tam

inA

or�

-car

oten

eco

mpa

red

topl

aceb

o.Fa

wzi

etal

220

Tanz

ania

;urb

anse

ttin

g;D

BRC

TH

IV-in

fect

edw

omen

(n�

1075

)w

ere

rand

omiz

edin

to4

grou

ps:g

roup

1(n

�26

9)re

ceiv

eda

daily

dose

ofvi

tam

inA

,gro

up2

(n�

269)

was

give

nm

ultiv

itam

ins

excl

udin

gvi

tam

inA

,gro

up3

(n�

270)

rece

ived

mul

tivita

min

sin

clud

ing

vita

min

A,a

ndgr

oup

4(t

heco

ntro

lgro

up,n

�26

7)w

asgi

ven

plac

ebo.

Ris

kof

LBW

(�25

00g)

was

decr

ease

dby

44%

(OR

:0.5

6;C

I:0.

38–0

.82;

P�

.003

);ri

skof

seve

repr

eter

mbi

rth

(�34

wk)

was

redu

ced

by39

%(O

R:0

.61;

CI:

0.38

–0.9

6;P

�.0

3),a

ndri

skof

SGA

was

redu

ced

by43

%(O

R:0

.57;

CI:

0.39

–0.8

2;P

�.0

02)

ingr

oups

2an

d3

(mul

tivita

min

-sup

plem

ente

dgr

oups

with

and

with

out

vita

min

A);

vita

min

Agi

ven

alon

eha

dno

effe

cton

inci

denc

eof

feta

ldea

ths

(OR

:0.8

9;C

I:0.

58–1

.36;

P�

.59)

,LB

W(O

R:1

.14;

CI:

0.68

–1.9

4;P

�.6

2),V

LBW

,pre

term

birt

hs,

orSG

Abi

rths

(OR

:0.8

3;C

I:0.

58–1

.18;

P�

.29)

.K

umw

enda

etal

224

Mal

awi;

urba

nho

spita

lset

ting;

RC

T

HIV

-infe

cted

wom

en(n

�69

7)w

ere

rand

omiz

edto

rece

ive

daily

dose

sof

iron

and

fola

teei

ther

alon

e(c

ontr

olgr

oup)

orco

mbi

ned

with

vita

min

A(3

mg

ofre

tinol

equi

vale

nt),

from

18to

28w

kof

gest

atio

nun

tilde

liver

y.

Inth

evi

tam

inA

and

cont

rolg

roup

s,th

em

ean

birt

hw

eigh

tw

as28

95�

31an

d28

05�

32g,

resp

ectiv

ely

(P�

.05)

,and

the

prop

ortio

nof

LBW

infa

nts

was

14.0

%an

d21

.1%

(P�

.03)

invi

tam

inA

-sup

plem

ente

dan

dco

ntro

lgro

ups,

resp

ectiv

ely;

the

prop

ortio

nof

anem

icin

fant

sat

6w

kpo

stpa

rtum

was

23.4

%an

d40

.6%

inth

e2

grou

ps,r

espe

ctiv

ely

(P�

.001

).Su

harn

oet

al16

8Se

eTa

ble

7

*D

ata

are

from

the

sam

etr

ial



was concluded, however, that there was insufficientevidence to fully evaluate the effect of zinc supple-mentation during pregnancy. Moreover, it is impor-tant to point out that most of the studies in theCochrane review were from developed countrieswith relatively few malnourished women, and therehave been a number of more recent studies fromdeveloping countries that may shed more light onthis subject (reviewed below). On the other hand,other reviewers94,228 have also concluded that it isunlikely that zinc supplementation alone will influ-ence birth weight or improve pregnancy outcomes indeveloping countries. Recent studies have focusedalso on combinations of zinc with other nutrientssuch as vitamin A. However, the results have beeninconsistent. In 1 instance, although a positive im-pact of either zinc or vitamin A supplementation onmaternal iron status during pregnancy was ob-served, the combination of the 2 failed to improveserum ferritin or Hb concentrations.229

community-based evidence. Our search identi-fied 6 studies for additional review, only 1 of whichtook place in a rural setting (Table 11). The majorityof the studies were undertaken in urban slums andwere comprised of subjects who had low zinc intakeand were considered to be at high risk of zinc defi-ciency during pregnancy.

The additional studies of zinc supplementation inpregnancy reviewed here also produced mixed re-sults. Only 1 study from Chile230 showed a decreasein prematurity rates in the supplemented group,along with a reduction in rate of LBW; the latterresult was also seen in a study from the UnitedStates.231 Overall, the majority of the studies failed toshow a significant impact of zinc supplementationon birth weight, preterm delivery rate, and eitherneonatal or perinatal mortality. A study from theurban slums of Bangladesh,232,233 however, showed asignificant impact of zinc supplementation on ratesof infectious disease morbidity (eg, diarrhea, dysen-tery, and impetigo) during the first 6 months of lifeamong LBW infants born to zinc-supplementedmothers.

There are also ancillary studies that have evalu-ated the effect of zinc supplementation on maternaland infant outcomes that may be of value. Sazawal etal234 showed a significant (68%) reduction in mortal-ity during the first 9.5 months of life among SGAinfants born to zinc-supplemented women in urbanIndia. Thus, although it is unlikely that isolated zincsupplementation among malnourished women willimpact LBW rates, it is possible that zinc supplemen-tation may improve infectious disease morbidity andboost immunity beyond the neonatal period. Recentdata on combined vitamin A and zinc supplementa-tion of pregnant women in central Java indicate thatthis combination supplement may significantly re-duce anemia and rates of puerperal sepsis,229 high-lighting the need to address multiple micronutrientdeficiencies in at-risk populations. Moreover, datafrom Bangladesh showed that offspring of womensupplemented with zinc from 4 months’ gestation todelivery had lower scores on mental and psychomo-

tor development indices at 13 months of age.235,236

These findings suggest the need for caution in ad-dressing micronutrient deficiencies in malnourishedpopulations with single-nutrient solutions.

conclusions. It is highly unlikely that isolatedzinc deficiency exists in at-risk populations; signifi-cant iron deficiency, among other nutrient defi-ciencies, frequently coexists with subclinical zincdeficiency. Evidence for benefit of isolated zinc sup-plementation on pregnancy outcome is relativelyweak, and at present we cannot recommend isolatedzinc supplementation in pregnancy. However, allattempts should be made to replenish zinc stores orprovide at least the RDA of zinc in malnourishedpopulations. Whenever possible, this must be donethrough formulations that address other key micro-nutrient deficiencies that coexist with zinc defi-ciency, such as iron and vitamin A deficiencies, andimplemented through multiple program pathways.237

Multiple-Micronutrient Supplementationbackground. With the emerging evidence of mul-

tiple micronutrient deficiencies in pregnancy, espe-cially in HIV-endemic areas,238,239 there has beenmuch interest recently in interventions employingmultiple micronutrient supplements in pregnancy byusing the standard UNICEF multiple-micronutrientsupplements. Randomized trials of multiple micro-nutrient interventions in pregnancy are currently un-derway in a number of developing countries withmalnourished populations, such as Nepal, Pakistan,Guinea Bissau, Bangladesh, and Indonesia.

The 10-year observational Camden Study on theimpact of multivitamin supplementation on preg-nancy and perinatal outcomes was conducted in apoor US urban setting240 and included 1430 pregnantwomen in various stages of pregnancy. Risk of LBWwas reduced approximately twofold with supple-ment use during the first (OR: 0.63; CI: 0.39–1.0)and/or second (OR: 0.57; CI: 0.38–0.86) trimester.The use of prenatal supplements during the first andsecond trimester was also associated with an approx-imately twofold reduction in risk of preterm delivery(ORs: 0.53 and 0.71; CIs: 0.35–0.81 and 0.5–1.01; forfirst and second trimester use, respectively). Thus, itseemed that the decrease in rates of LBW or VLBWwas not due to a decrease in rates of IUGR but ratherto a decrease in the rates of preterm infants. In an-other large trial in the United States reported in 1989,use of multivitamins lacking in folic acid content hadno impact on prevalence of NTDs.186

In contrast to the above-mentioned studies, whichfocused on micronutrient supplementation, there area large number of studies that have used combina-tions of food and energy supplements and micronu-trients. Most of these data are derived from pro-grams and interventions in developed countries. TheUS government’s Women, Infants, and ChildrenSupplemental Nutrition Program (WIC) provideslow-income mothers with vouchers for milk, eggs,cheese, fruit juice, cereals, legumes, and peanut but-ter.241 Although this intervention has not been as-sessed prospectively in a randomized trial, it hasbeen estimated retrospectively242 that participation



TA

BL

E11

.A

nten

atal

Zin

cSu

pple

men

tati

on

Sour

ceL

ocat

ion

and

Typ

eof

Tri

alIn

terv

enti

onM

ater

nal

Out

com

ePe

rina

tal/

Neo

nata

lO

utco

me

Cas

tillo

-Dur

anet

al23

0C

hile

;urb

ansl

um;R

CT

Preg

nant

adol

esce

nts

�20

wk

gest

atio

n(n

�80

4)w

ere

adm

inis

tere

dei

ther

zinc

supp

lem

enta

tion

(S)

at20

mg/

d(n

�40

1)or

plac

ebo

(P)

(n�

403)

until

deliv

ery.

No

effe

ctof

supp

lem

enta

tion

was

seen

onm

ean

birt

hw

eigh

tof

the

infa

nts;

how

ever

,the

prop

ortio

nof

LBW

birt

hs(�

2500

g)in

the

Sgr

oup

was

sign

ifica

ntly

low

erth

anin

the

Pgr

oup

(6/2

49vs

16/2

58,P

�.0

36).

The

LBW

rate

and

the

pret

erm

birt

hra

tede

crea

sed

by66

%an

d51

%,r

espe

ctiv

ely,

inth

eS

grou

p.M

ultip

lere

gres

sion

sfo

und

asi

gnifi

cant

effe

ctof

mat

erna

lnut

ritio

nals

tatu

s(P

�.0

11)

and

zinc

supp

lem

enta

tion

(P�

.05)

onbi

rth

wei

ght.

Dijk

huiz

enet

al71

6In

done

sia;

rura

lset

ting;

RC

TW

omen

betw

een

10–2

0w

kge

stat

ion

(n�

229)

wer

eas

sign

edto

1of

the

4gr

oups

.All

grou

psre

ceiv

edir

on�

fola

te.1

rece

ived

�-c

arot

ene

(n�

42),

1w

asgi

ven

zinc

at30

mg/

d(n

�42

)an

d1

was

give

nal

l4su

pple

men

ts(n

�42

).Th

eco

ntro

lgro

upre

ceiv

edon

lyir

on�

fola

te(n

�38

).

Com

bina

tion

ofpr

olon

ged

labo

ran

d/or

reta

ined

plac

enta

,bot

hin

dica

tive

ofut

erin

eco

ntra

ctile

dysf

unct

ion,

was

sign

ifica

ntly

high

erin

the

zinc

grou

pco

mpa

red

toal

loth

ergr

oups

.

The

mal

ein

fant

sbo

rnto

the

wom

ensu

pple

men

ted

with

�-

caro

tene

�zi

nc�

iron

�fo

late

wer

esi

gnifi

cant

lyhe

avie

r(m

ean

wei

ght

3.4

�0.

3kg

)th

anth

ein

fant

sin

the

zinc

�ir

on�

fola

tegr

oup

(3.0

�0.

6kg

,P�

.05)

orth

ein

fant

sin

the

zinc

�ir

on�

fola

tegr

oup

(3.1

�0.

4kg

).

Mer

iald

i717

Peru

;urb

ansh

anty

tow

n;R

CT

Wom

enre

ceiv

ing

iron

and

fola

te(n

�24

2)w

ere

recr

uite

dbe

twee

n10

–16

wk

preg

nanc

yan

dra

ndom

ized

tore

ceiv

eei

ther

25m

gzi

nc/d

vers

uspl

aceb

o.

Inad

ditio

nto

birt

hw

eigh

t,gr

owth

offe

tala

nato

mic

alpa

ram

eter

sw

asas

sess

edby

ultr

ason

ogra

phy.

Apo

sitiv

eef

fect

ofm

ater

nalz

inc

supp

lem

enta

tion

onfe

tal

fem

urdi

aphy

sis

grow

thw

asde

tect

ed.N

oef

fect

was

seen

onbi

rth

wei

ght.

Ose

ndar

pet

al*2

33Ba

ngla

desh

;urb

ansl

um;

PCS

See

belo

w,O

send

arp

etal

.In

fant

sof

mot

hers

taki

ngzi

ncsu

pple

men

tsha

dsi

gnifi

cant

lyfe

wer

epis

odes

ofac

ute

diar

rhea

(16%

risk

redu

ctio

n,R

R:

0.84

;CI:

0.72

–0.9

8;P

�.0

37),

dyse

nter

y(6

4%ri

skre

duct

ion;

RR

:0.3

6;C

I:0.

52–0

.84;

P�

.019

)an

dim

petig

o(4

7%ri

skre

duct

ion;

RR

:0.5

3;C

I:0.

34–0

.82;

P�

.005

).Im

pact

was

isol

ated

toLB

Win

fant

s.O

send

arp

etal

*232

Bang

lade

sh;u

rban

slum

;R

CT

Preg

nant

wom

enat

12–1

6w

kge

stat

ion

(n�

559)

wer

egi

ven

eith

er30

mg/

dof

zinc

(n�

269)

orpl

aceb

o(n

�29

0)un

tilde

liver

y.

Supp

lem

enta

tion

had

noef

fect

onm

ater

nalp

regn

ancy

wei

ght

gain

orm

idup

per

arm

circ

umfe

renc

e.Th

ere

was

noef

fect

ofth

esu

pple

men

ton

rate

sof

preg

nanc

ylo

ss.

No

sign

ifica

ntef

fect

oftr

eatm

ent

was

obse

rved

onbi

rth

wei

ght,

gest

atio

nala

ge,l

engt

h,an

dhe

ador

ches

tci

rcum

fere

nce.

No

diff

eren

ces

wer

eob

serv

edin

the

rate

ofLB

Wby

supp

lem

ent

grou

p(R

R:1

.12;

CI:

0.90

–1.4

1).

Cau

lfiel

det

al71

8Pe

ru;u

rban

hosp

ital

sett

ing;

RC

TPr

egna

ntw

omen

10–2

4w

kge

stat

ion

(n�

1295

)w

ere

rand

omly

assi

gned

tore

ceiv

eei

ther

iron

and

fola

teal

one

orir

on,f

olat

epl

us15

mg/

dof

zinc

.

No

sign

ifica

ntad

ditio

nale

ffec

tof

the

zinc

supp

lem

ent

was

seen

onbi

rth

wei

ght,

orLB

Wor

pret

erm

deliv

ery

rate

s.

Gar

get

al71

9In

dia;

urba

nho

spita

lse

ttin

g;R

CT

Wom

enin

the

inte

rven

tion

grou

p(n

�10

6)w

ere

give

n45

mg/

dof

zinc

supp

lem

ent;

cont

rols

(n�

62)

rece

ived

nosu

pple

men

ts.

Mat

erna

lsup

plem

enta

tion

sign

ifica

ntly

incr

ease

dbi

rth

wei

ght

ofin

fant

s(P

�.0

01);

the

degr

eeof

incr

ease

was

dire

ctly

rela

ted

toth

edu

ratio

nof

supp

lem

enta

tion.

Infa

nts

born

tom

othe

rssu

pple

men

ted

for

6–9

mo

wer

esi

gnifi

cant

lyhe

avie

r(3

.45

�0.

04kg

)th

anth

ose

born

follo

win

gsu

pple

men

tatio

nfo

r1–

3m

o(2

.98

�0.

07kg

);th

em

ean

birt

hw

eigh

tfo

rco

ntro

lsw

as2.

65kg

.Ges

tatio

nala

geof

supp

lem

ente

dba

bies

was

sign

ifica

ntly

high

erth

anco

ntro

lsw

hen

give

nfo

rm

ore

than

3m

o(P

�.0

1)an

dw

asre

late

dto

dura

tion

ofzi

ncsu

pple

men

tatio

n(P

�.0

5).

*D

ata

from

the

sam

etr

ial



in the WIC program was associated with reducedpreterm delivery and fetal death and small but sig-nificant increases in mean birth weight (�22.7 g) andhead circumference. Retrospective analyses of moth-ers with poor diets who had received supplementalmilk, eggs, and oranges revealed a higher mean birthweight compared with siblings (�107 g).243 Theseeffects were relatively greater in thinner mothers. Ina large Hungarian trial of micronutrient supplemen-tation,192 there was no impact on birth weight.

Data from developing countries are limited andrelate to disparate programs of varying sizes. Insome cases, the trial design does not allow for clearconclusions to be drawn regarding the contributionof multiple micronutrients to the impact of the inter-vention. Two large trials in Guatemala115 and Tai-wan108 evaluated the benefit to pregnant women ofincreased energy intake as well as multiple mineraland vitamin supplements. The trial in Guatemalaadministered 2 supplements, 1 caloric and 1 protein-caloric, to malnourished pregnant women duringpregnancy to examine effects of supplementation onbirth weight. Both supplements were fortified withcomparable amounts of micronutrients including vi-tamin C (4 mg), calcium (0.0–0.4 g), phosphorus(0.0–0.3 g), thiamine (1.1 mg), riboflavin (1.5 mg),vitamin A (18.5 mg), iron (�5 mg), and fluoride(0.2 mg). There was no difference in birth weightbetween the 2 intervention groups, but infants bornto the subgroup of mothers who had consumed�83 750 J of either supplement during pregnancyhad a correspondingly greater increment in birthweight. In contrast, the increment in birth weightwas comparable in Taiwan, where the supplementswere given to well-nourished women and were notsignificantly fortified with micronutrients.108 Al-though the results from the Guatemala trial havebeen attributed to increased energy intake, there mayalso have been an effect of the micronutrients presentin both supplements. However, because of the openstudy design, it is difficult to measure the attribut-able effects of micronutrient supplementation onbirth weight in this trial.

A small trial in a poor rural population of Thailandstudied the effect on birth weight of improving themothers’ dietary quality.110 Mothers (n � 43) wererandomized into 3 groups to receive either extra food(preprepared cooked food containing legumes, ses-ame, peanuts, and sugar; or rice, oil, peanuts,shrimp, and sugar) or no extra food. Newborn in-fants born to mothers who had received the food-based interventions were �250 g heavier than infantsborn to controls. In the Narangwal Study from In-dian Punjab,244 villages were assigned to 1 of 4groups, receiving (1) nutrition care, (2) health care,(3) both, or (4) neither. In nutrition villages, pregnantmothers received bulgar wheat porridge, sugar, milkpowder, and oil. Although birth weight was notrecorded, stillbirths were reduced by 40% comparedwith control villages.

community-based evidence. As indicated above,several community trials of multiple micronutrientsare currently underway. Preliminary results from a

community-based cluster-randomized trial of multi-ple micronutrient supplements in Pakistan suggest asignificant impact on birth weight (Z.A.B., unpub-lished observations, 2004).

The major published data available on efficacy ofmultiple micronutrient supplementation in develop-ing countries currently derive from a study in HIV-infected women in Tanzania220 and another trial inThe Gambia of high-energy biscuits that also con-tained extra micronutrients, calcium and iron104

(see Tables 6 and 12). The results of the Gambianstudy have been presented already (see “BalancedProtein-Energy Supplementation”). In Tanzania, af-ter micronutrient supplementation, the mean birthweight was higher among HIV-negative infants(�100 g; P � .01), and there were significant reduc-tions in rates of LBW (�44%), preterm births before34 weeks’ gestation (�39%), and IUGR (10% vs 18%;P � .002). Moreover, use of multiple micronutrientsand multivitamins resulted in significantly fewer fe-tal deaths (5.9% vs 9.6%; P � .02). Recent unpub-lished data from the same group indicate that mul-tiple RDAs of multivitamins and micronutrients maybe required in HIV-infected pregnant women (W.Fawzi, MD, DrPH, verbal communication, 2002).

In a trial in Chile,100 mothers received either pow-dered milk or milk fortified with vitamins and min-erals. Mean birth weight was higher in the fortifica-tion group (�73 g; P � .05), and the percentage ofIUGR infants was lower (32% vs 44%; P � .05). InBogota, Colombia, birth weight was greater (�95 g)in urban slum families who were randomly allocatedto receive extra food (milk, fortified bread, and veg-etable oil).245 Similarly, infants born to women inSouth Africa who received a supplement of maizeporridge and skim milk fortified with vitamins A, B1,B2, and calcium were an average of 300 g heavierthan infants born to unsupplemented women.246

However, in Mexico, other studies of multiple-micro-nutrient supplementation during pregnancy havefailed to demonstrate any benefit regarding in-creased birth weight250 or maternal hematologicparameters.251

conclusions. Although benefit from supplement-ing pregnant women with multiple vitamins andmicronutrients seems plausible, particularly due toincreased nutritional requirements during preg-nancy and lactation,94 there has been concern thatmultiple-micronutrient–supplement formulations de-signed for developed countries may be inappropriatefor women in developing countries with significantlydifferent staple diets.247 It is thus very important thatthese studies be undertaken with full attention topotential adverse outcomes. Recent data from Nepalon the use of multiple micronutrients suggest thatmultiple micronutrients may confer no additionaladvantage over iron-folate supplements but rathermay alter the distribution of birth weight, leadingto higher PMR and NMR due to obstructed la-bor.169,248,249 Current field trials of multiple micronu-trient supplementation in pregnancy will provideuseful data on efficacy, impact on LBW and otherperinatal outcomes, and potential interactions of mi-



cronutrients. Pending these data, however, the wide-spread use of these multiple micronutrients cannotbe recommended.

Infection Control and PreventionInterventions in Malaria-Endemic AreasMalaria continues to be a major health problem in

endemic countries and a matter of concern within the

global community. Globally, malaria affects almost10% of the world’s population, and of the nearly 500million cases, 1 million may die annually.252 In areasof high malarial transmission, chronic and repeatedmalaria infections greatly increase the risk of mater-nal anemia.253–257 In areas of low malarial transmis-sion, immunity is low, and malaria in pregnancy canrapidly progress to complications such as severe ane-

TABLE 12. Antenatal Multiple-Micronutrient Supplementation



Fawzi et al220 Tanzania; urbansetting; DBRCT

Women (n � 1075) wererandomized into 4 groups.Group 1 (n � 269) received adaily dose of vitamin A, group2 (n � 269) was givenmultivitamins excludingvitamin A, group 3 (n � 270)received multivitaminsincluding vitamin A, andgroup 4 (the control group, n� 267) was given placebo.

The multivitamin had abeneficial effect onHb levels and CD4cell counts.

A 40% decrease in fetal deaths wasseen in groups 2 and 3,(multivitamin-supplementedgroups) (RR: 0.61; CI: 0.39–0.94; P� .02). There was also a 43%decrease in the rate of SGA birthsin these groups (OR: 0.57; CI:0.39–0.82; P � .002). Multivitaminssignificantly decreased theincidence of LBW by 44% (OR:0.56; CI: 0.38–0.82; P � .003), andthe risk of VLBW by 58%.Although the risk of extremelypreterm births was reduced by39% in group 2, the overall effectof multivitamins on pretermbirths was not significant.

Mora et al245 See Table 6Ceesay et al104 See Table 6Mardones-

Santanderet al100

Chile; urbansetting; PCS

Pregnant women attending 9prenatal clinics were givenpowdered milk, powderedmilk with high milk-fatcontent, or multiple-micronutrient–fortifiedpowdered milk.Nonconsumers of anysupplement were controls.

Maternal iron statuswas significantlybetter in the fortifiedgroup than in thepowdered-milkgroup or the controlgroup.

Mean birth weight was higher(mean weight difference 73 g;P � 0.05) in the multiple-micronutrient–fortified groupthan in the group givenpowdered milk and even higherwhen compared to controls(mean weight difference 335 g;P � 0.05).

Ross et al246 South Africa;rural setting;RCT

Nutritionally deficient pregnantwomen �20 wk gestation (n �171) were randomized to 1 of4 groups receiving (1) amicronutrient-fortified high-bulk supplement (high inniacin and iron); (2) amicronutrient-fortified low-bulk supplement (high invitamin A, calcium, thiamine,and riboflavin; (3) a zincsupplement; or (4) placebo.

Mean birth weight was 9.5% greaterin the low-bulk fortified-supplement group comparedwith the high-bulk supplementgroup and 6.5% greater in thelow-bulk fortified-supplementgroup than the placebo group.There was no effect on length ofgestation.

Ramakrishnanet al251*

Mexico; ruralsetting; RCT

Pregnant women (n � 420) wererandomized to receive iron-only supplements (Fe) ormultiple-micronutrientsupplements containing iron(MM) to assess the effect onmaternal iron status.

Mean Hb (g/L) waslower for the MMgroup (104.2; 95% CI:102.5, 106.0)compared to the Fegroup (108.1; 95% CI:106.4, 109.8) afteradjusting for baselineserum ferritin.

Ramakrishnanet al250*

Mexico; ruralsetting; RCT

Pregnant women (n � 873) wererecruited before 13 wkgestation and receivedsupplements 6 d/wk at home,as well as routine antenatalcare, until delivery. Bothsupplements contained 60 mgof Fe, but the MM group alsoreceived 1–1.5 times therecommended dietaryallowances of severalmicronutrients.

Mean birth weight and birth lengthdid not significantly differbetween the 2 groups.




TA

BL

E13

.A

nten

atal

Mal

aria

Che

mop

roph

ylax

isor

IPT

Sour

ceL

ocat

ion

and

Typ

eof

Tri

alIn

terv

enti

onM

ater

nal

Out

com

ePe

rina

tal/

Neo

nata

lO

utco

me

Wol

feet

al28

3K

enya

;per

iurb

anse

ttin

g;co

st-e

ffec

tiven

ess

eval

uatio

nof

trea

tmen

tre

gim

en

Usi

ngex

istin

gda

tafr

omw

este

rnK

enya

(not

anin

terv

entio

nst

udy,

now

omen

enro

lled)

,art

icle

com

pare

d4

stra

tegi

esof

mal

aria

prop

hyla

xis

usin

gSP

for

effe

ctiv

enes

san

dco

st-e

ffec

tiven

ess,

incl

udin

g2-

dose

SP(1

500

mg

sulfa

doxi

ne,7

5m

gpy

rim

etha

min

e),m

onth

lySP

,HIV

test

ing

and

anSP

regi

men

,and

aca

sem

anag

emen

tap

proa

chin

clud

ing

SP.

The

mon

thly

SPre

gim

enw

asth

em

ost

effe

ctiv

est

rate

gyfo

rre

duci

ngLB

Was

soci

ated

with

mal

aria

.The

2-do

seSP

and

mon

thly

SPre

gim

ens

wou

ldpr

even

t17

2an

d22

9m

ore

LBW

birt

hsou

tof

aco

hort

of10

000

preg

nant

wom

en,r

espe

ctiv

ely,

com

pare

dw

ithth

eca

sem

anag

emen

tap

proa

ch.A

tH

IVse

ropr

eval

ence

rate

s�

10%

,the

mon

thly

SPre

gim

enis

the

mos

tco

st-

effe

ctiv

est

rate

gy.A

tH

IVse

ropr

eval

ence

rate

s�

10%

,the

2-do

seSP

regi

men

wou

ldbe

the

mos

tco

st-

effe

ctiv

eop

tion.

Ndy

omug

yeny

iet

al27

9U

gand

a;ru

rals

ettin

g;D

BRC

TPr

imig

ravi

dae

inth

eir

first

and

seco

ndtr

imes

ter

(n�

860)

wer

era

ndom

ized

to:g

roup

A(n

�28

4),o

ral3

00m

gC

Qba

se,p

lace

boir

onan

dpl

aceb

ofo

late

;gro

upB

(n�

282)

,pla

cebo

ofal

l3in

terv

entio

ns;g

roup

C(n

�29

4),w

eekl

yor

alel

emen

tali

ron

120

mg/

dan

dfo

licac

id5

mg,

plac

ebo

CQ

.

The

risk

ofbe

ing

anem

icat

deliv

ery

was

low

erin

grou

pA

(25.

6%)

and

grou

pC

(24.

7%)

com

pare

dto

grou

pB

(38.

1%)

(P�

.01)

.

Wom

enin

grou

pA

had

asi

gnifi

cant

lylo

wer

freq

uenc

yof

LBW

(2%

)th

anw

omen

ingr

oup

B(9

%)

(P�

.009

).Th

ere

was

nosi

gnifi

cant

diff

eren

cein

the

freq

uenc

yof

LBW

betw

een

grou

psB

and

C,o

rgr

oups

Aan

dC

.

Shul

man

etal

274

Ken

ya;r

ural

sett

ing;

DBR

CT

Wom

enpr

esen

ting

betw

een

16–3

0w

kof

gest

atio

n(n

�12

64)

wer

egi

ven

1,2,

or3

dose

sof

SP,

depe

ndin

gon

stag

eof

preg

nanc

y(n

�64

0).

Ano

ther

grou

p(n

�62

4)w

asgi

ven

plac

ebo.

Low

erra

tes

ofse

vere

anem

ia(�

8g/

dL)

wer

ese

enin

the

inte

rven

tion

grou

p,w

ithan

over

all

prot

ectiv

eef

ficac

yof

39%

(P�

.000

1).A

prot

ectiv

eef

fect

ofin

terv

entio

nw

asal

sose

enfo

rm

alar

ialp

aras

item

ia(8

5%,P

�.0

001)

and

plac

enta

lpar

asite

infe

ctio

n(1

7%,P

�.0

27).

Alth

ough

the

inte

rven

tion

decr

ease

dth

est

illbi

rth

rate

(22%

),th

ere

sults

did

not

reac

hsi

gnifi

canc

e.Th

ere

was

noef

fect

onpr

emat

ure

birt

hra

tes.

The

inte

rven

tion

decr

ease

dN

MR

by38

%,b

utth

isre

sult

was

not

sign

ifica

nt;s

imila

rly,

the

PMR

decr

ease

d32

%,b

utth

isre

sult

was

not

sign

ifica

nt.

Pari

seet

al28

0K

enya

;per

iurb

anse

ttin

g;R

CT

2077

preg

nant

wom

enw

ere

enro

lled

inth

est

udy.

The

Cas

eM

anag

emen

t(C

M)

grou

p(n

�73

6)re

ceiv

edm

edic

atio

non

lyup

onpr

esen

tatio

nw

ithfe

ver

and

para

site

mia

.The

2-do

segr

oup

(n�

680)

rece

ived

SPat

enro

llmen

tan

dag

ain

inth

eth

ird

trim

este

r.Th

eth

ird

grou

p(n

�66

1)re

ceiv

edSP

mon

thly

.

Com

pare

dto

the

CM

grou

p(2

7%),

the

2-do

se(1

2%)

and

mon

thly

trea

tmen

tgr

oups

(9%

)ha

dlo

wer

rate

sof

plac

enta

lmal

aria

infe

ctio

n(P

�.0

01).

The

mon

thly

regi

men

was

mor

eef

fect

ive

inde

crea

sing

thir

dtr

imes

ter

anem

iara

tes

and

also

inde

crea

sing

plac

enta

lin

fect

ion

amon

gH

IV-p

ositi

vew

omen

(P�

.002

).

No

diff

eren

ces

betw

een

grou

psw

ere

seen

inra

tes

ofst

illbi

rths

,pre

mat

ure

birt

hs,s

pont

aneo

usab

ortio

ns,o

rne

onat

alm

orta

lity.

The

still

birt

hra

tew

as2.

1%in

the

CM

grou

p,2.

6%in

the

2-do

segr

oup,

and

37%

inth

em

onth

lygr

oup.

The

NM

Rw

as1.

2%in

the

CM

grou

p,1.

3%in

the

2-do

segr

oup,

and

0.3%

inth

em

onth

lygr

oup.

Birt

hw

eigh

tw

assi

gnifi

cant

lyin

crea

sed

inth

e2-

dose

SPan

dth

em

onth

lySP

grou

ps(m

ean

birt

hw

eigh

t30

79�

585

gan

d31

98�

528

g,re

spec

tivel

y)co

mpa

red

toth

eC

Mgr

oup

(318

3�

534

g).

Ver

hoef

fet

al28

2M

alaw

i;ru

rals

ettin

g;Q

TA

llpr

egna

ntw

omen

(n�

525)

rece

ived

SP(1

500

mg

sulfa

doxi

ne,7

5m

gpy

rim

etha

min

e)at

the

time

ofpr

egna

ncy

regi

stra

tion

and

betw

een

wee

ks28

–34

ofge

stat

ion;

som

ere

ceiv

edan

addi

tiona

ldos

eif

they

deve

lope

dm

alar

ia.

Thos

ew

hore

ceiv

ed2

or3

dose

sde

liver

edsi

gnifi

cant

lyhe

avie

rba

bies

(mea

ndi

ffer

ence

257

g,P

�.0

1)th

anth

ose

who

rece

ived

only

1do

se.T

hein

cide

nce

ofLB

Win

prim

igra

vida

ean

dm

ultig

ravi

dae

who

had

been

give

n2

dose

sw

asha

lfth

atse

enin

subj

ects

who

had

been

give

n1

dose

(33.

9%vs

13.5

%,P

�.0

09,a

nd13

.9%

vs6.

5%,P

�.0

2,re

spec

tivel

y).

Bouv

ier

etal

281

Mal

awi;

rura

lset

ting;

RC

TIn

the

obse

rvat

iona

lpha

se,w

omen

(n�

126)

wer

egi

ven

mal

aria

chem

opro

phyl

axis

asC

Q30

0m

g/w

kin

asi

ngle

dose

alon

gw

ithir

on.I

nth

ein

terv

entio

nph

ase

the

drug

regi

men

was

com

bine

dPR

OG

200

mg/

dan

dC

Q30

0m

g/w

k.

Inth

eob

serv

atio

nalp

hase

,he

mat

ocri

tde

crea

sed

duri

ngth

era

iny

seas

on,w

here

asin

the

inte

rven

tion

phas

e,he

mat

ocri

tin

crea

sed

rega

rdle

ssof

seas

on.

An

annu

alcy

cle

ofbi

rth

wei

ght

was

obse

rved

inth

eob

serv

atio

nph

ase,

whe

reas

nose

ason

alva

riat

ion

inbi

rth

wei

ght

was

obse

rved

inth

ein

terv

entio

nph

ase.



TA

BL

E13

.C

onti

nued

Sour

ceL

ocat

ion

and

Typ

eof

Tri

alIn

terv

enti

onM

ater

nal

Out

com

ePe

rina

tal/

Neo

nata

lO

utco

me

Stek

etee

etal

258

Mal

awi;

rura

lse

ttin

g;R

CT

1766

preg

nant

wom

enw

ere

enro

lled

:gro

upA

rece

ived

CQ

25m

g/kg

div

ided

over

2d

,fo

llow

edby

300

mg

wee

kly;

grou

pB

rece

ived

CQ

25m

g/kg

div

ided

over

2d

and

repe

ated

ever

y4

wk;

grou

pC

rece

ived

CQ

300

mg

ofba

sew

eekl

y;an

dgr

oup

Dre

ceiv

eda

sing

lem

eflo

quin

e(M

Q)

dos

eof

750

mg

follo

wed

by25

0m

gw

eekl

y.

The

mul

tiva

rian

cem

odel

sugg

este

dth

atim

prov

emen

tof

birt

hw

eigh

tam

ong

offs

prin

gof

wom

enus

ing

MQ

was

dir

ectl

yre

late

dto

the

dru

g’s

effe

cton

clea

ranc

eof

plac

enta

lin

fect

ion.

MQ

use

was

asi

gnif

ican

tpr

otec

tive

fact

orfo

rL

BW

(OR

:1.4

;CI:

1.03

–1.9

;P�

.03)

.In

cont

rast

,po

orly

effe

ctiv

em

alar

iapr

even

tion

(ie,

CQ

use)

was

sign

ific

antl

yas

soci

ated

wit

hIU

GR

-LB

W(O

R:1

.63;

CI:

1.16

–2.2

9).P

ropo

rtio

nof

LB

Win

MQ

trea

ted

mot

hers

was

12.5

%vs

15.5

%in

thos

etr

eate

dw

ith

CQ

(P�

.05)

.

Cot

etal

284

Cam

eroo

n;ru

ral

and

peri

urba

nse

ttin

g;R

CT

266

preg

nant

wom

enw

ere

enro

lled

.The

CQ

grou

p(n

�13

6)re

ceiv

eda

300

mg

oral

dos

ew

eekl

yun

til

del

iver

y.T

heco

ntro

ls(n

�13

5)w

ere

untr

eate

d.

CQ

prop

hyla

xis

prov

ided

a32

%pr

otec

tive

effe

ctag

ains

tm

alar

ial

plac

enta

lin

fect

ion

(OR

:0.6

8;C

I:0.

46–0

.99,

P�

.043

).

Infa

nts

born

tom

othe

rsw

hore

ceiv

edC

Qpr

ophy

laxi

sw

ere

sign

ific

antl

yhe

avie

r(m

ean

dif

fere

nce:

207.

5g;

P�

.02)

than

the

infa

nts

born

toun

trea

ted

mot

hers

.The

rew

asa

62%

dec

reas

ein

the

LB

Wra

ted

ueto

the

inte

rven

tion

(OR

:0.

38;C

I:0.

16–0

.89;

P�

.017

).G

reen

woo

det

al27

6T

heG

ambi

a;ru

ral

sett

ing;

RPC

TD

ata

wer

eco

llect

edfr

ompr

egna

ntw

omen

(n�

406)

dur

ing

the

cour

seof

2co

ntro

lled

tria

lsof

chem

opro

phyl

axis

wit

hM

alop

rim

(pyr

imet

ham

ine

25m

gan

dd

apso

ne10

0m

g)ad

min

iste

red

ever

y2

wk

afte

rre

gist

rati

onw

ith

aT

BA

(n�

200)

.The

cont

rol

grou

p(n

�20

6)w

asun

trea

ted

.

No

dif

fere

nce

was

obse

rved

inm

ater

nal

mor

talit

y(5

/20

6in

the

Mal

opri

mgr

oup

vs4/

200

amon

gth

ose

rece

ivin

gpl

aceb

o).

Still

birt

hra

tes

amon

gof

fspr

ing

ofw

omen

who

rece

ived

chem

opro

phyl

axis

wer

eab

out

one-

half

that

ofof

fspr

ing

ofw

omen

who

rece

ived

plac

ebo

(62/

1000

vs11

6/10

00),

alth

ough

the

dif

fere

nce

was

not

stat

isti

cally

sign

ific

ant.

PMR

and

NM

Rd

ecre

ased

by33

%an

d14

%,r

espe

ctiv

ely,

inth

ein

terv

enti

ongr

oup.

Men

end

ezet

al15

5T

heG

ambi

a;ru

ral

sett

ing;

DB

RC

TPr

imig

ravi

dw

omen

(n�

230,

enro

lled

over

3y)

wer

era

ndom

ized

tore

ceiv

ew

eekl

yei

ther

1ta

blet

ofM

alop

rim

(pyr

imet

ham

ine1

2.5

mg

and

dap

sone

100

mg)

orpl

aceb

o.

Wom

enin

the

inte

rven

tion

grou

pha

d86

%le

sspa

rasi

tem

iaco

mpa

red

toth

epl

aceb

ogr

oup

(P�

.01)

.

Wom

enta

king

Mal

opri

mga

vebi

rth

tohe

avie

rba

bies

(mea

nd

iffe

renc

e:15

3g;

P�

.02;

�2 1

df�

1.2)

than

the

plac

ebo.

The

rew

asno

effe

cton

the

still

birt

hra

te.

Schu

ltz

etal

287

Mal

awi;

rura

lse

ttin

g;R

CT

357

preg

nant

wom

enw

ere

recr

uite

dfo

rth

etr

ial.

The

CQ

/C

Qgr

oup

(n�

104)

rece

ived

anin

itia

ltr

eatm

ent

dos

eof

CQ

(25

mg

base

/kg

)fo

llow

edby

CQ

(300

mg

base

)w

eekl

y.T

heSP

/C

Qgr

oup

(n�

117)

rece

ived

anin

itia

ltr

eatm

ent

dos

eof

SP(1

500

mg

sulf

adox

ine,

75m

gpy

rim

etha

min

e)fo

llow

edby

CQ

(300

mg

base

)w

eekl

yun

til

del

iver

y.T

heSP

/SP

grou

p(n

�13

6)re

ceiv

edan

init

ial

trea

tmen

td

ose

ofSP

wit

ha

seco

ndd

ose

atth

ebe

ginn

ing

ofth

eth

ird

trim

este

r.

The

SP/

SPre

gim

enw

asth

em

ost

effe

ctiv

ean

dco

nsis

tent

ind

ecre

asin

gpa

rasi

tem

iain

preg

nant

wom

enby

83%

.The

SP/

SPre

gim

enw

asal

soth

em

ost

effe

ctiv

ein

dec

reas

ing

plac

enta

lin

fect

ion

(32%

,26%

and

9%in

the

CQ

/C

Q,S

P/C

QSP

/SP

grou

ps,

resp

ecti

vely

,P�

.006

).

The

rew

ere

nosi

gnif

ican

td

iffe

renc

esin

birt

hw

eigh

t,L

BW

orpr

emat

urit

yra

tes

amon

gth

e3

regi

men

s.

Mut

abin

gwa

etal

278

Tan

zani

a;ru

ral

sett

ing;

RC

T31

2w

omen

wer

een

rolle

din

the

stud

yan

dra

ndom

lyas

sign

edto

1of

3pr

ophy

laxi

sre

gim

ens

wer

eem

ploy

ed.T

heC

Qgr

oup

(n�

107)

rece

ived

CQ

300

mg

once

wee

kly,

and

the

PRO

Ggr

oup

(n�

116)

rece

ived

PRO

G20

0m

gd

aily

.The

CQ

�PR

OG

grou

p(n

�89

)re

ceiv

eda

com

bina

tion

ofC

Qan

dPR

OG

inth

eab

ove

dos

ages

.

The

mea

nH

bof

prim

igra

vid

aew

ashi

ghes

tin

the

PRO

Ggr

oup

(10

�1.

6g/

dL

afte

r2

mo

ofbe

ginn

ing

prop

hyla

xis)

and

low

est

inth

eC

Qgr

oup

(9.4

�1.

6g/

dL

).

Prim

igra

vid

aein

the

CQ

�PR

OG

grou

pd

eliv

ered

heav

ier

infa

nts

(mea

nw

eigh

t:2.

89�

0.52

kg)

than

prim

igra

vid

aein

the

PRO

Ggr

oup

(2.7

9�

0.42

kg)

orC

Qgr

oup

(2.7

1�

0.34

),re

spec

tive

ly.T

heC

Qgr

oup

had

the

high

est

freq

uenc

yof

LB

W(2

1%)

com

pare

dto

12%

for

the

PRO

Ggr

oup

and

15%

for

the

CQ

�PR

OG

grou

p.



TA

BL

E13

.C

onti

nued

Sour

ceL

ocat

ion

and

Typ

eof

Tri

alIn

terv

enti

onM

ater

nal

Out

com

ePe

rina

tal/

Neo

nata

lO

utco

me

Nyi

rjes

yet

al27

7Z

aire

;rur

alse

ttin

g;PC

SSt

udy

com

pare

dpr

egna

ntfe

mal

espr

esen

ting

for

intr

apar

tum

man

agem

ent

(n�

302)

who

took

CQ

pren

atal

lyve

rsus

thos

ew

hod

idno

tta

keC

Q.

CQ

prop

hyla

xis

prev

ente

dm

ater

nal

mal

aria

(RR

:0.4

;CI:

0.16

–1.0

).C

Qpr

ophy

laxi

spr

even

ted

feta

lm

alar

ia(R

R:0

.2;

CI:

0.09

–0.4

4),s

igni

fica

ntly

dec

reas

edth

eri

skof

peri

nata

lm

orta

lity

by62

%(O

R:0

.38;

CI:

0.16

–0.

80)

and

red

uced

LB

Wby

61%

.C

otet

al72

0W

est

Afr

ica;

urba

nse

ttin

g;R

CT

Wom

enin

the

trea

tmen

tgr

oup

(n�

745)

wer

egi

ven

CQ

300

mg

once

wee

kly;

cont

rols

(n�

719)

wer

egi

ven

nom

alar

ial

prop

hyla

xis.

The

RR

ofpl

acen

tal

infe

ctio

nin

the

trea

ted

grou

pd

ecre

ased

wit

hin

crea

sing

dur

atio

nof

prop

hyla

xis;

wom

entr

eate

d�

12w

kha

dan

RR

of0.

14(C

I:0.

06–

0.3;

P�

.000

04).

The

rew

asno

effe

ctof

trea

tmen

ton

birt

hw

eigh

tan

dL

BW

rate

s.

Gre

enw

ood

etal

275

The

Gam

bia;

rura

lse

ttin

g;R

PCT

Onc

ea

wom

anre

port

edto

the

TB

Ath

atsh

ew

aspr

egna

nt,s

hew

asal

loca

ted

tore

ceiv

e1

tabl

etof

Mal

opri

m(p

yrim

etha

min

e25

mg

and

dap

sone

100

mg)

ever

y2

wk

orpl

aceb

o.12

08pr

egna

ntw

omen

part

icip

ated

inth

est

udy,

and

1049

rece

ived

Mal

opri

mfr

oma

TB

A.

Lev

elof

para

site

mia

was

low

er,

19%

vs38

%,i

npr

imig

ravi

dae

and

8%vs

20%

amon

gm

ulti

grav

idae

rece

ivin

gch

emop

roph

ylax

isw

ith

Mal

opri

mco

mpa

red

topl

aceb

o.T

hem

edic

atio

nal

sow

asef

fect

ive

inin

crea

sing

the

mea

nhe

mat

ocri

t,fr

om26

.6to

30.1

g/d

L(P

�.0

5)in

prim

igra

vid

wom

en.

Com

pare

dto

wom

enw

hod

idno

tre

port

toth

eT

BA

and

wer

eno

tgi

ven

med

icat

ion

orpl

aceb

o,th

ew

omen

who

rece

ived

Mal

opri

mha

da

69%

low

erst

illbi

rth

rate

.The

red

ucti

onw

asle

ssm

arke

dw

hen

the

Mal

opri

mgr

oup

was

com

pare

dto

plac

ebo

(11%

low

erst

illbi

rth

rate

).E

arly

NM

Rw

as25

%an

d32

%fo

rth

eM

alop

rim

and

plac

ebo

grou

ps,r

espe

ctiv

ely.

Lat

eN

MR

was

10%

and

13%

for

the

Mal

opri

man

dpl

aceb

ogr

oups

,res

pect

ivel

y.W

omen

usin

gM

alop

rim

del

iver

edhe

avie

rba

bies

(mea

nd

iffe

renc

ein

birt

hw

eigh

tam

ong

trea

ted

prim

igra

vid

ae15

9g;

CI:

8–31

0g)

.The

rew

asa

72%

red

ucti

onin

the

LB

Wra

team

ong

the

offs

prin

gof

Mal

opri

m-

trea

ted

wom

en(P

�.0

5).

van

Eijk

etal

286

Ken

ya;r

ural

hosp

ital

sett

ing;

PCS

Dat

aon

freq

uenc

yof

IPT

usin

gSP

and

birt

hou

tcom

es(v

ital

stat

us,b

irth

wei

ght,

gend

er,

and

pres

ence

ofco

ngen

ital

abno

rmal

itie

s)w

ere

colle

cted

over

a12

-mo

peri

odfo

r23

02co

nsec

utiv

ed

eliv

erie

sin

apr

ovin

cial

hosp

ital

inw

este

rnK

enya

.

IPT

(�1

dos

eof

SP)

was

asso

ciat

edw

ith

are

duc

tion

inpl

acen

tal

mal

aria

(OR

:0.5

6;C

I:0.

39–0

.95)

.T

here

was

asu

bsta

ntia

lre

duc

tion

inpa

rasi

tem

iaam

ong

mos

tw

omen

who

had

rece

ived

IPT

;ho

wev

er,w

omen

who

had

rece

ived

IPT

but

rem

aine

dpa

rasi

tem

icha

dsi

mila

rpa

rasi

ted

ensi

tyto

wom

enw

hoha

dne

ver

rece

ived

IPT

.

IPT

(�1

dos

eof

SP)

was

asso

ciat

edw

ith

are

duc

tion

inL

BW

(OR

:0.6

5;C

I:0.

45–0

.95)

.1d

ose

ofIP

Tw

asas

soci

ated

wit

ha

mea

nin

crea

sein

birt

hw

eigh

tof

54g

(CI:

12–1

20g;

P�

.11)

.�

2d

oses

ofIP

Tw

ere

asso

ciat

edw

ith

am

ean

incr

ease

inbi

rth

wei

ght

of12

8g

(CI:

42–2

13g;

P�

.004

).



mia, cerebral malaria, and death.258–262 Primigravi-dae also tend to have a much higher prevalence anddensity of parasitemia than both nonpregnantwomen and multigravidae.256,263 Where malaria isendemic, women become more susceptible to infec-tion during pregnancy, but this susceptibility de-creases with successive pregnancies.256,264 Given thepropensity of parasitized red blood cells to sequesterin the placenta, malaria is associated with maternalanemia, increased risk of preterm birth, LBW, andneonatal mortality.265–268 The estimated populationattributable risk of LBW among primigravidae withmalaria is 10% to 40%.269 Recent data also suggestthat HIV infection may impair the ability to acquirepregnancy-specific immunity, thus increasing thelikelihood of complications of malaria among HIV-positive multigravidae in endemic areas.258–262,270

Malaria Chemoprophylaxis or IPTbackground. Although malaria chemoprophy-

laxis with chloroquine (CQ) has been the main strat-egy for management of malaria in endemic areas,there is uncertainty about its efficacy in comparisonwith IPT. A review of 15 trials by the Cochranecollaboration271 revealed that women who weregiven regular, routine antimalarial drugs had lessrisk of developing severe anemia and had fewerepisodes of fever antenatally than those who did notreceive prophylactic therapy. Newborn infants ofthese women also had higher birth weight comparedwith those born to women who did not receive pro-phylaxis; this effect was principally seen among pri-migravidae. However, there was no overall benefit ofantimalarial chemoprophylaxis on perinatal andneonatal mortality.

Several studies indicated that IPT can be givenunder directly observed therapy in antenatal clinicprograms and achieve high program effectiveness.272

Additional support for the effectiveness of IPT wasgiven in a meta-analysis by Newman et al,273 whichfound that the 7 most promising drug regimens formalaria prevention in terms of program effectivenessand efficacy, were all IPT as opposed to chemopro-phylaxis because of their ease of delivery and lowercost. IPT with sulfadoxine-pyrimethamine (SP) inareas without SP-resistant strains was ranked highestin terms of effectiveness due to the low cost, wideavailability, easy deliverability, and acceptability ofSP.

community-based evidence. A variety of antena-tal antimalarial chemotherapy drugs (eg, SP, CQ,proguanil [PROG], and Maloprim [pyrimethamine-dapsone]) given for prophylaxis have been evaluatedin developing countries (Table 13). Decreases in ratesof stillbirths and perinatal, neonatal, and/or infantmortality were observed in some instances,274–277 butoverall, the impact on these outcomes was not sig-nificant. The preterm birth rate also was not im-pacted. An increase in birth weight, however, wasobserved in nearly all studies155,258–262,275,277–284;thus, the evidence for a beneficial effect of chemo-prophylaxis on birth weight in endemic areas isstrong. Chemoprophylaxis also had a uniformly sig-

nificant effect in reducing maternal anemia, para-sitemia, and/or placental parasite load.

Studies comparing various drug regimens re-vealed that mefloquine was of particular benefit inclearance of parasitemia and in some neonatal healthoutcomes (eg, reduction in LBW rate) in areas withhigh rates of CQ resistance.258–262,285 Another studyby van Eijk286 in Kenya found IPT (at least 1 dose ofSP) during pregnancy was associated with a signifi-cant reduction in the LBW rate (OR: 0.65; CI:0.45–0.95).

A growing body of research evaluates the cost-effectiveness of different drug prophylaxis/IPT reg-imens. In Malawi, Schultz et al287 found that �2doses of trimethroprim/sulfamethoxazole duringpregnancy were equally effective as 1 dose. Anotherstudy evaluated the cost-effectiveness of 3 ap-proaches to malaria chemoprophylaxis comparedwith case management of malarial disease in areaswith variable HIV prevalence and high malaria en-demicity.283 In malaria-endemic areas with high HIVprevalence (�10%), administering 1 dose of SP pro-phylaxis per month was more cost-effective, whereasin malaria-endemic areas with low HIV prevalence(�10%), administering 2 doses of SP prophylaxisduring pregnancy was more cost-effective. In anysetting, the monthly and the 2-dose strategies weremore cost-effective than the case managementapproach.

conclusions. Malaria prophylaxis during preg-nancy is an important intervention in communitysettings in malaria-endemic areas, because it servesto reduce maternal anemia and parasitemia and im-prove birth weight. In general, however, little im-provement in survival of offspring (eg, stillbirths,PMRs, and NMRs) has been demonstrated, althoughmost studies lacked sufficient power to evaluatemortality outcomes. Case management strategiesalone are less effective than continuous or intermit-tent prophylaxis, and in areas with high rates of CQresistance, alternative agents such as mefloquine aresuperior.

There is a real need to undertake concerted re-search in this area by means of large-scale effective-ness trials of various prophylaxis strategies for ma-laria, particularly IPT. Pending evidence to thecontrary or for alternative strategies, the current pol-icies of antimalarial prophylaxis in high-risk popu-lations must continue.

Malaria Prevention Using ITNsbackground. Consistent use of bed nets impreg-

nated with permethrin, an insecticide, reduces thefrequency of bites from infected mosquitoes and canconsequently reduce rates of malarial infection andparasitemia. A meta-analysis of RCTs evaluatingITNs showed that when ITNs were compared withuntreated nets or no nets at all, the efficacy for pre-vention of moderate to severe malarial infection was17%, and child mortality was reduced 23%.271 Onaverage, 6 lives were saved annually for every 1000children protected with ITNs. ITNs also reduced theincidence of mild malarial episodes by 48% and 34%



compared with no net and untreated bed net (NIB)controls, respectively.

There is some evidence that the widespread use ofITNs, particularly in high-transmission, malaria-en-demic areas, may reduce rates of maternal anemiaand placental infection, consequently reducing therisk of adverse birth outcomes such as stillbirth andLBW due to prematurity and IUGR.288 There arealso significant data showing that ITNs can reducechild morbidity and mortality, particularly in infants.In a high-transmission area of western Kenya, anRCT by Hawley et al289 reported that ITNs reducedall-cause postneonatal mortality (deaths of infants

between 1 and 11 months old) by 23%. However, nosingle research study has had sufficient size to ob-serve a specific significant impact of ITNs on neona-tal mortality; benefits of ITNs on neonatal mortalityhave been largely inferred from pregnancy outcomedata.290

community-based evidence. Table 14 summa-rizes the major studies evaluating efficacy of ITN useamong pregnant women in malaria-endemic areas.Overall, use of bed nets was effective,272,291 althoughnot uniformly so,292 in reducing rates of maternalanemia. In Kenya, the incidence of severe maternal

TABLE 14. Antenatal Malaria Prevention Using ITNs


Intervention Maternal Outcome Perinatal/NeonatalOutcome

Shulman290 Kenya; rural setting;RCT

503 pregnant women and theirhouseholds participated inthe study. In theintervention households,nets measuring 190 � 180 �150 cm were impregnatedwith permethrin to achievea target dose of 0.5 gpermethrin per m2 ofnetting.

No significant effect of theimpregnated netintervention wasobserved in maternal Hblevels or rates of anemiaor placental infection.

No significant effect ofthe intervention wasseen on stillbirth rate(RR: 0.67; CI: 0.19–2.32; P � .51). Nosignificant effect onPMR or birth weightwas seen.

D’Alessandroet al272

The Gambia; ruralsetting; RCT

651 primigravidae lived instudy villages; 308 lived inintervention villages, whichwere provided withinsecticide-treated bed nets(20% permethrin). 343women lived in controlvillages without bed nets.

The prevalence of severeanemia was significantlylower (73% reduction) invillages with treated bednets compared withcontrol villages, but thiseffect was limited to thedry season.

No significantdifferences betweenthe 2 village groupswere seen for PMR orbirth weight. Thepreterm birth rate waslower (72% reduction)in villages withtreated bed netsduring the rainyseason (OR: 0.28; CI:0.08–0.97; P � 0.02).

Dolan et al291 Thailand; ruralsetting; DBRCT

348 pregnant women wererandomly assigned to 1 of 4groups: 1) 111 women to thepermethrin-impregnated(500 mg/m2) bed net (PIB)group, 2) 112 women to theNIB group, 3) 30 women tothe no bed net group, and 4)88 women in the family-sizenonimpregnated bed net(FNIB) (These 88 womenwere originally assigned tothe no bed net group, butobtained family-size bednets and were reclassified.)

Women who used theirown FNIB or a PIB hadfewer episodes of malariathan the combined NIBand no-net groups (PIB33%, P � .04; FNIBgroup 32%, P � .07).Thus, the risk of at least1 attack of malaria was1.67 (CI: 1.07–2.61) timeshigher in the NIB andno-net group than in thepooled PIB and FNIBgroup (P � .03 allowingfor parity). Use of PIBswas associated with areduction in anemia inall camps and in womenof all gravidae,independent of the effecton parasitemic malaria.

No significant effect ofthe nets was observedon IMR, birth weightor gestational age atbirth.

ter Kuileet al288

Kenya; rural setting;RCT

All households in 40 of 79villages were randomized toreceive ITNs. Women werefollowed monthly throughpregnancy to monitor healthand birth outcomes.

No significant effect ofthe ITNs was foundon preterm birth rates.LBW was loweramong ITN users thancontrols (protectiveefficacy: 28%; CI: 2–47%). Adjusted meanbirth weight was 77.6g higher during first 4pregnancies of womenin ITN villages (P �.0008).



anemia (Hb �8 g/dL) was slightly lower in thebed-net group (15% vs 20%), but the overall inci-dence of anemia was the same (92% in the bed-netgroup and 91% in the control group).292

The impact of ITNs on pregnancy outcomes wasunclear, but overall, the results were not significant.The ITN intervention showed a trend toward re-duced stillbirth rates and PMRs in Kenya,292 whereasa study in The Gambia failed to show an impact onperinatal mortality.272 This latter study, however,revealed a decrease in premature birth rate and acorresponding increase in birth weight, althoughonly among primigravid women.272 Another studyin Thailand showed no impact of ITN use on gesta-tional age at birth,291 and overall, no study showed asignificant impact on birth weight.272,291,292 The im-pact of the intervention on IMR also was mixed.291,293

conclusions. There is strong evidence for the ef-ficacy of ITNs in reducing childhood mortality andmorbidity from malaria. There is encouraging evi-dence for improved pregnancy outcomes and re-duced perinatal or neonatal mortality from represen-tative settings in developing countries. Some of thevariability in data may be a result of the considerableheterogeneity in the treatment groups. Although theoverall effect of this intervention shows promisingtrends, particularly in reducing rates of maternalanemia, the number of well-designed studies withthe requisite power to assess outcomes of interest islimited. Cost-effectiveness studies and operations re-search are needed to help make this interventionmore feasible at scale. Currently, we are able to rec-ommend both chemoprophylaxis and PIB use inpregnancy in malaria-endemic areas, which, coupledwith the recommendation for use in childhood,makes the case for family-centered approaches.10

Dewormingbackground. A conservative estimate by the

WHO suggests that at any given time, nearly 44million pregnant women globally may be infectedwith hookworms.294 In endemic areas, hookworminfestation is known to be a major contributing factorin development of anemia in women of reproductiveage.295

Although the current recommendation duringpregnancy is to use a single-dose regimen of eithermebendazole or albendazole for treating hookworminfestation,296 in combination with iron-folate sup-plements, there have been almost no systematic stud-ies of the impact of this intervention on pregnancyoutcomes.

community-based evidence. Our review re-vealed little community-based information on theimpact of maternal deworming. A recentintervention trial in rural India297 used a combina-tion strategy with iron supplementation, de-worming, and nutrition education using speciallydeveloped information, education, and communica-tion materials (Table 15). Significant decreases in theprevalence of maternal iron-deficiency anemia andincreases in mean Hb levels were seen and weregreater the earlier the intervention was introduced inpregnancy.

Despite the lack of community-based data, therewere a few smaller studies on the impact of deworm-ing in developing-country settings. In a nonrandom-ized study in Sri Lanka167 women who received de-worming along with iron folate as part of theirantenatal care had better Hb and higher birth weightinfants than women who did not. In Bangladesh, arandomized trial using a 2 � 2 factorial design testedthe effects of iron and antihelminthic treatment onHb and wormload in female tea-plantation work-ers.298 The group receiving both iron supplementsand antihelminthics had the largest increments in Hbconcentration compared with the controls. Preva-lence and intensity of Ascaris, Trichuris trichiura, andhookworm infestation declined in the 2 groups thatreceived antihelminthic treatment.

The only studies of the impact of deworming onneonates undertaken in developing countries werein Nepal and Sri Lanka299 (Table 15). This retro-spective observational study in an urban hospitalsetting in Sri Lanka found that antihelminthic ther-apy (mebendazole) was associated with significantlylower stillbirth rates and PMRs (OR: 0.55; CI: 0.4–0.77). Taking medication was also associated with adecrease in the proportion of women delivering aVLBW infant (1.1% vs 2.3%). Women who took me-bendazole during the first trimester, however, had ahigher proportion of congenital malformations com-pared with the untreated women. In Nepal, anotherstudy found that 2 doses of albendazole during preg-nancy was associated with a 46% decrease in neona-tal mortality (RR: 0.54; CI: 0.37–0.78).300

conclusions. These data suggest that dewormingin areas of high endemicity may reduce rates ofmaternal anemia and lead to improved pregnancyoutcomes (eg, stillbirth, PMR, and LBW rates). How-ever, there is a need to formally evaluate the benefitsand potential complications of antihelminthic ther-apy in community-based effectiveness trials in di-verse programmatic settings and locations.

UTIs and Reproductive Tract InfectionsAscending bacterial infections of the genitourinary

tract can be a significant underlying factor in manylate fetal deaths301 as well as spontaneous onset ofpreterm labor.302 Although colonization of the lowergenital tract by fecal flora can occur commonly with-out adverse consequences,303,304 infection of the am-niotic fluid, the interior of the placenta, and/or thefetus is associated with serious complications.301

Chorioamnionitis or amniotic fluid infection hasbeen identified as an important cause of fetal deathin several studies from developing countries.305–309

The following section will review the evidence ofinterventions pertaining to common genitourinaryinfections in pregnancy.

Syphilis Screening and Treatmentbackground. Although syphilis is widely recog-

nized as an important cause of morbidity amongthe adult population, there are few reliable esti-mates of the exact contribution of syphilis to theburden of perinatal and neonatal mortality in devel-oping countries. Congenital syphilis continues to



be a major public health problem in developingcountries.310

Untreated syphilis was found to cause fetal deathin �22% of pregnancies among infected Africanmothers; the risks for antepartum and intrapartumstillbirths among infected women were 18-fold and8-fold higher, respectively.311 In Malawi, a longitu-dinal population-based study revealed a population-attributable risk for syphilis of 26% among all fetaldeaths and 38% for antepartum fetal deaths.312 Syph-ilis was the attributed cause of 10% of 315 late fetaldeaths in a case-control study (CCS) in Port Moresby,Papua New Guinea.79 Syphilis is particularly com-mon in Africa, with prevalence estimates around10%.313–315 Syphilis is also recognized as a majorcontributor to perinatal and infant mortality in theAfrican region.311,315–317 In Ethiopia, �6% of perina-tal mortality was due to syphilis.318 In Zambia, al-most 9% of infants seen at the University TeachingHospital were found to have congenital syphilis,319

and in the Central Hospital of Maputo, Mozambique,�1% of neonates had congenital syphilis.320 The re-ported seroprevalence of syphilis among women at-tending antenatal clinics in African, Asian, and LatinAmerican countries ranges from 4% to 19%.321–324

Some reports also indicate that the incidence of con-genital syphilis may be increasing in many develop-ing countries; to illustrate, a 10-fold increase in con-genital syphilis was reported in the Brazilian FederalDistrict from 1980 (0.17 cases per 100 000 inhabitants)to 1984 (1.7 per 1000).325 Early diagnosis and treat-ment sharply reduce the risk of fetal death due tosyphilis.326

The potential benefits of large-scale population-based intervention strategies for the prevention, de-tection, and treatment of congenital syphilis isstrengthened by the experience in developed coun-tries. To illustrate, the prevalence of congenital syph-ilis was reduced by as much as 70% by a large screen-ing and intervention program in Milwaukee.327

TABLE 15. Antenatal Deworming



Abel etal297

India; rural setting; PCS Iron supplementation anddeworming were providedto all pregnant women inthe intervention area(n � 458) from the fourthmonth of pregnancy. Anintensive information,education, and communi-cation effort was carriedout that provided facts onanemia and dietmodification to eachpregnant woman, using a1-to-1 approach in thecommunity, and a groupmethod in the mobileclinics. This was carriedout for a period of 18 mo.Pregnant controls in aneighboring village(n � 387) received onlythe previously availablestandard of care.

A significant decrease in theprevalence of anemia wasfound, from 56% to 25%(P � .001), 73% to 49%(P � .001) and 69% to57% (P � .01) amongwomen treated from thefirst, second and thirdtrimesters, respectively, inthe intervention area.Significant (P � .001)increases of 0.85 g/dL inmean Hb level (CI: 0.79–0.91), 0.59 g/dL (CI: 0.57–0.61) and 0.36 g/dL(CI: 0.32–0.40) were alsoobserved in treatedwomen in the first,second and thirdtrimesters, respectively.

De Silvaet al299

Sri Lanka; urban hospitalsetting; retrospectivecohort study (RCS)

All women recruited(n � 7087) werequestioned directly abouttheir use ofantihelminthics duringthat pregnancy. 5275 hadtaken mebendazole and1737 had not takenantihelminthics and werein the control group.

The stillbirth, PMRs andVLBW rates were 45%,45% (P � .004) and53% (P � .003) lower,respectively, amongthe women who tookmebendazole duringpregnancy.

Atukoralaet al167

See Table 7

Christianet al300

Nepal; rural setting; PCS Pregnant women received1 of 5 groups ofmultinutrientsupplements, some ofwhich containedalbendazole doses. Womenwho received albendazoleduring the secondtrimester were comparedto women who did notreceive albendazole.

NMR of infants born towomen treated withalbendazole comparedto those born tocontrols declined by46% (RR: 0.54; CI:0.37–0.78).



community-based evidence. We identified 5 in-tervention trials in developing countries that ad-dressed the issue of antenatal screening and treat-ment of syphilis and evaluated the impact onpregnancy outcomes (Table 16). Of these, only one316

was rural-based, and one was in a periurban/subur-ban setting328,329; the rest were urban hospital-basedstudies. We included these hospital-based studies inthe current review because they were conducted indeveloping countries and represented typical catch-ment populations.

The data reviewed suggested that serologic andclinical testing for syphilis did not have sufficientsensitivity for case detection, especially when con-ducted by paramedical or nursing staff.316,330 Onestudy reported sensitivity of 50% and specificity of91% for on-site rapid plasma reagin (RPR) testingdone by nurses.330 Universal screening, preferablyon site, is the appropriate option,331 because it en-sures immediate treatment. High-risk patients, how-ever, should be rescreened in the third trimester.

The studies reviewed demonstrated a significant re-duction in the incidence of congenital syphilis amongcases identified and treated antenatally.316,328,329 Still-

births and perinatal mortality328,329,332 were reducedin offspring of mothers diagnosed and subsequentlytreated for syphilis during pregnancy. An observa-tional study also demonstrated that women who werediagnosed and treated for syphilis during pregnancyhad lower risk (OR: 0.25; P � .0001) of giving birth to anLBW infant.332 Screening for and treatment of syphilisis thus a cost-effective means of reducing fetaldeaths.332

conclusions. Case identification and treatment ofmaternal syphilis have significant benefits in im-proving perinatal and neonatal outcomes, particu-larly in endemic areas such as sub-Saharan Africa.There are major challenges, however, in implement-ing diagnostic testing and treatment in programswhile assuring quality and access. Additional oper-ational research is needed on how to make accuratetesting and effective treatment feasible and availableat scale.

Antibiotics for Asymptomatic Bacteriuriabackground. Although treatment for symptom-

atic UTIs in pregnancy is standard-of-care in devel-oped countries, and the role of specific antimicrobial

TABLE 16. Antenatal Syphilis Screening and Treatment



Patel et al330 South Africa; urbanhospital setting;QT

5 nurses were trained by alaboratory technician toperform and read the RPRcard test. The result of on-site RPR testing (n � 513women) was compared withRPR test results reported bythe laboratory.

Sensitivity for the onsite testwas 50%, specificity was91%, positive predictivevalue was 33% andnegative predictive valuewas 95%.

Osmanet al329

Mozambique;urban hospitalsetting; QT

Before the intervention, nursesreferred patients (n � 918)to an STD doctor. Theintervention trained thenurses to be autonomous indiagnosis and treatment ofsyphilis with benzathinepenicillin.

Perinatal mortality was significantlyhigher in the control group (P �.03; 98.7% of the neonates werealive on the seventh day of life inthe intervention group, vs 95.8%in the control group).

Temmermanet al332

Kenya; urbanhospital setting;CCS

Of the 22 466 neonates born towomen giving birth in thehospital, 12 414 were testedfor syphilis. Cases weredefined as mothers whosebabies were stillborn orLBW.

Syphilis treatment duringpregnancy reduced therate of adverse outcomesignificantly from 26% to15%.

Women who were seropositive anduntreated were 4 times morelikely to have adverse pregnancyoutcomes, including a stillbirth(OR: 3.34, P � .03) or LBW baby(OR: 4.01, P � .0001).

Hira et al328 Zambia; periurbansetting; QT

Staff was trained in methodsof health education, clinicalevaluation, on-site RPRserological testing forsyphilis and on-sitetreatment. Number ofpatients was 3005. Prior tothe intervention, the studycenter had 491 patients andthe control center had 434.Post-intervention, the studycenter had 806 patients andthe control center had 1274.

Adverse pregnancy outcomes(abortion, stillbirth, preterm birth,LBW and congenital syphilis)were evaluated as a combinedentity. Adverse outcomes weresignificantly lower in the studyarea (28%) compared to thecontrol area (72%) (P � .001).

Guinnesset al316

Swaziland; ruralsetting; QT

Nurses performed the RPRtest for 283 deliveries andthe women who werepositive were treated in thiscommunity-based screeningand treatment protocol.

Screening averted 35% of congenitalsyphilis but missed the remaining65%.



therapy in pregnancy is well established,333 there areno community-based studies of treatment for asymp-tomatic bacteriuria during pregnancy and its impacton pregnancy outcomes. Most available data arefrom facility-based cases and from developed coun-tries. In a meta-analysis of the available evidence onbenefits of antimicrobial therapy for asymptomaticbacteriuria in pregnancy, mostly from developedcountries but also including poor populations,Smaill334 convincingly established a reduction inrates of pyelonephritis (OR: 0.25; CI: 0.19–0.32) and apooled relative risk (RR) of 0.64 (CI: 0.50–0.82) forthe composite outcome of preterm delivery or LBWbased on 10 controlled clinical trials.

community-based evidence. No studies wereidentified that reported data from developing-coun-try community settings on the impact of antimicro-bial therapy on asymptomatic bacteriuria in preg-nancy. There is considerable evidence, however, thatbacteriuria and occult UTIs are widespread in devel-oping countries.335–337

conclusions. Although there are no valid studiesof screening and treatment strategies for maternalasymptomatic bacteriuria from community-basedsettings in developing countries, there is little reasonto doubt the association of occult and overt UTIs indeveloping countries with increased risk of preterm/LBW births and neonatal infections, given thestrength of the relationship in developed coun-tries.338,339 Moreover, although the choice of antibi-otics may vary, there is little reason to suggest thatthe findings of the main Cochrane review334 wouldnot be applicable in diverse settings.340 Given theavailable evidence and biological plausibility, allclinically symptomatic UTIs must be appropriatelytreated, and attempts must be made to collect evi-dence from representative developing-country set-tings regarding the importance of treating asymp-tomatic bacteriuria. However, the logistics, technicalrequirements, and frequency of screening necessaryfor diagnosis of asymptomatic bacteriuria in devel-oping-country settings are formidable barriers towide-scale implementation of this intervention.Thus, pending additional evidence as to the feasibil-ity and cost-effectiveness of this strategy, and requireadditional operational and cost-effectiveness research.

Antibiotics for Bacterial Vaginosisbackground. Although reliable estimates of the

global burden of bacterial vaginosis are unavailable,the disorder is known to affect a large proportion ofwomen in developed countries (�16%) and is asso-ciated with adverse pregnancy outcomes includingincreased risk of preterm (�40%) or LBW birth andpremature rupture of membranes (PROM).326,341–344

Some studies have found high rates of bacterial vagi-nosis in developing countries,345 but data are lackingon the attributable risk of bacterial vaginosis forLBW.

Despite strong epidemiologic evidence of an asso-ciation between bacterial vaginosis and pretermbirth, randomized trials of topical clindamycin, aswell as systemic antibiotic treatments using any of a

variety of agents (eg, erythromycin, amoxicillin, ormetronidazole), have not shown clear evidence ofbenefit.302 This may be explained partly by hetero-geneity in the patient populations studied. In theonly published trial among low-risk (ie, no history ofprior preterm birth) asymptomatic women, screen-ing and treatment with oral clindamycin resulted innearly a 50% reduction in the rate of preterm birth orPROM.346 It has been suggested, however, that datashowing benefit from a prospective double-blind,placebo-controlled trial are needed before such anapproach is undertaken routinely in low-risk preg-nant women.347

Antibiotic treatment of bacterial vaginosis amonghigh-risk pregnant women in developed countrieswho previously had a preterm delivery has signifi-cantly reduced the risk of subsequent pretermbirth.348–351 Although Duff et al348 used oral amoxi-cillin for therapy, Hauth et al350 used a combinationof metronidazole and erythromycin for 7 days, andothers349,351 used a short course of metronidazolealone. However, in a recent large trial from the NIHMaternal-Fetal Medicine Network,352 a 2-dose regi-men of oral metronidazole had no effect on rates offetal and neonatal deaths, LBW, or VLBW. Moreover,a slightly increased risk of preterm birth was foundfor women with a prior history of preterm birth.Another recent publication from the same NIH net-work also reported an increased risk of preterm birthafter treatment of women with asymptomatic tricho-moniasis with metronidazole.353 Similarly, a com-bined intravenous (IV)-oral regimen of ampicillin/amoxicillin plus erythromycin had no impact on fetalor infant death rates.354 In the latter trial, however,treatment resulted in reduced incidence of neonatalnecrotizing enterocolitis, intraventricular hemor-rhage (IVH), respiratory distress, and early-onsetsepsis. A single trial of treatment for reproductivetract infection due to chlamydia failed to reduce therisk of preterm birth.355 In a recent review of theimpact of treatments for bacterial vaginosis, Joesoefet al347 concluded that women with symptomaticdisease, whether pregnant or not, should be treatedwith metronidazole (500 mg, twice daily, for 7 days),clindamycin vaginal cream (2%, once daily as a noc-turnal intravaginal application, for 7 days), or met-ronidazole vaginal gel (0.75%, twice daily, for 4days), because these regimens have produced equiv-alent cure rates. For pregnant women with high riskof an adverse pregnancy outcome (ie, those with aprior preterm birth) but asymptomatic disease, intra-vaginal clindamycin is not recommended, because ithas failed to reduce rates of preterm births. If screen-ing of high-risk women is undertaken, it is best per-formed early in the second trimester, and treatmentdoses should be limited to 250 mg of metronidazoleorally 3 times daily for 4 days to minimize any po-tential teratogenic effects of treatment.

community-based evidence. Diagnostic methodsfor bacterial vaginosis in developing-country set-tings have not been standardized, and relatively fewstudies have explored the impact of potential inter-ventions to screen and treat bacterial vaginosis on



pregnancy outcomes (Table 17). A double-blind, ran-domized, placebo-controlled trial at 7 maternity clin-ics in Indonesia reported an 85.5% cure rate amongwomen with bacterial vaginosis treated with 2% clin-damycin vaginal cream, but rates of preterm deliveryor LBW were not affected.344 It was suggested thatsystemic treatment may be needed to eradicate up-per reproductive tract disease and to reduce pretermbirths. Treatment of women in South Africa in earlypreterm labor for 4 days with metronidazole andampicillin (agents effective against, but not specifi-cally targeted toward, bacterial vaginosis in this trial)increased pregnancy duration, but preterm birthrates were not reported, and no reduction was ob-served in either neonatal deaths or length of hospitalstay.356

Broad-spectrum treatment for reproductive tractinfections in rural Uganda with single doses ofazithromycin, cefixime, and metronidazole resultedin a significant decrease in preterm birth rates(�23%) and early NMRs (�17%).345 In Kenya, a sin-gle oral dose of cefetamet-pivoxil given to women at28–32 weeks of pregnancy with a prior history ofstillbirth or an LBW infant resulted in significantreductions in the LBW rate (probably due largely topreterm birth) and postnatal endometritis and anincreased birth weight.357 No effect was observed onstillbirths or preterm birth rates. Large losses to fol-low-up, however, suggest the need for cautious in-terpretation of these results. Finally, a recently pub-

lished trial in pregnant women living in the Rakaidistrict of Uganda revealed that those randomized toreceive a single presumptive treatment dose ofazithromycin, cefixime, and metronidazole hadborderline significant reductions in risk of pretermbirth (OR: 0.72; CI: 0.56–1.05) and early neonataldeath (OR: 0.83; CI: 0.71–0.97).358 Each of these treat-ment regimens broadly decreased rates of maternalreproductive tract infections including bacterialvaginosis.

conclusions. Available data suggest that antibi-otic therapy effective for treatment of bacterial vagi-nosis may be beneficial in decreasing rates of LBWand, possibly, prematurity and neonatal mortality.345

Treatments that are effective against bacterial vagi-nosis may also decrease rates of other reproductivetract infections such as those due to chlamydia andgonorrhea. However, results have been mixed, withsome studies showing increased risk for pretermbirths. Moreover, there are still major problems withoperationalizing these interventions into programs,because simple, affordable diagnostic methods andadditional large-scale effectiveness studies of simpletreatment regimens, particularly in asymptomaticpregnant women, are needed. It is also unclear if theprevention of preterm births between 34 and 37weeks’ gestation will substantially improve perinataland newborn survival in public health programs.359

These interventions merit additional exploration be-

TABLE 17. Antibiotics for UTIs and STDs

Source Location and Type ofTrial


Wawer et al345 Uganda; rural setting;RCT

Intervention clusters(n � 12 733,of whom 6609 were HIV-negative) were given broad-spectrum treatment for STDsthat consisted ofazithromycin (1000 mg singledose), cefixime (400 mg singledose) and metronidazole (2 gsingle dose). The controlgroup (n � 6124) wasuntreated.

Prevalence rates of STDsother than bacterialvaginosis at thepostnatal visit weresignificantly lower inthe intervention groupthan the controlgroup.

There was a 17% reductionin early neonatal deathand a 23% reduction inthe risk of pretermdelivery in theintervention group.

Gichangi et al357 Kenya; urban hospitalsetting; DBRCT

Pregnant women (n � 320)were randomized into anintervention group (n � 160)and placebo (n � 160). Theintervention group received asingle oral dose of 2 gcefetamet-pivoxil; the controlgroup received placebo.

In the interventiongroup there were 65%fewer Neisseriagonorrheae cervicalinfections (P � .04).

The intervention had noeffect on the rate ofstillbirth. However,infants born to mothersin the interventiongroup were heavier(mean difference 155 g)than those in the placebogroup (P � .04). Theintervention reducedLBW by 44% (18.7% vs32.8%) but had no effecton gestational age(P � .01).

Joesoef et al344 Indonesia; multicentertrial, urban hospitalsetting; RCT

Women seeking prenatal care at14–26 wk of gestational ageand who had bacterialvaginosis (n � 681) weregiven either 2% clindamycinvaginal cream (n � 340) or aplacebo cream (n � 341).

2 weeks after thecompletion of thetreatment, 88.5% ofthe women werecured.

The rate of pretermdelivery (�37 wk) was15% for clindamycinpatients and 13.5% forplacebo patients (OR:1.1; CI: 0.7–1.7). The rateof LBW was 9% forclindamycin patients and6.8% for placebo patients(OR: 1.3; CI: 0.8–2.4).



fore their inclusion in routine antenatal carestrategies.

Antibiotics for Preterm Laborbackground. A major proportion of the burden of

global perinatal mortality relates to preterm births.The WHO estimates that prematurity is responsiblefor nearly one fourth of neonatal deaths.29 Of �13million annual preterm births globally,360 one thirdmay be related to spontaneous preterm delivery,361

and in 30% to 40% of preterm births, labor mayfollow spontaneous rupture of membranes.362

Several studies have identified an association ofsubclinical chorioamnionitis (ie, intact membranes)with spontaneous preterm labor.363–365 Pathogenswere identified from the amniotic fluid in 10% to15% of cases presenting with preterm labor.366 Ameta-analysis by the Cochrane collaboration ofavailable RCTs of antibiotic therapy in preterm laborwith intact membranes367 identified that antibiotictreatment in this situation prolonged pregnancy(weighted mean difference: 5.4 days; CI: 0.9–9.8days). There was no effect on reduction of pretermbirths or on respiratory distress syndrome or neona-tal sepsis, but a significant reduction in neonatalnecrotizing enterocolitis (OR: 0.33; CI: 0.13–0.88) wasnoted. The meta-analysis also revealed, however, anincrease in mortality related to preterm birth or itssequelae (OR: 2.43; CI: 0.92–6.43) and an increase inperinatal mortality in the group receiving antibiotics(OR: 3.36; CI: 1.21–9.32). With regard to maternaleffects, antibiotic treatment significantly reducedmaternal infection (chorioamnionitis/endometritis)(OR: 0.68; CI: 0.48–0.98). Still, overall evidence forthe benefit of antibiotic therapy in preterm labor withintact membranes is lacking.

This large meta-analysis has been supplementedby the large multicenter ORACLE II trial in severaldeveloped countries, including almost 6300 pregnantwomen randomized into 1 of 4 groups to receive(1) erythromycin, (2) amoxicillin/clavulanic acid,(3) both, or (4) placebo.368 The primary outcome mea-sure was a composite of neonatal death, chronic lungdisease, or major cerebral abnormality on ultra-sonography before discharge from hospital. Therewas no evidence that any of the antibiotic regimensimproved maternal or neonatal outcomes; length ofpregnancy, length of maternal hospital stay, mode ofdelivery, birth weight, proportion of infants admit-ted to intensive care or ventilated, proportion ofinfants with a positive blood culture, or compositeneonatal outcome were not different among thetreatment groups.

community-based evidence. Most of the trials ofantibiotic therapy for preterm labor have been con-ducted in developed countries; none have been re-ported from community settings. Table 18 illustratesthe 2 studies of antibiotic therapy for preterm laborin developing-country settings.356,369 The overallconclusions from these studies were consistent withthe results of the Cochrane meta-analysis367 and theORACLE II study.368

conclusions. No clear overall benefit for routineantibiotic therapy for preterm labor with intact mem-

branes has been demonstrated in trials in urban set-tings in developed or developing countries. Thus,this intervention cannot be recommended routinely.Evidence indicates that antibiotic administration inpreterm labor should be considered only when thereare clear indications of associated infection, or, asindicated in the next section, in the presence of otherrisk factors.

Antibiotics for PPROMbackground. The most common antecedent of

preterm labor is PPROM.362 The culture positivityrate of amniotic fluid for microbial organisms in suchcases ranges from 32% to 35%.366 The mechanismsunderlying PPROM may include local subclinicalinfection and inflammation leading to weakening ofthe amniotic membranes.370

A Cochrane meta-analysis of the benefits of anti-biotic therapy in cases of PPROM concluded thatmaternal antibiotic therapy in this situation is effec-tive in prolonging pregnancy and reducing maternaland neonatal infection-related morbidities.371 A sub-sequent large (n � 4826) multicenter RCT (ORACLEI)372 in urban centers in multiple developed and de-veloping countries (eg, Argentina, Brazil, Lithuania,Malaysia, South Africa, Sri Lanka) corroborated thefindings of the meta-analysis and suggested thatadministration of erythromycin to women withPPROM was associated with significant health ben-efits for the newborn. Fewer infants (P � .08) tendedto have the primary composite outcome (ie, death,chronic lung disease, or major abnormality on cere-bral ultrasonography) in the erythromycin group(151 of 1190 [12.7%]) than in the placebo group (186of 1225 [15.2%]). Among the 2260 singletons in thiscomparison, significantly fewer had the compositeprimary outcome in the erythromycin group (125 of1111 [11.2%] vs 166 of 1149 [14.4%]; P � .02). Use oferythromycin was also associated with prolongationof pregnancy beyond 48 hours of presentation (P �.004), reduction in neonatal treatment with surfactant(P � .05), reduced requirement for supplementaloxygen (P � .02), and fewer positive blood cultures(P � .02). An important additional finding amongsingletons was a lower rate of neonatal cranial ultra-sonographic abnormalities at discharge (P � .04) anda probable reduction in childhood disability. Al-though co-amoxiclav only and co-amoxiclav pluserythromycin were associated with prolongation ofpregnancy, they were also associated with a signifi-cantly higher rate of neonatal necrotizing enterocoli-tis (4-fold higher with co-amoxiclav alone versusplacebo, 2.5-fold higher with any co-amoxiclav thanwith none).

community-based evidence. The findings of 3other studies373–375 conducted in developing coun-tries largely corroborated findings from studies indeveloped countries, although data are very limited(Table 19). No trials of antibiotic therapy for PPROM,however, have been conducted solely in communitysettings in developing countries.

conclusions. Although there are few systematicstudies of antibiotic therapy for PPROM in develop-ing countries and none in community settings, the



evidence in support of the efficacy of this interven-tion in diverse settings is strong, and the practice isroutine under certain circumstances (eg, early gesta-tion). Thus, it is prudent to incorporate antibiotictherapy of PPROM in intervention programs wher-ever feasible, such as at referral-level health facilities.A domiciliary cadre of trained birth attendants po-tentially can be trained to recognize PPROM andprovide referral and, possibly, initial antimicrobialtherapy. However, this application of the interven-

tion requires additional assessment and may be dif-ficult in community settings.

TT Immunization and Clean Deliverybackground. Despite major advances in under-

standing of the pathogenesis and risk factors for neo-natal tetanus, it is still prevalent in many developingcountries. A UNICEF/WHO/United Nations Popula-tion Fund survey in June 2000 indicated that 57 coun-tries globally had not achieved elimination of neonatal

TABLE 18. Antibiotics for Preterm Labor



Oyarzunet al369

Chile; urban hospitalsetting; DBRPCT

Women in labor between 22–36wk gestation (n � 173) weregiven either amoxicillin anderythromycin for 6–7 d, plustocolytic therapy (n � 83) orplacebo (n � 90).

No significant differences betweenthe groups were found formaternal outcomes, includingduration of randomization-to-delivery interval and frequencyof clinical chorioamnionitis andendometritis. Rate of cesareansection was significantly higherin the placebo group (28% vs12%).

No significant differences inneonatal outcomes weredetected between thegroups, including pretermbirth rate, neonatal death,respiratory distresssyndrome, proven sepsis,and birth weight.Suspected sepsis wassignificantly more frequentin the placebo group (6/90vs 0/78).

Normanet al356

South Africa; multi-center trial, urbanslum setting; RCT

The study recruited 81 pregnantwomen. The study group(n � 43) received ampicillinand metronidazole for 5 d. Thecontrol group (n � 38)received no antibiotics. In allwomen, contractions weresuppressed with hexoprenalineand indomethacin for 24 hours,and betamethasone was givenfor fetal lung maturity.

In those receiving ampicillin andmetronidazole, pregnancy wassignificantly prolonged (median15 d vs 2.5 d, P � .04) withsignificantly more women stillpregnant after 7 d (63% vs 37%,P � .03, OR: 0.34; CI: 0.13–0.94).

Significantly fewer infants inthe treatment groupdeveloped necrotizingenterocolitis than in thecontrol group (0 vs 5,P � .02).

Kenyonet al368

(ORACLEII trial)

Multicountry, multi-center trial; urbanhospital setting;RPCT

6241 women in suspected orconfirmed preterm labor (�37wk gestation) wererandomized to receive 1 of 4possible treatments: 1) co-amoxiclav (325 mg) pluserythromycin (250 mg); 2) co-amoxiclav plus erythromycinplacebo; 3) erythromycin plusco-amoxiclav placebo; or 4)co-amoxiclav placebo pluserythromycin placebo andfollowed for pregnancyoutcomes.

Antibiotic use did not prolongpregnancy, as most women didnot deliver within 48 h (89.9%) or7 d (84.6%).

The only significant differ-ence between the antibioticand placebo groups wasthat the rate of oxygendependence at 36 wk post-conception was higherwith erythromycin usethan with no erythromycin(P � .04). Importantly, theother measure of chroniclung disease (oxygendependence at �28 d ofage) did not show thisdifference. There were nosignificant differencesbetween the babies withrespect to birth weight orneed for intensive orspecial care. Co-amoxiclavwas associated with a(non-significant) doublingof risk for suspected orproven necrotizingenterocolitis. In thesubgroup of infants bornat �32 wk gestation, therewere no significantdifferences in neonataloutcomes (respiratorydistress syndrome, oxygendependence, sepsis,necrotizing enterocolitis,abnormal cerebral scan, ordeath) between antibioticsand placebo.



tetanus (defined as �1 case of neonatal tetanus per1000 live births in every district).376 Globally, onlyabout half of women of reproductive age are ade-quately immunized, and it has been estimated thatneonatal tetanus accounts for as much as 7% of neona-tal deaths globally.29 Others point out that maternaltetanus may account for at least 5% of all maternaldeaths.377 An additional 90 000 mothers die annuallyfrom puerperal infections caused by unclean delivery

practices, and sepsis accounts for 14% of all obstetricaldeaths.378 Thus, it is imperative that preventive strate-gies such as clean delivery and maternal tetanusimmunization form a cornerstone of maternal care inany setting and that they go hand-in-hand.

Given the scientific basis for the importance oftetanus prevention and basic hygiene, it would nowbe unethical to conduct RPCTs assessing such ap-proaches. Hence, much of the evidence for benefit of

TABLE 19. Antibiotics for PPROM



Kenyon et al(ORACLEI trial)372

Urban hospitals inmany countries;RPCT

4826 women with PPROM wererandomly assigned to receive250 mg erythromycin (n �1197), 325 mg co-amoxiclav(250 mg amoxicillin plus 125mg clavulanic acid; n � 1212),both (n � 1192), or placebo(n � 1225), 4 times daily for10 d or until delivery.

Significant prolongation of gestation(40.7% females delivered within 48hours in the placebo group, vs34.8% in the erythromycin-onlygroup, P � .004)was seen inmothers treated with anyerythromycin combination but nodifferences were seen in rates ofmaternal infection or the need foradditional antibiotics.

The primary outcome measurewas a composite of neonataldeath, chronic lung disease, ormajor cerebral abnormality onultrasonography beforedischarge from hospital. Fewerinfants (P � .08) had theprimary composite outcome inthe erythromycin group (151of 1190 (12.7%)) than in theplacebo group (86 of 1225(15.2%)). Among the 2260singletons in this comparison,significantly fewer had thecomposite primary outcome inthe erythromycin group (125of 1111 (11.2%) vs 166 of 1149(14.4%), P � .02).

Magwaliet al375

Zimbabwe; HarareMaternityHospital; RCT

171 women with PPROMbetween 26 and 36 wkgestation were randomized toa course of prophylacticAugmentin or no prophylacticantibiotic treatment at all. Thecalculated sample size was 72women per group.

Although there was a trend towardshigher maternal febrile morbidityrate in the group that did notreceive any treatment, no statisticalsignificance was achieved. Thewomen in the treatment group hada significant prolongation ofgestation, with 37.8% of themothers in the this groupdelivering within 48 hours ofmembrane rupture compared to58.1% in the group that received notreatment (P � .01247). Moremothers in the treatment group(43.9%) delivered between 48 hoursand 7 d after membrane rupture incomparison to the group that didnot receive treatment (34.9%). 18%of the mothers in the treatmentgroup delivered after 7 d, asopposed to only 7% of the mothersin the group that received notreatment. (P � 0.01247).

The neonatal outcome measuresin the 2 groups (birth weight,Apgar scores, neonatal deathdue to sepsis, neonatal sepsis)were not significantly different.

Ovalle-Sallaset al374

Sites in Chile andUSA; urbanhospital setting;DBRCT

79 women with PPROM wererandomized to receive IVclindamycin plus gentamicinfor 7 d (n � 39) or to thecontrol group (n � 40).

Overall, incidence of prematurebirths (�37 wk and �34 wkgestation) was not reduced. Inwomen with infections whoreceived antibiotics, incidence ofpremature birth (�34 wk) wassignificantly lower than thosepatients who received placebo(44.4% vs 88.9%, respectively, P �.05).

Neonatal outcomes were unclearas the trial was stopped afterthe treatment group showedbetter maternal outcomes atintermediate analysis.

Almeidaet al373

Mozambique;urban hospitalsetting; RCT

106 third trimester pregnantwomen with PPROM wererandomized to either oralamoxicillin (0.75 g, 3 timesdaily) (n � 50) or placebo(n � 56) in a blinded fashion.The patients were hospitalizedin bed for 7 d unlesscontractions started anddelivery ensued.

The average rupture-to-expulsioninterval was 68.4 hours in theplacebo group and 91.7 hours inthe amoxicillin group (P � .28).Hospital stay (3.0 vs 4.3 d, P � .03)was prolonged by 43% in theamoxicillin group. Rates ofhemorrhage and postnatalendometritis-myometritis did notdiffer significantly in the 2 groups.

Birth weight and stillbirth ratedid not differ significantly inthe 2 groups. There was atrend towards longer durationof hospital stay amongnewborns in the amoxicillingroup suffering neonatal death(1.6 vs 6.5 d, P � .06).



clean delivery practices and TT immunization comesfrom observational studies.

community-based evidence. In addition to a hostof descriptive studies, our search identified 8 trialsconducted in rural settings in developing countries thatprovided pertinent information on perinatal and/orneonatal outcomes after various interventions to pre-vent tetanus and provide clean delivery (Table 20).These studies variably involved maternal TT immuni-zation and/or preparation for clean delivery by equip-ping mothers and/or birth attendants with delivery

kits and promoting hand-washing for delivery andnewborn care. Thus, in general, it was not possible todetermine the impact of individual interventions onoutcomes.

Significant reductions in rates of neonatal tetanus,as well as associated case fatality and NMRs, wereobserved in newborns of mothers immunized fortetanus during pregnancy.379–385 Even partial immu-nization was found to be effective in decreasing mor-tality and tetanus morbidity.379,382 Hand-washingand use of clean instruments for cord cutting were

TABLE 20. Antenatal TT Immunization and Clean Delivery


Intervention Perinatal/Neonatal Outcome

Meegan et al389 Kenya and Tanzania;rural setting; QT

During the project period, there were 29 689births in the intervention areas and 88 471births in the control areas. In theintervention area, TBAs carried outantenatal, intrapartum and postnatal care.They were supplied with delivery kits andtaught clean delivery. No immunization wasperformed.

The neonatal tetanus rate decreased by99% in the intervention area comparedto the control area (0.75/1000 live birthsin intervention areas, range � 0–3, vs82/1000 live births in control areas,range � 74–93).

Tsu393 Nepal; rural setting;PCS

Women from 3 districts who had delivered alive infant (defined as a baby that survivedat least 24 hours after birth and was born athome between 7 and 28 d prior to theinterview) were enrolled in 1 of 4 possiblecohorts: 1) kit user with trained attendant(n � 420), 2) kit user with untrained or noattendant (n � 398), 3) kit non-user withtrained attendant (n � 404), and 4) kit non-user with untrained or no attendant(n � 438). The data were then collected byinterviews with the mothers of thenewborns, members of the mother’shousehold who were present for thedelivery or were caretakers of the newbornalong with some trained TBAs who hadparticipated in some eligible deliveries. Fieldinterviewers also directly observed thenewborn’s abdomen and cord area, andconsulted with medical records if available.

The cohorts in which kits were used hadless than half the infection rate (0.45;CI: 0.25–0.81) of kit non-users who didnot use a new or boiled blade and cleancutting surface (after adjusting forconfounders), but there was nosignificant difference between kit usersand any other group of kit non-users,suggesting that clean cord-cutting is oneof the most important practices inpreventing infection.

Gupta et al379 India; rural setting;PCS

TT vaccination was given IM to pregnantwomen (n � 1760) as a single booster doseto those who had received 2 doses in thepreceding 3 y (n � 762), and TT vaccine wasgiven in 2 doses at 1-month intervals tothose who had no previous history of TTimmunization (n � 696) or who werepreviously only partially immunized(n � 230).

Neonatal tetanus prevention attributableto TT vaccine was 88% for completeimmunization and 59% for partialimmunization compared tononimmunized. The risk of neonataltetanus among children born to womenwho received any dose of TT was one-fifth that of children born to non-immunized mothers (OR: 0.12; CI: 0.02–0.41).

Chongsuvivatwonget al386

Thailand; Province-based study; RCT

An education program for TBAs (n � 214)consisting of general midwifery and cord-care training was carried out in 2 provinceswith a total population of 500 000, and massTT immunization was carried out in just 1of the provinces. Pre- and post-interventionsurveys were carried out with randomlyselected groups of 210 respondents; theproject also interviewed 112 TBAs.

The incidence of neonatal tetanus in bothprovinces declined sharply, suggestingthat reinforcement of routine servicesand hygienic practices was of primaryimportance. The NNT death rate was0.2–0.4/1000 at the end of theintervention, reduced 8- to 10-fold.

Kapoor et al380 India; rural setting;QT

TT immunization of pregnant women wasinitiated in 28 villages in 1970. The secondstrategy adopted in this area after 1982 wasto distribute to every registered pregnantwomen a sterilized delivery kit containinggauze pieces, half a razor blade and thread.

There was a gradual and sustainedreduction in the neonatal mortality ratefrom 42.3/1000 live births in 1972 to17.9/1000 in 1987. During this period,neonatal deaths due to tetanusdisappeared (14.6/1000 live births in1972, 0/1000 in 1987). A 22% reductionin NMR occurred after the introductionof birth kits; there is a mixed impact ofbirth kits and TT.



associated with reduced neonatal tetanus inci-dence,381,386–389 neonatal mortality,380,384 and sep-sis.271,384 Training TBAs in clean delivery and mak-ing clean delivery kits available were seen asimportant factors in infection control.380,385,390–392 Inrural Nepal, failure to wash hands before cutting thecord or use of dirty cloths on the umbilical cord wereassociated with 60% and 70% increased risk of cordinfection, respectively.393 Moreover, failure to use aboiled blade among nonusers of clean delivery kitsled to a 2.3-fold increase in risk of cord infectionsabove that of clean delivery kit users. However, useof a clean home-delivery kit did not affect cord in-fection rates, provided that a new or boiled bladewas used to cut the cord.

Thus, tetanus immunization of pregnant women incombination with promotion of hand-washing andclean delivery, including clean umbilical cord care,was protective against neonatal tetanus and resultedin reduced neonatal mortality and morbidity.Among these interventions, the most persuasive dataon efficacy are in support of tetanus vaccinationstrategies. In Sri Lanka, the introduction of the Ex-panded Programme on Immunization (EPI) vaccina-tion strategy in 1978 led to a dramatic reduction inneonatal tetanus infection rates, which declined from2.16 to 0.06 cases per 1000 live births.394 Similar datawere reported from Burma,384 showing that neonataltetanus mortality rates in non-EPI areas were three-fold higher than in EPI areas (9 vs 3 per 1000 livebirths). Similarly, in Bangladesh, national TT-immu-nization programs reduced the incidence of deathsfrom neonatal tetanus by 90% (from 41 to 4 deathsper 1000 live births) over a decade.29 One study (in 2sites) compared the efficacy of 3 intervention strate-

gies for reducing neonatal tetanus: maternal TT ad-ministration, hospital-based delivery (to promoteclean delivery), and home delivery by a trained birthattendant.395,396 Overall, maternal tetanus vaccina-tion is most effective and seems to offer the singlebest option for reducing neonatal tetanus in devel-oping countries (Table 21).

conclusions. Evidence for the benefits of mater-nal tetanus immunization on neonatal outcomes isincontrovertible. Maternal TT immunization must bean essential part of antenatal care packages as well asmass-vaccination programs. Similarly, the impor-tance of clean delivery practices, including clean um-bilical cord care, must be underscored. Availablecommunity-based data suggest that the best gainsmay be obtained from implementation of a combina-tion of maternal immunization and clean deliveryand cord-cutting practices. On the other hand, thereis no evidence that as a single intervention, use of aclean delivery kit is necessarily the most appropriateway of ensuring that caregivers and birth attendantspay sufficient attention to antisepsis, nor is thereevidence that kit use impacts umbilical cord infectionor NMRs. To the contrary, behavior-change commu-nications strategies to promote clean delivery prac-tices, including clean cord cutting, should be im-

TABLE 21. Comparative Efficacy of Interventions to PreventNeonatal Tetanus

TTAdministration

(2 Doses)

HospitalDelivery

Home Deliveryby a Trained

Birth Attendant

Burma395 91.4% 84.6% 32.8%Egypt721 87% 77% 49%

TABLE 20. Continued



Rahman et al383 Bangladesh; ruralsetting; PCS toevaluate an RCT

Neonatal mortality was compared betweenoffspring born from 1978–1979 to womenwho had received TT prior to pregnancy(n � 956), offspring born to women givenTT while pregnant from 1978–1979 (n �934), offspring born to partiallyimmunized women (n � 729), andoffspring born to non-immunized mothers(n � 7237). Women given TT were given 2doses if immunized in 1974 (non-pregnant), 3 doses of TT if immunized in1978 (pregnant) and 2 doses if immunizedin 1979.

Decreases in the stillbirth rate (44%,P � .05) and neonatal mortality rate(49%, P � .01) were observed in thewomen who were fully immunized inpregnancy compared to those whowere not.

Rahman 384 Bangladesh; ruralsetting; RCT

In 3 unions, TBAs were trained in bettermaternal nutrition and hygienic newborncare practices; in 3 unions TT was given;and 3 unions were kept as controls. 1760pregnant women were cared for by theseTBAs.

NMR decreased 72% and 54% in theTBA training areas and the TT areas,respectively. The NMR was 38.9/1000in TT unions compared to 85.2/1000in control unions. Tetanus neonatorumdecreased by 94.6% in the TT area.

Black et al382 Bangladesh; ruralsetting; DBRCT

Volunteer non-pregnant women (n � 46 443)received 0.5 mL of cholera toxin oraluminum phosphate-adsorbed TT by IMinjection. 13 220 women received 1 TTinjection, while 33 175 received 2 TTinjections. Neonatal mortality in infantsborn 9–32 mo after immunization werecompared.

In the first cohort of offspring, neonatalmortality was significantly lower(33%, P � .01) among those born after9–32 mo of mass tetanus vaccinationin women (68.4/1000 in controls vs44.1/1000 in TT area). 75% of thereduction in mortality was attributableto the reduction in neonatal mortalitybetween 4–14 d of age (P � .001).



plemented in tandem with introduction of cleandelivery kits.

We would underscore the need to undertake anevaluation of the most appropriate, cost-effective,and sustainable clean delivery strategies in commu-nity and rural settings in developing countries. Ifclean delivery kits are promoted, it should be done inthe context of broader behavior-change communica-tions regarding clean delivery practices.

Maternal Pneumococcal Immunizationbackground. The high incidence of infectious dis-

eases in young infants in developing countries sug-gests that, in many cases, maternally derived anti-bodies fail to provide adequate protection. Forexample, infection with Streptococcus pneumoniae ac-counts for 10% of deaths (220 000 deaths per year)globally in infants �90 days old.397 Although livevaccines are contraindicated in pregnancy, killed ortoxoid-based vaccines can be administered safely topregnant women. These strategies may be particu-larly useful in circumstances in which backgroundrates of routine vaccinations in childhood and ado-lescence are low or available vaccines are contrain-dicated or ineffective in young infants. Thus, al-though the benefits of maternal immunizationagainst tetanus are well established, there is verylittle information about the impact of other vaccinesadministered during pregnancy on pregnancyoutcomes.

community-based evidence. One study of mater-nal pneumococcal vaccination has been reported(Table 22).398 Although the study took place in anurban hospital setting, we included it in our evalua-tion because it was based in a developing countryand evaluated neonatal outcomes. Infants born tomothers who had been immunized had higher bloodlevels of antibodies sufficiently protective againstpneumococcal antigens. However, the newborn an-tibody levels depended primarily on the durationbetween immunization and delivery. Although thelevel of antibodies decreased at 22 weeks of agepostnatally, levels were still higher than those foundamong control infants. Additional data on clinicaloutcomes and long-term protection were notavailable.

conclusions. Although limited and preliminary,these data provide intriguing information on thepossibility of maternal immunization as a strategyfor addressing early neonatal pneumococcal infec-

tions. Development of maternal immunizationsagainst group B streptococcus and Haemophilus influ-enzae is similarly of great interest and under activeinvestigation.

Promotion of Smoking Cessation During Pregnancybackground. Inhalation of smoke by pregnant

women may occur through several modes of expo-sure and has been consistently associated with IUGRand LBW in offspring. It is widely recognized thatsmoking increases the risk of LBW births almosttwofold399 and significantly increases perinatalmortality.400

Infants born to women who smoke during preg-nancy weigh, on average, 200 g less than infantsborn to comparable women who do not smoke.399

Smoking has been shown to double the risk of LBWbirths among those mothers who smoke, comparedwith those who do not. In addition to LBW, in-creased risk of spontaneous abortion, as well as in-creased risk of perinatal and neonatal mortality, havealso been associated with maternal smoking.401 Re-cently, new evidence has shown an association ofindoor air pollutants with LBW and increased infantand perinatal mortality.402,403 One study reportedan association of LBW with use of wood as cookingfuel.401,402 The study by Boy et al,402 using a retro-spective cohort design, evaluated the association be-tween LBW and exposure to wood smoke by askingwomen at the time of delivery about the type ofcooking fuel they used. The study results were po-tentially confounded, however, failing to control fordifferences in nutritional and socioeconomic vari-ables associated with different fuel types. The studyalso included a combination of hospital and homedeliveries but excluded women living in remote ar-eas due to inaccessibility.

Smoking and use of tobacco in other forms arewidespread in developing countries. Moreover, ex-posure to outdoor air pollution, indoor air pollution,and indoor smoke (either from cooking or secondaryinhalation of cigarette smoke) is widespread andmay impact the mother as well as the young infant,leading to increased risk for asthma and respiratoryinfections.404

A systematic review of RCTs evaluating the im-pact of cessation of smoking during pregnancy405

found that there were significant reductions in theincidence of LBW (OR: 0.8; CI: 0.67–0.95) and pre-term births (OR: 0.83; CI: 0.69–0.99) and an increase

TABLE 22. Maternal Pneumococcal Immunization



Shahid et al398 Bangladesh; urbanhospital setting;DBRCT

Women 30–34 wk gestation(n � 70) were ran-domized to receive 23-valent pneumococcalpolysaccharide vaccine(n � 36) or meningo-coccal vaccine containingpolysaccharide of groupsA, C, Y and W-135.(n � 37).

Among pneumococcal vaccinerecipients, there was a highcorrelation between thematernal pre-immunizationand post-immunizationantibody titers (Spearmancorrelation RS � 0.66 fortype 6B and RS � 0.77 fortype 19F, P � .0001).

Infant cord blood antibody titerswere highly correlated withthe maternal titers at delivery.In the 22 wk postnatalsample, 71% and 63% of theinfants of women who hadreceived pneumococcalvaccine had antibody�0.15 �g/ml for types 6B and19F, respectively (P � .01).



in mean birth weight of 28 g (CI: 9–49 g). However,there was no difference in the incidence of VLBWbirths or overall perinatal mortality, and the largestcluster-randomized trial reviewed showed no evi-dence of an impact of reduction in smoking on ad-justed mean birth weight.406

community-based evidence. Measures to reduceenvironmental exposure to smoke may improvepregnancy outcomes, but this has not been system-atically studied in developing countries. In light ofthis lack of data, evidence available concerning theimpact of smoking cessation on pregnancy outcomesis presented from developed countries in an attemptto evaluate the impact of this potential intervention.

We identified 8 additional studies, the majority ofwhich were undertaken in the United States, pri-marily in urban settings, that reported on the im-pact of smoking cessation on pregnancy outcomes(Table 23). No rural-based study could be identified.The studies included several smoking-cessationcounseling methods ranging from individual coun-seling by midwives to counseling by a trained pro-fessional counselor, often supplemented with litera-ture about the potential adverse effects of smokingon the developing fetus. In all but 1 study,407 inter-ventions to promote smoking cessation in pregnantwomen were found to be significantly effective inincreasing smoking-cessation rates408–411 and reduc-ing the number of cigarettes smoked by moth-ers.410,412 In most studies, however, no impact wasseen on pregnancy outcomes, including birthweight,407,408,410,413,414 rates of LBW or preterm de-livery,407–410,413,414 or perinatal mortality.409,413,414

Some studies, however, demonstrated an increase inbirth weight and/or length in subgroups thatstopped smoking due to the intervention.409,411,412

conclusions. Overall, smoking-cessation pro-grams have produced mixed results on pregnancyoutcomes such as preterm birth or LBW rates andhave failed to impact fetal, perinatal, or infant mor-tality despite success in reducing smoking rates. Fewstudies have evaluated the impact of environmentalsmoke and indoor air pollution on pregnancy out-comes in developing countries or interventions toredress them. However, a recent review by the WHOhas highlighted the potential adverse health conse-quences from indoor air pollution.403

Clearly, it is prudent to discourage maternalsmoking due to the general harm that smoking doesto health. It is important, however, that additionalstudies on reducing exposure to tobacco and indoorair pollution be conducted in developing countrieswith direct assessment of pregnancy outcomes andhealth of the young infant, because evidence for theimpact of interventions on health status of newbornsin developing countries is currently unavailable.Presently, despite the paucity of studies from com-munity settings in developing countries, there isenough information from developed countries to de-cry the use of any form of tobacco (smoked orchewed) and addictive drugs such as alcohol, co-caine, and narcotics in pregnancy.

Maternal Care Packagesbackground. The aforementioned studies and

interventions largely evaluated solitary maternalinterventions. In most programmatic settings, how-ever, such interventions typically are provided incombinations and as part of a package of care. Al-though most such studies have examined maternaloutcomes, data also exist on the impact of pack-ages of maternal interventions on perinataloutcomes.

community-based evidence. Our review yielded2 studies that provided maternal care packages in arural community-based setting and reported infor-mation on perinatal outcomes (Table 24). These pro-grams were directed toward identification of high-risk patients and their early referral to a higher-levelfacility and provided a multilayered series of mea-sures including essential antenatal care, improvedreferral systems, and facility-based care. The com-munity-based maternal care programs showed sig-nificant reductions in perinatal (34%) and neonatal(43%) mortality.415 Recent data from a perinatal in-tervention trial in Macedonia indicate that it may bepossible to reduce perinatal and neonatal mortalityat the national scale through a concerted perinataleducation and health worker–training programwithin the health system setting.416 Similarly, theintroduction of a structured primary care program inBolivia,417 coupled with maternal care, led to a dra-matic reduction in infant and young child mortality.In 1992–1993, the annual rates of mortality of chil-dren �5 years old were 205.5 per 1000 and 98.5 per1000 in the comparison and intervention areas, re-spectively. The absolute difference in mortality of107.0 deaths per 1000 (CI: 72.7–141.3 per 1000) rep-resented 52.1% (CI: 35.2–68.8%) lower mortalityamong children �5 years old in the interventionareas, compared with the control areas.

conclusions. Available evidence suggests thatmuch of the improvement in perinatal and neonatalhealth in the West preceded the advent of neonatalintensive care and is largely owed to better andcomprehensive maternal care and reproductivehealth.418 In developing countries, the dependence ofinfant survival on the survival and health of themother has been clearly demonstrated.14 Data on theimpact of maternal interventions on perinatal andneonatal outcomes are scarce but robust. Investing inimproved maternal health must form the cornerstoneof strategies aimed at improving perinatal and neo-natal outcomes.

Intrapartum Interventions

Maternal Vaginal and Newborn Skin Antisepsisbackground. Maternal intrapartum and postpar-

tum infections are a major cause of maternal morbid-ity and mortality in developing countries. A largeproportion of early-onset neonatal infections in de-veloping countries may also be related to verticallytransmitted infections from the maternal genitaltract.21,25,419 Thus, there is interest in evaluating low-cost strategies for preventing and reducing infectious



TA

BL

E23

.Pr

omot

ion

ofSm

okin

gC

essa

tion

Dur

ing

Preg

nanc

y

Sour

ceL

ocat

ion

and

Typ

eof

Tri

alIn

terv

enti

onM

ater

nal

Out

com

ePe

rina

tal/

Neo

nata

lO

utco

me

Seck

er-W

alke

ret

al41

0U

S;ur

ban

sett

ing;

RC

T39

9pr

egna

ntw

omen

wer

ere

crui

ted

for

the

stud

y.In

terv

entio

nw

omen

(n�

197)

rece

ived

stru

ctur

edad

vice

from

thei

rph

ysic

ian

and

refe

rral

toin

divi

dual

beha

vior

-ch

ange

coun

selin

gby

trai

ned

nurs

esdu

ring

pren

atal

care

,whi

leco

ntro

lwom

en(n

�20

2)re

ceiv

edbr

ief

advi

ceto

stop

smok

ing

and

aqu

it-sm

okin

gbo

okle

tat

thei

rfir

stvi

sit.

40%

ofth

ew

omen

inth

ein

terv

entio

ngr

oup

stop

ped

smok

ing

orre

duce

dsm

okin

gto

�50

%of

the

base

line

leve

lcom

pare

dto

25%

inth

eco

ntro

lgr

oup

(P�

.02)

.

No

sign

ifica

ntef

fect

ofth

ein

terv

entio

nw

asse

enon

birt

hw

eigh

tor

LBW

rate

.

Wis

borg

etal

407

Den

mar

k;ur

ban

hosp

itals

ettin

g;Q

T

3156

wom

enw

ere

recr

uite

dfo

rth

est

udy.

Preg

nant

wom

enin

the

inte

rven

tion

grou

p(n

�52

7)w

ere

give

nin

divi

dual

advi

cean

da

leaf

let

abou

tsm

okin

gce

ssat

ion

atth

efir

stan

tena

talv

isit

at�

16w

kof

gest

atio

n.Th

eco

ntro

lgro

up(n

�26

29)

did

not

rece

ive

this

educ

atio

n.

No

diff

eren

cew

asfo

und

betw

een

the

inte

rven

tion

and

cont

rolg

roup

sin

the

rate

ofsm

okin

gce

ssat

ion.

No

effe

ctw

asse

enon

birt

hw

eigh

tor

pret

erm

birt

hra

te.

Lefe

vre

etal

414

Swed

en;m

ulti-

cent

ertr

ial;

RC

TW

omen

inth

esc

reen

edgr

oup

(n�

7617

)ha

d2

scre

enin

gul

tras

ound

eval

uatio

nsto

asse

ssfe

talg

row

thto

asse

ssw

heth

ersc

reen

ing

wou

lden

cour

age

smok

ing

cess

atio

n.N

osp

ecifi

csm

okin

gce

ssat

ion

educ

atio

nw

aspr

ovid

edin

addi

tion

toth

eul

tras

ound

.Aco

ntro

lgro

up(n

�75

34)

was

also

stud

ied

for

preg

nanc

you

tcom

es.

No

effe

ctof

the

inte

rven

tion

was

seen

onPM

R,b

irth

wei

ght

orpr

eter

mbi

rth

rate

com

pare

dto

the

cont

rolg

roup

.

Liet

al41

1U

S;ur

ban

sett

ing;

RC

T81

4pr

egna

ntw

omen

wer

ere

crui

ted

for

the

stud

y.W

omen

inth

ein

terv

entio

nar

m(n

�40

0)w

ere

give

nfo

rmal

coun

selin

g,a

guid

eboo

k,in

divi

dual

smok

ing

cess

atio

nre

info

rcem

ent

and

soci

alsu

ppor

t.W

omen

inth

eco

ntro

lgr

oup

(n�

414)

rece

ived

stan

dard

wri

tten

risk

info

r-m

atio

nan

dve

rbal

advi

ceto

quit.

Wom

enin

both

grou

psw

ere

subd

ivid

edan

dan

alyz

edac

cord

ing

toou

tcom

e:qu

itter

s(q

uit

smok

ing)

,red

ucer

s(r

educ

ednu

mbe

rof

daily

ciga

rett

es),

and

no-c

hang

ers

(no

chan

gein

smok

ing

habi

ts).

Mot

hers

who

had

neve

rsm

oked

wer

eus

edas

aco

mpa

riso

ngr

oup.

Ther

ew

asa

14.3

%qu

itra

team

ong

wom

enin

the

expe

rim

enta

lgro

upve

rsus

an8.

4%qu

itra

team

ong

the

cont

rolw

omen

.

The

mea

nbi

rth

wei

ght

ofin

fant

sbo

rnto

the

quitt

ers

was

241

ghe

avie

rth

anth

ose

born

tono

-cha

nger

s(P

�.0

08)

and

redu

cers

(P�

.04)

.The

adju

sted

mea

nbi

rth

wei

ght

ofin

fant

sbo

rnto

redu

cers

was

92g

heav

ier

than

amon

gth

eno

-cha

nger

s(P

�.0

8).

LBW

was

abou

t2

times

mor

elik

ely

amon

gth

ein

fant

sbo

rnto

redu

cers

orno

-cha

nger

s,co

mpa

red

with

mot

hers

who

neve

rsm

oked

,but

this

findi

ngw

asno

tst

atis

tical

lysi

gnifi

cant

.Int

eres

tingl

y,th

epe

rcen

tage

ofLB

Win

fant

sbo

rnto

the

quitt

ergr

oup

(10.

3%)

was

com

para

ble

toth

egr

oup

that

had

neve

rsm

oked

(10.

9%).

Had

dow

etal

413

US;

mul

tiple

site

s;R

CT

Wom

enin

the

inte

rven

tion

grou

p(n

�28

48pr

egna

ntsm

oker

s)w

ere

obje

ctiv

ely

educ

ated

(usi

ngse

rum

cotin

ine

leve

lfor

inte

rpre

tatio

nof

the

need

for

smok

ing

cess

atio

n),a

ndef

fect

iven

ess

ofed

ucat

ion

was

asse

ssed

byco

mpa

riso

nof

cotin

ine

leve

lsin

wom

enin

the

inte

rven

tion

and

cont

rolg

roup

s.

No

effe

ctw

asse

enon

NM

R,I

MR

,bir

thw

eigh

t,LB

Wor

pret

erm

birt

hra

tes.

Hja

lmar

son

etal

408

UK

;urb

anho

spita

lse

ttin

g;R

CT

Wom

enin

the

inte

rven

tion

arm

(n�

492)

wer

egi

ven

asm

okin

gce

ssat

ion

man

ualt

hat

deta

iled

way

san

dm

eans

ofta

cklin

gan

dqu

ittin

gth

eha

bit.

Con

trol

s(n

�23

1)w

ere

not

give

nth

em

anua

l.

50%

mor

ew

omen

inth

ein

terv

entio

nar

mha

dco

ntin

uous

abst

inen

ce.

No

sign

ifica

ntef

fect

was

seen

onbi

rth

wei

ght,

LBW

orpr

eter

mbi

rth

rate

s.

Mac

arth

uret

al41

2U

K;u

rban

hosp

ital

sett

ing;

RC

TSu

pple

men

tary

heal

thed

ucat

ion

abou

tsm

okin

gin

preg

nanc

yw

asgi

ven

tow

omen

inth

ein

terv

entio

ngr

oup

(n�

493)

whi

leth

ose

inth

eco

ntro

lgro

up(n

�48

9)re

ceiv

edon

lyro

utin

ead

vice

.

The

mea

nda

ilyre

duct

ion

innu

mbe

rof

ciga

rett

essm

oked

inth

ein

ter-

vent

ion

grou

pan

dth

eco

ntro

lgro

upw

as2.

2vs

1.1,

resp

ectiv

ely.

The

first

born

infa

nts

inth

ein

terv

entio

ngr

oup

wer

ehe

avie

r(m

ean

diff

eren

ce68

g,P

�.0

6)an

dlo

nger

(mea

ndi

ffer

ence

0.75

cm,P

�.0

1)th

anth

ose

inth

eco

ntro

lgro

up.

Sext

onet

al40

9U

S;ur

ban

sett

ing;

RC

T93

5pr

egna

ntsm

oker

sw

ere

rand

omiz

edto

eith

eran

inte

rven

tion

ora

cont

rolg

roup

.Wom

enin

the

inte

rven

tion

grou

pre

ceiv

edst

aff

assi

stan

cefo

rsm

okin

gce

ssat

ion.

The

repo

rted

quitt

ing

rate

atth

eei

ghth

mon

thof

preg

nanc

yw

as43

%co

mpa

red

with

20%

inth

etr

eatm

ent

and

cont

rolg

roup

s,re

spec

tivel

y(P

�.0

1).

No

diff

eren

ces

wer

ese

enin

feta

lmor

talit

yor

gest

a-tio

nala

ge,b

utin

fant

sin

the

inte

rven

tion

arm

wer

ehe

avie

r(m

ean

diff

eren

ce92

g,P

�.0

5)an

dlo

nger

(mea

ndi

ffer

ence

0.6

cm,P

�.0

5)th

anth

ose

inth

eco

ntro

lgro

up.



complications of maternal infections, particularly forsettings in which antenatal care may be suboptimal.

A number of antiseptic solutions or preparationsmay be used for cleansing the maternal genital tract.However, the largest body of evidence and greatestmeasure of support is available for chlorhexidine.This agent has been used successfully in developedcountries for prophylaxis against newborn coloniza-tion and infection with group B streptococcus.420–426

community-based evidence. A study of the im-pact of chlorhexidine cleansing was undertaken inMalawi at an urban hospital, at a cost of $0.10 foreach maternal-infant pair treated (Table 25).427,428

Cleansing of the maternal vaginal canal and thenewborn skin with a 0.25% chlorhexidine solutionresulted in significant reductions in serious postpar-

tum maternal infections (P � .02), neonatal admis-sions (OR: 0.8; CI: 0.79–0.97), neonatal sepsis (OR:0.5; CI: 0.32–0.76), overall neonatal mortality (28.6 vs36.9 per 1000; P � .05), and neonatal mortality due toinfections (OR: 0.5; CI: 0.29–0.88; 2.4 vs 7.3 per 1000;P � .005).428 Among women with rupture of mem-branes for �4 hours, the HIV transmission rate wasdecreased by 40% (OR: 0.6; CI: 0.4–0.9). In anotherstudy in Kenya comparing intrapartum vaginal la-vage with 0.2% and 0.4% chlorhexidine, there was nooverall effect on transmission of HIV, but risk oftransmission was significantly reduced (OR: 0.1; CI:0.0–0.9) in the subgroup that had not had rupture ofmembranes before treatment.429

conclusions. Chlorhexidine cleansing of both thematernal vaginal canal and the newborn skin has

TABLE 24. Maternal Care Packages



Yan415 China; rural setting;PCS

7 of 29 townships were randomlyselected. A 3-tiered (village,township and county) MaternalChild Health network wasdeveloped in which MCHworkers at the communitylevel identified pregnantwomen and gave themantenatal care, with regularhigher-level supervision fromdoctors. A ward for newbornswas available at the countylevel for emergencies.

NMR declined by 34% (from26.6/1000 to 17.5/1000 births).

Antia725 India; rural setting(Mandwa Project);QT

In collaboration with the eldersof the villages, VHWs weretrained to take care of pregnantwomen. A small health centerwith 10 beds (chiefly formaternity care and tubectomy)was established.

Increased case detection andtreatment of leprosy andtuberculosis by VHWswas recorded.

The IMR was lower in theproject area compared to thenational figures (74/1000 livebirths vs 114 nationally forthat time period).

TABLE 25. Maternal Vaginal and Newborn Skin Antisepsis



Taha et al428 Malawi; urbanhospital setting;QT

3635 women giving birth to 3743babies were enrolled in theintervention phase and 3330women giving birth to 3417babies were enrolled in thenon-intervention phase.Intervention consisted ofmanual cleansing of the birthcanal and cleansing the babyimmediately after birth with a0.25% chlorhexidine gluconatesolution.

The intervention resulted inreductions in maternaladmissions related todelivery (29.4 vs 40.2/1000; P � .02) andserious postnatalmaternal infections(1.7/1000 vs 5.1/1000;P � .02).

The intervention resulted ina 22% reduction in NMRand a 67% reduction ininfection-associatedmortality. The interventionreduced serious neonatalillness requiring admission(16.9% vs 19.3%; P � .01),neonatal sepsis (7.8/1000vs 17.9/1000; P � .0002),overall neonatal mortality(28.6/1000 vs 36.9/1000;P � .06) and mortality dueto infections (2.4/1000 vs7.3/1000; P � .005).

Biggar et al427 Malawi; urbanhospital setting;RCT

The infection status of infants of3327 control women (con-ventional delivery procedures)was compared with that of3637 infants of intervention-delivered women. Interventionconsisted of washing thematernal birth canal and theskin of the newborn with0.25% chlorhexidine.

Among women with ruptureof membranes �4 h beforedelivery, chlorhexidinereduced HIV transmissionrate by 40% (OR: 0.6; CI:0.4–0.9). No significanteffect for women withrupture of membranes�4 h before delivery.



significant potential for reducing maternal and neo-natal postnatal infections and improving neonatal(and perhaps maternal) survival. Additional efficacytrials are needed to confirm these results in differentfacility-based settings, to determine the relative con-tribution of vaginal cleansing versus newborn skincleansing, and, thence, to evaluate potential applica-tions and feasibility of the intervention within thecommunity. Evaluation of higher-concentration for-mulations of chlorhexidine to prevent sepsis withpathogens from the skin34 and to prevent maternal-to-child transmission of HIV (1%) is also warranted.

Postnatal Interventions

Newborn Resuscitationbackground. Intrapartum hypoxia and birth as-

phyxia are widely regarded as major causes of mor-bidity and mortality in developing countries.18,430,431

“Birth asphyxia,” or failure to establish breathing atbirth, comprises just a portion of the burden of earlydeath from hypoxia. For every neonatal death fromasphyxia, there seems to be �1 additional fresh still-birth that occurs due to intrapartum hypoxia,430,432

although the precise burden of stillbirths is not yetwell defined.24 Evidence for birth asphyxia as a ma-jor cause of neonatal mortality is well established;however, the contribution of intrapartum hypoxiaand birth asphyxia to cerebral palsy and the overallburden of handicap in developing countries is un-clear.433 Previously, it was estimated that for everycase of mortality due to asphyxia or intrapartumhypoxia, another 4 newborns survived but sufferedsequelae.434 However, more recent data suggest thatthis may be a gross overestimation of the burden ofhandicap, because most newborns in developingcountries with severe asphyxia die.435

Approaches to improving birth asphyxia-relatedoutcomes may include prevention through improvedantenatal care such as birth preparedness; intrapar-tum care such as the presence of a skilled birthattendant and fetal monitoring (eg, use of a parto-graph); or improved management (eg, resuscitation)of newborns who do not breathe adequately at birth.The effects of antenatal and intrapartum interven-tions on birth asphyxia outcomes are thoroughlycovered in another recent review24 and are not in-cluded here. The important role of neonatal resusci-tation in immediate newborn care is well acceptedand forms a cornerstone of immediate newborn carein developed countries,17 but there are particularchallenges to making this intervention feasible indeveloping-country settings.24 Recent treatment mo-dalities and interventions such as cerebral hypother-mia436,437 therapy and the role of medications afterasphyxia are experimental and have been evaluatedonly in controlled facility-based situations in devel-oped countries. This latter group of interventions isnot evaluated further in this review.

community-based evidence. In addition to a hostof descriptive studies, we identified 13 studies fromdeveloping countries that evaluated various aspectsof neonatal resuscitation, including the feasibility ofprimary care staff undertaking the interventions

(Table 26, see also Table 39). The reported literaturesuggests that TBAs who are trained and supervisedin newborn resuscitation are capable of learningand properly using simple resuscitation tech-niques431,438–442 and that resuscitation may be per-formed successfully by trained health providers us-ing relatively simple equipment (eg, mouth-to-maskbreathing)443,444 and room air.445,446 Training hospi-tal staff in India447 and China448 in newborn resusci-tation has been shown to reduce asphyxia-relateddeaths.

It has been suggested that neonatal mortality maybe reduced after training of TBAs in resuscita-tion,431,449 although little objective data have beenprovided. Daga et al449 reported 41% and 62% reduc-tions in NMR and PMR, respectively, over a 3-yearperiod after introduction of a TBA training programthat included mouth-to-mouth resuscitation of as-phyxiated infants as a central component of the in-tervention. Successful resuscitation was reported in83% of cases of asphyxia. Kumar and Aggarwal431,442

reported a 70% reduction in asphyxia-specific mor-tality (P � .05) among infants delivered by selectedTBAs in rural India trained in advanced resuscitationusing a mucus sucker and bag-and-mask comparedwith TBAs trained in mouth-to-mouth resuscitation.Newborns delivered by TBAs with advanced train-ing also had a 19% lower PMR and a 20% lowercase-fatality rate, although neither outcome effectwas significant. A recent meta-analysis of the impactof TBA training programs found that neonatal mor-tality due to asphyxia was reduced 11% (CI: 2–21%).3It was uncertain, however, which aspects of thetraining programs were responsible for the impact.Most programs did not describe neonatal resus-citation as an important part of training; thus, theimpact seemed to be due to interventions other thanresuscitation.

In northern India, CHWs were trained in essentialnewborn care (clean cord cutting, maintenance ofwarmth, breastfeeding promotion); identificationand special care, including referral when indicated,of at-risk infants; and mouth-to-mouth resuscitationof asphyxiated infants.438 The NMR fell 25%, from51.9 to 38.8 per 1000 live births, during the 2 years ofthe study, with 18% of deaths attributed to asphyxia.In another trial of home-based neonatal essential andemergency care, including resuscitation of asphyxi-ated newborns by trained CHWs, Bang et al439 re-ported a 48% reduction in asphyxia-specific neonatalmortality. Training of grassroots-level health provid-ers in a Chinese province in methods of newbornresuscitation resulted in a reduction in the case fatal-ity rate from 7.1% to 0.45%.450

No data exist on the role of family members inneonatal resuscitation.

conclusions. Our review suggests that, althoughobjective evidence for the impact of community-based resuscitation efforts is sparse, there are prom-ising approaches and techniques that seem effectivein preliminary trials. Trained TBAs or CHWs seem tobe capable of learning resuscitation skills and savingnewborn lives; however, the feasibility of scaling up



TA

BL

E26

.N

ewbo

rnR

esus

cita

tion

Sour

ceL

ocat

ion

and

Typ

eof

Tri

alIn

terv

enti

onM

ater

nal

Out

com

ePe

rina

tal/

Neo

nata

lO

utco

me

Sibl

eyan

dSi

pe3

24co

untr

ies

in3

regi

ons;

met

a-an

alys

isof

60st

udie

s

TBA

trai

ning

was

asso

ciat

edw

itha

sign

ifica

nt(7

%;C

I:4–

9%)

redu

ctio

nin

birt

h-as

phyx

ia-s

peci

ficne

onat

alm

orta

lity.

This

tran

slat

edin

to11

%(C

I:2–

21%

)fe

wer

neon

atal

deat

hsdu

eto

birt

has

phyx

iaam

ong

wom

enca

red

for

orliv

ing

inar

eas

serv

edby

trai

ned

TBA

s.D

eora

riet

al44

7In

dia;

14m

edic

alco

llege

hosp

itals

;PC

STh

eN

atio

nalN

eona

talR

esus

cita

tion

Prog

ram

,beg

unin

1990

,tra

ined

and

cert

ified

150

pedi

atri

cian

san

dnu

rses

astr

aine

rsin

neon

atal

resu

scita

tion,

prov

idin

gth

emw

ithes

sent

iale

quip

men

t.Th

ese

core

facu

ltytr

aine

dan

addi

tiona

l12

000

heal

thca

repr

ovid

ers

over

2y,

and

resu

scita

tion

tech

niqu

esw

ere

adde

dto

med

ical

and

nurs

ing

curr

icul

a.Ea

chho

spita

lpr

ovid

edba

selin

eda

tafo

r3

mo

prio

rto

the

inte

rven

tion,

asw

ella

s12

mo

ofin

terv

entio

nda

ta.

Alth

ough

over

alln

eona

talm

orta

lity

was

not

affe

cted

,asp

hyxi

a-re

late

dde

aths

decl

ined

sign

ifica

ntly

(P�

.01)

.

Bang

etal

439

Indi

a;ru

rals

ettin

g;R

CT

VH

Ws

in39

inte

rven

tion

villa

ges

wer

etr

aine

dto

prov

ide

apa

ckag

eof

mat

erna

land

new

born

care

serv

ices

,inc

ludi

nghe

alth

educ

atio

n,cl

ean

deliv

ery,

neon

atal

resu

scita

tion

usin

gtu

bean

dm

ask

vent

ilatio

n,br

east

feed

ing

prom

otio

n,pr

even

tion

and

man

agem

ent

ofhy

poth

erm

ia,a

ndde

tect

ion

and

trea

tmen

tof

loca

linf

ectio

ns(e

g,sk

in,u

mbi

lical

cord

,as

wel

las

seps

is)

with

cotr

imox

azol

ean

dge

ntam

icin

.Res

ults

wer

eco

mpa

red

to47

cont

rol

villa

ges

with

out

trai

ned

VH

Ws.

Trai

ned

VH

Ws

care

dfo

r16

76ne

onat

esfr

om19

95–9

8.

Ther

ew

asa

48%

nons

igni

fican

tde

crea

sein

the

birt

h-as

phyx

ia-

spec

ific

NM

Rin

the

inte

rven

tion

area

(10.

5to

5.5

per

1000

live

birt

hs)

from

the

first

toth

eth

ird

year

ofth

ein

terv

entio

npe

riod

.O

vera

llN

MR

drop

ped

from

62/1

000

atba

selin

eto

25.5

/100

0at

inte

rven

tion

end.

Still

birt

hsdr

oppe

dfr

om32

/100

0to

25.9

/100

0,an

dPM

Rdr

oppe

dfr

om68

.3/1

000

to47

.8/1

000.

Saug

stad

etal

445

6co

untr

ies;

RC

T60

9in

fant

sw

ere

enro

lled.

New

born

sre

quir

ing

resu

scita

tion

wer

era

ndom

ized

tore

ceiv

ero

omai

r(n

�28

8)or

100%

oxyg

enre

susc

itatio

n(n

�32

1).

No

diff

eren

cew

asse

enin

mor

talit

yin

the

first

7d

or28

dof

life

betw

een

the

2gr

oups

.The

rew

ere

22%

few

erin

fant

sw

ithA

pgar

scor

es�

7af

ter

5m

inin

the

room

-air

grou

pco

mpa

red

toth

eox

ygen

resu

scita

tion

grou

p(2

4.8%

vs31

.8%

).M

edia

ntim

eto

first

brea

thw

as1.

1(C

I:1.

0–1.

2)m

inin

the

room

air

grou

p,vs

1.5

(CI:

1.4–

1.6)

min

inth

eox

ygen

grou

p.K

umar

723

Indi

a;ru

rals

ettin

g;co

mm

unity

-bas

edsu

rvei

llanc

e

Asu

rvei

llanc

esy

stem

was

crea

ted

for

the

trac

king

ofbi

rths

and

neon

atal

deat

hsin

54vi

llage

s.O

ver

ape

riod

of18

mo,

trai

ned

field

wor

kers

inte

rvie

wed

the

fam

ilym

embe

rw

how

aspr

esen

tat

the

time

ofch

ildbi

rth

in20

41de

liver

ies

with

in2

wk

afte

rth

ebi

rth.

The

inte

rvie

ws

docu

men

ted

birt

hhi

stor

y,us

eof

resu

scita

tion,

and

the

trai

ning

stat

usof

the

TBA

,in

clud

ing

adva

nced

trai

ning

.

Birt

has

phyx

iapr

eval

ence

was

0.9%

amon

gba

bies

deliv

ered

byTB

As

with

adva

nced

trai

ning

,in

com

pari

son

to2.

4%in

babi

esde

liver

edby

conv

entio

nally

trai

ned

TBA

s(P

�.0

5).T

hem

orta

lity

rate

spec

ific

tobi

rth

asph

yxia

was

70%

less

amon

gba

bies

deliv

ered

byTB

As

with

adva

nced

trai

ning

,in

com

pari

son

toco

nven

tiona

llytr

aine

dTB

As

(P�

.05)

.The

fata

lity

rate

ofas

phyx

iaca

ses

was

20%

low

eran

dth

ePM

Rw

as19

%lo

wer

(49.

4/10

00vs

61.0

/100

0)am

ong

new

born

sde

liver

edby

TBA

str

aine

din

use

ofre

susc

itatio

neq

uipm

ent

com

pare

dto

thos

ede

liver

edby

TBA

str

aine

dfo

rm

outh

-to-

mou

thbr

eath

ing.

How

ever

,giv

enth

esm

alls

tudy

size

the

diff

eren

cedi

dno

tre

ach

stat

istic

alsi

gnifi

canc

e(P

�.0

5).



TA

BL

E26

.C

onti

nued

Sour

ceL

ocat

ion

and

Typ

eof

Tri

alIn

terv

enti

onM

ater

nal

Out

com

ePe

rina

tal/

Neo

nata

lO

utco

me

Zhu

etal

448

Chi

na;h

ospi

tal-

base

dse

ttin

g;PC

SA

pros

pect

ive

stud

yof

4751

new

born

s,36

6of

who

mw

ere

asph

yxia

ted

and

man

aged

bya

neon

atal

resu

scit

atio

npr

ogra

m(N

RPG

),w

asco

nduc

ted

over

the

peri

odof

2y.

Thi

sgr

oup

was

com

pare

dto

aco

ntro

lgr

oup

com

pris

edof

1722

live-

born

infa

nts

und

erth

etr

adit

iona

lre

susc

itat

ion

(TR

)pr

ogra

m.

The

NR

PGw

asas

soci

ated

wit

ha

65.7

%re

duc

tion

inea

rly

NM

R.

Dur

ing

impl

emen

tati

onof

the

NR

PG,o

nly

16in

fant

sou

tof

4751

birt

hs(0

.34%

)d

ied

wit

hin

7d

,and

2of

the

dea

ths

occu

rred

inth

ed

eliv

ery

room

.In

cont

rast

,17

new

born

sou

tof

1722

birt

hs(0

.99%

)m

anag

edw

ithi

nth

eT

Rpr

ogra

m,d

ied

wit

hin

7d

,wit

h10

ofth

ose

dea

ths

occu

rrin

gin

the

del

iver

yro

om.A

sa

resu

ltof

the

impl

emen

tati

onof

the

NR

PG,n

eona

tal

mor

talit

yw

asre

duc

edal

mos

t3

tim

es(�

2�

10.5

4,P

�.0

1).2

0of

21in

fant

sw

ith

seve

reas

phyx

iaw

ere

norm

al;1

had

cere

bral

pals

y.K

amen

ir44

4K

enya

;rur

alho

spit

al;

open

stud

y:re

tro-

spec

tive

char

acte

ri-

zati

onof

del

iver

ypr

acti

ces

and

risk

fact

ors

and

pros

pect

ive

asse

ssm

ent

ofre

susc

itat

ion

prac

tice

s

The

use

ofro

utin

ere

susc

itat

ion

(inc

lud

ing

nasa

lan

dor

opha

ryng

eal

suct

ioni

ng,d

ryin

gan

dpr

oper

stim

ulat

ion)

was

eval

uate

din

new

born

sw

ith

birt

has

phyx

ia(n

�87

8).I

fth

ene

onat

ere

mai

ned

apne

ic,b

ag-a

nd-m

ask

vent

ilati

onw

asst

arte

dw

ith

oxyg

en(i

fav

aila

ble)

.

4%of

878

new

born

sw

ith

asph

yxia

suff

ered

unfa

vora

ble

outc

omes

.R

isk

fact

ors

for

unfa

vora

ble

outc

ome

incl

uded

del

iver

ies

othe

rth

ansp

onta

neou

sno

rmal

vagi

nal

del

iver

ies

and

infa

nts

wei

ghin

g�

2000

gat

birt

h.

Mas

saw

eet

al44

3In

dia

and

Tan

zani

a;ur

ban

univ

ersi

tyho

spit

als;

QT

Mou

th-t

o-m

ask

(MM

)ve

ntila

tion

was

com

pare

dw

ith

bag-

and

-mas

k(B

M)

vent

ilati

on.1

74ba

bies

wer

est

udie

d;5

4(3

0M

Man

d24

BM

)w

ere

born

inB

omba

yan

d12

0(5

6MM

and

64B

M)

inD

ar-e

s-Sa

laam

.

The

MM

and

the

BM

met

hod

sw

ere

equa

llyef

fect

ive

inre

susc

itat

ion

ofas

phyx

iate

dne

onat

es.T

here

wer

eno

sign

ific

ant

dif

fere

nces

betw

een

the

2tr

eatm

ent

grou

psby

Apg

arsc

ore

�4

at5

and

10m

in,n

umbe

rof

babi

esw

ith

firs

tbr

eath

�5

min

,nu

mbe

rof

babi

esw

ith

hear

tra

te�

130

beat

s/m

in,o

rnu

mbe

rof

babi

esw

ith

puls

eox

imet

erva

lues

�75

%at

5m

in.A

t5

min

,75

%of

all

infa

nts

had

Apg

arsc

ores

�4.

InT

anza

nia,

low

resp

irat

ory

freq

uenc

yw

asas

soci

ated

wit

ha

slow

incr

ease

inhe

art

rate

abov

e13

0be

ats/

min

.K

umar

431

Ind

ia;r

ural

sett

ing;

PCS

Sim

plif

ied

met

hod

sof

resu

scit

atio

nw

ere

taug

htto

TB

As.

An

add

itio

nal

grou

pre

ceiv

edad

vanc

edtr

aini

ngon

use

ofth

em

ucus

extr

acto

ran

dba

g-an

d-m

ask

vent

ilati

on.O

vera

ll,20

41bi

rths

wer

ere

port

ed,s

ome

del

iver

edby

conv

enti

onal

lytr

aine

dT

BA

s(n

�96

8)an

dso

me

byT

BA

sw

ith

adva

nced

trai

ning

(n�

911)

.58

infa

nts

wer

eas

phyx

iate

dor

wer

efr

esh

still

birt

hs,2

0of

who

mw

ere

del

iver

edby

TB

As

wit

had

vanc

edtr

aini

ng.

The

rew

asa

19%

dec

reas

ein

the

PMR

amon

gth

egr

oup

ofw

omen

who

wer

eat

tend

edto

byad

vanc

edtr

aine

dT

BA

s.

Xia

oyu4

50C

hina

;rur

alse

ttin

g;PC

SA

nex

plor

ator

yst

udy

was

crea

ted

toin

trod

uce

mod

ern

resu

scit

atio

nto

gras

sroo

tsm

ater

nal

and

child

heal

thpe

rson

nel.

Thi

sst

udy

incl

uded

trai

ning

cour

ses

onne

onat

alre

susc

itat

ion,

on-t

he-

spot

teac

hing

wit

hre

peti

tion

ofke

yte

chni

cal

proc

edur

es,a

ndth

eop

erat

iona

lse

quen

ceof

the

5st

eps

ofth

eA

BC

DE

prot

ocol

(air

way

,bre

athi

ng,

circ

ulat

ion,

dis

abili

ty,a

ndex

posu

re).

The

appr

oach

high

light

edth

eim

port

ance

ofen

dot

rach

eal

intu

bati

onan

dpr

acti

cing

neon

atal

resu

scit

atio

nin

pers

on.I

nal

l,22

3ne

wbo

rns

inth

epr

ovin

cew

ere

resu

scit

ated

.

Asp

hyxi

a-sp

ecif

icN

MR

was

red

uced

86%

byus

ing

the

stri

ctap

plic

atio

nof

the

AB

CD

Epr

otoc

ol.2

23ne

wbo

rns

wer

ere

susc

itat

edin

the

6m

oof

the

proj

ect

wit

hon

ly1

dea

th(a

mor

talit

yra

teof

0.45

%).

All

14ca

ses

(100

%)

ofse

vere

lyas

phyx

iate

dba

bies

reco

vere

dan

dw

ere

wel

lup

ond

isch

arge

.By

cont

rast

,out

of18

4ba

bies

man

aged

prio

rto

intr

oduc

tion

ofth

epr

ogra

m,1

3ba

bies

die

dof

asph

yxia

(am

orta

lity

rate

of7.

1%)

(�2

�13

.29;

P�

.01)

.



TA

BL

E26

.C

onti

nued

Sour

ceL

ocat

ion

and

Typ

eof

Tri

alIn

terv

enti

onM

ater

nal

Out

com

ePe

rina

tal/

Neo

nata

lO

utco

me

Ram

jiet

al44

6In

dia

;urb

anho

spit

alse

ttin

g;R

CT

Con

secu

tive

asph

yxia

ted

new

born

infa

nts

(n�

84)

wer

eal

loca

ted

toei

ther

resu

scit

atio

nw

ith

room

air

(n�

42)

or10

0%ox

ygen

(n�

42).

No

sign

ific

ant

dif

fere

nces

wer

eno

ted

betw

een

the

2gr

oups

for

anu

mbe

rof

outc

omes

incl

udin

gd

urat

ion

ofas

sist

edve

ntila

tion

,ti

me

tofi

rst

brea

than

dcr

yan

dne

onat

alm

orta

lity.

Prel

imin

ary

dat

ain

dic

ated

that

resu

scit

atio

nw

ith

room

air

may

beas

effe

ctiv

eas

100%

oxyg

enin

neon

atal

resu

scit

atio

n.D

aga

etal

449 *

Ind

ia;r

ural

sett

ing;

trai

ning

asse

ssm

ent

ofco

mm

unit

y-ba

sed

wor

kers

InG

anja

dpr

imar

yhe

alth

cent

er,t

rain

ing

ofD

ais

and

Ang

anw

adi

wor

kers

inne

wbo

rnca

rein

clud

edpr

ovis

ion

ofw

arm

th,r

esus

cita

tion

ofas

phyx

iate

dne

wbo

rns,

and

iden

tifi

cati

onan

dre

ferr

alof

infa

nts

wit

hfo

otle

ngth

�6.

5cm

ina

popu

lati

onof

2000

0.A

ngan

wad

iw

orke

rssu

ppor

ted

dais

inid

enti

fyin

gan

dm

akin

gre

ferr

als

and

wor

ked

asa

link

betw

een

dais

and

auxi

liary

nurs

e-m

idw

ives

.

The

rew

asa

70%

dec

reas

ein

IMR

over

ape

riod

of4

y.

Dag

aet

al44

1 *In

dia

;rur

alse

ttin

g;PC

SD

ais

(TB

As)

(n�

67)

wor

king

ina

popu

lati

onof

2224

0w

ere

trai

ned

topr

ovid

ew

arm

thto

the

infa

nt,m

outh

-to-

mou

thre

susc

itat

ion

ofas

phyx

iate

din

fant

s,id

enti

fica

tion

ofL

BW

and

pret

erm

infa

nts

and

safe

tran

spor

tati

onof

high

-ri

skin

fant

sto

prim

ary

heal

thce

nter

s.D

ais

repo

rted

30ne

onat

alas

phyx

iaca

ses

over

the

3y

afte

rth

epr

ogra

mbe

gan.

Ant

enat

alre

gist

rati

onof

preg

nant

wom

enin

1990

show

eda

30%

incr

ease

over

pre-

prog

ram

leve

ls.

The

PMR

dec

lined

by61

.5%

,sti

llbir

thra

ted

ecre

ased

by51

.4%

,an

dN

MR

dec

reas

edby

42%

com

pare

dto

base

line

over

ape

riod

of3

y.

Prat

inid

hiet

al43

8In

dia

;rur

alse

ttin

g;PC

SC

HW

sin

ast

udy

popu

lati

onof

4700

0w

ere

trai

ned

inri

skid

enti

fica

tion

(LB

W,s

mal

lsi

ze,p

rete

rmbi

rth,

feed

ing

prob

lem

s)an

din

vari

ous

tech

niqu

esof

new

born

care

,inc

lud

ing

mou

th-t

o-m

outh

resu

scit

atio

n.H

ome

visi

tsw

ere

don

efo

rsc

reen

ing

and

man

agem

ent

ofat

-ris

kin

fant

s(n

�85

1)an

ded

ucat

ing

mot

hers

onne

wbo

rnca

re.F

ollo

w-u

pvi

sits

wer

efi

xed

for

the

8th

and

29th

day

afte

rbi

rth.

Whe

non

ly1

risk

fact

orw

aspr

esen

t,th

eT

BA

sre

com

men

ded

dom

icili

ary

care

byth

em

othe

ran

dC

HW

usin

ghe

alth

educ

atio

nun

der

the

supe

rvis

ion

ofnu

rse

ord

octo

r.W

hen

mor

eth

an1

risk

fact

orw

aspr

esen

t,th

em

othe

ran

d/

orin

fant

wer

ere

ferr

edto

the

hosp

ital

for

inpa

tien

tca

re.T

here

wer

e30

83liv

ebi

rths

reco

rded

dur

ing

the

3y

prio

rto

the

stud

y,an

d29

90liv

ebi

rths

reco

rded

dur

ing

the

2st

udy

year

s.

Ris

k-d

etec

tion

rate

byC

HW

sw

as78

%.T

hest

illbi

rth

rate

dec

reas

edby

25%

,fro

m28

.4/

1000

birt

hsto

21.5

/10

00bi

rths

.T

heN

MR

dec

reas

edby

25%

from

base

line

over

the

2-y

proj

ect

peri

od,f

rom

51.9

/10

00liv

ebi

rths

to38

.8/

1000

live

birt

hs,b

utth

isd

iffe

renc

ew

asno

tst

atis

tica

llysi

gnif

ican

t.

*D

ata

are

from

the

sam

etr

ial.



this approach is unclear. Supervision and mainte-nance of skills are critical to success, yet formidablechallenges exist at the community level. One of theareas of greatest need is to define feasible and cost-effective approaches by which community-level pro-viders (eg, CHWs, community midwives) can pre-vent fresh stillbirths and neonatal deaths due tointrapartum hypoxia/birth asphyxia. There is evi-dence that mouth-to-mask and bag-and-mask resus-citation are comparable techniques with regards toneonatal mortality and morbidity443; however, theformer was deemed more difficult for the providers.Thus, where feasible, bag-and-mask ventilation hasbecome the preferred modality.439 The evidence thatroom air could be satisfactorily used for neonatalresuscitation suggests that staff training alone andbasic equipment may be sufficient for resuscitationof asphyxiated newborns in developing-countrycommunities.

Additional large-scale effectiveness trials of suchintervention strategies with defined criteria and end-points are clearly required. More precise definitionsof “birth asphyxia” at the community level and im-proved verbal autopsy instruments for identificationof fresh stillbirths and neonatal deaths due to acuteintrapartum hypoxia are urgently needed. The cost-effectiveness and sustainability of resuscitation usingvarious modalities and the impact of resuscitation ondisability rates have not been documented. The var-ious components of resuscitation, such as the role ofpharyngeal suction, the need for oxygen therapy,and the role of ventilation by bag-and-mask (as op-posed to other modalities such as tube-and-mask,mouth-to-mask, or mouth-to-mouth), also requireadditional scientific scrutiny. Meanwhile, it is rea-sonable to promote the basic elements of newbornresuscitation (namely, drying, stimulation, andwarming) as a routine part of newborn care and tofurther evaluate the potential role of trained TBAsand CHWs in using equipment such as the bag-and-mask to ventilate nonbreathing newborns. In addi-tion, all skilled attendants should be trained to pro-vide proper newborn stimulation and ventilation.

Delayed Umbilical Cord Clampingbackground. There is little evidence as to

whether the current practice of umbilical cord clamp-

ing soon after birth, to prevent polycythemia on theone hand or anemia on the other, is based on solidscientific criteria. Some studies have demonstratedthat delayed cord clamping (after the cord stopspulsating) may increase neonatal blood volume byapproximately one third.451,452 Thus, delayed cordclamping may increase the newborn infant’s iron re-serves and reduce the incidence of iron-deficiency ane-mia in infancy, an issue of considerable public healthimportance.453

community-based evidence. In the course of in-troducing a program to reduce tetanus incidence inHaiti, it was observed that nearly half of familiesceased cutting the cord at home after delivery andinstead bundled the infant, placenta, and cord andsought cord care at the hospital.385 Despite delays incord cutting for up to several hours, neonatal mor-tality declined dramatically and no adverse effectson neonatal outcomes were noted.

We identified only 1 comparative trial from a devel-oping country that evaluated this issue in an urbanhospital setting (Table 27). Delaying umbilical cordclamping until the cord ceased pulsating was associ-ated with an increase in red blood cell count in theneonatal circulation regardless of whether the infantwas level with or below the level of the placenta.454

conclusions. Despite the potential importance ofthe timing of cutting the cord, there are little data onits relationship to neonatal health. No data are avail-able from developing countries on whether the com-mon practice of waiting until after the placenta hasbeen delivered to cut the cord has a beneficial oradverse impact on neonatal circulation and health.However, the scarce available evidence does indicatethat delayed cord clamping may be beneficial, be-cause it may allow for greater transfer of placentalblood to the newborn’s circulation before cutting thecord. Additional studies are needed to resolve thisissue, because it may have important implications forprevention of anemia in infancy in susceptible pop-ulations in developing countries.

Umbilical Cord Antisepsisbackground. The importance of clean cord cut-

ting for the prevention of neonatal tetanus was con-sidered in the section on tetanus prevention (see “TT

TABLE 27. Delayed Umbilical Cord Clamping


Intervention Perinatal/NeonatalOutcome

Grajeda et al454 Guatemala; urban hospitalsetting; RCT

88 infants were enrolled, out of which73 were followed up at 2 mo ofage. 69 blood samples wereavailable. In group 1 (n � 29) thecord was clamped immediatelyafter birth. Group 2 infants (n � 30)had the cord clamped afterpulsations stopped but with theinfants kept at the same level as themother’s heart. Group 3 infants(n � 29) had delayed cordclamping at cessation of pulsationsbut the infants were placed at alower level.

Infants in the groups withdelayed cord clampinghad significantly higherhematocrit values 2 moafter delivery comparedwith the control group.There were no differ-ences between groups 2and 3.



Immunization and Clean Delivery” and Table 20).Studies of the impact of topical applications of anti-septics to the umbilical cord in the postnatal periodgenerally have shown reductions in bacterial coloni-zation compared with no treatment.34 However, im-pact has varied with the specific antimicrobial com-pound, the mode or frequency of application, andthe degree of contamination in the environment. Ingeneral, chlorhexidine seems to be the most favor-able choice of antiseptic because of its broad spec-trum of activity, residual effects on the skin, lowtoxicity, and results showing overall reductions incord colonization after treatment.34,455

Evidence for the effect of topical antiseptic treat-ment of the cord in reducing local cord or skin infec-tions or sepsis is less clear because of conflictingreports.34 Thus, several neonatal skin care reviewshave called for discontinuation of routine topicalantiseptic care of the cord based on conflicting evi-dence regarding which topical agent most effectivelydecreases colonization and the lack of a clear rela-tionship between colonization and infection.456,457

However, in many cases, when antiseptic treatmentshave been discontinued, increased rates of cord col-onization and infection have been found.458,459 Ingeneral, wide variation in trial design, small samplesizes with insufficient power to assess impact oninfections, and inconsistent comparisons betweendifferent regimens have contributed to the largelyconflicting literature on the impact of cord-careregimens.

A Cochrane review on different methods of neo-natal cord care, including the use of topical antimi-crobials, suggested that the incidence of omphalitisand skin infections within 6 weeks of observationwas not affected by use of antiseptics.460 There was atrend toward reduced microbial colonization withthe use of antibiotics compared with no treatment.Antiseptics such as chlorhexidine, however, pro-longed the time to cord separation. It was concludedthat simply keeping the cord clean was probably aseffective as topical antibiotic use. However, this re-view was based on 10 RCTs, all conducted in devel-oped countries, none of which reported any systemicinfections or neonatal deaths.

In addition to the aforementioned Cochrane re-view, the WHO461 also reviewed the evidence froma host of descriptive and physiologic studies andconcluded that medical cord-care practice recom-mendations are overwhelmingly based on researchfrom hospital nurseries in developed countries. Itwas concluded that, compared with no treatment,application of a topical antimicrobial to the cordstump reduced umbilical colonization by harmfulbacteria in hospital nurseries. Chlorhexidine, tinc-ture of iodine, povidone-iodine solutions, tripledye and silver sulfadiazine were all found to beeffective in reducing microbial colonization, al-though caution was issued regarding the need tobalance their use with the potential for emergenceof antimicrobial resistance. In general, dry cord carewas recommended along with the use of soap-and-water solution to clean the cord when visibly soiled.

community-based evidence. There is a largebody of literature outlining the increased risk of um-bilical infection, mainly due to Clostridium tetani col-onization, after application of unclean substancessuch as ash or mud to the cord of neonates in devel-oping countries.462–464 Observational studies haveshown that application of antimicrobial agents to thecord after cutting was protective against tetanus andresulted in reduced neonatal mortality and morbid-ity381,387,463 (Table 28). In a large CCS of tetanusdeaths in Pakistan, Bennett et al388 showed that in-fants who received applications of antimicrobials(type unspecified), both at birth and subsequently,were at significantly less risk of death than thosewho received dry cord care alone (OR: 0.2; CI: 0.06–0.58), even after adjusting for use of unclean sub-stances such as cow dung, ash, or ghee (clarifiedbutter) (OR: 0.4; CI: 0.21–0.77). In Papua NewGuinea, neonates who received daily application of10% acriflavine in spirit to the umbilical cord were9.4 (P � .02) and 6.7 (P � .01) times less likely to havesepsis or fever, respectively, than those who did notreceive the antiseptic applications.463

conclusions. There is no definitive answer to thequestion of what constitutes the best form of cordcare after birth, particularly in domiciliary settings.More research is needed on this issue, especially insituations of limited resources and high potential forenvironmental contamination of the cord. Althoughthere is evidence to suggest that cord antisepsis maybe beneficial, there is insufficient evidence to recom-mend the widespread use of topical antimicrobialson the cord stump. Nevertheless, the decision to usethem may depend on local circumstances.

Rooming-in and keeping the infant with themother also significantly reduces the incidence ofcolonization with pathogenic bacteria and cord infec-tions.465,466 Thus, early colonization of the newbornwith commensal flora from the mother, facilitated byearly and prolonged contact as with KMC, may beprotective and may have contributed to the reduc-tion in infections observed in cohorts who practicedKMC (see “KMC” and Table 35). Similarly, for homedeliveries and for cord care after discharge fromhospital in developed countries, clean cord careseems to be sufficient, and the application of antisep-tics may not be required. However, for developing-country communities in which the majority of new-borns are delivered at home and may face severeimmunologic challenge from pathogenic bacteria inthe environment, the benefits of topical antisepticsremain unknown. Provision of clean delivery kitsmay increase compliance with clean cord cutting andtying after birth, although this has not been demon-strated (see “TT Immunization and Clean Delivery”and Table 20). Topical antimicrobial applicationsmay be a useful strategy for other instances in whichpotentially harmful practices such as cow dung ap-plication on the cord stump are widely prevalent,even if just for replacement of the harmful prac-tices.388,461 If chosen, the antimicrobial should have abroad spectrum of activity against Gram-positive



and Gram-negative pathogens and should be inex-pensive, culturally acceptable, and widely available.

Additional research is needed to define best cord-care practices in developing-country communities.Meanwhile, efforts must focus also on providing cul-turally appropriate behavior-change communica-tions to communities and caregivers on proven in-terventions such as clean cord-care practices,particularly hand-washing and use of a clean bladeto cut the cord.

Hypothermia Prevention and Managementbackground. Hypothermia in the newborn pe-

riod is widely regarded as a major contributory causeof morbidity in developing countries,467 although ithas been poorly documented. Hypothermia has beenassociated with increased risk of infection, coagula-tion defects, acidosis, delayed fetal-to-newborn cir-culatory adjustment, hyaline membrane disease, andbrain hemorrhage.468,469 At its extreme, hypothermiamay manifest as neonatal cold injury470,471 and beassociated with significant mortality.439 This risk is

particularly marked among LBW infants in develop-ing countries and is seen commonly even in coun-tries with tropical climates. Studies in Nepal usingcontinuous temperature monitoring have indicatedthat thermal stress may be extremely commonamong newborn infants; 80% of hospital-born infantsbecame hypothermic soon after birth.472 In anothermaternity hospital study in Nepal, 85% of newbornshad a temperature �36°C within 2 hours of birth.473

Similarly, in facility-based studies in Ethiopia,474

Zambia,475 and Zimbabwe,476 one half to two thirdsof newborns had hypothermia.

Few studies have addressed practices during andafter birth, which can put the newborn at risk forhypothermia. In a multicenter evaluation in LatinAmerica, Asia, and Africa, 65% and 73% of healthfacility workers had adequate knowledge regardingcauses and prevention, respectively, of hypothermia,yet interventions to prevent hypothermia were sel-dom provided (eg, 0% warmed the delivery room,11% dried and wrapped the infant, and 50% and 61%

TABLE 28. Umbilical Cord Antisepsis



Parashar et al381 Bangladesh; ruralsetting; CCS

For each neonatal tetanus case identified(n � 359), 3 living infants matched forgender, residence and date of birthwere selected as controls.

Application of antibiotics to the umbilicalstump at delivery decreased neonataltetanus by 79% (OR: 0.21; P � .019).Hand-washing by the deliveryattendant and prior immunization withTT were protective in multivariateanalyses, decreasing tetanus rates by36% (OR: 0.64; P � 0.005) and 50%(OR: 0.50; P � .001), respectively.

Bennett et al388 Pakistan; ruralsetting; CCS

Cases (n � 211) were uncircumcisedinfants diagnosed as having definiteor probable neonatal tetanus. 3controls who survived the neonatalperiod and whose mothers had notreceived TT were selected per case(n � 633).

Reported hand-washing by the deliveryattendant was highly protective againstneonatal tetanus, decreasing rates by56% (OR: 0.44; CI: 0.29–0.67; P �.0002). Continuous use of antimicrobialagents (antibiotics and antiseptics) onthe cord at delivery and during thefirst few days after delivery was highlyprotective in univariate testing(matched OR: 0.2; CI: 0.11–0.64;P � .003), and remained significantlyprotective when other delivery andcord-care practices were controlled for.

Garner et al463 Papua New Guinea;rural setting; QT

Newborns (n � 132) were divided into3 groups. Infants in the control group(n � 64) were born to motherswithout birth kits. Infants born theinitial part of the intervention phase(n � 33) were born to mothers whohad been supplied with a new razorblade and acriflavine 10% in spirit.Infants born during the later part ofthe intervention phase (n � 34) wereborn to mothers who had beensupplied with a razor blade,acriflavine 10%, and plastic umbilicalcord clamps which the women hadbeen instructed how to use.

Of the 9 cases of neonatal sepsis, 8 werein the control group and 1 was in theintervention group (P � .02).

Traverso et al387 Pakistan; ruralsetting; CCS

Tetanus cases (n � 102) were identifiedat an urban hospital and 3 controls(n � 306) were selected from thevillage of each case.

Delivery conducted by formally trainedpersonnel was associated with adecreased risk of neonatal tetanus(OR: .28; CI: 0.08–1.0). Use of a topicalantibiotic was also associated with aprotective effect, decreasing neonataltetanus rates by 66% (OR: 0.34; CI:0.18–0.65).



provided special protection during transport or toLBW infants, respectively).477 In 1 village-basedstudy in India, 11% of 189 neonates were found to behypothermic based on a single temperature readingtaken within the first 24 hours after birth. Only 58%of newborns were wiped soon after delivery, thehead was covered in 59% in winter and 10.5% insummer, no newborns were kept in skin-to-skin con-tact, and the room temperature was �24°C in 41% ofhouseholds.442 Early bathing and removal of vernixare also recognized but understudied risk factors forhypothermia. However, based on expert opinion, theWHO recommends that bathing be delayed for atleast 6 hours after birth to minimize the risk of coldstress during the period of maximum physiologictransition of the newborn.469

Although it has been demonstrated that physiciansand trained assistants in health facilities are capableof perceiving newborn body temperature by touchwith reasonable accuracy,441,478,479 in a community-based study in India, mothers using touch alone hadlimited ability to detect hypothermia.480 Overall,only one fourth (24.6%) of hypothermic infants werecorrectly identified as such by the mothers. A tem-perature-indicator device (ThermoSpot) placed onthe newborn’s skin has been shown to provide anaccurate indicator of the presence of hypothermiaand is acceptable to mothers in hospital set-tings,476,481–484 but additional research on its utilityand use in resource-poor communities, especially bymothers, is needed.

It is now well recognized that rapid rewarming ofhypothermic newborns in developed-country set-tings may be entirely safe and result in lower mor-tality compared with slower methods of rewarm-ing.485,486 The WHO has recognized the importanceof thermal care of the newborn by including it as apriority behavior in the Mother-Infant Package ofinterventions in developing countries,14 and the Sav-ing Newborn Lives initiative of Save the Children/USA considers it an essential newborn care prac-tice.17 Accepted ways to achieve optimal thermalcontrol of the newborn include warming of the room,immediate drying and wrapping after birth, imme-diate and frequent breastfeeding, delay in bathinguntil the infant is physiologically stable, close contactwith the mother such as skin-to-skin contact, andappropriate swaddling and dressing, including theuse of head cover. However, few studies have beenconducted to evaluate feasible strategies for the pre-vention, recognition and management of hypother-mia in developing countries. Recent emphasis hasbeen on the use of skin-to-skin contact or KMC in theprevention of hypothermia, and this strategy will beconsidered separately (see “KMC”). Other potentialinterventions (eg, incubators, semipermeable plasticsheeting) commonly used to reduce transepidermalwater loss (TEWL) and heat loss, particularly in pre-term infants in developed countries,455 have not beenadequately evaluated in developing-country set-tings, are largely impractical for community settings,and will not be considered further here.

community-based evidence. Topical applicationof products such as paraffin-based ointments or oilsaimed to modify skin-barrier function and reducetransepidermal loss of water and heat455,487–490 hasbeen evaluated in 2 hospital-based studies in devel-oping countries473,491 (Table 29). In India, topicaltherapy of preterm infants with corn oil every 4hours resulted in a significant reduction (P � .001) inneed for an external source of heat to maintain nor-mal body temperature.491 Another study in Nepalfound that traditional oil massage with mustard oil,swaddling with a plastic swaddler, or KMC wasequally effective in preventing hypothermia duringthe first 24 hours after birth.473

During neonatal transport, a simple Styropor boxshowed promise for preventing hypothermia.492,493

Other trials have included elements of hypothermiaprevention in a package of postnatal interven-tions,438,439,494,495 but the specific impact of the ther-mal control interventions could not be determined.

conclusions. Despite the critical importance ofmaintaining warmth for the newborn, there is a re-markable paucity of studies evaluating and com-menting on community-based prevention, recogni-tion, and/or management of hypothermia. There is aclear need for additional research on appropriate andcost-effective strategies for prevention of hypother-mia in health system settings, including transportand care of high-risk LBW infants. Approaches suchas training TBAs, CHWs, and caregivers to recognizeand respond appropriately to warning signs of hy-pothermia, for example, have not been reported. Al-though heated cots have been scientifically evaluatedin institutional settings,485 these are too expensive foruse in most developing countries.494 The Styroporbox offers an excellent example of simple technologythat could be used as an alternative to transportincubators in developing countries, but additionalevaluation is warranted to compare it with otheroptions. Preliminary studies suggest that topical ap-plications of skin barrier–modifying products maybe beneficial for preservation of body heat.473,491

Nevertheless, most cases of hypothermia likely couldbe prevented by a few simple behaviors such asimmediate drying and wrapping of the infant, in-cluding the head; immediate breastfeeding; delay inthe initiation of bathing; close contact with the moth-er; and keeping the room warm and the infantproperly clothed and/or wrapped. In addition, KMCmay be a highly cost-effective intervention for pre-venting hypothermia, especially of LBW infants (see“KMC”). Nevertheless, even when health workerspossess knowledge of how to prevent hypothermia,this knowledge may not be translated into prac-tice.477 Thus, there is a great need to develop andevaluate the impact of culturally appropriate behav-ior-change communications to promote these health-ful domiciliary thermal control practices.

Hypoglycemia Prevention and Managementbackground. Hypoglycemia after birth is a major

cause of morbidity, particularly among intrauterinegrowth-restricted and preterm infants. The risk ofhypoglycemia is significantly greater among preterm



infants because of their reduced energy and glycogenreserves and inability to mobilize alternative meta-bolic fuels.496 Hypoglycemia is also relatively com-mon among LBW infants and macrosomic infants ofdiabetic mothers.497 Prevention and management ofneonatal hypoglycemia have been the subjects of amajor review by Williams.498

Little data on hypoglycemia incidence or impacton newborn health are available from developingcountries, in part due to difficulties in detection,which requires proper processing in a well-function-ing laboratory. Studies in Nepal indicated that 38%of term infants had hypoglycemic episodes afterbirth.499,500

The most cost-effective strategy for preventing hy-poglycemia is early feeding (continued every 2 to 3hours on demand day and night) with breast milk,which is superior to milk formula in that it canpromote relatively greater ketogenesis501 and has arelatively lower insulinogenic effect.502 The benefitsof breast milk feeding are comparatively moremarked in preterm infants.503 Although nasogastricfeeding may be difficult in primary care settings,there is evidence that cup and/or spoon feeding inlarger preterm infants may be entirely feasible.504,505

Although the cornerstone of prevention and man-

agement of neonatal hypoglycemia is early co-lostrum and breast milk feeding, other options maybe needed for circumstances in which breast milk isnot available, insufficient, or contaminated withHIV.

community-based evidence. We identified 2studies from India for which alternative strategiesfor prevention of hypoglycemia were used506,507

(Table 30). Both studies, conducted in urban hospitalsettings, indicated that the addition of granulatedsugar to formula feeds in a sterile fashion was satis-factory for prevention and treatment of hypoglyce-mia. The osmolality of the sugar-fortified formula,however, was significantly higher than standardfeeds. Thus, in the absence of follow-up informationon risk of enteric infections and diarrhea with suchan approach, sugar fortification of formula cannot berecommended for widespread implementation inprograms.

conclusions. The use of artificial formulas forti-fied with glucose polymers508 or lipids such as me-dium-chain triglycerides,509,510 has been shown to beeffective in the prevention and management of hy-poglycemia in developed-country studies. It ishighly unlikely, however, that these strategies would

TABLE 29. Hypothermia Prevention and Management



Gosavi et al493 India; urban setting;PCS

32 high-risk newborn infants of varyingdisease severity were transportedusing the Styropor box.

None developed hypothermia (allhad temperatures �36.5°Cbefore and after transport). Noother complications occurredduring the transportation.

Daga et al492 India; rural setting;PCS

1 normal birth weight infant (2.6 kg)and a LBW infant (1.8 kg) weresequentially studied using KMC andthen a Styropor box.

The temperature of the 2 infantswas comparable using the 2methods of care.

Johanson et al473 Nepal; rural setting;RCT

A prospective observational study ofpostnatal neonatal body temperaturewas followed by a randomizedcontrolled intervention study usingKMC, traditional “oil massage” and a“plastic swaddler.” There were 500infants in the initial observationstudy and 300 in the interventionstudy. In the observation study, 85%(420/495) of infants had temperatures�36°C at 2 h and nearly 50% (198/405) had temperatures �36°C at 24 h(14% were �35°C). Most of theinfants who were cold at 24 h hadinitially become cold at the time ofdelivery (only 7 infants had beenboth well-dried and wrapped). In theintervention study, all infants weredried and wrapped before randomassignment to 1 of the 3 interventionmethods.

All 3 methods were found to beequally effective. Overall, 38%(114/298) and 18% (41/231) ofthe infants had a temperature�36°C at 2 and 24 h,respectively. None were�35°C.

Fernandez et al491 India; rural setting;PCS, controlsmatched forweight andgestational age

Preterm babies (n � 25) between 28and 36 wk gestational age weretreated with corn oil applied every4 h to the entire body. An equalnumber of preterm infants (n � 25)were matched for weight andgestational age and served as thecontrol group.

The study group required the useof the warmer to maintainrectal body temperature forsignificantly less time (P �.001) than the controls. Serumtriglyceride levels also rosesignificantly (P � .01) in thestudy group from baseline to72 h after initiation oftreatment.



find favor in developing countries with significantlyhigher risk of infection. Although the data on use ofsugar-fortified milk supplements in India are inter-esting,506,507 these studies did not evaluate the effi-cacy of this intervention in comparison with co-lostrum and breast milk feeding, nor would it havebeen ethical to do so. These alternative strategies tobreast milk feeding cannot be recommended pres-ently but merit careful controlled evaluation withappropriate endpoints and outcomes as possible in-terventions in settings in which breast milk feeding isnot possible or is inadvisable. Thus, consonant withthe recommendations of the WHO,498 the mainstayof prevention and treatment of hypoglycemia in de-veloping countries must clearly remain early andexclusive breastfeeding and the use of expressedbreast milk in other circumstances.

Breastfeedingbackground. Breastfeeding is a foundational

practice for appropriate care and feeding of newborninfants. A wide variety of benefits of breastfeedinghave been well documented, including reduced riskof hypothermia, hypoglycemia, necrotizing enteroco-litis, omphalitis, acute respiratory infections (ARIs),diarrhea, and septicemia.511 Benefits of breastfeedingon infant and child health and development havealso been extensively reviewed recently.60,512 Severalstudies have effectively demonstrated reduced ratesof morbidity and mortality in early infancy withexclusive breastfeeding in community settings in de-veloping countries.513–519 However, there are rela-tively few studies that have principally evaluated theimpact of breastfeeding on neonatal outcomes. Huff-man et al511 recently reviewed the evidence in thisregard and concluded that early and exclusivebreastfeeding played an important role in reducingneonatal mortality, particularly after the first week oflife. There is also additional evidence linking lack of

exclusive breastfeeding with increased risk of early-onset520 and late-onset neonatal sepsis.521

community-based evidence. Our review identi-fied several studies that addressed the issue of im-pact of breastfeeding practices on perinatal and neo-natal outcomes in developing countries. Thesestudies were mostly from developing-country hospi-tal-born cohorts220,505,522–525 (Table 31), although 2other reports principally provided community-baseddata.526,527 These studies consistently showed re-duced rates of morbidity (eg, diarrhea, pneumonia)and mortality associated with breastfeeding. Thelargest systematic review available on the relation-ship of breastfeeding to infectious disease outcomesin developing countries is the WHO multicountrypooled analysis.528 This study clearly indicated thatthose infants �2 months old who were not breastfedhad significantly higher mortality (OR: 5.8; CI: 3.4–9.8) due to infectious diseases than breastfed infants.

Although a number of barriers to effective breast-feeding in the neonatal period have been identi-fied,529–532 including reduced tendency to breastfeedLBW infants,533 there is ample evidence that cultur-ally appropriate behavior-change communicationsstrategies can increase immediate and exclusivebreastfeeding rates59,61,62,531,532,534–536 (Table 32). Inparticular, recent data from India convincingly dem-onstrate the benefit of community strategies for pro-motion of breastfeeding in terms of reducing diar-rhea morbidity and preventing growth faltering.537

conclusions. Despite the lack of RCTs evaluatingthe impact of breastfeeding on neonatal outcomes,there is overwhelming evidence in support of a va-riety of perinatal and neonatal health benefits frombreastfeeding. All newborn infants in developingcountries, especially in domiciliary settings, must re-ceive colostrum and exclusive breastfeeding, andintervention programs must place substantial em-

TABLE 30. Hypoglycemia Prevention and Management



Singhal et al507* India; urban hospitalsetting; RCT

110 term SGA infants without complicationswere enrolled in the study. Infants in theintervention group (n � 55) receivedfortified feeds that were prepared byadding 1 sachet containing 1.5 g ofpowdered sugar to 30 mL of standardmilk formula. Infants in the control group(n � 55) received standard milk formula.

There was an 85% decrease in theincidence of hypoglycemiaamong newborn infants in theintervention group comparedto controls (P � .01). No othermorbidity was reported.

Singhal et al506* India; urban hospitalsetting; RCT

130 term large-for-gestational-age infantswithout complications were enrolled inthe study. Infants in the interventiongroup (n � 65) received fortified feedsthat were prepared by adding 1 sachetcontaining 1.5 g of powdered sugar to30 mL of standard milk formula. Infantsin the control group (n � 65) receivedstandard milk formula.

There was a 77% reduction inrates of hypoglycemia amongintervention group infantscompared to controls. Theincidence of hypoglycemia wasreduced significantly (P � .05)by the sugar-fortified feeds.The mean blood sugar level inbabies receiving fortified feedswas significantly (P � .001)higher at all ages as comparedto those receiving standardfeeds. No other morbidity wasreported.




TABLE 31. Breastfeeding



WHO CollaborativeStudy Team528

Studies from Brazil, The Gambia,Ghana, Pakistan, ThePhilippines and Senegal;pooled analysis of existingstudies

Multiple studies evaluatingexclusive breastfeeding versusmixed or formula feeding andproviding data on outcomes ininfancy were combined in thispooled analysis.

Pooled OR for mortality in nonbreastfedindividuals was 5.8 (CI: 3.4–9.8) forinfants �2 mo of age.

Coutsoudis et al724 South Africa; urban hospitalsetting; PCS

549 women-infant pairs werestudied; 1 cohort of infants wasnever breastfed (n � 156), asecond cohort was exclusivelybreastfed (n � 103), and a thirdcohort received mixed feeding(n � 288). Among 549 infants, 3distinct groups were studied: (1)those who were exclusivelybreastfed, (2) those who weregiven mixed feeds, and (3) thosewho were never breastfed.

Breastfed and nonbreastfed infants hadcomparable rates of HIV infection(acquired from mothers who were HIVpositive). Exclusive breastfeeding to3 mo or longer was associated with asignificantly lower risk of infection(hazard ratio: 0.52; CI: 0.28–0.98,P � .04). Never having breastfed alsocarried a risk of infection similar tomixed feeding (0.85, CI: 0.51–1.42,P � 0.53).

Kumar et al525 India; urban hospital setting; PCS 50 newborns were enrolled in thestudy. group 1 (n � 25) wasexclusively breastfed (EBF);group 2 (n � 25) was mixed fed(MF).

Although birth weights in both groupswere comparable, group 1 infants hadsignificantly greater weight gain fromthe second month of life onwards(P � .01). EBF infants had 40 sicknessepisodes per 100 child-mo as comparedto 69 per 100 child-mo in group 2.Diarrheal episodes were 20 per 100child-mo in group 2 as compared to 6per 100 child-mo in group 1 (P � .01).Cases of pneumonia, otitis media andhospitalization were also more commonamong MF infants than EBF infants (4 vs3; 3 vs 1; and 5 vs 2, respectively),although the sample size was too smallto measure significance.

Augustine et al524 India; hospital setting; RCS Mortality and morbidity statisticsusing the records (n � 169) ofall the infants (60 preterm and109 term) admitted during thefirst 7 d of life in a ruralpediatrics ward over 1 y wereanalyzed.

Exclusive breastfeeding was associatedwith the lowest rate of mortality (29%),compared with infants not yet fed (64%)or those receiving sugar water or cow’smilk with or without breastfeeding(43%).

Singh et al526 India, urban and rural settings;Cross-sectional survey

Data from a cross-sectional surveyon the relationship betweencolostrum and neonatal andpost-neonatal deaths, in which acohort of newborn infants(n � 826) were followed up,was analyzed for mortalitypatterns.

In the rural high socioeconomic group, noneonatal deaths were found among thesubgroup that received colostrum, and8.2% of neonates died in the subgroupthat did not receive colostrum. As ageneral conclusion, 1.7% of urbanneonates and 2.6% of urban postneonateswho received colostrum died, comparedwith 6.3% and 7.1%, respectively, whodid not receive colostrum. In rural areas,comparable neonatal and postneonataldeath rates were 1.7% and 1.7%,respectively, for infants who receivedcolostrum, and 7.4% and 3.5%,respectively, for infants who did notreceive colostrum.

Victora et al523 Brazil; urban hospital setting;CCS

Cases (n � 357) in which younginfants had died of infectionswere identified throughcoroner’s records. 2 controlswere matched to each case forage and socioeconomic status(n � 714).

Compared with infants who wereexclusively breastfed, those who weremixed-fed (given formula or cow’s milkin addition to breast milk) had 4.2 timesthe risk of death from diarrhea (CI:1.7–10.1), whereas those who werecompletely weaned (not receiving anybreast milk) had a risk 14.2 times higher(CI: 5.9–34.1). Compared with exclusivelybreastfed infants, mixed-fed infants andcompletely weaned infants had 1.6 times(CI: 0.7–3.6), and 3.6 times (CI: 1.7–7.5)the risk of death from respiratoryinfections.



phasis on this component. Several authors have dem-onstrated that it is possible to improve breastfeedingpractices through culturally appropriate behavior-change communications. There is a need, however,for enhanced understanding of barriers to early, ex-clusive breastfeeding in many settings, leading toeffective development and implementation of breast-feeding promotion programs.

Prevention and Treatment of Ophthalmia Neonatorumbackground. Ophthalmia neonatorum is caused

primarily by Neisseria gonorrhea or Chlamydia tracho-matis. The prevalence of ophthalmia neonatorum indeveloping countries is uncertain. Before the intro-duction of preventive therapy, this condition wasrecognized to affect between 1% and 15% of new-born infants in Europe and 9% of newborn infantsin the United States.538 Transmission rates frommothers with gonococcal cervicitis to their offspringrange from 30% to 50% in the absence of prophylacticocular treatment,14 whereas transmission rangesfrom 18% to 61% for chlamydia.539 In developingcountries such as Kenya, where rates of maternal

STDs are high, rates of gonococcal ophthalmia rangefrom 15% to 34%.540,541

A variety of interventions have been evaluated forprevention of ophthalmia neonatorum in developedcountries, mostly in hospital settings, and havelargely focused on the prevention of gonococcal in-fections.542–545 These studies have suggested that sil-ver-nitrate, tetracycline, or erythromycin ointmentsgiven prophylactically are equivalent in effica-cy.546,547 The WHO recommends 1% silver-nitratesolution, 1% tetracycline ointment, or 2.5% povi-done-iodine within 1 hour of delivery.14,16,548 TheAmerican Academy of Pediatrics also considers0.5% erythromycin ointment as standard therapy, inaddition to silver nitrate or tetracycline.549 In thepresence of proper prophylactic treatment, gonococ-cal ophthalmia develops infrequently. However, sil-ver nitrate is not effective if used after infection hasbeen established, which may help to explain thehigher risk of gonococcal eye infections observed inassociation with PROM. For infants born to motherswith confirmed gonorrhea, parenteral treatment with

TABLE 31. Continued



Habicht et al527 Malaysia; urban and ruralsettings; RCS

Sample consisted of 5357 birthsbetween 1940 and 1975, withan average IMR of 47.8 infantdeaths per 1000 live births.Sample was restricted tobabies who could have liveduntil the end of the at-riskperiod. 83% of infants werefully breastfed.

The effects of breastfeeding on infantmortality were reported for 3 sub-periods of infancy: 8–28 d, 29 d-6 mo,and 7–12 mo. Reductions in deaths per1000 infants per added month of fullbreastfeeding were 68.6 (CI: 41.1–96.1),24.9 (CI: 13.4–36.4), and 3.4 (CI: 1.6–5.2)for the periods 8–28 d, 29 d-6 mo, and7–12 mo, respectively. Reductions indeaths per 1000 infants per addedmonth of partial breastfeeding were21.9 (CI: –7.9–51.7), 11.2 (CI: 0.2–22.2),and 1.7 (CI: 0.4–3.0) for the 3 timeperiods, respectively.

Narayanan et al505 India; urban hospital setting;RCT

A total of 226 LBW infants wererandomly assigned to 1 of 4intervention groups. Group 1(n � 57) was fed rawexpressed human milk only;group 2 (n � 56) was fedhuman milk subjected toholder pasteurization. Group 3(n � 56) was fed raw milkfrom 9 am until 9 pm (5 feeds)and formula for 3 feeds atnight. Group 4 (n � 57) wasfed pasteurized human milkfrom 9 am until 9 pm andformula for the remaining 3feeds.

There was a 68% reduction in neonatalinfection rates in group 1 (10.5%)compared to group 4 (33%); however,sample size was too small for statisticalsignificance.

Clavano522 Philippines; urban hospitalsetting; PCS

9886 newborn infants werefollowed up to assess therelationship of neonatalmortality and diarrhea tofeeding patterns.

Of the 138 infants with diarrhea, 90%were formula fed, 6% were partiallybreastfed, and 4% were exclusivelybreastfed. Mode of infant feeding wasalso significantly related to mortality.Of the 67 infants who died, 96% wereformula fed, 1% was partiallybreastfed, and 3% were exclusivelybreastfed. After rooming-in and formalbreastfeeding policies were introduced,the proportion of infants exclusivelybreastfeeding increased by 135%, andthe incidence of death among clinicallyinfected newborns dropped by 95.3%.



TABLE 32. Breastfeeding Promotion and Education



Jakobsenet al62

West Africa;suburbansetting; RCT

1226 pregnant women who gave birthto 1250 children were randomized to1 of 2 groups. Mothers in theintervention group (n � 659) weregiven individual health educationsessions at the local health center ontheir routine visits (ie, first antenatalvisit, vaccination sessions in weeks 6,10 and 14, and when the childreached the age of 9 mo). Mothers inthe control group (n � 591) were notexposed to these messages.Intervention messages encouragedbreastfeeding for 2 y, avoidance ofintroducing weaning foods andwater for 4–5 mo, and continuedbreastfeeding for ill children. Family-planning information was alsoincluded.

Significantly more mothers inthe intervention group hadhad an intrauterine deviceinserted (rate ratio � 2.45(CI: 1.27–4.70) comparedwith the control group.However, mothers who hadan intrauterine deviceinserted did not breastfeedlonger compared withmothers who reported notto have had an intrauterinedevice inserted.

Introduction of weaning foods wassignificantly delayed in theintervention group comparedwith the control group (incidencerate 1.22; CI: 1.05–1.42). Morechildren in the intervention group(31.1%) compared to controls(24.8%) were almost exclusivelybreastfed at 4 mo of age (OR:1.18; CI: 1.03–1.38).

Morrowet al61

Mexico;periurbansetting; RCT

130 pregnant women were randomizedto 1 of 3 groups. Group 1 (n � 44)received 6 visits by a peer counselor.Group 2 (n � 52) received 3 visits bya peer counselor. Group 3 (n � 34)was the control group; women inthis group were referred to aphysician for lactation problems.

There were significantly morefemales who were ex-clusively breastfeeding theirinfants in the interventiongroups than the controlgroup. Duration of anybreastfeeding (�3 mo or�6 mo) was significantlygreater in the interventiongroup than the controlgroups (P � .024).

Control infants had a significantlyhigher risk of diarrhea than thosein the intervention groups (OR:2.1; 90% CI: 1.11–4.04; P � .029).

Lutteret al59

Brazil; urbanhospitalsetting; PCS

442 women were enrolled in the study;236 women delivered at the programhospital and 206 delivered at thecontrol hospital. The interventionhospital had an active breastfeedingpromotion program, characterized byrooming-in, early initiation ofbreastfeeding and breastfeedingassistance and talks duringhospitalization. The control hospitalhad no such programs.

The median duration of exclusivebreastfeeding was 75 d amongwomen in the program hospitalcompared with 22 d in the controlhospital. At 1 mo, the probabilityof exclusive breastfeeding was0.64 in the program groupcompared with 0.39 in the controlgroup.

Haideret al531*

Bangladesh;urban setting;cross-sectionalknowledge,attitudes, andpracticessurvey

Study area covered 30 000 households;a random sample of households withinfants ages 0–6 mo (n � 1100) wereselected. Mothers were interviewedat home. The WHO classification wasused to describe infant feedingpractices.

93% of mothers fed their infantscolostrum during the first 3 d oflife. However, only 8% fedcolostrum as the first food. 97%claimed they had heard the term“exclusive breastfeeding” andunderstood it but 70% wronglyperceived it to mean giving thebaby breast milk and water. 33%and 5% of infants in their firstand fifth mo of life, respectively,were breastfed exclusively.

Haideret al532*

Bangladesh;urban setting;community-based RCT

40 adjacent zones were randomized tocontrol or intervention groups.Women were enrolled during the lasttrimester of pregnancy. 15 house-based counseling visits werescheduled, with 2 visits in the lasttrimester, 3 during the earlypostnatal period (1 within 48 h, 1 onday 5, and 1 between days 10 and14), and every 2 wk until the infantwas 5 mo old. Peer counselors werelocal mothers who received 10 d oftraining.

363 women were enrolled in eachgroup. Peer counselingsignificantly improvedbreastfeeding practices. For theprimary outcome, the prevalenceof exclusive breastfeeding at 5 mowas 202/228 (70%) for theintervention group and 17/285(6%) for the control group (meandifference: 64%; CI: 57–71%; P �.0001). For the secondaryoutcomes, mothers in theintervention group initiatedbreastfeeding earlier than controlmothers and were less likely togive prelacteal and postlactealfoods.




ceftriaxone is recommended as first-line therapy bythe WHO548 and the American Academy of Pediat-rics,549 although penicillin G can be considered ifresistant strains are unlikely. Silver nitrate has beenshown to be equivalent to tetracycline or erythromy-cin ointments for prevention of chlamydia conjunc-tivitis.539,545 Thus, systemic treatment with erythro-mycin is also needed to prevent progression tochlamydial pneumonia.

community-based evidence. Our search identi-fied 2 studies from developing countries (see Table33) on prevention and treatment of ophthalmia neo-natorum. Both were undertaken in urban hospitalsettings.541,550 In 1 study, povidone-iodine was supe-rior to silver-nitrate drops and to costlier erythromy-cin ointment.550 In another study, tetracycline oint-ment was superior to silver-nitrate drops.541

conclusions. Evidence for the benefit of prophy-laxis with silver-nitrate drops, tetracycline ointment,or erythromycin ointment to prevent ophthalmianeonatorum is well established, based largely ondeveloped-country studies. Scarce data from devel-oping countries suggest that povidone-iodine is alsoeffective, and this agent is recognized by the Amer-ican Academy of Pediatrics as a first-line prophylac-tic therapy. Thus, choice of agent may be less im-portant than wide-scale implementation of thisintervention, which must form part of primary neo-natal management packages, particularly in areasendemic for sexually transmitted diseases (STDs)due to gonorrhea. In areas with low rates of gono-coccal infections in pregnancy, the benefits of routineprophylaxis relative to other newborn interventionsmay be questionable. Although local gonococcal re-sistance rates should be considered, local availability,

acceptability, and cost of the various agents gener-ally will be primary determinants in choice of agent.Overall, the evidence in support of prophylaxis forchlamydial infections is less clear.

Vitamin K Prophylaxisbackground. Vitamin K prophylaxis in neonates

for prevention of IVH is well established. TwoCochrane reviews have addressed the role of vitaminK prophylaxis in neonatal health. Puckett and Of-fringa,551 in a Cochrane review of prophylactic vita-min K for prevention of IVH, also concluded that asingle dose (1.0 mg) of intramuscular (IM) vitamin Kafter birth is effective in the prevention of classichemorrhagic disease of the newborn (HDN), and IMor oral (1.0 mg) vitamin K prophylaxis improvesbiochemical indices of coagulation status at 1 to 7days. However, neither IM nor oral vitamin K hasbeen tested in randomized trials with respect to theeffect on late HDN, and oral vitamin K, in eithersingle or multiple doses, has not been tested in ran-domized trials for its effect on either classic or lateHDN. Crowther and Henderson-Smart,552 in aCochrane review of vitamin K administration topregnant women before preterm delivery, concludedthat antenatal vitamin K was associated with a non-significant trend toward reducing all grades ofperiventricular hemorrhage (RR: 0.82; 95% CI: 0.67–1.00), particularly severe (grades 3–4) disease (RR:0.75; CI: 0.45–1.25). This trend disappeared whenpoorer-quality trials were excluded. It was con-cluded that vitamin K administered to women beforevery preterm birth has not been shown to signifi-cantly prevent periventricular hemorrhage.

Although well established in developed countries,the incidence and implications of vitamin K defi-

TABLE 33. Prevention and Treatment of Ophthalmia Neonatorum



Isenberg et al550 Kenya; urbansetting; RCT

3117 newborns were enrolled andrandomized to 1 of 3 groups toreceive 2.5% povidone-iodinesolution (n � 1076), 0.5%erythromycin ophthalmicointment, (n � 1112), or a dropof 1% silver-nitrate ophthalmicsolution (n � 929).

13% of the infants in the povidone-iodinegroup, vs 17.5% of the infants in thesilver-nitrate group (P � .001), and15.2% of the infants in the erythro-mycin group (P � .01) had infectiousconjunctivitis. Povidone-iodine wasmore effective against Chlamydiatrachomatis than was silver nitrate(P � .001) or erythromycin (P � .008).Noninfectious conjunctivitis was lessfrequent among povidone-iodine-treated infants (9.7%) than thosetreated with silver-nitrate (13.9%,P � .001) or erythromycin group(13.3%, P � .004). Incidence of Ngonorrhoeae and Staphylococcus aureusinfections was similar in the 3 groups.

Laga et al541 Kenya; urbansetting; QT

2732 infants were enrolled in 1 of2 groups to have either 1%silver nitrate (n � 1233) or 1%tetracycline (n � 1499) instilledin their eyes. A historicalcohort of newborns who didnot receive eye prophylaxiswere included as controls.

Infection rates of gonococcal ophthalmianeonatorum in newborns exposed to Ngonorrhoeae at birth were 7.0% in thosereceiving silver nitrate and 3.0% inthose receiving tetracycline(CI: �3.4 to 11.4%). As compared withhistorical controls, the incidence ofgonococcal ophthalmia neonatorumdecreased 83% and 93% among infantstreated with silver nitrate and thosetreated with tetracycline, respectively.



ciency in developing countries has received littleattention. In a national hospital-based survey inThailand, the incidence of vitamin K deficiency wasfound to be 35 cases per 100 000 births.553 In contrast,a community hospital–based study in rural Thailandfound the incidence to be as high as 72 cases per100 000 births.554 There are very few community-based data on vitamin K deficiency or HDN fromeither Africa or south Asia.

community-based evidence. Our review couldnot identify any community-based vitamin K inter-vention trials.

conclusions. Vitamin K prophylaxis is a well-established intervention for prevention of IVH andearly HDN in newborns in developed countries. Thepreferred route is IM administration, whereas theefficacy of the oral route needs additional investiga-tion. There is no definitive evidence that administer-ing prenatal vitamin K to pregnant women enteringpreterm labor can prevent IVH in the neonate, butthis intervention warrants additional investigation.Moreover, studies are needed in developing-countrycommunity settings to identify cost-effective meth-ods of implementing vitamin K interventions.555

Hepatitis B Vaccinationbackground. Disease due to hepatitis B virus

(HBV) infection is one of the most prevalent publichealth problems worldwide. An estimated 350 mil-lion people are HBV carriers, and �1 million deathseach year are attributed to the virus.556,557 The hep-atitis B surface antigen (HBsAg) carrier rate in thegeneral population in Taiwan was reported to be15% to 20%,558 whereas the prevalence of HBsAgpositivity in Bangladesh was highest in 5- to 9-year-olds (8.5%), and simultaneous infection with hepati-tis D virus was common (24.4%), particularly in olderage groups.559 Dusheiko et al560 reported a hepatitisB carrier rate of 5.5% to 14% in the southern Africanpopulation.

Hepatitis B immunization prevents chronic HBVcarrier states and, subsequently, chronic liver diseasein adulthood. However, Wong et al561 found that theincidence of preterm birth, PPROM, SGA births, neo-natal jaundice, fetal distress, perinatal asphyxia, con-genital abnormalities, gastrointestinal tract abnor-malities, and perinatal mortality were similar amongpregnant women with and without serologic HBsAgpositivity. Thus, the benefits of hepatitis B vaccina-tion are long-term and most evident well after theneonatal period.

community-based evidence. Seven studies wereidentified that discussed programs implementinghepatitis B vaccination in developing-country com-munities; the 2 main studies were conducted in Tai-wan and The Gambia558,562–567 (Table 34). In general,incorporation of hepatitis B immunization into theEPI program of various countries was highly suc-cessful and effective. Hepatitis B vaccination pro-grams were not only effective in preventing perinataland early horizontal transmission of HBV but alsoresulted in decreasing mortality due to fulminanthepatitis and reduced development of childhoodhepatocellular carcinoma. One study in South Afri-

ca,568 however, determined that the intervention wasnot effective. Some reasons for the lack of effective-ness were given, and the authors proposed that theproblems they identified were likely to be faced byother programs in developing nations. Importantly,they found that program participants had difficultyin complying with WHO recommendations that thefirst dose of hepatitis B vaccine be given as soon aspossible after birth. Administering all 3 doses withinreasonable limits of the recommended vaccineschedule is often difficult in rural areas, where dis-tances to clinics are great and mobile outreach clinicsare impeded by poor road conditions. Moreover, thebirth dose may be particularly difficult given variousproscriptions against seeking care for newborns out-side the home in the formal health sector, particu-larly for early postnatal care.439,569–571 Providing ap-propriate training to village health workers (VHWs)in The Gambia to deliver heat-stable vaccines in aUniject device, however, was shown to be a cost-effective and reliable method for hepatitis B vaccina-tion in rural settings.572 This alternative means ofdelivering the intervention may facilitate resolutionof the above-mentioned problems with traditionalhepatitis B vaccination and may be an effective wayto improve coverage and seroconversion rates.

conclusions. Hepatitis B vaccine is administereduniversally at birth in developed countries. Consid-ering the high risk of hepatitis B infection in manydeveloping countries and the efficacy of the hepatitisB vaccine, policies regarding routine immunizationneed to be developed and implemented urgently indeveloping countries endemic for HBV infection. It isimportant to recognize, however, that evaluating thesuccess of integrating hepatitis B vaccination into theEPI is more complex than for many other diseases.Because many infected children do not have anyrecognizable signs of illness, serologic surveys,rather than disease surveillance, will be needed tomonitor the success of immunization programs. Theappropriate field methodology for such surveysmust be developed and standardized. The true ben-efit of such programs also requires measurement ofthe reduction in long-term sequelae of infection.

The total financial burden derived from both thedirect and indirect costs of HBV detection andvaccination must be weighed critically in light ofmultiple health problems and competing agendas inresource-poor countries. However, support from do-nor agencies such as the Bill & Melinda Gates Foun-dation and the Children’s Vaccine Program, whichhas initiated aid for hepatitis B vaccination in a num-ber of countries, could help to make immunizationagainst HBV, a proven intervention for benefit laterin life, a reality.

Neonatal Vitamin A Supplementationbackground. The significant public health bene-

fits of vitamin A supplementation on child mortalityin developing countries210 are well established.There is much interest in the potential benefit ofneonatal vitamin A supplementation on neonataland infant outcomes, given the widespread subclin-ical vitamin A deficiency that exists among pregnant



TABLE 34. Neonatal Hepatitis B Vaccination


Intervention Program Impact

Chen et al562;Kao et al567;Chen et al563

Taiwan; community-based QT

In July 1984, a nationwide vaccination programwas started. Preparatory work was done 1 ybefore the program started. It includededucation of doctors, allied health workers andpublic laboratory facilities in testing HBsAgand hepatitis B virus “e” antigen (HBeAg). Aconfidential registration data bank was formed.Infants born to HBsAg-positive mothers weregiven the vaccine at ages 1, 5, and 9 wk andthen a booster at 12 mo. Infants born to mothertesting positive for both antigens received ahepatitis B immunoglobulin within 24 hoursalong with the vaccines.

87% of the children received at least 3 dosesof vaccine. The overall prevalence rate ofHBsAg decreased from 9.8% in 1984 to1.3% in 1994 in children �10 y of age. Astatistically significant decrease wasobserved in every age group between 1 and10 y. The overall prevalence rate of HBeAgwas 26% in 1984, 15% in 1989 and 4% in1994. The average annual incidence ofhepatocellular carcinoma in Taiwanesechildren aged 6–14 y declined from 0.7 per100 000 children in 1981–1986 to 0.36 per100 000 in 1990–1994. The average mortalityassociated with fulminant hepatitis ininfants from 1975–1984 was 5.36/100 000infants (range: 2.9–6.7), which decreased to1.71/100 000 (range: 0.3–4.6) for the period1985–1998. Thus, the mortality ratedecreased threefold after the vaccinationprogram.

The GambiaHepatitisStudyGroup566;The GambiaHepatitisStudyGroup565;Montesano564

The Gambia;community-basedQT

Children (n � 60 000) were divided into 2groups. 1 group received routine EPIvaccination, and the other received routine EPIvaccination plus the hepatitis B vaccine. Thiscontinued from 1986–1990, after which allchildren were given hepatitis B vaccine and itbecame a part of the standard EPI. The firstdose was given at the same time as BCGvaccine (ie, during the first month of life). Thesecond dose was given along with a first doseof triple antigen at 2 mo of age or later. Thethird dose was given with the third dose oftriple antigen at 4 mo of age or later. Aminimum interval between doses was fixed at4 wk. A cohort of 1000 children was followedlong-term to measure immunogenicity.

The coverage survey showed that 98% ofchildren aged 12–18 mo had received thefirst dose, 94% had received the seconddose and 74% had received the third. In94% of children, immunization producedprotective levels of antibodies at 1 y of age.As a result of the vaccination program,protection against infection was �80% andthe protection against carriage status was�90%.

Odusanyaet al725

Nigeria; community-based QT

Immunization and preprimary health careservices were commenced for children (n �327) in a rural community in Nigeria. HepatitisB vaccine was administered at birth and thenin a single injection with diphtheria-tetanus-pertussis vaccine at 6 wk of age and another at3 mo of age.

2 y after the program was started, theimmunization coverage rates were 94% forBCG, 88% for diphtheria-tetanus-pertussisvaccine (third dose) and 82% for measles.Hepatitis B coverage for all 3 doses was58%.

Ariwan726 Indonesia; community-based QT

Healthy Start Program in Indonesia involved avisit by the village midwife between 1 and 7 dafter birth to provide Hepatitis B vaccinationand maternal care/education.

Hepatitis B vaccination rate increased from0% in 1990 to 84% in 1996. IMR decreasedfrom 73/1000 to 55/1000 live births, due toother aspects of the maternal care andeducation program.

Poovorwanet al727

Thailand; community-based QT

Hepatitis B vaccination has been an integral partof the EPI since 1992. In each of 5representative provinces, 400–488 healthyimmunocompetent infants ranging from 6 to18 mo of age were evaluated by examiningtheir sera for viral hepatitis markers.

The coverage rate of hepatitis B vaccine was71.2% to 94.3%. The number of individualsundergoing complete course of hepatitis Bvaccination increased fourfold. Only 0.7%of children born after the implementationof the EPI strategy were HBV carriers.

Sutantoet al572

Indonesia; community-based QT

During the study, village midwives (n � 110)used Uniject devices, rather than the standarddisposable or reusable syringes. The midwiveswere provided with these devices in anoutreach carrier box. After they picked upthese boxes the midwives were allowed tokeep them under ambient conditions (27°Caverage, range of 25�32°C) for 1 mo. Nursesadministered 10 000 sterile injections in home-based settings during the trial.

Potency test of the Uniject compared to thestandard injection after 1 month showedthat the TT vaccine had lost 6% of itspotency, while the hepatitis B vaccine hadsuffered a 1% drop in potency.Seroconversion rates were identical inchildren immunized using the Uniject vsthe standard vaccine method. Total cost perchild immunized using the Uniject devicewas US $6.57 vs US $7.19 using a standarddisposable syringe at the health center.

Schoubet al568

South Africa; ruralsetting; community-based QT

All infants born in 1989 in a self-governingregion of the Transvaal were immunizedaccording to 2 schedules: (1) early schedule �birth, 3 mo and 6 mo and (2) late schedule � 3mo, 4.5 mo and 6 mo. Those who were notgiven vaccine according to any 1 of theseschedules were classified as unscheduled.

Only 6.6% of vaccine recipients werevaccinated according to the schedules,whereas 93.4% were given vaccines in anunscheduled manner. There was nodifference in seroconversion to the surfaceantigen between the 2 groups.



women and lactating mothers in many developing-country settings, the demonstrated positive impactof antenatal vitamin A supplementation on maternalmortality218 (see “Antenatal Vitamin A Supplemen-tation”), and the demonstrated benefits of vitamin Asupplementation during infancy and childhood.94

Vitamin A supplementation has been investigatedin selected subsets of newborns in developed coun-tries, such as in VLBW infants, to prevent chroniclung disease.573–575 However, the benefit and cost-effectiveness of such interventions in developing-country settings, in which survival rates of VLBWinfants are low, remain unclear; few VLBW infantssurvive to develop chronic lung disease.419

community-based evidence. We identified 3studies that evaluated the impact of vitamin A sup-plementation during the neonatal period on neonataland/or infant outcomes in developing countries inboth rural and urban settings (Table 35). No studyshowed a reduction in mortality during the neonatalperiod,576,577 and the 3 studies showed a mixed im-pact on infant mortality.576–578 Results are forthcom-ing from an additional large-scale trial of neonatalvitamin A supplementation in Zimbabwe.

conclusions. Neonatal supplementation with vi-tamin A holds some promise for reducing mortalityduring or shortly after the neonatal period, and thereis some evidence of reduced mortality during infan-cy.576 These results, coupled with emerging evidencethat antenatal vitamin A supplementation may re-duce maternal mortality,218 indicate the need for ad-ditional research in this area. A critical factor may bethe timing of vitamin A supplementation, with po-tentially increased impact with early administra-tion.577 The role of supplementation in populationswith relatively high rates of HIV infection also re-quires investigation.

KMCbackground. KMC refers to the technique of po-

sitioning and skin-to-skin care of LBW infants afterbirth and was pioneered by Rey579 from Colombia.After initial skepticism,580 the technique has nowfound widespread acceptance and has been usedextensively in both developed and developing coun-tries. Despite numerous intervention studies evalu-ating the technique, however, very few have usedrigorous (eg, RCT) designs. A recent Cochrane re-view of KMC581 identified 3 studies that met criteriafor inclusion in their analysis, but the analysis wasbased largely on just 1 RCT.582 The review found thatneonates who were given KMC versus standard carehad lower risk of nosocomial infection (OR: 0.49; CI:0.25–0.93), severe illness (OR: 0.3; CI: 0.14–0.67), andlower respiratory tract disease (OR: 0.37; CI: 0.15–0.89) at 6-month follow-up. In addition, the propor-tion of infants who were not exclusively breastfeed-ing at discharge was reduced (OR: 0.41; CI: 0.25–0.68), as was the rate of maternal dissatisfaction withmethod of care (OR: 0.41; CI: 0.22–0.75). Infantsgiven KMC also had gained more weight per day bydischarge (weighted mean difference: 3.6 g per day;CI: 0.8–6.4). Despite these interesting data, becausethe results were largely based on a solitary RCT, itwas concluded that there were insufficient data torecommend its routine use in LBW infants.

Since the Cochrane review noted above, Charpaket al583 reported another RCT of KMC impact in ahospital setting. The intervention consisted of con-tinuous skin-to-skin contact and nearly exclusivebreastfeeding, and infants were followed until 12months’ corrected age. Risk of death tended to belower in the infants given KMC (OR: 0.57; CI: 0.17–1.18) but was not significantly lower than the controlinfants given routine incubator care. Infection rates

TABLE 35. Neonatal Vitamin A Supplementation



Humphrey etal576

Indonesia, urbanhospital setting;RPCT

2067 newborn infants were allocatedto receive 1 oral dose of 52 �molvitamin A plus 23 �mol vitamin E(n � 1034) or placebo (�0.1 �molvitamin A plus 23 �mol vitaminE) (n � 1033), on the first day oflife.

Supplementation had no effect on NMR.The probability of survival during thefirst year of life, however, was greaterin the vitamin A-supplemented thanthe control infants (OR: 0.36; CI: 0.16–0.87).

West et al578 Nepal, ruralsetting;DBRPCT

Newborns (n � 11 918) wererandomized to either an inter-vention group (n � 6086) or aplacebo group (n � 5832).Newborns in the interventiongroup received a single oral doseof vitamin A (15 000 retinolequivalent [RE] [�3 drops of oil]);newborns in the placebo groupreceived 75 RE (250 IU).

The cumulative OR of death after 4 moof the supplementation trial suggesteda modest protective effect for vitaminA (OR: 0.9) but this pattern thenchanged in the second year and theOR returned to 1.11 (CI: 0.86–1.42)after 24 mo; no impact overall onneonatal mortality was observed.

Rahmathullahet al577

India; ruralsetting; DBRCT

Infants (n � 11 619) were randomlyassigned to receive either 24 000IU of vitamin A twice within a24-h interval beginning within48 h of birth, or to receive aplacebo.

Infants in the vitamin A group had a22% reduction in total mortality (CI:4–37%) compared with those in theplacebo group. Vitamin A had animpact on mortality between 2 wk and3 mo after treatment, with noadditional impact after 3 mo.



were similar, although severity was judged to behigher in the control infants, and there was no dif-ference in disability rates. Total days in the hospitalwere decreased in infants who weighed �1500 g atbirth and received KMC.

An earlier 2-cohort study by Charpak et al584 re-ported a higher RR of mortality in infants givenKMC (RR: 1.9; CI: 0.6–5.8), although this risk re-versed after adjusting for weight at birth and gesta-tional age (RR: 0.5; CI: 0.2–1.2). Infants given KMCgrew less in the first 3 months and had a higherproportion of developmental delay at 1 year. Nosignificant differences were found in overall mortal-ity rates between the 2 groups. The cohorts in thestudy, recruited from 2 different tertiary hospitals,showed many social and economic differences.

A nonrandomized, controlled study conducted ina remote Zimbabwe mission hospital without incu-bator care reported that survival of infants bornweighing �1500 g improved from 10% to 50%,whereas that of infants 1500 to 1999 g improved from70% to 90%.585 Similar results were shown by an-other nonrandomized study from a secondary hos-pital in Mozambique. Of 32 infants weighing �1800g, survival was 73% in 22 infants given KMC and20% in 10 infants not given KMC (P � .01). It also hasbeen suggested that infants given KMC dischargedduring the cold season may be more vulnerable tosevere illness, especially lower respiratory tract in-fections, than those discharged during the warm sea-son.586 Another study on a small number of newborninfants with mild respiratory distress suggested thatearly introduction of KMC might have beneficialeffects.587

Ramanathan et al588 reported from their RCT inDelhi, India, that KMC neonates demonstrated betterweight gain after the first week of life (15.9 � 4.5 vs10.6 � 4.5 g/day in KMC and control groups, respec-tively; P � .05) and earlier hospital discharge (27.2 �7 vs 34.6 � 7 days in KMC and control groups,respectively; P � .05). The number of mothers exclu-sively breastfeeding at 6-week follow-up in the KMCgroup was double that of the control group (12 of 14[86%] vs 6 of 14 [43%]; P � .05).

community-based evidence. There are no dataon the impact of KMC from community-based set-tings in developing countries. All studies reportedfrom developing countries were conducted amongmedically stable LBW infants (typically �2000 g) inurban hospital settings (Table 36).

Two reports describe adaptation of the techniquefor use in the community.589 In a pilot study inBangladesh, CHWs were trained in community-based application of KMC (CKMC) and subse-quently taught expectant and new mothers how togive KMC. Women were interviewed 1 month post-partum to evaluate their experience with CKMC. Inall, 77% of mothers initiated skin-to-skin care, and85% with LBW infants did so (37% were LBW).CKMC was adopted quickly and popularly in thiscommunity. Similarly, in Uttar Pradesh, India, pro-motion of CKMC through a community mobilizationand behavior change communication program re-

sulted in adoption of skin-to-skin care of newbornsby �70% of families.590

conclusions. Studies to date have suggested thatKMC had a variety of beneficial effects on the healthof LBW infants, including increased weight gain andexclusive breastfeeding rates and lower risk of nos-ocomial infection and severe illness. The literaturealso suggests that KMC promotes mother-infantbonding, improves newborn thermal control, short-ens duration of hospital stay, results in earlier stabi-lization of physiologic and behavioral functioning,and may improve survival. This last claim is un-proven and may be unrealistic, given that studieshave all included medically stable infants.

Limited experience with adaptation of skin-to-skincare for use in the community in developing coun-tries has been encouraging. There is an urgent need,however, to adapt the approach for various culturalsettings and to further evaluate its acceptability,safety, and efficacy in developing-country commu-nity settings.26,591 We feel that KMC must be viewedin its wider application to include even culturallyappropriate variants such as co-bedding with themother and skin-to-skin practice by caregivers otherthan the mother herself, particularly the father, be-cause fathers, too, can provide effective thermal con-trol in newborn infants to reduce the risk of hypo-thermia.592 Given the paucity of resources and risk ofhypothermia in domiciliary settings in developingcountries, this intervention (or mode of newborncare) may be suitable for essential care of all new-borns in the community as well as for high-risksituations such as the care of preterm, VLBW infants,and transport of sick newborns.

Topical Emollient Therapybackground. The skin barrier of preterm infants

is compromised for a variety of reasons includingdevelopmental immaturity455 and lack of vernix,which is produced near term during gestation andserves as a naturally protective cutaneous biofilm.593

In addition, the skin barrier of preterm infants iseasily injured,594 and particularly in developing-country situations, the skin barrier of even term in-fants may be compromised as a result of intrauterinemalnutrition.595 Thus, the skin may serve as an im-portant portal of entry for serious bacterial infec-tions.596 However, evidence from laboratory studiesin animals and clinical trials in humans suggests thatit may be possible to enhance skin-barrier functionthrough application of topical emollients or oils,thereby improving health outcomes.

In experimental conditions, compromised epider-mal barrier function and dermatitis in essential fattyacid (EFA)-deficient rodents was reversed by topicalapplication of either purified EFAs or EFA-rich veg-etable oils, particularly sunflower-seed oil.597,598 Cu-taneous absorption and metabolism of linoleic acidseemed largely responsible for the beneficial effectsof sunflower-seed oil.490,598–600 Topical therapy withsunflower-seed oil has also been found to correctcutaneous signs of EFA deficiency (ie, normalizationof TEWL, resolution of dermatitis) in humans, in-cluding premature infants.598,601,602 It has been ad-



vanced that these beneficial effects may be due toincreased metabolism of lipids in the epidermis, in-cluding active fatty-acid transport by keratinocytes,which makes it possible for even the immature epi-

dermis of preterm infants to metabolize lipids de-rived from topically applied emollients and to utilizethem as nutritional building blocks for the formationof a healthy, functional epidermal barrier.455,490,603,604

TABLE 36. KMC



Cattaneoet al731

Ethiopia,Indonesia, andMexico;multicenterurban hospitalsetting; RCT

149 LBW neonates wererandomly assigned to KMC(almost exclusive skin-to-skincare after stabilization), and136 LBW neonates to con-ventional methods of care(warm room or incubatorcare).

Overall scores on mother’s senseof competence were better inthe KMC than in the controlgroup (weighted meandifference: 0.31; CI: 0.13–0.50)On the other hand, overallscores on mother’s perceptionof social support during theinfant’s stay in the NICU wereworse in the KMC group thanin the control group (weightedmean difference �0.18; CI:�0.35 to �0.01).

Lower rates of not exclusivelybreastfeeding were recorded atdischarge (RR 0.41; CI: 0.25–0.68).

Charpaket al582*

Colombia; urbanhospitalsetting; RCT

In an urban tertiary hospital,746 newborns were ran-domized to receive eitherKMC (n � 382) or traditionalcare (n � 364).

Significantly lower rates of nosocomialinfection (OR: 0.49; CI: 0.25–0.93) andsevere illness at 6 mo follow-up werefound in the KMC group. The risk ofdying was similar in both groups(RR � 0.59; CI: .22–1.6), and nodifferences were found in growthindices. Hospital stay after eligibilitywas shorter in the KMC group,primarily for infants �1800 g.

Sloanet al732

Ecuador; urbanhospitalsetting; RCT

300 newborns were randomizedto KMC (n � 140) orincubator/thermal cot care(n � 160).

Trial was stopped early because of asignificantly lower rate of severemorbidity in the KMC group(P � .005 at 6 mo).

Charpaket al583*

Colombia; urbanhospitalsetting; RCT

A group of 746 newborns wererandomized when eligible forminimum care, into KMC (n� 382) and “traditional” care(n � 364). Information onvital status was available for93% of infants at 12 mo ofcorrected age. KMC consistedof skin-to-skin contact onmother’s chest 24 h/d, nearlyexclusive breastfeeding, andearly discharge, with closeambulatory monitoring.Controls remained inincubators until the usualdischarge criteria were met.

There was a “trend” toward lower riskof death among the KMC group,although the result was notsignificant (KMC: 11 [3.1%] of 339infants died; control: 19 [5.5%] of 324infants died; RR: 0.57; CI: 0.17–1.1).Growth index of head circumferencewas significantly greater in the KMCgroup, but the developmental indicesof both groups were similar. Infantsweighing �1500 g at birth and givenKMC spent less time in the hospitalthen the group given standard care.There were similar numbers ofinfections in both groups, butinfections in the KMC group were oflesser severity.

Ramanathanet al588

India; urbanhospitalsetting;RCT

28 neonates with birth weights�1500 g were randomizedinto 2 groups, 1 receivingKMC and 1 receivingincubator care. The KMCgroup (n � 14) received KMCfor 4 h per day in not morethan 3 sittings. Infantsreceived KMC after shiftingfrom NICU and at home.Control infants receivedstandard care. A Likert scalewas used to assess mothers’/nurses’ attitudes towardsKMC.

The number of mothers exclu-sively breastfeeding at 6-wkfollow-up was double in theKMC as compared to thecontrol group (12/14 [86%] vs6/14 [43%], respectively;P � .05).

KMC neonates demonstrated betterweight gain after the first week of life(15.9 � 4.5 g/d vs 10.6 � 4.5 g/d inthe KMC and control groups,respectively; P � .05) and earlierhospital discharge (27.2 � 7 vs34.6 � 7 d in KMC and controlgroups, respectively, P � .05).

Kambaramiet al730

Zimbabwe;urban hospitalsetting; RCT

74 infants (37 in each of 2groups) were consecutivelyallocated to receive eitherKMC or incubator care.

KMC infants gained twice as muchweight per day (20.8 vs 10.2 g;P � .0001); had shorter stays in thehospital (16.6 vs 20.7 d; P � .0457);and had a better survival rate (0% vs9% deaths; sample size too small forsignificance).



Darmstadt et al490 used a hairless-mouse model todemonstrate that the impact of topical therapy variesmarkedly with the product applied. A single appli-cation of sunflower-seed oil significantly improvedskin-barrier function within 1 hour, and the effectwas sustained 5 hours after application. In contrast,other vegetable oils tested (mustard, olive, and soy-bean oils) significantly worsened skin-barrier func-

tion (ie, increased TEWL) 1 and 5 hours after appli-cation and delayed recovery of a compromised skinbarrier compared with control- or Aquaphor-treatedskin. Moreover, a single application of mustard oilresulted in a variety of adverse ultrastructuralchanges in the epidermis under transmission elec-tron microscopy, suggestive of toxicity.490,605 Thus,given the widespread use of mustard oil for massage

TABLE 36. Continued



Lincetto etal731*

Mozambique;urban hospitalsetting; PCS

2 cohorts of LBW (�2000 g) infants(n � 246) were enrolled. The firstcohort (n � 149) was selectedduring the cold season and asecond cohort (n � 122) wasenrolled in the hot season. Theintervention encouraged KMC athome. In 64%, routine follow-upexams after discharge wereperformed until infants reached aweight of 2500 g.

No seasonal differences in weight gainor the risk of complications werefound in infants treated with KMCin the hospital. Risk of seriouscomplications, including death (RR:1.96; P � .02) and readmission (RR:2.77; P � .04), was higher after dis-charge in the colder season becauseof mother’s incomplete compliancewith KMC and exposure to lowtemperatures.

Bergman585 Zimbabwe;missionhospitalsetting; PCS

KMC was introduced as theexclusive means of treating LBWinfants (n � 126), withoutincubators and standardequipment for care of LBWneonates. Survival with KMCwere compared to hospitalsurvival data prior to theintervention, from 1983 to 1987.

The survival of babies born �1500 gimproved from 10% to 50%,whereas that of infants 1500–1999 gimproved from 70% to 90%. Overallsurvival rate was 63%.

Christenssonet al475

Zambia; urbanuniversityhospitalNICU; RCT

80 consecutive low-riskhypothermic infants withadmission weight of �1500 gadmitted in the NICU wererandomly assigned treatmentwith skin-to-skin (STS) care bythe mother (n � 41) or in anincubator (n � 39).

At 240 min, 90% of babies in STSgroup reached normal temperature(37.1–37.2°C), compared with 60% inthe incubator group (P � .0001).

Lincetto etal586*

Mozambique;provincialhospitalsetting; PCS

KMC for LBW infants (n � 32) wasintroduced with limitedresources and without anintensive care unit. Care includedpost-discharge follow-up visits toall infants �1800 g. Mothers ofinfants in the intervention group(n � 22) were taught skin-to-skincare and asked to observe othermothers to facilitate compliance.Control infants (n � 10) weregiven standard non-KMC care inthe maternity ward.

Out of 32 LBW infants (�1800 g)admitted in 3 mo, survival was 73%in KMC and 20% in non-KMCinfants (P � .01).

Charpaket al584

Colombia; 2tertiary carehospitals, 1offeringtraditionalcare and theother KMC;PCS

332 newborns weighing �2000 gand eligible for care in theminimal care unit were enrolledin either of 2 groups: 1 receivingKMC (n � 162), and 1 controlgroup (n � 170). KMC infantseligible for MCU weredischarged regardless ofgestational age or weight, werekept 24 h/d in an uprightposition, attached to the mother’schest and receiving skin-to-skincontact until KMC was nottolerated anymore. Controlbabies from another facility werekept in incubators until theywere discharged. Both groupswere followed periodically up tothe age of 1 y.

RR of death was found to be higher inKMC infants (RR: 1.9), though thereverse was found after adjustingfor weight at birth and gestationalage (RR: 0.5). There were nosignificant differences in overallmortality rates among the 2 groups.KMC infants grew less in first 3 moand had a higher proportion ofdevelopmental delay at 1 y.




of newborns, as documented in south Asia,488,606 itseems that millions of newborns each year are beingtreated with potentially toxic applications to theskin.488,490

In clinical trials in developed countries, topicalapplications of epidermal barrier-enhancing emol-lients have been shown to decrease TEWL, improveskin condition, and minimize skin injury in ex-tremely preterm infants.490,594,607–609 However, topi-cal therapy has had a variable impact on risk ofinfection. A pilot study at Stanford University (Stan-ford, CA) showed that twice-daily application of theointment Aquaphor for the first 2 weeks of life topremature infants �33 weeks’ gestational age re-sulted in improved skin condition and a reduction inepisodes of culture-proven sepsis due to coagulase-negative staphylococci.608 In another trial, Aquaphortherapy in neonates weighing �1500 g decreased thenosocomial bloodstream infection rate to 5.4 per 1000patient-days, compared with 12.7 per 1000 patient-days during the preceding 16 months.610 In contrast,a small CCS suggested that extremely preterm in-fants weighing �1000 g who were treated with top-ical petrolatum ointment were at increased risk forCandida infections,611 and another report suggestedthat topical applications of Aquaphor might serve asa source for nosocomial infections with coagulase-negative staphylococci and Gram-negative organ-isms.612 A recently completed multicenter US-basedtrial showed that Aquaphor therapy increased riskof sepsis with coagulase-negative staphylococciamong neonates weighing 501 to 750 g (OR: 1.60; CI:1.07–2.39); no effects were seen in neonates weighing751 to 1000 g.613 A Cochrane review of studies indeveloped countries concluded that prophylactictopical-ointment therapy increases the risk of coag-ulase-negative staphylococcal infection and any nos-ocomial infection.614

community-based evidence. Oil massage of neo-nates is a nearly universal practice in south Asia.488,606

The typical timing (from the first days of life), fre-quency (typically 1–3 times daily), pattern (total body),

and duration (throughout infancy and early childhood)of use suggest that the practice is an important event indaily child care, that significant time and resources aredevoted to it, and that exposure of the infant’s skin tothe oil is significant. The most commonly used oil forinfant massage is mustard oil, which, as noted above, ispotentially toxic490 and, when tainted with seeds of theweed Argemone mexicana, may cause the neurologicsyndrome coined “epidemic dropsy.”615,616

In India, topical therapy with corn oil improvedthermoregulation (see also “Hypothermia Preven-tion and Management”) of 5- to 7-week-old infants inthe community,491 and growth of hospitalized pre-term infants was improved with sesame-seed oiltherapy.617 In Nepal, it was found that traditional oilmassage with mustard oil, swaddling with a plasticswaddler, or KMC was equally effective in pre-venting hypothermia during the first 24 hours afterbirth (see Table 29).473

No community-based data are available on theeffect of topical therapy on neonatal infections ormortality. Hospital-based studies have been con-cluded in Egypt and Bangladesh, and data have beenreported from Egypt (Table 37). In Egypt, topicalapplication of sunflower-seed oil 3 times daily topreterm infants �34 weeks gestational age (n � 51)at the Kasr El-Aini neonatal intensive care unit atCairo University significantly improved skin condi-tion (P � .037) and reduced the incidence of nosoco-mial infections (incidence rate ratio: 0.46; CI: 0.26–0.81; P � .007; adjusted for weight on admission,gestational age, and gender) compared with infantsnot receiving topical prophylaxis (n � 52) (see ref732). In Bangladesh, preliminary data analysis sug-gested that emollient therapy of preterm infants �33weeks gestational age 3 times daily with sunflower-seed oil or Aquaphor reduced the odds of neonatalmortality (OR: 0.75; C:I 0.57–0.98). Among infantsweighing �1250 g, emollient therapy reduced theodds of mortality by 35% (OR: 0.63; CI: 0.46–0.87).618

conclusions. Emollient therapy is a promisingintervention, particularly for LBW infants in devel-

TABLE 37. Neonatal Topical Emollient Therapy



Ahmed et al618 Bangladesh; urbanhospital setting;RCT (preliminaryreport)

In a Special Care Nursery, neonates(n � 385) were randomized totopical application of sunflower oilor Aquaphor 3 times a day, or tothe control group, which receivedstandard nursery care.

Treated infants (sunflower-seed oiland Aquaphor-treated infantscombined) had a significantlyreduced risk of neonatalmortality (25%) (OR: 0.75;CI: 0.57–0.98). Among infantsweighing �1250 g, emollienttherapy reduced the odds ofmortality by 37% (OR: 0.63; CI:0.46–0.87).

Darmstadt et al732 Egypt; urbanhospital setting;RCT

Neonates were randomized to 3 timesdaily topical application ofsunflower oil (n � 51) or to usualnursery practice (control group)(n � 52).

Treatment with sunflower-seed oilresulted in a significant improve-ment in skin condition (P �.037) and a highly significantreduction in the incidence ofnosocomial infections (adjustedincidence ratio: 0.46; CI: 0.26–0.81; P � .007) compared withinfants not receiving topicalprophylaxis (n � 52).



oping countries, although data to date are hospital-based. Data thus far are also restricted to LBW in-fants �33 to 34 weeks’ gestational age, in whom theskin barrier is likely to be most highly compromised.Additional research is needed on the impact of thistherapy in the community and among all newborns(without restriction based on birth weight).

Emollient therapy is inexpensive (� $0.20 [US dol-lars] for a course of therapy in a preterm infantweighing 1.5 kg)490 and technologically simple andcan be delivered readily by health care workers andcaregivers. Research suggests that the practice of oilmassage is nearly universal in south Asia, and care-givers indicate a willingness to modify their behaviorand apply an alternative oil that has skin barrier–enhancing properties.488,606 Behavior-change com-munications may be an effective way to introducethe intervention more broadly by modifying what isalready a well-entrenched traditional practice. Fur-thermore, in many communities, cadres of workerssuch as the domin or naun in India (and similarly, inNepal) are designated to apply oil massage in thecommunity to newborns, and working with themcould enhance the feasibility and scalability of theintervention.606

Choice of emollient is important. Emollients con-taining a physiologic balance of epidermal lipids(3:1:1:1 molar ratio of cholesterol/ceramide/palmi-tate/linoleate) are optimal for barrier repair,604,619

and tests of impact on neonatal outcomes are war-ranted using topical products with optimized effectson skin-barrier function. However, natural vegetableoils are readily available worldwide and may pro-vide a simpler, inexpensive alternative that also war-rants additional investigation as a more readily scal-able and affordable intervention, particularly for usein the community.

Hyperbilirubinemia Screeningbackground. Hyperbilirubinemia is widespread

among newborns. Almost 97% of all healthy new-born infants have biochemical hyperbilirubinemia(ie, serum bilirubin �17.1 �mol/L), and almost 65%may be overtly jaundiced. Jaundice tends to occurin term infants with serum bilirubin levels above�85 �mol/L.620 Although hyperbilirubinemia maynot be a major direct cause of mortality in developingcountries, delayed detection and therapy may pre-dispose infants to bilirubin encephalopathy and ker-nicterus,621,622 both of which are common causes ofhandicaps in developing countries.419,623

It is not the purpose of this review to discuss thenumerous options for management of hyperbiliru-binemia but rather to evaluate strategies for its earlyrecognition in community settings. Management ofthis condition has been reviewed extensively, andinterventions such as phototherapy and medicaltherapy (eg, phenobarbital, exchange transfusion,and management of underlying conditions such assepsis, hemoglobinopathies, and blood-group in-compatibilities) largely pertain to hospital set-tings.624 Nevertheless, of the known risk factors forjaundice and bilirubin encephalopathy, prematurityand past history of hemolytic disease of the newborn

are important. Frequent and exclusive breastfeedingmust be encouraged, because it has been shown toreduce the propensity for hyperbilirubinemia, due atleast in part to improved hydration status.625,626

Early detection of jaundice in developing countrieswould be significantly facilitated if first-line healthworkers at peripheral health facilities were able toaccurately identify young infants with clinically sig-nificant jaundice who needed care at a referral-levelfacility. Several studies, mostly from developedcountries, have found that bilirubin levels deter-mined with a simple hand-held icterometer corre-lated to varying degrees with measures made byusing more expensive and sophisticated instrumentsand standard techniques (ie, serum bilirubin mea-surement).627–632 However, there has been consider-able debate, even in developed countries, aboutwhether the presence of jaundice can be determinedreliably by using visual assessment and what cadresof individuals can perform visual assessments reli-ably. Although some investigators have found thatclinical estimation of jaundice in newborns by neo-natologists633 or physicians or nurses631,634,635 washighly correlated with serum bilirubin levels, othershave questioned the ability of health care providers,including physicians and nurses, to diagnose clinicaljaundice by visual estimation.628,636

community-based evidence. Our review identi-fied 5 studies from developing countries that evalu-ated strategies for the early detection of neonatalhyperbilirubinemia (Table 38); however, none ofthese studies were in a rural setting (all were con-ducted in urban hospitals). Only 1 of these studiestested the accuracy of visual estimation of jaundice;Riskin et al633 reported that neonatologists’ visualestimates were highly correlated to measurementsusing a spectrophotometer (r: 0.682; P � .001).

All the other studies tested the accuracy of nonin-vasive techniques (eg, use of a hand-held icterometeror bilirubinometer) relative to laboratory serum bil-irubin measurement. Bilgen et al632 and Kumar etal627 compared transcutaneous bilirubin readings byusing a bilirubinometer to serum bilirubin, and bothreported strong evidence of a linear correlation (r �0.83 and 0.91, respectively). Two studies, 1 in Turkeyand 1 in India, also found a significant positive cor-relation between icterometer measurement and se-rum bilirubin levels.630,632 However, in Pakistan,Bhutta and Yusuf629 found that transcutaneous bil-irubinometry was much less effective than standardbiochemical testing in accurately identifying hyper-bilirubinemia, particularly among dark-skinnedinfants.

conclusions. Simple methods including visualassessment of degree of jaundice by health workersand use of simple, hand-held icterometers holdpromise for use at the community and primary carelevels in developing countries, but additional studiesare warranted. The role of visual inspection in thehome by caregivers, particularly in areas with a rel-atively high prevalence of hyperbilirubinemia, hasnever been evaluated.



TBA/CHW Trainingbackground. Although there is consensus that in-

creased access to skilled health care, including askilled birth attendant during delivery, is widelydesired and likely to improve pregnancy out-comes,16,24 approximately half of women and new-borns lack access to skilled care.18 Moreover, nearlytwo thirds of births worldwide take place in thehome, and a large percentage of births in developingcountries take place in the hands of untrained ortrained TBAs. Given the projected absence of ade-quate cadres of skilled birth attendants in manycountries for years to come and a potential role forTBAs and/or CHWs in health service delivery fornewborn care, we evaluated the available evidencefor the impact of TBA and CHW training on perina-tal outcomes. CHWs typically differ from TBAs ineducation (higher) and age (younger) and are per-ceived by many as being more amenable to train-ing.24 However, we considered them together for thepurposes of this review.

community-based evidence. A number of studieshave assessed changes in knowledge and attitudesafter the training of TBAs.440,637–642 An observationalstudy showed that trained TBAs were more likelythan untrained TBAs to give advice on breastfeedingas well as on immunizations and oral rehydration

therapy.643 Trained TBAs also had better skills inneonatal resuscitation relative to untrained TBAs butdid not have any impact on maternal behaviors.Kumar et al640 stressed that continued training ofTBAs was important in sustaining improved mater-nal and newborn care practices.

Table 39 (see also Tables 26, 41, and 42) summa-rizes the evidence from studies that examined theeffect of TBA and CHW training on perinatal andneonatal health status outcomes in community set-tings. The CHWs were community based in all set-tings and had varying levels of training and linkageto the health system. Their relationship to the healthsystem ranged from close association that facilitatedreferral of high-risk infants and/or those with com-plications to settings in which care was entirely homebased.439

The 1 randomized trial in The Gambia644 showed areduction in neonatal mortality of 61% in the villagesto which primary health care was introduced andTBAs were trained. However, neonatal mortalityalso fell by 35% in the control villages, rendering the26% comparative fall in neonatal mortality nonsig-nificant. There was a significant increase in the num-ber of women receiving antenatal visits and tetanusimmunization during pregnancy, although the num-ber of women whose births were conducted by a

TABLE 38. Neonatal Hyperbilirubinemia Screening


Intervention Neonatal HyperbilirubinemiaDetection

Bilgen et al632 Turkey; urbanhospital setting;QT

Transcutaneous measurements were obtainedfrom newborns (n � 96) with a MinoltaAirShield bilirubinometer and an Ingramicterometer and compared to serumbilirubin measures.

A linear correlation existed betweenserum bilirubin values and thereadings on both the Minoltabilirubinometer (r � 0.83) and theIngram icterometer (r � 0.78).

Riskin et al633 Israel; urbanurban hospitalsetting; PCS

4 neonatologists were asked to estimatevisually the level of bilirubin in a group ofterm clinically jaundiced infants (n � 283)before discharge from the nursery on thethird day of life. Their clinical estimationwas compared with actual measurement ofserum total bilirubin from samples drawnsimultaneously.

Clinical visual estimation of serumtotal bilirubin had a highcorrelation to actual serumbilirubin levels as measured byspectrophotometer (r � 0.682;P � .001).

Kumar et al627 India; urbanhospital setting;RCT

Term babies (n � 100) were randomlyselected for estimation of transcutaneousbilirubin (by Minolta AirShieldbilirubinometer 101) and serum bilirubinlevels by conventional diazo method.Babies needing phototherapy (n � 40)were randomized into 2 groups: (1) thosehaving their transcutaneous bilirubin levelmeasured from a covered area on theforehead, and (2) those who did not have acovered area on the forehead.

There was a linear correlation (r �0.9090; P � .001) between serumbilirubin level and transcutaneousbilirubin level. In the babiesrequiring phototherapy, therewas significant (P � .001) linearcorrelation between prephoto-therapy and postphototherapyreadings, group 1 (r � 0.69) andgroup 2 (r � 0.80).

Bhutta et al629 Pakistan; urbanhospital setting;QT

Transcutaneous bilirubinometry wascompared with standard laboratoryestimation of bilirubin level in normaljaundiced newborns (n � 65).

Transcutaneous bilirubinometrywas not very effective inrecognizing hyperbilirubinemia incomparison with biochemicaltesting, especially among darkinfants. Although the correlationbetween the 2 methods wassignificant (r � 0.66; P � .01), thescatter was wide and the speci-ficity was only 53%.

Narayanan et al630 India; urbanhospital setting;QT

Effectiveness of the icterometer wasevaluated in detecting neonatalhyperbilirubinemia in newborn infants(n � 158).

There was positive correlationbetween the readings on the noseand serum bilirubin levels(P � .05).



trained midwife fell by 7% during the trial, con-founding precise interpretation of the effect of TBAtraining. In another study, perinatal mortality de-creased by 19% among offspring of women deliveredby TBAs who had been trained in essential newborncare and advanced resuscitation, including use of themucus extractor and bag-and-mask ventilation.639

In an uncontrolled study in India,438 CHWs iden-

tified and managed high-risk neonates in the home,including preterm and LBW infants and those withfeeding problems, illness, or a history of prolongedand difficult labor. Interventions included resuscita-tion of asphyxiated newborns, including cleaning ofthe mouth and pharynx and mouth-to-mouth respi-ration; antiseptic cord care, including cutting; pro-motion of breastfeeding; minimal handling to reduce

TABLE 39. TBA/CHW Training



Alisjahbana733 Indonesia; urbanslums; PCS

TBAs (n � 20) were providedwith a weighing scale withcolor codes and also apictorial chart depictingmaternal and child con-ditions. Using these, theTBAs determined the risk ofa poor pregnancy outcomefor the mothers and theirneonates (n � 178 mother-infant pairs).

The weights determined by theTBAs using the colored scalewere as accurate as theweights determined bymidwives.

Kumar639 See Table 26Rashid

et al643Bangladesh;

rural setting;CCS

TBAs in the intervention group(n � 28) were provided 3mo of basic training and 36mo of supervised follow-uptraining. They were com-pared to a cohort of un-trained TBAs (n � 27).

Although there was a significantdifference between the trained anduntrained TBAs in the adviceprovided to mothers regardingimmunization, oral rehydrationtherapy use and breastfeedingpractices, there was no associatedchange in behavior of the mothers.

Greenwoodet al644

The Gambia;rural setting;RCT

TBAs were trained to provideantenatal care (especiallymalarial prophylaxis), riskassessment and subsequentreferral, intrapartum care(especially safe and cleandelivery) and neonatal care.Mortality and morbiditydata for 1913 women overthe 3 y of the program werecompared to baseline datafrom 673 women.

There was a 62% decrease inmaternal mortality in theintervention area comparedwith the preinterventionperiod. Such a decrease wasnot observed in the non-intervention areas. There wasalso a significant increase inmaternal tetanusimmunization rates, maternalantenatal attendance andmothers who went to atrained TBA for delivery.

There was a significant 33% decreasein NMR in the intervention areacompared to the noninterventionarea. This change was mainlyattributed to a decrease in rates oflate neonatal deaths, especiallythose of infectious origin.

Daga et al441* See Table 26Daga et al449* See Table 26Smith et al648 Ghana; rural

setting; cross-sectionalhouseholdsurvey withTBA clients

The program was designed toimprove workingrelationships between TBAsand other health providersin 2 districts of Brong-AhafoRegion. TBAs receivedcomprehensive trainingcovering care duringantepartum, intrapartum,and postnatal periods andwere provided with a kitcontaining midwifery andpublic health items. 2-weektraining containedinstructions on care/management of normalpregnancy, recognition ofcomplications and referral,care of the newborn and 5nonobstetric primary healthcare topics; family planning,infant feeding, growthmonitoring, immunizationand control of diarrhealdiseases.

Training was protective againstpostnatal fever (OR: 0.30; CI:0.14–0.65), retained placenta(OR: 0.35; CI: 0.13–0.96), andlabor �18 h (OR: 2.57; CI:1.13–5.81).

Authors concluded that “theevidence for a beneficial impact ofTBA training on the health ofmothers and newborns is notcompelling. Some moderatebeneficial effects may beforthcoming, but it is unlikely thatTBA training will result in largereductions in maternal andperinatal morbidity and mortality.”




the risk of infection; postnatal home visitation andprovision of anticipatory guidance about routinenewborn care; and referral of sick newborns and

those with feeding problems to a hospital. Home careof preterm and LBW infants also included keepingthe room warm and feeding with a dropper if nec-

TABLE 39. Continued



Janowitzet al647

Brazil; ruralhospitalsetting; RCS

Data was collected from arandom sample of womenwho had recently deliveredover a 1-y period (n �1961) to compare outcomesof deliveries by familymembers, by TBAs, and inhospitals. TBAs in thestudy were trained in localhospitals or in the home ofa centrally located TBAfollowing the officialtraining plan of the localteaching hospital, and hadregular supervision from aphysician or nurse. TBAsprimarily providedassistance only duringdelivery but were alsogiven training in prenataland postpartum care, aswell as referral criteria.

Over the course of 1 y, 139 outof 1057 women who werebeing seen by TBAs plannedto deliver at home but werereferred to a hospitalinstead. An additional 119women gave birth at homewithout the assistance of aTBA.

While the survival rate amongbirths to women having their firstdelivery in a hospital was 4 timesthat of a first delivery in a home,hospital intervention did nothave a statistically significantimpact on the survival of infantsborn to women who were mostlikely to be referred.

Berggrenet al385

Haiti; ruralsetting; RCS

Training programs for TBAswith Haiti’s maternal andchild health family-planning program wereintegrated in order toreduce maternal andnewborn mortality. As aresult, a manual fortraining to be used byprofessionals was createdbased uponrecommendations bypersonnel from governmentand private institutionswho had extensiveexperience in trainingindigenous women asmidwives. The number ofmidwives reached by thetraining program increasedfrom 36 in 1968 to 175 in1970.

In the Trou Chouchou area,maternal mortality seemedto be unaffected.

Neonatal mortality was eliminatedonce the integrated trainingprogram had been implemented.Training of traditional midwivesat the Albert Schweitzer Hospitalresulted in an increasedacceptance of prenatalconsultations (80% after theprogram was introduced,compared with 37% at baseline)along with fewer deaths fromneonatal tetanus (1 neonataldeath from tetanus post-introduction, vs 15 neonataldeaths from tetanus at baseline;sample too small to reachsignificance).

Sibley andSipe3;Bang etal439

See Table 26 Neonatal, infant and PMRs in theintervention area weresignificantly reduced (62%, 46%,and 71%, respectively) comparedto the control area. Case fatalityresulting from neonatal sepsissignificantly declined, from 16.6%before treatment to 2.8% aftertreatment by VHWs.

Pratinidhiet al438

See Table 26

Rahman384 See Table 20Meegan

et al389See Table 20

O’Rourke etal642

Guatemala;rural setting;trainingprogramreport

A TBA training program toimprove relations betweenhospital staff and TBAs, aswell as to improvestandards of care, wasimplemented at a ruralhospital. Mothers referredby TBAs were surveyed toevaluate the program.

Apgar scores improved and PMRdeclined, but the change was notstatistically significant.



essary. In the absence of management of sepsis,newborn mortality declined by 25% during the inter-vention year compared with the year before imple-mentation of the program.

Similar but less stringently evaluated success hasbeen reported by Daga et al441,449 from the RuralNeonatal Care Project in Dahanu, India, where NMRand PMR fell 41% and 62%, respectively, over a3-year period. The project used TBAs as the meansfor delivery of neonatal care. Maintenance of the“warm chain,” resuscitation of asphyxiated new-borns, and appropriate, early referral of high-riskmothers and newborns (eg, LBW) were recognizedas the most important interventions.441 Anganwadiworkers also played an important role in bridgingthe gap between TBAs and formal health workers.449

Overall, the IMR in the block to which these serviceswere introduced declined 70% over a 4-year period,from 148 to 45 per 1000.449

As part of the Narangwal Nutrition Study, a con-trolled trial in a rural area of Punjab, India, the RuralHealth Research Centre trained a cadre of villagersas family health workers (FHWs) (lady health visi-tors) to provide health services in the villages inwhich they lived. All were given special training incomprehensive primary care and surveillance tech-niques and provided prevention, early diagnosis,management of illnesses (including penicillin injec-tions for suspected pneumonia in infants), and refer-ral services in village clinics and homes.645 FHWswere the principal providers of nutrition and medi-cal care in the study and were data collectors forlongitudinal fertility and dietary surveys. They haddetailed, written standing orders and were givenweekly supervision by a doctor and a public healthnurse to provide support, facilitate referrals, and en-sure quality data collection. FHWs spent at least halftheir time performing home visits, which helpedthem identify pregnancies early and provide basicprenatal care to all women in service villages.244,645

The interventions significantly reduced perinatalmortality. Analysis of service records revealed that90% of patient contacts were handled capably byFHWs, and the remainder were safely referred tophysicians.244

In another study in India, Society for Education,Action and Research in Community Health(SEARCH) trained TBAs in elements of clean deliv-ery and CHWs in a variety of antenatal and postnatalinterventions including provision of health educa-tion to caregivers in the home, promotion of breast-feeding, management of birth asphyxia, identifica-tion of premature and LBW (ie, high-risk) infants,prevention and management of hypothermia, treat-ment of skin and cord infections, identification ofsick newborns, and administration of oral and inject-able antibiotics for cases of suspected sepsis.646 Al-though the precise impact of the TBAs versus theCHWs cannot be determined, there was little to noinput from the formal health sector; thus, the impactapparently was largely facilitated by these trainedvillage-level workers alone. The NMR fell by 62%,the PMR fell by 71%, and stillbirths were reduced by18%.

A community-based trial conducted in Tanzaniaand Kenya demonstrated that, for areas in whichmaternal immunization was not a feasible method ofdecreasing tetanus morbidity and mortality, mea-sures such as TBA training for safe and clean deliv-ery and cord care were effective in decreasing peri-natal, neonatal, and infant mortality389 (Table 20).

Janowitz reported that trained TBAs in rural Brazilidentified and appropriately referred most high-riskwomen.647 In another study evaluating the impact ofTBA training in Ghana, Smith et al648 found reduc-tions in intrapartum fever, labor �18 hours, andretained placenta, but the number of perinatal deathsand tetanus cases were too low to evaluate. Clients oftrained TBAs were more likely than those of un-trained TBAs to be referred for tetanus immuniza-tion (58% vs 28%, respectively), and immunizationrates in both groups were high (87% vs 77%, respec-tively). In this setting, in which health providersother than TBAs were the primary antenatal carecontacts, the value of training TBAs in clean cordcare was questioned.648

O’Rourke642 reported a twofold increase over anearly 3-year period in referrals of women withpregnancy complications after introduction of a TBAtraining program in Guatemala. A 27% nonsignifi-cant reduction in PMR was reported using a before-after comparison.

A recent meta-analysis of TBA training showedthat there were statistically significant decreases of11% in neonatal complications and 6% in perinataldeaths in areas served by trained TBAs comparedwith areas without trained TBAs.3

conclusions. The overall effects of TBA trainingon perinatal and neonatal outcomes were generallybeneficial. Little evidence was found to support thewidely held notion that TBA training is futile. On thecontrary, training was associated with improved be-haviors (ie, advice, skills) among birth attendants,although it did not necessarily translate into im-proved maternal behaviors regarding newborn care.In some cases, caregiver behaviors did improve: themeta-analysis by Sibley and Sipe3 showed improve-ments in overall TBA and caregiver knowledge aswell as behaviors. Although there has been littleevaluation of the impact of TBA training on perinataland neonatal health outcomes, the weight of evi-dence suggests that TBAs may make positive contri-butions to newborn care, and further definition ofTBA and CHW roles and evaluation of the impact ofTBA training on perinatal and neonatal outcomes aremerited.

TBA training need not be seen in isolation fromother interventions, particularly interventions thatimprove the skills of other caregivers. The Gadchirolitrial successfully used both TBAs and CHWs work-ing as a team along with the mother and other care-givers, particularly the mother-in-law, to improvedomiciliary care for the mothers and newborn in-fants.439 These interventions, therefore, may be con-sidered for settings in which TBAs and CHWs willremain an important group of care providers for theforeseeable future. However, additional research is



needed to define the interventions that TBAs and/orCHWs can be trained to provide most effectively inprograms at scale, to delineate criteria for selection ofTBAs to be trained, to develop appropriate preser-vice and in-service training programs, and to evalu-ate the cost-effectiveness of such training programs.

Pneumonia Case Managementbackground. The true incidence of neonatal

pneumonia or ARI at the community level is notknown; however, a significant proportion of new-born infants diagnosed with sepsis or serious infec-tions may have associated pneumonia, and the clin-ical presentations of these conditions have significantoverlap in the newborn. Recommended treatment ofserious systemic bacterial infections, including pneu-monia, in neonates in developed countries includesparenteral administration of antibiotics in a healthcare facility. Similarly, in developing countries, theWHO recommends parenteral antibiotic therapy (eg,benzylpenicillin or ampicillin plus an aminoglyco-side such as gentamicin) in a health facility for treat-ment of serious neonatal infections.16,25,649,650 In re-source-poor countries, however, the majority ofbirths and neonatal deaths take place in the home,and families are often reluctant to seek care outsidethe home for neonatal illness.439,569–571 In these set-tings, facility-based care with use of parenteral anti-biotics is currently infeasible for many neonates, andalternative management strategies are needed. In-sight into the role of oral antibiotic therapy as apotentially simpler and more feasible regimen fortreatment of serious neonatal infections for situationsin which referral to a health facility for quality careand parenteral therapy is not possible may be gainedby reviewing pneumonia case management trialsthat included treatment of neonates.

community-based evidence. We identified 6 re-ports654 and 2 meta-analyses of community-basedtrials of pneumonia case management that includedneonates (Table 40). Sazawal and Black651,652 con-ducted a meta-analysis of community-based inter-vention trials on case management of pneumonia inunselected preschool-aged children in developingcountries. The impact of pneumonia case manage-ment on total neonatal mortality and pneumonia-specific neonatal mortality was determined. The 5studies comprising the neonatal mortality analysiscompared concurrent control and treatmentgroups.653–660 In 4 of the studies, neonates with sus-pected pneumonia were treated with oral co-trimox-azole,653–656,659,661 and another used both injectablepenicillin and oral ampicillin.660 Uncorrected analy-sis showed a 27% reduction in all-cause neonatalmortality (CI: 18–35).652 After correction for per-ceived biases that may have affected the study re-sults, the estimated reduction in total and pneumo-nia-specific neonatal mortality was 30% (CI: 15–42%;OR: 0.70, 0.59–0.84) and 42% (CI: 20% to 64%; OR:0.56, 0.37–0.83), respectively.

In an additional study in Indonesia, use of ampi-cillin plus supportive care (eg, continued breastfeed-ing, clearing of the nose, fever control) in childrenwith pneumonia had no measurable impact on cure

rates of mild disease at 1-week follow-up, and didnot halt progression to moderate disease at 1 weekcompared with the use of supportive care alone inthe control group.662

conclusions. Several studies have shown that itis possible to train CHWs to recognize and success-fully treat pneumonia in newborn infants in commu-nity-based settings.653–660,663 The impact of suchtraining on reducing neonatal mortality and morbid-ity in community settings in developing countries issignificant. Available data clearly indicate that a casemanagement approach emphasizing essential new-born care along with prompt recognition of seriousbacterial infections and treatment with oral antibiot-ics is superior to no case management. In fact, thereduction in neonatal mortality found in the meta-analysis was comparable to estimates of the propor-tion of neonatal deaths due to infections (ie, 32%18),suggesting that a substantial proportion of thesedeaths were averted through case management thatprimarily included oral antibiotic therapy in thehome. However, the impact has its limits, given thatit is impossible in many cases to distinguish pneu-monia from sepsis on clinical grounds, and theseconditions frequently coexist during the neonatal pe-riod. Thus, pneumonia case management alone, in-cluding use of oral antibiotics, will lead to under-treatment of many infants with more widespreadinfections (ie, sepsis) that require parenteral therapy.It was this realization that led the group at SEARCHto provide more comprehensive detection and man-agement of serious neonatal infections with a com-bination of oral and parenteral antibiotics (see “Neo-natal Care Packages”). A host of factors, includingfeasibility and cost-effectiveness of implementation,as well as ethical standards, must be consideredwhen weighing the relative merits of neonatal pneu-monia versus more comprehensive sepsis casemanagement.

Neonatal Care Packagesbackground. Given that the majority of births

take place at home in developing countries, fre-quently in the hands of relatively untrained birthattendants, there is considerable interest in the po-tential of developing locally adapted packages ofnewborn care in community-based settings. Whereasthe previous sections reviewed trials that focused onuse of trained TBAs/CHWs and on pneumonia casemanagement, this section discusses more compre-hensive packages of care that may or may not haveincluded management of sick newborns.

community-based evidence. We identified 9studies that were undertaken in community settingsand included a comprehensive neonatal care planrather than solitary interventions (Table 41). Severalof these studies used TBAs and/or CHWs to providecare and thus were discussed in detail in “TBA/CHW Training” (Table 39).

In India, Pratinidhi et al438 used VHWs to identifyand manage high-risk neonates in the home. Inter-ventions addressed birth asphyxia, hypothermia pre-vention, clean cord care, breastfeeding promotion,postnatal visitation, and identification and referral of



TABLE 40. Neonatal Pneumonia Case Management



Bang etal658*


The project trained 86 TBAs in diagnosis andmanagement of under-5 children withpneumonia (n � 2568 attacks). They wereinitially trained in visual judgment of fastbreathing or difficult breathing; later in theproject, 10 TBAs were trained in the use ofsimple breath counters. Training in safe andhygienic delivery and newborn care wasalso given. Community acceptance of casemanagement by different workers wasassessed. Surveys were conducted every 6mo and causes of death in children weredetermined by verbal autopsy; see also Banget al.656,657

A significant decline in neonatal mortality due topneumonia (44%) was noted. The proportion oferror-free case management by TBA increasedcontinuously as the program progressed from56.7% in the first year, to 68.6% in the secondyear, and 83.4% in the third year (P � .0001).The proportion of correct diagnoses by 10 TBAsincreased from 60% to 82% when the breathcounter was used.

Bang etal657*


TBAs in the intervention area were trained todiagnose and treat pneumonia, and care-seeking for suspected pneumonia wasencouraged in families. Management ofneonatal pneumonia cases (n � 65) includedcontinued breastfeeding, administration ofcotrimoxazole syrup, and paracetamol forfever. Severe cases were promptly referred.

Case management led to a 40% reduction inpneumonia-specific mortality in the neonatalperiod, and a 78% reduction in mortality in thesecond month of life. 80% reduction in mortalitywas observed during the rest of infancy.

Bang etal656*


TBAs, VHWs and paramedics diagnosed andtreated pneumonia in the community. TBAswere also trained and provided withnecessary equipment to conduct safedeliveries and to resuscitate asphyxiatednewborns using mouth-to-mouth breathing.2180 women in villages were interviewed toevaluate the effectiveness of the program.

Pneumonia-specific mortality rate decreased by53.8% in the intervention area, and the case-specific fatality rate declined by 62% in theintervention area compared to the control area.

Datta etal568

India; rural setting;PCS

Primary health care workers in theintervention area of 38 villages wereidentified and trained in recognition,treatment and referral of ARI. Additionaltraining material including self-learningeducational aids were developed and fieldtested to ensure their suitability andsimplicity. Infants with moderate or severeARI were administered penicillin orally in 2equal daily doses of 125 mg for 5 d.

Primary health care workers were contacted in 38%of cases of ARIs in the intervention area asopposed to 1% in the control area. Meanduration of infections in the intervention areawas significantly lower (P � .01), whereas casefatality was �33% of that in the control area.

Pandeyet al659

Jumla, Nepal; ruralsetting; controlledintervention trial.

13 404 children were enrolled. Active casedetection by CHWs trained in WHOpneumonia case management strategy.Principal criteria for establishing a diagnosisof pneumonia were a respiratory rate of 50breaths per minute or greater or chestindrawing, measured through a speciallydeveloped battery-operated beeper timer toassess respiratory rate accurately. Emphasiswas placed on maternal education for earlydetection of signs of pneumonia. Parents ofpresumed cases were given freecotrimoxazole pediatric suspension (8 mgtrimethoprim/kg per d for 5 d) withdetailed instructions and workers watchedas parents administered the first dose.Workers revisited households on the thirdand sixth days to assess compliance and thechild’s status. Children who did notimprove after 4 doses of cotrimoxazole, orrelapsed, were treated with chloramphenicolsuspension as the sole second-line antibiotic.

Case referral by mothers increased from 15%during the first year to 56% by the third year.80% of pneumonia episodes were appropriatelydetected and treated. A statistically significantdecreased risk of death (RR: 0.72; CI: 0.63–0.82)was observed and by the third year of theprogram. The greatest RR reduction was seen ininfants between 6 and 11 mo and those between1 wk and 5 mo. In addition to reduction in thedeaths resulting from pneumonia, significantreductions in deaths due to diarrhea andmeasles were recorded, indicating that reductionin pneumonia morbidity had considerablecarryover effects.

Sazawalet al652

India, Tanzania,Pakistan,Bangladesh, andNepal; meta-analysis ofcommunity-basedneonatalpneumonia casemanagement trials

Meta-analysis found reduction in total under-5mortality of 27% (CI: 18–35%), 20% (CI: 11–28%)and 24% (CI: 14–33%) among neonates, infantsand children 0–4 y of age, respectively.Pneumonia-specific mortality was reduced by42% (OR: 22–57%), 36% (20–48%) and 36%(20–49%), in the same groups, respectively.




TABLE 41. Neonatal Care Packages



Bang etal439

See Tables 26 and39

O’Rourkeet al734

Bolivia; ruralsetting; PCS

Groups followed a 4-stage communityaction cycle to identify needs andpractical solutions to problems thatled to maternal and perinatalmortality. The impact of thisintervention was evaluated bycomparing PMRs and obstetricbehavior among community women(n � 409) before and after theintervention.

There was an increase inthe users of contraceptivesfrom an estimatedbaseline of �0–1% to 27%.There was a 71% increasein tetanus immunizationin women. An increase inattendance at antenatalclinics was observed.Women resorted more totrained Dais (trained birthattendants) as opposed tountrained Dais fordelivery.

There was a 63% decrease inPMR after the interventioncompared to baseline (44deaths/1000 births vs 117deaths/1000 births,respectively; P � .001).

Manandharet al665;Osrin etal69

Nepal; ruralsetting; RCT

Local female facilitators inintervention villages were trained tolead discussions within villagedevelopment committees andwomen’s groups about perinatalhealth issues. The groups thendeveloped participatory actionplans to solve perinatal healthproblems in an iterative process toreach 28 000 married women ofreproductive age. Common goals ofthe action plans includedsurveillance of birth outcomes,caregiver recognition of dangersigns, improved health workerknowledge and skills, cleandelivery, early breastfeeding, andimproved referrals.

Maternal mortality wassignificantly reduced inthe intervention clusterscompared to the controlclusters (OR: 022; CI:0.05–0.90).

A 30% reduction in IMRwas observed in theintervention clusterscompared to the controlclusters (OR: 0.70; CI:0.53–0.94).

Daga492 India; rural setting;PCS

6 babies with a foot length of �6 cm,whose mothers were not willing forhospital care, were managed athome using trained health workers.

All 6 newborn infantssurvived and did notrequire admission to thehospital.

Schieberet al735

Guatemala; ruralsetting; CCS

100 cases were obtained through arandom sample of all perineonataldeaths. 120 controls were selectedby enrolling the next registeredbirth after the study case in whichthe infant lived for at least 28 d).Mothers were interviewed, and datawere reviewed by physicians toidentify probable cause of death. 2analyses were performed, first toidentify predictors of perineonatalmortality and, second, to comparebaseline characteristics betweenwomen who elected traditional asopposed to modern care.

Population-basedattributable risks relatedto prematurity,malpresentation, andprolonged labordemonstrated that thesecomplications account forsignificant proportions ofobserved perineonatalmortality.

Daga etal449*

See Tables 26 and39

Daga etal441*

See Tables 26 and39

Bartlett etal664

Guatemala; ruralsetting; PCS

All pregnant women identified in thecommunity over a period of 1 ywere enrolled. Newborns (n � 320)were seen weekly in the first monthand biweekly in months 2 and 3.Mothers were taught routinenewborn care by fieldworkers. Aphysician examined the infantsbiweekly in the first month andmonthly in months 2 and 3. Forminor illnesses, physicians providedtreatment in the community. Forcomplicated cases, physiciansprovided immediate treatment inthe community, followed byreferral.

Mortality rate in the first 3mo of life was reduced by85%, compared tohistorical controls.



sick newborns and those with feeding problems tohealth facilities. Newborn mortality declined by 25%during the intervention year, compared with the yearbefore implementation of the program (Tables 26and 39).

In Narangwal, a town in Rural Punjab, India, 3service groups received nutrition care (nutrition ed-ucation, surveillance, and food supplementation of1674 J and 11 g/day of protein through special feed-ing centers), medical care (immunization, health ed-ucation, and early diagnosis and treatment of illnessthrough frequent surveillance), or both nutrition careand medical care.118,244,645 A fourth group served asthe control group. Results showed significant im-provements in growth (weight and height) and Hblevels of children. Medical care significantly reducedpostneonatal morbidity and mortality in the 1- to3-year-old age groups and decreased illness durationof all 6 conditions examined in this study. Prenatalnutrition care to pregnant women was most effectivein preventing perinatal deaths, followed by medicalcare for infants (Table 41).

Two decades ago in a rural area of The Gambia, apackage of primary health care interventions was

introduced by the Gambian government to improvematernal and neonatal birth outcomes.644 A VHWand a TBA from each village were trained to provideantenatal care (especially malarial prophylaxis), riskassessment and subsequent referral, intrapartumcare (especially safe and clean delivery), and neona-tal care. Alongside significant declines in maternalmortality over the 3 years after the program wasintroduced (from 2716 to 1051 of 100 000; �2: 5.9; P �.05), the program documented statistically signifi-cantly lower rates of neonatal deaths in interventionvillages relative to nonintervention villages (46.6 vs69.6 of 1000, respectively; �2: 4.3; P � .05), primarilydue to a reduction in late neonatal deaths due toinfections.

In Guatemala, pregnant women were taught rou-tine infant care and care seeking for illness, andwhen symptoms of severe illnesses were detected,immediate empiric treatment was begun in the com-munity with accompanied referral to an area hospi-tal.664 The mortality rate among infants enrolled inthe study was reduced by 85% compared with his-torical controls.

In rural India, Daga et al441,449,492 emphasized re-

TABLE 41. Continued



Pratinidhiet al438

Greenwoodet al644

See Tables 26 and39

See Table 39See Table 6

Kielmannet al118*

Morbidity duration wasshorter in villages withMC (or NUT � MC) thanin NUT or control villages(P � .02). Postneonatalmortality was lower invillages receiving MC orNUT � MC (23.3/1000and 35.2/1000 live birthsfor MC and NUT � MCgroups, respectively) thanthe control villages (50–52/1000). Differencebetween combinedmortality of villagesreceiving medical care(MC and NUT � MC) andthe control villages wasstatistically significant(P � .05). The differencebetween 1 and 7 dmortality in the 3intervention groupsrelative to the controlgroup was highlysignificant (1 d: 28/1000;7 d: 52.1/1000;P � .005).

Kielmannet al244*


Groups and interventions were thesame as Kielmann et al (See Table6).

Mothers in the interventiongroup prolonged breast-feeding by about 2 mobecause of nutritioneducation given to them.

IMR was lowest in the MCgroup (70/1000 livebirths), second in the MC� NUT group (81/1000live births) and third inthe NUT group (89/1000live births) as compared tothe control group (129/1000 live births).




suscitation of asphyxiated newborns, prevention ofhypothermia, and referral of sick newborns andachieved 41% and 62% reductions in NMR and PMR,respectively, compared with baseline data over a3-year period. The administration cost of the pro-gram was considered “affordable,” because it wasbased on inexpensive domiciliary neonatal care byTBAs supported by relatively low-cost facilities forneonatal care. This intervention was acceptable to thecommunity and the TBAs because it was simple tocarry out and did not involve technology that causes“culture shock”441 (Tables 26, 39, and 41).

An uncontrolled trial (the “Warmi” Project) to im-prove maternal and child health in the rural provinceof Inquisivi, Bolivia, included training TBAs, offeringtetanus immunizations to women of reproductiveage, ensuring a clean birthing surface and hygieniccord care, providing antibiotic eye drops, encourag-ing immediate breastfeeding, and ensuring thermalcontrol.642 Family-planning services and communityfunds and mechanisms for transport of women inneed of care to secondary health facilities were es-tablished also. The program emphasized communityparticipation in identifying and solving problems.This approach led to a 67% reduction in perinatalmortality over a 3-year period, an increased propor-tion of women receiving prenatal care, and an in-creased number of women initiating breastfeedingon the first day after birth compared with baseline(from 25% to 50%).

In rural India, SEARCH trained VHWs in interven-tion areas to provide a package of home-based neo-natal care, including health education to pregnantwomen, diagnosis and management of birth as-phyxia, identification of high-risk (premature andLBW) neonates for more intensive surveillance, tem-perature maintenance, promotion of breastfeeding,administration of vitamin K, treatment of skin infec-tions, and identification of sick newborns suspectedof having septicemia, meningitis, and/or pneumoniaand administration of antibiotics (oral co-trimox-azole and IM gentamicin) in the home.439 Neonatalmortality due to sepsis was reduced 76% and overallneonatal mortality declined by 62%, compared withthe control, nonintervention area.

More recently, in the rural district of Makwanpur,Nepal, Osrin et al69 conducted a cluster-randomizedtrial to evaluate a community-based participatoryintervention to improve essential newborn care. Theintervention, which covered a population of 28 000married women of reproductive age, trained localfemale facilitators in intervention villages to leaddiscussions within village-development committeesabout perinatal health issues. These facilitators alsohelped the groups develop participatory action plansin an iterative process. Mothers’ groups within eachvillage-development committee met monthly toidentify perinatal health problems, design and im-plement solutions, monitor birth outcomes, andshare results with others. Common goals of the ac-tion plans included community surveillance of birthsand birth outcomes, improved caregiver recognitionof danger signs, proper care seeking, improvedknowledge and skills of health workers, clean deliv-

ery practices, increased rates of early breastfeeding,and improved referral patterns. Infant mortality wasreduced by 30% (OR: 0.70; CI: 0.53–0.44) and mater-nal mortality was reduced by 78% (OR: 0.22; CI:0.05–0.90).665

conclusions. These data strongly support imple-mentation of packages of essential newborn care. Forsettings in which health facility capacity is limitedand referral is not possible, the data also support theutilization of trained community-based workers inthe screening and preventive and curative care ofnewborn infants. These interventions have not beenevaluated in effectiveness trials using available staffand infrastructure. Thus, there are legitimate con-cerns regarding the replicability and sustainability ofthese comprehensive programs of care. Additionalcost-effectiveness evaluations of this approach mustbe undertaken in larger trials in diverse geographiclocations.

Creation of linkages to referral facilities and facil-itation of referral pathways and improvement inquality of care at referral facilities were essentialaspects of most community-based newborn care pro-grams reporting improved survival. Nevertheless,some programs118,244,439,642,665 reduced neonatalmortality with little health systems, capacity, primar-ily through provision of improved antenatal andneonatal care in a context of community empower-ment.

Care in Peripheral Health Facilitiesbackground. In many settings, a primary strategy

to improve neonatal outcomes will be to increasehealthful essential newborn care practices in thehome, while also seeking to improve care seeking fornewborn illness, and stabilization and referral of sickinfants to health care facilities as close to the com-munity as possible, where better-quality care may beavailable.

community-based evidence. We identified 9studies that described and evaluated various postna-tal interventions in primary or secondary health fa-cilities in community settings (Table 42). Most of thefacilities described were community-based and hadlimited resources. Also included was a study thatevaluated postdischarge follow-up of neonates in acommunity setting.666

Various approaches to improving neonatal careand outcomes were evaluated in these studies, whichfocused primarily on training caregivers and healthproviders in essential newborn care. Although nostudy used a randomized, controlled design, im-provements were reported in survival441,494,667–669

and immunization rates,669 and reductions werenoted in length of hospitalization and number ofreadmissions after discharge.666

conclusions. Strengthening health facilities andhealth systems is an essential aspect of community-based programs to improve neonatal health. How-ever, given the barriers that exist to care seeking forneonatal illness,439,569–571 it is unlikely that utilizationof facility-based services for newborn care in manycommunities will increase in the absence of demandcreation, improved referral pathways, and quality of



care. Although it is unlikely that interventions basedin facility settings alone will make a major differencein situations in which the bulk of deliveries takeplace in domiciliary settings, they are nevertheless anessential adjunct to such care.

There are few existing guidelines for the care ofsick newborn infants in health care institutions withlimited facilities and manpower.548,670 Thus, there isan urgent need to improve and strengthen referralpathways and linkages between domiciliary-basedinterventions and facility-based care for sick neo-nates. There are considerable opportunities to reduce

cost of care and improve efficiency by such an ap-proach. Despite the limited information, we wouldstrongly endorse incorporation of basic guidelinesfor improved secondary care of newborns in healthfacilities as an essential adjunct to community-basedand domiciliary interventions.

COST-EFFECTIVENESS OF COMMUNITY-BASEDINTERVENTIONS

background. During the last few decades, signif-icant reductions in perinatal and neonatal mortalityhave been achieved in the context of studies evalu-

TABLE 42. Neonatal Care in Peripheral Health Facilities



Arif et al736 Pakistan; urbanhospital setting;RCT

362 babies were randomly assigned toeither an incubator or heated cot bythe mother’s side (n � 151), or to thespecial care baby unit (n � 211).Mothers were trained by nurses.

Weight gain from admission todischarge was significantlyhigher among newborn infants inthe maternal care group(P � .001). Mortality in the fairand poor babies was significantlylower (44%) in the maternal caregroup (P � .001). Overall,mortality in the maternal caregroup was 57% lower than thespecial care unit, irrespective ofcondition at admission.

Bose et al666 India; rural setting;RCS

At the Community Health andDevelopment Hospital, a low-costnursery for newborns (n � 175) wasestablished with facilities for photo-therapy, heat cradles and other basicnursing facilities.

Of 175 newborns admitted, 6 diedand 8 were transferred tonurseries in a tertiary care orgovernment hospital.

Borulkaret al737


Staff at a community hospital wastaught special care for newborns(n � 2266), comprised of provision ofoptimal warmth, resuscitation ofasphyxiated neonates, proper feedingand oxygen administration whenindicated.

For a 5-y period, survival was61.5% in the 1000–1500 g birthweight category and 92.5% in the1520–2000 g category ofadmissions. No comparablehistorical data was reported.

Dutt et al669 India; rural setting;PCS

Jawaharlal Institute Rural HealthCenter served as a treatment as wellas a referral center in the rural areafor births (n � 356).

94% of the deliveries wereconducted by trainedpersonnel, as compared to16% in 1967. All motherswere immunized againsttetanus.

IMR decreased by 68%(comparative neonatal mortalityrates were not specified). In thepostnatal period, �98% of thechildren had received the fullcourses of DPT, oral polio, andBCG immunizations. There wereno deaths due to neonataltetanus.

van der Meiet al667

Ghana; ruralhospital setting;PCS

Care of neonates (n � 567) was doneunder limited resources by trainingmothers and nurses in basic newborncare

Survival rates of infants weighing�1500 g and for those weighing1500–2000 g were 52% and 90%,respectively.

Wilkinson668 South Africa; ruralhospital setting;QT

The pilot study was based on ananalysis of consecutive hospital andclinic deliveries (n � 640). Asubsequent study included 2193consecutive hospital and clinic births.Basic interventions (including acommunity obstetrics guide for localmidwives, a policy change to admitall pre-eclamptic women with highdiastolic blood pressure to thehospital, a managed system for bloodcollection and result delivery, andtraining for midwives and skilledbirth attendants in neonatalresuscitation and management oflabor emergencies) were introducedin the hospital to tackle problemsidentified through a pilot study.

A 32% decrease in NMR was seenduring the intervention period.



ating a variety of interventions in developing-coun-try communities, as reviewed above. For some inter-ventions or packages of care, promising results inefficacy trials have included, or subsequently led to,evaluations of their cost-effectiveness.671,672 In gen-eral, however, attempts at costing interventionshave included posthoc data collection or extrapola-tion, which are not optimal for this kind ofanalysis.673

Studies in developed countries have shown thatearly initiation of antenatal care is among the mostcost-effective strategies for reducing neonatal mor-tality. In contrast, neonatal intensive care, despite itsimpact on mortality of the sickest patients, has beenfound to be one of the least cost-effective and feasibleapproaches.81,674 Several interventions in pregnancysuch as the preterm birth–prevention program inurban Los Angeles675 and management of gesta-tional diabetes mellitus676 have also been shown tobe cost-effective. Well-defined efforts in developedcountries to estimate the cost-effectiveness of variousother interventions have been reported, such as pre-vention of preterm births,675 reorganization of peri-

natal services,677 and establishment of perinataldatabases.678 In addition, certain specific postnatalinterventions such as neonatal resuscitation,679 indo-methacin therapy for premature infants,680 surfac-tant replacement therapy,681 prevention of respira-tory syncytial virus infections,682 or therapy forapnea of prematurity683 have been shown to be cost-effective interventions in the context of developedcountries. Most available information on cost-effec-tiveness of interventions refers to hospital or facility-based interventions, and these cost-effectivenessevaluations have used variable methods; only rarelyhave maternal and infant benefits together beenevaluated.

community-based evidence. Cost-effectivenessmodels and analyses of interventions to improveperinatal and neonatal outcomes have been infre-quently applied in emerging economies684 and arerarer still in developing countries. Examples of cost-effectiveness assessment of intervention strategies indeveloping countries have largely included hospital-based interventions such as routine ultrasonogra-

TABLE 42. Continued



Bhakooet al738

India; urbanhospital setting;RCS

Changes were made to the admissionand discharge criteria in theneonatal special care unit (NSCU),to encourage keeping the baby withthe mother and early hospitaldischarge for home care. Ratherthan be kept in the NSCU (wheremothers were not allowed), morehigh-risk babies stayed with theirmothers. Outcomes for infants keptin the NSCU (n � 165) werecompared to infants who werecared for alongside their mothers (n� 127).

After changes were made to theadmission and dischargecriteria, fewer newborn infantswho were cared for outside theNSCU alongside their mothersdied (7/127), as compared tothose admitted to the NSCU(57/165). Neonatal mortality inbabies weighing �2 kg declinedsignificantly over a 13-y period(7.94% in 1986 vs 12.88% in1973; P � .005), and over thesame period, mortality fellamong preterm babies from26.88 to 11.5% (P � .001). Thiswas achieved despite a two- tothreefold increase in the high-risk babies and without anyincrease in the number ofneonatal special care beds ornurses.

Daga etal494

India; urbanhospital setting;QT

A conservative neonatal care unit wasestablished in J J Hospital, Bombay,having 4 main features: 1) roomwarming, 2) exclusivebreastfeeding, 3) maternalinvolvement in infant care, and 4)minimum handling and minimumintervention. Birth outcomes for acohort of infants (n � 21) in 3different weight groups (1000–1250g, 1260–1500 g, and 1510–2000 g)were measured and compared tohistorical controls.

3 neonates were admitted withbirth weights of 1000–1250 g,and the survival was 33% inthis group compared to aprevious best of 50%. In thebirth weight group of 1260–1500 g, 8 neonates wereadmitted and their survival was75%, compared to 66% as aprevious best in this weightgroup. For the weight group1510–2000 g (n � 10 admitted),a 90% survival rate was found,compared to 92.5% as aprevious best.

Cooperet al739

South Africa; urbanhospital setting;RCT

A group of VLBW infants (n � 19)was fed a formula specificallydeveloped for such infants, whileanother group (n � 20) was fedexpressed breast milk (EBM).

Time to reach a weight of 1800 gwas 28 d for the formula-fedgroup, vs 40 d for thosereceiving EBM. The allocationgroups were not strictlyrandomized for severity ofillness.



phy,685 surfactant replacement therapy,686 preven-tion of respiratory syncytial virus infections,687 and(more recently) the successful WHO-modified ante-natal care package.84,688 In other instances, organiza-tion of basic neonatal care services in referral hospi-tals has been shown to be cost-effective.689

Cost-effectiveness data are also available from de-veloping-country communities for selected interven-tions focusing on antenatal interventions to preventinfectious complications of pregnancy and theirimpact on neonatal outcomes. These interventionsinclude antenatal screening and treatment for syph-ilis,324 syphilis prevention,690 toxoplasmosis treat-ment,691,692 TT-immunization administration,693 ma-laria-prevention programs,694 and screening andtreatment for asymptomatic bacteriuria.695 Theseevaluations reported a wide range of cost-benefitratios, with estimates ranging from $14 to $115 perneonatal death or adverse outcome averted(Table 43).

There has been much interest in recent years in thecost-effectiveness of community-based strategies forperinatal care. Such data, however, are almost exclu-sively available from developed countries and in-clude the institution of community-based nurse-mid-wifery services, culturally adapted perinatal care,696

and earlier discharge and community care for pre-mature infants.697,698 The strongest body of evidencefor the impact and cost-effectiveness of community-based interventions in developed countries pertainsto smoking-cessation699–701 and other LBW-preven-tion programs (for example, nutrition-related inter-ventions such as the WIC program),702–704 indicatingsubstantial benefits.

In contrast, there are almost no systematic studiesof the cost-effectiveness of community-based inter-ventions to improve perinatal and neonatal out-comes in developing countries, with the exception ofstudies in Nepal and rural India439,665 and severalevaluations of malaria prophylaxis and thera-py.283,694,705 Studies to evaluate even well-estab-lished strategies for the improvement of perinataland neonatal health outcomes, such as breastfeedingpromotion in large effectiveness trials, are only nowbeing commissioned.706

conclusions. Critical steps in the development ofeffective community-based health interventions in-

clude the demonstration of efficacy and effective-ness. To convince policy makers to support theprogrammatic implementation of promising inter-ventions, however, cost-effectiveness data areneeded to inform the feasibility of the intervention atscale and its expected benefits relative to other ser-vices. Even for interventions considered to be ofproven benefit for perinatal and/or neonatal health,little such data are available and are primarily fromfacility-based evaluations and, moreover, from de-veloped countries. Cost-effectiveness data for com-munity-based perinatal/neonatal health interven-tions in developing countries are almost nonexistent.However, with the recent development of guidelinesfor costing of maternal and newborn interventionsby the WHO709 and the Saving Newborn Lives Ini-tiative (Saving Newborn Lives, Saving NewbornLives Initiative: Project Costing Guidelines, Wash-ington, DC, Save the Children/USA, unpublisheddata) and with the establishment of Marginal Bud-geting for Bottlenecks and CHOICE activities to in-form public health resource allocation, more suchdata are expected.707,708 For example, several trials ofthe impact of antenatal, intrapartum, and postnatalinterventions on perinatal and neonatal health thatare currently underway have included evaluation ofcost-effectiveness as a key objective. Moreover, dataare beginning to emerge from Marginal Budgetingfor Bottlenecks on the marginal costs of introducingcertain interventions relevant to perinatal and neo-natal health to existing programs.

SUMMARY

Implications for ProgramsThis review of evidence from developing-country

community-based trials for the impact of antenatal,intrapartum, and postnatal interventions on perina-tal and neonatal health outcomes has highlighted thepaucity of available information, particularly fromRCTs. Only 31 studies were RCTs that reported pri-mary perinatal/neonatal health status outcomes (eg,stillbirth rate, PMR, NMR), and only 40 were RCTsthat reported secondary perinatal/neonatal healthoutcomes (eg, LBW, preterm births, morbidities, orbreastfeeding rates) (Table 1). Still fewer (only 10)were interventions conducted in health system set-tings or were effectiveness trials.

TABLE 43. Available Cost-effectiveness Data for Antenatal, Intrapartum, and Postnatal Interventions

Intervention Cost-Effectiveness Data Evidence of Cost-Effectiveness*

Malaria chemoprophylaxis or intermittenttherapy

2-dose regimen of SP cost $9.66 per case of LBWaverted705

3

Malaria prevention using PIBs $4–43 per discounted life-year lost740 3Syphilis screening and treatment $1.05 per neonatal discounted life-year gained324 4TT immunization $27–115 per neonatal death averted693 4Package of antenatal, intrapartum and

postnatal interventions1 y of care cost $5.30 per newborn and averted 1

neonatal death for every 18 neonates cared for4393

Participatory neonatal care women’sgroups

$3442 per neonatal death averted ($4397 includinghealth services strengthening); $111 per life-yearsaved665

3

* Rankings are based on a scale of 1–4 as follows: 1 indicates evidence of poor cost-effectiveness (leave out of programs); 2, uncertainevidence of cost-effectiveness (need additional data before including in programs); 3, some evidence of cost-effectiveness (may includein programs, but further evaluation is warranted); 4, strong evidence of cost-effectiveness (include in programs). Note that cost-effectiveness data from community-based trials was available only for selected interventions (see cost-effectiveness section).



To broaden the relevance of the conclusions thatcan be drawn from the available data, we placed theevidence in the context of biological plausibility, datafrom studies in developed countries, programmaticexperience, and recommendations by the WHO andother leading child health agencies. Recommenda-tions based on this review are broadly applicable todeveloping-country communities but are particu-larly germane to the most impoverished populationswith high NMRs (eg, �40 per 1000 live births).

It is clear that the level of evidence for benefit of anumber of interventions (Table 2) warrants theirbroad programmatic implementation (Fig 1). Inter-estingly, this group of evidence-based interventionsclosely resembles those identified through a strategicplanning process at the international and multiple-country levels and outlined in a conceptual frame-work for community-based maternal and newborncare recently advanced by Save the Children/USA.17

Moreover, these elements of essential newborn careare highlighted in recent recommendations for rou-tine and sick newborn care by the WHO.14–16 Thus,there seems to be broad convergence of expert opin-ion and the evidence base regarding priority inter-ventions to advance perinatal and neonatal healthand survival at the community level in developingcountries. Considering past experience of childhealth programs in implementation of various inter-ventions, and current recommendations of the WHOand leading child health agencies, a few additionalinterventions not covered in this review have beenadded to Fig 1 (marked with an asterisk). Theseinterventions include birth preparedness and recog-nition of and appropriate response to danger signs inthe antenatal and intrapartum periods; skilled healthcare at delivery (evidence reviewed elsewhere); earlypostnatal visitation for provision of anticipatoryguidance and recognition and management of ma-ternal and newborn illness; and birth spacing. Manyof these interventions have been included in compre-hensive packages of maternal and newborn interven-tions but have not been rigorously evaluated per sefor their specific contribution to the total impact ofthe package of care.

Effective interventions span maternal and neonatalcare, as anticipated when one considers that preg-nancy-related causes, delivery-related causes, andinfections each account for approximately one thirdof neonatal deaths.18 Moreover, although not empha-sized here, many of the interventions of proven ben-efit for neonates also lead to improved maternalhealth (detailed in Tables 4–42),45,710 which serves toillustrate the importance of integrating maternal andneonatal care while avoiding vertical programs foreither the mother or the newborn. Although data oncost-effectiveness are particularly lacking, an ap-proach that integrates maternal and neonatal healthinto Safe Motherhood and Child Survival programsand bridges the gap between these programs will notonly establish continuity of care across the life cyclebut will also enhance the cost-effectiveness of inter-vention packages.

Research GapsAlthough a number of interventions have been

shown to reduce perinatal and/or neonatal mortalityand to have the potential to reduce global neonatalmortality by approximately half through wide-scaleimplementation of evidence-based interventions,10

there are fundamental gaps in our knowledge of howto most effectively improve perinatal and neonataloutcomes in developing-country communities.711 Al-though we know that implementation of comprehen-sive neonatal care programs can reduce perinataland/or neonatal mortality substantially, failure toempower and mobilize communities to accept inter-ventions712 and lack of understanding of communitypractices and culture are major barriers. Thus, thereis an urgent need to adapt and evaluate culturallyand regionally appropriate packages of interventionsin a variety of settings. Pivotal questions regardingimplementation of neonatal health care programsthat demand additional operational research include:Which cadre of health workers in various settingscan most effectively deliver the needed services formothers and newborns at the community level? Howwill these workers be trained and supervised in asustainable manner at scale, and what are the mosteffective methods for preservice and in-service train-ing? What will be the scope of their service delivery(eg, with regard to client age, breadth of services, andgeographic reach)? Is a team of skilled birth atten-dants and newborn care providers needed at thecommunity level to provide simultaneous care forthe mother and newborn during the critical intrapar-tum period, particularly to address birth asphyxia?

The conceptual framework for newborn care at thecommunity level advanced by Save the Children/USA17 calls for provision of both preventive andcurative care, particularly for birth asphyxia and in-fections. However, in many settings, provision ofcurative care for these major causes of neonatal mor-tality is beyond the capacity of current health caresystems. Thus, critical unanswered questions are:Can effective implementation of a behavior-changecommunications package at the domiciliary level,without active case management of newborn illnessby health workers, improve neonatal outcomes?What is the added benefit and cost-effectiveness ofactive identification and management of neonatalillness, particularly serious bacterial infections andintrapartum hypoxia/birth asphyxia? What are themost feasible and effective ways to deliver life-sav-ing newborn resuscitation and antibiotic therapy inthe community? How can barriers to care seeking fornewborn illness be overcome most effectively so thathome-based care and care seeking can be effectivelylinked with referral-level care at facilities? What isthe impact and cost-effectiveness of postnatal visita-tion for promotion of healthful behaviors and recog-nition of neonatal illness? Can the same worker ad-dress the postnatal needs of both mothers andnewborns? What is the optimal timing and numberof routine visits with a health care provider?

Skilled care during delivery is universally recog-nized as a major long-term priority for improving the



care of mothers and newborns, and plans for advanc-ing health system capabilities for providing this careare paramount. Based on a consideration of the factthat most births and neonatal deaths in developingcountries still occur at home during the early neona-tal period, due to birth asphyxia and/or infectionsand primarily among LBW infants, the followingemerge as major research gaps:

1. Understanding and improving household andcommunity practices and their determinants: Lo-cal formative research is needed to better under-stand local beliefs and practices and the reasonsbehind them so that effective behavior-changestrategies can be developed and evaluated.23 Thismust be followed by appropriate research to de-velop intervention strategies to improve care-seeking behaviors at the household and commu-nity levels.

2. Improving the capacity of health systems for pro-viding essential preventive and special curativeneonatal health care: As noted above, some of themost challenging questions in neonatal health carerelate to how to deliver services most effectivelyto newborns in an integrated way within existingprograms for maternal and child health.20–22 Al-though difficult, answering these questions re-quires that many packages and combinations ofinterventions be tested through effectiveness trialsin health system settings.

3. Preventing and improving recognition and man-agement of birth asphyxia: Identification of sus-tainable interventions for management of intra-partum hypoxia/birth asphyxia is urgentlyneeded at the community level.24 Solutions mustallow for immediate response at the time of deliv-ery in a cost-effective manner and necessarily willrequire integration with skilled health care formothers at delivery and links with referral facili-ties.16

4. Preventing and improving recognition and man-agement of infections: There is an urgent need toidentify how the burden and severity of maternalinfections relate to perinatal outcomes. These in-fections may range from subclinical intrauterineinfection and bacterial vaginosis to overt genitaltract infections that may lead to preterm labor.The true burden of bacterial neonatal infections incommunity settings is also unclear, because manyclinical bacterial infections may represent viralinfections. Narrowing this information gap is vi-tal: to devise optimal antibiotic-treatment strate-gies for neonatal infections,25 we need to know theagents of life-threatening infections in the commu-nity and their antibiotic-susceptibility patterns.26

There is additional need for validated algorithmsfor accurate and rapid identification of infectedneonates by CHWs and caregivers. We also mustadvance antibiotic-treatment strategies for seriousinfections, which may include simplified antibiot-ic-delivery systems and/or regimens. The poten-tial development and evaluation of simplified oraltreatment regimens that include oral administra-

tion will be a major advance for public healthprograms.

5. Preventing and improving care for LBW infants:Given that the majority of newborns who die inmany developing countries are LBW, improvedstrategies for both prevention of and care for LBWinfants are urgently needed. These strategies in-clude interventions to reduce preterm births, re-duce the incidence of IUGR, or a combination ofboth. Prevention may be achieved by improvedmaternal nutrition and detection and treatment ofmaternal infections. Improved postnatal care ofLBW infants may be achieved in part by behavior-change communications, topical emollient ther-apy, breastfeeding promotion, and widespreadimplementation of culturally adapted methodsfor skin-to-skin care (or variations thereof) withthe mother and, when indicated, other householdmembers. The development, validation, and avail-ability of low-cost technology for the care of LBWinfants in primary and secondary health care fa-cilities are important adjuncts to community-based management strategies.

6. Improving information on the magnitude andcauses of neonatal mortality: Lack of accurateglobal, regional, and country-specific data on themagnitude and causes of perinatal and neonatalmorbidity and mortality currently is limiting ad-vocacy and program planning in newborn health.Strengthening of information systems, includingbirth and death registration and dissemination ofinformation at local levels about causes of new-born morbidity and mortality (and their determi-nants), are needed to guide resource allocationand program and research priorities. Moreover, asprograms incorporate newborn care, their impactmust be monitored and accurate data fed back tothose involved in health policy and program de-cision-making to enable them to use scarce re-sources more effectively. Integral to documentingand monitoring newborn health status is the needfor improved verbal autopsy instruments to en-able more accurate determination of causes ofperinatal and neonatal deaths in the communityand to assess the contribution of sociocultural andlogistic factors. Perinatal audit may also be a pow-erful tool for identifying avoidable factors indeaths, and mobilizing change in communities toimprove maternal and neonatal health care.

7. Cost-effectiveness analyses: Assessment of cost-effectiveness must be incorporated into neonatalhealth research to guide selection of interventionsand stimulate investment in neonatal health.

8. Development of indicators and simple manage-ment tools for assessing and monitoring healthsystem performance for perinatal and newborncare at the national level: An important impedi-ment to wider implementation of neonatal healthprogramming is lack of inclusion of perinatal andneonatal health indicators among global indica-tors for measuring progress in Child Survival (eg,Millennium Development goals). Moreover, pro-grams too often fail to monitor adequately anddemonstrate the effectiveness of their programs.



Tools for rapid situational analysis, prioritizationof program activities, and accurate monitoringand documentation of program effectiveness areurgently needed.

A major factor currently limiting our ability toidentify effective interventions is the wide variationin study designs and indicators for assessing impactand the almost complete absence of cost-effective-ness data. In 2001, a group of neonatal health re-searchers met to discuss a common agenda andmethodologies for neonatal health research in devel-oping-country communities.27 This review furtherhighlights the need, as recommended at the 2001meeting, for dialogue among researchers, policymakers, program managers, and donors regardingthe selection of research priorities; the use of com-mon (and, whenever possible, rigorous) study de-signs; and for sharing of data-collection instrumentsand research results. Reporting of indicators of neo-natal health at global and national levels is essentialfor monitoring large-scale success.

CONCLUSIONSA paucity of community-based data are available

from developing countries on health status impact,principally perinatal and/or neonatal mortality, formany interventions that are currently being consid-ered for inclusion in neonatal health programs. How-ever, a review of the evidence and consideration ofthe broader context of knowledge, experience, andrecommendations regarding these interventions en-abled us to categorize the interventions according tothe strength of the evidence base and confidencewith which the intervention could be included nowin programs. As a result, a package of priority inter-ventions for inclusion in programs was identified,and research priorities for advancing the state-of-the-art in neonatal health care were formulated. Thus,this review can serve as a guide for development ofevidence-based maternal and newborn health careprogramming at the community level and for selec-tion of research to advance the state-of-the-art incommunity-based neonatal care. It also may facilitatedialogue with policy makers about the importance ofinvesting in neonatal health.

Clearly there is ample evidence for benefit of anumber of antenatal, intrapartum, and postnatal in-terventions. Operational questions of how to imple-ment the intervention(s) in an affordable and accept-able manner at scale were an overriding concern formany interventions. Thus, although there is greatneed for continued research on the cost-effectivenessof a number of interventions, particularly for preven-tion and care of LBW infants, prevention of freshstillbirths, and prevention and management of birthasphyxia, it must not hamper implementation now ofmany interventions of known impact at a widerscale. For many of these proven interventions, how-ever, critical questions remain regarding how to im-plement them in a cost-effective manner at scale.Close monitoring of the impact of these programs isimperative. Finally, close communication betweenprogram managers who document experience with

intervention implementation; the researchers whogenerate answers to operational questions and de-vise new, innovative interventions and approaches;and the donors who promote the research will becritical to advancing maternal and neonatal healthcare at the community level.

ACKNOWLEDGMENTSThis review was supported by the Saving Newborn Lives Ini-

tiative of Save the Children/USA through a grant from the Bill &Melinda Gates Foundation and the Department of Child andAdolescent Health and Development, World Health Organization.

We thank Dr Jose Martines of the World Health Organizationand Anne Tinker of Save the Children/USA for technical inputand encouragement during the review process; Chitra Krishnanfor assistance in locating relevant articles for review; Fauzia AmanMalik for technical and editorial assistance; and Sarah Hollandand Charlotte Storti for editorial expertise. We also thank thefollowing individuals who reviewed the manuscript and providedtechnical and editorial review on behalf of the World HealthOrganization: Dr Martin Weber, Dr Jelka Zupan, and Dr Berna-dette Daelmans (World Health Organization, Geneva, Switzer-land); Professor Robert Black (Department of International Health,Johns Hopkins University Bloomberg School of Public Health,Baltimore, MD); Dr Douglas McMillan (University of Calgary/Calgary Health Region, Calgary, Alberta, Canada); and Dr StevenWall and Dr Vinod Paul (Save the Children/USA, Washington,DC).

REFERENCES1. Koblinsky M, Campbell O, Heichelheim J. Organizing delivery care:

what works for safe motherhood? Bull World Health Organ. 1999;77:399–406

2. Graham W, Bell J, Bullough C. Can skilled attendance at deliveryreduce maternal mortality in developing countries? In: Studies inHealth Services, Organisation, and Policy. Antwerp, Belgium: PrinceLeopold Institute of Tropical Medicine; 2001:97–129

3. Sibley L, Sipe T. Traditional Birth Attendant Training Effectiveness: AMeta-analysis. Final Technical Report. Washington, DC: Academy forEducational Development/Sara Project; 2002

4. Bale J, Stoll B, Lucas A, eds. Improving Birth Outcomes: Meeting theChallenge in the Developing World. Committee on Improving Birth Out-comes, Board on Global Health. Washington, DC: The National Acad-emies Press; 2003

5. Concato J, Horwitz R. Beyond randomised versus observational stud-ies. Lancet. 2004;363:1660–1661

6. Horwitz R. Complexity and contradiction in clinical trial research.Am J Med. 1987;82:498–510

7. Rabeneck L, Viscoli C, Horwitz R. Problems in the conduct and anal-ysis of randomized controlled trials: are we getting the right answersto the wrong questions? Arch Intern Med. 1992;152:507–512

8. Victora C, Habicht J-P, Bryce J, Black R. Beyond randomized controlledtrials (Cochrane Review). Oxford, United Kingdom: Update Software;2004

9. International Nutritional Anemia Consultative Group. Efficacy andeffectiveness interventions to control iron deficiency and iron defi-ciency anemia. Washington, DC: ILSI Human Nutrition Institute; 2004

10. Jones G, Steketee RW, Black RE, Bhutta ZA, Morris SS; Bellagio ChildSurvival Study Group. How many child deaths can we prevent thisyear? Lancet. 2003;362:65–71

11. Kramer M. Intrauterine growth and gestational duration determinants.Pediatrics. 1987;80:502–522

12. Zhu B-P, Rolfs R, Nangle B, Horan J. Effect of the interval betweenpregnancies on perinatal outcomes. N Engl J Med. 1999;340:589–594

13. Rutstein S. Birth spacing and newborn health. In: American PublicHealth Association, 130th Annual Meeting and Exposition, Program andAbstracts. 2002; Washington, DC: American Public Health Association;2002

14. World Health Organization. Mother-Baby Package: Implementing SafeMotherhood in Countries: A Practical Guide. Maternal Health and SafeMotherhood Programme. Geneva, Switzerland: World HealthOrganization; 1994

15. World Health Organization. Pregnancy, Childbirth, Postpartum and New-born Care: A Guide for Essential Practice. Geneva, Switzerland: WorldHealth Organization; 2003



16. Integrated Management of Pregnancy and Childbirth. Managing Com-plications in Pregnancy and Childbirth: A Guide for Midwives and Doctors.Geneva, Switzerland: World Health Organization; 2000

17. Marsh D, Darmstadt G, Moore J, Daly P, Oot D, Tinker A. Advancingnewborn health and survival in developing countries: a conceptualframework. J Perinatol. 2002;22:572–576

18. Saving Newborn Lives. State of the World’s Newborns. Washington, DC:Save the Children/USA; 2001

19. Bhutta Z, Darmstadt G, Ransom E. Using Evidence to Save NewbornLives. Policy brief. Washington, DC: Population Reference Bureau;2003

20. Darmstadt G, Lawn J, Costello A. Advancing the state of the world’snewborns. Bull World Health Organ. 2003;81:224–225

21. Bhutta Z, Ali N, Hyder A, Wajid A. Perinatal and newborn care inPakistan: seeing the unseen. In: Bhutta Z, ed. Maternal and Child Healthin Pakistan: Challenges and Opportunities. Karachi, Pakistan: OxfordUniversity Press; 2004:19–46

22. Bhutta Z, De Silva H, Awasthi S, et al. Maternal and child health: isSouth Asia ready for change? BMJ. 2004;328:816–819

23. Parlato R, Darmstadt G, Tinker A. Planning and Using QualitativeResearch to Improve Newborn Care Practices: A Guide for Program Manag-ers. Washington, DC: Saving Newborn Lives, Save the Children/USA;2004

24. Lawn J, Darmstadt G. To breathe or not to breathe: research andprogram priorities to improve management of birth asphyxia at com-munity level in developing countries. J Perinatol. 2005; In press

25. Zaidi A, Ali S, Darmstadt G, Bhutta Z. Serious bacterial infectionsamong neonates and young infants in developing countries: evalua-tion of etiology, antimicrobial resistance and therapeutic managementstrategies in community settings. Pediatr Infect Dis J. 2005; In press

26. Darmstadt G, Black R, Santosham M. Research priorities and postpar-tum-care strategies for the prevention and treatment of neonatal infec-tions in less developed countries. Pediatr Infect Dis J. 2000;19:739–750

27. Coco G, Darmstadt G, Kelley L, Martines J, Paul V, eds. Perinatal andneonatal health interventions research. Report of a meeting, April29–May 3, 2001, Kathmandu, Nepal. J Perinatol. 2002;22(suppl 2):S1–S41

28. Hyder A, Morrow R, Wali S, McGuckin J. Burden of Disease for NeonatalMortality in South Asia and Sub-Saharan Africa. Washington, DC: Savethe Children Federation–USA; 2001

29. World Health Organization. Estimates. In: State of the World’s New-borns. Washington, DC: Saving Newborn Lives, Save the Children/USA; 2001:1–49

30. Stoll B. The global impact of neonatal infection. Clin Perinatol. 1997;24:1–21

31. Hill K. Approaches to the Measurement of Childhood Mortality: A Compar-ative Review. Baltimore, MD: Johns Hopkins University School of Hy-giene and Public Health; 1992

32. Child Health Research Project. Reducing Perinatal and Neonatal Mortal-ity. Baltimore, MD: Johns Hopkins School of Public Health; 1999

33. Kramer M, Liu S, Luo Z, et al. Analysis of perinatal mortality and itscomponents: time for a change? Am J Epidemiol. 2002;156:493–497

34. Mullany L, Darmstadt G, Tielsch J. Role of antimicrobial applicationsto the umbilical cord in neonates to prevent bacterial colonization andinfection: a review of the evidence. Pediatr Infect Dis J. 2003;22:996–1002

35. Moss W, Darmstadt G, Marsh D, Black R, Santosham M. Researchpriorities for the reduction of perinatal and neonatal morbidity andmortality in developing country communities. J Perinatol. 2002;22:484–495

36. Donnay F. Maternal survival in developing countries: what has beendone, what can be achieved in the next decade. Int J Gynaecol Obstet.2000;70:89–97

37. Frey-Hobcraft J, McDonald J, Rutstein S. Socio-economic factors ininfant and child mortality: a cross-national comparison. Popul Stud(Camb). 1984;38:192–223

38. Bicego G, Boerma J. Maternal education and child survival: a compar-ative study of survey data from 17 countries. Soc Sci Med. 1993;36:1207–1227

39. Victora C, Huttly S, Barros F, Lombardi C, Vaughan J. Maternaleducation in relation to early and late child health outcomes: findingsfrom a Brazilian cohort study. Soc Sci Med. 1992;34:899–905

40. Van Ginneken J, Lob-Levyt J, Gove S. Potential interventions for pre-venting pneumonia among young children in developing countries:promoting maternal education. Trop Med Int Health. 1996;1:283–294

41. Harrison KA. Child-bearing, health and social priorities: a survey of22 774 consecutive hospital births in Zaria, Northern Nigeria. Br JObstet Gynaecol. 1985;92(suppl 5):1–119

42. Harrison K. The importance of the educated healthy woman in Africa.Lancet. 1997;349:644–647

43. Jewkes R, Wood K. Competing discourses of vital registration andpersonhood: perspectives from rural South Africa. Soc Sci Med. 1998;46:1043–1056

44. Ross S. Promoting Quality Maternal and Newborn Care: A Reference Man-ual for Program Managers. Washington, DC: CARE; 1999

45. Bhutta Z, Darmstadt G, Ransom E, Starrs A, Tinker A. Basing newbornand maternal health policies on evidence. In: Shaping Policy for Mater-nal and Newborn Health. Baltimore, MD: JHPIEGO; 2003:5–12

46. World Health Organization. Perinatal Mortality–A Listing of AvailableInformation. Maternal Health and Safe Motherhood Programme. Ge-neva, Switzerland: World Health Organization; 1996

47. Caldwell J, McDonald P. Influence of maternal education on infant andchild mortality: levels and causes. Health Policy Educ. 1982;2:251–267

48. Ware H. Effects of maternal education, women’s roles and child careon child mortality. In: Mosley W, Chen L, eds. Child Survival: Strategiesfor Research. Cambridge, United Kingdom: Cambridge UP; 1984:191–214

49. Lindenbaum S, Chakraborty M, Elisa M. The Influence of MaternalEducation on Infant and Child Mortality in Bangladesh. Dhaka,Bangladesh: Eastern Commercial Service; 1985

50. Cleland J. Maternal education and child survival: further evidence andexplanations. In: Caldwell J, Findley S, Caldwell P, et al, eds. What WeKnow About Health Transition: The Cultural, Social and Behavioural De-terminants of Health. Canberra, Australia: Health Transition Center,Australian National University; 1990:400–419

51. Elo I. Utilisation of maternal health-care services in Peru: the role ofwomen’s education. Health Transit Rev. 1992;2:49–69

52. Office of Population Research. World fertility surveys. Princeton, NJ:Office of Population Research, Princeton University; 1973–1977. Avail-able at: http://opr.princeton.edu/archive/wfs. Accessed December28, 2004

53. Hobcraft J. Women’s education, child welfare and child survival: areview of the evidence. Health Transit Rev. 1993;3:159–175

54. Cleland J, van Ginneken J. Maternal education and child survival indeveloping countries: the search for pathways of influence. Soc SciMed. 1988;27:1357–1368

55. Bicego G, Boerma I. Maternal education and child survival: a compar-ative analysis of DHS data. In: Demographic and Health Surveys WorldConference Proceedings. Vol 1. Columbia, MD: IRD/Macro Interna-tional; 1991:177–204

56. Bender D, McCann M. The influence of maternal intergenerationaleducation on health behaviors of women in peri-urban Bolivia. Soc SciMed. 2000;50:1189–1196

57. Kramer M. Nutritional advice in pregnancy (Cochrane Review). Ox-ford, United Kingdom: Update Software; 2002a

58. Kafatos A, Vlachonikolis I, Codrington C. Nutrition during pregnancy:the effects of an educational intervention program in Greece. Am J ClinNutr. 1989;50:970–979

59. Lutter C, Perez-Escamilla R, Segall A, Sanghvi T, Teruya K, WickhamC. The effectiveness of a hospital based program to promote exclusivebreastfeeding among low income women in Brazil. Am J Public Health.1997;87:659–663

60. Leon-Cava N, Lutter C, Ross J, Martin L. Quantifying the Benefits ofBreastfeeding: A Summary of the Evidence. Washington, DC: Pan Amer-ican Health Organization and LINKAGES; 2002

61. Morrow A, Guerrero M, Shults J, et al. Efficacy of home-based peercounselling to promote exclusive breastfeeding: a randomised con-trolled trial. Lancet. 1999;353:1226–1231

62. Jakobsen M, Sodemann M, Molbak K, Alvarenga I, Aaby P. Promotingbreastfeeding through health education at the time of immunizations:a randomized trial from Guinea Bissau. Acta Paediatr. 1999;88:741–747

63. Siegel E, Gillings D, Campbell S, Guild P. Controlled evaluation ofrural regional perinatal care: developmental and neurologic outcomesat 1 year. Pediatrics. 1986;77:187–195

64. Harris K, Wilson C, Sheppard Brown N, Keys L, Wenz C, Mendler V.A perinatal education consortium: improved resource utilization. JObstet Gynecol Neonatal Nurs. 1999;28:486–492

65. Frank J, Rhodes T, Edwards W, et al. The New Hampshire PerinatalProgram: twenty years of perinatal outreach education. J Perinatol.1999;19:3–8

66. Woods D, Theron G. The perinatal education program. S Afr Med J.1994;84:61

67. Woods D, Theron G. The impact of the perinatal education programmeon cognitive knowledge in midwives. S Afr Med J. 1995;85:150–153

68. Bolam A, Manandhar D, Shrestha P, Ellis M, Costello A. The effects ofpostnatal health education for mothers on infant care and family



planning practices in Nepal: a randomized controlled trial. BMJ. 1998;316:805–811

69. Osrin D, Mesko N, Shrestha B, et al. Implementing a community-basedparticipatory intervention to improve essential newborn care in ruralNepal. Trans R Soc Trop Med Hyg. 2003;97:18–21

70. Osrin D, Tumbahangphe K, Shrestha D, et al. Cross sectional, commu-nity based study of care of newborn infants in Nepal. BMJ. 2002;325:1063

71. Standing T, el-Sabagh N, Brooten D. Maternal education during theperinatal period. Clin Perinatol. 1998;25:389–402

72. Bergsjo P, Villar J. Scientific basis for the content of routine antenatalcare. II. Power to eliminate or alleviate adverse newborn outcomes;some special conditions and examinations. Acta Obstet Gynecol Scand.1997;76:15–25

73. Hill Z, Kirkwood B, Edmond K. Family and Community Practices ThatPromote Child Survival, Growth and Development: A Review of the Evi-dence. Geneva, Switzerland: World Health Organization; 2004

74. Villar J, Bergsjo P. Scientific basis for the content of routine antenatalcare. I. Philosophy, recent studies, and power to eliminate or alleviateadverse maternal outcomes. Acta Obstet Gynecol Scand. 1997;76:1–14

75. Carroli G, Villar J, Piaggio G, et al. WHO systematic review of ran-domised controlled trials of routine antenatal care. Lancet. 2001;357:1565–1570

76. Greenwood A, Greenwood B, Bradley A, et al. A prospective survey ofthe outcome of pregnancy in a rural area of the Gambia. Bull WorldHealth Organ. 1987;65:635–643

77. Maine D, Rosenfield A. The Safe Motherhood Initiative: why has itstalled? Am J Public Health. 1999;89:480–482

78. Kapoor S, Anand K, Kumar G. Risk factors for stillbirths in a secondarylevel hospital at Ballabgarh, Haryana: a case control study. IndianJ Pediatr. 1994;61:161–166

79. Amoa A, Kluffo C, Moro M, Kariwiga G, Mola G. A case-control studyof stillbirths at the Port Moresby General Hospital. P N G Med J.1998;4:126–136

80. Conde-Agudelo A, Belizan J, Diaz-Rossello J. Epidemiology of fetaldeath in Latin America. Acta Obstet Gynecol Scand. 2000;79:371–378

81. McDonagh M. Is antenatal care effective in reducing maternal mor-bidity and mortality? Health Policy Plan. 1996;11:1–15

82. McDuffie RJ, Beck A, Bischoff K, Cross J, Orleans M. Effect of fre-quency of prenatal care visits on perinatal outcome among low-riskwomen. A randomized controlled trial. JAMA. 1996;275:847–851

83. Villar J, Bakketeig L, Donner A, et al. The WHO antenatal care ran-domized controlled trial: rationale and study design. Paediatr PerinatEpidemiol. 1998;12(suppl):27–58

84. Villar J, Ba’aqeel H, Piaggio G, et al. WHO antenatal care randomisedtrial for the evaluation of a new model of routine antenatal care[published correction appears in Lancet. 2001;358:1556]. Lancet. 2001;357:1551–1564.

85. Langer A, Villar J, Romero M, et al. Are women and providers satisfiedwith antenatal care? Views on a standard and a simplified, evidence-based model of care in four developing countries. BMC Womens Health.2002;2:7

86. de Onis M, Villar J, Gulmezoglu M. Nutritional interventions to pre-vent intrauterine growth retardation: evidence from randomised con-trolled trials. Eur J Clin Nutr. 1998;52(suppl 1):S83–S93

87. Jackson A, Bhutta Z, Lumbiganon P. Nutrition as a preventive strategyagainst adverse pregnancy outcomes. J Nutr. 2003;133:1589S–1591S

88. Sachdev H. Effect of supplementation with vitamin A or beta caroteneon mortality related to pregnancy. No magic pills exist for reducingmortality related to pregnancy [letter]. BMJ. 1999;319:1202

89. United Nations Administrative Committee on Coordination, Sub-Committee on Nutrition. Low Birthweight: Report of a Meeting in Dhaka,Bangladesh, on 14 –17 June 1999. Pojda J, Kelly L, eds. Geneva,Switzerland: ACC/SCN. Nutrition Policy Paper No. 18.

90. Bhargava S, Sachdev H, Fall C, et al. Relation of serial changes inchildhood body-mass index to impaired glucose tolerance in youngadulthood. N Engl J Med. 2004;350:865–875

91. Villar J, Merialdi M, Gulmezoglu A, et al. Nutritional interventionsduring pregnancy for the prevention or treatment of maternal mor-bidity and preterm delivery: an overview of randomized controlledtrials. J Nutr. 2003;133(5 suppl 2):1606S–1625S

92. Merialdi M, Carroli G, Villar J, et al. Nutritional interventions duringpregnancy for the prevention or treatment of impaired fetal growth: anoverview of randomized controlled trials. J Nutr. 2003;133(5 suppl2):1626S–1631S

93. De Brouwere V, Van Lerberghe W. Safe Motherhood Strategies: A Reviewof the Evidence. Antwerp, Belgium: ITG Press; 2001

94. Allen L, Gillespie S. Preventing and Treating Vitamin A Deficiency. WhatWorks? A Review of the Efficacy and Effectiveness of Nutrition Interventions.Geneva, Switzerland: ACC/SCN; 2001. ACC/SCN Nutrition PolicyPaper No. 19, ADB Nutrition and Development Series No. 5

95. Kramer M. Balanced protein/energy supplementation in pregnancy(Cochrane Review 1999). Oxford, United Kingdom: Update Software;2001b

96. Ross R, Perlzweig W, Taylor H, McBryde A, Yates A, Kondritzer A. Astudy of certain dietary factors of possible etiologic significance intoxemias of pregnancy. Am J Obstet Gynecol. 1938;35:426–440

97. Elwood P, Haley T, Hughes S, Seetnam P, Gray O, Davies D. Childgrowth (0–5 years), and the effect of entitlement to a milk supplement.Arch Dis Child. 1981;56:831–835

98. Viegas O, Scott P, Cole T, Eaton P, Needham P, Wharton B. Dietaryprotein energy supplementation of pregnant Asian mothers at Sor-rento, Birmingham. II: Selective during third trimester only. BMJ.1982;285:592–595

99. Viegas O, Scott P, Cole T, Mansfield H, Wharton P, Wharton B. Dietaryprotein energy supplementation of pregnant Asian mothers at Sor-rento, Birmingham. I: Unselective during second and third trimesters.BMJ. 1982;285:589–592

100. Mardones-Santander F, Rosso P, Stekel A, et al. Effect of a milk-basedfood supplement on maternal nutritional status and fetal growth inunderweight Chilean women. Am J Clin Nutr. 1988;47:413–419

101. Rush D, Stein D, Susser M. A randomized controlled trial of prenatalnutrition supplementation in New York City. Pediatrics. 1980;65:683–697

102. Iyengar L. Effects of dietary supplements late in pregnancy on theexpectant mother and her newborn. Indian J Med Res. 1967;55:85–89

103. Rush D. Effects of changes in protein and calorie intake during preg-nancy on the growth of the human fetus. In: Chalmers I, Enkin M,Keirse M, eds. Effective Care in Pregnancy and Childbirth. Oxford, UnitedKingdom: Oxford UP; 1989:92–101

104. Ceesay S, Prentice A, Cole T, et al. Effects on birth weight and perinatalmortality of maternal dietary supplements in rural Gambia: 5 yearrandomised controlled trial. BMJ. 1997;315:786–790

105. Mora J, Clement J, Christiansen N, Suescun J, Wagner M, Herrera M.Nutritional supplementation and the outcome of pregnancy. III. Peri-natal and neonatal mortality. Nutr Rep Int. 1978;18:167–175

106. Campbell-Brown M. Protein energy supplements in primigravidwomen at risk of low birth weight. In: Campbell D, Gillmer M, eds.Nutrition in Pregnancy. Proceedings of the 10th Study Group of the RoyalCollege of Obstetrics and Gynaecology, London, 1982; London, UnitedKingdom: Royal College of Obstetrics and Gynaecology; 1983:85–98

107. Adair L, Pollitt E. Outcome of maternal nutritional supplementation: acomprehensive review of the Bacon Chow study. Am J Clin Nutr.1985;41:948–978

108. McDonald E, Pollitt E, Mueller W, Hsueh A, Sherwin R. The BaconChow study: maternal nutritional supplementation and birth weightof offspring. Am J Clin Nutr. 1981;34:2133–2144

109. Kusin J, Kardjati S, Houtkooper J, Renqvist U. Energy supplementationduring pregnancy and postnatal growth. Lancet. 1992;340:623–626

110. Tontisirin K, Booranasubkajorn U, Hongsumarn A, Thewtong D. For-mulation and evaluation of supplementary foods for Thai pregnantwomen. Am J Clin Nutr. 1986;43:931–939

111. Villar J, Rivera J. Nutritional supplementation during two consecutivepregnancies and the interim lactation period: effect on birth weight.Pediatrics. 1988;81:51–57

112. Prentice A, Whitehead R, Watkinson M, Lamb W, Cole T. Prenataldietary supplementation of African women and birth weight. Lancet.1983;8323(1):489–492

113. Prentice A, Cole T, Foord F, Lamb W, Whitehead R. Increased birthweight after prenatal dietary supplementation of rural African women.Am J Clin Nutr. 1987;46:912–925

114. Qureshi S, Rao N, Madhavi V, Mathur Y, Reddi Y. Effect of maternalnutrition supplementation on the birth weight of the newborn. IndianPediatr. 1973;10:541–544

115. Lechtig A, Habicht J, Delgado H, Klein R, Yarbrough C, Martorell R.Effect of food supplementation during pregnancy on birthweight.Pediatrics. 1975;56:508–520

116. Kardjati S, Kusin J, DeWith C. Energy supplementation in the lasttrimester of pregnancy in East Java. I. Effect on birthweight. Br J ObstetGynaecol. 1988;95:783–794

117. Girija A, Geervani P, Rao N. Influence of dietary supplementationduring pregnancy on lactation performance. J Trop Pediatr. 1984;30:79–83



118. Kielmann A, Taylor C, DeSweemer C, et al. The Narangwal experi-ment on interactions of nutrition and infections: II. Morbidity andmortality effects. Indian J Med Res. 1978;68(suppl):21–41

119. World Health Organization, Maternal Health and Safe MotherhoodProgramme, Nutrition Programme. The Prevalence of Anemia in Women.2nd ed. Geneva, Switzerland: World Health Organization; 1992

120. Rush D. Nutrition and maternal mortality in the developing world.Am J Clin Nutr. 2000;72(suppl):212S–240S

121. Guidotti R. Anaemia in pregnancy in developing countries. BJOG.2000;107:437–438

122. AbouZahr C, Royston E. Maternal Mortality: A Global Factbook. Geneva,Switzerland: World Health Organization; 1991

123. Allen L. Anemia and iron deficiency: effects on pregnancy outcome.Am J Clin Nutr. 2000;71(suppl 5):S1280–S1284

124. Allen L. Pregnancy and iron deficiency: unresolved issues. Nutr Rev.1997;55:91–101

125. World Health Organization/Food and Agriculture Organization, In-ternational Conference on Nutrition. Nutrition and Development: AGlobal Assessment. Rome, Italy: World Health Organization/Food andAgriculture Organization; 1992

126. Stolzfus R, Dreyfuss M. Guidelines for the Use of Iron Supplements toPrevent and Treat Iron Deficiency Anemia. Washington, DC: ILSI Press;1998

127. Kolsteren P, de Souza S. Micronutrients and pregnancy outcome. In:De Brouwere V, Van Lerbeghe W, eds. Safe Motherhood Strategies: AReview of the Evidence. Antwerp, Belgium: ITG Press; 2001:55–76

128. Gallego EB. Severe Anemia in Pregnancy. Report of a workshop held atthe Institute of Child and Mother Health in Dhaka, Bangladesh. Gal-lego EB, ed. Micronutrient Initiative 2000. Ottawa, ON, Canada: Inter-national Development Research Centre; 2000

129. Garn S, Ridela S, Petzoid A, Falkner F. Maternal hematologic levelsand pregnancy outcomes. Semin Perinatol. 1981;5:155–162

130. Murphy J, O’Riordan J, Newcombe R, Coles E, Pearson J. Relation ofhaemoglobin levels in first and second trimesters to outcome of preg-nancy. Lancet. 1986;1:992–994

131. Leiberman E, Ryan K, Monson R, Schoenbaum S. Association of ma-ternal hematocrit with premature labor. Am J Obstet Gynecol. 1988;159:107–114

132. Knottnerus J, Delgado L, Knipschild P, Essed G, Smits F. Haematologicparameters and pregnancy outcome: a prospective cohort study in thethird trimester. J Clin Epidemiol. 1990;43:461–466

133. Blankson M, Goldenberg R, Cutter G, Cliver S. The relationship be-tween maternal hematocrit and pregnancy outcome: black and whitedifferences. J Natl Med Assoc. 1993;85:130–134

134. Llewelyn-Jones D. Severe anaemia in pregnancy (as seen in Kuala-Lumpur, Malaysia). Aust N Z J Obstet Gynaecol. 1965;5:191–197

135. Steer P. Maternal hemoglobin concentration and birth weight. Am JClin Nutr. 2000;71(suppl):1285S–1287S

136. Rasmussen K. Is there a causal relationship between iron deficiency oriron deficiency anemia and weight at birth, length of gestation andperinatal mortality? J Nutr. 2001;131(suppl):590S–603S

137. Chi I, Agoestina T, Harbin J. Maternal mortality at twelve teachinghospitals in Indonesia—an epidemiological analysis. Int J GynaecolObstet. 1981;19:259–266

138. Mahomed K. Iron supplementation in pregnancy (Cochrane Review).Oxford, United Kingdom: Update Software; 2000a

139. Hemminki E, Rimpela U. A randomized comparison of routine versusselective iron supplementation during pregnancy. J Am Coll Nutr.1991;10:3–10

140. Fleming A, Ghatoura G, Harrison K, Briggs N, Dunn D. The preven-tion of anaemia in pregnancy in primigravidae in the guinea savannaof Nigeria. Ann Trop Med Parasitol. 1986;80:211–233

141. Agarwal K, Agarwal D, Mishra KP. Impact of anaemia prophylaxis inpregnancy on maternal haemoglobin, serum ferritin & birth weight.Indian J Med Res. 1991;94:277–280

142. Fullerton W, Turner A. Exchange transfusion in treatment of severeanaemia in pregnancy. Lancet. 1962;282:75–78

143. Brabin B, Hakimi M, Pelletier D. An analysis of anemia and pregnancy-related maternal mortality. J Nutr. 2001;131:604S–614S

144. Scholl T, Hediger M. Anemia and iron deficiency anemia: compilationof data on pregnancy outcome. Am J Clin Nutr. 1994;59(suppl):492S–510S

145. Ramakrishnan U, Manjrekar R, Rivera J, Gonzales-Cosio T, MartorellR. Micronutrients and pregnancy outcome: a review of the literature.Nutr Res. 1999;19:103–159

146. Klebanoff M, Shiono P, Shelby J, Trachtenberg A, Graubard B. Anemiaand spontaneous preterm birth. Am J Obstet Gynecol. 1991;164:59–63

147. Lu Z, Goldenberg R, Cliver S, Cutter G, Blankson M. The relationshipbetween maternal hematocrit and pregnancy outcome. Obstet Gynecol.1991;77:190–194

148. Scholl T, Hediger M, Fischer R, Shearer J. Anemia versus irondeficiency: increased risk of preterm delivery in a prospective study.Am J Clin Nutr. 1992;55:985–988

149. Zhou L, Yang W, Hua J, Deng C, Tao X, Stoltzfus R. Relation ofhemoglobin measured at different times in pregnancy to preterm birthand low birth weight in Shanghai, China. Am J Epidemiol. 1998;148:998–1006

150. Cuervo L, Mahomed K. Treatments for iron deficiency anaemia inpregnancy (Cochrane Review). Oxford, United Kingdom: UpdateSoftware; 2001

151. Mahomed K. Folate supplementation in pregnancy (Cochrane Re-view). Oxford, United Kingdom: Update Software; 2001a

152. Mahomed K. Iron and folate supplementation in pregnancy (CochraneReview). Oxford, United Kingdom: Update Software; 2001b

153. Sood S, Ramachandran K, Mathur M, et al. W. H. O. sponsoredcollaborative studies on nutritional anaemia in India. 1. The effects ofsupplemental oral iron administration to pregnant women. Q J Med.1975;44:241–258

154. Menendez C, Todd J, Alonso P, et al. The effect of iron supplementa-tion during pregnancy, given by traditional birth attendants, on theprevalence of anemia and malaria. Trans R Soc Trop Med Hyg. 1994;88:590–593

155. Menendez C, Todd J, Alonso P, Lulat S, Francis N, Greenwood B.Malaria chemoprophylaxis, infection of the placenta and birth weightin Gambian primigravidae. J Trop Med Hyg. 1994;97:244–248

156. Preziosi P, Prual A, Galan P, Daouda H, Boureima H. Hercberg Effectof iron supplementation on the iron status of pregnant women: con-sequences for newborns. Am J Clin Nutr. 1997;66:1178–1182

157. Thane-Toe, Thein-Than. The effect of oral iron supplementation onferritin levels in pregnant Burmese women. Am J Clin Nutr. 1982;35:95–99

158. Afifi A. Plexafer-F in the treatment of latent iron deficiency in preg-nancy. J Int Med Res. 1978;6:34–40

159. Reddaiah V, Raj P, Ramachandran K, et al. Supplementary iron dose inpregnancy anaemia prophylaxis. Indian J Pediatr. 1989;56:109–114

160. Ridwan E, Schultink W, Dillon D, Gross R. Effects of weekly ironsupplementation on pregnant Indonesian women are similar to thoseof daily supplementation. Am J Clin Nutr. 1996;63:884–890

161. Tan C, Ng K. The effect of oral iron on the haemoglobin concentrationduring the second half of pregnancy. In: 27th British Congress of Ob-stetrics and Gynaecology; 1995 July 4–7. Dublin, Ireland: Royal College ofObstetricians and Gynaecologists; 1995:A101

162. Dommisse J, Bell D, Du Toit E, Midgley V, Cohen M. Iron-storagedeficiency and iron supplementation in pregnancy. S Afr Med J. 1983;64:1047–1051

163. Ekstrom E, Kavishe F, Habicht J, Frongillo E, Rasmussen K, Hemed L.Adherence to iron supplementation during pregnancy in Tanzania:determinants and hematologic consequences. Am J Clin Nutr. 1996;64:368–374

164. Sloan N, Jordan E, Winikoff B. Does Iron Supplementation Make a Dif-ference? Mothercare Working Paper No. 15. Arlington, VA: John Snow,Inc; 1992

165. Cook J, Reddy M. Efficacy of weekly compared with daily iron sup-plementation. Am J Clin Nutr. 1996;62:117–120

166. Yip R. Iron supplementation during pregnancy: is it effective? Am JClin Nutr. 1996;63:853–855

167. Atukorala S, de Silva L, Dechering W, Dassanaeike T, Perera R. Eval-uation of effectiveness of iron-folate supplementation and antihelmin-thic therapy against anemia in pregnancy—a study in the plantationsector of Sri Lanka. Am J Clin Nutr. 1994;60:286–292

168. Suharno D, West KJ, Muhilal C, Karyadi D, Hautvast J. Supplementa-tion with vitamin A and iron for nutritional anaemia in pregnantwomen in West Java, Indonesia. Lancet. 1993;342:1325–1328

169. Christian P, Shrestha I, LeClerq C, et al. Supplementation with micro-nutrients in addition to iron and folic acid does not further improve thehematologic status of pregnant women in rural Nepal. J Nutr. 2003;133:3492–3498

170. Byles A, D’Sa A. Reduction of reaction to iron-dextran infusion usingchloroquine. Br Med J. 1970;3(723):625–627

171. Oppenheimer S, Macfarlane S, Moody J, Harrison C. Total dose ironinfusion, malaria and pregnancy in Papua New Guinea. Trans R SocTrop Med Hyg. 1986;80:818–822

172. World Health Organization/United Nations Children’s Fund, Inte-grated Management of the Sick Child Initiative. Integrated Management



of Childhood Illness Guidelines. Geneva, Switzerland: World HealthOrganization; 2001

173. Baker H, Thind I, Frank O, DeAngelis B, Caterini H, Louria D. Vitaminlevels in low-birth-weight newborn infants and their mothers. Am JObstet Gynecol. 1977;129:521–524

174. Scholl T, Hediger M, Schall J, Khoo C, Fischer R. Dietary and serumfolate: their influence on the outcome of pregnancy. Am J Clin Nutr.1996;63:520–525

175. Tamura T, Goldenberg R, Johnston K, Cliver S, Hoffman H. Serumconcentrations of zinc, folate, vitamins A and E, and proteins, and theirrelationships to pregnancy outcome. Acta Obstet Gynecol Scand Suppl.1997;165:63–70

176. Rondo P, Abbott R, Rodrigues L, Tomkins A. Vitamin A, folate, andiron concentrations in cord and maternal blood of intra-uterine growthretarded and appropriate birth weight babies. Eur J Clin Nutr. 1995;49:391–399

177. Rondo P, Tomkins A. Folate and intrauterine growth retardation. AnnTrop Paediatr. 2000;20:253–258

178. Rao S, Yajnik C, Kanade A, et al. Maternal fat intakes and micronutri-ent status are related to fetal size at birth in rural India; the PuneMaternal Nutrition Study. J Nutr. 2001;131:1217–1224

179. Iyengar L, Apte S. Prophylaxis of anemia in pregnancy. Am J Clin Nutr.1970;23:725–730

180. Baumslag N, Edelstein T, Metz J. Reduction of incidence of prematu-rity by folic acid supplementation in pregnancy. Br Med J. 1970;1(687):16–17

181. Lopez A, Murray C. The global burden of disease, 1990–2020. Nat Med.1998;4:1241–1243

182. Suarez L, Hendricks K, Cooper S, Sweeney A, Hardy R, Larsen R.Neural tube defects among Mexican Americans living on the US-Mexico border: effects of folic acid and dietary folate. Am J Epidemiol.2000;152:1017–1023

183. MRC Vitamin Study Research Group. Prevention of neural tubedefects: results of the Medical Research Council Vitamin Study. Lancet.1991;338:131–137

184. Berry J, Li Z, Erickson D, et al. Prevention of neural tube defects withfolic acid in China. N Engl J Med. 1999;341:1485–1490

185. Department of Health. Folic Acid and the Prevention of Disease: Report ofthe Committee on Medical Aspects of Food and Nutrition Policy. London,United Kingdom: The Stationery Office; 2000

186. Milunsky A, Jick H, Jick S, et al. Multivitamin/folic acid supplemen-tation in early pregnancy reduces the prevalence of neural tube de-fects. JAMA. 1989;262:2847–2852

187. Lumley J, Watson L, Watson MC. Periconceptional supplementationwith folate and/or multivitamins for preventing neural tube defects(Cochrane Review). Oxford, United Kingdom: Update Software; 2001a

188. Indian Council of Medical Research Collaborating Centres and CentralTechnical Co-ordinating Unit. Multicentric study of efficacy ofpericonceptional folic acid containing vitamin supplementation in pre-vention of open neural tube defects from India. Indian J Med Res.2000;112:206–211

189. Rolschau J, Kristoffersen K, Ulrich M, Grinsted P, Schaumburg E,Foged N. The influence of folic acid supplement on the outcome ofpregnancies in the county of Funen in Denmark. Part I. Eur J ObstetGynecol Reprod Biol. 1999;87:105–110

190. Laurence K, James N, Miller M, Tennant G, Campbell H. Double-blindrandomised controlled trial of folate treatment before conception toprevent recurrence of neural-tube defects. Br Med J (Clin Res Ed).1981;282:1509–1511

191. Czeizel A, Dudas I. Prevention of the first occurrence of neural-tubedefects by periconceptional vitamin supplementation. N Engl J Med.1992;327:1832–1835

192. Czeizel A. Prevention of congenital abnormalities by periconceptionalmultivitamin supplementation. BMJ. 1993;306:1645–1648

193. Shaw G, Lammer E, Wasserman R, Malley D, Tolarova M. Risks oforofacial clefts in children born to women using multivitamins con-taining folic acid periconceptionally. Lancet. 1995;346:393–396

194. Yang Q, Khoury M, Olney R, Mulinare J. Does periconceptional mul-tivitamin use reduce the risk for limb deficiency in offspring? Epide-miology. 1997;8:157–161

195. Li D, Daling R, Mueller B, Hickok D, Fantel A, Weiss N. Periconcep-tional multivitamin use in reduction to the risk of congenital urinarytract anomalies. Epidemiology. 1995;6:212–218

196. El-Khairy L, Vollset S, Refsum H, Ueland P. Plasma total cysteine,pregnancy complications, and adverse pregnancy outcomes: theHordaland Homocysteine Study. Am J Clin Nutr. 2003;77:467–472

197. Ozerol E, Ozerol I, Gokdeniz R, Temel I, Akyol O. Effect of smoking onserum concentrations of total homocysteine, folate, vitamin B12, and

nitric oxide in pregnancy: a preliminary study. Fetal Diagn Ther. 2004;19:145–148

198. Lopez-Quesada E, Vilaseca M, Lailla J. Plasma total homocysteine inuncomplicated pregnancy and in preeclampsia. Eur J Obstet GynecolReprod Biol. 2003;108:45–49

199. Lopez-Quesada E, Vilaseca M, Vela A, Lailla J. Perinatal outcomeprediction by maternal homocysteine and uterine artery Doppler ve-locimetry. Eur J Obstet Gynecol Reprod Biol. 2004;113:61–66

200. Guerra-Shinohara E, Paiva A, Rondo P, Yamasaki K, Terzi C,D’Almeida V. Relationship between total homocysteine and folatelevels in pregnant women and their newborn babies according tomaternal serum levels of vitamin B12. BJOG. 2002;109:784–791

201. Hasan B, Bhutta Z. Periconceptional use of folic acid in pregnancy[commentary]. WHO Reprod Health Libr. 2002:5

202. Recommendations of the International Task Force for Disease Eradi-cation. MMWR Recomm Rep. 1992;42(RR16):1–25

203. Mahomed K, Gulmezoglu A. Maternal iodine supplements in areas ofdeficiency. In: Cochrane Database Syst Rev. 2000:CD000135

204. Pharoah P, Buttfield I, Hetzel B. Neurological damage to the fetusresulting from severe iodine deficiency during pregnancy. Lancet. 1971;1:308–310

205. Pharoah P, Connolly K. A controlled trial of iodinated oil for theprevention of endemic cretinism: a long-term follow-up. Int J Epide-miol. 1987;16:68–73

206. Pharoah P, Connolly K. Effects of maternal iodine supplementationduring pregnancy. Arch Dis Child. 1991;66:145–147

207. Thilly C, Swennen B, Moreno-Reyes R, Hindlet J, Bourdoux P, Van-deroas J. Maternal, fetal and juvenile hypothyroidism, birthweight andinfant mortality in the etiopathogenesis of the IDD spectrum in Zaireand Malawi. In: Stanbury J, ed. The Damaged Brain of Iodine Deficiency.New York, NY: Cognizant Communication; 1994:241–250

208. Thilly C, Delange F, Lagasse R, et al. Fetal hypothyroidism and ma-ternal thyroid status in severe endemic goiter. J Clin Endocrinol Metab.1978;47:354–360

209. Thilly C, Lagasse R, Roger G, Bourdoux P, Ermans A. Impaired fetaland postnatal development and high perinatal death rate in a severeiodine deficient area. In: Stockigt J, ed. Thyroid Research VIII. Proceed-ings of the 8th International Thyroid Congress. Canberra, Australia: Aus-tralian Academy of Sciences; 1980:20–23

210. Beaton G, Martorell R, Aronson K, et al. Effectiveness of vitamin Asupplementation in the control of young child morbidity and mortalityin developing countries. ACC/SCN Nutrition Policy Discussion PaperNo. 13. Geneva, Switzerland: World Health Organization; 1993

211. Christian P, West KJ, Khatry S, et al. Night blindness of pregnancy inrural Nepal-nutritional and health risks. Int J Epidemiol. 1998;27:231–237

212. Christian P, West KJ, Khatry S, et al. Maternal night blindness in-creases risk of mortality in the first 6 months of life among infants inNepal. J Nutr. 2001;131:1510–1512

213. Crosby W, Metcoff J, Costiloe J, et al. Fetal malnutrition: an appraisalof correlated factors. Am J Obstet Gynecol. 1977;128:22–31

214. Ghebremeskel K, Burns L, Burden T, et al. Vitamin A and relatedessential nutrients in cord blood: relationships with anthropometricmeasurements at birth. Early Hum Dev. 1994;39:177–188

215. Metcoff J, Costiloe J, Crosby W, et al. Maternal nutrition and fetaloutcome. Am J Clin Nutr. 1981;34:708–721

216. Howells D, Haste F, Rosenberg D, Brown I, Brooke O. Investigation ofvitamin A nutrition in pregnant British Asians and their infants. HumNutr Clin Nutr. 1986;40C:43–50

217. Neel N, Alvarez J. Chronic fetal malnutrition and vitamin A in cordserum. Eur J Clin Nutr. 1990;44:207–212

218. West KJ, Katz J, Khatry K, et al. Double blind, cluster randomized trialof low dose supplementation with vitamin A or B carotene on mor-tality related to pregnancy in Nepal. BMJ. 1999;318:570–575

219. Katz J, West KJ, Khatry S, et al. Maternal low-dose vitamin A orbeta-carotene supplementation has no effect on fetal loss and earlyinfant mortality: a randomized cluster trial in Nepal. Am J Clin Nutr.2000;71:1570–1576

220. Fawzi W, Msamaga G, Spiegelman D, et al. Randomised trial of effectsof vitamin supplements on pregnancy outcomes and T cell counts inHIV-1 infected women in Tanzania. Lancet. 1998;351:1477–1482

221. Coutsoudis A, Bobat R, Coovadia H, Kuhn L, Tsai W, Stein Z. Theeffects of vitamin A supplementation on the morbidity of childrenborn to HIV-infected women. Am J Public Health. 1995;85:1076–1081

222. Coutsoudis A, Pillay K, Spooner E, Kuhn L, Coovadia H. Randomizedtrial testing the effect of vitamin A supplementation on pregnancyoutcomes and early mother-to-child HIV-1 transmission in Durban,



South Africa. South African Vitamin A Study Group. AIDS. 1999;13:1517–1524

223. Semba R. Overview of the potential role of vitamin A in mother-to-child transmission of HIV-1. Acta Paediatr Suppl. 1997;421:107–112

224. Kumwenda N, Miotti P, Taha T, et al. Antenatal vitamin A supple-mentation increases birth weight and decreases anemia among infantsborn to human immunodeficiency virus-infected women in Malawi.Clin Infect Dis. 2002;35:618–624

225. King J. Determinants of maternal zinc status during pregnancy. Am JClin Nutr. 2000;71(suppl 5):1334S–1343S

226. Keen C, Clegg M, Hanna L, et al. The plausibility of micronutrientdeficiencies being a significant contributing factor to the occurrence ofpregnancy complications. J Nutr. 2003;133:1597S–1605S

227. Mahomed K. Zinc supplementation in pregnancy. Cochrane DatabaseSyst Rev. 2000:CD000230

228. Osendarp S, West C, Black R. The need for maternal zinc supplemen-tation in developing countries: an unresolved issue. J Nutr. 2003;133:817S–827S

229. Nurdiati D. Nutrition and reproductive health in central Java, Indo-nesia [medical dissertations]. Umea, Sweden: University of Umea; 2001

230. Castillo-Duran C, Marin V, Alcazar L, Irurralde H, Ruz M. Controlledtrial of zinc supplementation in Chilean pregnant adolescents. NutrRes. 2001;21:715–724

231. Goldenberg R, Tamura T, Neggers Y, et al. The effect of zinc supple-mentation on pregnancy outcome. JAMA. 1995;274:463–468

232. Osendarp S, van Raaij J, Arifeen S, Wahed M, Baqui A, Fuchs G. Arandomized, placebo-controlled trial of the effect of zinc supplemen-tation during pregnancy on pregnancy outcome in Bangladeshi urbanpoor. Am J Clin Nutr. 2000;71:114–119

233. Osendarp S, van Raaij J, Darmstadt G, Baqui A, Hautvast J, Fuchs G.Zinc supplementation during pregnancy and effects on growth andmorbidity in low birthweight infants: a randomised placebo controlledtrial. Lancet. 2001;357:1080–1085

234. Sazawal S, Black R, Menon V, et al. Zinc supplementation in infantsborn small for gestational age reduces mortality: a prospective, ran-domized, controlled trial. Pediatrics. 2001;108:1280–1286

235. Hamadani J, Fuchs G, Osendarp S, Huda S, Grantham-McGregor S.Zinc supplementation during pregnancy and effects on mental devel-opment and behaviour of infants: a follow up study. Lancet. 2002;360:290–294

236. Hamadani J, Fuchs G, Saskia J, et al. Randomized controlled trial of theeffect of zinc supplementation on the mental development of Bang-ladeshi infants. Am J Clin Nutr. 2001;74:381–386

237. International Zinc Nutrition Consultative Group. Assessment of zincdeficiency in populations and options for its control. Technical Docu-ment No. 1. Food Nutr Bull. 2004;25:S92–S203

238. Friis H, Michaelson K. Micronutrients and HIV infection: a review. EurJ Clin Nutr. 1998;52:157–163

239. Bhutta Z. Sind Maternal Micronutrient Data and Intervention Plan. Na-tional Nutrition Consultation. Bhurban, Pakistan: Planning Commis-sion Government of Pakistan; 2001

240. Scholl T, Hediger M, Bendich A, Schall J, Smith W, Krueger P. Use ofmultivitamin/mineral prenatal supplements: influence on the out-come of pregnancy. Am J Epidemiol. 1997;146:134–141

241. Ahluwalia B, Hogan K, Grummer-Strawn L, Coulville W, Peterson A.The effect of WIC participation on small for gestational age births:Michigan, 1992. Am J Public Health. 1998;8:1374–1377

242. Rush D. Maternal nutrition and perinatal survival. Nutr Rev. 2001;59:315–326

243. Higgins A, Moxley J, Pencharz P, Mikolainis D, Dubois S. Impact of theHiggins Nutrition Intervention Program on birthweight: a within-mother analysis. J Am Diet Assoc. 1989;89:1097–1103

244. Kielmann A, Taylor C, Parker R. The Narangwal Nutrition Study: asummary review. Am J Clin Nutr. 1978b;31:2040–2052

245. Mora J, de Paredes B, Wagner M, et al. Nutritional supplementationand the outcome of pregnancy. I. Birth weight. Am J Clin Nutr. 1979;32:455–462

246. Ross S, Nel E, Naeye R. Differing effects of low and high bulk maternaldietary supplements during pregnancy. Early Hum Dev. 1985;10:395–302

247. Gopalan C. Multiple micronutrient supplementation in pregnancy.Nutr Rev. 2002;60(5 pt 2):S2–S6

248. Christian P, Khatry S, Katz J, et al. Effects of alternative maternalmicronutrient supplements on low birth weight in rural Nepal: doubleblind randomised community trial. BMJ. 2003;326:571

249. Christian P, West KJ, Khatry S, et al. Effects of maternal micronutrientsupplementation on fetal loss and infant mortality: a cluster-randomized trial in Nepal. Am J Clin Nutr. 2003;78:1194–1202

250. Ramakrishnan U, Gonzalez-Cossio T, Neufeld L, Riveria J, Martorell R.Multiple micronutrient supplementation during pregnancy does notlead to greater infant birth size than does iron-only supplementation:a randomized controlled trial in a semirural community in Mexico.Am J Clin Nutr. 2003;77:720–725

251. Ramakrishnan U, Neufeld L, Gonzalez-Cossio T, et al. Multiple micro-nutrient supplements during pregnancy do not reduce anemia orimprove iron status compared to iron-only supplements in semiruralMexico. J Nutr. 2004;134:898–903

252. Weiss U. Malaria [editorial]. Nat Insight. 2002;415:669253. Diagne N, Rogier C, Cisse B, Trape J. Incidence of clinical malaria in

pregnant women exposed to intense perennial transmission. Trans RSoc Trop Med Hyg. 1997;91:166–170

254. Diagne N, Rogier C, Sokhna C, et al. Increased susceptibility to malariaduring the early postpartum period. N Engl J Med. 2000;343:598–603

255. Menendez C. Malaria during pregnancy: a priority area of malariaresearch and control. Parasitol Today. 1995;11:178–183

256. Duffy P, Fried M. Malaria during pregnancy: parasites, antibodies andchondroitin sulphate A. Biochem Soc Trans. 1999;27:478–482

257. Fried M, Duffy P. Adherence of Plasmodium falciparum to chondroitinsulfate A in the human placenta. Science. 1996;272:1502–1504

258. Steketee R, Wirima J, Hightower A, Slutsker L, Heymann D, Breman J.The effect of malaria and malaria prevention in pregnancy on offspringbirthweight, prematurity, and intrauterine growth retardation in ruralMalawi. Am J Trop Med Hyg. 1996;55(1 suppl):33–41

259. Steketee R, Wirima J, Slutsker L, Breman J, Heymann D. Comparabilityof treatment groups and risk factors for parasitemia at the first ante-natal clinic visit in a study of malaria treatment and prevention inpregnancy in rural Malawi. Am J Trop Med Hyg. 1996;55(1 suppl):17–23

260. Steketee R, Wirima J, Slutsker L, Khoromana C, Heymann D, BremanJ. Malaria treatment and prevention in pregnancy: indications for useand adverse events associated with use of chloroquine and meflo-quine. Am J Trop Med Hyg. 1996;55(1 suppl):50–56

261. Steketee R, Wirima J, Slutsker L, Heymann D, Breman J. The problemof malaria and malaria control in pregnancy in sub-Saharan Africa.Am J Trop Med Hyg. 1996;55(1 suppl):2–7

262. Steketee R, Wirima J, Bloland P, et al. Impairment of a pregnantwoman’s acquired ability to limit Plasmodium falciparum by infectionwith human immunodeficiency virus type-1. Am J Trop Med Hyg.1996;55(1 suppl):42–49

263. McGregor I. Epidemiology, malaria and pregnancy. Am J Trop MedHyg. 1984;33:517–525

264. Miller L, Smith J. Motherhood and malaria. Nat Med. 1998;4:1244–1245265. Brabin B, Piper C. Anaemia- and malaria-attributable low birthweight

in two populations in Papua New Guinea. Ann Hum Biol. 1997;24:547–555

266. Brabin B. The Risks and Severity of Malaria in Pregnant Women. Geneva,Switzerland: World Health Organization; 1991

267. Brabin B. Malaria in pregnancy: current issues. Afr Health. 1997;19:19–20

268. Miller L, Baruch D, Marsh K, Doumbo O. The pathogenic basis ofmalaria. Nat Rev. 2002;415:673–679

269. Brabin B. An assessment of low birthweight risk in primiparae as anindicator of malaria control in pregnancy. Int J Epidemiol. 1991;20:276–283

270. Verhoeff F, Brabin B, Hart C, Chimsuku L, Kazembe P, Broadhead R.Increased prevalence of malaria in HIV-infected pregnant women andits implications for malaria control. Trop Med Int Health. 1999;4:5–12

271. Garner P, Gulmezoglu A. Interventions to prevent malaria duringpregnancy in endemic malarious areas: prevention versus treatmentfor malaria in pregnant women. In: Garner P, Gelband H, Olliaro P,Salinas R, Volmink J, Wilkinson D, eds. Cochrane Database Syst Rev.2003;(1):CD000169

272. D’Alessandro U, Langerock P, Bennett S, Francis N, Cham K, Green-wood B. The impact of a national impregnated bed net programme onthe outcome of pregnancy in primigravidae in The Gambia. Trans R SocTrop Med Hyg. 1996;90:487–492

273. Newman R, Parise M, Slutsker L, Nahlen B, Steketee R. Safety, efficacyand determinants of effectiveness of antimalarial drugs duringpregnancy: implications for prevention programmes in Plasmodiumfalciparum-endemic sub-Saharan Africa. Trop Med Int Health. 2003;8:488–506

274. Shulman C, Dorman E, Cutts F, et al. Intermittent sulphadoxine-pyrimethimine to prevent severe anaemia secondary to malaria inpregnancy: a randomised, placebo-controlled trial. Lancet. 1999;353:632–636

275. Greenwood B, Greenwood A, Snow R, Byass P, Bennett S, Hatib NJieA. The effects of malaria chemoprophylaxis given by traditional birth



attendants on the course and outcome of pregnancy. Trans R Soc TropMed Hyg. 1989;83:589–594

276. Greenwood A, Menendez C, Todd J, Greenwood B. The distribution ofbirth weights in Gambian women who received malaria hemoprophy-laxis during their first pregnancy and in control women. Trans R SocTrop Med Hyg. 1994;88:311–312

277. Nyirjesy P, Kavasya T, Axelrod P, Fischer P. Malaria duringpregnancy: neonatal morbidity and mortality and the efficacy of chlo-roquine chemoprophylaxis. Clin Infect Dis. 1993;16:127–132

278. Mutabingwa T, Malle L, de Geus A, Oosting J. Malaria chemosuppres-sion in pregnancy. II. Its effect on maternal haemoglobin levels, pla-cental malaria and birth weight. Trop Geogr Med. 1993;45:49–55

279. Ndyomugyenyi R, Magnussen P. Chloroquine prophylaxis, iron-folicacid supplementation or case management of malaria attacks in pri-migravidae in western Uganda: effects on maternal parasitaemia andhaemoglobin levels and on birthweight. Trans R Soc Trop Med Hyg.2000;94:413–418

280. Parise M, Ayisi J, Nahlen B, et al. Efficacy of sulfadoxine-pyrimeth-amine for prevention of placental malaria in an area of Kenya with ahigh prevalence of malaria and human immunodeficiency virus infec-tion. Am J Trop Med Hyg. 1998;59:813–822

281. Bouvier P, Breslow N, Doumbo O, et al. Seasonality, malaria, andimpact of prophylaxis in a West African village. II. Effect on birth-weight. Am J Trop Med Hyg. 1997;56:384–389

282. Verhoeff F, Brabin B, Chimsuku L, Kazembe P, Russell W, BroadheadR. An evaluation of the effects of intermittent sulfadoxine-py-rimethamine treatment in pregnancy on parasite clearance and risk oflow birth weight in rural Malawi. Ann Trop Med Parasitol. 1998;92:141–150

283. Wolfe E, Parise M, Haddix A, et al. Cost-effectiveness of sulfadoxine-pyrimethamine for the prevention of malaria-associated low birthweight. Am J Trop Med Hyg. 2001;64:178–186

284. Cot M, LeHesran JY, Miailhes P, Esveld M, Etya’ale D, Breart G.Increase of birth weight following chloroquine chemoprophylaxis dur-ing the first pregnancy: results of a randomized trial in Cameroon.Am J Trop Med Hyg. 1995;53:581–585

285. Sowunmi A, Oduola O, Ogundahunsi F, et al. Randomised trial ofartemether versus artemether and mefloquine for the treatment ofchloroquine/sulfadoxine-pyrimethamine-resistant falciparum malariaduring pregnancy. J Obstet Gynaecol. 1998;18:322–327

286. van Eijk A, Ayisi J, ter Kuile F, et al. Effectiveness of intermittentpresumptive treatment with sulphoxadine-pyrimethamine for controlof malaria in pregnancy in western Kenya: a hospital-based study. TropMed Int Health. 2004;9:351–360

287. Schultz L, Steketee R, Macheso A, Kazembe P, Chitsulo L, Wirima J.The efficacy of antimalarial regimens containing sulfadoxine-pyrimethamine and/or chloroquine in preventing peripheral and pla-cental Plasmodium falciparum infection among pregnant women inMalawi. Am J Trop Med Hyg. 1994;51:515–522

288. ter Kuile F, Terlouw D, Phillips-Howard P, et al. Reduction of malariaduring pregnancy by permethrin-treated bed nets in an area of intenseperennial malaria transmission in Kenya. Am J Trop Med Hyg. 2003;68(4suppl):50–60

289. Hawley W, ter Kuile F, Steketee R, et al. Implications of the WesternKenya Permethrin-Treated Bed Net Study for policy, program imple-mentation, and future research. Am J Trop Med Hyg. 2003;68(suppl4):168–173

290. Steketee RW, Nahlen BL, Parise ME, Menendez C. The burden ofmalaria in pregnancy in malaria-endemic areas. Am J Trop Med Hyg.2001;64:28–35

291. Dolan G, ter-Kuile F, Jacoutot V, et al. Bed nets for the prevention ofmalaria and anaemia in pregnancy. Trans R Soc Trop Med Hyg. 1993;87:620–626

292. Shulman C, Dorman E, Talisuna A, et al. A community randomizedcontrolled trial of insecticide-treated bednets for the prevention ofmalaria and anemia among primigravid women on the Kenyan coast.Trop Med Int Health. 1998;3:197–204

293. Hill A, MacLeod W, Joof D, Gomez P, Walraven G. Decline of mor-tality in children in rural Gambia: the influence of village-level pri-mary health care. Trop Med Int Health. 2000;5:107–118

294. World Health Organization. The World Health Report 1996: FightingDisease, Fostering Development. Geneva, Switzerland: World HealthOrganization; 1996

295. Pawlowski Z, Schad G, Stott G. Hook Worm Infection and Anaemia:Approaches to Prevention and Control. Geneva, Switzerland: WorldHealth Organization; 1991

296. Albonico M, Smith P, Hall A, Chwaya H, Alawi K, Savioli L. Arandomized controlled trial comparing mebendazole and albendazole

against Ascaris, Trichuris and hookworm infections. Trans R Soc TropMed Hyg. 1994;88:585–589

297. Abel R, Rajaratnam J, Kalaimani A, Kirubakaran S. Can iron status beimproved in each of the three trimesters? A community-based study.Eur J Clin Nutr. 2000;54:490–493

298. Gilgen D, Mascie-Taylor C. The effect of anthelmintic treatment onhelminth infection and anaemia. Parasitology. 2001;122:105–110

299. De Silva N, Sirisena J, Gunasekera D, Ismail M, de Silva H. Effect ofmebendazole therapy during pregnancy on birth outcomes. Lancet.1999;353:145–149

300. Christian P, Khatry SK, West KP. Antenatal anthelmintic treatment,birthweight, and infant survival in rural Nepal. Lancet. 2004;364:981–983

301. Maleckiene L, Nadisauskiene R, Stankeviciene I, Cizauskas A, Berg-strom S. A case-referent study on fetal bacteremia and late fetal deathof unknown etiology in Lithuania. Acta Obstet Gynecol Scand. 2000;79:1069–1074

302. Brocklehurst P, Hannah M, McDonald H. Interventions for treatingbacterial vaginosis in pregnancy. Cochrane Database Syst Rev. 2000:CD000262

303. Morgan D, Aboud C, McCaffrey I, Bhide S, Lamont R, Taylor-Robinson D. Comparison of Gram-stained smears prepared from blindvaginal swabs with those obtained at speculum examination for theassessment of vaginal flora. Br J Obstet Gynaecol. 1996;103:1105–1108

304. Priestley C, Jones B, Dhar J, Goodwin L. What is normal vaginal flora?Genitourin Med. 1997;73:23–28

305. Naeye R, Tafari N, Judge D, Gilmour D, Marboe C. Amniotic fluidinfections in an African city. J Pediatr. 1977;90:965–970

306. Ross S, Macpherson T, Naeye R, Khatree M, Wallace J. Causes of fetaland neonatal mortality in a South African black community. S Afr MedJ. 1982;61:905–908

307. Moyo S, Hagerstrand I, Nystrom L, et al. Stillbirths and intrauterineinfection, histologic chorioamnionitis and microbiological findings. IntJ Gynaecol Obstet. 1996;54:115–123

308. Folgosa E, Gonzalez C, Osman N, Hagerstrand I, Bergstrom S, LjunghA. A case control study of chorioamniotic infection and histologicalchorioamnionitis in stillbirth. APMIS. 1997;105:329–336

309. Matthews J, Mathai M, Peedicayil A, Matthews K, Ponnaiya J, JasperM. Subclinical chorioamnionitis as a causal factor in unexplainedstillbirths. Int J Gynaecol Obstet. 2001;74:195–197

310. Mascola L, Pelosi R, Blount J, Alexander C, Cates WJ. Congenitalsyphilis revisited. Am J Dis Child. 1985;139:575–580

311. Schultz K, Cates WJ, O’Mara P. Pregnancy loss, infant death andsuffering: legacy of syphilis and gonorrhoea in Africa. Genitourin Med.1987;63:320–325

312. McDermott J, Steketee R, Larsen S, Wirima J. Syphilis-associated peri-natal and infant mortality in rural Malawi. Bull World Health Organ.1993;71:773–780

313. Ratnam A, Din S, Hira S, et al. Syphilis in pregnant women in Zambia.Br J Vener Dis. 1982;58:355–358

314. Liljestrand J, Bergstrom S, Nieuwenhuis E, Hederstedt B. Syphilis inpregnant women in Mozambique. Genitourin Med. 1985;6:355–358

315. Mabey D. Syphilis in sub-Saharan Africa. Afr J Sex Transmi Dis. 1986;2:61–64

316. Guinness L, Sibandze S, McGrath E, Cornelis A. Influence of antenatalscreening on perinatal mortality caused by syphilis in Swaziland.Genitourin Med. 1987;64:294–297

317. Humphrey M, Bradford D. Congenital syphilis: still a reality in 1996.Med J Aust. 1996;165:382–385

318. Judge D, Tefari N, Naeye R, Narboe C. Congenital syphilis and peri-natal mortality. Pediatr Pathol. 1986;5:411–420

319. Chattopadhyay B. Prevention of congenital syphilis. Br J Hosp Med.1988;40:68–70

320. Bastos dos Santos R, Folgosa E, Fransen L. Reproductive tract infec-tions in Mozambique: a case study of integrated services. In: GermainA, Holmes K, Piot P, Wasserheit J, eds. Reproductive Tract Infections.Global Impact and Priorities for Women’s Health. New York, NY: PlenumPress; 1992:343–360

321. Prabhakar P, Bailey A, Smikle M, McCaw-Binns A, Ashley D. Sero-prevalence of Toxoplasma gondii, rubella virus, cytomegalovirus herpessimplex virus (TORCH) and syphilis in Jamaican pregnant women.West Indian Med J. 1991;40:166–169

322. Luthra U, Mehta S, Bhargava N, et al. Reproductive tract infections inIndia: the need for comprehensive reproductive health policy andprograms. In: Germain A, Holmes K, Piot P, Wasserheit J, eds. Repro-ductive Tract Infections. Global Impact and Priorities for Women’s Health.New York, NY: London, United Kingdom: Plenum Press; 1992



323. Maggwa A, Ngugi E. Reproductive tract infections in Kenya: insightsfor action and research. In: Germain A, Holmes K, Piot P, WasserheitJ, eds. Reproductive Tract Infections. Global Impact and Priorities for Wom-en’s Health. New York, NY: Plenum Press; 1992:275–295

324. Schultz K, Scholte J, Berman S. Maternal health and child survival:opportunities to protect both women and children from adverse con-sequences of reproductive tract infections. In: Germain A, Holmes K,Piot P, Wasserheit J, eds. Reproductive Tract Infections. Global Impact andPriorities for Women’s Health. New York, NY: Plenum Press; 1992:145–183

325. Faundes A, Tanaka A. Reproductive tract infections in Brazil: solutionsin a difficult economic climate. In: Germain A, Holmes K, Piot P,Wasserheit J, eds. Reproductive Tract Infections. Global Impact and Prior-ities for Women’s Health. New York, NY: Plenum Press; 1992:253–273

326. Goldenberg R, Andrews W, Yuan A, MacKay H, St Louis M. Sexuallytransmitted diseases and adverse outcomes of pregnancy. Clin Perina-tol. 1997;24:23–41

327. Swain G, Kowalewski S, Schubot D. Reducing the incidence of con-genital syphilis in Milwaukee: a public/private partnership. Am JPublic Health. 1998;88:1101–1102

328. Hira S, Bhat G, Chikamata D, et al. Syphilis intervention in pregnancy:Zambian demonstration project. Genitourin Med. 1990;66:159–164

329. Osman N, Challis K, Cotiro M, Nordahl G, Bergstrom S. Maternal andfetal characteristics in an obstetric cohort in Mozambique. Afr J ReprodHealth. 2000;4:110–119

330. Patel A, Moodley D, Moodley J. An evaluation of on-site testing forsyphilis. Trop Doct. 2001;31:79–82

331. Delport S, Van den Berg J. On site screening for syphilis at an antenatalclinic. S Afr Med J. 1998;88:43–44

332. Temmerman M, Gichangi P, Fonck K, et al. Effect of a syphilis controlprogramme on pregnancy outcome in Nairobi, Kenya. Sex TransmInfect. 2000;76:117–121

333. Vazquez J, Villar J. Treatments for symptomatic urinary tract infectionsduring pregnancy (Cochrane Review). Oxford, United Kingdom: Up-date Software; 2001

334. Smaill F. Antibiotics for asymptomatic bacteriuria in pregnancy (Co-chrane Review). Oxford, United Kingdom: Update Software; 2002

335. Dietrich M, Hoosen A, Moodley J, Moodley S. Urogenital tract infec-tions in pregnancy at King Edward VIII Hospital, Durban, SouthAfrica. Genitourin Med. 1982;68:39–41

336. Orrett F, Balbirsingh M, Carrington L. Socio-biological associations ofbacteriuria in pregnancy. West Indian Med J. 1995;44:28–31

337. Versi E, Chia P, Griffiths D, Harlow B. Bacteriuria in pregnancy: acomparison of Bangladeshi and Caucasian women. Int Urogynecol JPelvic Floor Dysfunct. 1997;8:8–12

338. Qureshi R, Khan K, Darr O, Khattak N, Farooqui B, Rizvi J. Bacteriuriaand pregnancy outcome: a prospective hospital-based study in Paki-stani women. J Pak Med Assoc. 1994;44:12–13

339. Raghavan M, Mondal G, Bhat B, Srinivasan S. Perinatal risk factors inneonatal infections. Indian J Pediatr. 1992;59:335–340

340. Tolosa J. RHL commentary: antibiotic versus no treatment for asymp-tomatic bacteriuria in pregnancy. Geneva, Switzerland: World HealthOrganization; 2002

341. Lamont R, Fisk N. The role of infection in the pathogenesis of pretermlabour. In: Studd J, ed. Progress in Obstetrics and Gynaecology. Edin-burgh, United Kingdom: Churchill Livingstone; 1993:135–158

342. Hay P, Lamont R, Taylor-Robinson D, Morgan D, Ison C, Pearson J.Abnormal bacterial colonisation of the genital tract and subsequentpreterm delivery and late miscarriage. BMJ. 1994;308:295–298

343. Hillier S, Nugent R, Eschenbach D, et al. Association between bacterialvaginosis and preterm delivery of a low-birth-weight infant. The Vag-inal Infections and Prematurity Study Group. N Engl J Med. 1995;333:1737–1742

344. Joesoef M, Hillier S, Wiknjosastro G, et al. Intravaginal clindamycintreatment for bacterial vaginosis: effects on preterm delivery and lowbirth weight. Am J Obstet Gynecol. 1995;173:1527–1531

345. Wawer M, Sewankambo N, Serwadda D, et al. Control of sexuallytransmitted diseases for AIDS prevention in Uganda: a randomisedcommunity trial. Lancet. 1999;353:525–535

346. McGregor J, French J, Parker R, et al. Prevention of premature birth byscreening and treatment for common genital tract infections: results ofa prospective controlled evaluation. Am J Obstet Gynecol. 1995;173:157–167

347. Joesoef M, Schmid G, Hillier S. Bacterial vaginosis: review of treatmentoptions and potential clinical indications for therapy. Clin Infect Dis.1999;28(suppl 1):S57–S65

348. Duff P, Lee M, Hillier S, Herd L, Krohn M, Eschenbach D. Amoxicillintreatment of bacterial vaginosis during pregnancy. Obstet Gynecol.1991;77:431–435

349. Morales W, Schorr S, Albritton J. Effect of metronidazole in patientswith preterm birth in preceding pregnancy and bacterial vaginosis: aplacebo-controlled, double-blind study. Am J Obstet Gynecol. 1994;171:345–349

350. Hauth J, Goldenberg R, Andrews W, DuBard M, Copper R. Reducedincidence of preterm delivery with metronidazole and erythromycin inwomen with bacterial vaginosis. N Engl J Med. 1995;333:1732–1736

351. McDonald H, O’Loughlin J, Vigneswaran R, et al. Impact of metroni-dazole therapy on preterm birth in women with bacterial vaginosisflora (Gardnerella vaginalis): a randomised, placebo controlled trial. Br JObstet Gynaecol. 1997;104:1391–1397

352. Carey J, Klebanoff M, Hauth J, et al. Metronidazole to prevent pretermdelivery in pregnant women with asymptomatic bacterial vaginosis.N Engl J Med. 2000;342:534–540

353. Klebanoff M, Carey J, Hauth J, et al. Failure of metronidazole toprevent preterm delivery among pregnant women with asymptomaticTrichomonas vaginalis infection. N Engl J Med. 2001;345;487–493

354. Mercer B, Miodovnik M, Thurnau G, et al. Antibiotic therapy forreduction of infant morbidity after preterm premature rupture of themembranes. A randomized controlled trial. National Institute of ChildHealth and Human Development Maternal-Fetal Medicine UnitsNetwork. JAMA. 1997;278:989–995

355. Martin D, Eschenbach D, Cotch M, et al. Double-blind placebo-controlled treatment trial of Chlamydia trachomatis endocervical infec-tions in pregnant women. Infect Dis Obstet Gynecol. 1997;5:10–17

356. Norman K, Pattinson R, de Souza J, de Jong P, Moller G, Kirsten G.Ampicillin and metronidazole treatment in preterm labour: a multi-centre, randomized controlled trial. Br J Obstet Gynaecol. 1994;101:404–408

357. Gichangi P, Ndinya-Achola J, Ombete J, Nagelkerke N, TemmermanM. Antimicrobial prophylaxis in pregnancy: a randomized, placebo-controlled trial with cefetamet-pivoxil in pregnant women with a poorobstetric history. Am J Obstet Gynecol. 1997;177:680–684

358. Gray R, Wabwire-Mangen F, Kigozi G, et al. Randomized trial ofpresumptive sexually transmitted disease therapy during pregnancyin Rakai, Uganda. Am J Obstet Gynecol. 2001;185:1209–1217

359. Thinkhamrop J. Commentary: interventions for treating bacterial vagi-nosis in pregnancy. Geneva, Switzerland: World Health Organization;2002

360. Hall M, Danielian P, Lamont R. The importance of preterm birth. In:Elder M, Romero R, Lamont R, eds. Preterm Labour. Edinburgh, UnitedKingdom: Churchill Livingstone; 1997:1–28

361. Berkovitz G, Papiernik E. Epidemiology of preterm birth. EpidemiolRev. 1993;15:414–443

362. Romero R, Athayde N, Maymon E, Pacora P, Bahado-Singh R. Prema-ture rupture of membranes. In: Reece A, Hobbins J, eds. Medicine ofthe fetus and mother. Philadelphia, PA: Lippincott-Raven; 1999:1581–1625

363. Russell P. Inflammatory lesions of the human placenta: clinical signif-icance of acute chorioamnionitis. Am J Obstet Gynecol. 1979;2:127–137

364. Naeye R, Peters E. Causes and consequences of premature rupture offoetal membranes. Lancet. 1980;1:192–194

365. Guzick D, Winn K. The association of chorioamnionitis with pretermdelivery. Obstet Gynecol. 1985;65:11–15

366. Gomez R, Romero R, Mazor M, Ghezzi F, David C, Yoon B. The role ofinfection in preterm labour and delivery. In: Elder M, Romero R,Lamont R, eds. Preterm Labour. Edinburgh, United Kingdom: ChurchillLivingstone; 1997:85–126

367. King J, Flenady V. Antibiotics for preterm labour with intact mem-branes (Cochrane Review). Oxford, United Kingdom: UpdateSoftware; 2002

368. Kenyon S, Taylor D, Tarnow-Mordi W. Broad-spectrum antibiotics forspontaneous preterm labour: the ORACLE II randomised trial. Lancet.2001;357:989–994

369. Oyarzun E, Gomez R, Rioseco A, et al. Antibiotic treatment in pretermlabor and intact membranes: a randomized, double-blinded, placebo-controlled trial. J Matern Fetal Med. 1998;7:105–110

370. Bell S. Mechanisms underlying prelabour rupture of the fetal mem-branes. In: Kingdom J, Jauniaux E, O’Brien S, eds. The Placenta: BasicScience and Clinical Practice. London, United Kingdom: RCOG Press;2000:187–204

371. Kenyon S, Boulvain M, Neilson J. Antibiotics for preterm prematurerupture of membranes (Cochrane Review). Oxford, United Kingdom:Update Software; 2002



372. Kenyon S, Taylor D, Tarnow-Mordi W. Broad-spectrum antibiotics forpreterm, prelabour rupture of fetal membranes: the ORACLE I ran-domised trial. Lancet. 2001;357:979–988

373. Almeida L, Schmauch A, Bergstrom S. A randomised study on theimpact of peroral amoxicillin in women with prelabour rupture ofmembranes preterm. Gynecol Obstet Invest. 1996;41:82–84

374. Ovalle-Salas A, Gomez R, Martinez M, et al. Antibiotic therapy inpatients with preterm premature rupture of membranes: a prospective,randomized, placebo-controlled study with microbiological assess-ment of the amniotic cavity and lower genital tract. Prenat NeonatalMed. 1997;2:213–222

375. Magwali T, Chipato T, Majoko F, Rusakaniko S, Mujaji C. Prophylacticaugmentin in prelabor preterm rupture of the membranes. Int J Gynae-col Obstet. 1999;65:261–265

376. United Nations Children’s Fund/World Health Organization/UnitedNations Population Fund. Maternal and Neonatal Tetanus Elimination by2005: Strategies for Achieving and Maintaining Elimination. Geneva,Switzerland: World Health Organization; 2000

377. Fauveau V, Mamdani M, Steinglass R, Koblinsky M. Maternal tetanus:magnitude, epidemiology and potential control measures. Int J Gynae-col Obstet. 1993;40:3–12

378. World Health Organization. Mother and Baby Package. Geneva,Switzerland: World Health Organization; 1993

379. Gupta S, Keyl P. Effectiveness of prenatal tetanus toxoid immunizationagainst neonatal tetanus in a rural area in India. Pediatr Infect Dis J.1998;17:316–321

380. Kapoor S, Reddaiah V, Lobo J. Control of tetanus neonatorum in arural area. Indian J Pediatr. 1991;58:341–344

381. Parashar U, Bennett J, Boring J, Hlady W. Topical antimicrobialsapplied to the umbilical cord stump: a new intervention against neo-natal tetanus. Int J Epidemiol. 1998;27:904–908

382. Black R, Huber D, Curlin G. Reduction of neonatal tetanus by massimmunization of non-pregnant women: duration of protection fromone or two doses of aluminum-adsorbed tetanus toxoid. Bull WorldHealth Organ. 1980;58:927–930

383. Rahman M, Chen L, Chakraborty J, Yunus M, Faruque A, ChowdhuryA. Use of tetanus toxoid for the prevention of neonatal tetanus. 1.Reduction of neonatal mortality by immunization of non-pregnant andpregnant women in rural Bangladesh. Bull World Health Organ. 1982;60:261–267

384. Rahman S. The effect of traditional birth attendants and tetanus toxoidin reduction of neonatal mortality. J Trop Pediatr. 1982;28:163–165

385. Berggren G, Verly A, Garnier N, Peterson W, Ewbank D, DieudonneW. Traditional midwives, tetanus immunization, and infant mortalityin rural Haiti. Trop Doct. 1983;13:79–87

386. Chongsuvivatwong V, Bujakorn L, Kanpoy V, Treetrong R. Control ofneonatal tetanus in southern Thailand. Int J Epidemiol. 1993;22:931–935

387. Traverso H, Kamil S, Rahim H, Samadi A, Boring J, Bennett J. Areassessment of risk factors for neonatal tetanus. Bull World HealthOrgan. 1991;69:573–579

388. Bennett J, Macia J, Traverso H, Banoagha S, Malooly C, Boring J.Protective effects of topical antimicrobials against neonatal tetanus. IntJ Epidemiol. 1997;26:897–903

389. Meegan M, Conroy R, Lengeny S, Renhault K, Nyangole J. Effect onneonatal tetanus mortality after a culturally-based health promotionprogramme. Lancet. 2001;358:640–641

390. Nessa S, Arco E, Kabir I. Birth kits for safe motherhood in Bangladesh.World Health Forum. 1992;13:66–69

391. Amooti-Kaguna B, Nuwaha F. Factors influencing choice of deliverysites in Rakai district of Uganda. Soc Sci Med. 2000;50:203–213

392. Beun M, Wood S. Acceptability and use of clean home delivery kits inNepal: a qualitative study. J Health Popul Nutr. 2003;21:367–373

393. Tsu V. Nepal Clean Home Delivery Kit: Evaluation of Health Impact.Washington, DC: Program for Appropriate Technology in Health; 2000

394. Galazka A, Gasse F, Henderson R. Neonatal tetanus and the globalexpanded programme on immunization. In: Kessel E, Awan A, eds.Maternal and Child Care in Developing Countries: Assessment, Promotion,Implementation: Proceedings of the Third International Congress for Mater-nal and Neonatal Health. Ott, Switzerland: Ott Verlag Thun; 1989:109–125

395. Stroh G, Aye K, Thaung U, Kyaw L. Measurement of mortality fromneonatal tetanus in Burma. Bull World Health Organ. 1987;65:309–316

396. Expanded Programme on Immunization. Neonatal tetanus mortalitysurveys: Africa. Wkly Epidemiol Rec. 1987;62:305–308

397. World Health Organization. Meeting on maternal and neonatal pneu-mococcal immunization. Wkly Epidemiol Rec. 1998;73:187–188

398. Shahid N, Steinhoff M, Hoque S, Begum T, Thompson C, Siber G.Serum, breast milk, and infant antibody after maternal immunizationwith pneumococcal vaccine. Lancet. 1995;346:1252–1257

399. Department of Health and Human Services. The Health Consequences ofSmoking for Women. Washington, DC: Department of Health and Hu-man Services, Surgeon General; 1980

400. Butler NR, Goldstein H, Ross EM. Cigarette smoking in pregnancy: itsinfluence on birth weight and perinatal mortality. Br Med J. 1972;2(806):127–130

401. Walsh RA. Effects of maternal smoking on adverse pregnancyoutcomes: examination of criteria of causation. Hum Biol. 1994;66:1059–1092

402. Boy E, Bruce N, Delgado H. Birth weight and exposure to kitchenwood smoke during pregnancy in rural Guatemala. Environ HealthPerspect. 2002;110:109–114

403. Bruce N, Perez-Padilla R, Albalak R. Indoor air pollution in developingcountries: a major environmental and public health challenge. BullWorld Health Organ. 2000;78:1078–1092

404. Brunekreef B, Holgate S. Air pollution and health. Lancet. 2002;360:1233–1242

405. Lumley J, Oliver S, Waters E. Interventions for promoting smokingcessation during pregnancy (Cochrane Review). Oxford, UnitedKingdom: Update Software; 2001b

406. Kendrick J, Zahniser S, Miller N, et al. Integrating smoking cessationinto routine public prenatal care: the Smoking Cessation in Pregnancyproject. Am J Public Health. 1995;85:217–222

407. Wisborg K, Henriksen T, Secher N. A prospective intervention study ofstopping smoking in pregnancy in a routine antenatal care setting. Br JObstet Gynaecol. 1998;105:1171–1176

408. Hjalmarson A, Hahn I, Svanberg B. Stopping smoking in pregnancy:effect of a self-help manual in controlled trial. Br J Obstet Gynaecol.1991;98:260–264

409. Sexton M, Hebel R. A clinical trial of change in maternal smoking andits effect on birth weight. JAMA. 1984;251:911–915

410. Secker-Walker R, Soloman L, Flynn B, Skelly J, Mead P. Reducingsmoking during pregnancy and postpartum: physician’s advice sup-ported by individual counseling. Prev Med. 1998;27:422–430

411. Li C, Windsor R, Perkins L, Goldenberg R, Lowe J. The impact oninfant birth weight and gestational age of cotinine-validated smokingreduction during pregnancy. JAMA. 1993;269:1519–1524

412. Macarthur C, Newton J, Knox E. Effect of anti-smoking health educa-tion on infant size at birth: a randomized controlled trial. Br J ObstetGynaecol. 1987;94:295–300

413. Haddow E, Knight G, Kloza E, Palomaki G, Wald N. Cotinine-assistedintervention in pregnancy to reduce smoking and low birthweightdelivery. Br J Obstet Gynaecol. 1991;98:859–865

414. Lefevre L, Evans J, Ewingman B. Is smoking an indication for prenatalultrasonography? RADIUS Study Group. Arch Fam Med. 1995;4:120–123

415. Yan R. How Chinese clinicians contribute to the improvement ofmaternity care. Int J Gynaecol Obstet. 1989;30:23–26

416. Jeffery H, Kocova M, Tozija F, et al. The impact of evidence-basededucation on a perinatal capacity-building initiative in Macedonia.Med Educ. 2004;38:435–447

417. Perry H, Shanklin D, Schroeder D. Impact of a community-basedcomprehensive primary healthcare programme on infant and childmortality in Bolivia. J Health Popul Nutr. 2003;21:383–395

418. Bergstrom S, Hojer B, Liljestrand J, Tunell R. Perinatal Health Care WithLimited Resources. London, United Kingdom: MacMillan; 1994

419. Bhutta Z. Effective interventions to reduce neonatal mortality andmorbidity from perinatal infection. In: Costello A, Manandhar D, eds.Improving Newborn Infant Health in Developing Countries. London,United Kingdom: Imperial College Press; 2000:289–308

420. Dykes A, Christensen K, Christensen P, Kahlmeter G. Chlorhexidinefor prevention of neonatal colonization with group B streptococci. II.Chlorhexidine concentrations and recovery of group B streptococcifollowing vaginal washing in pregnant women. Eur J Obstet GynecolReprod Biol. 1983;16:167–172

421. Dykes A, Christensen K, Christensen P. Chlorhexidine for preventionof neonatal colonization with group B streptococci. IV. Depressedpuerperal carriage following vaginal washing with chlorhexidine inlabour. Eur J Obstet Gynecol Reprod Biol. 1987;24:293–297

422. Christensen K, Christensen P, Dykes A, Kahlmeter G. Chlorhexidinefor prevention of neonatal colonization with group B streptococci. III.Effect of vaginal washing with chlorhexidine before rupture of themembranes. Eur J Obstet Gynecol Reprod Biol. 1985;19:231–236



423. Kollee L, Speyer I, Kuijck M, et al. Prevention of group B streptococcitransmission during delivery by vaginal application of chlorhexidinegel. Eur J Obstet Gynecol Reprod Biol. 1989;31:47–51

424. Burman L, Christensen P, Christensen K, et al. Prevention of excessneonatal morbidity associated with group B streptococci by vaginalchlorhexidine disinfection during labour. Lancet. 1992;340:65–69

425. Henrichsen T, Lindemann R, Svenningsen L, Hjelle K. Prevention ofneonatal infections by vaginal chlorhexidine in labour. Acta Paediatr.1994;83:923–926

426. Adriaanse A, Kollee L, Muytjens K, Nijuis J, de Haan A, Eskes T.Randomised study of vaginal chlorhexidine disinfection during laborto prevent vertical transmission of group B streptococci by vaginalchlorhexidine disinfection during labor to prevent vertical transmis-sion of group B streptococci. Eur J Obstet Gynecol Reprod Biol. 1995;61:135–141

427. Biggar R, Miotti P, Taha T, et al. Perinatal intervention trial in Africa:effect of a birth canal cleansing intervention to prevent HIV transmis-sion. Lancet. 1996;347:1647–1650

428. Taha T, Biggar R, Broadhead R, et al. Effect of cleansing the birth canalwith antiseptic solution on maternal and newborn morbidity andmortality in Malawi: clinical trial. BMJ. 1997;315(7102):216–220

429. Gaillard P, Mwanyumba F, Verhofstede C, et al. Vaginal lavage withchlorhexidine during labour to reduce mother-to-child HIVtransmission: clinical trial in Mombasa, Kenya. AIDS. 2001;15:389–396

430. Ellis M, Manandhar N, Manandhar D, Costello A. Risk factors forneonatal encephalopathy: the Kathmandu case-control study. BMJ.2000;320:1229–1236

431. Kumar R. Birth asphyxia in a rural community of North India. J TropPediatr. 1995;41:5–7

432. Dommisse J. The causes of perinatal deaths in the Greater Cape Townarea: a 12-month survey. S Afr Med J. 1991;80:270–275

433. Bhutta Z. Priorities in newborn care and development of clinicalneonatology in Pakistan: where to now? J Coll Physicians Surg Pak.1997;7:231–234

434. Swyer P. The Intensive Care of the Newly Born: Physiological Principles andPractice. Basel, Switzerland: Karger Publishers; 1975

435. Ellis M, Manandhar N, Shrestha P, Shrestha L, Manandhar D, CostelloA. Outcome at 1 year of neonatal encephalopathy in Kathmandu,Nepal. Dev Med Child Neurol. 1999;41:689–695

436. Gunn A, Gluckman P, Gunn T. Selective head cooling in newborninfants after perinatal asphyxia: a safety study. Pediatrics. 1998;102:885–892

437. Battin M, Dezoete J, Gunn T, Gluckman P, Gunn A. Neurodevelop-mental outcome of infants treated with head cooling and mild hypo-thermia after perinatal asphyxia. Pediatrics. 2001;107:480–484

438. Pratinidhi A, Shah U, Shrotri A, Bodhani N. Risk approach strategy inneonatal care. Bull World Health Organ. 1986;64:291–297

439. Bang A, Bang R, Baitule S, Reddy M, Deshmukh M. Effect of home-based neonatal care and management of sepsis on neonatal mortality:field trial in rural India. Lancet. 1999;354:1955–1961

440. Kumar R. Effect of training on the resuscitation practices of traditionalbirth attendants. Trans R Soc Trop Med Hyg. 1994;88:159–160

441. Daga S, Daga A, Dighole R, Patil R, Dhinde H. Rural neonatal care:Dahanu experience. Indian Pediatr. 1992;29:189–193

442. Kumar R, Aggarwal A. Body temperatures of home delivered new-borns in North India. Trop Doct. 1998;28:134–136

443. Massawe A, Kilewo C, Irani S, et al. Assessment of mouth-to-maskventilation in resuscitation of asphyxic newborn babies. A pilot study.Trop Med Int Health. 1996;1:865–873

444. Kamenir S. Neonatal resuscitation and newborn outcomes in ruralKenya. J Trop Pediatr. 1997;43:170–173

445. Saugstad O, Rootwelt T, Aalen O. Resuscitation of asphyxiated new-born infants with room air or oxygen: an international controlled trial:the Resair 2 study. Pediatrics. 1998;102(1). Available at: www.pediatrics.org/cgi/content/full/102/1/e1

446. Ramji S, Ahuja S, Thirupuram S, Rootwelt T, Rooth G, Saugstad O.Resuscitation of asphyxic newborn infants with room air or 100%oxygen. Pediatr Res. 1993;34:809–812

447. Deorari A, Paul V, Singh M, Vidyasagar D. The national movement ofneonatal resuscitation in India. J Trop Pediatr. 2000;46:315–317

448. Zhu X, Fang H, Zeng S, Li Y, Lin H, Shi S. The impact of the neonatalresuscitation program guidelines (NRPG) on the neonatal mortality ina hospital in Zhuhai, China. Singapore Med J. 1997;38:485–487

449. Daga S, Daga A, Dighole R, Patil R. Anganwadi worker’s participationin rural newborn care. Indian J Pediatr. 1993;60:627–630

450. Xiaoyu Z. Neonatal resuscitation. World Health Forum. 1993;14:289–290451. Yao A, Moinian M, Lind J. Distribution of blood between infant and

placenta after birth. Lancet. 1969;2(7626):871–873

452. Nelle M, Zilow E, Bastert G, Linderkamp O. Effect of Leboyer child-birth on cardiac output, cerebral and gastrointestinal blood flow ve-locities in full term neonates. Am J Perinatol. 1995;12:212–216

453. Pisacane A. Neonatal prevention of iron deficiency. BMJ. 1996;312:136–137

454. Grajeda R, Perez-Escamilla R, Dewey K. Delayed clamping of theumbilical cord improves hematological status of Guatemalian infantsat 2 months of age. Am J Clin Nutr. 1997;65:425–431

455. Darmstadt G, Dinulos J. Neonatal skin care. Pediatr Clin North Am.2000;47:757–782

456. Medves J, O’Brien B. Cleaning solutions and bacterial colonization inpromoting healing and early separation of the umbilical cord inhealthy newborns. Can J Public Health. 1997;88:380–382

457. Dore S, Buchan D, Coulas S, et al. Alcohol versus natural drying fornewborn cord care. J Obstet Gynecol Neonatal Nurs. 1998;27:621–627

458. Allen K, Ridgway E, Parsons L. Hexachlorophene powder and neona-tal staphylococcal infection. J Hosp Infect. 1994;27:29–33

459. Paes B, Jones C. An audit of the effect of two cord-care regimens onbacterial colonization in newborn infants. QRB Qual Rev Bull. 1987;13:109–113

460. Zupan J, Garner P. Topical umbilical cord care at birth. CochraneDatabase Syst Rev. 2000:CD001057

461. World Health Organization. Care of the Umbilical Cord: A Review of theEvidence. Geneva, Switzerland: World Health Organization; 1998

462. Mull D, Anderson J, Mull J. Cow dung, rock salt, and medical inno-vation in the Hindu Kush of Pakistan: the cultural transformation ofneonatal tetanus and iodine deficiency. Soc Sci Med. 1990;30:675–691

463. Garner P, Lam D, Baea M, Edwards K, Heywood P. Avoiding neonataldeath: an intervention study of umbilical cord care. J Trop Pediatr.1994;40:24–28

464. Bennett J, Azhar N, Rahim F, et al. Further observations on ghee as arisk factor for neonatal tetanus. Int J Epidemiol. 1995;24:643–647

465. Rush J. Rooming-in and visiting on the ward: effects on newborncolonization rates. Infect Control. 1987;2(suppl 3):10–15

466. Montgomery T, Wise R, Lang W, Mandle R, Fritz M. A study ofstaphylococcal colonization of postpartum mothers and newborn in-fants. Comparison of central care and rooming-in. Am J Obstet Gynecol.1959;78:1227–1233

467. World Health Organization. Thermal Control of the Newborn: A PracticalGuide. Geneva, Switzerland: World Health Organization; 1993

468. Tunell R. Hypothermia: epidemiology and prevention. In: Costello A,Manandhar D, eds. Improving Newborn Health in Developing Countries.London, United Kingdom: Imperial College Press; 2000:207–220

469. World Health Organization. Thermal Protection of the Newborn: A Prac-tical Guide. Geneva, Switzerland: World Health Organization; 1997

470. Mann T. Hypothermia in the newborn: a new syndrome? Lancet.1955;1:613–614

471. Elliott RI, Mann TP. Neonatal cold injury due to accidental exposure tocold. Lancet. 1957;272(6962):229–234

472. Ellis M, Manandhar N, Shakya U, Manadhar D, Fawdry A, Costello A.Postnatal hypothermia and cold stress among newborn infants inNepal monitored by continuous ambulatory recording. Arch Dis ChildFetal Neonatal Ed. 1996;75:f42–f45

473. Johanson R, Spencer S, Rolfe P, Jones P, Malla D. Effect of post-deliverycare on neonatal body temperature. Acta Paediatr. 1992;81:859–863

474. Tafari N, Gentz J. Aspects on rewarming newborn infants. In: Sterky G,Tafari N, Tunnel R, eds. Breathing and Warmth at Birth. Sarec ReportNo. R2. Montreal, QC, Canada: Sarec; 1985:53–58

475. Christensson K, Bhat GJ, Amadi BC, Eriksson B, Hojer B. Randomisedstudy of skin-to-skin versus incubator care for rewarming low-riskhypothermic neonates. Lancet. 1998;352:1115

476. Kambarami R, Chidede O, Pereira N. ThermoSpot in the detection ofneonatal hypothermia. Ann Trop Paediatr. 2002;22:219–223

477. Dragovich D, Tamburlini G, Alisjahbana A, et al. Thermal control ofthe newborn: knowledge and practice of health professional in sevencountries. Acta Paediatr. 1997;86:645–650

478. Johanson R. Diagnosis of hypothermia—a simple test? J Trop Pediatr.1993;39:313–314

479. Singh M, Rao G, Malhotra A, Deorari A. Assessment of newbornbaby’s temperature by human touch: a potentially useful primary carestrategy. Indian Pediatr. 1992;29:449–452

480. Kumar R, Aggarwal A. Accuracy of maternal perception of neonataltemperature. Indian Pediatr. 1996;33:583–585

481. Morley D, Blumenthal I. A neonatal hypothermia indicator. Lancet.2000;355(9204):659–660

482. Morley D, Kennedy N. Hypothermia: prevention at community level.Trop Doct. 2002;32:22–24



483. Kennedy N, Gondwe L, Morley D. Temperature monitoring withThermoSpots in Malawi. Lancet. 2000;355(9212):1364

484. Pejavar RK, Nisalga R, Gowda B. Temperature monitoring in new-borns using ThermoSpot. Indian J Pediatr. 2004;71:795–796

485. Sarman I, Can G, Tunell R. Rewarming preterm infants on a heated,water filled mattress. Arch Dis Child. 1989;64:687–692

486. Tafari N, Gentz J. Aspects of rewarming newborn infants with severeaccidental hypothermia. Acta Paediatr. 1974;63:595–600

487. Rutter N, Hull D. Reduction of skin water loss in the newborn. I. Effectof applying topical agents. Arch Dis Child. 1981;56:669–672

488. Darmstadt GL, Saha SK. Traditional practice of oil massage of neonatesin Bangladesh. J Health Popul Nutr. 2002;20:184–188

489. Brice JE, Rutter N, Hull D. Reduction of skin water loss in the newborn.II. Clinical trial of two methods in very low birthweight babies. ArchDis Child. 1981;56:673–675

490. Darmstadt G, Mao-Qiang M, Chi E, et al. Impact of topical oils on theskin barrier: possible implications for neonatal health in developingcountries. Acta Paediatr. 2002;91:1–9

491. Fernandez A, Patkar S, Chawla C, Taskar T, Prabhu S. Oil applicationin preterm babies, a source of warmth and nutrition. Indian Pediatr.1987;24:1111–1117

492. Daga S, Dighole R, Patil R. Managing very low birth weight babies athome in a rural area. World Health Forum. 1996;17:289–290

493. Gosavi D, Swaminathan M, Daga S. Appropriate technology in trans-portation of sick newborns in developing countries. Trop Doct. 1998;28:101–102

494. Daga S. Reduction in neonatal mortality by simple interventions. J Bio-soc Sci Suppl. 1989;10:127–136

495. Daga S, Daga A. Reduction in neonatal mortality with simple inter-ventions. J Trop Pediatr. 1989;35:191–196

496. Hume R, Burchell A. Abnormal expression of glucose-6-phosphatasein preterm infants. Arch Dis Child. 1993;68:202–204

497. Cornblath M, Odell G, Levin E. Symptomatic neonatal hypoglycemiaassociated with toxemia of pregnancy. J Pediatr. 1959;55:545–562

498. Williams A. Hypoglycaemia of the newborn: a review of the literature.Bull World Health Organ. 1997;75:261–290

499. Anderson S, Shakya K, Shrestha L, Costello A. Hypoglycemia: a com-mon problem among uncomplicated newborn infants in Nepal. J TropPediatr. 1993;39:273–277

500. Costello A, Pal D. Neonatal hypoglycaemia. In: Costello A, ManandharD, eds. Improving Newborn Infant Health in Developing Countries.Singapore: World Scientific Publications; 2000:347–358

501. Hawdon JM, Ward Platt MP, Aynsley-Green A. Patterns of metabolicadaptation for preterm and term infants in the first neonatal week.Arch Dis Child. 1992;67:357–365

502. Lucas A, Boyes S, Bloom S, Aynsley-Green A. Metabolic and endocrineresponses to a milk feed in six-day-old term infants: differences be-tween breast and cow’s milk formula feeding. Acta Paediatr. 1981;70:195–200

503. Lucas A. AIDS and human milk bank closures. Lancet. 1987;1(8541):1092–1093

504. Lang S, Lawrence CJ, Orme RL. Cup feeding: an alternative method ofinfant feeding. Arch Dis Child. 1994;71:365–369

505. Narayanan I, Prakash K, Murthy NS, Gujral VV. Randomised con-trolled trial of effect of raw and holder pasteurized human milk and offormula supplements on incidence of neonatal infection. Lancet. 1984;2(8412):1111–1113

506. Singhal P, Singh M, Paul V, Malhotra A, Deorari A, Ghorpade M. Acontrolled study of sugar-fortified milk feeding for prevention ofneonatal hypoglycaemia. Indian J Med Res. 1991;94:342–345

507. Singhal PK, Singh M, Paul VK, et al. Prevention of hypoglycemia: acontrolled evaluation of sugar fortified milk feeding in small-for-gestational age infants. Indian Pediatr. 1992;29:1365–1369

508. Costalos C, Russell G, Al-Rahim Q, Tarlow M, Lloyd D, Ross K. Effectson plasma glucose concentration of light-for-date infants and infants ofdiabetic mothers of feeds supplemented with a glucose polymer. ActaPaediatr. 1985;74:382–385

509. Sann L, Divry P, Lasne Y, Ruitton A. Effect of oral lipid administrationon glucose homeostasis in small-for-gestational age infants. Acta Pae-diatr. 1982;71:923–927

510. Sann L, Mousson B, Rousson M, Maire I, Bethenod M. Prevention ofneonatal hypoglycaemia by oral lipid supplementation in low birthweight infants. Eur J Pediatr. 1988;147:158–161

511. Huffman S, Zehner E, Victora C. Can improvements in breast-feedingpractices reduce neonatal mortality in developing countries? Mid-wifery. 2001;17:80–92

512. Kramer M. High protein supplementation in pregnancy (CochraneReview). Oxford, United Kingdom: Update Software; 2001a

513. Lopez-Alarcon M, Villalpando S, Fajardo A. Breast-feeding lowers thefrequency and duration of acute respiratory infection and diarrhea ininfants under six months of age. J Nutr. 1997;127:436–443

514. Zaman K, Baqui A, Yunus M, Bateman O, Chowdhury H, Black R.Acute respiratory infections in children: a community-based longitu-dinal study in rural Bangladesh. J Trop Pediatr. 1997;43:133–137

515. Leach A, McArdle T, Banya W, et al. Neonatal mortality in a rural areaof the Gambia. Ann Trop Paediatr. 1999;19:33–43

516. Raisler J, Alexander C, O’Compo P. Breastfeeding and infant illness: adose-response relationship? Am J Public Health. 1999;89:25–30

517. Perera B, Ganesan S, Jayarasa J, Ranaweera S. The impact of breast-feeding practices on respiratory and diarrhoeal disease in infancy: astudy from Sri Lanka. J Trop Pediatr. 1999;45:115–118

518. Cesar J, Victora C, Barros F, Santos I, Flores J. Impact of breastfeedingon admission for pneumonia during postnatal period in Brazil: nestedcase-control study. BMJ. 1999;318:1316–1320

519. Arifeen S, Black RE, Antelman G, Baqui A, Caulfield L, Becker S.Exclusive breastfeeding reduces acute respiratory infection and diar-rhea deaths among infants in Dhaka slums. Pediatrics. 2001;108(4).Available at: www.pediatrics.org/cgi/content/full/108/4/e67

520. Bhutta Z, Yusuf K. Early-onset neonatal sepsis in Pakistan: a case-control study of risk factors in a birth cohort. Am J Perinatol. 1997;14:579–583

521. Ashraf RN, Jalil F, Zaman S, et al. Breastfeeding and protection againstneonatal sepsis in a high-risk population. Arch Dis Child. 1991;66:488–490

522. Clavano N. Mode of feeding and its effect on infant mortality andmorbidity. J Trop Pediatr. 1982;28:287–293

523. Victora CG, Smith PG, Vaughan JP, et al. Evidence for protection bybreast-feeding against infant deaths from infectious diseases in Brazil.Lancet. 1987;2(8554):319–321

524. Augustine T, Bhatia B. Early neonatal morbidity and mortality patternin hospitalized children. Indian J Matern Child Health. 1994;5:17–19

525. Kumar P, Nangia S, Saili A, Dutta A. Growth and morbidity patternsof exclusively breast-fed preterm babies. Indian Pediatr. 1999;36:296–300

526. Singh K, Srivastava P. The effect of colostrum on infant mortality:urban rural differentials. Health Pop. 1992;15:94–100

527. Habicht JP, DaVanzo J, Butz WP. Does breastfeeding really save lives,or are apparent benefits due to biases? Am J Epidemiol. 1986;123:279–290

528. WHO Collaborative Study Team on the Role of Breastfeeding on thePrevention of Infant Mortality. Effect of breastfeeding on infant andchild mortality due to infectious diseases in less developed countries:a pooled analysis [published correction appears in Lancet. 2000;355:1104]. Lancet. 2000;355:451–455

529. Talukder M. The importance of breastfeeding and strategies to sustainhigh breastfeeding rates. In: Costello A, Manandhar D, eds. ImprovingNewborn Infant Health in Developing Countries. Singapore: World Scien-tific Publications; 2000:309–342

530. Jalil F. Community-based strategies in improve newborn and infantcare practices. In: Costello A, Manandhar D, eds. Improving NewbornInfant Health in Developing Countries. Singapore: World ScientificPublications; 2000:409–423

531. Haider R, Kabir I, Ashworth A. Are breastfeeding promotion messagesinfluencing mothers in Bangladesh? Results from an urban survey inDhaka, Bangladesh. J Trop Pediatr. 1999;45:315–318

532. Haider R, Ashworth A, Kabir I, Huttly S. Effect of community-basedpeer counselors on exclusive breastfeeding practices in Dhaka,Bangladesh: a randomized controlled trial. Lancet. 2000;356:1643–1647

533. Adair L, Popkin B. Low birth weight reduces the likelihood of breast-feeding among Filipino infants. J Nutr. 1996;126:103–112

534. Haider R, Kabir I, Fuchs G, Habte D. Neonatal diarrhea in the diarrheatreatment center in Bangladesh: clinical presentation, breastfeedingmanagement and outcome. Indian Pediatr. 2000b;37:37–43

535. Bonuck K, Arno P, Memmott M, Freeman K, Gold M, McKee D.Breast-feeding promotion interventions: good public health and eco-nomic sense. J Perinatol. 2002;22:78–81

536. Academy for Educational Development, LINKAGES Project. LINK-AGES results. Presentation at: USAID stakeholder meeting; December14, 2000; Washington, DC: Academy for Educational Development;2000

537. Bhandari N, Bahl R, Mazumdar S, et al. Effect of community-basedpromotion of exclusive breastfeeding on diarrhoeal illness and growth:a cluster randomised controlled trial. Lancet. 2003;361:1418–1423

538. Howe L. Crede’s method for prevention of purulent ophthalmia ininfancy in public institutions. Trans Am Ophthalmol Soc. 1987;8:52–57



539. Weiss A. Conjunctivitis in the neonatal period (ophthalmia neonato-rum). In: Long S, Prober C, Pickering LK, eds. Principles and Practice ofPediatric Infectious Disease. 2nd ed. New York, NY: ChurchillLivingstone; 2000:550–560

540. Fransen L, Nsanze H, Klaus V. Ophthalmia neonatorum in Nairobi,Kenya: the roles of Neiserria gonorrheae and Chlamydia trachomatis.J Infect Dis. 1986;153:862–869

541. Laga M, Plummer F, Piot P, et al. Prophylaxis of gonococcal andchlamydial ophthalmia neonatorum: a comparison of silver nitrate andtetracycline. N Engl J Med. 1988;318:653–657

542. Mathieu P. Comparison study: silver nitrate and oxytetracycline innewborn eyes. Am J Dis Child. 1958;95:609–611

543. Christian JR. Comparison of ocular reactions with the use of silvernitrate and erythromycin ointment in ophthalmia neonatorum pro-phylaxis. J Pediatr. 1960;57:55–60

544. Hammerschlag M, Chandler J, Alexander E, et al. Erythromycin oint-ment for ocular prophylaxis of neonatal chlamydial infection. JAMA.1980;244:2291–2293

545. Hammerschlag M, Cummings C, Roblin P, Williams T, Delke I. Effi-cacy of neonatal ocular prophylaxis for the prevention of chlamydialand gonococcal conjunctivitis. N Engl J Med. 1989;320:769–772

546. Greenberg M, Vandow J. Ophthalmia neonatorum: evaluation of dif-ferent methods of prophylaxis in New York City. Am J Public Health.1961;51:836–845

547. Oriel J. Ophthalmia neonatorum: relative efficacy of current prophy-lactic practices and treatment. J Antimicrob Chemother. 1984;14:209–219

548. World Health Organization. Managing Newborn Problems: A Guide forDoctors, Nurses and Midwives. Geneva, Switzerland: World HealthOrganization; 2003

549. American Academy of Pediatrics. Gonococcal infections. In: PickeringLK, ed. 2000 Red Book: Report of the Committee on Infectious Diseases. 25thed. Elk Grove Village, IL: American Academy of Pediatrics; 2000:254–260

550. Isenberg S, Apt L, Wood M. A controlled trial of povidone-iodine asprophylaxis against ophthalmia neonatorum. N Engl J Med. 1995;332:562–566

551. Puckett R, Offringa M. Prophylactic vitamin K for vitamin K deficiencybleeding in neonates. Cochrane Database Syst Rev. 2000:CD002776

552. Crowther C, Henderson-Smart D. Vitamin K prior to preterm birth forpreventing neonatal periventricular hemorrhage. Cochrane DatabaseSyst Rev. 2001;(1):CD000229

553. Ungchusak K, Tishyadhigama S, Choprapowan C, SawadiwutipongW, Varintarawat S. Incidence of idiopathic vitamin K deficiency ininfants: a national, hospital-based survey in Thailand, 1983. J MedAssoc Thai. 1988;71:417–421

554. Khanjanathiti P, Nanna P, Chairasamee H. Reduction of infant mor-bidity and mortality. Rama Med J. 1987;10:155–161

555. Victora C, Philip V. Vitamin K prophylaxis in less developed countries:policy issues and relevance to breastfeeding promotion. Am J PublicHealth. 1998;88:203–209

556. Simkiss D. Hepatitis B vaccination. J Trop Pediatr. 2002;48:256–257557. Murray C, Lopez A. Mortality by cause for eight regions of the world:

Global Burden of Disease Study. Lancet. 1997;349:1269–1276558. Chen D, Lai M. Epidemiological hepatitis B virus infection in Taiwan.

J Gastroenterol. 1984;1:1–9559. Zaki H, Darmstadt G, Baten A, Ahsan C, Saha S. Seroepidemiology of

hepatitis B and delta virus infections in Bangladesh. J Trop Pediatr.2003;49:371–374

560. Dusheiko G, Brink B, Conradie J, Marimuthu T, Sher R. Regionalprevalence of hepatitis B, delta, and human immunodeficiency virusinfection in southern Africa: a large population survey. Am J Epidemiol.1989;129:138–145

561. Wong S, Chan L, Yu V, Ho L. Hepatitis B carrier and perinatal outcomein singleton pregnancy. Am J Perinatol. 1999;16:485–488

562. Chen D, Hsu N, Sung J, et al. A mass vaccination program in Taiwanagainst hepatitis B virus infection in infants of hepatitis B surfaceantigen-carrier mothers. JAMA. 1987;257:2597–2603

563. Chen H, Chang M, Ni Y, et al. Seroepidemiology of hepatitis B virusinfection in children: ten years of mass vaccination in Taiwan. JAMA.1996;276:906–908

564. Montesano R. Hepatitis B immunization and hepatocellular carci-noma: The Gambia Hepatitis Intervention Study. J Med Virol. 2002;67:444–446

565. The Gambia Hepatitis Study Group. Hepatitis B vaccine in the ex-panded programme of immunisation: the Gambian experience. Lancet.1989;1(8646):1057–1060

566. The Gambia Hepatitis Study Group. The Gambia Hepatitis Interven-tion Study. Cancer Res. 1987;47:5782–5787

567. Kao J, Hsu H, Shau W, Chang M, Chen D. Universal hepatitis Bvaccination and the decreased mortality from fulminant hepatitis ininfants in Taiwan. J Pediatr. 2001;139:349–352

568. Schoub B, Johnson S, McAnerney J, Blackburn N, Kew M, McCutcheonJ. Integration of hepatitis B vaccination into rural African primaryhealth care programmes. BMJ. 1991;302:313–316

569. Datta N, Kumar V, Kumar L, Singhi S. Application of case manage-ment to the control of acute respiratory infections in low-birth-weightinfants: a feasibility study. Bull World Health Organ. 1987;65:77–82

570. Zoysa I, Bhandari N, Aktari N, Bhan M. Careseeking for illness inyoung infants in an urban slum in India. Soc Sci Med. 1998;47:2101–2111

571. Ahmed S, Sobhan F, Islam A, e-Khuda B. Neonatal morbidity andcare-seeking behaviour in rural Bangladesh. J Trop Pediatr. 2001;47:98–105

572. Sutanto A, Suarnawa I, Nelson C, Steward T, Soewarso T. Homedelivery of heat-stable vaccines in Indonesia: outreach immunizationwith a prefilled, single-use injection device. Bull World Health Organ.1999;77:119–126

573. Shenai J, Chytil F, Stahlman M. Vitamin A status of neonates withbronchopulmonary dysplasia. Pediatr Res. 1985;19:185–188

574. Shenai J. Vitamin A. In: Tsang R, Lucas A, Uauy R, Zlotkin S, eds.Nutritional needs of the preterm infant: scientific basis and practicalguidelines. Baltimore, MD: Williams and Wilkins; 1993:87–100

575. Atkinson S. Special nutritional needs of infants for prevention of andrecovery from bronchopulmonary dysplasia. J Nutr. 2001;131:942S–946S

576. Humphrey J, Agoestina T, Wu L, et al. Impact of neonatal vitamin Asupplementation on infant morbidity and mortality. J Pediatr. 1996;128:489–496

577. Rahmathullah L, Tielsch J, Thulasiraj R, et al. Impact of supplementingnewborn infants with vitamin A on early infant mortality: communitybased randomised trial in southern India. BMJ. 2003;327:254

578. West K, Katz J, Shrestha R, et al. Mortality of infants �6 mo of agesupplemented with vitamin A: a randomized, double-masked trial inNepal. Am J Clin Nutr. 1995;62:143–148

579. Rey M. Manejo racional del nino prematuro [Rational management ofthe premature infant]. In: I Curso de Medicina Fetal y Neonatal. Bogota,Colombia; 1983:137–151

580. Whitelaw A, Sleath K. Myth of the marsupial mother: home care ofvery low birth weight babies in Bogota, Colombia. Lancet. 1985;1(8439):1206–1208

581. Conde-Agudelo A, Diaz-Rossello J, Belizan J. Kangaroo mother care toreduce morbidity and mortality in low birthweight infants (CochraneReview). Oxford, United Kingdom: Update Software; 2002

582. Charpak N, Ruiz-Pelaez J, Figueroa C, Charpak Y. Kangaroo motherversus traditional care for newborn infants �2000 grams: a random-ized, controlled trial. Pediatrics. 1997;100:682–688

583. Charpak N, Ruiz-Pelaez J, Figueroa C, Charpak Y. A randomized,controlled trial of kangaroo mother care: results of follow-up at 1 yearof corrected age. Pediatrics. 2001;108:1072–1079

584. Charpak N, Ruiz-Pelaez J, Charpak Y. Rey-Martinez Kangaroo MotherProgram: an alternative way of caring for low birth weight infants?One year mortality in a two cohort study. Pediatrics. 1994;94:804–810

585. Bergman N, Jurisoo L. The “kangaroo-method” for treating low birthweight babies in a developing country. Trop Doct. 1994;24:57–60

586. Lincetto O, Nazir A, Cattaneo A. Kangaroo mother care with limitedresources. J Trop Pediatr. 2000;46:293–295

587. Anderson G. Current knowledge about skin-to-skin (kangaroo) carefor preterm infants. J Perinatol. 1991;11:216–226

588. Ramanathan K, Paul V, Deorari A, Taneja U, George G. Kangaroomother care in very low weight infants. Indian J Pediatr. 2001;68:1019–1023

589. Quasem I, Sloan N, Chowdhury A, Ahmed S, Winikoff B, ChowdhuryA. Adaptation of kangaroo mother care for community-based appli-cation. J Perinatol. 2003;23:646–651

590. Kumar V, Yadav R, Darmstadt G. Lessons from implementation ofskin-to-skin care in rural India for the development of strategies forcommunity KMC. Presented at: First International Congress of Kan-garoo Mother Care; November 10, 2004; Rio de Janeiro, Brazil

591. Sachdev H. Kangaroo mother care method to reduce morbidity andmortality in low-birth-weight infants [commentary]. Geneva, Swit-zerland: World Health Organization; 2002

592. Christensson K. Fathers can effectively achieve heat conservation inhealthy newborn infants. Acta Paediatr. 1996;85:1354–1360

593. Yoshio H, Tollin M, Gudmundsson G, et al. Antimicrobial polypep-tides of human vernix caseosa and amniotic fluid: implications fornewborn innate defense. Pediatr Res. 2003;53:211–216



594. Pabst R, Starr K, Qaiyumi S, Schwalbe R, Gewolb I. The effect ofapplication of Aquaphor on skin condition, fluid requirements, andbacterial colonization in very low birth weight infants. J Perinatol.1999;19:278–283

595. Darmstadt G. The skin and nutritional disorders of the newborn. EurJ Pediatr Dermatol. 1998;8:221–228

596. Darmstadt G, Saha S, Ahmed A, Khatun M, Chowdhury M. The skinas a potential portal of entry for invasive infections in neonates.Perinatology. 2003;5:205–212

597. Prottey C, Hartop P, Press M. Correction of the cutaneous manifesta-tions of essential fatty acid deficiency in man by application of sun-flower seed oil to the skin. J Invest Dermatol. 1975;64:228–234

598. Prottey C, Hartop P, Black J, McCormack J. The repair of impairedepidermal barrier function in rats by the cutaneous application oflinoleic acid. Br J Dermatol. 1976;84:13–21

599. Houtsmuller U, van der Beek A. Effects of topical application of fattyacids. Prog Lipid Res. 1981;20:219–224

600. Elias P, Brown B, Ziboh V. The permeability barrier in essential fattyacid deficiency: evidence for a direct role for linoleic acid in barrierfunction. J Invest Dermatol. 1980;74:230–233

601. Press M, Hartop P, Prottey C. Correction of essential fatty-acid defi-ciency in man by the cutaneous application of sunflower-seed oil.Lancet. 1974;1(7858):597–598

602. Friedman Z, Shochat S, Maisels M, Marks K, Lamberth E. Correction ofessential fatty acid deficiency in newborn infants by cutaneous appli-cation of sunflower-seed oil. Pediatrics. 1976;58:650–654

603. Schurer N, Schliep V, Williams M. Differential utilization of linoleicand arachidonic acid by cultured human keratinocytes. Skin Pharmacol.1995;8:30–40

604. Elias P, Mao-Qiang M, Thornfeldt C, Feingold K. The epidermal per-meability barrier: effects of physiologic and non-physiologic lipids. In:Hoppe U, ed. The Lanolin Book. Hamburg, Germany: Beiersdorf AG;1999:253–279

605. Jonas M, Darmstadt G, Rubens C, Chi E. Ultrastructure of invasion ofgroup A streptococcus into human keratinocytes. Presented at: The 14thInternational Congress of Electron Microscopy; August 31–September 4,1998; Cancun, Mexico

606. Mullany L, Darmstadt G, Khatry S, Tielsch J. Traditional massage ofnewborns in Nepal: implications for trials of improved practice. J TropPediatr. 2005: In press

607. Lane A, Drost S. Effects of repeated application of emollient cream topremature neonates’ skin. Pediatrics. 1993;92:415–419

608. Nopper A, Horii K, Sookdeo-Drost S, Wang T, Mancini A, Lane A.Topical ointment therapy benefits premature infants. J Pediatr. 1996;128:660–669

609. Edwards W, Conner J, Soll R. The effect of Aquaphor original emol-lient ointment on nosocomial sepsis rates and skin integrity in infantsof birth weight 501 to 1000 grams [abstract]. Pediatr Res. 2001;49(4suppl):2330

610. Wallace M, Lindado S, Bedrick A, Moravec C, Nieto S. Decreasingbloodstream infection rates in very low birth weight infants withtopical ointment therapy [abstract]. Infect Control Hosp Epidemiol. 1998;19:689

611. Campbell J, Zaccaria E, Baker C. Systemic candidiasis in extremely lowbirth weight infants receiving topical petrolatum ointment for skincare: a case-control study. Pediatrics. 2000;105:1041–1045

612. Ramsey K, Malone S, Fey P, et al. Aquaphor as a source of colonizationand subsequent blood stream infections among very low birthweightneonates[abstract 53]. Infect Control Hosp Epidemiol. 1998;19:689

613. Edwards W, Conner J, Soll R. The effect of prophylactic ointmenttherapy on nosocomial sepsis rates and skin integrity in infants withbirth weights of 501 to 1000 g. Pediatrics. 2004;113:1195–1203

614. Conner J, Soll R, Edwards W. Topical ointment for preventing infectionin preterm infants (Cochrane Review). In: The Cochrane Library. Chich-ester, United Kingdom: John Wiley & Sons, Ltd; 2004

615. Sood N, Sachdev M, Mohan M, Gupta S, Sachdev H. Epidemic dropsyfollowing transcutaneous absorption of Argemone mexicana oil. Trans RSoc Trop Med Hyg. 1985;79:510–512

616. Singh N, Anuradha S, Dhanwal D, et al. Epidemic dropsy—a clinicalstudy of the Delhi outbreak. J Assoc Physicians India. 2000;48:877–880

617. Agarwal K, Gupta A, Pushkarna R, Bhargava S, Faridi M, Prabhu M.Effect of massage and use of oil on growth, blood flow and sleeppattern in infants. Indian J Med Res. 2000;112:212–217

618. Ahmed A, Saha S, Chowdhury M, Law P, Ahmed S, Darmstadt G.Topical skin barrier therapy and infection control measures improvesurvival of hospitalized very low birth weight neonates. In: Proceedingsof the 10th Asian Conference on Diarrhoeal Diseases and Nutrition. Dhaka,Bangladesh: ICDDR, B; 2004

619. Mao-Qiang M, Feingold K, Thornfeldt C, Elias P. Optimization ofphysiological lipid mixtures for barrier repair. J Invest Dermatol. 1996;106:1096–1101

620. Maisels M, Gifford K. Normal serum bilirubin levels in the newbornand the effect of breast-feeding. Pediatrics. 1986;78:837–843

621. Cashore W. The neurotoxicity of bilirubin. Clin Perinatol. 1990;17:437–447

622. Connolly A, Volpe J. Clinical features of bilirubin encephalopathy. ClinPerinatol. 1990;17:371–380

623. Arif K, Bhutta Z. Risk factors and spectrum of neonatal jaundice in abirth cohort in Karachi. Indian Pediatr. 1999;36:487–493

624. Maisels M. Neonatal jaundice. In: Sinclair J, Bracken M, eds. A guide toeffective care in pregnancy and childbirth. Oxford, United Kingdom:Oxford University Press; 1992

625. De Carvalho M, Klaus M, Merkatz R. The frequency of breastfeedingand serum bilirubin concentration. Am J Dis Child. 1982;136:737–738

626. Yamauchi Y, Yamanouchi I. Breastfeeding frequency during the first24 hours after birth in full-term neonates. Pediatrics. 1990;86:171–175

627. Kumar A, Faridi M, Singh N, Ahmad S. Transcutaneous bilirubinom-etry in the management of bilirubinemia in term neonates. Indian J MedRes. 1994;99:227–230

628. Tayaba R, Gribetz D, Gribetz I, Holzman I. Noninvasive estimation ofserum bilirubin. Pediatrics. 1998;102(3). Available at: www.pediatrics.org/cgi/content/full/102/3/e28

629. Bhutta Z, Yusuf K. Transcutaneous bilirubinometry in Pakistani new-born infants: a preliminary report. J Pak Med Assoc. 1991;41:155–156

630. Narayanan I, Banwalikar J, Mehta R, et al. A simple method of eval-uation of jaundice in the newborn. Ann Trop Paediatr. 1990;10:31–34

631. Madlon-Kay D. Home health nurse clinical assessment of neonataljaundice: comparison of 3 methods. Arch Pediatr Adolesc Med. 2001;155:583–586

632. Bilgen H, Ince Z, Ozek E, Bekiroglu N, Ors R. Transcutaneous mea-surement of hyperbilirubinaemia: comparison of the Minolta jaundicemeter and the Ingram icterometer. Ann Trop Paediatr. 1998;18:325–328

633. Riskin A, Kuglman A, Abend-Weinger M, Green M, Hemo M, BaderD. In the eye of the beholder: how accurate is clinical estimation ofjaundice in newborns? Acta Paediatr. 2003;92:574–576

634. Ebbeson F. The relationship between the cephalo-pedal progress ofclinical icterus and the serum bilirubin concentration in newborninfants without blood type sensitization. Acta Obstet Gynecol Scand.1975;54:329–332

635. Madlon-Kay D. Recognition of the presence and severity of newbornjaundice by parents, nurses, physicians, and icterometer. Pediatrics.1997;100(3). Available at: www.pediatrics.org/cgi/content/full/100/3/e3

636. Moyer V, Ahn C, Sneed S. Accuracy of clinical judgment in neonataljaundice. Arch Pediatr Adolesc Med. 2000;154:391–394

637. Akpala C. An evaluation of the knowledge and practices of trainedtraditional birth attendants in Bodinga, Sokoto State, Nigeria. J TropMed Hyg. 1994;97:46–50

638. Begum J, Kabir I, Mollah A. The impact of training traditional birthattendants in improving MCH care in rural Bangladesh. Asia Pac JPublic Health. 1990;4:142–144

639. Kumar R. Training traditional birth attendants for resuscitation ofnewborns. Trop Doct. 1995;25:25–30

640. Kumar R, Thakur J, Aggarwal A. Effect of continuing training onknowledge and practices of traditional birth attendants about maternaland newborn care. Indian J Public Health. 2000;44:118–123

641. Daga A, Daga S, Dighole R, Patil R, Patil M. Evaluation of trainingprogramme for traditional birth attendants in newborn care. IndianPediatr. 1997;34:1021–1024

642. O’Rourke K. The effect of hospital staff training on management ofobstetrical patients referred by traditional birth attendants. Int J Gynae-col Obstet. 1995;(48 suppl):S95–S102

643. Rashid M, Tayakkanonta K, Chongsuvivatwong V, Geater A, BechtelG. Traditional birth attendants’ advice toward breast-feeding, immu-nization and oral rehydration among mothers in rural Bangladesh.Women Health. 1999;28(3):33–44

644. Greenwood A, Bradley A, Byass P, et al. Evaluation of a primary healthcare programme in The Gambia. I. The impact of trained traditionalbirth attendants on the outcome of pregnancy. J Trop Med Hyg. 1990;93:58–66

645. Taylor C, Kielmann A, DeSweemer C, et al. The Narangwal experi-ment on interactions of nutrition and infections: I. Project design andeffects upon growth. Indian J Med Res. 1978;(68 suppl):1–20

646. Bang A, Bang R, Baitule S, Deshmukh M, Reddy M. Burden of mor-bidities and the unmet need for health care in rural neonates—a



prospective observational study in Gadchiroli, India. Indian Pediatr.2001;38:952–965

647. Janowitz B, Bailey P, Dominik R, Araujo L. TBAs in rural northeastBrazil: referral patterns and perinatal mortality. Health Policy Plan.1988;3:48–58

648. Smith J, Coleman N, Fortney J, Johnson J, Blumhagen D, Grey T. Theimpact of traditional birth attendant training on delivery complica-tions in Ghana. Health Policy Plan. 2000;15:326–331

649. World Health Organization. Programme of Acute Respiratory Infections.Report of the fourth meeting of technical advisory group 6–10 March1989. Geneva, Switzerland: World Health Organization; 1989

650. World Health Organization. Acute Respiratory Infections in Children:Case Management in Hospitals in Developing Countries. Geneva,Switzerland: World Health Organization; 1990

651. Sazawal S, Black R. Meta-analysis of intervention trials on case man-agement of pneumonia in community settings. Lancet. 1992;340:528–533

652. Sazawal S, Black R, Pneumonia Case Management Trials Group. Effectof pneumonia case management on mortality in neonates, infants, andpreschool children: a meta-analysis of community-based trials. LancetInfect Dis. 2003;3:547–556

653. Mtango F, Neuvians D. Acute respiratory infections in children underfive years. Control project in Bagamoyo District, Tanzania. Trans R SocTrop Med Hyg. 1986;80:851–858

654. Khan A, Khan J, Akbar M, Addiss D. Acute respiratory infections inchildren: a case management intervention in Abbottabad District, Pa-kistan. Bull World Health Organ. 1990;68:577–585

655. Khan A. Case-management of acute respiratory infection in children ofAbbottabad District, Pakistan: an intervention study. Bull Int UnionTuberc Lung Dis. 1990;65(4):25–28

656. Bang AT, Bang RA, Tale O, et al. Reduction in pneumonia mortalityand total childhood mortality by means of community-based interven-tion trial in Gadchiroli, India. Lancet. 1990;336:201–206

657. Bang AT, Bang RA, Morankar VP, Sontakke PG, Solanki JM. Pneumo-nia in neonates: can it be managed in the community? Arch Dis Child.1993;68:550–556

658. Bang AT, Bang RA, Sontakke PG. Management of childhood pneumo-nia by traditional birth attendants. The SEARCH Team. Bull WorldHealth Organ. 1994;72(6):897–905

659. Pandey MR, Daulaire NM, Starbuck ES, Houston RM, McPherson K.Reduction in total under-five mortality in western Nepal throughcommunity-based antimicrobial treatment of pneumonia. Lancet. 1991;338:993–997

660. Fauveau V, Stewart M, Chakraborty J, Khan S. Impact on mortality ofa community-based programme to control acute lower respiratorytract infections. Bull World Health Organ. 1992;70:109–116

661. Roesin R, Sutanto A, Sastra K, Winarti. ARI intervention study inKediri, Indonesia (a summary of study results). Bull Int Union TubercLung Dis. 1990;65(4):23

662. Sutrisna B, Frerichs R, Reingold A. Randomized, controlled trial ofeffectiveness of ampicillin in mild acute respiratory infections in In-donesian children. Lancet. 1991;338:471–474

663. Cherian T, John TJ, Simoes E, Steinhoff MC, John M. Evaluation ofsimple clinical signs for the diagnosis of acute lower respiratory tractinfection. Lancet. 1988;2(8603):125–128

664. Bartlett AV, Paz de Bocaletti ME, Bocaletti MA. Neonatal and earlypostneonatal morbidity and mortality in a rural Guatemalancommunity: the importance of infectious diseases and their manage-ment. Pediatr Infect Dis J. 1991;10:752–757

665. Manandhar DS, Osrin D, Shrestha BP, et al. Effect of a participatoryintervention with women’s groups on birth outcomes in Nepal: clus-ter-randomized controlled trial. Lancet. 2004;364:970–979

666. Bose A, Sinha S, Choudhary N, Aruldas K, Moses PD, Joseph A.Experiences of neonatal care in a secondary level hospital. IndianPediatr. 1999;36:802–806

667. van der Mei J. Survival chances of low birth weight infants in a ruralhospital in Ghana. Trop Geogr Med. 1994;46:313–317

668. Wilkinson D. Perinatal mortality—an interventional study. S Afr MedJ. 1991;79:552–553

669. Dutt D, Srinivasa D. Impact of maternal and child health strategy onchild survival in a rural community of Pondicherry. Indian Pediatr.1997;34:785–792

670. World Health Organization. Pocket Book of Hospital Care for Children:Guidelines for the Management of Common Illnesses With Limited Re-sources. Geneva, Switzerland: World Health Organization; 2004

671. Brosnan CA, Swint JM. Cost analysis: concepts and application. PublicHealth Nurs. 2001;18:13–18

672. Walsh JA, Measham AR, Feifer CN, Gertler PJ. The impact of maternalhealth improvement on perinatal survival: cost-effective alternatives.Int J Health Plann Manage. 1994;9:131–149

673. Hutubessy R, Bendib L, Evans D. Critical issues in the economicevaluation of interventions against communicable diseases. ActaTropica. 2001;78:191–206

674. Joyce T, Corman H, Grossman M. A cost-effectiveness analysis ofstrategies to reduce infant mortality. Med Care. 1988;26:348–360

675. Ross MG, Sandhu M, Bemis R, Nessim S, Bragonier JR, Hobel C. TheWest Los Angeles Preterm Birth Prevention Project: II. Cost-effectiveness analysis of high-risk pregnancy interventions. Obstet Gy-necol. 1994;83:506–511

676. Kitzmiller JL, Elixhauser A, Carr S, et al. Assessment of costs andbenefits of management of gestational diabetes mellitus. Diabetes Care.1998;21(suppl 2):B123–B130

677. Pollack L. An effective model for reorganization of perinatal services ina metropolitan area: a descriptive analysis and historical perspective. JPerinatol. 1996;16:3–8

678. Dye TD, Wojtowycz MA, Aubry RH. A cost evaluation of implement-ing a quality-oriented, regional perinatal data system. J Public HealthManag Pract. 1997;3(2):37–40

679. Chapalain MT. Perinatality: French cost-benefit studies and decisionson handicap and prevention. Ciba Found Symp. 1978;59:193–206

680. Moya MP, Goldberg RN. Cost-effectiveness of prophylactic indometh-acin in very-low-birth-weight infants. Ann Pharmacother. 2002;36:218–224

681. Phibbs C, Phibbs R, Wakeley A, Schlueter M, Sniderman S, Tooley W.Cost effects of surfactant therapy for neonatal respiratory distresssyndrome. J Pediatr. 1993;123:953–962

682. Stevens TP, Sinkin RA, Hall CB, Maniscalco WM, McConnochie KM.Respiratory syncytial virus and premature infants born at 32 weeks’gestation or earlier: hospitalization and economic implications of pro-phylaxis. Arch Pediatr Adolesc Med. 2000;154:55–61

683. Zupancic J, Richardson D, O’Brien B, Eichenwald E, Weinstein M.Cost-effectiveness analysis of predischarge monitoring for apnea ofprematurity. Pediatrics. 2003;111:146–152

684. Stembera Z. Cost-benefit in perinatology—II. Its application to condi-tions in the Czech Republic [in Czech]. Ceska Gynekol. 1998;63:269–275

685. Geerts LT, Brand EJ, Theron GB. Routine obstetric ultrasound exami-nations in South Africa: cost and effect on perinatal outcome–a pro-spective randomised controlled trial. Br J Obstet Gynaecol. 1996;103:501–507

686. Pejaver R, al Hifzi I, Aldussari S. Surfactant replacement therapy—economic impact. Indian J Pediatr. 2001;68:501–505

687. Farina D, Rodriguez S, Bauer G, et al. Respiratory syncytial virusprophylaxis: cost-effective analysis in Argentina. Pediatr Infect Dis J.2002;21:287–291

688. Mugford M, Hutton G, Fox-Rushby J. Methods for economic evalua-tion alongside a multicentre trial in developing countries: a case studyfrom the WHO Antenatal Care Randomised Controlled Trial. PaediatrPerinat Epidemiol. 1998;12(suppl 2):75–97

689. Duke T, Willie L, Mgone JM. The effect of introduction of minimalstandards of neonatal care on in-hospital mortality. P N G Med J.2000;43:127–136

690. Stray-Pedersen B. Cost-benefit analysis of a prenatal preventive pro-gramme against congenital syphilis. NIPH Ann. 1980;3(1):57–66

691. Lappalainen M, Sintonen H, Koskiniemi M, et al. Cost-benefit analysisof screening for toxoplasmosis during pregnancy. Scand J Infect Dis.1995;27:265–272

692. Stray-Pedersen B, Jenum P. Economic evaluation of preventive pro-grammes against congenital toxoplasmosis. Scand J Infect Dis Suppl.1992;84:86–96

693. Berman P, Quinley J, Yusuf B, et al. Maternal tetanus immunization inAceh Province, Sumatra: the cost-effectiveness of alternative strategies.Soc Sci Med. 1991;33:185–192

694. Goodman CA, Mills AJ. The evidence base on the cost-effectiveness ofmalaria control measures in Africa. Health Policy Plan. 1999;14(4):301–312

695. Rouse D, Andrews W, Goldenberg R, Owen J. Screening and treatmentfor asymptomatic bacteriuria of pregnancy to prevent pyelonephritis:a cost-effectiveness and cost-benefit analysis. Obstet Gynecol. 1995;86:119–123

696. Affonso DD, Korenbrot CC, De AK, Mayberry LJ. Use of care, out-comes and costs of a culturally-based perinatal program for AsianAmerican and Pacific Islander women in Hawaii. Asian Am Pac IslJ Health. 1999;7:10–24



697. Merritt TA, Raddish M. A review of guidelines for the discharge ofpremature infants: opportunities for improving cost-effectiveness. JPerinatol. 1998;18(suppl):S27–S37

698. Casiro OG, McKenzie ME, McFadyen L, et al. Earlier discharge withcommunity-based intervention for low birth weight infants: a random-ized trial. Pediatrics. 1993;92:128–134

699. Lightwood JM, Phibbs CS, Glantz SA. Short-term health and economicbenefits of smoking cessation: low birth weight. Pediatrics. 1999;104:1312–1320

700. Marks J, Koplan J, Hogue C, Dalmat M. A cost-benefit/cost-effectiveness analysis of smoking cessation for pregnant women. Am JPrev Med. 1990;6:282–289

701. Hueston WJ, Mainous AG 3rd, Farrell JB. A cost-benefit analysis ofsmoking cessation programs during the first trimester of pregnancy forthe prevention of low birthweight. J Fam Pract. 1994;39:353–357

702. Schramm W. WIC prenatal participation and its relationship to new-born Medicaid costs in Missouri: a cost/benefit analysis. Am J PublicHealth. 1985;75:851–857

703. Schramm W. Prenatal participation in WIC related to Medicaid costsfor Missouri newborns: 1982 update. Public Health Rep. 1986;101:607–615

704. Buescher PA, Larson LC, Nelson MD Jr, Lenihan AJ. Prenatal WICparticipation can reduce low birth weight and newborn medical costs:a cost-benefit analysis of WIC participation in North Carolina. J AmDiet Assoc. 1993;93:163–166

705. Schultz L, Steketee R, Chitsulo L, Wirima J. Antimalarials duringpregnancy: a cost-effectiveness analysis. Bull World Health Organ. 1995;73:207–214

706. Global Forum for Health Research. The 10/90 Report on Health Research.Geneva, Switzerland: Global Forum; 2002

707. Soucat A, Van Leberghe W, Dion F, et al. 2002 Marginal Budgeting forBottlenecks: a new costing and resource allocation practice to buyhealth results. Using health sectors budget expansion to progresstowards the Millennium Development goals in sub-Saharan Africa.Washington, DC: African Region Human Development, World Bank;2002

708. Tan-Torres Edejer T, Baltussen R, Adam T, et al, eds. Making Choices inHealth: WHO Guide to Cost-Effectiveness Analysis. Geneva, Switzerland:World Health Organization; 2003

709. World Health Organization. Mother-baby package costing spread-sheet. Geneva, Switzerland: World Health Organization; 1999

710. Bang RA, Bang AT, Reddy MH, Deshmukh MD, Baitule SB, Filippi V.Maternal morbidity during labour and the puerperium in rural homesand the need for medical attention: a prospective observational studyin Gadchiroli, India. BJOG. 2004;111:231–238

711. Claeson M, Gillespie D, Mshinda H, Troedsson H, Victoria CG; Bella-gio Study Group on Child Survival. Knowledge into action for childsurvival. Lancet. 2003;362:323–327

712. Taylor C, Taylor-Ide D. Just and Lasting Change: When Communities OwnTheir Futures. Baltimore, MD: Johns Hopkins University Press; 2002

713. Munjanja SP, Lindmark G, Nystrom L. Randomised controlled trial ofa reduced-visits programme of antenatal care in Harare, Zimbabwe.Lancet. 1996;348:364–369

714. Srinivasan V, Radhakrishna S, Sudha R, et al. Randomized controlledfield trial of two antenatal packages in rural south India. Indian J MedRes. 1995;102:86–94

715. Delong G, Leslie P, Wang S, et al. Effect on infant mortality of iodina-tion of irrigation water in the severely iodine deficient areas of China.Lancet. 1997;350:771–773

716. Dijkhuizen M, Wieringa F. Vitamin A, iron and zinc deficiency inIndonesia [dissertation]. Wageningen, Netherlands: University ofWageningen; 2001

717. Merialdi M. The Effect of Maternal Zinc Supplementation During Preg-nancy on Fetal Growth and Neurobehavior Development. Baltimore, MD:Johns Hopkins University; 2001

718. Caulfield L, Zavaleta N, Figueroa A, Leon Z. Maternal zinc supple-mentation does not affect size at birth or pregnancy duration in Peru.J Nutr. 1999;129:1563–1568

719. Garg HK, Singhal KC, Arshad Z. A study of the effect of oral zincsupplementation during pregnancy on pregnancy outcome. IndianJ Physiol Pharmacol. 1993;37:276–284

720. Cot M, Roisin A, Barro D, et al. Effect of chloroquine chemoprophy-laxis during pregnancy on birth weight: result of a randomized trial.Am J Trop Med Hyg. 1992;46:21–27

721. Expanded Programme on Immunization. Neonatal tetanus mortalitysurveys: Egypt. Wkly Epidemiol Rec. 1987;62:332–335

722. Antia NH. The Mandwa project: an experiment in community partic-ipation. Int J Health Serv. 1988;18:153–164

723. Kumar R. Effectiveness of training traditional birth attendants formanagement of asphyxia neonatorum using resuscitation equipment.Prenat Neonatal Med. 1998;3:255–260

724. Coutsoudis A, Pillay K, Spooner E, Kuhn L, Coovadia HM. Influenceof infant-feeding patterns on early mother-to-child transmission ofHIV-1 in Durban, South Africa: a prospective cohort study. SouthAfrican Vitamin A Study Group. Lancet. 1999;354:471–476

725. Odusanya OO, Alufohai JE, Meurice FP, Clemens R, Ahonkhai VI.Short term evaluation of a rural immunization program in Nigeria.J Natl Med Assoc. 2003;95:175–179

726. Ariwan I. Improving neonatal health in South-East Asia region. Reportof a regional consultation. Delhi, India; April 1–5, 2002

727. Poovorwan Y, Theamboonlers A, Vimolket T, et al. Impact of hepatitisB immunisation as part of the EPI. 2000;19:943–949

728. Cattaneo A, Davanzo R, Worku B, et al. Kangaroo mother care for lowbirthweight infants: a randomized controlled trial in different settings.Acta Paediatr. 1998;87:976–985

729. Sloan NL, Camacho LW, Rojas EP, Stern C. Kangaroo mother method:randomised controlled trial of an alternative method of care for stabi-lised low-birthweight infants. Maternidad Isidro Ayora Study Team.Lancet. 1994;344:782–785

730. Kambarami R, Chidede O, Kowo D. Kangaroo care versus incubatorcare in the management of well preterm infants—a pilot study. AnnTrop Paediatr. 1998;18:81–86

731. Lincetto O, Vos ET, Graca A, Macome C, Tallarico M, Fernandez A.Impact of season and discharge weight on complications and growthof Kangaroo Mother Care treated low birthweight infants in Mozam-bique. Acta Paediatr. 1998;87:433–439

732. Darmstadt GL, Badrawi N, Law PA, et al. Topically applied sunflowerseed oil prevents invasive bacterial infections in preterm infants inEgypt: a randomized, controlled clinical trial. Pediatr Infect Dis J. 2004;23:719–725

733. Alisjahbana A, Widjaya J, Sukadi A. A method of reporting andidentifying high-risk infants for traditional birth attendants. J TropPediatr. 1984;30:17–22

734. O’Rourke K, Howard-Grabman L, Seoane G. Impact of communityorganization of women on perinatal outcomes in rural Bolivia. RevPanam Salud Publica. 1998;3:9–14

735. Schieber B, O’Rourke K, Rodriguez C, Bartlett A. Risk factor analysis ofperi-neonatal mortality in rural Guatemala. Bull Pan Am Health Organ.1994;28:229–238

736. Arif M, Arif K. Low birthweight babies in the Third World: maternalnursing versus professional nursing care. J Trop Pediatr. 1999;45:278–280

737. Borulkar PD, Borulkar SP, Dhole RK, Daga SR. Special care for new-borns at a community hospital: a 5-year experience. Trop Doct. 1998;28:201–203

738. Bhakoo ON, Khajuria R, Desai A, Narang A. Lessons from improvedneonatal survival at Chandigarh. Indian Pediatr. 1989;26:234–240

739. Cooper PA, Rothberg AD, Pettifor JM, Bolton KD, Devenhuis S. Eval-uation of a milk formula modified specifically for very low birthweight infants. J Pediatr Gastroenterol Nutr. 1984;3:749–754

740. Goodman C, Coleman P, Mills A. The cost-effectiveness of antenatalmalaria prevention in sub-Saharan Africa. Am J Trop Med Hyg. 2001;64:45–56



DOI: 10.1542/peds.2004-1441 2005;115;519-617 Pediatrics

Zulfiqar A. Bhutta, Gary L. Darmstadt, Babar S. Hasan and Rachel A. Haws Outcomes in Developing Countries: A Review of the Evidence

Community-Based Interventions for Improving Perinatal and Neonatal Health

& ServicesUpdated Information

http://www.pediatrics.org/cgi/content/full/115/2/S1/519including high-resolution figures, can be found at:

References

http://www.pediatrics.org/cgi/content/full/115/2/S1/519#BIBLfree at: This article cites 559 articles, 159 of which you can access for

Citations

icleshttp://www.pediatrics.org/cgi/content/full/115/2/S1/519#otherartThis article has been cited by 23 HighWire-hosted articles:

Rs)3Peer Reviews (PPost-Publication

http://www.pediatrics.org/cgi/eletters/115/2/S1/519R has been posted to this article: 3One P

Subspecialty Collections

nhttp://www.pediatrics.org/cgi/collection/premature_and_newbor

Premature & Newbornfollowing collection(s): This article, along with others on similar topics, appears in the

Permissions & Licensing

http://www.pediatrics.org/misc/Permissions.shtmltables) or in its entirety can be found online at: Information about reproducing this article in parts (figures,

Reprints http://www.pediatrics.org/misc/reprints.shtml

Information about ordering reprints can be found online:

by on April 13, 2010 www.pediatrics.orgDownloaded from

http://www.pediatrics.org/cgi/content/full/115/2/S1/519

http://www.pediatrics.org/cgi/content/full/115/2/S1/519#BIBL

http://www.pediatrics.org/cgi/content/full/115/2/S1/519#otherarticles

http://www.pediatrics.org/cgi/eletters/115/2/S1/519

http://www.pediatrics.org/cgi/collection/premature_and_newborn

http://www.pediatrics.org/misc/Permissions.shtml

http://www.pediatrics.org/misc/reprints.shtml


3.2.pdf - OSSyR - Observatorio de Salud Sexual y Reproductiva

Documents

Transcript of 3.2.pdf - OSSyR - Observatorio de Salud Sexual y Reproductiva