Cytomegalovirus infection and responsiveness to influenza vaccination in elderly residents of long-term care facilities

(1)

Cytomegalovirus infection and responsiveness to influenza vaccination in elderly residents of long-term care facilities

Elzen, W.P.J. den; Vossen, A.C.M.T.; Cools, H.J.M.; Westendorp, R.G.J.; Kroes, A.C.M.;

Gussekloo, J.

Citation

Elzen, W. P. J. den, Vossen, A. C. M. T., Cools, H. J. M., Westendorp, R. G. J., Kroes, A. C.

M., & Gussekloo, J. (2011). Cytomegalovirus infection and responsiveness to influenza vaccination in elderly residents of long-term care facilities. Vaccine, 29(29-30), 4869-4874.

doi:10.1016/j.vaccine.2011.03.086

Version: Not Applicable (or Unknown)

License: Leiden University Non-exclusive license Downloaded from: https://hdl.handle.net/1887/117588

Note: To cite this publication please use the final published version (if applicable).

(2)

ContentslistsavailableatScienceDirect

Vaccine

j o ur na l ho me p ag 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

Cytomegalovirus infection and responsiveness to inﬂuenza vaccination in elderly residents of long-term care facilities

Wendy P.J. den Elzen

^a,∗

, Ann C.M.T. Vossen

^b

, Herman J.M. Cools

^a

, Rudi G.J. Westendorp

^c,d

, Aloys C.M. Kroes

^b

, Jacobijn Gussekloo

^a

aDepartmentofPublicHealthandPrimaryCare,LeidenUniversityMedicalCenter,Leiden,TheNetherlands

bDepartmentofMedicalMicrobiology,LeidenUniversityMedicalCenter,Leiden,TheNetherlands

cDepartmentofGerontologyandGeriatrics,LeidenUniversityMedicalCenter,Leiden,TheNetherlands

dNetherlandsConsortiumforHealthyAgeing,Leiden,TheNetherlands

a r t i c l e i n f o

Articlehistory:

Received2December2010

Receivedinrevisedform16March2011 Accepted22March2011

Available online 15 April 2011

Keywords:

Cytomegalovirus Cytomegalovirusinfections Inﬂuenza

Inﬂuenzavaccines Aged

Long-termcarefacility

a b s t r a c t

Ampleevidencesuggeststhatinfectionwithcytomegalovirus(CMV)leadstoacceleratedagingofthe immunesystemandmaycontributetopoorresponsivenesstoinﬂuenzavaccinationinolderpersons.

TheobjectiveofthisstudywastoinvestigatewhetherCMVinfection,acquiredearlierinlife,affectsthe responsetoinﬂuenzavaccinationinarandomizedcontrolledtrialamongolderpersonsinlong-termcare facilities.

Duringthe1997–1998influenzaseason,731residents(medianage83[interquartilerange78–88], 75.4%female)in14long-termcarefacilitiesintheNetherlandswererandomlyassignedtoreceive15 or30␮gofinactivatedinfluenzavaccine,followedbya15␮gboostervaccineoraplacebovaccineat day84.Bloodsampleswerecollectedatday0,day25,day84andday109.Seroresponsestoinfluenza vaccinationweremeasuredbyhemagglutination-inhibitionteststotheA/H3N2strainatalltimepoints.

Subsequently,baselinelevelsofIgGanti-CMVantibodiesweremeasuredusinganautomatedchemilu- minescentmicroparticleimmunoassay.ParticipantswithCMVantibodylevel≥6AU/mLwereconsidered toharborCMVinfection.

Atbaseline,nodifferencesinpre-vaccinationgeometricmeanantibodytiters(GMT)wereobserved betweenparticipantswith(n=571,78.1%)orwithoutCMVinfection(n=160,21.9%).Duringfollow-up, participantswithandwithoutCMVinfectionhadsimilarresponsestoinﬂuenzavaccinationasmeasured withchangesinGMT(linearmixedmodel,adjustedforgender,age,pre-vaccinationGMTandvaccination strategy,p=0.46).Analogously,noassociationwasfoundbetweenCMVinfectionandamorethan4-fold increaseinantibodytiter(GeneralizedEstimatingEquations,adjustedOR1.14[95%CI0.80;1.64])oran antibodytiter≥40(adjustedOR1.24[95%CI0.86;1.80]).

Inconclusion,CMVinfectiondidnotexplainpoorresponsivenesstoinﬂuenzavaccinationinresidents oflong-termcarefacilities.

1. Introduction

Agingisassociated withan increasedsusceptibilitytoinfec- tionsandareducedresponsetovaccination[1–4].Theage-related declineintheprotectiveimmuneresponsecanmostlybeexplained byreplacementofnaïveTcellsbymemoryTcellsandadecrease inthediversityandfunctionoftheTcellpopulation[3–5].

Accumulating evidence suggests that infection with cytomegalovirus(CMV)contributestotheage-associatedchanges inimmunity[5–8].CMVseroprevalencevariesbetweencountries

∗ Correspondingauthor.Tel.:+31715268444;fax:+31715268259.

E-mailaddress:w.p.j.den elzen@lumc.nl(W.P.J.denElzen).

butingeneralriseswithadvancingagefrom60%inpersonsaged 40–49to>90%inpersonsaged80andover[9–11].Onceinfected withCMV,theimmunesystemisnotabletoeliminatethevirus[5], resultinginlatentCMVinfection.Mostinfected personsremain freeofclinicalsymptomsbecauseofefﬁcientCMVimmunosurveil- lance[12].ThepresenceofCMVinfectionisconsideredtobethe drivingforcebehindtheoligoclonalexpansionsofTcellsobserved inolderpersons[5,12,13].

Although this immunological imprint of CMV infection on theTcellpopulationis widelyrecognized,theclinicaleffectof CMVinfectioninolderpersonsislargelyunknown.Interestingly, Trzonkowski and colleagues observed higher concentrations of anti-CMVantibodiesinnursinghomeresidentsandstaffthatdid notshowaserologicalresponsetoinﬂuenzavaccination[14].It 0264-410X/$–seefrontmatter © 2011 Elsevier Ltd. All rights reserved.

doi:10.1016/j.vaccine.2011.03.086

(3)

4870 W.P.J.denElzenetal./Vaccine29 (2011) 4869–4874

maythereforebehypothesizedthattheresponsetoinﬂuenzavac- cinationwillbelowerinolderpersonswithCMV infectiondue toafunctionalimpairmentofimmuneeffectormechanisms.The demonstrationofsuchaneffectwouldofferopportunitiesforopti- mizingvaccinationstrategiesbypriordeterminationoftheCMV statusinthispopulation.Therefore,weinvestigatedwhetherthe presenceofCMVinfectionaffectstheresponsetoinﬂuenzavacci- nationinalargerandomizedcontrolledtrialamongolderpersons inlong-termcarefacilitiesintheNetherlands.

2. Materialandmethods 2.1. Studypopulationanddesign

The current study is embeddedin a randomized controlled multicentertrialofinﬂuenzavaccinationstrategiesinlong-term carefacilitiesthatwasconductedintheNetherlandsduringthe 1997–1998inﬂuenza season.The trial hasbeen described pre- viously in detail [15,16]. In short, 2444 residents of 14 Dutch long-term care facilities were invited by mail to participate.

Informedconsentwasobtainedfrom815residentsortheirlegal representatives.Thestudywasapproved by theMedical Ethics CommitteeoftheLeidenUniversityMedicalCenter.Participants wererandomlyassignedtooneoffourtreatmentarmsbystrat- iﬁedblockrandomization usingrandomnumber tablesforeach long-termcarefacility.Thefourtreatmentarmswere:(1)a15␮g doseofvaccineonday0andaplacebovaccineonday84,(2)a 15␮gdoseofvaccineonday0anda15␮gdoseofboostervaccine onday84,(3)a30␮gdoseofvaccineonday0andaplacebovac- cineonday84,and(4)a30␮gdoseofvaccineonday0anda15␮g doseofboostervaccineonday84.Theinitialvaccinedose(15or 30␮g)wasgivenasoneortwo0.5mLintramuscularinjectionsin thesamearm.

2.2. Vaccines

Thetrivalentsplitvirusvaccineforthe1997–1998inﬂuenza season (Pasteur Mérieux, Connaught, France) contained an amount of split virus equivalent to 15mg hemagglutinin of A/Nanchang/933/95(H3N2), A/Johannesburg/82/96(H1N1), and B/Harbin/7/94.Anidenticallotnumberwasusedforallvaccina- tions.The placebo vaccineexistedof phosphate bufferedsaline (PBS).

2.3. Laboratorymeasurements

Bloodsampleswerecollectedatday0,day25,day84andday 109andstoredat−20^◦C.In2007,pre-andpost-vaccinationtiters weremeasured. Alltests wereperformed in triplicate.Because influenzaA/H3N2wasthepredominantvirussubtypesince1968 until2009,weperformedthehemagglutinationinhibition(HAI) test tothe A/H3N2 strain.The assay conditionsincluded using turkeyerythrocytesforagglutination,afiltrateofVibriocholerae fimbriaeassourceofreceptordestroyingenzymeandfourhemag- glutinatingunitsofthevaccinestrainA/Nanchang/933/95(H3N2), whichwaspropagatedonMadineDarbycaninekidneycells[17].

Pre-vaccinationantibodytitersandserologicalresponsestovac- cinationarepresentedandanalyzedinthreeways:(1)geometric meanantibodytiters(GMT)with95%conﬁdenceintervals,(2)HAI antibodytiter≥4-foldcomparedtoday0,and(3)seroprotection deﬁnedasHAIantibodytiter≥40[18].

In2009,IgGanti-CMVantibodiesweremeasuredusinganauto- matedchemiluminescentmicroparticleimmunoassay(Architect, AbbottLaboratories,AbbottPark,IL).ParticipantswithaCMVanti- bodylevel ≥6AU/mLwereconsideredtoharbor CMVinfection.

ParticipantswithCMVinfectionwerefurtherdividedinagroup

ofparticipantswithaCMVantibodylevelof6.0–249.9AU/mLand agroupofparticipantswithaCMVantibodylevel≥250AU/mL.

2.4. Otherclinicalparametersatbaseline

Demographicandmedicaldatawerecollectedatday0from multidisciplinarypatientﬁles.Thesix-itemlistbyKatzetal.was usedtoassessdisabilityinactivitiesindailyliving(ADL)[19].

2.5. Statisticalanalysis

Baselinedifferences incontinuousdatabetweenparticipants inthefourtreatmentarmsweretestedwithJonkheere-Terpstra tests.DifferencesincategoricaldataincludingCMVinfectionwere testedwithChisquaretests.Differencesinpre-vaccinationGMT were tested with One Way ANOVA. Differences between participants withand without CMV infection were analyzed with Mann-WhitneyUtestsforcontinuousdata,Chisquaretestsfor categorical dataand independentt-tests for differencesin pre- vaccinationGMT.

Withineachtreatmentarm,seroresponsestoinﬂuenzavacci- nationatday25,day84 andday109werecompared between participantswithandwithoutCMVinfection.DifferencesinGMT weretestedwithindependentt-tests.Differencesinthepercent- ageofparticipantswithanantibodytiter≥4-foldcomparedtoday0 anddifferencesinseroprotectionrate(percentagewithanantibody titer≥40)weretestedwithChisquaretests.

The effect of CMV infection on GMT prior or after vaccina- tionduringfollow-upwasinvestigatedwithlinearmixedmodel analysis.Predictedmeansand 95%CIwereadjustedforgender, age,pre-vaccinationGMTandvaccinationstrategy.AGeneralized EstimatingEquationsapproachforbinarydatawasusedtoinves- tigatethe effectof CMV infection onan antibodytiter ≥4-fold andtoinvestigatetheeffectonseroprotectionrate(antibodytiter

≥40).Again,theseanalyseswereadjustedfor gender,age,pre- vaccinationGMTandvaccinationstrategy.

DatawereanalyzedusingSPSS16.0 forWindows(SPSSInc., Chicago,IL).

3. Results

Thenumbersofparticipantsineachtreatmentgroupateach timepointduringfollow-uparepresentedinFig.1.Atday0,blood sampleswerecollectedfor731outof815participantsthatgave informedconsent.Themedianageofthetotalstudypopulationwas 83years(interquartilerange[IQR]78–88)and75.4%ofthepopula- tionwasfemale.CMVinfection(CMVantibodylevel≥6.0AU/mL) wasobservedin571participants(78.1%).Table1showsthechar- acteristicsofthestudypopulationforeachofthefourtreatment arms.Therewerenodifferencesinsocio-demographiccharacter- istics,clinicalcharacteristics,ADLdisabilityscore,pre-vaccination GMTandCMVinfectionbetweenthegroups.

Table2describesthecharacteristicsof thestudypopulation dependingonthepresenceofCMVinfection.Characteristicsofpar- ticipantswithCMVinfectiondidnotdiffersubstantiallyfromthose withoutCMVinfectionandvaccinationstrategieswereequallydis- tributed.Therewasnodifferenceinpre-vaccinationGMTbetween bothgroups (p=0.32).Moreover,weobservednodifferencesin pre-vaccinationGMTbetweenthoseparticipantswithCMVanti- bodylevels<6AU/mL(mean23.1[95%CI18.2–29.2]),thosewith CMV antibodylevelsbetween6.0 and 249.9AU/mL (mean19.7 [95%CI16.5–23.6])andthosewithCMVantibodylevel≥250AU/mL (mean20.6[95%CI17.9–23.8],One-WayANOVA,p=0.54).Inaddi- tion,no differencein pre-vaccinationGMT wasfound between thoseparticipantswithCMVantibodylevel≥250AU/mLandthose withCMVantibodylevelbetween6.0and249.9AU/mL(p=0.85).

(4)

Eligibleresidents N=2444

Informedconsent N=815

30µg + placebobooster

N=201 15µg +

15µg booster N=204 15 µg +

placebo booster N=206

30µg + 15 µg booster

N=204

Day0

Day25

Day84

Day109

N=176

CMV+, N=136 CMV-, N=40

N=184

CMV+, N=142 CMV-, N=42

N=182

CMV+, N=143 CMV-, N=39

N=189

CMV+, N=150 CMV-, N=39

N=170

CMV+, N=132 CMV-, N=38

N=177

CMV+, N=135 CMV-, N=42

N=176

CMV+, N=137 CMV-, N=39

N=184

CMV+, N=146 CMV-, N=38

N=157

CMV+, N=122 CMV-, N=35

N=168

CMV+, N=131 CMV-, N=37

N=169

CMV+, N=132 CMV-, N=37

N=171

CMV+, N=136 CMV-, N=35

N=155

CMV+, N=119 CMV-, N=36

N=168

CMV+, N=130 CMV-, N=38

N=158

CMV+, N=122 CMV-, N=36

N=163

CMV+, N=130 CMV-, N=33

Fig.1.NumberofparticipantsandCMVstatusduringfollow-up.

InTable3 wecompared seroresponsestoinﬂuenzavaccina- tionatday25,day84andday109betweenparticipantswithand withoutCMVinfection,separatelywithineachtreatmentarmof thestudy.We didnotobserveanydifferencesinGMT between participantswithorwithoutCMVinfectionatanytimepoint.

InadditiontoanalyzingGMTastheoutcomevariable,wealso analyzedseroresponse(antibodytiter≥4-foldcomparedtoday0) andseroprotectionrate(antibodytiter≥40)asoutcomeswithin eachtreatmentgroup(Table3).Exceptforthoseparticipantsthat hadreceivedadoseof30␮gandplaceboboosteratday109,we observednostatisticallysigniﬁcantdifferencesinseroresponseand seroprotection ratebetween participantswithor without CMV infectionatanytimepointduringfollow-up.

Finally, we performed additional analyses beyond the vari- oustreatmentstrategies. We used linearmixed modelanalysis toinvestigatetheeffectofCMVinfectiononGMTseroresponsive-

nessuponvaccinationwithinﬂuenzaafteradjustmentforunequal distributionsofgender,age,pre-vaccinationGMTandvaccination strategy.NodifferenceinGMTaftervaccinationduringfollow-up wasfoundbetweenparticipantswithandwithoutCMVinfection (Fig.2,p=0.46).Analogously,noassociationwasfoundbetween CMVinfectionandamorethan4-foldincreaseinantibodytiter frombaseline(adjustedOR1.14(95%CI0.80;1.64,p=0.47),orwhen anantibodytiter≥40wastakenasanendpoint(adjustedOR1.24 (95%CI0.86;1.80,p=0.25).

4. Discussion

Inthepresentstudy,wehaveshownthatCMVinfection,deﬁned asaCMVantibodylevel≥6.0AU/mL,doesnotaffecttheresponse toinﬂuenzavaccinationinolderindividualsinlong-termcarefacilities. Thislack of effectis independent of thedose, number of

Table1

Socio-demographic,functionalandclinicalcharacteristicsoftheparticipantsatbaselinedependingonvaccinationstrategy.

Vaccinationstrategy p-value

Dose15␮g+placebo (n=176)

Dose15␮g+15␮g booster(n=184)

Dose30␮g+placebo booster(n=182)

Dose30␮g+15␮g booster(n=189)

Age(years) 83(78–88) 83(77–87) 83(77–88) 84(78–88) 0.86

Females 125(71.0%) 143(77.7%) 139(76.4%) 144(76.2%) 0.47

Lengthofstay(months) 22(8–48) 22(10–46) 23(10–42) 20(9–44) 0.63

Katz-score 8(6–10) 8(4–9) 8(5–9) 7(4–9) 0.08

Numberofmedicaments 5(3–6) 4(3–6) 4(2–6) 4(3–6) 1.00

Diagnosisofdementia 122(69.3%) 126(68.5%) 122(67.0%) 126(66.7%) 0.94

Inﬂuenzastatus

Pre-vaccinationGMT 22.7(18.1–28.4) 22.0(18.0–27.0) 18.4(15.2–22.4) 20.6(17.1–24.8) 0.50

Highpre-vaccinationtiter(≥40) 70 (39.8%) 72 (39.1%) 66(36.3%) 75(39.7%) 0.89

CMVinfection(≥6AU/mL) 136(77.3%) 142(77.2%) 143(78.6%) 150(79.4%) 0.95

Continuousdataarepresentedasmedianwithcorrespondinginterquartilerange.DifferencesweretestedwithJonkheereTerpstratests(pfortrend).Pre-vaccinationGMT ispresentedasmeanwithcorresponding95%conﬁdenceinterval.DifferencesweretestedwithOne-WayANOVA.Categoricaldataarepresentedasnumber(percentage).

Differencesweretestedwithchisquaretests.

(5)

Table2

CharacteristicsofthestudypopulationdependingonthepresenceofCMVinfection.

CMVinfection p-value

Yes No

Level≥6AU/mL(n=571) Level<6.0AU/mL(n=160)

Age(years) 83(78–88) 83(77–87) 0.39

Females 439(76.9%) 112(70.0%) 0.07

Lengthofstay(months) 23(10–45) 19(9–39) 0.10

Katz-score 8(5–9) 7(4–9) 0.28

Numberofmedicaments 4(3–6) 4(2–6) 0.02

Diagnosisofdementia 389(68.1%) 107(66.9%) 0.77

Inﬂuenzastatus

Pre-vaccinationGMT 20.3(18.1–22.6) 23.1(18.2–29.2) 0.32

Highpre-vaccinationtiter(≥40) 217 (38.0%) 66 (41.2%) 0.46

Vaccinationstrategy

Dose15␮g+placebo 136(23.8%) 40(25.0%)

Dose15␮g+15␮gbooster 142(24.9%) 42(26.2%) 0.95

Dose30␮g+placebobooster 143(25.0%) 39(24.4%)

Dose30␮g+15␮gbooster 150(26.3%) 39(24.4%)

Continuousdataarepresentedasmedianwithcorrespondinginterquartilerange.DifferencesweretestedwithMann-WhitneyUtests.Pre-vaccinationGMTispresented asmeanwithcorresponding95%conﬁdenceinterval.Differencesweretestedwithindependentt-tests.Categoricaldataarepresentedasnumber(percentage).Differences weretestedwithchisquaretests.

vaccinations,CMVantibodylevelandvariousparametersofcomor- bidity.

ThislackofimpactofCMVinfectionontheresponsetoinfluenza vaccinationisunexpected.Influenzavaccinationisclinicallyeffec- tivein 70–90% of younger adults, but in only 17–53%of older persons[20].AgingisassociatedwithlownumbersofCD8+naïveT cellsandincreasednumbersofmemorycells[21].CMVinfectionis consideredamajordriverofoligoclonalexpansionsofCD8Tcellsin oldage,andtheCD57+CD8Tcellpoolinparticular[5–7,12,13,22], whichisthoughttorepresentahighlydifferentiatedpopulation oflatememoryTcells[23–25].Othershavethereforepostulated thatCMVinfectioniscausallyrelatedtotheage-associateddys- functionoftheimmunesystem[5,12,13],exemplifiedbyalower numberofcirculatingnaïveTcells,whichwouldimplythatthe decreasedresponsetoinfluenzavaccinationinolderpersonscould beattributedtoCMVinfection[4].

Thehypothesisthatpastand/orpersistentCMVinfectionmod- ulatesresponsivenesstoinﬂuenzavaccinationwassupportedby a trial by Trzonkowski et al. in 154 nursing home residents and staff who all received inﬂuenza vaccines containing anti-

0 10 20 30 40 50 60 70 80 90 100

23 25 82 84 107 109

Day

Predictedmean(95% CI) GMT

25 84 109 p=0.46

■ CMV positive(n=571)

♦^{CMV negati}^ve^(n=160)

0 10 20 30 40 50 60 70 80 90 100

23 25 82 84 107 109

Day

Predictedmean(95% CI) GMT

25 84 109 p=0.46

■ CMV positive(n=571)

♦^{CMV negati}^ve^(n=160)

Fig.2.EffectofCMVinfectiononGMTseroresponsetovaccination,adjustedfor gender,age,pre-vaccinationGMTandvaccinationstrategy.Predictedmeanand 95%CI,andp-valuewereobtainedbylinearmixedmodelanalysis.

gensofinfluenzastrainsA/Beijing/262/95(H1N1),A/Sydney/5/97 (H3N2)andB/Beijing/184/93[14].Strongreciprocalcorrelations were observed between anti-CMV antibodies and titers of all anti-hemagglutininsrangingfrom−0.41foranti-H3to−0.74for anti-H1.Inaddition,olderparticipantswhodidnotshowasero- logicalresponsetotheinfluenzavaccinehadhigherconcentrations ofanti-CMVantibodiescomparedtotheircounterpartswhodid respondtothevaccine.Theauthorsconcludedfromtheseobser- vationsthatCMVinfectionmayhave hada negativeimpacton theeffectiveness ofinfluenza vaccination.However, ourresults are notin linewiththese earlier findings.Differencesin study design, participantselection andrecruitment,and health status oftheparticipantsmaywellhavecontributedtothediscrepancy betweenourstudy and thestudy by Trzonkowskiet al.In the latter,age and thepresence of diseasehave likely affectedthe responsetoinfluenzavaccinationasCMVseroprevalenceishigh- estinoldersubjectswithcomorbidity.Ourdatapresentedhereare freefromconfoundingbyageanddiseaseastherandomizeddesign achievedsimilarpre-vaccinationtitersand CMVseroprevalence betweenthegroups.Anotherexplanationmaybethatouranalysis wasrestrictedtotheserologicalresponsetotheinfluenzastrain A/H3N2,becausethestudybyTrzonkowskiobservedtheweakest correlationbetweenCMVantibodylevelandtheH3N2titer[14].

This study has several strengths. First, we made use of a largerandomizedcontrolledtrialoninfluenzavaccinationamong older persons in long-term care-facilities, allowing us to effi- cientlyinvestigatewhetherCMVinfectionaffectedtheoutcome ofinfluenzavaccination[16].Inaddition,weusedthreewaysto assessseroresponsivenesstovaccination(GMT,HAIantibodytiter rise≥4-foldcomparedtoday0,andHAIantibodytiter≥40)at multipletimepointsduring3monthsoffollow-up,whichenabled ustothoroughlyandextensivelystudytheresponsetoinfluenza vaccination. As the current trial had sufficient power to show improvementsinprotectiveseroresponsesduringthefollow-up periodbetweenthefourtreatmentgroups[16],weconsideritlikely thatitcouldalsoshowaneffectofCMVinfectiononprotective seroresponses.Inthegroupofparticipantsthathadreceivedadose of30␮gandplacebobooster,asinglep-valuebelow0.05wasfound forthepercentageofparticipantswithanantibodytiter≥40atday 109,butthismaybeconsideredachancefindingasthisassociation wasnotfoundinanyothertreatmentgroupandtheseroprotection ratewashigherintheCMVpositiveratherthanintheCMVneg- ativegroup,whichisunexpected.Lastly,asaresultofthedesign

(6)

W.P.J.denElzenetal./Vaccine29 (2011) 4869–48744873 Seroresponsestoinﬂuenzavaccinationatday25,84and109dependingondoseandCMVinfection.

Dose15␮g+placebo p-value Dose15␮g+15␮gbooster p-value Dose30␮g+placebobooster p-value Dose30␮g+15␮gbooster p-value

CMVinfection CMVinfection CMVinfection CMVinfection

Yes No Yes No Yes No Yes No

n 136 40 142 42 143 39 150 39

GMT

Day25 56.7(44.4–72.4) 74.3(42.0–131.2) 0.33 52.5(40.4–68.3) 67.2(40.5–111.6) 0.37 74.6(58.9–94.4) 53.5(31.9–89.6) 0.21 71.3(56.7–89.6) 74.1(45.3–121.1) 0.88 Day84 47.2(37.2–60.0) 64.9(37.1–113.7) 0.24 42.5(32.8–55.1) 63.1(37.5–106.1) 0.16 53.2(42.2–67.0) 37.3(22.5–62.0) 0.17 57.4(45.2–73.0) 59.3(37.0–94.9) 0.91 Day109 47.1(36.8–60.5) 57.6(33.4–99.2) 0.46 52.5(40.9–67.3) 59.5(35.3–100.4) 0.64 52.2(40.9–66.6) 32.9(19.2–56.6) 0.09 66.6(52.4–84.6) 66.4(42.1–104.7) 0.99

≥4-foldincrease

Day25 53(40.2%) 10(26.3%) 0.12 40(29.6%) 13(31.0%) 0.87 66(48.2%) 17(43.6%) 0.61 63(43.2%) 17(44.7%) 0.86

Day84 39(32.0%) 6(17.1%) 0.09 36(27.5%) 12(32.4%) 0.56 53(40.2%) 11(29.7%) 0.25 48(35.3%) 14(40.0%) 0.61

Day109 35(29.4%) 7(19.4%) 0.24 44(33.8%) 14(36.8%) 0.73 45(36.9%) 11(30.6%) 0.49 55(42.3%) 16(48.5%) 0.52

Titer≥40

Day25 87(65.9%) 26(68.4%) 0.77 89(65.9%) 28(66.7%) 0.93 106(77.4%) 25(64.1%) 0.09 106(72.6%) 27(71.1%) 0.85

Day84 77(63.1%) 24(68.6%) 0.55 73(55.7%) 25(67.6%) 0.20 89(67.4%) 20(54.1%) 0.13 96(70.6%) 25(71.4%) 0.92

Day109 77(64.7%) 22(61.1%) 0.69 80(61.5%) 24(63.2%) 0.86 84(68.9%) 18(50.0%) 0.04 96(73.8%) 25(75.8%) 0.82

GMTispresentedasmeanwithcorresponding95%conﬁdenceinterval.Differencesweretestedwithindependentt-tests.Categoricaldataarepresentedasnumber(percentage).Differencesweretestedwithchisquaretests.

(7)

ofthestudy,wewerealsoabletoassessanypotentialeffectof vaccinationdoseandboostervaccination.

Alimitationof ourstudyis thattheanalysis waslimitedto theserological response tothe single inﬂuenza strain A/H3N2.

Theresultscan thereforenot begeneralized toother inﬂuenza virusstrainsorotherviruses.However,theserologicalresponse toA/H3N2hasproventobehighlyrelevantforolderpersons,as thisstrainhas,sinceitsappearancein1968,mostoftenbeenimpli- catedininﬂuenza-relatedmorbidityandmortalityinolderpersons.

Anotherlimitationofourstudyisthatthemainoutcomeofour studywasthehumoralmemoryresponsetoinﬂuenzavaccination.

Primaryimmuneresponsestoneo-antigensorcellularimmunity werenotstudiedandcouldthereforestillbeaffectedbyCMVinfec- tion.Inaddition,ourstudypopulationconsistedofolderpersons livinginlong-termcarefacilities.Althoughtheclinicalrelevance ofinﬂuenzavaccinationmaybehighestinthisparticularsubgroup ofolderpersons,theresultsofourstudymaynotbegeneralized toolderpersonsinthegeneralpopulationatlargeduetopossible differencesinhealthandimmunestatus.

Ourstudypopulationconsistedof olderindividualsliving in long-term care facilities with many comorbid conditions and impaired immune response. Our study therefore provides an excellentopportunitytoinvestigatetheeffectofCMVstatuson responsivenesstoinfluenzavaccination.Althoughourresultsdo notexcludearole ofCMVinfectioninthedevelopmentofage- related changesin the immune systemin olderpersonsin the generalpopulationatlarge,thepresentdatafromalargerandom- izedcontrolledtrialdo notsuggestthatCMV infectionhadany negativeeffectontheimmuneresponsetoinfluenzavaccination, whichisanimportantandclinicallyrelevanttriggeroftheimmune system,inolderpersonsinlong-termcarefacilities.Forthatrea- son,thesefindingsalsodonotsupportarolefordeterminingCMV statusinanefforttooptimizeindividualvaccinationstrategiesin olderpersonsinlong-termcarefacilities.Thesefindingswarrant furthervalidationinotherstudypopulations.

Contributors:ProfessorGussekloohad fullaccesstoallofthe datainthestudyandtakesresponsibilityfortheintegrityofthedata andtheaccuracyofthedataanalysis(guarantor).Studyconceptand design:denElzen,Cools,KroesandGussekloo.Acquisitionofdata:

Vossen,Cools,KroesandGussekloo.Analysisandinterpretationof data: denElzen, Vossen,Cools, Westendorp,Kroesand Gussek- loo.Draftingofthemanuscript:denElzen,VossenandGussekloo.

Criticalrevisionofthemanuscriptforimportantintellectualcontent:

denElzen,Vossen,Cools,Westendorp,KroesandGussekloo.Sta- tisticalanalysis:denElzenandGussekloo.Obtainedfunding:Cools, KroesandGussekloo.Administrative,technical,ormaterialsupport:

Vossen,CoolsandGussekloo.Studysupervision:Cools,Kroesand Gussekloo.Allauthorshaveapprovedtheﬁnalarticle.

Competinginterests:Nonedeclared.

Funding:Thisstudy wasfunded byhetPraeventiefonds. The sponsorhadnoroleinstudydesignandthecollection,analysis, andinterpretationofdataandthewritingofthearticleandthe decisiontosubmititforpublication.Allresearcherswereindepen- dentfromfundersandsponsors;allresearchershadaccesstoall thedata.

References

[1] Webster RG. Immunity to inﬂuenza in the elderly. Vaccine 2000;18(16):1686–9.

[2] YoshikawaTT.Epidemiologyanduniqueaspectsofagingandinfectiousdis- eases.ClinInfectDis2000;30(6):931–3.

[3] Miller RA. The aging immune system: primer and prospectus. Science 1996;273(5271):70–4.

[4]Kovaiou RD,Herndler-BrandstetterD,Grubeck-LoebensteinB. Age-related changesinimmunity:implicationsforvaccinationintheelderly.ExpertRev MolMed2007;9(3):1–17.

[5] PawelecG,DerhovanessianE,LarbiA,StrindhallJ,WikbyA.Cytomegalovirus andhumanimmunosenescence.RevMedVirol2009;19(1):47–56.

[6] KhanN,ShariffN,CobboldM,etal.Cytomegalovirusseropositivitydrivesthe CD8Tcellrepertoiretowardgreaterclonalityinhealthyelderlyindividuals.J Immunol2002;169(4):1984–92.

[7]LooneyRJ,FalseyA,CampbellD,etal.RoleofcytomegalovirusintheTcell changesseeninelderlyindividuals.ClinImmunol1999;90(2):213–9.

[8]PawelecG,AkbarA,CarusoC,SolanaR,Grubeck-LoebensteinB,WikbyA.

Humanimmunosenescence:isitinfectious?ImmunolRev2005:205257–68.

[9]BateSL,DollardSC,CannonMJ.CytomegalovirusseroprevalenceintheUnited States:thenationalhealthandnutritionexaminationsurveys,1988–2004.Clin InfectDis2010;50(11):1439–47.

[10]CannonMJ,SchmidDS,HydeTB.Reviewofcytomegalovirusseroprevalence and demographiccharacteristicsassociated withinfection.RevMed Virol 2010;20(4):202–13.

[11]StarasSA,DollardSC,RadfordKW,FlandersWD,PassRF,CannonMJ.Sero- prevalenceofcytomegalovirusinfectionintheUnitedStates,1988–1994.Clin InfectDis2006;43(9):1143–51.

[12] KochS,LarbiA,OzcelikD,etal.Cytomegalovirusinfection:adrivingforcein humanTcellimmunosenescence.AnnNYAcadSci2007:111423–35.

[13]OlssonJ,WikbyA,JohanssonB,LofgrenS,NilssonBO,FergusonFG.Age-related changeinperipheralbloodT-lymphocytesubpopulationsandcytomegalovirus infectionintheveryold:theSwedishlongitudinalOCTOimmunestudy.Mech AgeingDev2000;121(1–3):187–201.

[14]Trzonkowski P, Mysliwska J, Szmit E, et al. Association between cytomegalovirus infection, enhanced proinﬂammatory response and low levelofanti-hemagglutininsduringtheanti-inﬂuenzavaccination—animpact ofimmunosenescence.Vaccine2003;21(25–26):3826–36.

[15]Roos-VanEijndhovenDG,CoolsHJ,WestendorpRG,TenCate-HoekAJ,Knook DL, RemarqueEJ.Randomizedcontrolledtrialofseroresponsesto double doseandboosterinﬂuenzavaccinationinfrailelderlysubjects.JMedVirol 2001;63(4):293–8.

[16]CoolsHJ,GusseklooJ,RemmerswaalJE,RemarqueEJ,KroesAC.Beneﬁtsof increasingthedoseofinﬂuenzavaccineinresidentsoflong-termcarefacilities:

arandomizedplacebo-controlledtrial.JMedVirol2009;81(5):908–14.

[17] KellyH,AttiaJ,AndrewsR,HellerRF.Thenumberneededtovaccinate(NNV) andpopulationextensionsoftheNNV:comparisonofinﬂuenzaandpneu- mococcalvaccineprogrammesforpeopleaged65yearsandover.Vaccine 2004;22(17–18):2192–8.

[18] Beyer WE, Palache AM, Luchters G, Nauta J, Osterhaus AD. Seropro- tection rate, mean fold increase, seroconversion rate: which parameter adequately expresses seroresponse to inﬂuenza vaccination? Virus Res 2004;103(1–2):125–32.

[19]KatzS,DownsTD,CashHR,GrotzRC.Progressindevelopmentoftheindexof ADL.Gerontologist1970;10(1):20–30.

[20]GoodwinK,ViboudC,SimonsenL.Antibodyresponsetoinﬂuenzavaccination intheelderly:aquantitativereview.Vaccine2006;24(8):1159–69.

[21]GoronzyJJ,LeeWW,WeyandCM.AgingandT-celldiversity.ExpGerontol 2007;42(5):400–6.

[22]DerhovanessianE,MaierAB,BeckR,etal.Hallmarkfeaturesofimmunosenes- cenceareabsentinfamiliallongevity.JImmunol2010;185(8):4618–24.

[23]GratamaJW,KardolM,NaipalAM,etal.Theinﬂuenceofcytomegaloviruscar- rierstatusonlymphocytesubsetsandnaturalimmunity.ClinExpImmunol 1987;69(1):16–24.

[24]WangEC,Taylor-WiedemanJ,PereraP,FisherJ,BorysiewiczLK.Subsetsof CD8+,CD57+cellsinnormal,healthyindividuals:correlationswithhuman cytomegalovirus(HCMV)carrierstatus,phenotypicandfunctionalanalyses.

ClinExpImmunol1993;94(2):297–305.

[25]Wang EC, Moss PA, Frodsham P, Lehner PJ, Bell JI, Borysiewicz LK.

CD8highCD57+Tlymphocytesinnormal,healthyindividualsareoligoclonal andrespondtohumancytomegalovirus.JImmunol1995;155(10):5046–56.