Effect of withholding breastfeeding on the immune response to a live oral rotavirus vaccine in North Indian infants

(1)

ContentslistsavailableatScienceDirect

Vaccine

jou rn al h om ep a g e :w w w . e l s e v i e r . c o m / l o c a t e / v a c c i n e

Effect of withholding breastfeeding on the immune response to a live oral rotavirus vaccine in North Indian infants ^夽

Temsunaro Rongsen-Chandola

^a

, Tor A. Strand

^b,c,∗

, Nidhi Goyal

^a

, Elmira Flem

^d

, Sudeep Singh Rathore

^a

, Alok Arya

^a

, Brita Askeland Winje

^e

, Robin Lazarus

^f

, Elango Shanmugasundaram

^f

, Sudhir Babji

^f

, Halvor Sommerfelt

^d,g

, Kirsti Vainio

^h

, Gagandeep Kang

^f

, Nita Bhandari

^a

aCentreforHealthResearchandDevelopment,SocietyforAppliedStudies,45KaluSarai,NewDelhi110016,India

bInnlandetHospitalTrust,Lillehammer,Norway

cCentreforInterventionScienceinMaternalandChildHealth,CentreforInternationalHealth,UniversityofBergen,Norway

dDepartmentofVaccines,DivisionofInfectiousDiseaseControl,NorwegianInstituteofPublicHealthOslo,Norway

eDepartmentofInfectiousDiseasesEpidemiology,DivisionofInfectiousDiseaseControl,NorwegianInstituteofPublicHealthOslo,Norway

fTheWellcomeTrustResearchLaboratory,ChristianMedicalCollege,Vellore632004,TamilNadu,India

gDepartmentofInternationalPublicHealth,NorwegianInstituteofPublicHealth,Oslo,Norway

hDepartmentofVirology,DivisionofInfectiousDiseaseControl,NorwegianInstituteofPublicHealthOslo,Norway

a r t i c l e i n f o

Articlehistory:

Keywords:

Rotavirus Rotarix^® Vaccine Immuneresponse Withholdbreastfeeding Encouragedbreastfeeding

a b s t r a c t

Interferencefromtransplacentalandbreastmilkantibodiesmayimpedetheperformanceoforallive vaccines.Theeffectofbreastfeeding ontheimmunogenicityofRotarix^®,atwo-doseoralmonova- lentrotavirusvaccine,wasexaminedinacommunity-basedtrialinNewDelhi,India.Fourhundred mother–infantpairswererandomizedintotwoequalgroups.Infantswereaged6–7weeksatenroll- ment.Motherswereencouragedtoeitherbreastfeedortowithholdbreastfeedingduringthe30min priortoandaftereachvaccinedosewasadministered.Wecollectedbloodspecimensfrominfantsat enrollmentand4weeksafterthesecondvaccinedose.Bloodandbreastmilkspecimenswereobtained frommothersatbaselineandbreastmilkspecimenswerecollectedatthetimeofthesecondvaccine dose.Seroconversionwasdeﬁnedasinfantserumanti-VP6IgAantibodylevelof≥20IU/mL4weeks afterthesecondvaccinedoseanda≥4-foldrisefrombaseline.Therewasnodifferenceinthepropor- tionwhoseroconvertedbetweenthetwogroups(26%vs27%;p=0.92).ThelevelsofinfantserumIgA, maternalserumandbreastmilkIgAandIgGanti-rotavirusantibodiespredictedtheanti-rotavirusIgA levelininfantsatend-studyandexplainedapproximately10%ofthevariabilityoftheimmuneresponse (r²=0.10,p<0.001).

Inthispopulation,theimmuneresponsetoRotarix^®wasnotenhancedbywithholdingbreastfeeding aroundthetimeofvaccination.Maternalanti-rotavirusantibodiesexplainedlittleofthevariabilityinthe immuneresponsetothevaccine.Factorsotherthanmaternalanti-rotavirusantibodiesprobablyexplain whyinfantsinlow-andmiddle-incomesettingsrespondpoorlytoliveoralrotavirusvaccines.

1. Introduction

Rotavirusistheleadingcauseoffatal andseverediarrheain children[1].InIndia,itisresponsibleforalmost100,000deaths

夽Trialregistration:ClinicalTrialRegistry-India(CTRI/2012/10/003057),Clinical- trials.gov(NCT01700127).

∗Correspondingauthorat:CentreforInterventionScienceinMaternalandChild Health,CentreforInternationalHealth,UniversityofBergen,Norway.

E-mailaddress:[email protected](T.A.Strand).

annually[2].TheWHOhasrecommendedinclusionofrotavirus vaccinesinallnationalimmunizationprograms.Currentlythereare twolicensedrotavirusvaccinesavailable;Rotarix^®,GSKBiologicals andRotaTeq^®,Merck&Co.Bothvaccineshavedemonstratedhigh efficacy(>90%)againstsevererotavirusdiseasesandrotavirusasso- ciatedhospitalizationinclinicaltrialsinhigh-andmiddle-income countries[3–5].However,trialsofthesetwovaccinesconducted indevelopingsettingsinAfricaand Asiashowedlowerefficacy, ofapproximately60%[6–9].Mostrecently,theindigenouslyman- ufacturedlive,oral116Emonovalenthuman–bovinevaccinehas completedanefficacytrialandisexpectedtobelicensedinIndia http://dx.doi.org/10.1016/j.vaccine.2014.04.078

(2)

soon.Theefﬁcacyofthe116Evaccinewas54%[10]whichissimilar tothatofRotateq^®andRotarix^®inthesesettings.

Otherliveoral vaccines have alsoperformedpoorly in low- income countriesas compared tomore afﬂuent countries[11].

Currentevidence indicatesthat decreased vaccine performance couldbeattributedto severalfactorsincludingchild or mater- nalmalnutrition, environmentalenteropathy, interferencefrom maternal antibodiesand presence of other intestinal infections [11].Presenceofrotavirusantibodiesinbreastmilkandtranspla- centalmaternalantibodiesisassociatedwithimpairedresponsesto rotavirusvaccines[12–14].Indianwomenseemtohavehighercon- centrationsofrotavirusneutralizingantibodiesinbreastmilkthan womeninindustrializedcountries[15].Invitrostudiesoftheneu- tralizingeffectofbreastmilkhavesuggestedthatwithholdingof breastfeedingaroundthetimeofrotavirusvaccineadministration couldimprovetheimmuneresponsetothevaccine[15].

Previoustrialsofrotavirusvaccineshadnotshownanydiffer- enceintheimmune response tovaccineregardlessof whether breastmilkwasgivenornot atthetimeofvaccineadministra- tion.In thosetrials informationonbreastfeedingwasavailable, however,breastfeedingwasself-reportedbymothersandthedura- tionbetweenbreastfeedingand vaccinationwasnotadequately assessed[16,17].ArecentstudyfromSouthAfricareportedthat abstentionfrombreastfeedinganhourbeforeandaftereachvac- cinationhad nosubstantialeffectontheimmuneresponsetoa rotavirusvaccineinHIV-uninfectedinfants[18].

Without clearevidence, it is difﬁcult todetermine whether rotavirusantibodiesinbreastmilkinterferewithimmuneresponse tooralrotavirusvaccinesininfants.Itisimportanttoexplorethis association,asitmayhelpimprovetheimpactofthevaccines.We thereforemeasuredwhetherwithholdingbreastfeedingaroundthe timeofvaccinationwithRotarix^®wouldimproveitsimmunogenic- ity. We also exploredtheassociation betweenmaternal serum andbreastmilkanti-rotavirusantibodyconcentrationswiththe immuneresponseininfantsaftertwodosesofthisvaccine.

2. Materialandmethods 2.1. Studydesignandparticipants

Thetrialwasconductedintypicalurbanresettlementneighbor- hoodsofSouthDelhi,India.

Infantsagedlessthan7weekswereidentifiedthroughahouse- holdsurvey.Familiesofinfantsaged6–7weekswereinvitedto thestudyclinicfor screeningandenrollment.Informedwritten consentwasobtainedfromallparentsandalsospecificallyfrom themothers.AllenrolledinfantsreceivedtwodosesofRotarix^®at 6–7weeksandat10–14weeksofagealongwithotherchildhood vaccines(Diphtheria,Pertussis,Tetanus,Haemophilusinfluenzae B,HepatitisBandoralPolio).

Atthestudyclinicafterconsentwasobtained,aphysicianexam- inedtheinfant.Mother–infantpairswereenrollediftheparents gaveconsent,infantswereaged6–7weeks,theweightforagewas

>−3SDoftheWHOchildgrowthstandards,andthefamilyhadno planstomoveoutofthestudyareaforthenext4months.Infants wereexcludediftheywerenotbreastfed,hadalreadyreceiveda rotavirusvaccine,hadimmunodeﬁciencydisease,chronicenteric disease,and/oranyotherconditionaswarrantingexclusionbythe investigator.Infantsweretemporarilyexcludediftheyhaddiarrhea oranyillnessrequiringhospitalreferralonthedayofenrollment.

Eligibleinfantswereeitherallocatedtothegroupwheremoth- erswererequestedtowithholdbreastfeedingfor30minbeforeand aftervaccineadministrationortothegroupwheremotherswere encouragedtobreastfeedtheirinfantsaroundthetimeofvacci- nation.Thereweretwoseparatelocationsinthestudyclinicfor

thetwogroupstoensurethatinstructionsforbreastfeedingwere followedbymothers.Clinicalcoordinatorssupervisedeacharea.

Activitieswereconductedinthefollowingorder:30minofwith- holdingorencouragingbreastfeeding;administrationofRotarix^®; 30minofwithholdingorencouragingbreastfeeding;administra- tionofotherchildhoodvaccines;observationfor30mintoassess for immediateadverseevents. Thestudyteamdocumented the timebreastfeedingstartedandendedaswellasthetimewhenthe othervaccineswereadministered.

2.2. Follow-upactivities

Infantswereobservedforimmediateadverseeventsinthestudy clinicandreferredtothehospital,ifrequired.Familiesofinfants werecontactedweeklyaftereachdoseoftheRotarix^®toascertain presenceofsignsandsymptomsofanyillnessrequiringhospital referralincludingintussusception,orotherseriousadverseevents.

Minorillnessesnotrequiringhospitalreferralweremanagedby thestudyphysician.Seriousadverseeventswerereportedtothe relevantEthicsCommittees.

2.3. Randomizationandblinding

Therandomization listwasgenerated bya statisticianinde- pendentof thestudyteamin Stata11 (StataCorpLP, TX,USA).

Eligibleinfantswererandomlyallocatedtoeithergroupthrough seriallynumberedopaquesealedenvelopeswiththegroupallo- cationwritteninsidetheenvelope.Thelaboratoryassessingthe immuneresponseswasblindedtothegroupallocation.

2.4. Assessmentofimmuneresponse

Atenrollment,bloodandbreastmilkspecimenswereobtained frommothersandbloodandstoolspecimenswereobtainedfrom theinfants. Atthetime oftheseconddoseofRotarix^®,abreast milkspecimenwasobtainedfromthemother.Fourweeksafter theseconddoseofRotarix^®,bloodspecimenwasobtainedfrom eachinfant.

2.5. Laboratorymethods

ThespecimensweretestedattheWellcomeTrustResearchLab- oratoryatChristianMedicalCollege,Vellore.TheIgAandIgGtiters weredeterminedby comparingtheopticaldensity values form samplewellswiththestandardcurvebasedonderivedunitsofIgA arbitrarilyassignedtopooledhumanserumsamples,aspreviously described[19].

2.6. Statisticalanalyses

StatisticalanalyseswerecarriedoutinStata11.0(StataCorpLP, TX,USA).Descriptivemeasuresofcontinuousvariableswerepre- sentedasmeansandstandarddeviationsforsymmetricaldata,and asmediansandinterquartilerangesforskeweddata.TheSpearman rank-ordercorrelationtestwasusedforcomparingmedianvalues.

Seroconversionwasdefinedasinfantserumanti-VP6IgAantibody levelof≥20IU/mL4weeksafterthesecondvaccinedoseanda≥4- foldrisefrombaseline.Wemeasuredtheeffectoftheinterventions andotherexposuresontheproportionwhoseroconvertedandon thelog-transformedendstudyantibodylevelsoftheinfants.The relationshipbetweenmaternalandchildantibodiesandtheseout- comeswereexaminedincrudeandmultivariatelogisticandlinear regressionmodels.Inthesemodels,weinitiallyincludedvariables thatweresignificantona0.05level(fromthecrudemodels),we keptthosethatremainedsignificantandaddedtheotherexposure variablesoneatatimeandretainedsignificantvariablesforthe

(3)

ﬁnalmodel.Theratiobetweenproportionsanditscorresponding conﬁdenceintervalwascalculatedusingthebinregcommandin stata.

2.7. Ethicalconsiderations

EthicalclearancewasobtainedfromSocietyforAppliedStudies, EthicsReviewCommittee,ChristianMedicalCollege,Institutional EthicsCommitteeandSouth-EastRegionalEthicalCommitteeof Norway.Thisstudywasconductedincompliancewiththeprotocol, GoodClinicalPracticesandotherrelevantregulatoryguidelines.

3. Results

Ofthe533infantsscreenedforeligibility,400wereenrolled andrandomizedintotwoequalgroups.Allinfantsreceivedtheﬁrst doseofRotarix^®and391receivedbothdoses;fourfamiliesmoved outofthestudyareaandﬁverefusedtheseconddose(Fig.1).

Bothbaselineandendstudybloodspecimenwereavailablefor 388infants.

The baseline characteristics were comparable between the groups(Table1).Theproportionofinfantsbeingexclusivelybreast- fed was highoverall; 75% in the group that was requested to withholdbreastfeedingaroundthetimeofvaccinationand80%in thegroupthatwasencouragedtobreastfeed.Themean(SD)age ofinfantsatthetimeofvaccinationwas6.9(0.56)and11.2(0.62) monthsfortheﬁrstandseconddoses,respectively.

Theinfantandmaternalanti-rotavirusantibodylevelsinthe serum and breast milk were similar between the two groups (Table2).Allexceptonemotherinthegroupthatwaswithhold- ingbreastfeedingadheredtotheinstructions.Infantsinthegroup withholdingbreastfeedingwerenotbreastfedforamean(SD)dura- tionof49 (11.1)and 46(10.9)minafterreceivingtheﬁrst and seconddosesofRotarix^®,respectively.

Theproportionsofinfantswhoseroconvertedatstudyendwere similarinthetwogroups;26%ofinfantsinthegroupwherebreast- feedingwaswithheldand27%inthegroupwhereinfantswere breastfed(p=0.920) (Table3).The ratioof the proportion that seroconvertedinthetwogroupswas0.98(95%CI0.70,1.38).The

Fig.1.Trialproﬁle.

maternalserumIgAandIgGatbaselineandbreastmilkIgAand IgGwerealsosigniﬁcantlyassociatedwiththeimmuneresponse (Table4).Whiletheinfantbaselineantibodylevelwaspositively associated,maternalantibodieswerenegativelyassociated with theimmuneresponse.Theadjustedmodel,includinginfantbase- lineserumIgA,breastmilkIgAandbreastmilkIgGconﬁrmedthese associations(Table4).

The odds (95% CI)of seroconversion showed similar results withhigheroddsofseroconversionwithincreasinglevelsofinfant

Table1

Baselinecharacteristicsofparticipantsinthetwostudygroups.

Breastfeedingwithheld(n=200) Breastfeedingencouraged(n=200) Infantcharacteristics

Ageatenrollment(days) Mean(SD)

48(4.0) 49(3.8)

Birthweight^a(kg) Mean(SD)

2.8(0.4) 2.8(0.5)

Weightatscreening(kg) Mean(SD)

4.4(0.6) 4.4(0.5)

Sex

Girls 97(48.5) 95(47.5)

Exclusivelybreastfed 150(75.0) 160(80.0)

Socioeconomiccharacteristics

Homebirth 61(30.5) 52(26.0)

Typeoffamily

Nuclear 112(56.0) 115(57.5)

Joint 88(44.0) 85(42.5)

Numberofsiblings Mean(SD)

0.95(0.96) 1.1(1.1)

Maternalage(years) Mean(SD)

24.4(3.5) 24.8(3.9)

Motherswhohavenotattendedschool 48(24.0) 45(22.5)

Fatherswhohavenotattendedschool 22(11.0) 22(11.0)

FamilywhoowncolorTV,cooler,orscooter 182(91.0) 179(89.5)

Annualfamilyincome,kRupeesmedian/interquartilerange 84(60,120) 84(72,120)

Allaren(%)exceptwhenspeciﬁedotherwise.

aReliableinformationonbirthweightwasavailablefor137(68.5%)infantsinthewithholdingofbreastfeedinggroupand143(71.5%)infantsinthegroupencouragedto breastfeed.

(4)

Table2

Infantandmaternalanti-rotavirusantibodyconcentrationsinbreastfeedingencouragedandwithheldinfants(logIU).

Breastfeedingwithheld Breastfeedingencouraged

Median(IQR) Median(IQR)

Infant

SerumIgAbaseline 1.95(0.33,5.48) 2.04(0.59,4.37)

SerumIgApostdose2 6.77(1.95,35.38) 6.06(1.72,29.02)

Maternal

SerumIgAbaseline 210.19(105.16,384.56) 164.99(94.96,371.05)

SerumIgGbaseline 11,562.03(5402.71,20,575.12) 9801.49(4577.05,23,997.40)

BreastmilkIgAbaseline 17.63(9.13,44.16) 18.23(9.68,36.63)

BreastmilkIgGbaseline 14.26(0.00,52.79) 15.85(0.00,46.13)

BreastmilkIgAatdose2 17.46(9.44,34.81) 17.43(9.22,28.94)

BreastmilkIgGatdose2 16.69(0.00,46.66) 15.82(0.00,53.50)

p>0.05forallcomparisons.

Table3

Immuneresponsesfourweeksafterdose2ofRotarix^®inseronegativeinfants(serumanti-rotavirusIgA<20IU/mLatbaseline).

BreastfeedingwithheldN=172 BreastfeedingencouragedN=184 pvalue

4-Foldrise(seroconversion)^aN% 45(26.1) 49(26.6) 0.92

2-Foldrise^b N% 49(28.5) 51(27.7) 0.87

aSeroconversionwasdeﬁnedasinfantserumanti-VP6IgAantibodylevelof≥20IU/mL4weeksafterthesecondvaccinedoseanda≥4-foldrisefrombaseline.

bDeﬁnedasinfantserumanti-VP6IgAantibodylevelof≥20IU/mL4weeksafterthesecondvaccinedoseanda≥2-foldriseintheanti-VP6IgAantibodylevelfrom baseline.

Table4

Anti-rotavirusantibodylevelsininfantsandmothersaspredictorsforanti-rotavirusserumIgAlevels4weeksafterdose2ofRotarix^®ininfantswhowereseronegativeat baseline.

Explanatoryvariables Crudeestimates Adjustedestimates^a

Co-efﬁcient(95%CI) pvalue R-Squared Co-efﬁcient(95%CI) pvalue

InfantserumIgAbaseline 0.50(0.30,0.70) 0.000 0.063 0.49(0.29,0.69) <0.001

Maternal

BaselinemotherserumIgA −0.17(−0.33,−0.02) 0.031 0.013 – –

BaselinemotherserumIgG −0.22(−0.37,−0.06) 0.006 0.020 – –

BaselinebreastmilkIgA −0.23(−0.38,−0.07) 0.004 0.023 – –

BaselinebreastmilkIgG −0.14(−0.22,−0.06) 0.001 0.030 – –

BreastmilkIgAatdose2 −0.26(−0.42,−0.11) 0.001 0.031 −0.20(−0.36,−0.05) 0.010

BreastmilkIgGatdose2 −0.15(−0.23,−0.07) 0.000 0.038 −0.10(−0.19,−0.02) 0.013

Logtransformedanti-rotavirusconcentrations.

aTheadjustedestimatesaretheresultsofamultiplelinearregressionmodelwhereN=356,p<0.001adjustedr²=0.1047,onlythevariablesthatremainedsigniﬁcant wereretainedintheﬁnalmodel.

serum IgA at baseline and lower odds of seroconversion with increasinglevelsofmaternalantibodies(Table5).

4. Discussion

Weexaminedtheeffectoftemporarilywithholdingbreastfeed- ing onthe immune response tothe liveoral rotavirusvaccine Rotarix^® in a randomized community trial. Despite excellent

Table5

Anti-rotavirusantibodyconcentrationininfantsandmothersaspredictorsforsero- conversion4weeksafterdose2ofRotarix^®ininfantsseronegativeatbaseline.

Explanatoryvariables Crudeestimates

Oddsratio(95%CI) pvalue Infant

InfantserumIgAbaseline 1.24(0.92–1.68) 0.162 Maternal

BaselinemotherserumIgA 0.82(0.65–1.03) 0.094 BaselinemotherserumIgG 0.75(0.60–0.93) 0.011 BaselinebreastmilkIgA 0.76(0.59–0.97) 0.029 BaselinebreastmilkIgG 0.84(0.75–0.96) 0.007 BreastmilkIgAatdose2 0.70(0.54–0.91) 0.007 BreastmilkIgGatdose2 0.87(0.77–0.98) 0.021

compliancetothebreastfeedinginstructionsinthegroupswhere breastfeedingwaswithheldaswellasthegroupwherebreastfeed- ingwasencouraged,theproportionofinfantswhoseroconverted wassimilarinthetwogroups.Theseresultsaresimilartothose reportedfromsimilarstudiesinSouthAfricaandPakistan[18,21].

Theoverallseroconversionrateinourstudywaslow,andfac- torsotherthanmaternalantibodiesarelikelytoberesponsiblefor thepoorimmunogenicityofthevaccine.ArecentRotarix^®trialin southIndiaexaminedtheeffectofprobioticandzincsupplemen- tationontheimmuneresponsetooralrotavirusandoralpoliovirus vaccines.Thisstudyreporteda35%seroconversionrateininfants whoreceivedthevaccinewithprobioticsupplementationand28%

ininfantswhoreceivedthevaccineandaplacebo.Inchildrenwho receivedthevaccinewithzincsupplementationtheseroconversion ratewas34%comparedto29%inthegroupreceivingthevaccine andaplacebo[20].TheinfantsinthestudyinsouthIndiawere ofthesameageastheinfantsinourstudyandinboth studies childhoodvaccinesweregivenalongwithRotarix^®.Alowsero- conversionratewasalsoseeninPakistanwheretheyexamined theimmuneresponsetoRotarix^®afterwithholdingbreastfeeding foranhourbeforeandaftergivingthevaccine[21].

AnimmunogenicitystudyofRotarixinIndiareporteda58.3%

seroconversionrate[22].Inthisstudy,themeanageofinfantsatthe timeofreceivingtheﬁrstandseconddosesofthevaccinewere8.7

(5)

and13.4weeksofage,comparedto6weeksinourstudyandinthe southIndianstudy.Also,inthisimmunogenicitystudy,aninterval oftwoweekswasmaintainedbetweenotherchildhoodvaccines andtherotavirusvaccinewhereasinourstudyandinthesouth IndianstudythechildhoodvaccinesweregivenalongwithRotarix.

SimilarﬁndingswereseenwiththeIndianrotavirusvaccine,ORV 116E,wheretheimmuneresponseinthephaseIa/IIbwasmuch higherthanreportedinthephaseIII(90%vs.40%)[10,23].Inthe phase1a/IIbtrial,infantswerearound8weeksoldatthetimeof receivingtheﬁrstdoseofthevaccineand therewasin interval oftwoweeksbetweenchildhoodvaccinesand116Ewhileinthe phaseIIItrial,infantswerearound6weeksoldandreceivedthe childhoodvaccinesalongwiththerotavirusvaccine.

Itis possiblethat inboth Rotarixand 116Eimmunogenicity studiestheslightlyhigherageatvaccinationand/ormaintainingan intervalbetweenchildhoodvaccinesandrotavirusvaccinespartic- ularlytheliveoralpoliovaccine,mayhaveimprovedtheimmune response.Ithasbeendescribedbeforethatco-administrationof oralpoliovirusvaccineinterferes withtheimmune responseto rotavirusvaccines[19,24,25],althoughpolioseroconversionrates arenotaffected.

Otherstudieshavereportedinverseassociationseenbetween maternalserumandbreastmilkIgAandIgGlevelsofinfantIgAlev- elspostdose2.The116Evaccineshowedaninverserelationship betweenlevelsofpre-existingrotavirusIgGandimmuneresponse tothevaccine[26].Inourstudy,preexisitngantibodiesatbaseline explainedonlyabout10%ofthevariabilityintheimmuneresponse tothevaccine.Althoughmaternalantibodiesimpairtheimmuno- genicity,otherfactorsseemtobemoreimportantandcontribute tothepoorimmuneresponse.

The protective role of maternal antibodies against rotavirus infectionisnotclear[13,14,27]althoughitissuggestiveofpro- tection[28,29].InthepreviouslymentionedstudyinPakistan,the seroconversionratewashigherinthegroupthatwasbreastfed aroundthetimeofvaccination,althoughthedifferencewasnotsta- tisticallysigniﬁcant.Evenifwithholdingbreastmilkatthetimeof vaccinationcouldmodifytheimmuneresponse,theimpactwould beminimalasthematernallevelsexplainedonlyafractionofthe variabilityintheimmuneresponses.

Alimitationofourstudywasthat thedurationof withhold- ingbreastfeedingaroundthetimeofvaccinationwasrestrictedto 30minbeforeandaftereachdose.Studieshaveshownthathalfgas- tricemptyingtimeininfantswhowerebreastfedvariedbetween 47and56min[30–32].Therefore,halfofthebreastmilkwould stillbepresentintheinfant’sstomachafteranhour.Withhold- ingbreastfeedingforanhourbeforeandaftervaccinationwould havebeenappropriatebutwasnotfeasibleinthisstudysetting.

Thistimeintervalwasalsousedinthepreviouslymentionedphase 1a/IIoralrotavirusvaccine116Etrialwhich demonstratedgood immunogenicity[23].However,therecentstudyfromsouthAfrica suggestthatincreasingthewindowfor withholdingbreastfeed- ingdoesnoteffecttheimmuneresponse[18].Additionally,inthis study,onlyinfantswhowerecurrentlybreastfedwereenrolled.It ispossiblethatmaternalantibodiestransferredtransplacentallyor throughbreastmilktotheinfantmayinterferewiththeimmune responseevenifmotherswithholdbreastfeedingaroundthetime ofvaccination.Theprevalenceofexclusivebreastfeedingwashigh atbaseline,whichisconsistentwithpreviousobservationsinthis population[33].

Seroconversion is not a direct indicator of clinical vaccine efﬁcacybut itis neverthelessimportant asa proxy for vaccine uptake.Mechanismsotherthanantibodylevelsmayexplainthe low immune response to rotavirus vaccines. It is worthwhile toexplore whetherinterferencewithotherintestinal infections or micronutrient deﬁciencies may modify immune responses [34,35].

In conclusion,withholding breastfeeding aroundthetime of vaccinationdidnotimprovetheimmuneresponsetoRotarix^®in Indianinfants.Thissuggeststhattheinterferenceofbreastmilk withthevaccine‘take’asassumedpreviouslymaynotbeofprac- ticalclinicalrelevance.

Conﬂictofinterest

Noneoftheauthorsdeclareanyconﬂictofinterest

Acknowledgements

This study was funded by the Research Council of Norway (projectnumber201208/S50).Wethanktheentirestudyteamfor theirsigniﬁcantcontributiontothesuccessofthisstudy.Wealso thankPankajVohra,thesafetyadvisor.Wegratefullyacknowledge alltheparticipantsfortheirwillingnesstocontributetoresearch.

References

[1]TateJE,BurtonAH,Boschi-PintoC,SteeleAD,DuqueJ,ParasharUD,etal.2008 estimateofworldwiderotavirus-associatedmortalityinchildrenyoungerthan 5yearsbeforetheintroductionofuniversalrotavirusvaccinationprogrammes:

asystematicreviewandmeta-analysis.LancetInfectDis2012;12:136–41.

[2]MorrisSK,AwasthiS,KheraA,BassaniDG,KangG,ParasharUD,etal.Rotavirus mortalityinIndia:estimatesbasedonanationallyrepresentativesurveyof diarrhoealdeaths.BullWorldHealthOrgan2012;90:720–7.

[3]Ruiz-PalaciosGM,GuerreroML,Bautista-MárquezA,Ortega-GallegosH,Tuz- DzibF,Reyes-GonzálezL,etal.Doseresponseandefﬁcacyofalive,attenuated humanrotavirusvaccineinMexicaninfants.Pediatrics2007;120:e253–61.

[4]VesikariT,KarvonenA,PrymulaR,SchusterV,TejedorJC,CohenR,etal.Efﬁcacy ofhumanrotavirusvaccineagainstrotavirusgastroenteritisduringtheﬁrst2 yearsoflifeinEuropeaninfants:randomised,double-blindcontrolledstudy.

Lancet2007;370:1757–63.

[5]VesikariT,ItzlerR,MatsonDO,SantoshamM,ChristieCD,CoiaM,etal.Efﬁcacy ofapentavalentrotavirusvaccineinreducingrotavirus-associatedhealthcare utilizationacrossthreeregions(11countries).IntJInfectDis2007;11:S29–35.

[6]MadhiSA,CunliffeNA,SteeleD,WitteD,KirstenM,LouwC,etal.Effectof humanrotavirusvaccineonseverediarrheainAfricaninfants.NEnglJMed 2010;362:289–98.

[7]ArmahGE,SowSO,BreimanRF,DallasMJ,TapiaMD,FeikinDR,etal.Efﬁcacyof pentavalentrotavirusvaccineagainstsevererotavirusgastroenteritisininfants indevelopingcountriesinsub-SaharanAfrica:arandomised,double-blind, placebo-controlledtrial.Lancet2010;376:606–14.

[8]ZamanK,DangDA,VictorJC,ShinS,YunusM,DallasMJ,etal.Efﬁcacyofpen- tavalentrotavirusvaccineagainstsevererotavirusgastroenteritisininfantsin developingcountriesinAsia:arandomised,double-blind,placebo-controlled trial.Lancet2010;376:615–23.

[9]BreimanRF,ZamanK,ArmahG,SowSO,AnhDD,VictorJC,etal.Analysesof healthoutcomesfromthe5sitesparticipatingintheAfricaandAsiaclinical efﬁcacytrialsoftheoralpentavalentrotavirusvaccine.Vaccine2012;30:A24–9.

[10]BhandariN,Rongsen-ChandolaT,BavdekarA,JohnJ,AntonyK,TanejaS,etal.

Efﬁcacyofamonovalenthuman–bovine(116E)rotavirusvaccineinIndian infants:arandomised,double-blind,placebo-controlledtrial.Lancet2014, http://dx.doi.org/10.1016/S0140-6736(13)62630-6.

[11]SackD,QadriF,SvennerholmA-M.Determinantsofresponsestooralvaccines indevelopingcountries.AnnNestle[Fr]2008:1–9.

[12]EmmettPM,RogersIS.Gastroenteritis,diarrhoeaandbreastfeeding.EarlyHum Dev1997;49(Suppl.):S83–103.

[13]GlassRI,StollBJ,WyattRG,HoshinoY,BanuH,KapikianAZ,etal.Observations questioningaprotectiveroleforbreast-feedinginsevererotavirusdiarrhea.

ActaPaediatrScand1986;75:713–8.

[14]ClemensJ,RaoM,AhmedF,WardR,HudaS,ChakrabortyJ,etal.Breast-feeding andtheriskoflife-threateningrotavirusdiarrhea:preventionorpostpone- ment?Pediatrics1993;92:680–5.

[15]MoonSS,WangY,ShaneAL,NguyenT,RayP,DennehyP,etal.Inhibitoryeffect ofbreastmilkoninfectivityofliveoralrotavirusvaccines.PediatrInfectDisJ 2010;29(October(10)):919–23.

[16]GoveiaMG,DiNubileMJ,DallasMJ,HeatonPM,KuterBJ.Efﬁcacyofpentavalent humanbovine(WC3)reassortantrotavirusvaccinebasedonbreastfeeding frequency.PediatrInfectDisJ2008;27:656–8.

[17]PatelM,ShaneAL,ParasharUD,JiangB,GentschJR,GlassRI,etal.Oralrotavirus vaccines:howwellwilltheyworkwheretheyareneededmost?JInfectDis 2009;200:S39–48.

[18]GroomeMJ,MoonSS,VelasquezD,JonesS,KoenA,vanNiekerkN,etal.Effect ofbreastfeedingonimmunogenicityoforallive-attenuatedhumanrotavirus vaccine:arandomizedtrialinHIV-uninfectedinfantsinSoweto,SouthAfrica.

BullWorldHealthOrgan2014;92:238–45.

(6)

[19]WardRL,BernsteinDI,SmithVE,SanderDS,ShawA,EidenJJ,etal.Rotavirus immunoglobulinAresponsesstimulatedbyeachof3dosesofaquadrivalent human/bovinereassortantrotavirusvaccine.JInfectDis2004;189:2290–3.

[20]JohnJ,SarkarR,TateJE,FlemmingJ,MosesP,MuliylJ,etal.Effectofsupple- mentationwithzincandprobioticsonimmuneresponsestooralvaccines indevelopingcountries.In:PresentedattheVaccinesforEntericDiseases, November6–8,2013.2013.

[21]AliAS,KaziM,CorteseM,FlemingJ,ParaharU,JiangB,etal.Impactofwith- holdingbreastfeedingaroundthetimeofvaccinationontheimmunogenicity ofthehumanrotavirusvaccineinPakistan—arandomizedcontrolledtrial.In:

PresentedattheVaccinesforEntericDiseases,November6–8,2013.2013.

[22]NarangA,BoseA,PanditAN,DuttaP,KangG,BhattacharyaSK,etal.Immuno- genicity,reactogenicityandsafetyofhumanrotavirusvaccine(RIX4414)in Indianinfant.HumVaccine2009;5:414–9.

[23]BhandariN,SharmaP,TanejaS,KumarT,Rongsen-ChandolaT,AppaiahgariMB, etal.Adose-escalationsafetyandimmunogenicitystudyofliveattenuated oralrotavirusvaccine116Eininfants:arandomized,double-blind,placebo- controlledtrial.JInfectDis2009;200:421–9.

[24]VodopijaI,BaklaicZ,VlatkovicR,BogaertsH,DelemA,AndreFE,etal.Combined vaccinationwithliveoralpoliovaccineandthebovinerotavirusRIT4237strain.

Vaccine1986;4:233–6.

[25]MigasenaS,SimasathienS,SamakosesR,PitisuttithamP,SangaroonP,van SteenisG,etal.Simultaneousadministrationoforalrhesus-humanreassor- tanttetravalent(RRV-TV)rotavirusvaccineandoralpoliovirusvaccine(OPV) inThaiinfants.Vaccine1995;13:168–74.

[26]AppaiahgariMB,GlassR,SinghS,TanejaS,Rongsen-ChandolaT,Bhandari N,etal. Transplacentalrotavirus IgGinterfereswithimmuneresponseto

liveoralrotavirus vaccineORV-116E inIndian infants.Vaccine2014;32:

651–6.

[27]GoldingJ,EmmettPM,RogersIS.Gastroenteritis,diarrhoeaandbreastfeeding.

EarlyHumDev1997;49:S83–103.

[28]StollBJ.Theprotectiveeffectofhumanmilkagainstdiarrhea.Areviewof studiesfromBangladesh.ActaPaediatrScandSuppl1989;351:131–6.

[29]BlakePA,RamosS,MacDonaldKL,RassiV,GomesTA,IveyC,etal.Pathogen- speciﬁcriskfactorsandprotectivefactorsforacutediarrhealdiseaseinurban Brazilianinfants.JInfectDis1993;167:627–32.

[30]StaelensS,VandenDriesscheM,BarclayD,Carrié-FaesslerAL,HaschkeF, VerbekeK,etal.Gastricemptyinginformula-fedandbreast-fedinfantsmea- suredwiththe13C-octanoicacidbreathtest.JPediatrGastroenterolNutr 1999;29:46–51.

[31]CavellB.Gastricemptyingininfantsfedhumanmilkorinfantformula.Acta PaediatrScand1981;70:639–41.

[32]BilleaudC,GuilletJ,SandlerB.Gastricemptyingininfantswithorwith- outgastro-oesophagealreﬂuxaccordingtothetypeofmilk.EurJClinNutr 1990;44:577–83.

[33]BhandariN,MazumderS,BahlR,MartinesJ,BlackRE,BhanMK,etal.Aneduca- tionalinterventiontopromoteappropriatecomplementaryfeedingpractices andphysicalgrowthininfantsandyoungchildreninruralHaryana,India.J Nutr2004;134:2342–8.

[34]SabettaJR,DePetrilloP,CiprianiRJ,SmardinJ,BurnsLA,LandryML,etal.

Serum25-hydroxyvitamindandtheincidenceofacuteviralrespiratorytract infectionsinhealthyadults.PLoSONE2010;5:e11088.

[35]Tavera-MendozaLE,WhiteJH.Celldefensesandthesunshinevitamin.SciAm 2007;297(62–65),68–70,72.