Human Milk Feeding as a ProtectiveFactor for Retinopathy of Prematurity:A Meta-analysisJianguo Zhou, MDa, Vivek V. Shukla, MDb, Denny John, MBA, MPHc, Chao Chen, MD, PhDa

abstract CONTEXT: Studies have suggested that human milk feeding decreases the incidence of retinopathyof prematurity (ROP); however, conflicting results have been reported.

OBJECTIVE:The aim of this meta-analysis was to pool currently available data on incidence of ROPin infants fed human milk versus formula.

DATA SOURCES: Medline, PubMed, and EBSCO were searched for articles published throughFebruary 2015.

STUDY SELECTION: Longitudinal studies comparing the incidence of ROP in infants who were fedhuman milk and formula were selected. Studies involving donor milk were not included.

DATA EXTRACTION: Two independent reviewers conducted the searches and extracted data. Meta-analysis used odds ratios (ORs), and subgroup analyses were performed.

RESULTS: Five studies with 2208 preterm infants were included. Searches including variousproportions of human milk versus formula, any-stage ROP, and severe ROP were defined topool data for analyses. For any-stage ROP, the ORs (95% confidence intervals [CIs]) were asfollows: exclusive human milk versus any formula, 0.29 (0.12 to 0.72); mainly human milkversus mainly formula, 0.51 (0.26 to 1.03); any human milk versus exclusive formula, 0.54(0.15 to 1.96); and exclusive human milk versus exclusive formula, 0.25 (0.13 to 0.49). Forsevere ROP, they were 0.11 (0.04 to 0.30), 0.16 (0.06 to 0.43), 0.42 (0.08 to 2.18), and 0.10(0.04 to 0.29), respectively.

LIMITATIONS: Prospective randomized studies being impossible because of ethical issues, wechose observational studies for analysis. A few studies involving subgroup analyses presentedhigh heterogeneity.

CONCLUSIONS: Based on current limited evidence, in very preterm newborns, human milk feedingpotentially plays a protective role in preventing any-stage ROP and severe ROP.

aChildren’s Hospital of Fudan University, Shanghai, China; bThe Hospital for Sick Children, Toronto, Canada; and cPeople’s Open Access Education Initiative, Manchester, United Kingdom

Drs Zhou and Chen conceptualized the study; Drs Zhou, Shukla, and Chen planned the study; Drs Zhou and Shukla searched for articles and drafted the manuscript;Dr Chen supervised the progress of the study; Drs Zhou and Shukla and Mr John collected data; Mr John performed the meta-analysis and critically appraised themanuscript; and all authors revised the manuscript and consented to the final manuscript as submitted.

www.pediatrics.org/cgi/doi/10.1542/peds.2015-2372

DOI: 10.1542/peds.2015-2372

Accepted for publication Sep 9, 2015

Address correspondence to Chao Chen, Division of Neonatology, Children’s Hospital of Fudan University, 399 Wanyuan Road, Minhang District, Shanghai, China 201102.E-mail: chen6010@163.com

PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275).

Copyright © 2015 by the American Academy of Pediatrics

REVIEW ARTICLE PEDIATRICS Volume 136, number 6, December 2015 by guest on September 29, 2020www.aappublications.org/newsDownloaded from

Prematurity is a major contributor toglobal neonatal mortality.1 Withincreases in preterm births globally2

and major recent advances inmanagement for preterm neonates,the survival of the smallest andsickest neonates has significantlyincreased.3–8 This increase in survivalhas led to an equivalent increase inlong-term morbidities.7,9,10 Recentstudies have shown increasedincidence of retinopathy ofprematurity (ROP) in developed anddeveloping countries, such as theUnited States, Sweden, China, andTurkey.8,11–13 In extremely preterminfants with a gestational age of 22to 28 weeks, the incidence of ROPwas 59% (96% at 22 weeks and32% at 28 weeks) according to alarge US cohort study.8 In China, theincidence was .50% in infantswith birth weight ,1000 g basedon a multicenter epidemiologicstudy.12

Globally, ROP has become a leadingcause of childhood blindness inrecent times.9 ROP is a multifactorialdisease, and risk factors such as lowgestational age, oxygen therapy, andoxidative stress have been associatedwith its development.10,14 Humanmilk is recommended to all preterminfants after birth and has beenshown to be effective in preventingnecrotizing enterocolitis and late-onset sepsis.15,16 Human milk alsocontains a number of antioxidantcomponents that could be potentiallyprotective against ROP.

There has not been sufficient focus onthis subject, and scientific analysisregarding the possible beneficialeffect of breast milk on preventingROP is lacking. The data until nowhave been scattered and limited to afew clinical studies.17,18 Given theethical implications in conducting arandomized, controlled trial ofcomparing human milk feeding toformula feeding and the effects onROP, it is imperative thatobservational studies provide high-quality evidence for comparison.

We systematically reviewed theevidence from observationalstudies comparing human milk andformula feeding for preventing ROPand present the meta-analysisresults.

METHODS

Our study was undertaken toinvestigate whether human milk isprotective against any-stage ROP andsevere ROP in comparison withformula feeding. We have followedthe Meta-Analysis of ObservationalStudies in Epidemiology (MOOSE)guidelines in reporting our study.19

Systematic Search Strategy andStudy Selection

Studies published in English weresystematically identified though adatabase search of PubMed, Medline,and EBSCO from their earliestavailable dates up to February 28,

2015, using (human milk OR breastmilk) AND (retinopathy ofprematurity OR necrotizingenterocolitis) as search keywords. Weused necrotizing enterocolitis as asearch keyword to isolate studieswith ROP as secondary outcome, asthe topic of “necrotizing enterocolitisand human milk” is widely studied allover the world. In addition, wemanually searched relevant journalsrelated to pediatrics andophthalmology.

Study Selection Criteria

Two authors independentlyperformed study screening of allcitations by title and abstract in pairs.The full texts of these studies werethen retrieved, and 2 authorsindependently screened them forinclusion. In both stages,disagreements about inclusion wereresolved by discussion or byconsulting a third author.

FIGURE 1Summary of evidence search and study selection.

PEDIATRICS Volume 136, number 6, December 2015 e1577 by guest on September 29, 2020www.aappublications.org/newsDownloaded from

TABLE1

Characteristicsof

included

studies

Study

Country

StudySite

Study

Design

Sample

Size

StudyDuration

Patient

Characteristics

Feeding

Categories,n

ROPDiagnosis

GestationalAge,wk

BirthWeight,g

OtherFactorsReported

Hylander

etal,200121

usNICU,university

hospital

Cohort

174

January1992

toSeptem

ber1993

Prenatal

care,h

ealth

insurance,

maternalsm

oking,alcoholuse,

illegal

drug

use

Stages

1–4

Human

milk

28.06

2.2

1044

6251

100%

,17;

80%–90%,28;

20%–79%,39;

,20%,16

Form

ula

27.56

2.4

9486

223

74Furm

anet

al,200322

USNICU,university

hospital

Cohort

119

January1997

toFebruary

1999

Ethnicity,ventilator

dependence

Stages

1–4

Human

milk

286

2mL/kg/day:1–24,914

6205;25–49,988

6248;

.50,11636

225

mL/kg/day:1–24,

29;25–49,18;

.50,32

Form

ula

286

21103

6260

40Helleret

al,200718

USNICU,m

ulticenter

university

hospitals

Cohort

1057

October1999

toSeptem

ber2001

Pneumothorax,ethnicity,

maternalhypertension,day

offirstfeeding,human

milk

proportion,antenatalsteroids

Surgical

ROP

Human

milk

26.06

27756

134

788

Form

ula

26.26

27836

140

269

Maayan-Metzger

etal,

201223

Israel

NICU,university

hospital

Cohort

360

2006

to2008

Smallforgestationalage,gender,

multiplepregnancy,mechanical

ventilation

Stages

1–3

Human

milk

30.5(24–32)

1305

6388

.50%,188

Form

ula

31(24–32)

1425

6398

.50%,172

Manzoni

etal,201317

Italy

NICU,m

ulticenter

hospitals

Cohort

498

2004

to2008

Hyperglycemia

Human

milk

29.46

2.5

1125

6247

314

Allstages

ROP

Form

ula

29.26

2.8

1100

6272

184

ThresholdRO

P

e1578 ZHOU et al by guest on September 29, 2020www.aappublications.org/newsDownloaded from

We followed a priori study eligibilitycriteria for study selection. Any typeof observational study (cohort orcase-control) was included thatcompared human milk feeding andformula feeding. We excluded studiesthat reported only donor human milkfeeding.

We determined a priori to report thestudies on outcomes that reportedROP at any stage, including severeROP.

Data Extraction and QualityAssessment

Two authors extracted data from theincluded studies separately using astructured data extraction sheet. Thefollowing details were extracted fromeach study: authors, year ofpublication, geographical area, studysite, study design, population

(gestational age and birth weight),feeding type, ROP diagnosis, relativerisk/odds ratio (OR) and 95%confidence interval (CI), and relevantrisk factors for ROP besides feeding. Ifthe abstracted data differed betweenthe 2 authors, resolution wasconducted through discussion ordiscussion with a third author.

A critical appraisal was conducted forthe observational studies included inthe meta-analysis, using the CriticalAppraisal Skills Programme (UK)checklist, assessing the validity of theresults from each study on a scale ofhigh, medium, and satisfactory:20

high quality, the study wasprospective and scored well on mainquality parameters such as studymethod, result validity, precision ofoutcomes, and generalizability;medium quality, study method was

sound and results were presentedwith precision; satisfactory quality,the study did not score well or did notcontain any information on the mainquality parameters such as studymethod, result validity, precision ofoutcomes, or generalizability.

Data Synthesis and StatisticalAnalysis

Data were abstracted from all thestudies that met eligibility criteria. Allstatistical tests in the analysis were2-tailed, and P values of #.05 wereconsidered significant. Statisticalanalysis was done using SPSS(version 22, IBM SPSS Statistics,Chicago, IL). Estimates of associationbetween human milk feeding andROP risk were evaluated by ORs andcorresponding 95% CIs. I2 statisticswere applied for the assessment of

FIGURE 2Forest plots of the summary OR value with corresponding 95% CIs for the correlation between human milk feeding and any-stage ROP.

PEDIATRICS Volume 136, number 6, December 2015 e1579 by guest on September 29, 2020www.aappublications.org/newsDownloaded from

heterogeneity among studies in themeta-analysis using RevMan (version5.3.5, Nordic Cochrane Centre,Cochrane Collaboration, London, UK).Evidence summaries were preparedfor the included studies by usingpredetermined output tables.

Role of the Funding Source

This review was conducted as acollaboration of researchers withdiverse backgrounds (neonatology,gastroenterology, and public health).No funding was obtained forconducting this study, and all theauthors contributed throughvoluntary efforts.

RESULTS

Selection Results and IncludedStudies

We identified 1270 citations from theelectronic search of the databasesfrom earliest date until February 28,2015. After duplicate studies wereremoved, 728 studies were subjectedto title and abstract screening. Afterexcluding 418 studies and laterincluding 2 additional studies basedon manual searching from relevantjournals, 312 studies were subjectedto full-text review. Finally, weidentified 5 cohort studies forqualitative synthesis and meta-analysis after excluding 307studies.17,18,21–23 Fig 1 provides asummary of the evidence search andliterature review. The backgroundinformation of these studies ispresented in Table 1. Averagegestational age and birth weight ofparticipants ranged from 26 to 30.2weeks and 775 to 1376 g,respectively. Definitions of feedingtype and ROP varied across studies.To pool data, we classified feedinginto 6 categories: exclusive humanmilk (100% human milk feeding), anyhuman milk, mainly human milk(.50%), exclusive formula (100%formula feeding), any formula, andmainly formula (.50%). We definedROP as any-stage ROP or severe ROP

(including stage 3 or 4, surgical, andthreshold ROP).

Feeding and Any-Stage ROP

Based on the feeding categories, weconducted 4 groups of meta-analyses.The results are presented in Fig 2.Each group includes 2 to 4 studies asdetailed in Table 2. The estimatedORs (95% CIs) were 0.29 (0.12 to0.72), 0.51 (0.26 to 1.03), 0.54 (0.15to 1.96), and 0.25 (0.13 to 0.49),respectively, for exclusive human milkversus any formula, mainly humanmilk versus mainly formula, anyhuman milk versus exclusive formula,and exclusive human milk versusexclusive formula. Human milkfeeding acted as a protective factorfor any-stage ROP. Heterogeneitytests in 4 analyses showed I2 valuesof 53%, 74%, 90%, and 18%. Incomparing mainly human milk versusmainly formula and any human milkversus exclusive formula, theoutcomes of the individual studiesshowed poor consistency.

Feeding and Severe ROP

Four groups of meta-analyses weredone with the target outcome ofsevere ROP, with each groupcomprising 2 to 3 studies. The results

are presented in Fig 3. The estimatedORs (95% CIs) were 0.11 (0.04 to0.30), 0.16 (0.06 to 0.43), 0.42 (0.08to 2.18), and 0.10 (0.04 to 0.29),respectively, for exclusive human milkversus any formula, mainly humanmilk versus mainly formula, anyhuman milk versus exclusive formula,and exclusive human milk versusexclusive formula. Human milkfeeding acted as a protective factorfor severe ROP in all analyses exceptany human milk versus exclusiveformula. Heterogeneity tests in 4analyses showed I2 values of 0%,27%, 91%, and 0%. In comparing anyhuman milk versus exclusive formula,the outcomes of the individual studiesshowed poor consistency.

Quality Assessment of ObservationalStudies

The 5 observational studies includedin the meta-analysis were furtherassessed independently using theCritical Appraisal Skills Programchecklist for cohort studies(Table 3).20 Based on our assessment,1 study17 was rated high, 2studies18,21 were rated medium, and2 studies22,23 were rated satisfactory.This implies that the medium-qualitystudies scored well on all the

TABLE 2 ORs of group analyses

Group and Reference Patients, n Any-stage ROP Stage 3 or 4, surgical ROP

n OR (95% Cl) n OR (95% Cl)

Exclusive human milk versusany formulaHylander et al, 200121 17 vs 157 6/82 0.50 (0.18–1.42) 0/15 0.26 (0.02–4.58)Manzoni et al, 201317 314 vs 184 11/29 0.19 (0.09–0.40) 4/22 0.10 (0.03–0.28)

Mainly human milk versusmainly formulaHylander et al, 200121 45 vs 90 19/53 0.51 (0.25–1.05) 0/11 0.08 (0.00–1.32)Furman et al, 200322 32 vs 87 14/45 0.73 (0.32–1.64) No data —

Maayan-Metzger et al, 201223 188 vs 172 23/22 0.95 (0.51–1.78) 3/7 0.38 (0.10–1.50)Manzoni et al, 201317 314 vs 184 11/29 0.19 (0.09–0.40) 4/22 0.10 (0.03–0.28)

Any human milk versusexclusive formulaHylander et al, 200121 100 vs 74 41/47 0.40 (0.22–0.74) 6/9 0.46 (0.16–1.36)Furman et al, 200322 79 vs 40 44/15 2.10 (0.96–4.57) No data —

Heller et al, 200718 788 vs 269 No data — 130/33 1.41 (0.94–2.13)Manzoni et al, 201317 314 vs 184 11/29 0.19 (0.09–0.40) 4/22 0.10 (0.03–0.28)

Exclusive human milkversus exclusive formulaHylander et al, 200121 17 vs 74 6/47 0.40 (0.22–0.74) 0/9 0.20 (0.01–3.55)Manzoni et al, 201317 314 vs 184 11/29 0.19 (0.09–0.40) 4/22 0.10 (0.03–0.28)

e1580 ZHOU et al by guest on September 29, 2020www.aappublications.org/newsDownloaded from

checklist parameters related to studymethod and results validity. Whereasthe satisfactory-quality studies scoredwell on parameters related to studymethod and results validity, specificinformation pertaining to certainparameters was either unclear or notreported in the studies.

DISCUSSION

ROP is a vasoproliferative disorder ofimmature retina affecting the vastmajority of preterm newborns. Theincidence of severe ROP in very-preterm infants weighing ,1250 gcould be as high as 37%.24 Low birthweight and prematurity are stronglyassociated with increased risk for thedisease.25 ROP is currently thebiggest contributor to infantblindness in developed countries, asvision loss occurs secondary to retinal

detachment that may occur in the mostsevere cases.26 In addition, myopia,strabismus, and amblyopia also occurfrequently.27,28 The pathogenesis ofROP is multifactorial: besidesprematurity and low birth weight,factors such as high-concentrationoxygen therapy and suboptimalpostnatal nutrition could put infants atsignificant risk for this devastating eyedisease, as established by clinicalstudies and animal studies.9,29

The meta-analysis results of our studyindicate that the overall incidence ofROP was reduced among infants fedhuman milk compared with those fedformula, and exclusive or mainlyhuman milk feeding showed significantbenefits in preventing severe ROP.

The underlying physiologicmechanism through which breastmilk may protect against the

development of ROP may reflect theantioxidant30 and immune-protective31 properties of humanmilk. In vitro chemical analysis ofantioxidant content consistentlyshows that human milk containsvitamin C, vitamin E, and b-caroteneand has greater antioxidantproperties than formula.32,33 Inaddition to the antioxidant properties,human milk also containsimmunomodulatory substances suchas secretory immunoglobulin A,lactoferrin, lysozyme, cytokines,oligosaccharides, antioxidantenzymes, and cellularcomponents.33–35 These factors arethought to influence immunedefenses of the infant, which mayexplain the lower risk of necrotizingenterocolitis and sepsis amonginfants fed human milk.15,16

FIGURE 3Forest plots of the summary OR value with corresponding 95% CIs for the correlation between human milk feeding and severe ROP.

PEDIATRICS Volume 136, number 6, December 2015 e1581 by guest on September 29, 2020www.aappublications.org/newsDownloaded from

TABLE3

CriticalAppraisalSkillsProgram

checklistresults

Sectionand

number

Question

Option

Hylander

etal,200121

Furm

anet

al,200322

Helleret

al,200718

Maayan-Metzger

etal,

201223

Manzoni

etal,201317

Aretheresults

ofthestudyvalid?

1Didthestudyaddressa

clearlyfocused

issue?

Yes/Can’ttell/No

Yes

Yes

Yes

Yes

Yes

2Was

thecohort

recruitedin

anacceptableway?

Yes/Can’ttell/No

Yes

Yes

Yes

Yes

Yes

3Was

theexposure

accuratelymeasured

tominimizebias?

Yes/Can’ttell/No

Yes

Yes

Yes

Yes

Yes

4Was

theoutcom

eaccuratelymeasured

tominimizebias?

Yes/Can’ttell/No

Yes

Yes

Yes

Yes

Yes

5aHave

theauthors

identified

all

important

confoundingfactors?

Yes/Can’ttell/No

Yes

Yes.Theauthorshave

countedliquidfortifier

aspreterm

form

ula,and

onlytheactualvolumeof

maternalmilk

was

countedas

maternal

milk.

Yes

Yes

Yes

5bHave

they

taken

accountof

the

confoundingfactors

inthedesign

and/or

analysis?

Yes/Can’ttell/No

Yes.Associations

between

potentialvariablesand

presence

ofRO

Pwas

identified

throughbivariate

analysis.

Yes.Theauthorshave

calculated

themean

proportionof

maternal

milk

oftotalintake

(oral

plus

intravenous)

andof

oral

intake.

Yes

Yes

Yes

6aWas

thefollow-upof

subjects

complete

enough?

Yes/Can’ttell/No

No.M

issing

data,through

comprisingonly2.4%

ofthe

totalnumberof

data

elem

ents,occurredin

one-

sixthof

thecases.

Yes.Results

ofall119

infantsarereported.

Yes

Yes

Yes

6bWas

thefollow-upof

subjects

long

enough?

Yes/Can’ttell/No

Yes

Yes.Subjectswerefollowed

throughweeks

2,4,and

6.

Yes

Yes

Yes

Whataretheresults

ofthisstudy?

e1582 ZHOU et al by guest on September 29, 2020www.aappublications.org/newsDownloaded from

TABLE3

Continued

Sectionand

number

Question

Option

Hylander

etal,200121

Furm

anet

al,200322

Helleret

al,200718

Maayan-Metzger

etal,

201223

Manzoni

etal,201317

7Whataretheresults

ofthisstudy?

Explainin

1sentence

Incidenceof

ROPwas

significantlyreducedin

human

milk-fedVLBW

infants

comparedwith

exclusively

form

ula-fedVLBW

infants.

Ratesof

ROPdidnotdiffer

accordingto

the

amountsof

maternal

milk

received.

Neither

receiptnor

increasing

intake

ofhuman

milk

was

associated

with

adecreasedrisk

ofdeveloping

severe

ROP.

Lower

ratesof

ROPwere

also

detected

ininfants

born

at24–28

wks

who

werebreastfed,

butthe

results

didnotreach

statistical

significance

usingunivariate

analysis

(P,.06).U

sing

multivariate

analysis,

however,R

OPstageIII

amongthissubgroup

was

significantlylower

(P,.022).

Overall,RO

Pincidence(at

anystage)

was

significantlylower

ininfantsfedmaternal

milk

comparedwith

form

ula-fedneonates.

8Howprecisearethe

results?

Explain/

comment

Dose–response

effect

was

not

observed

whencategories

ofhuman

milk

wereentered

into

logisticregression

results,asdurationof

feedingwas

notincluded;

henceresults

would

bepreciseonlyto

someextent.

Results

arenotprecise,as

samplesize

was

not

largeenough

toadequatelyassess

ROP.

Results

areprecise,as

power

calculationwas

conductedforadequate

samplesize

ofincluded

infants.

Results

arepreciseto

someextent

dueto

the

fact

that

studylacked

preciseknow

ledge

regardingfeedingdays

andam

ounts,as

wellas

theuseof

human

milk

fortifier

consistingof

cowmilk

proteinin

the

HMgroup.

Good,asmultivariate

logisticregression

controlling

for

potentially

confounding

factorsto

ROPat

any

stageat

univariate

analysisshow

edtype

ofmilk

feedingretained

significance,maternal

milk

beingprotectiveat

P=.01.

9Do

youbelieve

the

results?

Yes/Can’ttell/No

Yes

Yes

Yes

Yes

Yes

Will

theresults

help

locally?

10Cantheresults

beappliedto

thelocal

population?

Yes/Can’ttell/No

No.The

studywas

limitedto

1tertiary

center

with

extensiveresourcesthrough

theMilk

Bank

andLactation

Center

availableto

mothers

who

choose

toprovide

human

milk

totheirVLBW

infants,thus

allowing

relativelyhigh

rate

ofprovidinghuman

milk

toVLBW

infantsin

thestudy

sample.

Yes

Yes

Yes

Yes

11Do

theresults

ofthis

studyfitwith

other

availableevidence?

Yes/Can’ttell/No

Yes

No.M

aternalm

ilkhasbeen

reported

toreduce

rate

ofandseverityof

ROPin

VLBW

infants.

No.H

Mhasbeen

associated

with

decreasedrisk

ofRO

P.

Yes

Yes

PEDIATRICS Volume 136, number 6, December 2015 e1583 by guest on September 29, 2020www.aappublications.org/newsDownloaded from

This meta-analysis did not includestudies on donor human milk. In2005, Schanler et al conducted arandomized trial of donor humanmilk versus preterm formula as asubstitute for mother’s milk in thefeeding of extremely low birthweight premature infants.36 Thestage 3 ROP incidences were 19%and 14% in donor milk and formulagroups, respectively. The study didnot reveal any benefit of donor milkin terms of preventing severeROP.36 Another study comparingdonor milk and formula feedingalso did not demonstrate anybenefit in preventing ROP.37 Thiscontrast may be possibly relatedto loss of the protective factors inbreast milk during processing andstorage.

The medical benefits of human milkand the recommendation of humanmilk as the preferred feeding sourcefor preterm infants limit prospectiverandomized studies; we thereforeselected observational studies for ourmeta-analysis. This selection could bea limitation to the current meta-analysis. Another limitation is that thedefinitions of infant feeding werebased on retrospective data and couldunderestimate or overestimate thevolume of human milk feeding, andthe duration of human milk feedingwas not reported in the studies. Ofnote, in the study of Furman et al,22

the volume of human milk fed wasunknown; to pool the data with theother studies, we included patientsfed with$50 mL/kg/day human milkin the “mainly human milk” group.The diagnosis of ROP also variedamong studies, further complicatinganalysis. We resolved this by poolingstage 3 and 4 ROP, surgical ROP, andthreshold ROP as a single definition ofsevere ROP. However, there was lackof long-term prognostic information.Our analysis showed highheterogeneity in some subgroupanalyses, so the results in thecorresponding analyses should beaccepted cautiously and notconsidered a definitive statement.TA

BLE3

Continued

Sectionand

number

Question

Option

Hylander

etal,200121

Furm

anet

al,200322

Helleret

al,200718

Maayan-Metzger

etal,

201223

Manzoni

etal,201317

12Whatarethe

implications

ofthis

studyforpractice?

Explain/

comment

Findings

suggestaprotective

effect

ofhuman

milk

feedings

againstROPafter

controlling

forpotential

confoundingvariables.

Results

cannot

begeneralized,assample

size

was

lowto

assess

theeffect

ofmaternal

milk

feedingon

neonatal

morbiditiesthatoccurat

lower

ratessuch

asRO

P.

Limitations

such

aslack

ofdata

collectionof

eye

exam

inations

after

discharge,missing

data

of.10%

ofsamplewho

didnotrequiresurgery

forRO

Pbefore

discharge

anddidnotreturn

for

follow-upvisits

at18

or30

mocorrectedage,

andabsenceof

defined

definitionof

ROPacross

studycenters.

Human

milk

isRO

Pprotectiveeven

ifpartially

administered.

Exclusivematernalmilk

feedingsincebirthmay

preventRO

Pofanystage

inVLBW

infantsin

the

NICU.

Overallquality

Medium

Satisfactory

Medium

Satisfactory

High

VLBW

,verylowbirthweight.

e1584 ZHOU et al by guest on September 29, 2020www.aappublications.org/newsDownloaded from

The heterogeneity among studiescould be attributable to two factors.First, variations in study duration:the studies included in meta analyseswere from 1992 to 2008. With theadvancement of respiratory support,standardization of oxygen utilization,and increased survival rate of verypreterm infants, the incidence of ROPvaried from study to study. Second,variation in gestational age and birthweight: the 2 most relevant riskfactors for ROP ranged from 28 to 30weeks and 775 to 1425 g,respectively. It has been observedthat a 1-week increment ingestational age could change ROPrate significantly.8,12 A random-effects model for conducting meta-analysis has been suggested as areasonable way of addressingheterogeneity in pooled studies, andin such cases narratively explainingthe reasons for heterogeneity hasbeen proposed.38 We justify usingour meta-analysis approach on thebasis of these methodologicalsuggestions.

Notwithstanding these limitations,our meta-analysis has severalstrengths. Ours is the first systematicanalysis of evidence to dateregarding the possible benefits ofhuman milk on ROP. We have donemultiple database searches,including manually searchingrelevant journals in pediatrics andophthalmology, to incorporatemaximum published studiesconcerning our focus. We havefollowed expected guidelines formeta-analysis of observationalstudies, which increases theapplicability of our results.

We have also reported qualityassessment of the included studiesin the meta-analysis. Our objectivein conducting this exercise was notto assess if the quality ofpublication should be a criterion forinclusion in the meta-analysis, suchthat only the high-quality studiesare included, but to present howconfident can one be in the resultspresented in the published studies.In doing so, it enables the

methodology in these studies to beunderstood and appraised for thereader.

CONCLUSIONS

In our meta-analysis, longitudinalstudies comparing the incidence ofROP in infants who were fed humanmilk versus formula were selected.Studies involved donor milk were notincluded. After unifying thedefinitions of feeding and ROPdiagnosis and pooling the data, wefound that human milk feedingpotentially plays a strong role inprotecting very preterm newbornsfrom any-stage ROP and severe ROP.

ABBREVIATIONS

CI: confidence intervalMOOSE: Meta-analysis of

Observational Studies inEpidemiology

OR: odds ratioROP: retinopathy of prematurity

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.

FUNDING: No external funding.

POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

REFERENCES

1. Oza S, Lawn JE, Hogan DR, Mathers C,Cousens SN. Neonatal cause-of-deathestimates for the early and late neonatalperiods for 194 countries: 2000-2013. BullWorld Health Organ. 2015;93(1):19–28

2. World Health Organization. Preterm Birth.Fact sheet 363, November 2014. Availableat: www.who.int/mediacentre/factsheets/fs363/en/. Accessed May 29, 2015

3. Hack M, Wright LL, Shankaran S, et al.Very-low-birth-weight outcomes of theNational Institute of Child Health andHuman Development Neonatal Network,November 1989 to October 1990. Am JObstet Gynecol. 1995;172(2 Pt 1):457–464

4. Fanaroff AA, Wright LL, Stevenson DK, et al.Very-low-birth-weight outcomes of theNational Institute of Child Health and

Human Development Neonatal ResearchNetwork, May 1991 through December1992. Am J Obstet Gynecol. 1995;173(5):1423–1431

5. Stevenson DK, Wright LL, Lemons JA,et al. Very low birth weight outcomes ofthe National Institute of Child Healthand Human Development NeonatalResearch Network, January 1993through December 1994. Am J ObstetGynecol. 1998;179(6 pt 1):1632–1639

6. Lemons JA, Bauer CR, Oh W, et al; NICHDNeonatal Research Network. Very lowbirth weight outcomes of the NationalInstitute of Child health and HumanDevelopment Neonatal Research Network,January 1995 through December 1996.Pediatrics. 2001;107(1). Available at: www.pediatrics.org/cgi/content/full/107/1/e1

7. Hintz SR, Kendrick DE, Vohr BR, PooleWK, Higgins RD; National Institute ofChild Health and Human DevelopmentNeonatal Research Network. Changesin neurodevelopmental outcomes at18 to 22 months’ corrected age amonginfants of less than 25 weeks’gestational age born in 1993-1999.Pediatrics. 2005;115(6):1645–1651

8. Stoll BJ, Hansen NI, Bell EF, et al Neonataloutcomes of extremely preterm infantsfrom the NICHD Neonatal ResearchNetwork. Pediatrics. 2010;126(3):443–456

9. Hartnett ME, Penn JS. Mechanisms andmanagement of retinopathy ofprematurity. N Engl J Med. 2012;367(26):2515–2526

10. Hartnett ME. Pathophysiology andmechanisms of severe retinopathy of

PEDIATRICS Volume 136, number 6, December 2015 e1585 by guest on September 29, 2020www.aappublications.org/newsDownloaded from

prematurity. Ophthalmology. 2015;122(1):200–210

11. Austeng D, Källen K, Hellström A, et al.Regional differences in screening forretinopathy of prematurity in infantsborn before 27 weeks of gestation inSweden—the EXPRESS study. ActaOphthalmol. 2014;92(4):311–315

12. Xu Y, Zhou X, Zhang Q, et al. Screening forretinopathy of prematurity in China: aneonatal units-based prospective study.Invest Ophthalmol Vis Sci. 2013;54(13):8229–8236

13. Cerman E, Balci SY, Yenice OS, KazokogluH, Celiker H, Eraslan M. Screening forretinopathy of prematurity in a tertiaryophthalmology department in Turkey:incidence, outcomes, and risk factors.Ophthalmic Surg Lasers Imaging Retina.2014;45(6):550–555

14. Karna P, Muttineni J, Angell L, KarmausW. Retinopathy of prematurity and riskfactors: a prospective cohort study. BMCPediatr. 2005;5(1):18

15. Herrmann K, Carroll K. An exclusivelyhuman milk diet reduces necrotizingenterocolitis. Breastfeed Med. 2014;9(4):184–190

16. Neu J, Walker WA. Necrotizing enterocolitis.N Engl J Med. 2011;364(3):255–264

17. Manzoni P, Stolfi I, Pedicino R, et al;Italian Task Force for the Study andPrevention of Neonatal Fungal Infections,Italian Society of Neonatology. Humanmilk feeding prevents retinopathy ofprematurity (ROP) in preterm VLBWneonates. Early Hum Dev. 2013;89(suppl 1):S64–S68

18. Heller CD, O’Shea M, Yao Q, et al; NICHDNeonatal Research Network. Human milkintake and retinopathy of prematurity inextremely low birth weight infants.Pediatrics. 2007;120(1):1–9

19. Stroup DF, Berlin JA, Morton SC, et al.Meta-analysis of observational studies inepidemiology: a proposal for reporting.Meta-analysis Of Observational Studiesin Epidemiology (MOOSE) group. JAMA.2000;283(15):2008–2012

20. Critical Appraisal Skills Programme(CASP). May 2013. Available at:

www.casp-uk.net/#!casp-tools-checklists/c18f8. Accessed June 24, 2015

21. Hylander MA, Strobino DM, Pezzullo JC,et al. Association of human milkfeedings with a reduction in retinopathyof prematurity among very lowbirthweight infants. J Perinatol. 2001;21(6):356–362

22. Furman L, Taylor G, Minich N, Hack M. Theeffect of maternal milk on neonatalmorbidity of very low-birth-weightinfants. Arch Pediatr Adolesc Med. 2003;157(1):66–71

23. Maayan-Metzger A, Avivi S, Schushan-Eisen I, Kuint J. Human milk versusformula feeding among preterm infants:short-term outcomes. Am J Perinatol.2012;29(2):121–126

24. Good WV, Hardy RJ, Dobson V, et al; EarlyTreatment for Retinopathy ofPrematurity Cooperative Group. Theincidence and course of retinopathy ofprematurity: findings from the earlytreatment for retinopathy ofprematurity study. Pediatrics. 2005;116(1):15–23

25. Lad EM, Hernandez-Boussard T, MortonJM, Moshfeghi DM. Incidence ofretinopathy of prematurity in the UnitedStates: 1997 through 2005. Am JOphthalmol. 2009;148(3):451–458

26. Darlow BA, Hutchinson JL, Henderson-Smart DJ, Donoghue DA, Simpson JM,Evans NJ; Australian and New ZealandNeonatal Network. Prenatal risk factorsfor severe retinopathy of prematurityamong very preterm infants of theAustralian and New Zealand NeonatalNetwork. Pediatrics. 2005;115(4):990–996

27. Quinn GE, Dobson V, Davitt BV, et al.Progression of myopia and high myopiain the Early Treatment for Retinopathy ofPrematurity study: findings at 4 to 6years of age. J AAPOS. 2013;17(2):124–128

28. Nudleman E, Robinson J, Rao P, DrenserKA, Capone A, Trese MT. Long-termoutcomes on lens clarity after lens-sparing vitrectomy for retinopathy ofprematurity. Ophthalmology. 2015;122(4):755–759

29. Hellström A, Smith LE, Dammann O.Retinopathy of prematurity. Lancet. 2013;382(9902):1445–1457

30. Cloetens L, Panee J, Åkesson B. Theantioxidant capacity of milk—theapplication of different methods in vitroand in vivo. Cell Mol Biol (Noisy-le-grand). 2013;59(1):43–57

31. Perrin MT, Fogleman A, Allen JC. Thenutritive and immunoprotective qualityof human milk beyond 1 yearpostpartum: are lactation-duration-based donor exclusions justified? J HumLact. 2013;29(3):341–349

32. Elisia I, Kitts DD. Differences in vitamin Eand C profile between infant formula andhuman milk and relative susceptibility tolipid oxidation. Int J Vitam Nutr Res.2013;83(5):311–319

33. Aycicek A, Erel O, Kocyigit A, Selek S,Demirkol MR. Breast milk providesbetter antioxidant power than doesformula. Nutrition. 2006;22(6):616–619

34. Baydas G, Karatas F, Gursu MF, et al.Antioxidant vitamin levels in term andpreterm infants and their relation tomaternal vitamin status. Arch Med Res.2002;33(3):276–280

35. L’Abbe MR, Friel JK. Superoxidedismutase and glutathione peroxidasecontent of human milk from mothersof premature and full-term infantsduring the first 3 months of lactation.J Pediatr Gastroenterol Nutr. 2000;31(3):270–274

36. Schanler RJ, Lau C, Hurst NM, Smith EO.Randomized trial of donor human milkversus preterm formula as substitutesfor mothers’ own milk in the feeding ofextremely premature infants. Pediatrics.2005;116(2):400–406

37. Cristofalo EA, Schanler RJ, Blanco CL,et al. Randomized trial of exclusivehuman milk versus preterm formuladiets in extremely premature infants.J Pediatr. 2013;163(6):1592–1595.e1

38. Schroll JB, Moustgaard R, Gøtzsche PC.Dealing with substantial heterogeneityin Cochrane reviews. Cross-sectionalstudy. BMC Med Res Methodol. 2011;11:22

e1586 ZHOU et al by guest on September 29, 2020www.aappublications.org/newsDownloaded from

DOI: 10.1542/peds.2015-2372 originally published online November 16, 2015; 2015;136;e1576Pediatrics 

Jianguo Zhou, Vivek V. Shukla, Denny John and Chao ChenMeta-analysis

Human Milk Feeding as a Protective Factor for Retinopathy of Prematurity: A

ServicesUpdated Information &

http://pediatrics.aappublications.org/content/136/6/e1576including high resolution figures, can be found at:

Referenceshttp://pediatrics.aappublications.org/content/136/6/e1576#BIBLThis article cites 33 articles, 8 of which you can access for free at:

Subspecialty Collections

http://www.aappublications.org/cgi/collection/breastfeeding_subBreastfeedinghttp://www.aappublications.org/cgi/collection/nutrition_subNutritionsubhttp://www.aappublications.org/cgi/collection/fetus:newborn_infant_Fetus/Newborn Infantfollowing collection(s): This article, along with others on similar topics, appears in the

Permissions & Licensing

http://www.aappublications.org/site/misc/Permissions.xhtmlin its entirety can be found online at: Information about reproducing this article in parts (figures, tables) or

Reprintshttp://www.aappublications.org/site/misc/reprints.xhtmlInformation about ordering reprints can be found online:

by guest on September 29, 2020www.aappublications.org/newsDownloaded from

DOI: 10.1542/peds.2015-2372 originally published online November 16, 2015; 2015;136;e1576Pediatrics 

Jianguo Zhou, Vivek V. Shukla, Denny John and Chao ChenMeta-analysis

Human Milk Feeding as a Protective Factor for Retinopathy of Prematurity: A

http://pediatrics.aappublications.org/content/136/6/e1576located on the World Wide Web at:

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

by the American Academy of Pediatrics. All rights reserved. Print ISSN: 1073-0397. the American Academy of Pediatrics, 345 Park Avenue, Itasca, Illinois, 60143. Copyright © 2015has been published continuously since 1948. Pediatrics is owned, published, and trademarked by Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it

by guest on September 29, 2020www.aappublications.org/newsDownloaded from