EVADOCprojectVietnam June-December2014 4

1

ModelingtheeffectivenessofavianinfluenzavaccinationstrategiesinVietnam:

evaluationoftheaddedvalueofdayoldchickvaccinationinhatcheries

GeneralcontextandobjectivesoftheprojectVaccination against avian influenza (AI) is currently being implemented worldwide mostly usinginactivatedvaccinesthatarenotapplicabletodayoldchicks(DOCs).SinceNovember2012,anovelrecombinantHVT(HerpesvirusofTurkey)AIvaccinehasbeenrecentlycommercializedandappliedinsomeindustrialhatcheriesinVietnam.Inprinciple,vaccinationofDOCsathatcheries,withtheuseoflimitedresources,wouldensuresufficientvaccinecoverageandprotectionlevels.Theobjectivesofthisstudyweretoassessthecost-effectivenessofAIDOCvaccinationinhatcheriesandfeasibilityofimplementingAIDOCvaccinationinthedifferentproductionsectorsinVietnam.

AmodeloftheVietnampoultryproductionnetworkwascombinedwithaflockimmunitymodeltosimulate thedistributionprofileofAI immunityaccording to thedifferent typesofpoultry farming(i.e. Grand Parents, breeders, layers or broilers) and production sectors (1 to 4); and vaccinationscenario (includingabsenceorpresenceofDOCvaccination).Themodelestimated thevaccinationcoverage(proportionofvaccinatedbirdsfromthetotalpopulationconsidered),theprotectivesero-conversionlevel(birdswithheamagluttinininhibition(HI)titersgreaterthan4log2)andthedurationofimmunity(numberofweekswheremorethan70%ofthebirdshadaprotectivesero-conversionlevel)foreachnodeofthenetworkandvaccinationscenario.

Publicprivatepartnership(PPP)inVeterinaryPublicHealthThiswork is the product of a public-private partnership between CIRAD, a French research publicinstituteforAgriculturaldevelopment,theNationalUniversityofAgricultureofVietnam(VNUA)andCEVA. PPP aims to create conditions of partnership between Public and Private Stakeholders: toreach a common goal by developing complementary means and skills within a given perimetersupportingrespectiveinterests.CEVAcontributestotheprojectintermsofvaccinedevelopmentandprivate sector understanding;while CIRAD and VNUAbrings in technical competency for scientificapproachforevaluationofcontrolstrategiesfromthelocaltonationallevel.

EVADOCprojectVietnam June-December2014 4

2

ProjectorganizationA firstmappingof thepoultry industrywasperformedbyCEVA in2013. In2014a field studywasimplemented in collaboration between VNUA, CIRAD and CEVA experts to collect additionalinformationonpoultryproduction(typeofpoultryproduction(s),farmtotalcapacityperyear,totalvolumeofDOCproductionperyear,originanddestinationoftheDOCs)andAIvaccinationstrategy((vaccine coverage, vaccine type, vaccination protocol and perception of AI vaccination (includingDOCvaccination).Quantitativedataonavianinfluenzavaccineefficacywerecollectedfromscientificandgreyliteratures(reportsonlaboratoryandfieldevaluationinVietnam)orobtainedfromVNUA,CIRADandNIVR.Thenetworkanalysis and immunitymodelingwas thenperformedbyCIRADandVNUA and the results were presented for validation during an expert workshop in October 2014gathering experts from public and private sectors, from laboratory, epidemiology and poultryfarmersfromallsectorsandfrominternationalorganisations.

VaccinationscenariosSix vaccination scenarios were identified based on centrality and connectivity analysis of thestructureofthepoultryproductionnetworkandageatvaccination(vaccinationofbirds>7daysoldand/or DOCs) (Table 1). DOC vaccination of long cycle birds (breeders and layers) was onlyconsideredasaprime-booststrategy(vaccinationatdayoldinhatcheriesandboostat20weeksinbreederfarms);whereasDOCbroilervaccinationwasconsideredasonesingleshotofvaccine.

Table1.Descriptionofthevaccinationprotocolsconsideredwithinthedifferentvaccinationscenariosconsideredandtestedbythemodel

DOCsource IntegratedhatcheriesNon-Integrated

hatcheriesLargeandmedium

Governmenthatcheries

Non-Integratedhatcheries

Small(nativebirds)

DOCdestinationIntegration(Sector1)

ContractFarms(Sector2and3)

ClientsSmallfarms(Sector3)

Mediumandsmallfarms

(Sector2and3)

Mediumandsmallfarms

(Sector2and3)

Smallfarms(Sector3)

Scenario1Farmvaccination

Farmvaccination

Farmvaccination

FarmvaccinationFarmvaccination

Farmvaccination

Scenario2 Hatchery HatcheryFarm

vaccinationFarmvaccination

Farmvaccination

Farmvaccination

Scenario3 Hatchery Hatchery Hatchery FarmvaccinationFarmvaccination

Farmvaccination

Scenario4 Hatchery Hatchery Hatchery HatcheryFarmvaccination

Farmvaccination

Scenario5 Hatchery Hatchery Hatchery Hatchery HatcheryFarmvaccination

Scenario6 Hatchery Hatchery Hatchery Hatchery HatcheryHatchery

EVADOCprojectVietnam June-December2014 4

3

NetworkanalysisofVietnamchickenproductionsystemModels of the Vietnam broiler and layer production networks were developed based on type ofpoultry production (i.e. Grand Parents, breeders, layers or broilers) and farm production sizes(numberofheadsandvolumeofproductionforaoneyearperiod)(Figure1).Eachnoderepresentedatypeofpoultryproduction;eachlinkrepresentedeitherthemovementofeggs(betweenbreederfarmsandhatcheries)orthemovementofDOCs(betweenhatcheriesandcommercialbroiler/layerfarms).

Figure 1. Vietnam broiler (A), layer (B) and native (C) poultryproduction network models. The types and colors of nodesrepresent the different types of productions: integrated (red);contract (green), government (light blue) or independent farms(black) of grand-parents (GP), commercial broilers (Bro) or layers(Lay). The size of the nodes represents the number of birds pernodes. The arrows represent the movement of eggs or day oldchicks (DOC- between the nodes and the size of the arrows thevolumeofeggsorDOCforeachmovement.

InterestofusingnetworkanalysisapproachtodisplaypoultryproductionsysteminVietnam:

• TopresentthedynamicmodelofDOCflowwithinthepoultryproductionnetwork

• Toidentifymostimportantactorstotargetsurveillanceandcontrolstrategies

• Tocombinewithothermodels(e.g.immunityor/andinfectionmodels)topredictoutcomes

ofsurveillanceandcontrolstrategies

A B

C

EVADOCprojectVietnam June-December2014 4

4

AddedvalueofDOCvaccinationforcommercialbroilersandlayersUnder the current strategy (scenario1= farmvaccinationwith inactivatedvaccines),onlya limitedpercentage of the total poultry population (<40%, only long cycle birds (GPs and breeders)) hassufficient level of immunity against avian influenza (vaccine coverage >80% and sero-protection>60%)(Figure2Aand2B).WiththeintroductionofhatcheryvaccinationofbroilerDOCproducedandraisedintheintegratedcompanies(Scenario3),themodelpredictsthattheimmunitylevelforthewholebroilerpopulationswillriseaboveprotectivelevels(Figure2Aand2B).Moreover,thedurationof sero-protection in in thebroiler population, under the current scenario (vaccinatedat the farmwithinactivatedvaccine))iscurrentlybelow20%ofthebirdsproductiontime(Figure2C).HatcheryvaccinationofDOCwillincreasethisprotectiontimeby20-40%accordingtothetypeofproduction(Scenario3and4)(Figure2C).

A B

C

Figure 2. Evolution of the vaccine coverage rate (A), the positive sero-conversion rate (B) and theduration of sero-protection (C) for broiler bird population, expressed as a percentage of the birdpopulationperpoultryproduction typeand according to thedifferent vaccinationscenarios tested inthemodel:Sc1=noDOCvaccination(farmvaccinationonlywith inactivatedvaccine);Sc2=vaccinationofDOC in integrated farms; Sc3= vaccinationofall DOC producedby integrationDOC (including soldclients); Sc4=Sc3+DOCvaccination in large/mediumhatcheries; Sc5=DOCvaccination inmediumallhatcheries.

EVADOCprojectVietnam June-December2014 4

5

Forthelayerpopulationsimilarresultsareobserved(datanotshown),withsufficientimmunityanddurationofimmunityabove60%fromScenario3.

Thesituationisdifferentforthenativepopulationwhichismainlyproducedbymediumandsmallbreederfarms:theimmunityinthetotalnativepopulationwillincreaseifhatcheryvaccinationwasintroduceduptothemediumsizeindependentbreederfarmsbutwillonlyreachsufficientlevelsisDOCsproducedbysmallbreederfarmswerevaccinated(Figure3).

Commercialpoultrybestscenario=Sc3-4(vaccinationofallDOCproducedbytheintegrationsandlargebreeders)

! Totalbroilerandlayerpopulationvaccinecoverage>80%

! Totalboilerpopulationsero-protection>60%;layerpopulationsero-protection~80%

! Totalbroilerpopulationdurationsero-protection>50%;layerdurationsero-protection~60%

Nativepoultrybestscenario=Sc5-6(vaccinationofDOCuptomediumandsmallbreeders)

! Totalnativepopulationvaccinecoverage=80%

! Totalnativepopulationsero-protection=60%

! Totalnativepopulationdurationsero-protection>40%

Figure3.Evolutionofthevaccinecoveragerate(A),thepositivesero-conversionrate(B)andthedurationofsero-protection(C)fornativebirdpopulation,expressedasapercentageofthebirdpopulation per poultry production type and according to thedifferent vaccination scenarios tested in the model: Sc1= noDOC vaccination (farm vaccination only with inactivatedvaccine); Sc2= vaccination of DOC in integrated farms; Sc3=vaccination of all DOC produced by integrationDOC (includingsold clients); Sc4= Sc3 + DOC vaccination in large/mediumhatcheries;Sc5=DOCvaccinationinmediumallhatcheries;Sc6=AllDOCvaccination.

A B

C

EVADOCprojectVietnam June-December2014 4

6

SpatialdistributionofAIimmunityaccordingtothedifferentvaccinationstrategiesAsthepoultrypopulationdensityisnothomogeneouswithinthecountryitisimportanttoassessthedistributionoftheimmunitylevelsatthespatiallevel(Figure4).ThespatialdistributionprofileofAIimmunitywasheterogeneousacrosstheprovinces,whichreflectedthisspatialclusterdistributionofthedifferentpoultryproductiontypes.However,thebestscenariotoensuresufficientprotectionintheareasclassifiedathighriskofinfectionwasthesameasforthehomogeneouspopulationmodeloutputsdescribedbefore(Scenario3and4)(Figure4).

Figure4.SpatialdistributionofthebirdpopulationimmunityagainstAIaccordingtothedifferentvaccinationscenariotestedinthemodel.Foreachscenariothepercentageofvaccinecoverage,protectivesero-conversionanddurationofimmunity(modeloutputs)wereweightedaccordingtothepoultrydensity(foreachproductiontype)perprovince. Thearea classifiedathigh riskofHPAI infection (according to the cumulatednumberofpoultryoutbreakssince2004)arecircledinred.

EVADOCprojectVietnam June-December2014 4

7

ConclusionThis study demonstrated the interest of combining network analysis and immunity modeling toassesstheefficacyofAIvaccinationscenarios.ThemodelpredictedthattargetingDOCAIvaccinationwould increase immunity levels in theoverallpoultrypopulationup to sufficient levels to improveHPAIdiseasecontrolincommercialpoultry.ThemodelpredictedthattargetingDOCAIvaccinationinintegratedhatcherieswouldincreaseimmunitylevelsintheoverallpoultrypopulationandespeciallyinsmallcommercialpoultryfarms.DOCvaccinationstrategywasshowntobemoreefficientthanthecurrent strategy using inactivated vaccines. Improving HPAI control in commercial poultry sectorcould have positive spill over effect on the epidemiological situation of the disease in native andbackyard poultry. This model could be applied for strategic management of other contagiousdiseasessuchasNewcastleDiseases.

• AIfarmvaccinationstrategy(inactivatedvaccine)(Sc1):! Allowsforsufficientcoverageandsero-protectioninlongcycleproduction(e.g.

breedersandlayers)! Provideslimiteddurationofprotectioninlongcyclebirds! Notefficientforbroilersandnativebirds

• DOCvaccination(vectoredvaccine)(Sc3-Sc5):! Improvesdurationofsero-protectioninlongcyclebirds! Providessufficientvaccinecoverageandimmunitylevelsinbroilerseveninsmall

farms! DOCvaccinationtargetedinlargeintegratedhatcheries(Sc3)allowsforsufficient

immunityanddurationofimmunityinthetotalpopulationespeciallyinhighHPAIriskareas.

AcknowledgmentsThis work was supported by funds from CEVA. The report was prepared by Marisa Peyre, CIRAD withcollaboration fromHiepDaoThi,VNUAandVuDinhTon,VNUA.Theauthorswould like to thankThiThanhPhamHoaandJuliePecqueurfortheirvaluablesupportandfacilitationoftheimplementationofthestudy.Wewould like to thank all the study participants and all the people involved in the expert meeting for theirvaluablecontribution.