Primary and secondary prevention interventions for cognitive decline and dementia

Overview of reviews

2016

2 Table of contents

Publishedby

Title

TheNorwegianInstituteofPublicHealthSectionforevidencesummariesintheKnowledgeCentrePrimaryandsecondarypreventioninterventionsforcognitivedeclineanddementia

Norwegiantitle Primær‐ogsekundærforebyggendetiltakforkognitivsviktogdemensResponsible CamillaStoltenberg,direktør

Authors GerdMFlodgren,projectleader,researcher,theKnowledgeCentreRigmorCBerg,HeadofUnit,forSocialWelfareResearchattheKnowledgeCentre

ISBN 978‐82‐8082‐745‐6 

Projectnumber 798Typeofpublication Overviewofreviews

Noofpages 69(110inklusivvedlegg)Client Nasjonalforeningenforfolkehelsen

MeSHterms Alzheimer’sdisease,dementia,cognition,cognitiveimpairment,cognitivedisorders,memorycomplaints,primaryprevention,secondaryprevention

Citation

FlodgrenGM,BergRC.Primaryandsecondarypreventioninterventionsforcognitivedeclineanddementia.[Primær‐ogsekundærforebyggendetiltakforkognitivsviktogdemens]Rapport−2016.Oslo:Folkehelseinstituttet,2016.

3 Table of contents

Tableofcontents

TABLEOFCONTENTS 3

KEYMESSAGES 5

EXECUTIVESUMMARY 6 Background 6 Objectives 6 Methods 6 Results 6 Discussion 8 Conclusions 8

HOVEDFUNN(NORSK) 9

SAMMENDRAG(NORSK) 10 Bakgrunn 10 Problemstillinger 10 Metoder 10 Resultat 10 Diskusjon 12 Konklusjon 12

PREFACE 13

OBJECTIVES 15

BACKGROUND 16 Descriptionofthecondition 16 Howtheinterventionsmaywork 18 Whyisitimportanttodothisoverviewofreviews? 21

METHODS 22 Objectives 22 Inclusioncriteria 22 Exclusioncriteria 23 Literaturesearch 23 Selectionofreviews 24 Dataextraction 24 Datasynthesis 25 Gradingoftheevidence 25

4 Table of contents

Ethics 26

RESULTS 27 Descriptionofincludedreviews 27 Primarypreventioninterventions 31 Secondarypreventioninterventions 37 Certaintyoftheevidence 40 Effectsofinterventions 41 Ethics 50

DISCUSSION 52 Mainresults 52 Certaintyoftheevidence 52 Strengthsandweaknesses 53 Potentialbiasesintheoverviewprocess 54 Overallcompletenessandapplicabilityoftheevidence 54 Agreementsordisagreementswithotheroverviewsofreviews 57 Applicationsforpractice 57 Needforfurtherresearch 57

CONCLUSION 59

REFERENCES 60

APPENDICES 71 1Glossary 71 2Searchstrategy 77 3Excludedreviewsandreasonsforexclusion 88 4Assessmentofmethodologicalquality 90 5Ongoingreviews 92 6Descriptionofincludedreviews 93 7GRADEevidenceprofiles 97

5 Keymessages

Keymessages

Dementiaisasyndromecharacterisedbydeteriorationinmemory,thinking,behaviour,andtheabilitytoperformeverydayactivities,whichultimatelymayleadtototaldependenceanddeath.Sincetheworld’spopulationissteadilygrowingolder,thenumberofpeoplewithdementiaisalsoincreasing.Itisthereforeofutmostimportancetoidentifyeffectivestrategiestopreventordelayitsonset.Thekeyfindingsofthisoverviewofreviewsarebasedonevidencefromeightsystematicreviews.Theresultsforthesingleinterventionstargetingcognitivelyhealthypeoplesuggestthatcomparedtocontrol:

Antihypertensivedrugsmayleadtoaslightdecreaseinincidenceofdementiainpeoplewithhypertension(lowcertaintyofevidence).

Statintherapyprobablyleadstolittleornodifferenceonincidenceofdementiainpeoplewith,oratriskof,cardiovasculardisease(moderatecertainty).

Omega‐3FattyAcids(FAs)probablyleadtolittleornoeffectoncognitivetestscores(moderatetohighcertainty).

Computerisedcognitivetrainingprobablyleadstoaslightimprovementincognitivetestscoresdirectlyafterthetraining(moderatecertainty).

Aerobicexercisemayleadtolittleornoeffectoncognitivetestscores(lowcertainty).

Theresultsfortheinterventionstargetingpeoplewithmildcognitiveimpairmentsuggestthatcomparedtocontrol:

Cholinesteraseinhibitorsprobablyleadtoaslightdecreaseindementiaincidence,buttosignificantlymoreadverseevents(moderatecertainty).

VitaminEprobablyleadstolittleornodifferenceinincidenceofAlzheimer’sdementia(moderatecertainty).

Omega‐3FAsprobablyleadtolittleornodifferenceincognitivetestscores(moderatetohighcertainty).

Wedidnotfindanyreviewsthatevaluatedtheeffectsofinterventionstargetingmorethanoneriskfactor,andwecanthereforenotsayanythingaboutthecombinedeffectsoftheseinterventions.

Title:Primaryandsecondarypreventioninterventionsforcognitivedeclineanddementia‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐

Typeofpublication:

Overviewofreviews‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐Doesn’tanswereverything: Nointerventionstargetingpeoplewithdementia.

Nohealtheconomicevaluation.

‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐Whoisresponsibleforthispublication?TheNorwegianInstituteofPublicHealthcompletedthisassignmentwhichwascom‐missionedbytheNationalAssociationofPublicHealth.‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐Whenweretheliteraturesearched?LiteraturesearcheswereconductedinJanuary2016.‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐

Peerreferees:ØyvindKirkevold,professorSeksjonforsykepleieAvdelingforhelse,omsorgogsykepleie,NorgesTeknisk‐Naturviten‐skapligeUniversitet.VeslemøyEgede‐Nissen,førs‐telektor,Instituttforsykepleieoghelsefremmendearbeid,HøgskoleniOsloogAkershus

6 Executivesummary

Executivesummary

Background

Dementiaisachronicsyndromecharacterisedbydeteriorationinmemory,thinking,behaviour,andtheabilitytoperformeverydayactivities,whichoftenleadstototalde‐pendenceanddeath.Theworld’spopulationissteadilygrowingolder,andasdementiaismoreprevalentinpeopleover70,andincreaseswithincreasingage,thenumberofpeoplewithdementiaisalsoincreasing.In2012,around71,000peopleinNorwayhadadementiadiagnosis,whichrepresents1.6%ofthetotalpopulation.

Objectives

Theaimofthisoverviewofreviewswastoanswerthefollowingtwoquestions:1)Whatisthedocumentedeffectivenessofinterventionstopreventcognitivedeclineorincidenceofdementiaincognitivelyhealthypeople(primaryprevention),2)Whatisthedocumentedeffectivenessofinterventionstoprevent(further)cognitivedeclineorprogressiontodementiainpeoplewithmildcognitiveimpairment(MCI)orotherearlysymptomsorsignsofdementia(secondaryprevention)?

Methods

WeconductedanoverviewofreviewsinaccordancewiththeKnowledgeCentre’shandbook.WesearchedineightdatabasesuptoFebruary2016forreviewsevaluatingtheeffectsofinterventionstopreventordelaycognitivedeclineordementiainpeoplewithorwithoutMCI.Twopeopleindependentlyscreenedalltitlesandabstracts,re‐viewedfulltexts,assessedreviewquality,andgradedthecertaintyoftheevidenceus‐ingtheGRADE(GradingofRecommendationsAssessment,DevelopmentandEvalua‐tion)tool.Oneauthorextracteddata,andanothercheckedthatitwascorrect.

Results

Weincludedeighthighqualityreviewspublishedbetween2009and2016.Fiveofthereviewsinvolvedprimarypreventioninterventionsforcognitivelyhealthypeople.ThreereviewsincludedsecondarypreventioninterventionsforpeoplewithMCIormemorycomplaints.Thereviewsevaluatedtheeffectsofpharmacologicaltherapies(3reviews),dietarysupplements(3reviews),aerobictraining(onereview),andcognitivetraining(onereview).Thecomparatorinthelattertwowaseitheranotheractiveinter‐ventionornointervention,andplaceboinalltheothers.

7 Executivesummary

Primarypreventioninterventions

Pharmacologicaltherapies

Antihypertensivedrugs.Weincludedonereview(4trials;n=15,936)concernedwiththeeffectsofantihypertensivedrugsondementiaincidenceinolderpeoplewithhyperten‐sion.TheparticipantswererecruitedinEurope,NorthAmerica,China,AustralasiaandTunisia.Thepooledresult(OddsRatio[OR]:0.89[0.74to1.07])indicatesthatantihy‐pertensivedrugsmayleadtoaslightdecreaseinincidenceofdementia,ascomparedtoplacebo,at1.8to4.5yearsfollowup.

Cholesterolloweringdrugs(Statins).Weincludedonereview(2trials;n=26,340)con‐cernedwiththeeffectsofstatinsonincidenceofdementiaincognitivelyhealthyolderpeoplewithevidenceof,orathighriskof,cerebrovasculardisease.ThestudieswereconductedintheUK,Ireland,andtheNetherlands.Theresultofonetrial(OR:1.00[0.61to1.65],n=20,536)indicatesthatstatintherapymayleadtolittleornodifferenceinincidenceofdementia,ascomparedtoplacebo.Theothertrial(n=5,804)reportednodifferenceincognitivetestscoresbetweengroupsatmean3.2yearsfollowup.

Dietarysupplements

Omega‐3FattyAcids(FAs).Weincludedonereview(3trials;n=4,080)whichevaluatedtheeffectsofOmega‐3FAsoncognitivedeclineincognitivelyhealthyparticipants.Thepooledresults(StandardisedMeanDifference[SMD][4tests]:0.06higherto0.04lowerscores;MeanDifference[MD][2tests]:0.12higherto0.07lowerscores)suggestthatOmega‐3FAsupplementationprobablyleadstolittleornodifferenceinoverallcognitivefunction,ascomparedtoplacebo,at6to40monthsfollowup.

Aerobicexercise

Weincludedonereview(12trials;n=754)whichevaluatedtheeffectsofsupervisedaerobicexerciseoncognitivefunctionincognitivelyhealthyolderpeople.ThestudieswereconductedintheUSA,CanadaandFrance.Thepooledresultssuggestthataerobicexercisemayleadtolittleornoeffectoncognitivetestscores(SMD:range0.09lowerto0.30higherscores;MD[2tests]:0.10to0.16ascomparedtonointerventionat8to24weeksfollowup).

Cognitivetraining

Weincludedonereview(52trials;n=4,885)whichevaluatedtheeffectsofcomputer‐isedcognitivetraining(CCT)oncognitivedeclineincognitivelyhealthypeople.TheparticipantswerefromtheUSA,Europe,Canada,Australia,Israel,China,Taiwan,SouthKorea,andJapan.Thepooledresultofthisreview(Hedge’sg:0.22[0.15to0.29])sug‐geststhatCCTprobablyleadstoasmallimprovementincognitivetestscoresdirectlyafterthetraining.Secondarypreventioninterventions

Pharmacologicaltherapies

Cholinesteraseinhibitors.Weincludedonereview(9trials;n=5,149)concernedwiththeeffectsofcholinesteraseinhibitorsforthepreventionofdementiainpeoplewithMCI.ThestudieswereconductedinUSA,Canada,Singapore,andGermany.Thepooled

8 Executivesummary

resultssuggestthatcholinesteraseinhibitorsmayleadtoaslightdecreaseinprogres‐siontodementia(RelativeRisk[RR]:0.84[0.70to1.02]),at3years,buttomoread‐verseeventsthanplacebo(RR:1.09[1.02to1.16]).

Dietarysupplements

VitaminE.Weincludedonereview(1trial;n=769)oftheeffectsofvitaminEonde‐mentiaincidenceinpeoplewithMCI.Theresultsofthissingletrial,conductedinUSAandCanada,suggestthatVitaminEsupplementationpossiblehaslittleeffectoninci‐denceofAD(HazardRatio:1.02[0.74to1.41])at36months,ascomparedtoplacebo.

Omega‐3FAs.Weincludedonereview(4trials;n=676)oftheeffectsofOmega‐3FAsoncognitivedeclineinpeoplewithMCI.ThestudieswereconductedintheNether‐lands,England,Wales,Japan,Israel,andtheUSA.ThepooledresultsshowthatOmega‐3FAsprobablyleadtolittleornodifferenceincognitivefunction(MD:0.16higherto0.05lowerscores)atmedian14.5to24weeks,ascomparedtoplacebo.

Discussion

Weincludedeighthighqualityreviews(86originalstudies)concernedwiththeeffectsofinterventionsaimedatpreventingcognitivedeclineanddementia.Resultsfromfourofthesereviewssuggestthatstatintherapy,Omega‐3FAsandvitaminEsupplementsprobablyleadtolittleornodifferenceincognitivefunctionorincidenceofdementia.Theresultsforcholinesteraseinhibitorsandantihypertensivedrugs,suggestthatthesedrugsmayleadtoaslightdecreaseinincidenceofdementia,butthatcholinesterasein‐hibitorsprobablyleadtomoreadverseeventsthanplacebo.CCTprobablyleadstoslightlyimprovedcognitivefunctiondirectlyafterthetraining,whileaerobicexercisemayleadtolittleornodifferenceincognitivefunction.

Wedidnotidentifyanyeligiblehighqualityreviewsconcernedwiththeeffectsofhealthylifestyles(otherthanaerobicexercise),e.g.changetoahealthydiet,decreasedalcoholuse,etc.,orotherriskfactors,e.g.depression,lackofsocialengagement,orloweducationalattainment.Wefoundnoreviewsassessingtheeffectsofinterventionstar‐getingmultipleriskfactorstopreventcognitivedeclineordementia.

Weunfortunatelystillhavelittleknowledgefromsystematicreviewsofeffectivepre‐ventiveinterventions,addressingsingleormultipleriskfactorsfordementia.

Conclusions

Wefoundnoconvincingevidencefortheeffectivenessoftheinterventionsincludedinthisoverviewofreviewsinpreventingcognitivedeclineordementia.Wideconfidenceintervalsandfeweventsinsomeoftheanalyseswarrantcautionwheninterpretingtheresults.Asprogressiontodementiaispartlydeterminedbyanumberofmodifiablefac‐torsrelatedtolifestyle,environment,depression,educationallevel,anddegreeofsocialinteraction,itispossiblethatpreventiveinterventionsmaybemoreeffectiveiftheytakeintoaccountthemultifacetedaetiologybehindthedisease,i.e.interventionsthattargetsmultipleriskfactors.

9 Hovedfunn(Norsk)

Hovedfunn(Norsk)

Demenserensykdomkjennetegnetvedsvekkethukommelse,tenkning,atferdogevnetilåutføredagligeaktiviteter,somtilsluttofteførertiltotalavhengighetogdød.Sidenverdensbefolk‐ningstadigblireldreforventesantallpersonermeddemensåøkedramatisk.Deterderforviktigåidentifisereeffektivestrate‐gierforåforebyggeellerutsettesykdomsdebuten.

Deviktigstefunneneidenneoversiktenoveroversiktererbasertpåevidensfraåtteoversikter.Forenkelt‐tiltakrettetmotkognitivtfriskeeldrepersonertyderresultatenepåatsammenlignetmedkontroll:• Blodtrykkssenkendelegemidlerførermuligenstilenliten

reduksjoniforekomstavdemenshoseldrepersonermedhøytblodtrykk(lavtillittileffektestimatene).

• Kolesterolsenkendelegemidler(statiner)førertroligtillitenelleringenforskjelliforekomstavdemenshoseldrepersonermedhjerte‐ogkarsykdommer(moderattillit).

• Omega‐3fettsyretilskuddførertillitenelleringenforskjellikognitivetestresultater(moderattilhøytillit).

• Datastyrtkognitivtreningførertroligtilenlitenbedringikognitivetestresultaterdirekteettertreningen(moderattillit).

• Aerobtreningførermuligenstillitenelleringenforskjellikognitivetestresultater(lavtillit).

Fortiltakrettetmotpersonermedmildkognitivsvikt,tyderresultatenepåatsammenlignetmedkontroll:

Kolinesterasehemmereførertroligtilenlitenreduksjoniforekomstavdemens,menførertilsignifikantflerebivirkninger(moderattillit).

VitaminEførertroligtillitenelleringengenerellforskjelliforekomstavAlzheimersdemens(moderattillit).

Omega‐3fettsyretilskuddførertroligtillitenelleringenforskjellikognitivetestresultater(moderattilhøytillit).

Vifantingenoversiktersomevaluerteeffektenavtiltakrettetmotmerennénrisikofaktor,ogvikandetderforikkesinoeomeffektenavåkombineredissetiltakene. 

Tittel:Primær‐ogsekundærforebyggendetiltakforkognitivsviktogdemens‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐

Publikasjonstype:

Oversiktoveroversikter‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐

Svarerikkepåalt:•Ingentiltakrettetmotpersonermeddemensdiagnose.•Ingenhelseøkonomiskevaluering‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐

Hvemstårbakdenne

publikasjonen?

Folkehelseinstituttethargjennom‐

førtoppdragetetterforespørselfra

NasjonalforeningenforFolkehelsen‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐

Nårblelitteratursøketgjennom‐

ført:

Søketterstudierbleavslutteti

januar2016.‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐

Eksternefagfeller:ØyvindKirkevold,ProfessorSeksjonforsykepleieAvdelingforhelse,omsorgogsykepleie,NorgesTeknisk‐Naturviten‐skapligeUniversitetVeslemøyEgede‐Nissen,førstelektor,Instituttforsykepleiereoghelsefremmendearbeid,HøgskoleniOsloogAkershus

10 Sammendrag (Norsk)

Sammendrag(Norsk)

Bakgrunn

Demenserettilstandkarakterisertvedsvekkelseavhukommelsen,tenkning,adferdogevnentilåutføredagligeaktiviteter,somtilsluttførertiltotalavhengighetogdød.Si‐denverdensbefolkningstadigblireldre,ogdemensermerutbredthospersonerover70,forventesantalletpersonermeddemensogsååøke.I2012haddeomkring71000personeriNorgeendemensdiagnose,hvilketrepresenterer1,6prosentavdentotalebefolkningen.

Problemstillinger

Måletmeddenneoversiktoveroversiktervaråbesvarefølgendetospørsmål:1)Hvaerdendokumenterteeffektenavforebyggendetiltakforåforebyggekognitivsviktogdemens(primærforebygging)?2)Hvaerdendokumenterteeffektenavtiltakforåfore‐bygge(ytterligere)kognitivsviktogprogresjontildemenshospersonermedmildkog‐nitivsviktellerandretidligesymptomerellertegnpådemens(sekundærforebygging)?

Metoder

VigjennomførteenoversiktoveroversikterihenholdtilKunnskapssenteretshåndbok.Visøkteiåttedatabaseropptiljanuar2016foroversiktersomevaluerteeffektenavtiltakforåhindreellerforsinkekognitivsvikt,Alzheimerssykdomellerandreformerfordemenshospersonermedellerutenmildkognitivsvikt.Uavhengigavhverandregjennomgikktopersoneralletitlerogabstrakt,vurderterelevanteoversikterifulltekst,vurdertedenmetodiskekvalitetenogbedømtetillittileffektestimatenevedhjelpavGRADE(GradingofRecommendationsAssessment,DevelopmentandEvaluation).Enforfatterhentetutdata,ogenforfatterkontrollerteriktighetenavdata.

Resultat

Viinkluderteåtteoversikteravhøykvalitetpublisertmellom2009og2016.Femavoversikteneomhandletprimærforebyggendetiltakrettetmotkognitivtfriskepersoner.Treoversikteromhandletsekundærforebyggendetiltakrettetmotpersonermedmildkognitivsvikt.Oversikteneevaluerteeffektenavfarmakologisketiltak(3oversikter),kosttilskudd(3oversikter),aerobtrening(1oversikt),ogkognitivtrening(1oversikt).

11 Sammendrag (Norsk)

Detosistnevntesammenlignetmedentenenannenaktivtiltakelleringentiltak,ogalledeandresammenlignetmedplacebo.

Primærforebyggendetiltak

Farmakologiskbehandling

Blodtrykkssenkendelegemidler.Viinkluderteénoversikt(4studier;n=15936)omef‐fektenavblodtrykkssenkendemidlerpådemenshoseldrepersonermedhypertensjon.DeltakernevarrekruttertfraEuropa,Nord‐Amerika,Kina,AustralasiaogTunisia.Detsamlederesultatetfradenneoversikten(Oddsratio[OR]:0,89[0,74til1,07])tyderpåatblodtrykkssenkendemidlermuligensførertilenlitenreduksjoniforekomstavde‐mens,sammenlignetmedplaceboved1,8til4,5årsoppfølging.

Kolesterolsenkendelegemidler(statiner).Viinkluderteénoversikt(2studier;n=26340)omeffektenavstatinerpåforekomstavdemenshoskognitivtfriskeeldrepersonermedhøyrisikofor,ellerdokumentertcerebrovaskulærsykdom.StudienebleutførtiStorbritannia,IrlandogNederland.Resultatenefraenstudie(OR1,00[0,61til1,65];n=20536),tyderpåatstatinertroligførertillitenelleringenforskjellifore‐komstavdemens,sammenlignetmedplacebo.Denandreinkludertestudien(n=5804)rapporterteingenforskjellermellomgruppeneikognitivetestresultatervedgjennom‐snittlig3,2årsoppfølging.

Kosttilskudd

Omega‐3Fettsyrer(FAs).Viinkluderteénoversikt(3studier;n=4080)omeffektenavOmega‐3FAspåkognitivsvikthoskognitivtfriskepersoner.Desamlederesultatene(Standardisertgjennomsnittsforskjell[SMD][4tester]:0,06høyeretil0,04lavereskåre;Gjennomsnittsforskjell[MD][2tester]:0,12høyeretil0,07lavereskåre)tyderpåatOmega‐3FAsførertillitenelleringeneffektpåkognisjonved6til40månedersoppfølging.

AerobtreningViinkluderteénoversikt(12studier;n=754)omeffektenavaerobtreningpåkognitivfunksjonhoskognitivtfriskeeldrepersoner.StudienevarutførtiUSA,CanadaogFrankrike.Aerobtreningførermuligenstillitenelleringenforskjellikognitivetestre‐sultater(SMD[8tester]:0,09laveretil0,30høyereskåre;MD[2tester]:0,10til0,16,sammenlignetmedingentiltakved8til24ukersoppfølging).

KognitivtreningViinkluderteénoversikt(52studier;n=4885)omeffektenavuliketyperavdatastyrtkognitivtrening(CCT)forforebyggingavaldersrelatertkognitivsvikthoskognitivtfriskepersoner.Desamlederesultatenefradenneoversikten(Hedge’sg:0,22[0,15til0,29))tyderpåatCCTtroligførertilenlitenforbedringikognitivfunksjondirekteet‐tertreningen.

Sekundærforebyggendetiltak

Farmakologiskbehandling

Kolinesterasehemmere.Viinkluderteénoversikt(9studier;n=5149)omeffektenavkolinesterasehemmerepåforekomstavdemenshospersonermedmildkognitivsvikt.StudienevargjennomførtiUSA,Canada,Singapore,Tysklandogiflereikke‐navngitteland.Desamlederesultateneviseratkolinesterasehemmeretroligharenliteneffektpå

12 Sammendrag (Norsk)

progresjontildemens(datafra3studier;Riskratio[RR]:0,84[0,70til1,02])ved3årsoppfølging,menharflereskadevirkningerennplacebo(RR:1,09[1,02til1,16]).

Kosttilskudd

VitaminE.Viinkluderteénoversikt(1studie;n=769)omeffektenavvitaminEpåfore‐komstavdemenshospersonermedmildkognitivsvikt.Resultatenefradenneenkelt‐studien,utførtiUSAogCanada,tyderpåatvitaminEharliteneffektpåforekomstavmuligellertroligAlzheimer(Hazardratio:1,02[0,74til1,41])ved36månedersoppføl‐ging,sammenlignetmedplacebo.

Omega‐3Fettsyrer(FAs).Viinkluderteénoversikt(4studier;n=676)omeffektenavOmega‐3FAspåkognitivsvikthospersonermedmildkognitivtsvikt.StudieneblegjennomførtiNederland,England,Wales,Japan,IsraelogUSA.DesamlederesultateneviseratOmega‐3FAstroligharlitenelleringeneffektpåkognitivtestskåre(MD0,16høyeretil0,05lavereskåre)vedenmedianoppfølgingstidpå14,5til24ukeroppføl‐ging,sammenlignetmedplacebo.

Diskusjon

Viinkluderteåtteoversikteravhøymetodiskkvalitet(86originalestudier)somopp‐summerteeffektenavtiltakforåforebyggekognitivsviktogdemens.

Resultaterfrafireavdisseoversiktenetyderpåatkolesterolsenkendelegemidler,Omega‐3fettsyrerogvitaminEtilskuddtroligførertillitenelleringenforskjellikogni‐tivfunksjonellerforekomstavdemens.Resultateneforkolinesterasehemmereogblod‐trykkssenkendemedisinertyderpåatdissemedisinenekanføretilensvaknedgangiforekomstenavdemens,menatkolinesterasehemmeresannsynligvisførertilflerebi‐virkningerennplacebo.CCTførertroligtillittbedrekognitivfunksjonrettettertre‐ning,mensaerobtreningmuligensførertillitenelleringenforskjellikognitivfunksjon.

Vifantingenoversikteravhøymetodiskkvalitetsomomhandleteffektenavendringtilensunnlivsstil,annetennaerobtrening,f.eks.suntkosthold,røykeslutt,elleroversik‐teravtiltakrettetmotandrerisikofaktorer,f.eks.depresjon,lavtutdanningsnivåellermangelpåsosialtilknytning.Vifanthelleringenoversikteromsammensattetiltakret‐tetmotflererisikofaktorerforåforebyggekognitivsviktellerdemens.

Vihardessverrefortsattlitekunnskapfrasystematiskeoversikteromeffektivetiltak,bådenårdetgjelderénogflererisikofaktorer,foråforebyggekognitivsviktellerde‐mens.

Konklusjon

Ingenavdeprimær‐ellersekundærforebyggendetiltakeneoppsummertideinklu‐derteoversikteneviseroverbevisendeeffektpåforebyggingavkognitivsvikt,ADogandreformerfordemens.Bredekonfidensintervallerogfåhendelserinoenavanaly‐senetilsieratmanmåvisevarsomhetvedtolkningavresultatene.Progresjontilde‐menserknyttettilenrekkemodifiserbarefaktorer,sliksomlivsstil,miljø,depresjon,utdanningsnivåogsosialtilknytning.Derforerdetmuligatforebyggendetiltakkanhabedreeffekthvisdetarhensyntildensammensatteetiologienbaksykdommen,dvs.hvistiltakrettesmotflererisikofaktorersamtidig.

13 Preface

Preface

TheKnowledgeCentreattheNationalInstituteofPublicHealthwasintheautumn2013commissionedbytheNationalAssociationofPublicHealthtoconductanover‐viewofreviewsevaluatingtheeffectivenessofinterventionsaimedatpreventingorde‐layingtheonsetofdementiaincognitivelyhealthypeopleandinpeoplewithmildcog‐nitiveimpairmentwithoutadementiadiagnosis.Inthisoverviewofreviewswesummariseandevaluatetheevidencefromeightsys‐tematicreviews(SRs)oftheeffectofvariousprimaryandsecondaryinterventions(i.e.pharmacologicaltherapy,dietarysupplements,cognitivetrainingandaerobicexercise)oncognitivefunctionandprogressiontodementiainpatientswithoutadementiadiag‐nosis.TheKnowledgeCentrefollowsacommonapproachinsummarisingresearch,docu‐mentedinthemanual"Howwesummariseresearch."Itmeansthatwemayusestand‐ardformulationswhenwedescribethemethods,resultsanddiscussionofthefindings.Contributorstotheproject:Projectleaderandresearcher:GerdMFlodgren,theKnowledgeCentre,Headofunitandresearcher:RigmorCBerg,theKnowledgeCentre,Internalcontributors:RigmorCBergservedasprojectleaderduringthefirstphasesoftheproject.Shedevel‐opedtheprojectprotocol,andledtheworkrelatedtoscreeningandqualityassessmentofliterature.GerdMFlodgrenthereaftertookontheroleasprojectleader,andleddataextraction,analysis,andwriteupofthereportforpublication.RigmorCBergcom‐mentedonearlyversionsofthereportandapprovedthefinalversion.WewishtoacknowledgeKristinThuveDahmandThereseDalsbøwhocontributedattheinitialstage(i.e.toscreening,reviewselection/qualityassessmentfortheinitialsearch),re‐searchlibrarianGyriHvalStraumannwhoperformedthesystematicsearch,andre‐searchlibrarianIngridHarboeforpeerreviewingthesearchstrategy.WealsowishtoacknowledgeKjetilBrurbergforhelpfulcommentsonthereport,andGunnVist,LivMereteReinar,whopeerreviewedthereport.Externalcontributors:Wealsowishtoacknowledgetheexternalpeerreferees:ØyvindKirkevold,ProfessorvedSeksjonforsykepleieAvdelingforhelse,omsorgogsykepleie,NorgesTeknisk‐Na‐turvitenskapligeUniversitet,andførstelektorVeslemøyEgede‐Nissen,vedInstituttforsykepleiereoghelsefremmendearbeid,HøgskoleniOsloogAkershus.

14 Preface

Declarationofinterest:Neithertheauthorsnortheexternalpeerrefereesstateanyconflictsofinterest.Theaimofthisreportistosupportwell‐informedevidence‐baseddecisionsinhealthcarethatleadtoimprovedqualityofservices.Wesuggestthatwhenmeetingwiththeindividualpatient,theresultsofthisoverviewshouldbeconsideredinconjunctionwithotherrelevantfactors,patientneedsandclinicalexperience.

SigneFlottorpHeadofDepartment

RigmorCBergHeadofUnit

GerdMFlodgrenProjectleader

15 Objectives

Objectives

Theaimofthisprojectwastoconductanoverviewofreviewsthatanswersthefollow‐ingquestions:

1) Whatisthedocumentedeffectivenessofinterventionstopreventcognitivedeclineandincidenceofdementiaincognitivelyhealthypeople(primaryprevention)?

2) Whatisthedocumentedeffectivenessofinterventionstoprevent(further)cognitivedeclineandprogressiontodementiainpeoplewithmildcognitiveimpairmentorotherearlysymptomsorsignsofdementia(secondaryprevention)?

16 Background

Background

Descriptionofthecondition

Dementiaisasyndromeinwhichthereisdeteriorationinmemory,thinking,behav‐iour,andtheabilitytoperformeverydayactivities.Itisconsideredoneofthemajorcausesofdisabilityanddependencyintheworld(1).Dementiamainlyaffectspeopleover70yearsofage,butisnotapartofnormalageing.Thecauseofdementiaisdeathofbraincells,whichinturnmaybecausedbyvariousconditions.ThemostcommoncauseisAlzheimer’sdisease,aneurodegenerativedisease,whichmayaccountforasmuchas60‐70%ofalldementiacases(1).Othertypesofdementiaarevascularde‐mentia,Lewybodydementia,andfrontotemporaldementia.Otherlesscommonde‐mentiacausesaresubstanceabuse,headinjury,metabolicdisease,vitamindeficiencyandotherdiseases(2).Themostcommondementiatypesaredescribedinmoredetailbelow.AglossaryisfoundinAppendix1.Differenttypesofdementia

Alzheimer’sdisease

Alzheimer’sdisease(AD)ischaracterisedbyacomplexseriesofbrainchangesthatde‐velopovermanyyears.Itinvolvessocalledamyloidplaquesandneurofibrillarytangleswhichdevelopinbrainstructuresthathelptoencodememories,andinareasthatareusedinthinkingandmakingdecisions(3,4).Theseevolvingchangesresultinaslowdeclineinmemory,thinking,andreasoningskills,thatoftenleadstototaldependencyanddeath(3).Ithasbeensuggestedthatthecauseofthepathophysiologicalbrainchangesismultifactorial,i.e.acombinationofgenetic,environmental,andlifestylefac‐tors(3).ClinicallyitisonlypossibletomakeaprobableADdiagnosis,andadefinitedi‐agnosiscanonlybeachievedpostmortem.Researchisbeingconductedintothepatho‐physiologyofAD,inordertofindsensitivebiomarkerstoprovidebetterdiagnostictoolsandcriteriaforearlyidentificationofAD(4).

Vasculardementia

Vasculardementia(VAD)isthesecondmostcommontypeofdementiaafterAD,andaccountsforaround15%ofdementiacases(5).ItisgenerallyagreedthattheaetiologybehindVADinvolvesvariouscardiovascularconditionscausingdamagetothebloodvesselsofthebrain(e.g.stroke,highbloodpressure;hardeningofthearteries;diabe‐tes)(6).Thereishowevernoconsensusonexactlyhowthecerebrovascularpathologi‐calchangestranslateintocognitiveimpairmentordementia,andthereisalsono

17 Background

agreedschemeforstagingordiagnosingVAD(5).ThecognitivechangesinVADvariesmorethaninAD,butcommonlyincludedeficitsinattention,informationprocessing,andexecutivefunction(5).ThiseffectivelymeansthattoolsusedtodiagnoseADmaynotbeeffectiveindiagnosingVAD.Lewybodydementia

Lewybodydementia(LBD)isanumbrellatermthatincludesclinicallydiagnosedde‐mentiawithLewybodies(DLB)andParkinson’sdiseasedementia(PDD),which,ac‐cordingtomanyresearchers,sharethesamepathology.WhileDLBusuallydevelopsconcomitantlywithorbeforeanysignsofParkinsonism,PDDdevelopsinpatientswithanalreadywellestablishedPD.Thepathologicalchanges,whicharecharacterisedbyabnormalclumpsofaproteincalledalpha‐synuclein(Lewybodies)inneuronsinthebrain,whichresultsinaprogressivecognitivedecline(7).DespiteimprovedconsensuscriteriaforthediagnosisofDLB,diagnosticaccuracyisstillmoderateinresearchandpoorinclinicalsettings,andDLBisoftenmisdiagnosedasAD(7).Frontotemporaldementia

Frontotemporaldementia(FTD), isagroupofneurodegenerativediseasescharacter‐isedbydeficitsinbehaviour,executivefunction,orlanguage,andmayoftenbemis‐takenforapsychiatricdisorder(8).FTD,whichisparticularlycommoninpeopleyoungerthan65years,maybecausedbyanumberofdifferentneuropathologicalcon‐ditions,allofwhicharecharacterisedbytheselectivedegenerationofthefrontalandtemporalcorticesofthebrain.Improvedtoolsforclinicalimaging,andmolecularchar‐acterisationhavemadeiteasiertodiagnosethedifferentsubtypesofFTD,andalsotodifferentiateFTDfrompsychiatricdisorders(8).Mixedtypedementia

Therearedifferenttypesofmixeddementia,butinthemostcommonformtheabnor‐malplaquesassociatedwithADexisttogetherwithbloodvesselproblemslinkedtoVAD.BrainchangestypicalforADmayalsocoexistwithLewybodiesandsometimesallthreeconditionsmayco‐exist.Thereissomeevidencefromautopsystudiessuggestingthatmixeddementiamaybemorecommonthanwhatwaspreviouslythought(9).Ontheotherhand,recentresultsfromacross‐sectionalstudysuggestlowprevalenceofmixeddementiainlate‐onsetAD(10).Howmanypeopleareaffectedbydementia?

AccordingtorecentfiguresfromtheWorldHealthOrganisation(WHO),47.5millionpeoplearoundtheworldareafflictedbydementia,andeveryyearthereareapproxi‐mately7.7millionnewcases(11).Theestimatedagestandardiseddementiapreva‐lenceforpeoplemorethan60yearsoldliesbetween5and7percentinmostcountries(12).In2012itwasestimatedthat1.6percentofpeopleinNorwaylivedwithdemen‐tia,whichissimilartotheEUaverageof1.5percent(13).Howmanypeoplethatatpre‐sentareundiagnosedandwholivewiththepre‐stagesofADorotherdementiasinthegeneralpopulationisunknown.InNorwaythereareresearchindicatingthatonlyabouthalfofpeoplewithdementialivinginnursinghomeshaveadiagnosis(14)andinhomecaretheproportionmaybeevenlower(15).

18 Background

Whataretheriskfactorsfordementia?

Oldageisthestrongestnon‐modifiableriskfactorforthedevelopmentofdementia(16).Otherriskfactorsincludefemalesex,genes,andfamilialdisposition.Sincethereistodatenocurativetreatmentfordementia,itisonlynaturalthatthemodifiableriskfactorsareofgreatinterest.Ithasbeensuggestedthatriskfactorsforvascularandmixedtypedementiainparticular,butalsoforAD,maybethesameasforcardiovascu‐lardisease(17,18),i.e.hypertension,highcholesterol,diabetes,inflammation,recur‐rentinfections,andfactorsrelatedtoanunhealthylifestyle,e.g.physicalinactivity,smoking,excessivealcoholconsumption,unhealthydiet,andobesity.Otherriskfactorsaredepression,loweducationalattainment,lowsocioeconomicstatus,andlackofso‐cialinteraction(19,20).Mildcognitiveimpairment(MCI)isoftenthoughtofasapre‐stagetoADandotherde‐mentias.MCIischaracterisedbyacognitivedeterioration,includingmemoryproblems,butwhichisnotsevereenoughtohamperaperson’sdaytodayactivities,andthusdonotmeetthedementiacriteria(21).AreviewoftheliteraturesuggestsaprevalenceofMCIbetween16and20percentinpeopleovertheageof60,accordingtothenewercriteria(22).MCIdoesnotalwaysprogresstodementia.Inarecentsystematicreviewmostincludedstudiesreportedprogressionratesbetween20to40%(10‐15%peryear)(22).MCIduetoADisonlyoneofmanysuggestedsubtypes.AmnesticMCI,isasubtypethatprimarilyaffectsthememory.Itmaysometimesbedifficulttodifferenti‐atenormalage‐relatedcognitivedeclinefromdifferenttypesofMCI,andMCIfromde‐mentia(23).UntilfairlyrecentlytherehasbeenlittleconsensusaboutwhichtoolsandcriteriatousefordiagnosingMCI.Intworecentpapers(23,24),consensuscriteriafordiagnosingMCIhasbeenputforward.Resultsfromalongitudinalcohortstudyofpeo‐plewithParkinson’sdisease(25)suggestthattheprognosticaccuracyofMCImaybeincreased,ifneuropsychologicaltestsarerepeatedlyadministeredovertime.Cognitiveimpairmentnodementia(CIND)isatermthatalsodescribespeoplewithbelownor‐malcognitivefunctioningwhodonotmeetdementiacriteria,butthedefinitionisbroaderthanthatofMCI(22).

Howtheinterventionsmaywork

Sinceitatpresentisnotpossibletocuredementia,itisnaturalthatfocusofanti‐de‐mentiastrategies,inadditiontofindingacure,isonpreventingitsonset.ADisaslowlyprogressingcondition,andearlypathophysiologicalchangesmaybepresentmanyyearsorevendecadesbeforetheactualdiagnosis(26).IfpeopleatriskfordevelopingADanddementiacanbeidentifiedatearlypre‐clinicalstages,awindowforearlyinter‐ventionsopensup(27).Toaccomplishthis,furtherresearchregardingthepathophysi‐ologicalprocessesbehindAD,andtheidentificationofADsensitivebiomarkers,isneededtodeterminewhichfactorsbestpredicttheriskofprogressionfrom“normal”cognitiontomildcognitiveimpairmentandADdementia(27).

19 Background

Thereare,asmentionedearlier,severalmodifiableriskfactorsfordementia,ofwhichcardiovascularandlifestylemodificationshavereceivedgreatinterest.Thus,manydementiapreventionstrategiesfocusoncardiovascularrisk‐factors,e.g.statinsforhighcholesterol,antihypertensivedrugsforhighbloodpressure,andcholin‐esteraseinhibitorsforinflammation.Below,wedescribesomeofthetheoriesbehindwhythese,andotherinterventions,areconsideredforthepreventionofcognitivede‐clineanddementia.Pharmacologicaltreatment

Bloodpressureloweringinterventions

Anumberoflongitudinalstudieshaveconsistentlyshowedarelationshipbetweenmid‐lifehypertensionandthedevelopmentofcognitiveimpairmentlaterinlife,espe‐ciallyifuntreated.Fortherelationshipbetweenlate‐lifehypertensionanddementiaorcognitiveimpairmenthowever,theresultsareinconsistent(28).Inadditiontopharma‐cologicaltreatment,hypertensionmayalsobereducedbyinterventionsincludingsaltrestriction,weightreduction,physicalexercise,andreducedalcoholconsumption(29).Cholesterolloweringdrugs(statins)

IncreasedserumcholesterollevelhasbeensuggestedtocontributetothepathologicalprocessesleadingtoAD.Statinsarecholesterolloweringdrugsofprovenbenefitinvasculardisease(30,31).Resultsfromrecentsystematicreviews,includingmeta‐anal‐yses,mostlyofobservationaldata,suggestinconsistenteffectsofstatinsonincidenceofdementia(32).Theseresultsshouldbeinterpretedwithcautionduetothehighhetero‐geneityfoundinthemeta‐analyses.Thereareknownadverseeffectsofstatinuse,butitisdebatedhowfrequentlyoccurringtheseeventsareamongusers(33).Cholinesteraseinhibitors

PatientswithADhavereducedcerebralproductionofcholineacetyltransferaseinthebrain,whichleadstoadecreaseinacetylcholinesynthesisandimpairedcorticalcholin‐ergicfunction(34).Cholinesteraseinhibitors,whichincreasecholinergictransmissionbyinhibitingcholinesterase,havebeenshowntobeofsomebenefitinpatientswithADaswellasothernon‐ADdementias,eventhoughtheeffectsinmostcaseshavebeenmodest(35).Peoplewithamnesticmildcognitiveimpairmentarealsobelievedtohaveacentralcholinergicdeficit(36).Ithasthereforebeensuggestedthatcholinesterasein‐hibitorscouldbeusedtodelayorevenpreventtheprogressionfromamnesticMCItoAD.Dietarypatternsanddietarysupplements

Mediterraneandiet

Ithasbeensuggestedthatvariousdietarypatternswithdifferingfoodandnutrientcompositions,mayelicitdifferenteffectsontheageingbrain.ThetraditionalMediterra‐neandiet(MD)ischaracterisedbyhighconsumptionofvegetables,fruits,oliveoil,leg‐umes,fish,wholegraincereals,nutsandseeds,andmoderateredwineconsumption.Itislowinprocessedfoods,dairyproducts,redmeat,andvegetableoils(37).Arecentre‐viewoftheliterature(38)includingmostlycohort‐andobservationalstudies,suggests

20 Background

thatMDmaybeeffectiveinreducingcognitivedeclineinolderage.Theauthorshow‐ever,questionthefeasibilityofintroducingthistypeofdietaryinterventioninWesterncountriesduetoculturaldifferences.

Omega‐3FattyAcids

Fattyacids(FAs)areamongthemostcrucialmoleculesforthebrain'sabilitytoper‐form,andthereisevidenceforarelationshipbetweeninadequatedietaryintakeoffattyacidsandimpairedbrainperformanceanddiseases(39).AnumberofpossiblemechanismsforaprotectiveroleofOmega‐3FAsindementiahavebeenputforward,butresultsfromtrialsinvestigatingtheeffectsofOmega‐3FAsoncognitioninpeoplewithandwithoutcognitiveimpairmenthavebeenmixed(40).VitaminE

VitaminEisadietarycompoundknownforitsabilitytoprotectcellsfromthenegativeeffectsoffreeradicals.EvidencethatfreeradicalsmaybeinvolvedinthepathologicalprocessesofAD(andothercognitiveimpairments)hasledtointerestintheuseofvita‐minEinthetreatmentofMCIandAD.ThereisalsosomeevidencesuggestinglowerlevelsofVitaminEinplasmaofpeoplewithADandinMCI(41),butwhetherthelowervitaminElevelisacauseoraneffectofpoordietaryintakeisdebated(42).Physicalexercise

Evidencefromameta‐analysis(43),andamorerecentlongitudinalstudy(44),sug‐geststhatthehighertheamountofregularexercise,thelowertheriskfordevelopingAD.Theevidenceforthepossiblecellularandmolecularmechanismssuggestedtoliebehindtheneuroprotectiveeffectofaerobicexercise,isfromanimalstudies.Thesug‐gestedmechanismsincludeincreasedcerebralbloodflowwhichinturntriggersdiffer‐entneurobiologicaleventsthatmayincreasethelevelsofdifferentgrowthfactorsthatareofimportancetotheneuroplasticityofthebrain,andtheantioxidantenzymelevels(45)Cognitivetraining,education,andsocialengagement

ThecognitivereservehypothesisaimstoexplainwhythosewithhigherIQ,education,andmoresociallyactivepeoplewhoparticipateinleisureactivitiesexhibitlesssevereclinicalorcognitivechangeswhenafflictedbyage‐relatedorADpathology(46,47).Thehypothesisproposesthatdifferencesbetweenindividualsinhowtasksarepro‐cessedbyourbrainsprovideareserveagainstbrainneuropathology.Thisreservemayallowformoreflexibleusageofthebrain,greaterneuralefficiencyandcapacity,andgreaterabilitytocompensatefordegeneratedbrainareasthroughtherecruitmentofadditionalbrainregions,andthusimprovedbrainfunction.IthasbeensuggestedthatevencognitivetraininginterventionsdeliveredlateinlifemayhaveaprotectiveeffectagainstADandagerelatedcognitivedecline(48).Thisoverviewofreviewsislimitedtoeffectsofinterventionsoncognitivelyhealthyolderpeople,andpeoplewithcognitiveimpairment.PeoplewithdiagnosedADorothertypesofdementiaarenotincluded.

21 Background

Whyisitimportanttodothisoverviewofreviews?

TheprevalenceofADandothertypesofdementiaiscontinuouslyincreasingasthepopulationintheworldissteadilyageing,andthereforealsothesocietalcostsofcaringforpeoplewithdementiaareincreasing(12).Astheretodateisnocurefordementia,thereisagreatneedtoidentifyeffectiveinterventionstodelayorpreventdementiaatanearlystageofthediseaseprocess.Inthisoverviewofreviewswesummarisetheevi‐dencefromsystematicreviewsoftheeffectivenessofinterventionstodelayorpreventtheonsetofdementiaincognitivelyhealthypeopleandinpeoplewithmildcognitiveimpairment.Ifeffectiveinterventionscanbeidentified,alotcanbesavedintermsofpersonalsuffering,healthcareuseandcostsforthesocietyasawhole.Also,helpingolderpeopletostaycognitivehealthy,wouldenablethislargeandgrowinggroupofpeopletoliveanactiveandindependentlife,andtocontributetothesocietylongafterretirement.

22 Methods

Methods

Weconductedanoverviewofreviewsevaluatingtheeffectivenessofinterventionsaimedatpreventingordelayingcognitivedeclineand/ortheonsetofdementia.Wein‐cludedreviewsofinterventionsthattargetedcognitivelyhealthypeople(primarypre‐vention)andpeoplewithcognitiveimpairment(secondaryprevention).Weexcludedreviews/studiesofinterventionstargetingpeoplewithadementiadiagnosis.Thisover‐viewofreviewswasconductedinaccordancewiththeguidanceforsummarisingevi‐dencedescribedintheKnowledgeCentre’shandbook(49).

Objectives

Tosummariseandcriticallyappraisetheexistingevidencefromsystematicreviewsoftheeffectivenessofinterventionstopreventcognitivedeclineanddementia.Specifi‐cally,weaimedtoanswerthefollowingquestions:

1. Whatisthedocumentedeffectivenessofinterventionstopreventcognitivedeclineandincidenceofdementiaincognitivelyhealthypeople(primaryprevention)?

2. Whatisthedocumentedeffectivenessofinterventionstoprevent(further)cognitivedeclineanddementiainpeoplewithmildcognitiveimpairmentorotherearlysymptomsorsignsofdementia(secondaryprevention)?

Inclusioncriteria

Weusedthefollowingcriteriawhenconsideringreviewsforinclusion:

Population Cognitivelyhealthyolderpeople Peoplewithmildcognitiveimpairment

Intervention Weincludedsystematicreviewsofstudiesevaluatingtheeffec‐tivenessofanyinterventionaimedatdelayingorpreventingtheonsetofdementiaincludingthefollowing:

pharmacologicaltherapy psychosocial dietary/nutritionalsupplements lifestylemodificatione.g.changesrelatedtounhealthy

diet,physicalinactivity,useofalcoholortobacco

23 Methods

Comparison Usualcare,nointerventionorotherintervention.Outcomes Developmentofdementiaandsymptomsofmildcognitiveim‐

pairment.Symptomsincluded,butwerenotlimitedto,cogni‐tion,dailyfunction,neuropsychiatricsymptomsofdementia(e.g.agitation,depressionandanxiety).

Studydesign Weconsideredsystematicreviewsofhighquality,inanylan‐guage,andpublished2009orlaterforinclusion.

Otherinclusioncriteriaandspecifications:Weconsideredareviewasbeingsystematicifitcontainedadescriptionof1)arobustsearchstrategy,2)criteriaforinclusionand3)assessmentofthequalityofincludedstudies.WeassessedthemethodologicalqualityofpossiblyeligiblestudieswiththeKnowledgeCentre’schecklistforsystematicreviews,andincludedonlyreviewsofhighmethodologicalquality.Inaddition,weincludedonlyfinalisedsystematicreviewsthatwecouldfindinfulltext.Incaseswheretherewasoverlapbetweenreviews(thesameincludedindividualstudies,approximatelythesameresearchquestionsasked),weuseddatafromthemostrecentlyupdatedreview(orthelargerandmoredetailedre‐view)andexcludedtheotherreviews.

Exclusioncriteria

Weappliedthefollowingexclusioncriteria:•Systematicreviewsconcerningmeasuresdirectedtowardsindividualswithdementiadiagnosis.•Systematicreviewsconcerningmeasuresdirectedtowardsrelativesofdementiasuf‐ferers.•Systematicreviewsconcerningpreventivecomplementary/alternativemeasuresagainstdementia(e.g.Acupuncture,aromatherapy),andalternativemeasuresthatarenotinterventionsdescribedintheinclusioncriteria.•Systematicreviewsoflowormoderatemethodologicalquality.•Primarystudiesandotherstudiesthatdonotsummarisetheeffectofpreventivemeasuresagainstdementia.•Abstractsandotherpublicationformatsthatarenotavailableinfulltextormissingdetailsofacompletedsystematicreview.

Literaturesearch

WesearchedthefollowingdatabasesforsystematicreviewsuptoFebruary2016:

OvidMEDLINE(R)In‐Process&OtherNon‐IndexedCitations,OvidMEDLINE(R)

Daily,OvidMEDLINE(R)

EMBASE(Ovid)

PsycINFO(Ovid) Cinahl

24 Methods

TheCochraneDatabaseofSystematicReviews(CDSR) CRD

WebofScience

Pubmed

Thedatabasesearchstrategywasdesignedbyandsearchesexecutedbyresearchli‐brarianGyriHvalStraumann.Thesearchstrategywaspeer‐reviewedbyresearchli‐brarianIngridHarboe.Thesearchwasadaptedtoeachdatabase.Weusedacombina‐tionofsubjectterms,textwords,andwhenavailableinthedatabases,filtersforsys‐tematicreviews.ThecompletesearchstrategyisavailableinAppendix2.Wesupple‐mentedthedatabasesearchbysearchingliteraturelistsofrelevantreviewsandin‐cludedstudies.

Selectionofreviews

Tworeviewersindependentlyreadallrecordsresultingfromthesearches.Weusedpre‐designedinclusion/exclusionformsforeachscreeninglevel:(i)thereviewti‐tle/abstract,(ii)thecompletefulltextofthereview,and(iii)thereview’smethodologi‐calquality(seedetailsbelow).Weresolveddisagreementsthroughdiscussionandsub‐sequentconsensus.Iftherewascompleteoverlapintermsofincludedstudiesbetweentwoormoreofthereviews,wereportedtheresultsfromthemostrecentreviewand/orthereviewwiththelargestnumberofincludedstudiesorthemostdetailedde‐scription.Itwasnotnecessarytocontacttheauthorsofanyreviewstoaidthedecisionprocess.Welistthereviewsconsideredinfull‐text,butsubsequentlyexcluded,inAp‐pendix3alongwiththereasonsforexclusion.Qualityassessmentaspartoftheselectionprocess:Tworeviewersindependentlyassessedthemethodologicalqualityofeachpossibleeli‐giblesystematicreviewusingtheKnowledgeCentre’schecklistforsystematicreviews(49).Thechecklistevaluatesthemethodsusedinareviewagainst10criteriatodeter‐minethedegreetowhichthereviewmethodsareunbiased.Areviewthatadequatelymet“allormostofthecriteria”wasconsideredtobeofhighquality,orifanyofthecri‐teriawerenotmet,ithadtobejudgedveryunlikelythatthiswouldaffectthereview’sconclusions.Weincludedonlysystematicreviewsofhighqualityandexcludedreviewsofmoderateorlowquality.Wedescribetheresultsofthequalityassessment,aswellasthechecklistitemsinAppendix4.

Dataextraction

Alldatawerereviewedandextractedbyonereviewer(GF)intoastandardiseddataex‐tractionform,whichwasthencheckedforaccuracybyanotherreviewer(RB).Thefol‐lowingdatawereextractedfromeachreview:citation,aimofthereview,theoryused/evidencebaseofintervention,numberofrelevantandnon‐relevantincludedstudies,studydesignoforiginalstudies,totalnumberofparticipants,baselinecharac‐teristicsofparticipants(age,gender,cognitivestatus,educationallevel,andethnicity),

25 Methods

country,typeandcomponentsofintervention,durationofinterventionandfollowup,comparators,methodsusedtoassessoutcomes(e.g.toolstoassesscognitivestatus,ortodiagnosedementia),effectsizesreported(andmeasuresofdispersion),statisticalmethodsusedandadjustmentsforconfoundingfactorsinmultivariatemodels,lossestofollowup,conflictofinterestofreviewauthorsandofauthorsoforiginalstudies(ifavailable),andtypeofreview(Cochraneornon‐Cochranereview).

Datasynthesis

Wepresenttheresultsseparatelyforprimarypreventioninterventionstargetingcog‐nitivelyhealthypeople,andsecondarypreventioninterventionstargetingpeoplewithcognitiveimpairmentand/ormemorycomplaints.Weorganisedthedatawithineachofthetwogroupsaccordingto:i)typeofinterven‐tionbeingevaluated(pharmacologicalinterventions,dietaryinterventions,cognitivetrainingandaerobicexercise),andii)typeofoutcomes(incidenceofdementia,cogni‐tivetestscores,adverseeffects,andotherclinicaloutcomes).Wereporttheresultsforthemainoutcomesintextandintables.Ifacombinedmeasureofperformanceonabatteryofcognitivetestswasreportedbythereviewauthors,weusedthissummaryestimatewhenreportingtheresults.Ifnosummaryofeffectestimatewasprovided,wereportedtherangeofeffects.Weconductednooverarchingmeta‐analysisoftheresultreportedintheincludedreviews,astheyallevaluateddifferenttypesofinterventions.

Gradingoftheevidence

Tworeviewauthors(GFandRB)usedtheGRADEtool(GradingofRecommendationsAssessment,DevelopmentandEvaluation)developedbytheGRADEworkinggroup(50)todeterminethecertaintyoftheestimatesofeffectsofinterventionsreportedintheincludedreviews,i.e.towhatdegreewecouldtrusttheresults.Weconsideredthecompileddocumentationforeachofthemainoutcomes(i.e.progressiontodementia,performanceofcognitivetests)usingGRADE.EvidencefromrandomisedcontrolledtrialsstartashighcertaintyevidencebutmaybedowngradeddependingonfivecriteriainGRADEthatareusedtodeterminethecer‐taintyoftheevidence:i)methodologicalstudyqualityasassessedbyreviewauthors,ii)degreeofinconsistency,iii)indirectness,iv)imprecision,andv)publicationbias.Up‐gradingofresultsfromobservationalstudiesispossibleaccordingtoGRADEifthereisalargeeffectestimate,oradose‐responsegradient,orifallpossibleconfounderswouldonlydiminishtheobservedeffectandthatthereforetheactualeffectmostlikelyislargerthanwhatissuggestedbythedata.InaccordancewiththeGRADEtool,wegradedthecertaintyoftheevidenceashigh,moderate,low,andverylow.ThesegradesofevidencearedefinedbytheGRADEWorkingGroupinthefollowingway:Highquality:Weareveryconfidentthatthetrueeffectliesclosetothatoftheestimateoftheeffect

26 Methods

Moderatequality:Wearemoderatelyconfidentintheeffectestimate:Thetrueeffectislikelytobeclosetotheestimateoftheeffect,butthereisapossibilitythatitissubstantiallydifferentLowquality:Ourconfidenceintheeffectestimateislimited:ThetrueeffectmaybesubstantiallydifferentfromtheestimateoftheeffectVerylowquality:Wehaveverylittleconfidenceintheeffectestimate:Thetrueeffectislikelytobesubstantiallydifferentfromtheestimateofeffect

Ethics

Inthisoverviewofreviewsweassessedwhetherandhowauthorsofincludedreviewsaddressedissuespertainingtoequity,benefitsandharms,andfinancialdisclosuresre‐portedintheoriginalstudies.Inaddition,wealsotooknotesonanyreportingofpa‐tientinvolvementindecisionsregardingthetrialdesign.

27 Results

Results

Descriptionofincludedreviews

Searchresults

Theliteraturesearchesandsearchingothersourcesyielded4,349uniquecitations(see

Figure1).Oftheseweexcluded4,247irrelevantcitationsonthebasisoftitleandab‐

stract.Weretrievedandevaluated102reviewsinfulltext.Forty‐oneofthe102re‐

viewswerepromotedtofurtherevaluationofmethodologicalquality(28,40,42,51‐

87),andtheotherswereeitherexcludedwithreasons(seeAppendix3),orlistedunder

ongoingstudies(seeAppendix5;73‐79).Wejudged10ofthe41reviewstobeofhigh

methodologicalqualityandeligibleforinclusioninthisoverviewofreviews(28,40,42,

51,57,65‐67,71‐73,78,85).Afterfurtherscrutinywefoundthattwoofthereviews

(52,57),whichevaluatedtheeffectsofcholinesteraseinhibitors,overlappedinterms

ofincludedstudieswiththereviewbyRussandcolleagues(78).Asthelatterreview

wasmoredetailedandincludedadditionalstudieswedecidedtoreporttheresults

fromthisreview.Twootherreviews(40,71)thatevaluatedtheeffectsofOmega3fatty

acidsupplements,hadtwooutofthreeincludedstudiestargetingcognitivelyhealthy

peopleincommon,andreportedresultsforoneuniquestudyeach.Sincetheunique

studyreportedinoneofthereviews(71)wassmall(49participants)andoflowqual‐

ity,andthestudyintheotherreview(40)waslarge(2,911participants),wedecidedto

includetheresultsforcognitivelyhealthypeoplereportedintheSydenhamreview

(40).Wereporttheresultsforpeoplewithmildcognitiveimpairment(MCI)fromthe

reviewbyMazereeuwandcolleagues(71).SeelistofexcludedreviewsinAppendix3.Characteristicsofincludedreviews

Weidentifiedeighthighqualityreviews(40,42,65,71‐73,78,85)thatweincludedinthisoverviewofreviews.Allthereviewsincludedrandomisedcontrolledtrialsonly.SixwereCochranereviews(40,42,72,73,78,85),andtwowerenon‐Cochranere‐views(65,71).Forfurtherdetailsseetable1andAppendix6.Populations

Fiveofthereviewsincludedstudiesthattargetedcognitivelyhealthypeople(40,65,72,73,85).Tworeviews(42,78)includedstudiesofpeoplewithmildcognitiveimpair‐ment(MCI).Inoneofthestudies(42)peoplewithMCIofamnestictypewereincludedwhiletheotherstudy(78)includedpeoplewithanytypeofMCI(howeverdefined).

28 Results

Onereview(71)includedbothcognitivelyhealthypeopleandthosewithcognitiveim‐pairment.Thelattergroupconstitutedpeoplewithmemorycomplaintsandobjectivecognitivedecline.Interventions

Threereviewsfocusedonpharmacologicalinterventions(72,73,78),andthreeevalu‐

atedtheeffectsofdietarysupplements(40,42,71).Onereviewwasconcernedwith

theeffectsofcognitivetraining(65),andonewiththeeffectsofaerobicexercise(85).

Noneoftheincludedreviewsevaluatedtheeffectsofotherlifestylechanges,e.g.change

toahealthydiet,reduceduseofalcoholortobacco,orinterventionstargetingother

riskfactorsfordementia,e.g.depression,lowlevelofeducation,orlackofsocialinter‐

action.

Comparators

Thecomparatorinterventionwasplaceboinallbuttworeviews(65,85),inwhichthe

activeinterventions(i.e.cognitivetrainingandaerobicexercise)werecomparedwith

eitheranotheractiveinterventionornointervention.Outcomes

Fouroftheincludedreviews(42,72,73,78)reportedondementiaincidence/progres‐siontodementia.Threeofthesealsoreportedcognitivetestscores(72,73,78).Theotherfourreviewsreportedonlycognitivetestscores(intotal>350differentcognitiveoutcomes).Oneofthereviews,however,didnotreportanybetween‐groupcompari‐sonsforthecognitivetestsscores(42).Fourreviews(40,72,73,78)reportedadverseeffectsofinterventions.Threereviews(72,73,85)reportedotherclinicaloutcomes(i.e.cholesterollevel,bloodpressure,andmeasuresofaerobiccapacity).Onereview(72)alsoreportedactivitylevelandinstrumentalactivitiesofdailyliving.

29 Results

Figure1PRISMAstudyflowdiagram(88)describingthereviewselectionprocess.

Toolsusedbyreviewauthorstodetermineriskofbiasandqualityoftheevidence

AllfiveincludedCochranereviewsusedtheCochraneriskofbiastool(89)toassesstheriskofbiasoftheincludedstudies.Inaddition,oneofthereviews(85)alsousedtheCLEARNPTtoolfornon‐pharmacologicalinterventions(90).Oneofthetwonon‐CochranereviewsusedthePEDroscale(91)andtheothernon‐CochranereviewusedboththeCochranetoolandthePEDroscale(65).NoneoftheincludedreviewsusedtheGRADEtool(50)toassessthecertaintyoftheincludedevidenceforaneffect,nordidtheyprovideasummaryoffindingstable.Toolsusedinoriginalstudiestodefineparticipantsascognitivelyimpaired

Theoriginalstudiessummarisedinthreeoftheincludedreviews(42,71,78)useddif‐ferenttestsandcriteriatodeterminethedegreeofcognitiveimpairmentofpartici‐pantsatbaseline.ThemostcommonlyusedtestswereClinicalDementiaRating(CDR)scale(92),themodifiedHachinskiIschemicScore(93)andtheMiniMentalStateExam‐ination(MMSE)(94).

Records screened (n =4, 349) 

Records identified through  database searching 

(n = 5, 824) 

Additional records identified through other sources 

(n = 3) 

Records after removal of duplicates (n =4, 349) 

Excluded records  (n = 4, 247) 

Full text reviews assessed for eligibility (n = 102  ) 

Reviews directly excluded (n=36) 

Full text reviews excluded with reasons  

(n = 49) Review protocols (n=9) 

Included reviews (n = 8) 

30 Results

Toolsusedinoriginalstudiestoassesseffectsoncognitivefunction

Around400differentcognitivetestswereusedtoassesstheeffectsoncognitivefunc‐tion,ofwhichtheMMSE(94)wasoneofthemostcommonlyusedtests.OthertestswereforexampletheADAS‐Cog,CDRSumofboxes,andtheSymbolDigitModalitiestest.Thenumberofdifferenttestsusedbysinglestudiesrangedfromonetomorethan10.Inoneofthereviews(65)onlythetotalnumberoftests(n=396)usedinthe52in‐cludedstudieswasreported,butnoinformationonthetypeandnumberoftestsusedinindividualstudieswasprovided.Littleinformationwasprovidedaboutthescalingandinterpretationofthedifferenttests(e.g.thedesireddirectionofeffect).ToolsusedinoriginalstudiestodetermineeffectsonincidenceADanddementia

Fouroftheincludedreviews(42,72,73,78)reportedonincidenceofdementia.Crite‐riathatwereused,withtheverificationofthediagnosisbyablindedexpertpanel,weretheDiagnosticandStatisticalManualofMentalDisorderscriteria(DSM‐IIIR)(95)theInternationalStatisticalClassificationofDiseasesandRelatedHealthProblemscrite‐ria(ICD10)(96),andtheNationalInstituteofNeurologicalandCommunicativeDisor‐dersandStrokeandtheAlzheimer'sDiseaseandRelatedDisordersAssociationcriteria(NINCSD‐ADRDA)(26).ThesecriteriaweresometimesusedtogetherwiththeDSM‐IVcriteria(97).Table1.Descriptionofthereviews(n=8)includedinthisoverviewofreviews.

Review Search date No of studies1 Population Intervention

Farina 2012(42) June 2012 3 (1) Cognitively im-paired

Vitamin E

Lampit 2014 (65) July 2014 52 (52) Cognitively healthy

Cognitive training

Mazereeuw 2012 (71)

September 2011 10 (7)2 Cognitively im-paired (and cog-nitively healthy)

Omega 3 Fatty Acids

McGuinness 2009 (73)

February 2008 4 (4) Cognitively healthy

Hypertensive drugs

McGuinness 2016 (72)

November 2015 2 (2) Cognitively healthy

Cholesterol lower-ing drugs (statins)

Russ 2012 (78) Not reported 8 (8) Cognitively im-paired

Cholinesterase in-hibitors

Sydenham 2012 (40)

April 2012 3 (3) Cognitively healthy

Omega 3 Fatty Acids

Young 2015 (85) August 2013 12 (12) Cognitively healthy

Aerobic exercise

1Fromtheincludedreviewsweonlyusedstudieswithpopulationsandinterventionsthatwererelevantforourresearchquestion.Thenumberswithinparenthesisgiveinformationonhowmanystudiesintheincludedreviewthatmetourinclusioncriteria.2FromthisreviewonlythreestudiestargetingpeoplewithMCIwasincludedinthisoverviewofreview,astheincludedstudiestargetinghealthypeopleoverlappedalmostentirelywiththeSydenhamreview.

Theeighthighqualitysystematicreviewsincludedintotal86uniqueoriginalstudies

(range1to52studies)thatwererelevantforourresearchquestion.

31 Results

Wefoundnoreviewsconcernedwiththeeffectsoflifestylechanges(otherthanaerobic

exercisetraining),i.e.healthyeating,reducedalcoholconsumption,smokingcessation,

orinterventionstargetingotherriskfactorsfordementia,likeforexamplemid‐life

depression,loweducationalattainment.Wealsodidnotfindanyreviewsconcerned

withmultifacetedinterventionstodelayorpreventcognitivedecline,ADorothertypes

ofdementia.

Primarypreventioninterventions

Sixreviews(fourCochranereviews(40,72,73,85)andtwonon‐Cochranereviews(65,

71)),evaluatedtheeffectivenessofinterventionsaimedatpreventingcognitivedecline

and/ordementiaincognitivelyhealthypeople.Thefocusofthreeofthereviews(72,

73,78)wasonpharmacologicalinterventions(i.e.antihypertensivedrugs,cholesterol

loweringdrugs,andcholinesteraseinhibitors).Onefocusedontheeffectsofdietary

supplements(40),oneontheeffectsofaerobicexercise(85),andonereviewwascon‐

cernedwiththeeffectsofcognitivetraining(65).

Thecharacteristicsoftheincludedreviewsaredescribedbelow.

Pharmacologicaltherapies

Antihypertensivedrugs

OnereviewbyMcGuinnessetalfrom2009(73)evaluatedtheeffectsofantihyperten‐

sivedrugtherapyforthepreventionofcognitivedeclineandincidenceofdementia.

ThereviewincludedfourRCTsofcognitivelyhealthyolderpeoplewithhypertension

withnoapparentpriorcerebrovasculardisease.Allfourtrialsmettheinclusioncriteria

ofouroverviewofreviews.Thereviewauthorsjudgedthetrialstobeofhighmethodo‐

logicalquality.Theincludedoriginalstudieswerepublishedbetween1991and2008.

Participants

Theaverageageofparticipants(n=15,936)was75.4years.MeanMMSEscoreatbase‐

linerangedfrom26to29acrossstudies.InoneofthetrialsbaselineMMSEscorewas

notreported,onlytheMMSEcut‐offforinclusion(>24).Theaveragelengthofeduca‐

tionintwoofthestudieswas11.7and12.3yearsrespectively,whileinonestudymore

than50%ofparticipantshadeithernoeducationorprimaryschooleducationonly.

Onestudydidnotreporteducationallevel.Meanbloodpressurelevelatbaselinewas

171/86mmHgacrossstudies.Theparticipantswererecruitedfromanumberofcoun‐

triesandgeographicregions:WesternandEasternEurope,NorthAmerica,China,anda

smallerproportionofparticipantsfromAustralasiaandTunisia.Themajorityweream‐

bulatorypatients,recruitedinthecommunityorinprimarycaresettings.

Intervention

32 Results

Allfourincludedstudiesreportedasteppedcareapproachtohypertensiontreatment,

andevaluatedtheeffectsofdifferentfirstlinedrugtherapies(calcium‐channelblock‐

ers,thiazidediuretics,orangiotensinIItypeIreceptorblockers).Thesecondlinedrugs

includeddiuretics,beta‐blockers,centrallyactingagents,andACEinhibitors.Forfur‐

therdetailsonthedrugregimenpleaseseethefulltextreview(73).Themeanduration

offollowuprangedfrom1.8to4.5yearsacrossstudies(median2.8years).

Comparison

Thecomparatorinterventionwasplacebo.Itshouldhoweverbenotedthatmanyofthe

participantsinthecontrolgroupwereprescribedothernon‐studyantihypertensive

drugsduringthestudyperiodastheirbloodpressureexceededpre‐setvalues.Inone

ofthetrials84percentofthecontrolgroupparticipantsand75percentofthetreat‐

mentgroupwerenotonassignedtherapyatstudyend.Inanothertrial44percentof

theplacebogroupparticipantsascomparedwith10%oftheinterventionparticipants

werenotonassignedtherapyatstudyend.

Outcomes

Theprimaryoutcomesinthisreviewwereincidenceofdementiaandchangeincogni‐

tivetestscores(onetest;MMSE),forwhichpooledresultswerereportedbythereview

authors.Secondaryoutcomeswerebloodpressurelevel,adverseeffectsrequiringdis‐

continuationoftreatment,andqualityoflife.Pooledeffectestimateswereprovidedfor

thefirsttwo,whilethequalityoflifedatacouldnotbeanalysed.Incidenceofdementia

wasasecondaryoutcomeinalltrialsincludedinthereview.

Table2.Characteristicsofthereviewonantihypertensivedrugtherapy(McGuinness2009).

No of studies Population Intervention Comparison Outcomes 4 trials N=15,936 cognitively

healthy hypertensive participants; mean age 75.4 years (range 70.3 to 83.6); mean BP 171/86 mmHg

Antihypertensive drug therapy1,2

Follow up: range 1.8 to 4.5 years (median 2.8 years)

Placebo3

Incidence of dementia, change in cognitive test scores,4 blood pressure level, incidence and se-verity of adverse effects requiring discontinuation of treatment5

1 The studies used different first-line drugs, and a number of different second-line drugs. 2 McGuinness and colleagues intended to include also non-pharmacological blood pressure lowering interventions, e.g. salt restriction, weight reduction, exercise, alcohol restriction, smoking cessation, but failed to find any eligible studies evaluating non-drug interventions for inclusion. 3 A large proportion of control group participants received other antihypertensive medication during the study. 4 Incidence of dementia and cognitive test scores were secondary outcomes in all four included studies. Only one test, the Mini Mental State examination (MMSE) were used to assess cognitive change. No information about the tests i.e. scaling or interpretation of the MMSE scores, was provided by the review authors.5Type of adverse effects that required discontinuation of treatment not further described.

Cholesterolloweringdrugs

OnereviewbyMcGuinnessetalfrom2016(72)summarisedtheeffectsofstatins(cho‐

lesterolloweringdrugs)forthepreventionofdementiaandcognitivedecline.There‐

viewincludedtwotrialsinvolvingcognitivelyhealthypeoplewithahistoryof,orat

33 Results

highriskforcerebrovasculardisease.Bothtrialswererelevantforourresearchques‐

tions.Thereviewauthorsjudgedthetrialstobeofhighmethodologicalquality.Both

trialswerepublishedin2002.

Participants

Theageofparticipants(n=26,340)inthetwotrialsrangedfrom40to82years.Oneof

thetrialsrecruitedpeoplewithcoronaryheartdisease,otherocclusivearterialdisease,

anddiabetesorhypertension(n=20,536patients40to80yearsold,n=5,806wereat

least70yearsold),andtheothertrialrecruitedpeople(n=5,804;age70‐82years)with

ahistoryof,oratriskofvasculardisease.Oneofthetrialsdidnotprovidebaseline

measuresofcognitivefunction,andtheothertrialreportedameanMMSEscoreof28

acrossgroups.Educationallevelorsocioeconomicclasswerenotreportedineither

study.Totalcholesterollevelrangedfrommean5.7to5.9mMol/Latstudyentry.One

ofthetrialswasconductedintheUK,andtheotherinScotland,Ireland,andtheNeth‐

erlands.Mostoftheparticipantswereambulatorypatients,recruitedeitherinacom‐

munityorinaprimarycaresetting.Intervention

Differenttypesofstatinswereusedinthetwotrials:inonetrialparticipantsreceived

simvastatin40mgdaily,andintheothertheparticipantsreceivedpravastatin40mg

daily.Meanfollowupwas3.2yearsinonestudyand5yearsintheother.

Comparison

Thecomparatorinterventionwasplacebo

Outcomes

Theoutcomesreportedinthereviewwere:incidenceofdementia,changeincognitive

testscores(4tests;MMSE;StroopColourWordTest;Picture‐WordLearningtest;Let‐

terDigitCodingtest)orininterviewassessedcognitivefunction(TICS‐m),cholesterol

level,andincidenceandseverityofadverseeffects.

Table3.Characteristicsofthereviewoncholesterolloweringtherapy(McGuinness2016).

No of studies Population Intervention Comparison Outcomes 2 trials N=26,340 cognitively

healthy people with a history of, or risk fac-tors for, vascular dis-ease, age between 40 and 82 years

Cholesterol lowering drugs (Statins1)

Follow up: 3.2 years and 5 years

Placebo Incidence of dementia2, change in cognitive test scores2; cholesterol level; incidence and se-verity of adverse effects

1 The two trials used different types of Statins (simvastatin and pravastatin) 2 Incidence of dementia was a secondary outcome in the trial reporting this outcome .No information was provided on what criteria that was used to diagnose dementia. 3 A description of the cognitive tests i.e. scaling and how to interpret them (i.e. if higher or lower score would indicate an improvement) was not provided for all tests: Mini Mental State Examination (MMSE) Score: Score out of 30 (higher score better) (measures global cognitive function); Stroop Colour Word Test: Total number of seconds required to complete the third Stroop card containing 40 items (Measures atten-tion); Picture-Word Learning Test 15 Picture Learning Test (measures immediate and delayed recall); Letter Digit Coding Test. Total number of correct entries completed in 60 seconds (measures processing speed); Modified Telephone Interview for Cognitive Status Score (Score out of 39) A TICS-m score below 22 out of 39 was pre-specified as indicative of some cognitive impairment..

34 Results

Dietarysupplements

Omega‐3FAs

OnereviewbySydenhametalfrom2012(40)evaluatedtheeffectsofOmega‐3FAsup‐

plementsoncognitivedeclineincognitivelyhealthypeople.Thereviewincludedinto‐

talthreetrialsthatwererelevantforourresearchquestion.Thereviewauthorsjudged

thetrialstobeatlowriskofbias.Thetrialswerepublishedbetween2008and2011.

Population

Thethreetrialsrecruitedintotal4,080cognitivelyhealthyparticipantsagedbetween

60and80years.PerformanceontheMMSEwasusedtodeterminetheeligibilityof

participants.ThestudieswereconductedintheNetherlands,England,andWales,and

inonestudythecountryoforiginwasunclear.Inoneofthestudiestheparticipants

wererecruitedusingadatabasewithvolunteers,inonestudyparticipantsweredrawn

frompatientlistsof20generalpractices,andinthethirdstudyparticipantswithprevi‐

ousmyocardialinfarctionwererecruitedthroughcardiologists.

Intervention

IntwostudiesinterventionparticipantsreceivedgelcapsulescontainingOmega‐3FAs.

Inonestudy,participantsreceivedmargarinespreadcontainingOmega‐3FAs.Forde‐

tailsondosespleaseseethefulltextreview.Thefollowuprangedfrom6to40months

(median24months).

Comparison

Thecomparisoninterventionwasmatchedplaceboinalltrials:capsuledwithhigh‐

oleicsunfloweroil(onestudy),withOmega‐9richoliveoil(onestudy),andplacebo

margarineinthethirdstudy.

Outcomes

ThisreviewaimedtoinvestigatetheeffectofOmega‐3FAsonincidenceofdementia

butnoneoftheincludedstudiesreportedthisoutcome.Thereportedoutcomeswere

cognitivetestscores,adverseeffects,andadherence.Theauthorscategorisedthecog‐

nitivetestsintosixsub‐categories(MMSE,immediateanddelayedrecall,verbalflu‐

ency,andexecutivefunction)andpooledtheseseparately.

Table4.CharacteristicsofthereviewontheeffectsofOmega‐3FAs(Sydenham2012).

No of studies Population Intervention Comparison Outcomes 3 trials1 N=4,080 cognitively

healthy people aged 60 years or more

Omega 3 FAs (sup-plements or pro-vided meals) Follow up: range 6 to 40 months

Placebo or usual diet

Cognitive tests scores2, adverse effects, and ad-herence to supplementa-tion

1 Two trials were included in both Mazareeuw 2012 and Sydenham 2012, and both reviews reported results from one unique study each. As the unique study reported in Mazereeuw 2012 was very small compared to the one reported in Sydenham 2012, it was excluded, and we here present the results reported in the review by Sydenham 2012. 2 A description of the cognitive tests i.e. scaling and how to interpret them (i.e. if higher or lower score would indicate an improvement) was not provided.

35 Results

Physicalexercise

Aerobicexercise

OneCochranereviewbyYoungandcolleaguesfrom2015(85)evaluatedtheeffectsofaerobicexerciseoncognitivefunctionincognitivelyhealthyolderpeople.There‐viewincludedintotal11trials(n=754),whichallwererelevantforourresearchquestion.Thereviewauthorsjudgedthemethodologicalqualityoftheincludedstud‐iestobeathightomoderateriskofbias,asassessedwiththeCochraneriskofbiastoolandtheCLEARNRTtool(meanscore:33;IQR:29to37;thescalegoesfrom14to48points,lowerscoresbetter).Thetrialswerepublishedbetween1990and2012.Participants

The11trialsrecruitedintotal754cognitivelyhealthyparticipants,withameanagerangingfrom60to91yearsacrossstudies.Themediannumberofparticipantsintheincludedstudieswas49(IQR:30to101participants).ThetrialswereconductedintheUSA,CanadaandFrance.Baselineaerobicandcognitivecapacityofparticipantswasnotreported. InterventionTheaerobicexercisewaseitherbriskwalking,jogging,orcycling,oracombination.Thefrequencyanddurationofsupervisedtrainingrangedfrom1to3one‐hourses‐sionsperweek,withamajorityofstudiesproviding3X60minutestrainingperweek.Thedurationoffollowuprangedfrom8to24weeks(median16weeks).ComparisonThecomparatorwaseithernointervention,oranactiveintervention(aflexibil‐ity/balanceprogramme,astrengthtrainingprogramme,orasocial/cognitivepro‐gramme).Inamajorityofstudiesthefrequencyand/ordurationofsessionswerenotmatchedtothoseoftheinterventionprogramme.

OutcomesTheoutcomesreportedinthisreviewwerecognitivetestscores(oneto14testsacrossstudies),measuresofaerobiccapacity(i.e.VO2max,steptests,differentwalk‐ingtestsorratepressureproduct),anddropoutrate.Thecognitivetestswerecatego‐risedinto11subdomainsandpooled.

36 Results

Table5.Characteristicsofthereviewontheeffectsofaerobicexerciseoncognition(Young2015).

No of studies Population Intervention Comparator Outcomes 11 trials1

N= 754 cognitively healthy participants; mean age ranged from 61 to 91 years

Aerobic exercise (i.e. brisk walking, jogging or cycling) Follow up: range 8 to 24 weeks (me-dian 16 weeks)

No intervention or other active intervention (i.e. strength training, flexibil-ity/balance training, or social/cognitive training).

Cognitive tests scores2, measures of aerobic capacity, and dropout rate.

1 One additional study reported a secondary analysis of a subgroup of participants from one of the included trials, but these results were not used in the review. 2 A description of the cognitive tests i.e. scaling and how to interpret them (i.e. if higher or lower score would indicate an improvement) was not provided for the cognitive tests:

Cognitivetraining

Weincludedonenon‐Cochranereview(65)thatevaluatedtheeffectivenessofcomput‐

erisedcognitivetraining(CCT)inattenuatingage‐relatedcognitivedeclineincogni‐

tivelyhealthyolderpeople.Thisreviewincludedintotal51randomisedstudies(52in‐

dependentcomparisons).Thereviewauthorsjudgedthemethodologicalqualityofthe

includedstudiestobemixed:34of51studieswerejudgedtobeathighriskofbias,as

assessedwiththeCochrane’sriskofbiastool.ThemeanPEDroscorewas6.2/9(SD

1.35),andaccordingtothistool35studieswerejudgedtobeathighriskofbias.The

trialswerepublishedbetween1997and2013.

Participants

Thetrialsrecruitedintotal4,885participants,withmeanagesrangingfrom60.7to

81.9yearsacrossstudies.Accordingtothereviewauthorstheparticipantslackedany

majorcognitive,neurological,psychiatric,and/orsensoryimpairment. The MMSE

scoresatbaselinerangedfrom>24to29.3acrossstudiesthatreportedabaseline

measureofcognitivefunction(39of51studies).Themediannumberofparticipantsin

theincludedstudieswas44(IQR:30to67participants).Theparticipantsweremainly

fromtheUSAandEurope,butalsotoalesserextentfromCanada.Australia,Israel,

China,TaiwanSpecialAdministrativeRegion,RepublicofKorea,andJapan.

Intervention

TheCCTprogramswerestandardized,computerisedtasksorvideogames.Thetotal

durationofthetrainingvariedfrom4to60hours,thenumberofsessionsvariedfrom

3to50(of15to120minutesduration),andtheintensityofintervention(frequencyof

sessionsperweek)rangedfromonetoseventimesperweek.Variouselectronicde‐

vices,e.g.personalcomputers,mobiledevices,orgamingconsoles,wereusedtodeliver

thetraining.ThereviewauthorscategorizedtheCCTprogramsinto5groups:multi‐do‐

maintraining(24studies),speedofprocessingtraining(ninestudies),working

memorytraining(ninestudies),attentiontraining(sixstudies),andinfourstudies

video‐gameswereused.Fifteenstudiesused‘in‐house’programs,andtheremaining36

studiesusedcommercialcognitivetrainingprograms,orvideo‐games.Theintervention

wasdeliveredeitherascenter‐basedgrouptraining(32studies;61.5%),orasunsuper‐

visedtrainingintheparticipant’shome(19studies;36.5%).Nodataonadherenceto

theinterventionwasprovided.

37 Results

Comparison

Thecomparisoninterventionswereeitheractive,i.e.anotherintervention(26studies;

50%),orpassive(nointervention).Theactivecontrolconditionswerenotfurtherde‐

scribed.Outcomes

Theprimaryoutcomeofthisreviewwaschangeincognitivetestscoresfrombaseline

toimmediatelypost‐training.Nolong‐termeffectswereevaluated.Intotal,396cogni‐

tiveoutcomeswerereported.Noinformationwasprovidedonwhattestshadbeen

usedinthedifferentstudies.

Table6.Characteristicsofthereviewontheeffectsofcognitivetraining(Lampit2009).

No of studies Population Intervention Comparator Outcomes 51 trials (52 com-parisons)

N=4,885 cognitively healthy participants; mean age ranged from 60 to 82 years

Computerized cognitive training (CCT);> 4 duration; home-based (unsupervised) or centre-based (group) training; 5 different types of CCT Follow up: 40 to 60 hours total duration (unclear number of weeks)

Active or pas-sive interven-tions

Cognitive tests scores1

1 The names of the cognitive tests, details on the scaling and how to interpret the test scores were not provided in the review, neither was the number of tests used in the included studies.

Secondarypreventioninterventions

Threereviews(twoCochranereviews(42,78)andonenon‐Cochranereview(71)),

evaluatedtheeffectsofsecondarypreventioninterventions,i.e.pharmacologicalther‐

apy(78),dietarysupplements(71),todelayorpreventtheonsetofdementiainpeople

withMCIorearlysymptomsandsignsofdementia.

Pharmacologicaltherapy

Cholinesteraseinhibitors

OneCochranereviewbyRussetalfrom2012(78)evaluatedtheeffectivenessofcho‐

linesteraseinhibitorsinpreventingordelayingprogressiontodementiainpeoplewith

MCI.Thereviewincludedninetrials(8publishedreports)whichallwererelevantfor

ourresearchquestions.Thereviewauthorsjudgedallincludedstudiestobe“wellcon‐

ductedandrobustRCTs”.Thetrialswerepublishedbetween2004and2012.

Participants

Theeightincludedstudiesrecruitedintotal5,149participantswithMCI(anywayde‐

fined)withameanagerangingfrom45to90years.Threeofthestudieswerecon‐

ductedintheUSA,oneinUSAandCanada,oneinGermany,oneinSingapore,andin

38 Results

twomultisitestudiestheparticipantswerefrom16and14countriesrespectively.The

mediannumberofparticipantsperstudywas769(range19to2,037participants).

Intervention

Threedifferentcholinesteraseinhibitorswereusedintheincludedstudies:donepezil

(threestudies),rivastigmine(twostudies),andgalantamine(fourstudies).Onetrial,

whichreportedontwodifferentdosages,wasreportedintwostudies.Inamajorityof

studies,theend‐dose(afterescalationfromalowerdose),was10‐12mg/day.

Comparator

Thecomparatorinterventionwasplacebo.Inonestudybothinterventionandcontrol

groupsalsoreceivedmultivitamins.Themedianfollowupwas76weeks(range16

weeksto48months).

Outcomes

Thereviewreportedonprogressiontodementia(primaryoutcome)andadverseef‐

fectsofthedrugtherapy.Otheroutcomeswerechangeinperformanceoncognitive

tests(5tests;MMSE,ADAS‐Cog,CDR‐sumofboxes,SymbolDigitModalities,ADCS‐

ADL)andmortality.Thereviewauthorsreportthepooledeffectsofincidenceofde‐

mentiaat3years,andpooledeffectsofadverseevents(anyadverseevent,seriousad‐

verseevents,mortality,andotheradverseevents)thatoccurredduringthestudy.

Table7.Characteristicsofthereviewontheeffectsofcholinesteraseinhibitors(Russ2012).

No of studies Population Intervention Comparator Outcomes 9 trials N= 5,149 participants

with MCI; age 45 to 90 years

Cholinesterase inhibi-tors1(end-dose typi-cally 10-12 mg/day) Follow up: range 16 weeks to 48 months (median 76 weeks)

Placebo Progression to dementia, side effects, change in cognitive test scores2 and mortality

1 Three different cholinesterase inhibitors were used in the included studies (Donepezil, Rivastigmine and Galantamine) 2 A description of the cognitive tests i.e. scaling and how to interpret them (i.e. if higher or lower score would indicate an improvement) was not provided for the cognitive tests:

Dietarysupplements

VitaminE

OneCochrane‐reviewbyFarinaetal(42)evaluatedtheeffectivenessofvitaminEsup‐

plementationinpreventingprogressiontodementiainpeoplewithMCI.Thisreview

includedthreetrials,ofwhichonethatrecruitedparticipantswithMCI,wasrelevant

forouroverviewofreviews.Thereviewauthorsjudgedthetrialtobeatlowriskof

bias.Theincludedstudywaspublishedin2005.

Population

Inthismulti‐centretrial,people(n=516)withamnestictypeofMCIandanaverageage

of73yearsparticipated.ThetrialwasconductedintheUSAandinCanada.

39 Results

Intervention

TheincludedtrialhadthreetreatmentgroupsbutonlydatafortheVitaminE(alpha‐

tocopherol;2000IU/day)groupandtheplacebogroupwereevaluated.Bothinterven‐

tionandcontrolgroupparticipantsalsoreceivedmultivitamins.Thedurationoffollow

upwas3years.

Comparator

Thecomparatorinterventionwasplacebo.

Outcomes

TheoutcomesreportedinthereviewwereprogressionfromMCItopossibleorproba‐

bleAD(primaryoutcome),adverseeventsanddeath.

Table8.CharacteristicsofthereviewontheeffectsofvitaminEonADincidence(Farina2012).

No of studies Population Intervention Comparator Outcomes

3 trials (of which one

was included in this

overview of reviews)

N=5161 participants

with amnestic MCI;

mean age 73 years,

46% females

Vitamin E (alpha-to-

copherol; 2000

IU/day)+ multivita-

mins

Follow up: 3 years

Placebo + multivita-

mins

Time to progression

from MCI to possible

or probable AD; cog-

nitive test scores2

1 Total number of participants, including the third study arm (n= 769). The third study arm was not included in Farina 2012.. 2 The authors of the original trial did not report any between group comparisons for any of the cognitive (secondary)outcomes. Nor did they

provide information on the cognitive tests i.e. scaling and interpretation.

Omega3FAs

OnereviewbyMazereeuwandcolleaguesfrom2012(71)evaluatedtheeffectsof

Omega‐3FAssupplementsoncognitivefunctioninpeoplewithcognitiveimpairment

butnodementia(CIND).Thereviewincludedfourtrials,encompassing650partici‐

pantswithMCIormemorycomplaints,whichwerealleligibleforinclusioninourover‐

viewofreviews.Thereviewauthorsjudgedthemethodologicalqualityoftheincluded

studiestobehigh.Thetrialswerepublishedbetween2006and2011.

Population

Participants(n=650)withCINDand/ormemorycomplaintswererecruited.Meanage

oftheparticipantsrangedfrom68.5to74.5acrossstudies.Twoofthestudiesrecruited

peoplewithCIND(inonestudythiswasnotconfirmedwithanycognitivetests).The

othertwostudiesrecruitedpeoplewithmemorycomplaints.MeanMMSEscoreranged

from27.6to28.2acrossstudies.Inonestudyparticipantsscored>25onmemoryCom‐

plaintQuestionnaireScale).ThestudieswereconductedintheNetherlands,England

andWales,Japan,Australia,Israel,andUSA.Themediannumberofparticipantsinthe

includedstudieswas86(range:21to487).

Intervention

40 Results

TheinterventionconsistedintheprovisionofOmega‐3FAsupplementstopeoplewith

cognitiveimpairment.Thedurationoffollowuprangedfrom12.8to27weeks(median

19.5weeks).

Comparator

Thecomparatorinterventionwasplacebo.Outcomes

Themainoutcomeswerechangeinperformanceonanumberofcognitivetestswhich

werecategorisedintosevencognitivesubdomains(MMSE,compositememory;imme‐

diaterecall;delayedrecall,recognition,attentionandprocessingspeed,working

memory,andexecutivefunction).

Table9.CharacteristicsofthereviewontheeffectsofOmega‐3FAs(Mazereeuw2012).

No of studies Population Intervention Comparator Outcomes 4 trials N= 650 participants

with CIND; mean age 65.7 to 74.6; MMSE 27.6-28.2. One study did not report MMSE scores

Omega-3 FAs. Dura-tion: 12.8 to 27 weeks (median 19.5 weeks) Follow up::range 12.8 to 27 weeks (median 19.5 weeks)

Placebo Cognitive test scores1.

1 A description of the cognitive tests i.e. scaling and how to interpret them (i.e. if higher or lower score would indicate an im-provement) was not provided for the cognitive tests.

Certaintyoftheevidence

Allstudiesintheincludedreviewswererandomisedcontrolledtrialswhichareconsid‐eredthehighestlevelofevidence.Insixoftheincludedreviews,theauthorsjudgedtheriskofbiasofincludedstudiestobelow.Inonereviewtheauthorsjudgedtherisktobemoderatetohigh(85),andinonereviewamajorityofstudieswerejudgedtobeathighrisk(65),whichresultedindowngradingoftheevidence.Thecertaintyoftheincludedevidenceforthemainoutcomes(dementiaandpossibleorprobableAD)asassessedusingtheGRADEtool,wasmoderateinthreeofthere‐viewsthatreportedthisoutcome(72,78,42),andlowinone(73).Forcognitivetestscores(whichwerereportedinallbutoneofthereviews)thecertaintyoftheevidencerangedfromlowinonereview,duetoriskofbiasandimprecision(85),topredomi‐nantlymoderatetohighcertaintyofevidence,intheotherincludedreviews.Inonere‐view(78),however,thecertaintyofevidencefortheeffectoncognitivetestscoresrangedfromverylowtohigh.Regardingtheresultsforadverseeffects(threereviewsreportedpooledresults)(72,73,78),thecertaintyofevidencewasmoderatetohigh.ForfurtherdetailsseeAppendix7Evidenceprofiles.

41 Results

Effectsofinterventions

Primarypreventioninterventions

Pharmacologicaltherapies

Antihypertensivedrugs

PooledresultsofincidenceofdementiafromthereviewbyMcGuinnessandcolleagues

(73)(4studies;n=15,427participants),suggestaslightdifferencebetweenpartici‐

pantswhoreceivedantihypertensivedrugsandthosewhoreceivedplacebo(OR:0.89

[0.74to1.07],p=0.21)at1.8to4.5yearsfollowup.Belowwereporttherangeofeffect

sizesforcognitivetestscores(MMSE),adverseeffects,andchangeinbloodpressure

levelacrossstudies,ratherthanthepooledeffectestimatesreportedbythereviewau‐

thors,intable10(allp<0.00001),astheheterogeneitywasveryhigh(I2=97‐98%).The

meandifferenceforthechangeinMMSEscores(3trials;n=10,640)rangedfrom1.80

to0.07;adverseeffects(3trials;n=12,091)rangedfromOR0.92to1.11.Themeande‐

creaseinbloodpressurerangedfrom3.2to15.0mmHgandfrom1.6to5.9mmHgfor

systolicanddiastolicbloodpressurerespectively.

Inconclusion,lowcertaintyofevidencesuggestthatantihypertensivedrugsmaylead

toaslightdecreaseinincidenceofdementiaascomparedwithplacebo.Theresults

shouldbeinterpretedwithcaution,duetothewideCI,andlargelossestofollowupin

someoftheincludedstudies.

Table10.Summaryoffindingstableoftheeffectsofantihypertensivedrugsoncognitiveimpair‐mentanddementiaincidence(McGuinness2009).

Antihypertensive drugs vs. placebo for prevention of cognitive impairment and dementia in patients without prior cerebrovascular disease

Patient or population: cognitively healthy participants without prior cerebrovascular disease Setting: Western and Eastern Europe, North America, China, Australasia and Tunisia Intervention: antihypertensive drugs Comparison: placebo

Outcomes Anticipated absolute effects* (95% CI) Relative ef-fect (95% CI)

№ of parti-cipants (studies)

Quality of the evi-dence (GRADE) Risk with placebo Risk with antihypertensive drugs

Dementia inci-dence1

Study population OR 0.89 (0.74 to 1.07)

15, 427 (4 RCT)

⨁⨁◯◯

LOW 2,3

34 per 1000 30 per 1000 (25 to 36)

Change in cog-nitive test score (MMSE)4

The mean cogni-tive test scores (MMSE) was 0

The mean change in cognitive test scores from baseline (MMSE) in the in-tervention group was 0.42 higher (0.3 higher to 0.53 higher)

- 10, 640 (3 RCT)

⨁⨁◯◯

LOW 5

Study population

42 Results

Antihypertensive drugs vs. placebo for prevention of cognitive impairment and dementia in patients without prior cerebrovascular disease

Patient or population: cognitively healthy participants without prior cerebrovascular disease Setting: Western and Eastern Europe, North America, China, Australasia and Tunisia Intervention: antihypertensive drugs Comparison: placebo

Outcomes Anticipated absolute effects* (95% CI) Relative ef-fect (95% CI)

№ of parti-cipants (studies)

Quality of the evi-dence (GRADE) Risk with placebo Risk with antihypertensive drugs

Adverse events6 186 per 1000

187 per 1000 (174 to 202)

OR 1.01 (0.92 to 1.11)

12, 091 (3 RCT)

⨁⨁◯◯

LOW 5,6

The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).

CI: Confidence interval; OR: Odds ratio; MD: Mean difference Note: The results on dementia incidence can be compared with the 80 per 1000 dementia incidence in Norway for people over the age of 80.

1. Follow up ranged from mean 2 to 4.5 years across three studies, and was average 1.8 years in one study. 2. Relatively wide CI. 3. Indirectness as many of the people in the control group also received non-study antihypertensive drugs. 4. MMSE: Mini Mental State Examination. Scaling and interpretation of scale not further described. 5. Very high heterogeneity (I2=97-98%).* 6. Adverse events that required a discontinuation of treatment. No further description of the adverse events provided.

Cholesterolloweringdrugs(statins)

TheresultsfromthereviewbyMcGuinnessandcolleaguesfrom2016(72)indicateno

differenceindementiaincidence(1study;n=20,536participants)betweenpartici‐

pantsreceivingstatinsandthosereceivingplacebo(OR:1.00(0.61to1.65))atmean5

yearsfollowup(SeeTable11).Further,theauthorsreportednobetweengroupdiffer‐

encesincognitivetestscores(MMSE,SCWT,Pictureword,Letterdigittests;rangeof

meandifferences:0.8higherto0.01lowerthancontrol),andnodifferencesfortele‐

phoneinterviewassessedcognitivestatus(TICS;meandifference0.02).Inaddition,

thereweresimilareffectsonadverseeffects(2trials;OR:0.94[0.83to1.05}).TheLDL

cholesterollevel(2trials;1.2mmol/L[‐1.24to‐1.15])waslowerintheintervention

groupascomparedtocontrol.

Adherencetothedrugtherapyintheinterventiongrouprangedfrom82to94%across

studies.Between3and5%ofinterventionparticipantswereonnon‐studystatins

aloneand2%onboth,andbetween10and17%ofcontrolpatientsreceivednon‐study

statins.Inonetrialapproximately25%oftheparticipantswerelosttofollowup,while

intheothertrialthelossestofollowupwereverysmall.Inbothtrialsanalysiswasby

intentiontotreat.

Inconclusion,thereismoderatecertaintyofevidenceindicatingthatstatintherapy

giventoolderpeoplewith,oratriskof,vasculardisease,probablyleadstolittleorno

differenceinincidenceofdementiaascomparedtoplacebo.Therewereveryfew

events(andwideCI),whytheresultsshouldbeinterpretedwithcaution.

43 Results

Table11.Summaryoffindingstableoftheeffectsofcholesterolloweringtherapyonincidenceofdementiaandcognitivedecline(McGuinness2016).

Cholesterol lowering therapy (statins) vs. placebo for the prevention of dementia and cognitive de-cline in cognitively healthy people

Patient or population: cognitively healthy people Setting: UK, Ireland and the Netherlands Intervention: statins (cholesterol lowering drugs) Comparison: placebo

Outcomes Anticipated absolute effects* (95% CI) Relative ef-fect (95% CI)

№ of partic-ipants (studies)

Quality of the evidence (GRADE)

Risk with pla-cebo

Risk with statins

Dementia incidence; mean follow up: 5 years

Study population OR 1.00 (0.61 to 1.65)

20, 536 (1 RCT)

⨁⨁⨁◯

MODERATE1

3 per 1000 3 per 1000 (2 to 5)

Cognitive test scores (MMSE); mean follow up: 3.2 years1

The mean cogni-tive test scores (MMSE) was 0

The mean change in cognitive test scores (MMSE) in the in-tervention group was 0.06 higher (0.04 lower to 0.16 higher)

- 5, 804 (1 RCT)

⨁⨁⨁⨁ HIGH

Cognitive test scores (Stroop, Picture world test, Letter Digit test)

The mean cogni-tive test scores was 0

The mean change in cognitive test scores in the intervention group ranged from 0.8 higher to 0.01 lower

- 5, 804 (1 RCT)

⨁⨁⨁◯

MODERATE 2

TICS-m2 scores at mean 5 years (final visit)

The mean TICS-m scores at final visit was 0

The mean TICS-m scores at fi-nal visit in the intervention group was 0.02 higher (0.12 lower to 0.16 higher)

- 20, 536 (1 RCT)

⨁⨁⨁⨁ HIGH

Adverse effect; at 3.2 to 5 years follow up

The mean adverse effects was 0

The mean adverse effects in the intervention group was 0.94 higher (0.83 higher to 1.05 higher)

- 26, 340 (2 RCT)

⨁⨁⨁⨁ HIGH

*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; OR: Odds ratio; MD: Mean difference

1 We downgraded the evidence due to few events and wide CI. 2 Change from second baseline visit to last, on treatment for MMSE (Score out of 30 (higher score better), Stroop test (Total number of sec-onds required to complete the third Stroop card containing 40 items; lower is better), Picture word (15 Picture Learning Test; unclear scaling), and Letter digit tests scores (Total number of correct entries completed in 60 seconds, higher is better). Mean follow up 3.2 years. Moderate quality for Stroop test, but high for the other three tests. 3 TICS-m: The modified Telephone Interview of Cognitive Status. Score out of 39. No further information on interpretation of test. 4 Adverse effects requiring discontinuation of treatment i.e. Rhabdomyolysis; creatinine kinase levels >10 times upper limit of normal values, a liver aminotransferase levels>3 times upper limits of normal values.

Dietarysupplements

Omega‐3FAs

ThereviewbySydenhamandcolleaguesfrom2012(40)reportednodifferencesfor

thepooledresults(3studies;n=4,080)ofeightcognitivetestscores(SeeTable12)be‐

tweenparticipantsreceivingOmega‐3FAsandthosereceivingplaceboat6to40

monthsfollowup.Inadditiontherewerelittleornodifferencesinadverseeffectsbe‐

tweengroups.

44 Results

Treatmentcompliancewasreportedtobehigh(81%and99%)acrosstreatmentandplacebogroup.

Inconclusion,hightomoderatecertaintyofevidencesuggestthatOmega‐3FAsproba‐blyleadlittleornodifferenceincognitivedecline,ascomparedwithplacebo,whengiventocognitivelyhealthyolderpeople.

Table12.SummaryoffindingstableoftheeffectsofOmega‐3FAsoncognitivedeclineanddementia(Sydenham2012).

Omega-3 FAs vs. placebo for prevention of cognitive decline and dementia in cognitively healthy older people

Patient or population: cognitively healthy older people Setting: England and Wales Intervention: omega-3 fatty acids Comparison: placebo

Outcomes Anticipated absolute effects* (95% CI) Rela-tive effect (95% CI)

№ of participants (studies)

Quality of the evidence (GRADE)

Risk with placebo

Risk with Omega-3 fatty acids

Cognitive test scores1 (MMSE). Follow up: range 24 to 40 months

The mean cognitive test scores - was 0

The mean MMSE scores in the intervention group was 0.07 lower (0.25 lower to 0.1 higher)

- 3, 221 (2 RCT)

⨁⨁⨁◯ MODERATE 2

Cognitive test scores1(Imme-diate recall, Delayed recall, Word recognition, Number of animals named). Follow up: range 6 to 24 months

The mean cognitive test scores was 0

The mean cognitive test scores3 in the intervention group was 0.06 standard deviations higher to 0.04 standard deviations lower

- 1, 043 (3 RCT)

⨁⨁⨁⨁ HIGH

Cognitive test score1 (Digit span forward and backward). Follow up: range 6 to 24 months

The mean cognitive test score was 0

The mean cognitive test score4 in the intervention group was 0.12 to 0.03 higher

- 1, 018 (3 RCT)

⨁⨁⨁⨁ HIGH

*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio; OR: Odds ratio;

1. A description of the cognitive tests i.e. scaling and how to interpret them (i.e. if higher or lower score would indicate an improve-ment) was not provided for the cognitive tests.

2. High I2=53% 3. We have reported the range of effects (SMD) since none of the individual test scores showed an effect of the inter-

vention. 4. We have reported the range of effects (MD) for two tests here.

Aerobictraining

ThereviewbyYoungetal.from2015(85),reportednodifferencesbetweengroupsfortheindividuallypooledtestscoresfor10cognitivecategories(11trials;n=754)forthecomparisonaerobicexercisevs.nointervention(SDM[8outcomes]:0.09lowerto0.30higher;MD[2outcomes],0.10to0.16).SeeTable13below.Theresultsforthecompar‐isonaerobicexerciseversusanyactiveintervention(SDM[10outcomes],range:0.26lowerto0.38higher;MD[1outcome]:0.15)(resultsnotshowninTable12).Forall

45 Results

outcomestheconfidenceintervalincludedthepointofnoeffect.FormoredetailsseeevidenceprofileinAppendix6.

Inconclusion,aerobicexercisemayleadtolittleornoeffectontheperformanceoncognitivetests.

Table13.Summaryoffindingstableoftheeffectsofaerobicexerciseoncognitivedecline(Young2015).

Aerobic exercise vs. no intervention for the prevention of cognitive decline

Patient or population: cognitively healthy older people Setting: USA, France and Canada Intervention: Aerobic exercise Comparison: no intervention

Outcomes3 Anticipated absolute effects* (95% CI) Rela-tive ef-fect (95% CI)

№ of participants (studies)

Quality of the evidence (GRADE)

Risk with no inter-vention

Risk with Aerobic exercise

Cognitive speed 1 - SMD 0.12 higher ( 0.16 lower to 0.41 higher) - 260 (5 RCTs)

⨁⨁◯◯

LOW 2,3

Verbal memory func-tions (immediate) 1

- SMD 0.09 higher ( 0.24 lower to 0.43 higher) - 137 (2 RCTs)

⨁⨁◯◯

LOW 2,3

Visual mamory func-tions (immediate) 1

- SMD 0.09 lower ( 0.57 lower to 0.40 higher) - 65 (1 RCT)

⨁⨁◯◯

LOW 2,3

Working memory 1 The mean working memory was 0

The mean working memory in the interven-tion group was 0,3 higher (0.54 lower to 1.15 higher)

- 137 (2 RCTs)

⨁⨁◯◯

LOW 2,3

Memory functions (delayed) 1

- SMD 0.09 higher ( 0.23 lower to 0.41 higher) - 152 (2 RCTs)

⨁⨁◯◯

LOW 2,3

Executive functions 1 - SMD 0.18 higher ( 0.16 lower to 0.53 higher) - 217 (3 RCTs)

⨁⨁◯◯

LOW 2,3

Cognitive inhibition1 - SMD 0.20 higher ( 0.06 lower to 0.47 higher) - 217 (3 RCTs)

⨁⨁◯◯

LOW 2,3

Visual attention 1 - SMD 0.05 higher (0.26 lower to 0.37 higher) - 155 (3 RCTs)

⨁⨁◯◯

LOW 2,3

Auditory attention1 The mean auditory at-tention was 0

The mean auditory attention in the interven-tion group was 0,16 higher (0.01 lower to 1.33 higher)

- 65 (1 RCT)

⨁⨁◯◯

LOW 2,3

Motor function 1 The mean motor func-tion was 0

The mean motor function in the intervention group was 0,1 higher (7.87 lower to 8.08 higher)

- 65 (1 RCT)

⨁⨁◯◯

LOW 2,3

*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; SMD: Standardised mean difference; MD: Mean difference

46 Results

Table13.Summaryoffindingstableoftheeffectsofaerobicexerciseoncognitivedecline(Young2015).

Aerobic exercise vs. no intervention for the prevention of cognitive decline

Patient or population: cognitively healthy older people Setting: USA, France and Canada Intervention: Aerobic exercise Comparison: no intervention

Outcomes3 Anticipated absolute effects* (95% CI) Rela-tive ef-fect (95% CI)

№ of participants (studies)

Quality of the evidence (GRADE)

Risk with no inter-vention

Risk with Aerobic exercise

1. A description of the cognitive tests i.e. scaling and how to interpret them (i.e. if higher or lower score would indicate an improvement) was not provided for the cognitive tests.

2. High to moderate risk of bias in at least one domain. 3. Fewer than 300 participants (imprecision).

Computerisedcognitivetraining

ThereviewbyLampitetalfrom2014(65),reportedasmalleffectofcomputerised

cognitivetraining(CCT)oncognitivetestscores(52studies;n=4,445participants)in

cognitivelyhealthyolderpeople(Hedge’sg;95%CI:0.22[0.15to0.29],p<0.001).

Therewaslowtomoderateheterogeneity(I2=29.92,p=0.03).Cognitivefunctionwas

measureddirectlyafterthetraining,whichvariedindurationfrom4to60hoursinto‐

tal.Seetable14.

Insummary,thereismoderatecertaintyofevidenceofasmallbeneficialeffectofCCT

oncognitivetestscores,directlyafterthetraining.

Table14.Summaryoffindingstableoftheeffectsofcomputerisedcognitivetrainingonage‐re‐latedcognitivedecline(Lampit2014).

Computerised cognitive training vs. active or passive intervention for prevention of age-related cog-nitive decline in cognitively healthy older people

Patient or population: cognitively healthy older people Setting: USA, Europe, Canada, Australia, Israel, China, Taiwan, Republic of Korea and Japan Intervention: Computerised cognitive training Comparison: active or passive intervention

Outcomes Anticipated absolute effects* (95% CI) Rela-tive ef-fect (95% CI)

№ of participants (studies)

Quality of the evidence (GRADE)

Risk with ac-tive or pas-sive interven-tion

Risk with Computerised cog-nitive training

Cognitive test scores1

The mean cog-nitive test scores was 0

The mean cognitive test scores in the intervention group was 0.22 higher (0.15 higher to 0.29 higher)

- 4, 885 (52 RCT)

⨁⨁⨁◯ MODERATE2

*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; MD: Mean difference

1. No information was provided on the type and number of cognitive tests used in the 51 included studies. No information on scaling and interpretation of test scores. Only short-term effects were evaluated.

2. Thirty-three of 51 included studies were at high risk of bias. Eighteen studies were at low risk.

47 Results

Secondarypreventioninterventions

Pharmacologicaltherapy

CholinesteraseinhibitorsThereviewbyRussetalfrom2012(78)reportedaslightdifferenceinprogressiontodementia(9studies;n=4,207)betweengroupsreceivingcholinesteraseinhibitorsandthosereceivingplaceboat3years(RR:0.84[0.70to1.02]).Further,thereviewauthorsreportnoeffectofcholinesteraseinhibitorsoncognitivetestscores,apartfromasmallandmostlikelyclinicallyinsignificantbeneficialeffectonasingletest.Thereweresig‐nificantlymoreadverseeventsinthecholinesteraseinhibitorgroups(RR:1.09[1.02to1.16]),butnomoreseriousadverseeventsordeaths.Resultsforanyadverseeventwereasfollows:gastrointestinalsideeffects:diarrhea(RR2.10[1.30to3.39]),nausea(RR2.97[2.57to3.42]),vomiting(RR4.42[3.23to6.05]);musclespasmsorlegcramps(RR7.52[4.34to13.02]);abnormaldreams(RR4.25[2.57to7.04]);insomnia(RR1.66[1.36to2.02]);syncopeordizziness(RR1.62[1.36to1.93]);andheadache(RR1.34[1.05to1.71]).Insummary,cholinesteraseinhibitorsprobablyleadtoaslightdecreaseinincidenceofdementia,ascomparedtoplacebo(moderatecertainty).However,moderatecertaintyofevidenceindicatehigherincidenceofanyadverseeventsinthecholinesterasegroup,butasimilareffectonseriousadverseevents.TheresultsforincidenceofdementiaandseriousadverseeffectsshouldbothbeinterpretedwithcautionduetothewideCI.Evidenceofmixedcertaintysuggestlittleornoeffectofcholinesteraseinhibitorsoncognitivetestscores.

Table15.Summaryoffindingstableoftheeffectsofcholinesteraseinhibitorsondementiain‐cidenceandcognitivedecline(Russ2012).

Cholinesterase inhibitors vs. placebo for prevention of dementia and cognitive decline in people with cognitive impairment no dementia (CIND)

Patient or population: people with mild cognitive impairment Setting: USA, Canada, Singapore and a number of other countries Intervention: Cholinesterase inhibitors Comparison: placebo

Outcomes Anticipated absolute effects* (95% CI) Relative effect (95% CI)

№ of par-ticipants (studies)

Quality of the evidence (GRADE)

Risk with placebo

Risk with Cholines-terase inhibitors

Conversion to dementia Follow up: mean 3 years

Study population RR 0.84 (0.70 to 1.02)

1, 530 (2 RCT)

⨁⨁⨁◯

MODERATE1

237 per 1000 199 per 1000 (166 to 241)

Cognitive test scores (ADAS-Cog1). Follow up: mean 2 years

The mean cognitive test scores was 0

The mean cognitive test scores in the intervention group was 0.78 lower (1.92 lower to 0.35 higher)

- 2, 675 (4 RCT)

⨁⨁◯◯

LOW 2,3

48 Results

Table15.Summaryoffindingstableoftheeffectsofcholinesteraseinhibitorsondementiain‐cidenceandcognitivedecline(Russ2012).

Cholinesterase inhibitors vs. placebo for prevention of dementia and cognitive decline in people with cognitive impairment no dementia (CIND)

Patient or population: people with mild cognitive impairment Setting: USA, Canada, Singapore and a number of other countries Intervention: Cholinesterase inhibitors Comparison: placebo

Outcomes Anticipated absolute effects* (95% CI) Relative effect (95% CI)

№ of par-ticipants (studies)

Quality of the evidence (GRADE)

Risk with placebo

Risk with Cholines-terase inhibitors

Cognitive test scores (CDR Sum of boxes1). Follow up: mean 1 years

The mean cognitive test scores was 0

The mean cognitive test scores in the intervention group was 0.1 lower (0.11 lower to 0.09 lower)

- 1, 269 (2 RCT)

⨁⨁⨁⨁ HIGH

Cognitive test scores- (Symbol digit modalities1). Follow up: mean 6 months

The mean cognitive test scores was 0

The mean cognitive test scores in the intervention group was 0.17 higher (2.87 lower to 3.21 higher)

- 312 (2 RCT)

⨁◯◯◯

VERY LOW2,3,

Cognitive test scores (MMSE); scale from 0 to 30. Follow up: mean 1 years

The mean cognitive test scores was 0

The mean cognitive test scores in the intervention group was 0.24 higher (0.13 lower to 0.61 higher)

- 1, 269 (2 RCT)

⨁⨁◯◯

LOW 2,3

Cognitive test scores 1(ADCS-ADL); scale from 0 to 78. Fol-low up: mean 1 years

The mean cognitive test scores was 0

The mean cognitive test scores in the intervention group was 0.15 higher (0.27 lower to 0.57 higher)

- 2, 408 (3 RCT)

⨁⨁⨁◯

MODERATE2

Serious adverse events4 Study population RR 0.97 (0.86 to 1.10)

4, 207 (5 RCT)

⨁⨁⨁◯

MODERATE3

190 per 1000 185 per 1000 (164 to 209)

Any adverse events5 Study population RR 1.09 (1.02 to 1.16)

4, 207 (5 RCT)

⨁⨁⨁◯

MODERATE3

825 per 1000 899 per 1000 (841 to 957)

*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio; MD: Mean difference

1. A description of the cognitive tests i.e. scaling and how to interpret them (i.e. if higher or lower score would indicate an improve-ment) was not provided for the cognitive tests. 2. Wide CI. 3. I2=79-98% and thus very high heterogeneity. 4. Serious adverse events (not further described) and death. 5. Any adverse events including gastrointestinal side effects (diarrhoea, vomiting and nausea), cardiac problems, muscle spasms or leg cramps, headache, syncope or dizziness, insomnia and abnormal dreams.

49 Results

Dietarysupplements

VitaminE

ThereviewbyFarinaetalfrom2012(42)reportedlittleornodifferenceinthepro‐gressiontopossibleorprobableADbetweentheVitaminEandtheplacebogroup.Atstudyendpoint(36months),76outof257participants(29.6%)inthevitaminEgroupand73outof259participants(28.2%)intheplacebogrouphadprogressedtoAD(HR:1.02:95%CI0.74to1.41,p=0.91)(seeTable16).Nobetween‐groupcomparisonswerereportedforthecognitivetestscores.Sixty‐sixparticipantsfromthetreatmentgroupand72fromtheplacebogroupdroppedoutduringthestudy,withthemainreasonsbeingdeath,adverseeffects,andwithdrawalofconsent.Noinformationwasprovidedregardingthedistributionofthereasonsfordiscontinuingtreatmentacrossgroups,oronadverseeffectsoftheintervention.

Insummary,thereismoderatecertaintyofevidencethatvitaminEprobablyleadstolittleornodifferenceinincidenceofpossibleorprobableAD,ascomparedtoplacebo.However,duetothewideCI,theresultsshouldbeinterpretedwithcaution.

Table16.SummaryoffindingstableoftheeffectsofvitaminEonprogressiontoAD(Farina2012)

Vitamin E vs. placebo for prevention of Alzheimer’s disease in people with MCI

Population: people with mild cognitive impairment Setting: USA and Canada Intervention: Vitamin E Comparison: Placebo

Outcomes Anticipated absolute effects* (95% CI)

Relative effect (95% CI)

№ of participants (studies)

Quality of the evidence (GRADE)

Risk with pla-cebo

Risk with Vita-min E

Possible or probable Alzheimer's disease. Follow up: mean 3 years

Study population HR 1.02 (0.74 to 1.41)

516 (1 RCT)

⨁⨁⨁◯

MODERATE 1,2

282 per 1000

287 per 1000 (217 to 373)

*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the rela-tive effect of the intervention (and its 95% CI).

CI: Confidence interval; HR: hazard ratio

1. Only one study. Relatively small groups. 2. Wide CI overlapping no effect. *As noted by the review authors, measures were taken at 6, 12, 18, 24, 30 and 36 months, but results were only reported for 12 and 36 months, and for the remainder only changes from baseline were provided.

Omega‐3FAs

ThepooledresultsfromthereviewbyMazereeuwetalfrom2011(68)suggestlittleornoeffectsofOmega‐3FAsoncognitivetestscores(4studies;n=650)atmedian14.5to24weeksfollowup.Onlyfortwoofthesevencognitivesubdomainswasthereabenefi‐cialeffectoftheintervention:immediaterecall(Hedge’sg(95%CI):0.16(0.01to0.32)),andattentionandprocessingspeed(Hedge’sg(95%CI):0.32(0.03to0.61).Seetable17.Therewerenodifferencesindropoutrateoradverseeventsbetweengroups.

50 Results

Insummary,evidence,mainlyofhighcertainty,suggestthatOmega‐3FAsprobablyleadtolittleornodifferenceincognitivetestscores,ascomparedtoplacebo.

Table17.SummaryoffindingstableontheeffectsofOmeaga‐3FAsoncognitivetestscores(Ma‐zereeuw2012).

Omega-3 FAs vs. placebo for improved cognitive function in people with CIND

Population: people with cognitive impairment no dementia (CIND) Setting: the Netherlands, England, Wales, Japan, Australia, Israel, and USA Intervention: omega-3 FAs Comparison: placebo

Outcomes Anticipated absolute effects* (95% CI) Rela-

tive ef-fect (95% CI)

№ of parti-cipants (studies)

Quality of the evidence (GRADE)

Risk with placebo Risk with Omega-3 fatty acids

Cognitive test scores (Composite memory1). Follow up: median 14.5 weeks

The mean Compo-site memory score was 0

The mean cognitive test scores in the intervention group was 0.1 higher (0.06 lower to 0.25 higher)

- 676 (4 RCT)

⨁⨁⨁⨁ HIGH

Cognitive test scores (Immediate and delayed recall1). Follow up: me-dian 19.5 weeks

The mean Immediate recall test score was 0

The mean cognitive test scores in the intervention group was 0.16 to 0.03 higher

- 676 (4 RCT)

⨁⨁⨁⨁ HIGH

Cognitive test scores (Recognition1)

The mean Recogni-tion test score was 0

The mean cognitive test in the intervention group was 0.03 lower (0.18 lower to 0.13 higher)

- 655 (3 RCT)

⨁⨁⨁⨁ HIGH

Cognitive test scores (At-tention and processing speed1)

The mean Attention and processing speed test scores were 0

The mean cognitive test in the intervention group was 0.32 higher (0.03 higher to 0.61 higher)

- 193 (3 RCT)

⨁⨁⨁◯

MODERATE2

Cognitive test scores (Working memory and executive function1)

The mean Working memory and execu-tive function test scores was 0

The mean cognitive test scores - in the intervention group was 0.04 higher (0.13 lower to 0.21 higher)

- 533 (2 RCT)

⨁⨁⨁⨁ HIGH

Cognitive test scores – (MMSE). Follow up: mean 24 weeks

The mean MMSE score was 0

The mean cognitive test scores in the intervention group was 0.06 lower (0.23 lower to 0.12 higher)

- 483 (1 RCT)

⨁⨁⨁⨁ HIGH

*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; MD: Mean difference 

1 A description of the cognitive tests i.e. scaling and how to interpret them (i.e. if higher or lower score would indicate an improvement) was not provided for the cognitive tests 2 Downgraded due to imprecision (wide CI).

Ethics

Noneoftheincludedreviewsdiscussedequity,orthelackofadiscussionofequityis‐sues,inthestudiesincludedinthereviews.Nordidtheyattempttodiscussthebalance

51 Results

betweenbenefitsandharms,takingintoaccountthecostsandcost‐effectivenessofin‐terventions.Inamajorityoftheincludedreviews,possiblefinancialinterestsandcon‐flictsofinterestrelatedtotheinterventionunderstudywerenotdiscussed.Forexam‐ple,inthereviewbyLampitandcolleagues(65),inwhichcommercialandsocalled‘in‐house’computerprogramswereusedtodeliverthecognitivetraining,noinformationonpossiblefinancialconflictsofinterestwasprovided(i.e.ifpeopleconductingthetri‐alswereinanywayassociatedwith,orfundedby,thepeopleproducingtheCCTpro‐grams).Noneoftheincludedreviewsdiscussedanyconsumerinvolvementintrialde‐sign,orthelackofit.

52 Discussion

Discussion

Mainresults

Inthisoverviewofreviewsweincludedeighthighqualityreviewssummarisingtheef‐fectsofinterventionsaimedatpreventingcognitivedecline,AD,andotherdementiasincognitivelyhealthypeople(primaryprevention),andinpeoplewithmildcognitiveim‐pairment(secondaryprevention).Moderatetohighcertaintyofevidencefromfourofthesereviewssuggeststhatstatintherapy(72)andthedietarysupplementsevaluated(40,42,71)probablyleadtolittleornodifferenceincognitivedeclineandincidenceofdementiaascomparedtoplacebo.Moderatecertaintyofevidenceforcholinesterasein‐hibitors(78),andlowcertaintyofevidenceforantihypertensivedrugs(73)indicatethatthesedrugsmayleadtoslightlydecreasedincidenceofdementia,ascomparedtoplacebo.WideCIsandfeweventsinsomeoftheanalyseswarrantcautionwheninter‐pretingtheresults.Moderatecertaintyofevidencefromonereview(65)suggeststhatcomputerisedcognitivetraining(CCT)probablyleadstoasmallimprovementincogni‐tivetestscoresdirectlyafterthetraining,butthelong‐termeffectsareunknown.Lowcertaintyofevidencefromonereview(85)indicatethataerobictrainingpossiblyleadstolittleornodifferenceincognitivefunction.Threeoftheincludedreviews(42,72,73)reportedsimilarratesofadverseeventsininterventionandcontrolgroups.Onereview(78)reportedhigherincidenceofadverseeventinthetreatment(cholinesterase)groupascomparedtoplacebo,andforseriousadverseeventstheinterventioneffectvaried.

Certaintyoftheevidence

Alloriginalstudiesincludedinthereviewswererandomisedcontrolledtrials,whichconstitutethehighestlevelofevidence.Thecertaintyoftheevidencefortheoutcomedementiaincidence,asassessedusingtheGRADEtool,wasmoderateinthreeofthefourreviewsthatreportedthisoutcome,andlowinonereview,Moderatecertainty,meansthatthetrueeffectislikelytobeclosetotheestimateoftheeffect,butthereisalsoapossibility(bothfortheinterventionssupportedbylowandmoderatecertaintyofevidence)thattheeffectsoftheevaluateddrugsandthedietarysupplementsmaybesubstantiallydifferent.Somecautionwheninterpretingtheresultsonincidenceofde‐mentia,isinorder,astheincidenceratewasratherlow(around3%),forsomeoftheanalysesand/ortheCIsrelativelywide.Forthecognitivetestscoresthequalityofevi‐dencerangedfromhightolow(duetoimprecision,inconsistencyand/orhighriskofbiasforsomeofthecognitiveoutcomes).

53 Discussion

Strengthsandweaknesses

Weconductedacomprehensivesearchforsystematicreviewsofinterventionstodelayorpreventcognitivedeclineanddementia,whichwasupdatedinFebruary2016.Therewerenolanguagerestrictions.Wesearchedthereferencelistsofincludedreviews,andthereferencelistsofotherrelevantpublicationsidentified.Thesearchstrategywasde‐velopedbyanexperiencedinformationscientist,whoalsoconductedthesearch.Inad‐dition,twoauthorsindependentlyscreenedallthereferencesforinclusion,whichmakesitlesslikelythatwemissedanyrelevantreviews.Wealsogradedthecertaintyoftheevidenceinduplicate.Weincludedonlyreviewsthatwejudgedtobeofhighquality,andtheyinturnincludedonlyRCTs.SixoftheeightincludedreviewswereCochranereviews.Allreviewswerepublishedbetween2009and2016.Therearesomelimitationswithouroverviewofreviews.Oneisthatweonlyincludedhighqualitysystematicreviews,asitcanbearguedthattheexclusionofmoderate‐tolowqualityreviewsmayhaveexcludeddatarelevantforthepresentwork.Anotherlimitationisthatwereliedentirelyonthedescriptionoftheoriginalpapersprovidedintheincludedsystematicreviews,whenconductingouroverviewofreviews.Athirdlimitation,relatedtotheoldsearchdateinoneoftheincludedreviews(literaturesearchin2008),isthatwemayhavemissedtoincludemorerecentlypublishedstudiesontheeffectsofantihypertensivedrugs.Alimitationwiththeavailableevidenceisthattherearenostudieswithlongerfollowuptimethan5years.Itmaybedoubtedthatthisissufficienttimeforchangetotakeplacegiventhenatureofthecondition(i.e.dementia)andtheeffectsoftheinterven‐tionsunderstudy.Evenifsomereviewsincludedstudiesofpeoplewithdementia,whichwasagroupofpeoplenotwithinthescopeofouroverviewofreviews,thisdidnotconstituteaprob‐lemaslongasresultsforeachgroup(cognitivelyhealthy,MCIanddementia)werere‐portedseparately.WehadtoexcludeonereviewtargetingpeoplewithParkinson’sdis‐ease(67),andfourotherreviews(58,63,66,77)ascognitivelyhealthypeopleandcog‐nitivelyimpairedpeoplewerenotanalysedseparately.Wecould,ifitisjudgedim‐portant,includeathirdmixedgroup,ifwedecidetoupdatethisoverviewofreviews.

Ethics

ThenumberofpeoplelivingwithADisexpectedtoincreasedramaticallyoverthenextdecades.Mostofthisincreasewillbeinlow‐andmiddleincomecountries,wheretheprevalencealreadyisthehighest(12).Despitethis,issuespertainingtoequitywerenotaddressedinanyoftheincludedreviews.Themajorityoftheincludedstudieswerefromhighincomecountries,whileinsomecasesitwasnotclearwherethestudieshadbeenconducted.

54 Discussion

Introducingandimplementingnewinterventions/treatmentscanbecostly,andpayingforanewtreatmentforonediseasegroupmaytakeawaymoneyfromothertreatmentalternatives,orfromotherpatientgroups.Ensuringthataninterventioniscost‐effec‐tive,andthatthebeneficialeffectsexceedsharms,isthereforeveryimportanttopre‐ventwasteofhealthcareresources(99).Noneoftheincludedreviewsreportedordis‐cussedthelackofdataoncostsandcost‐effectivenessoftheinterventionsevaluatedintheoriginalstudies.Therearehugefinancialinterestsinvolvedindruginterventionstopreventcognitivedecline,AD,anddementia.Ithasbeensuggestedthatresearchthatisfundedbydrugcompaniesismorelikelytoyieldpositiveoutcomesthanresearchwithothertypeoffunding.Thisbiascanbeintroducedinanumberofways(100).Itisthereforeim‐portanttodisclosethefundingsourcesforincludedstudies,andpossiblerelationshipsbetweenthoseconductingtheresearchandthepharmaceuticalcompanies(101).Onlyoneofthereviews(78)reportedthefinancialdisclosuresoftheoriginalstudies,whichallweresupportedbydrugcompanies.Whendesigningatrial,itisimportanttoensurethatoutcomesthatareofimportancetothepatientsaremeasured(102).Qualityoflifeisoneexampleofanimportantpa‐tient‐centredoutcome,whichwasonlyreportedintwooftheincludedreviews(72,73),functioninactivitiesofdailylivingisanother.Noneofthereviewsreportedonpa‐tientinvolvementinthetrialdevelopment,ordiscussedthelackofit.Involvingpa‐tientsinrefininganinterventionmayleadtoclinicaltrialendpointsthatbettercomplywiththeneedsandconcernsofpatientsandcaregivers(103).

Potentialbiasesintheoverviewprocess

Atleasttwopeopleindependentlyappliedeligibilitycriteriaandassessedthereviewsforinclusion.Oneauthorextracteddataintoastandardiseddataextractionsform,andanotherauthorcheckedtheaccuracyoftheextracteddata,whichshouldreducebiasintheoverviewprocess.Atleasttwoauthorsassessedthescientificqualityofreviewsac‐cordingtothetoolusedbytheKnowledgeCentre.However,excludingreviewsofmod‐eratetolowqualitymayhaveinducedbias.

Overallcompletenessandapplicabilityoftheevidence

OnlyfouroftheincludedreviewsreportedonAD(42)orincidenceofothertypesofde‐mentia(72,73,78),whileallreviewsreportedoncognitivetestscores.Itcanbeques‐tionedifcognitivetestscoresisagoodproxyforprogressiontodementia.Qualityoflifewasonlyreportedintworeviews(72,73),andADLfunctioninonereviewonly(72).Adverseeffectswerereportedinfourreviews(42,72,73,78).Eventhoughthepharmacologicalinterventions(antihypertensivedrugsandstatins)evaluatedintwooftheincludedreviews(72,73),showedbeneficialeffectsonthetar‐getedriskfactorsfordementia(i.e.bloodpressureandcholesterollevelrespectively),

55 Discussion

incidenceofdementiaandcognitivefunctiondidnotchangesignificantly.Theconfi‐denceintervals(forincidenceofdementia)wererelativelywideinbothstudies,whytheresultsshouldbeinterpretedwithcaution.McGuinnessandcolleagues(73),aimedtoincludealsonon‐pharmacologicalbloodpressureloweringinterventions(e.g.saltrestriction,weightreduction,exercise,re‐ducedalcoholintake),butcouldnotidentifyanysuchstudiesforinclusion.Someoftheincludedtrialssufferedfromlargelossestofollowup,andalargeproportionofcontrolparticipantsreceivingnon‐studyantihypertensivedrugs.Anotherlimitationwiththereviewonantihypertensivedrugswasthatonlyasingletest(MMSE)wasusedtoas‐sesstheeffectsoncognitivefunction.Alimitationwiththestatinreview(72)wastheunclearmethodsusedtodiagnosede‐mentiainsomeoftheincludedstudies,andthatonlypeopledeemedtobeatmoderatetohighriskofacardiovasculardiseasewereincluded.Alimitationwithboththereviewonantihypertensivedrugsandthereviewonstatins,isthatincidenceofdementiawasasecondaryoutcomeinallincludedstudiesInthereviewthatevaluatedtheeffectofaerobicexerciseoncognitivefunction(85),notalloftheincludedstudiesreportedimprovedaerobicfitnessasaresultoftheinter‐vention,andnonereportedabeneficialeffectoncognition.Inamajorityofstudies,thefrequency/durationofthecomparatorinterventions,werenotmatchedtothoseintheinterventiongroup.Inaddition,neithercognitivefunctionnoraerobicfitnessatbase‐linewerereported,whichmakesitdifficulttofullyappreciatetheresults.Itmaybedoubtedwhetherthefollowuptimewassufficientforchangetotakeplace(from8to26weeks),andwhetherthestudiesweresufficientlypowered(median49partici‐pants).Itmayalsobequestionedwhetherthetargetedpopulationwasoptimalforthistypeofinterventiontoshowaneffect,asitincludedveryoldpeopleupto92yearsofage.Itispossiblethatinterventionsaddressinglifestylerelatedriskfactorslikephysi‐calinactivity,aremorelikelytobeeffectiveinconservingcognitivefunctioniftheyaredeliveredearlierinlife.Thereisofcoursethentheproblemwithverylonginterventiontimes,astheoptimalageforoutcomeassessmentislaterinlife(104).Theonereview(65),whichsummarisedtheeffectsofCCToncognitivetestscoresfromanumberofsmallstudies(mediann=44participants),didnotreportanylong‐termef‐fectsonperformanceoncognitivetests.Inaddition,thereviewauthorsdidnotprovideanyinformationonwhattypeofcognitiveteststhathadbeenusedintheincludedstudies,nordidtheyprovideanydetailsontheactivecomparisoninterventionortheadherencetotheintervention,whichmakeinterpretationoftheresultsdifficult.Wedidnotfindanyhighqualityreviewsconcernedwiththeeffectsofhealthylifestylechangesoncognitivefunctionanddementiaincidence(apartfromthereviewonaero‐bicexercise),e.g.conversiontoahealthydiet,decreasedalcoholintake,smokingcessa‐tion,weightlossetc.Threeexcludedreviews(includingmostlycohortandcross‐sec‐tionalstudies)reportsomeevidenceofaneffectofincreasedfruitandvegetableintakeand/oradherencetoaMediterraneandiet(69,70,75).Wealsodidnotfindanyhigh

56 Discussion

qualityreviewsconcernedwiththeeffectsofinterventionstargetingotherriskfactorsfordementialikeforexamplemid‐lifedepression,lackofsocialinteraction,orlowedu‐cationalattainment.Oneexcludedreviewoflowquality(105)reportedoninterven‐tionstotreatdepressioninpeoplewithMCIanddementia,andtwoexcludedreviews(79,80)ofmoderatequalitywereconcernedwithcognitiveandphysicalleisureactivi‐ties.Ithasbeensuggestedthatloweducationalattainmentistheriskfactorcontrib‐utingmosttotheburdenofdementiaworldwide,andthatphysicalinactivityistheriskfactorthatcontributesthemosttothediseaseburdeninUSAandEurope(17).Thereisthereforeanurgentneedtoevaluateeffectiveinterventionstargetingthese,andotherriskfactors,inRCTsandlargescalepopulationbasedstudies,withlongfollowuptimes.Dementiasaresyndromesthatmostlikelyareofamultifactorialorigin(16).Itispossi‐blethatpreventiveinterventionsmaybemoreeffectiveiftheytakeintoaccountthemultifacetedaetiologybehindthedisease,i.e.interventionsthattargetmultipleriskfactorsmaybemoreeffective.However,allofthereviewsincludedinthisoverviewofreviews,evaluatedtheeffectivenessofasingleitemintervention,targetingonlyoneriskfactor.WeareawareoffourlargeEuropeantrialsofsocalledmultidimensionalin‐terventions(includingforexamplediet,exercise,cognitivetraining,andvascularriskmonitoring)forpeopleatriskfordementia.Threeofthesetrialsareongoing(106).Re‐sultsfromoneofthetrials,recentlypublished(107),suggestbeneficialeffectsofthemultidimensionalinterventiononcognitivefunction.Untilfairlyrecently(23,26)therehasbeenlittleornoconsensusonwhatcriteriaandtoolstouseforthediagnosisofMCI,ADandothertypesofdementia.Intheoriginalstudiesincludedinthereviewsnowunderconsideration,ofwhich54.6percentwerepublishedbeforethepublicationoftheconsensuscriteriaand45.4percentafter,differ‐entcriteriawereusedtodeterminetheonsetofdementiaandtodiagnoseMCI.Inaddi‐tion,morethan396differentcognitivetestswereusedtoassesstheeffectsofinterven‐tionsoncognitivefunction.Mostcommonlythetestswerenotwelldescribed,i.e.infor‐mationaboutthescaling,thedesireddirectionofeffect,andwhatwouldbeconsideredaclinicallyimportanteffectwasnotprovided.InonereviewallstudiesusedonlytheMiniMentalStateExamination(MMSE)toassesscognitivefunction,whileotherstudiesuseduptotendifferenttestsortestbatteriestoassesscognition.Achangein2pointsontheMMSEscalehasbeensuggestedtobeaclinicallysignificanteffect(108).ThechangeintheMMSErangedfrom0.07lowerto0.42higheracrosssixoftheincludedreviews.Resultsfromarecentsystematicreview(109)donotsupporttheuseoftheMMSEasasingleadministeredstand‐alonetesttoassessprogressiontodementia.Thereviewauthorsinsteadsuggesttouseasetoftestsadministeredovertime.Tofacili‐tatefutureevaluationsofthecomparativeeffectivenessofinterventions,researchersshouldattempttocometoaconsensusonwhattestbatteriestouseforassessmentofcognitivefunction,andtousepublishedconsensuscriteriaandmethodsfordiagnosingMCIanddementia.

57 Discussion

Agreementsordisagreementswithotheroverviewsofreviews

Weidentifiedoneoverviewofreviewsofinterventionstopreventcognitivedeclineanddementia(110).Thisoverviewwaspublishedin2010andwithaliteraturesearchfrom2009,andthusitisbecomingoutdated.Incontrasttoouroverviewofreviewsithadwiderinclusioncriteria,andalsolookedatfactorsassociatedwithriskreduction,andnotonlyeffectsofinterventionstodelayorpreventcognitivedeclineanddemen‐tia.Insomerespecttheireligibilitycriteriawerestricter,astheyincludedonlyoriginalstudieswithafollowupofatleast6monthsduration,whileweappliednorestrictionstothelengthoffollowup.Ourresultsofnon‐significanteffectsofvitaminE,antihyper‐tensivedrugs,andcholinesteraseinhibitorsondementiaincidence,ascomparedtopla‐cebo,areinagreementwiththeresultsreportedinthisotheroverviewofreviews.ItdidnotidentifyanystudiesevaluatingtheeffectivenessofOmega‐3FAsforinclusion,thereforecomparisonwithourresultcouldnotbemade.

Applicationsforpractice

Physicians,patients,andcarersshouldbemadeawareofthelackofconvincingevi‐denceforaneffectofOmega‐3FAs,antihypertensivedrugs,andstatinsondementiaincidenceandcognitivefunction,whendeliveredtocognitivelyhealthyolderpeople.Thisofcoursehavenobearingontheeffectsofantihypertensivedrugsandstatinsonotherconditions(e.g.cardiovasculardisease).ThepractitionershouldalsobemadeawareoftheevidenceoflittleornoeffectofOmega‐3FAs,vitaminE,andcholinester‐aseinhibitorsinpreventingcognitivedecline,ADorothertypesofdementia,inpeoplewithmildcognitiveimpairment.ThereissomeevidenceforasmalleffectofCCTontheperformanceincognitivetestsincognitivelyhealthypeople,butfurtherresearchofthelongtermeffectsisneeded.

Needforfurtherresearch

Furtherresearchisneededtoidentifyeffectiveinterventionstopreventcognitivede‐clineanddementia,whilewaitingforresearcherstodiscoveracurativetreatmentfordementia.Toincreasethecertaintyoftheresultsofthisreview,trialswithmorepre‐ciseeffectestimatesandnarrowerCIsareneeded.Sincesofaronlyeffectsofsinglefac‐etedinterventionshavebeenevaluatedandsummarisedinreviews,interventionsin‐volvingacombinationofhealthylifestylechanges,e.g.conversiontoahealthydiet,de‐creasedalcoholconsumption,smokingcessation,weightloss,cognitivetrainingetc.,areneeded.Alsostudiesevaluatingtheeffectivenessofinterventionstargetingotherriskfactorsfordementia,likeforexampleinterventionsforearlyidentificationandtreatmentofmid‐lifedepression,forincreasedsocialinteraction,andinterventionstoencourageeducationinpopulationswithloweducationalattainment.Physicalinactiv‐ityandloweducationalattainmentshouldbeaddressedinlargescalestudies,sincetheseriskfactorsareestimatedtocontributemosttotheburdenofdisease.

58 Discussion

ResearchersshouldaimtouseconsensusmethodsandcriteriatodiagnoseMCI,ADanddementia,andforassessingcognitivefunction,toenableeffectcomparisonsbetweenstudies.Theyshouldalsoaimtoimprovethereporting(describethetests/scalesusedtoassesscognitivefunction,givedetailsontheinterventionandthecomparatorinter‐vention),soastoprovidesufficientdetailtoenablereplication.Preferablyallstudiesshouldreportincidenceofdementia(asaprimaryoutcome),andnotonlyperformanceoncognitivetests.

59 Conclusion

Conclusion

Giventheevidencefromtheincludedsystematicreviews,thereisatpresentstillsomeuncertaintyabouttheeffectivenessoftheevaluatedinterventionsforpreventionofcognitivedeclineanddementia.Onlysingleiteminterventionshavesofarbeenevalu‐ated.However,sincetheaetiologybehinddementiamostlikelyismultifactorial,inter‐ventionsaddressingmorethanonemodifiableriskfactormaybeneeded.Thereisanurgentneedtoevaluateinterventionstargetingimportantriskfactors,usingRCTsandlargescalepopulationbasedstudies,withlongfollowuptimes.

60

References

1. WHO.Dementia.APublicHealthPriority.http://wwwwhoint/mental_health/publications/dementia_report_2012/en/2012:1‐103.

2. PlassmanBL,WilliamsJW,Jr.,BurkeJR,HolsingerT,BenjaminS.Systematic

review:factorsassociatedwithriskforandpossiblepreventionofcognitivedeclineinlaterlife.AnnInternMed2010;153(3):182‐193.

3. InternationalAsD.WorldAlzheimer'sReport.TheGlobalVoiceonDementiaLondon,UK:Alzheimer'sDiseaseInternational;2014 4. BallardC,GauthierS,CorbettA,BrayneC,AarslandD,JonesE.Alzheimer's

disease.TheLancet2011;377(9770):1019‐1031.5. O'BrienJT,ThomasA.Vasculardementia.TheLancet2015;386(10004):1698‐

1706.6. ClinicM.vascularDementia.http://www.mayoclinic.org/diseases‐

conditions/vascular‐dementia/basics/causes/con‐20029330.[Accessed03.05.16].Availablefrom.

7. WalkerZ,PossinKL,BoeveBF,AarslandD.Lewybodydementias.TheLancet

2015;386(10004):1683‐1697.8. BangJ,SpinaS,MillerBL.Frontotemporaldementia.TheLancet

2015;386(10004):1672‐1682.9. SchneiderJA,ArvanitakisZ,BangW,BennettDA.Mixedbrainpathologiesaccount

formostdementiacasesincommunity‐dwellingolderpersons.Neurology2007;69(24):2197‐2204.

10. ClausJJ,StaekenborgSS,RoordaJJ,StevensM,HerderscheeD,van

MaarschalkerweerdW,etal.LowPrevalenceofMixedDementiainaCohortof2,000ElderlyPatientsinaMemoryClinicSetting.JAlzheimersDis2016;50(3):797‐806.

11. OrganisationWH.Dementia.

http://www.who.int/mediacentre/factsheets/fs362/en/.[Accessed03.05.16].Availablefrom.

12. PrinceM,BryceR,AlbaneseE,WimoA,RibeiroW,FerriCP.Theglobalprevalence

ofdementia:asystematicreviewandmetaanalysis.Alzheimer's&Dementia2013;9(1):63‐75.e62.

61

13. http://www.alzheimer‐europe.org/Policy‐in‐Practice2/Country‐

comparisons/The‐prevalence‐of‐dementia‐in‐Europe/Norway.Norway.TheprevalenceofAlzheimer'sinEurope.Alzheimer'sEurope.[Accessed03.05.16].Availablefrom.

14. SelbaekGKØ,EngedalK.Theprevalenceofpsychiatricsymptomsand

behaviouraldisturbancesandtheuseofpsychotropicdrugsinNorwegiannursinghomes.IntJGeriatrPsychiatry2007;22(9):843‐849.

15. WergelandJN,SelbækG,HøgsetLD,SöderhamnU,KirkevoldØ.Dementia,

neuropsychiatricsymptoms,andtheuseofpsychotropicdrugsamongolderpeoplewhoreceivedomiciliarycare:across‐sectionalstudy.IntPsychogeriatr2014;26(03):383‐391.

16. InternationalAs.Demantiaandriskreduction.Ananalysisofprotectiveand

modifiablefactors.WorlsAlzheimer'sReport20142014.17. NortonS,MatthewsFE,BarnesDE,YaffeK,BrayneC.Potentialforprimary

preventionofAlzheimer'sdisease:ananalysisofpopulation‐baseddata.TheLancetNeurology2014;13(8):788‐794.

18. PurnellC,GaoS,CallahanCM,HendrieHC.Cardiovascularriskfactorsand

incidentAlzheimerdisease:Asystematicreviewoftheliterature.AlzheimerDisAssocDisord2009;23(1):1‐10.

19. MorettiL,CristoforiI,WeaverSM,ChauA,PortelliJN,GrafmanJ.Cognitivedecline

inolderadultswithahistoryoftraumaticbraininjury.TheLancetNeurology2012;11(12):1103‐1112.

20. YaffeK.Traumaticbraininjuryandneuropsychiatricoutcomes.TheLancet

Neurology2012;11(12):1020‐1021.21. McDadeEMPR.MildCognitiveImpairment:Epidemiology,pathologyandclinical

assessment.UpToDate,http://www.uptodate.com/contents/mild‐cognitive‐impairment‐epidemiology‐pathology‐and‐clinical‐assessment.[Accessed03.05.16].Availablefrom.

22. RobertsR,KnopmanDS.ClassificationandEpidemiologyofMCI.ClinGeriatrMed

2013;29(4):10.1016/j.cger.2013.1007.1003.23. AlbertMS,DeKoskyST,DicksonD,DuboisB,FeldmanHH,FoxNC,etal.The

diagnosisofmildcognitiveimpairmentduetoAlzheimer’sdisease:RecommendationsfromtheNationalInstituteonAging‐Alzheimer’sAssociationworkgroupsondiagnosticguidelinesforAlzheimer’sdisease.Alzheimer's&dementia:thejournaloftheAlzheimer'sAssociation2011;7(3):270‐279.

24. LitvanI,GoldmanJG,TrosterAI,SchmandBA,WeintraubD,PetersenRC,etal.

DiagnosticcriteriaformildcognitiveimpairmentinParkinson'sdisease:MovementDisorderSocietyTaskForceguidelines.MovDisord2012;27(3):349‐356.

62

25. PedersenK,LarsenJ,TysnesO,AlvesG.Prognosisofmildcognitiveimpairmentinearlyparkinsondisease:Thenorwegianparkweststudy.JAMANeurology2013;70(5):580‐586.

26. McKhannGM,KnopmanDS,ChertkowH,HymanBT,JackCR,KawasCH,etal.The

diagnosisofdementiaduetoAlzheimer’sdisease:RecommendationsfromtheNationalInstituteonAging‐Alzheimer’sAssociationworkgroupsondiagnosticguidelinesforAlzheimer’sdisease.Alzheimer's&dementia:thejournaloftheAlzheimer'sAssociation2011;7(3):263‐269.

27. SperlingRA,AisenPS,BeckettLA,BennettDA,CraftS,FaganAM,etal.Toward

definingthepreclinicalstagesofAlzheimer'sdisease:RecommendationsfromtheNationalInstituteonAging‐Alzheimer'sAssociationworkgroupsondiagnosticguidelinesforAlzheimer'sdisease.Alzheimer's&dementia:thejournaloftheAlzheimer'sAssociation2011;7(3):280‐292.

28. McGuinnessB,CraigD,BullockR,PassmoreP.Statinsforthepreventionof

dementia.CochraneDatabaseofSystematicReviews2009(2):CD003160.29. HedayatiSS,ElsayedEF,ReillyRF.Non‐pharmacologicalaspectsofblood

pressuremanagement:whatarethedata?KidneyInt2011;79(10):1061‐1070.30. EbrahimS,TaylorFC,BrindleP.Statinsfortheprimarypreventionof

cardiovasculardisease.BMJ2014;348.31. TaylorF,HuffmanMarkD,MacedoAnaF,MooreTheresaHM,BurkeM,Davey

SmithG,etal.Statinsfortheprimarypreventionofcardiovasculardisease.CochraneDatabaseofSystematicReviews:JohnWiley&Sons,Ltd;2013.

http://onlinelibrary.wiley.com/doi/10.1002/14651858.CD004816.pub5/abstract

32. WongWB,LinVW,BoudreauD,DevineEB.Statinsinthepreventionofdementia

andAlzheimer'sdisease:ameta‐analysisofobservationalstudiesandanassessmentofconfounding.PharmacoepidemiolDrugSaf2013;22(4):345‐358.

33. GodleeF.Adverseeffectsofstatins.BMJ2014;348.34. MufsonEJ,CountsSE,PerezSE,GinsbergSD.Cholinergicsystemduringthe

progressionofAlzheimer'sdisease:therapeuticimplications.ExpertRevNeurother2008;8(11):1703‐1718.

35. PressDAM.Cholinesteraseinhibitorsinthetreatmentofdementia.UpToDate:

http://www.uptodate.com/contents/cholinesterase‐inhibitors‐in‐the‐treatment‐of‐dementia[Accessed03.05.16].Availablefrom.

36. GauthierS,ReisbergB,ZaudigM,PetersenRC,RitchieK,BroichK,etal.Mild

cognitiveimpairment.TheLancet;367(9518):1262‐1270.37. WillettWC,SacksF,TrichopoulouA,DrescherG,Ferro‐LuzziA,HelsingE,etal.

Mediterraneandietpyramid:aculturalmodelforhealthyeating.TheAmericanJournalofClinicalNutrition1995;61(6):1402S‐1406S.

38. KnightA,BryanJ,MurphyK.IstheMediterraneandietafeasibleapproachto

preservingcognitivefunctionandreducingriskofdementiaforolderadultsin

63

Westerncountries?Newinsightsandfuturedirections.AgeingResearchReviews2016;25:85‐101.

39. TomaszW,EwaS‐W,JolantaP,StanisławL,KatarzynaC,RobertB,etal.Omega‐3

FattyAcidsandtheirRoleinCentralNervousSystem‐AReview.CurrMedChem2016;23(8):816‐831.

40. SydenhamE,DangourAD,LimWS.Omega3fattyacidforthepreventionof

cognitivedeclineanddementia.CochraneDatabaseofSystematicReviews2012;6:CD005379.

41. MangialascheF,XuW,KivipeltoM,CostanziE,ErcolaniS,PigliautileM,etal.

Tocopherolsandtocotrienolsplasmalevelsareassociatedwithcognitiveimpairment.NeurobiolAging2012;33(10):2282‐2290.

42. FarinaN,IsaacMG,ClarkAR,RustedJ,TabetN.VitaminEforAlzheimer's

dementiaandmildcognitiveimpairment.CochraneDatabaseofSystematicReviews2012;11:CD002854.

43. HamerM,ChidaY.Physicalactivityandriskofneurodegenerativedisease:a

systematicreviewofprospectiveevidence.PsycholMed2009;39(01):3‐11.44. BruijnRFAG,SchrijversEMC,GrootKA,WittemanJCM,HofmanA,FrancoOH,et

al.Theassociationbetweenphysicalactivityanddementiainanelderlypopulation:theRotterdamStudy.EurJEpidemiol2013;28(3):277‐283.

45. PaillardT.Preventiveeffectsofregularphysicalexerciseagainstcognitivedecline

andtheriskofdementiawithageadvancement.SportsMedicine‐Open2015;1(1):20.

46. AmievaH,MokriH,LeGoffM,MeillonC,Jacqmin‐GaddaH,Foubert‐SamierA,et

al.Compensatorymechanismsinhigher‐educatedsubjectswithAlzheimer’sdisease:astudyof20yearsofcognitivedecline.Brain2014;137(4):1167‐1175.

47. WilliamsK,KemperS.ExploringInterventionstoReduceCognitiveDeclinein

Aging.JPsychosocNursMentHealthServ2010;48(5):42‐51.48. TuckerAM,SternY.COGNITIVERESERVEINAGING.CurrentAlzheimerresearch

2011;8(4):354‐360.49. FolkehelseinstituttetKi.Slikoppsummererviforskning.

http://kilden.kunnskapssenteret.no/h%C3%A5ndb%C3%B8ker/h%C3%A5ndbok‐slik‐oppsummerer‐vi‐forskning.[Accessed03.05.16].Availablefrom.

50. groupGw.GRADEtool:http://gradeworkinggroup.org/toolbox/index.htm.

GRADEworkinggroup.[Accessed03.05.16].Availablefrom.51. BirksJ,EvansJG.Ginkgobilobaforcognitiveimpairmentanddementia.Cochrane

DatabaseofSystematicReviews2009(1).52. BirksJ,McGuinnessB,CraigD.Rivastigmineforvascularcognitiveimpairment.

CochraneDatabaseofSystematicReviews:JohnWiley&Sons,Ltd;2013. http://onlinelibrary.wiley.com/doi/10.1002/14651858.CD004744.pub3/abstrac

t

64

53. BlondellSJ,Hammersley‐MatherR,VeermanJL.Doesphysicalactivitypreventcognitivedeclineanddementia?:Asystematicreviewandmeta‐analysisoflongitudinalstudies.BMCPublicHealth2014;14:510.

54. ChengPF,ChenJJ,ZhouXY,RenYF,HuangW,ZhouJJ,etal.Dosoyisoflavones

improvecognitivefunctioninpostmenopausalwomen?Ameta‐analysis.Menopause‐theJournaloftheNorthAmericanMenopauseSociety2015;22(2):198‐206.

55. CooperC,LiR,LyketsosC,LivingstonG.Treatmentformildcognitiveimpairment:

systematicreview.BrJPsychiatry2013;203(3):255‐264.56. DangourAD,WhitehousePJ,RaffertyK,MitchellSA,SmithL,HawkesworthS,et

al.B‐vitaminsandfattyacidsinthepreventionandtreatmentofAlzheimer'sdiseaseanddementia:asystematicreview.JAlzheimersDis2010;22(1):205‐224.

57. DinizBS,PintoJA,Jr.,GonzagaML,GuimaraesFM,GattazWF,ForlenzaOV.To

treatornottotreat?Ameta‐analysisoftheuseofcholinesteraseinhibitorsinmildcognitiveimpairmentfordelayingprogressiontoAlzheimer'sdisease.EurArchPsychiatryClinNeurosci2009;259(4):248‐256.

58. ForbesSC,Holroyd‐LeducJM,PoulinMJ,HoganDB.EffectofNutrients,Dietary

SupplementsandVitaminsonCognition:aSystematicReviewandMeta‐AnalysisofRandomizedControlledTrials.CanadianGeriatricsJournalCGJ2015;18(4):231‐245.

59. GardT,HolzelBK,LazarSW.Thepotentialeffectsofmeditationonage‐related

cognitivedecline:asystematicreview.AnnNYAcadSci2014;1307:89‐103.60. GatesNJ,SachdevPS,FiataroneSinghMA,ValenzuelaM.Cognitiveandmemory

traininginadultsatriskofdementia:asystematicreview.BMCGeriatr2011;11:55.

61. GiebelC,ChallisD.Translatingcognitiveandeverydayactivitydeficitsinto

cognitiveinterventionsinmilddementiaandmildcognitiveimpairment.IntJGeriatrPsychiatry2014.

62. HindinSB,ZelinskiEM.Extendedpracticeandaerobicexerciseinterventions

benefituntrainedcognitiveoutcomesinolderadults:ameta‐analysis.JAmGeriatrSoc2012;60(1):136‐141.

63. JiaoJ,LiQ,ChuJ,ZengW,YangM,ZhuS.Effectofn‐3PUFAsupplementationon

cognitivefunctionthroughoutthelifespanfrominfancytooldage:asystematicreviewandmeta‐analysisofrandomizedcontrolledtrials.AmJClinNutr2014;100(6):1422‐1436.

64. KuiperJS,ZuidersmaM,VoshaarRCO,ZuidemaSU,vandenHeuvelER,StolkRP,

etal.Socialrelationshipsandriskofdementia:Asystematicreviewandmeta‐analysisoflongitudinalcohortstudies.AgeingResearchReviews2015;22:39‐57.

65. LampitA,HallockH,ValenzuelaM.Computerizedcognitivetrainingincognitively

healthyolderadults:asystematicreviewandmeta‐analysisofeffectmodifiers.PLoSMedicine/PublicLibraryofScience2014;11(11):e1001756.

65

66. LawLL,BarnettF,YauMK,GrayMA.Effectsofcombinedcognitiveandexercise

interventionsoncognitioninolderadultswithandwithoutcognitiveimpairment:asystematicreview.AgeingResearchReviews2014;15C:61‐75.

67. LeungIH,WaltonCC,HallockH,LewisSJ,ValenzuelaM,LampitA.Cognitive

traininginParkinsondisease:Asystematicreviewandmeta‐analysis.Neurology2015;85(21):1843‐1851.

68. LiMM,YuJT,WangHF,JiangT,WangJ,MengXF,etal.EfficacyofvitaminsB

supplementationonmildcognitiveimpairmentandAlzheimer'sdisease:asystematicreviewandmeta‐analysis.CurrentAlzheimerResearch2014;11(9):844‐852.

69. LoefM,WalachH.Fruit,vegetablesandpreventionofcognitivedeclineor

dementia:asystematicreviewofcohortstudies.JournalofNutrition,Health&Aging2012;16(7):626‐630.

70. LouridaI,SoniM,Thompson‐CoonJ,PurandareN,LangIA,UkoumunneOC,etal.

Mediterraneandiet,cognitivefunction,anddementia:asystematicreview.Epidemiology2013;24(4):479‐489.

71. MazereeuwG,LanctotKL,ChauSA,SwardfagerW,HerrmannN.Effectsofomega‐

3fattyacidsoncognitiveperformance:ameta‐analysis.NeurobiolAging2012.72. McGuinnessB,CraigD,BullockR,PassmoreP.Statinsforthepreventionof

dementia.CochraneDatabaseSystRev2016;1:Cd003160.73. McGuinnessB,ToddS,PassmoreP,BullockR.Bloodpressureloweringinpatients

withoutpriorcerebrovasculardiseaseforpreventionofcognitiveimpairmentanddementia.CochraneDatabaseofSystematicReviews2009(4):CD004034.

74. MuangpaisanW,BrayneC,Alzheimer'sSocietyVascularDementiaSystematic

ReviewG.Systematicreviewofstatinsforthepreventionofvasculardementiaordementia.Geriatrics&gerontologyinternational2010;10(2):199‐208.

75. OpieRS,RalstonRA,WalkerKZ.AdherencetoaMediterranean‐styledietcanslow

therateofcognitivedeclineanddecreasetheriskofdementia:asystematicreview.NutritionandDietetics2013;70(3):206‐217.

76. ReganB,VaranelliL.Adjustment,depression,andanxietyinmildcognitive

impairmentandearlydementia:asystematicreviewofpsychologicalinterventionstudies.IntPsychogeriatr2013;25(12):1963‐1984.

77. RouchL,CestacP,HanonO,CoolC,HelmerC,BouhanickB,etal.Antihypertensive

drugs,preventionofcognitivedeclineanddementia:asystematicreviewofobservationalstudies,randomizedcontrolledtrialsandmeta‐analyses,withdiscussionofpotentialmechanisms.CNSDrugs2015;29(2):113‐130.

78. RussTomC,MorlingJoanneR.Cholinesteraseinhibitorsformildcognitive

impairment.CochraneDatabaseofSystematicReviews:JohnWiley&Sons,Ltd;2012.

http://onlinelibrary.wiley.com/doi/10.1002/14651858.CD009132.pub2/abstract

66

http://onlinelibrary.wiley.com/store/10.1002/14651858.CD009132.pub2/asset/CD009132.pdf?v=1&t=i2rjjnmi&s=dcf78350aedf488bb38458f88ed1c2f56a2583ae

79. SternC,MunnZ.Cognitiveleisureactivitiesandtheirroleinpreventingdementia:

asystematicreview.InternationalJournalofEvidence‐BasedHealthcare2010;8(1):2‐17.

80. SternC,KonnoR.Physicalleisureactivitiesandtheirroleinpreventingdementia:

asystematicreview.InternationalJournalofEvidence‐BasedHealthcare2009;7(4):270‐282.

81. SwigerKJ,ManalacRJ,BlumenthalRS,BlahaMJ,MartinSS.Statinsandcognition:

asystematicreviewandmeta‐analysisofshort‐andlong‐termcognitiveeffects.MayoClinProc2013;88(11):1213‐1221.

82. TsengCN,GauBS,LouMF.Theeffectivenessofexerciseonimprovingcognitive

functioninolderpeople:asystematicreview.JNursRes2011;19(2):119‐131.83. ValenzuelaM,SachdevP.Cancognitivetrainingpreventtheonsetofdementia?A

systematicreviewofrandomizedclinicaltrialswithlongitudinalfollowup.Alzheimer'sandDementia2009;1):157‐158.

84. WilliamsKN,KemperS.Interventionstoreducecognitivedeclineinaging.J

PsychosocNursMentHealthServ2010;48(5):42‐51.85. YoungJ,AngevarenM,RustedJ,TabetN.Aerobicexercisetoimprovecognitive

functioninolderpeoplewithoutknowncognitiveimpairment.CochraneDatabaseofSystematicReviews2015;4:CD005381.

86. ZhengGH,LiuFW,LiSZ,HuangMM,TaoJ,ChenLD.TaiChiandtheProtectionof

CognitiveAbilityASystematicReviewofProspectiveStudiesinHealthyAdults.AmJPrevMed2015;49(1):89‐97.

87. MartinM,ClareL,AltgassenAM,CameronMH,ZehnderF.Cognition‐based

interventionsforhealthyolderpeopleandpeoplewithmildcognitiveimpairment.CochraneDatabaseofSystematicReviews2011(1).

88. MoherD,LiberatiA,TetzlaffJ,AltmanDG.Preferredreportingitemsfor

systematicreviewsandmeta‐analyses:ThePRISMAstatement.InternationalJournalofSurgery2010;8(5):336‐341.

89. Cochrane.Cochraneriskofbiastool.

http://ohg.cochrane.org/sites/ohg.cochrane.org/files/uploads/Risk%20of%20bias%20assessment%20tool.pdf.[Accessed03.05.16].Availablefrom.

90. BoutronI,MoherD,TugwellP,GiraudeauB,PoiraudeauS,NizardR,etal.A

checklisttoevaluateareportofanonpharmacologicaltrial(CLEARNPT)wasdevelopedusingconsensus.JClinEpidemiol2005;58(12):1233‐1240.

91. deMortonNA.ThePEDroscaleisavalidmeasureofthemethodologicalqualityof

clinicaltrials:ademographicstudy.AustJPhysiother2009;55(2):129‐133.92. HughesCP,BergL,DanzigerWL,CobenLA,MartinRL.Anewclinicalscaleforthe

stagingofdementia.TheBritishJournalofPsychiatry1982;140(6):566‐572.

67

93. FischerP,JellingerK,GattererG,DanielczykW.ProspectiveneuropathologicalvalidationofHachinski'sIschaemicScoreindementias.JournalofNeurology,Neurosurgery,andPsychiatry1991;54(7):580‐583.

94. TombaughTN,McIntyreNJ.TheMini‐MentalStateExamination:A

ComprehensiveReview.JAmGeriatrSoc1992;40(9):922‐935.95. SpitzerRLWJ,GibbonM,FirstMB.ThestructuredclinicalinterviewforDSM‐III‐R.

I:History,rationaleanddescription.ArchGenPsychiatry1992;49:624‐629.96. OrganisationWH.InternationalClassificationofDiseases(ICD).

http://www.who.int/classifications/icd/en/.[Accessed03.05.16].Availablefrom.

97. ClassroomAPCsVP.Diagnosticandstatisticsmanualofmentaldisorders,fourth

edition(DSM‐IV).http://allpsych.com/disorders/dsm/.[Accessed03.05.16].Availablefrom.

98. FarinaN,RustedJ,TabetN.Theeffectofexerciseinterventionsoncognitive

outcomeinAlzheimer'sdisease:asystematicreview.IntPsychogeriatr2014;26(1):9‐18.

99. KnappM.Thecost‐effectivenesschallenge:isitworthit?AlzheimersResTher

2015;7(1):1‐3.100. LexchinJ.ThoseWhoHavetheGoldMaketheEvidence:HowthePharmaceutical

IndustryBiasestheOutcomesofClinicalTrialsofMedications.ScienceandEngineeringEthics2011;18(2):247‐261.

101. RosemanM,TurnerEH,LexchinJ,CoyneJC,BeroLA,ThombsBD.Reportingof

conflictsofinterestfromdrugtrialsinCochranereviews:crosssectionalstudy.BMJ2012;345.

102. GuyattGH,OxmanAD,KunzR,AtkinsD,BrozekJ,VistG,etal.GRADEguidelines:

2.Framingthequestionanddecidingonimportantoutcomes.JClinEpidemiol2011;64(4):395‐400.

103. HolmKE,CasaburiR,CerretaS,GussinHA,HusbandsJ,PorszaszJ,etal.Patient

InvolvementintheDesignofaPatient‐CenteredClinicalTrialtoPromoteAdherencetoSupplementalOxygenTherapyinCOPD.ThePatient‐Patient‐CenteredOutcomesResearch2015:1‐9.

104. RichardE,AndrieuS,SolomonA,MangialascheF,AhtiluotoS,vanCharanteEPM,

etal.Methodologicalchallengesindesigningdementiapreventiontrials&#x2014;TheEuropeanDementiaPreventionInitiative(EDPI).JNeurolSci;322(1):64‐70.

105. ReganB,VaranelliL.Asystematicreviewofpsychologicalinterventionstudiesin

MildCognitiveImpairmentandearlydementia:impactonadjustment,depressionandanxiety.AustralasianJournalonAgeing2014;33:60‐60.

106. MangialascheF,KivipeltoM,SolomonA,FratiglioniL.Dementiaprevention:

currentepidemiologicalevidenceandfutureperspective.AlzheimersResTher2012;4(1):6‐6.

68

107. NganduT,LehtisaloJ,SolomonA,LevälahtiE,AhtiluotoS,AntikainenR,etal.A2yearmultidomaininterventionofdiet,exercise,cognitivetraining,andvascularriskmonitoringversuscontroltopreventcognitivedeclineinat‐riskelderlypeople(FINGER):arandomisedcontrolledtrial.TheLancet;385(9984):2255‐2263.

108. LopezMN,CharterRA,MostafaviB,NibutLP,SmithWE.PsychometricProperties

oftheFolsteinMini‐MentalStateExamination.Assessment2005;12(2):137‐144.

109. Arevalo‐RodriguezI,PedrazaOL,RodriguezA,SanchezE,GichI,SolaI,etal.

Alzheimer'sdiseasedementiaguidelinesfordiagnostictesting:asystematicreview.AmJAlzheimersDisOtherDemen2013;28(2):111‐119.

110. WilliamsJW,PlassmanBL,BurkeJ,BenjaminS.PreventingAlzheimer'sdisease

andcognitivedecline.EvidenceReport/TechnologyAssessment2010(193):1‐727.

111. AarslandD,SardahaeeFS,AnderssenS,BallardC,Alzheimer'sSocietySystematic

Reviewg.Isphysicalactivityapotentialpreventivefactorforvasculardementia?Asystematicreview.Aging&MentalHealth2010;14(4):386‐395.

112. BarnesD.RiskfactorreductionandAlzheimer'sdiseaseprevalence:Projected

effectandpracticalimplications.Alzheimer'sandDementia2012;1):P605.113. EtgenT,SanderD,BickelH,SanderK,ForstlH.VitaminDdeficiency,cognitive

impairmentanddementia:asystematicreviewandmeta‐analysis.DementGeriatrCognDisord2012;33(5):297‐305.

114. FaucounauV,WuYH,BoulayM,DeRotrouJ,RigaudAS.Cognitiveintervention

programmesonpatientsaffectedbyMildCognitiveImpairment:apromisinginterventiontoolforMCI?JournalofNutrition,Health&Aging2010;14(1):31‐35.

115. FotuhiM,MohasselP,YaffeK.Fishconsumption,long‐chainomega‐3fattyacids

andriskofcognitivedeclineorAlzheimerdisease:acomplexassociation.NatureClinicalPracticeNeurology2009;5(3):140‐152.

116. Franco‐MartinM,Parra‐VidalesE,Gonzalez‐PalauF,Bernate‐NavarroM,SolisA.

[Theinfluenceofphysicalexerciseinthepreventionofcognitivedeteriorationintheelderly:asystematicreview].RevNeurol2013;56(11):545‐554.

117. GizachewS,MakonnenE.TheroleofstatinsinAlzheimer'sdisease:Ameta‐

analysis.AfricanJournalofNeurologicalSciences2012;31(2).118. KimK,HanJW,KimYJ,ParkJH,LeeSB,LeeJJ,etal.Cognitivestimulationasa

therapeuticmodalityforearlydementiaandmildcognitiveimpairment:Ameta‐analysisstudy.Alzheimer'sandDementia2014;10:P602.

119. KueiderAM,ParisiJM,GrossAL,RebokGW.Computerizedcognitivetrainingwith

olderadults:asystematicreview.PLoSOne2012;7(7).120. LigthartSA,vanCharanteEPM,vanGoolWA,RichardE.Treatmentof

cardiovascularriskfactorstopreventcognitivedeclineanddementia:Asystematicreview.VascularHealthandRiskManagement2010;6(1):775‐785.

69

121. Lopez‐LeonS,Rivero‐FerrerE.Statinsforpreventionofalzheimer'sdisease:

Systematicreviewandmeta‐analysis.PharmacoepidemiolDrugSaf2013;22:58‐59.

122. MaidmentI,ChanWY,BuaN,MyintP,CampbellN,BoustaniM,etal.Anupdated

systematicreviewoftheimpactofmedicationswithanticholinergicactivityoncognitivefunctionandmortality.Alzheimer'sandDementia2013;1):P774.

123. MauerS,VergneD,GhaemiSN.Standardandtrace‐doselithium:asystematic

reviewofdementiapreventionandotherbehavioralbenefits.AustNZJPsychiatry2014;48(9):809‐818.

124. NaqviR,LibermanD,RosenbergJ,AlstonJ,StrausS.Preventingcognitivedecline

inhealthyolderadults(Structuredabstract).CMAJ2013.p.881‐885. http://onlinelibrary.wiley.com/o/cochrane/cldare/articles/DARE‐

12013037786/frame.html125. OoiCheowP,LokeSengC,YassinZ,HamidT‐A.Carbohydratesforimprovingthe

cognitiveperformanceofindependent‐livingolderadultswithnormalcognitionormildcognitiveimpairment.CochraneDatabaseofSystematicReviews:JohnWiley&Sons,Ltd;2011.

http://onlinelibrary.wiley.com/doi/10.1002/14651858.CD007220.pub2/abstract

126. OttBR,DaielloLA,DahabrehIJ,SpringateBA,BixbyK,MuraliM,etal.DoStatins

ImpairCognition?ASystematicReviewandMeta‐AnalysisofRandomizedControlledTrials.JGenInternMed2015;30(3):348‐358.

127. RuizAragonJ,VillegasPorteroR.VaccineforthepreventionofAlzheimer's

disease(Structuredabstract).HealthTechnologyAssessmentDatabase:AndalusianAgencyforHealthTechnologyAssessment(AETSA);2010.

http://onlinelibrary.wiley.com/o/cochrane/clhta/articles/HTA‐32010001469/frame.html

128. SantosC,CostaJ,SantosJ,Vaz‐CarneiroA,LunetN.Caffeineintakeanddementia:

Systematicreviewandmeta‐analysis.JAlzheimersDis2010;20(SUPPL.1):S187‐S204.

129. SimonSS,YokomizoJE,BottinoCMC.CognitiveinterventioninamnesticMild

CognitiveImpairment:Asystematicreview.NeurosciBiobehavRev2012;36(4):1163‐1178.

130. SofiF,MacchiC,AbbateR,GensiniGF,CasiniA.Effectivenessofthe

Mediterraneandiet:canithelpdelayorpreventAlzheimer'sdisease?JAlzheimersDis2010;20(3):795‐801.

131. ValenzuelaM,SachdevP.Cancognitiveexercisepreventtheonsetofdementia?

Systematicreviewofrandomizedclinicaltrialswithlongitudinalfollow‐up.AmJGeriatrPsychiatry2009;17(3):179‐187.

132. vanderSchaftJ,KoekHL,DijkstraE,VerhaarHJ,vanderSchouwYT,Emmelot‐

VonkMH.TheassociationbetweenvitaminDandcognition:asystematicreview.AgeingResearchReviews2013;12(4):1013‐1023.

70

133. WaldDS,KasturiratneA,SimmondsM.Effectoffolicacid,withorwithoutotherBvitamins,oncognitivedecline:meta‐analysisofrandomizedtrials.AmJMed2010;123(6):522‐527.e522.

134. WeihM,DegirmenciU,KreilS,KornhuberJ.Physicalactivityandalzheimer's

disease:Ameta‐analysisofcohortstudies.GeroPsych:TheJournalofGerontopsychologyandGeriatricPsychiatry2010;23(1):17‐20.

135. YangM,XuDD,ZhangY,LiuX,HoevenR,ChoWC.Asystematicreviewonnatural

medicinesforthepreventionandtreatmentofAlzheimer'sdiseasewithmeta‐analysesofinterventioneffectofginkgo.AmJChinMed2014;42(3):505‐521.

136. AbrahamRajeshP,DentonDavidA,Al‐AssafAalyaS,RutjesAnneWS,ChongLee

Y,MalikMuzaffarA,etal.Vitaminandmineralsupplementationforpreventionofdementiaordelayingcognitivedeclineinpeoplewithmildcognitiveimpairment.CochraneDatabaseofSystematicReviews2015(10).

137. ForbesScottC,ForbesD,ForbesS,BlakeCatherineM,ChongLeeY,Thiessen

EmilyJ,etal.Exerciseinterventionsformaintainingcognitivefunctionincognitivelyhealthypeopleinlatelife.CochraneDatabaseofSystematicReviews2015(5).

138. ForbesScottC,ForbesD,ForbesS,BlakeCatherineM,ChongLeeY,Thiessen

EmilyJ,etal.Exerciseinterventionsforpreventingdementiaordelayingcognitivedeclineinpeoplewithmildcognitiveimpairment.CochraneDatabaseofSystematicReviews2015(5).

139. HarrisonStephanieL,BirdiR,SmartChrisO,BrittainK,RutjesAnneWS,SiervoM,

etal.Dietaryinterventionsformaintainingcognitivefunctionincognitivelyhealthypeopleinmidlife.CochraneDatabaseofSystematicReviews2015(10).

140. JordanF,McGuinnessB,MurphyK,PassmoreP,KellyJohnP,DevaneD,etal.

Aspirinandanti‐inflammatorydrugsforthepreventionofdementiaDohigherprimarycarepracticeperformancescorespredictlowerratesofemergency

admissionsforpersonswithseriousmentalillness?Ananalysisofsecondarypaneldata.CochraneDatabaseofSystematicReviews:Reviews2015;Issue1(16).

141. KrauseD,RoupasP.Dietaryinterventionsasaneuroprotectivetherapyforthe

delayoftheonsetofcognitivedeclineinolderadults:anumbrellareviewprotocol.JBIDatabaseofSystematicReviews&ImplementationReport2015;13(2):74‐8310p.

142. SiervoMLJ,MunroA,TangEYH,RutjesAWS,StephanB..Dietaryinterventions

formaintainingcognitivefunctionincognitivelyhealthypeopleinlatelife.CochraneData‐baseofSystematicReviews2015(Issue10.).

143. TangEugeneYeeH,HarrisonStephanieL,AlbaneseE,GormanThomasJ,Rutjes

AnneWS,SiervoM,etal.Dietaryinterventionsforpreventionofdementiainpeoplewithmildcognitiveimpairment.CochraneDatabaseofSystematicReviews2015(10).

71

Appendices

1Glossary

Term Explanation

AD Alzheimer'sdisease(AD)isaprogressive,degenerativedisor‐derthatattacksthebrain'snervecells,orneurons,resultinginlossofmemory,thinkingandlanguageskills,andbehaviouralchanges.

ADAS‐Cog Alzheimer'sDiseaseAssessmentScale‐Cognitivesubscale(ADAS‐cog).

ADCS‐ADL ActivitiesofDailyLivingInventory.Aninventoryofinformantbaseditemstoassessactivitiesofdailylivingandinstrumentalactivitiesofdailyliving,i.e.functionalperformance,ofAlz‐heimer’sdisease(AD).

ADL Activitiesofdailyliving(ADL)scale.

Aerobictraining Aerobictraininginphysicalexerciseoflowtohighinten‐sitythatdependsprimarilyontheaerobicenergy‐generatingprocessi.e.thatthereissufficientamountofoxygentomeetthebody’senergydemands.

Alpha‐tocopherol α‐TocopherolisaformofvitaminEthatispreferentiallyab‐sorbedandaccumulatedinhumans.

AmnesticMCI MildCognitiveImpairmentthatprimarilyaffectsmemory.

Bias Abiasisasystematicerror,ordeviationfromthetruth,inre‐sultsorinferences.Biasescanoperateineitherdirection:differ‐entbiasescanleadtounderestimationoroverestimationofthetrueinterventioneffect.Biasescanvaryinmagnitude:somearesmall(andtrivialcomparedwiththeobservedeffect)andsomearesubstantial(sothatanapparentfindingmaybeentirelyduetobias).Evenaparticularsourceofbiasmayvaryindirection:biasduetoaparticulardesignflaw(e.g.lackofallocationcon‐cealment)mayleadtounderestimationofaneffectinonestudybutoverestimationinanotherstudy.Itisusuallyimpossibleto

72

knowtowhatextentbiaseshaveaffectedtheresultsofapartic‐ularstudy,althoughthereisgoodempiricalevidencethatpar‐ticularflawsinthedesign,conductandanalysisofrandomizedclinicaltrialsleadtobias.

CDR‐SumofBoxes TheClinicalDementiaRatingScaleSumofBoxes.

CI ConfidenceInterval

CIND Cognitiveimpairmentnodementia(CIND)describesindividualswhosecognitivefunctioningfallsbelownormalbutwhodonotmeetdementiacriteria.AcriterialessrestrictivethanMCI,asitdoesnotexcludepeoplebasedontheaetiologyoftheircogni‐tiveimpairment.

Cognitivefunction Anintellectualprocessbywhichonebecomesawareof,per‐ceives,orcomprehendsideas.Itinvolvesallaspectsofpercep‐tion,thinking,reasoning,andremembering.

Delayedrecall Theabilitytoremembersomethingafteraperiodofrestordis‐tractionranginganywherefromminutestodays.

Dementia Dementiaisalossofmentalabilitysevereenoughtointer‐ferewithnormalactivitiesofdailyliving,last‐ingmorethansixmonths,notpresentsincebirth,andnotasso‐ciatedwithalossoralterationofconsciousness.Therearedif‐ferenttypesofdementia,ofwhichthemostcommonlyoccur‐ringisAlzheimer’sdisease.

Dietarysupplement Adietarysupplementprovidenutrientsthatmayotherwisenotbeconsumedinsufficientquantitiesandincludevitamins,min‐erals,fiber,fattyacids,oraminoacids,amongothersubstances.

Digitspanforwardandbackward

Digit‐spantaskisusedtomeasureworkingmemory’snumberstoragecapacity.Participantsarepresentedwithaseriesofdig‐its(e.g.,'8,2,4')andmustimmediatelyrepeatthemback.Thelengthofthelongestlistapersoncanrememberisthatperson'sdigitspan.Whiletheparticipantisaskedtoenterthedigitsinthegivenorderintheforwarddigit‐spantask,inthebackwarddigit‐spantasktheparticipantneedstoreversetheorderofthenumbers.

Executivefunction Cognitivecontrolandsupervisoryattentionalsystem.Anum‐brellatermforthemanagement(regulation,control)ofcogni‐tiveprocesses,includingworkingmemory,reasoning,flexibil‐ity,andproblemsolvingaswellasplanningandexecution.

Frontotemporalde‐mentia

Frontotemporallobardegenerationisanumbrellatermforadi‐versegroupofuncommondisordersthatprimarilyaffectthe

73

frontalandtemporallobesofthebraini.e.theareasgenerallyassociatedwithpersonality,behaviourandlanguage.

GRADE GradingofRecommendations,Assessment,DevelopmentandEvaluation.Atoolthatisusedtoassessthecertaintyoftheevi‐denceinasystematicreview.

Hedge’sg Hedges'gpoolvariancesontheassumptionofequalpopulationvariances,usingn‐1foreachsampleinsteadofn(asinCohen’sd),whichprovidesabetterestimate,especiallythesmallerthesamplesizes.Providesthedifferencebetweenthetwomeans.

Heterogeneity Anykindofvariabilityamongstudiesinasystematicreviewmaybetermedheterogeneity.Variabilityintheparticipants,in‐terventionsandoutcomesstudiedmaybedescribedasclinicalheterogeneity,andvariabilityinstudydesignandriskofbiasmaybedescribedasmethodologicalheterogeneity.Variabilityintheinterventioneffectsbeingevaluatedinthedifferentstud‐iesisknownasstatisticalheterogeneity,andisaconsequenceofclinicalormethodologicaldiversity,orboth,amongthestud‐ies.Statisticalheterogeneitymanifestsitselfintheobservedin‐terventioneffectsbeingmoredifferentfromeachotherthanonewouldexpectduetorandomerror(chance)alone.

HR Hazardratio.Theinstantaneoushazardrateisthelimitofthenumberofeventsperunittimedividedbythenumberatrisk,asthetimeintervalapproaches0.

Hypertension Hypertension,alsoknownashighbloodpressureorarterialhy‐pertension,isachronicmedicalconditioninwhichthebloodpressureinthearteriesispersistentlyelevated.

Immediaterecall Theabilitytoremembereventsoccurringwithinthepastfewminutes.

Imprecision Imprecisioningeneral,iswhenstudiesincluderelativelyfewparticipants,andfewevents,andthereforehavewideconfi‐denceintervalsaroundtheestimateofeffect.

Indirectness Indirectnessofevidenceiswhenevidencecomesfromresearchthateitherdoesnotdirectlycomparetheinterventionsofinter‐estwithcontrol,orwhentheinterventionisnotappliedtothepopulationsofinterestorifastudymeasuresoutcomesthatarenotdirectmeasuresimportanttopatientsbutproxymeasuresorprocessmeasures.

Inconsistency Inconsistencyofrelative(ratherthanabsolute)treatmentef‐fectsinbinary/dichotomousoutcomesmaybedeterminedby

74

lookingatthe(dis)similarityofpointestimates,extentofover‐lapofconfidenceintervals,andstatisticalcriteriaincludingtestsofheterogeneity(I2).

Letterdigitsubstitu‐tiontest

Digitsymbolsubstitutiontest(DSST)isaneuropsychological‐testsensitivetobraindamage,dementia,ageanddepression.Itconsistsof(e.g.nine)digit‐symbolpairs(e.g.1/‐,2/┴...7/Λ,8/X,9/=)followedbyalistofdigits.Undereachdigitthesubjectshouldwritedownthecorrespondingsymbolasfastaspossible.Thenumberofcorrectsymbolswithintheallowedtime(e.g.90or120sec)ismeasured.

Levybodydementias LBDisanumbrellatermfortworelateddiagnoses.LBDreferstobothParkinson’sdiseasedementiaanddementiawithLewybodies.Theearliestsymptomsofthesetwodiseasesdiffer,butreflectthesameunderlyingbiologicalchangesinthebrain.Overtime,peoplewithbothdiagnoseswilldevelopverysimilarcog‐nitive,physical,sleep,andbehavioralsymptoms.

MCI Mildcognitiveimpairment.MCIcausescognitivechangesthatareseriousenoughtobenoticedbytheindividualsexperienc‐ingthemortootherpeople,butthechangesarenotsevereenoughtointerferewithdailylifeorindependentfunction.

MMSE TheMiniMentalStateExamination(MMSE)isthemostcom‐monlyusedtestforcomplaintsofproblemswithmemoryorothermentalabilities.Itcanbeusedbyclinicianstohelpdiag‐nosedementiaandtohelpassessitsprogressionandseverity.Itconsistsofaseriesofquestionsandtests,eachofwhichscorespointsifansweredcorrectly.TheMMSEtestsanumberofdif‐ferentmentalabilities,includingaperson'smemory,attentionandlanguage.

Omega‐3FattyAcids Omega‐3fattyacids(FAs)arepoly‐saturatedfattyac‐ids(PUFAs)withadouble‐bond(C=C)atthethirdcarbonatomfromtheendofthecarbonchain.Omega‐3fattyacidsaeim‐portantfornormalmetabolism.

OR Anoddsratio(OR)isameasureofassociationbetweenanexpo‐sureandanoutcome.Itrepresentstheoddsthatanoutcomewilloccurgivenaparticularexposure,comparedtotheoddsoftheoutcomeoccurringintheabsenceofthatexposure(i.e.inthecontrolgroup).

Primaryprevention Primarypreventionseekstopreventtheonsetofspecificdis‐easesviariskreduction:byalteringbehavioursorexposuresthatcanleadtodisease,orbyenhancingresistancetotheef‐fectsofexposuretoadiseaseagent.

75

Processingspeed Thepaceatwhichapersontakeininformation,makesenseofitandbegintorespond.

RR Riskratioorrelativerisk.Relativeriskisthera‐tiooftheriskofdiseaseamongthoseexposedtoariskfac‐tortotheriskamongthosenotexposed.

SMD Thestandardizedmeandifferenceisusedasasummarystatis‐ticinmeta‐analysiswhenthestudiesallassessthesameout‐comebutmeasureitinavarietyofways(forexample,allstud‐iesmeasuredepressionbuttheyusedifferentpsychometricscales).Inthiscircumstanceitisnecessarytostandardizetheresultsofthestudiestoauniformscalebeforetheycanbecom‐bined.Thestandardizedmeandifferenceexpressesthesizeoftheinterventioneffectineachstudyrelativetothevariabilityobservedinthatstudy.

Secondaryprevention Secondarypreventionincludesproceduresthatdetectandtreatpre‐clinicalpathologicalchangesandtherebycontroldiseaseprogression.

SD Thestandarddeviation(SD)isameasureusedtoquantifytheamountofvariationofasetofdatavalues.Ifcloseto‘0’itindicatesthatthedatapointstendtobeveryclosetothemeanofthedataset,whileahighstandarddeviationindicatesthatthedatapointsarespreadoutoverawiderrangeofvalues.

SDMT TheSymbolDigitModalitiestest(SDMT)involvesasimplesubsti‐tutiontask.Usingareferencekey,theexamineehas90secondstopairspecificnumberswithgivengeometricfigures.

Statins Statins(orHMG‐CoAreductaseinhibitors)areaclassofcholes‐terolloweringdrugsthatinhibittheenzymeHMG‐CoAreduc‐tasewhichplaysacentralroleintheproductionofcholesterol.

Strooptest TheStroopeffectisthefindingthatnamingthecolourofthefirstsetofwordsiseasierandquickerthanthesecond.Inpsychology,theStroopeffectisademonstrationofinterferenceinthereactiontimeofatask.

TICS‐m TheModifiedTelephoneInterviewforCognitiveStatus.

VAD Vasculardementiaiscausedbyreducedbloodsupplytothebrainduetodiseasedbloodvessels.Itisthesecondmostcom‐montypeofdementiaafterAD.ThereareseveraldifferenttypesofVADthatdifferinthecauseofthedamageandthepartofthebrainthatisaffected.ThedifferenttypesofVADhavesomesymptomsincommonandsomesymptomsthatdiffer.Theirsymptomstendtoprogressindifferentways.

76

Verbalfluencytest Verbalfluencytestsareakindofcognitivetestinwhichpartici‐pantsareaskedtosayasmanywordsaspossiblefromacate‐gory(e.g.animals,fruitsorwordsthatbeginwithaspecificlet‐ter)inagiventime(usually60seconds).

Workingmemory Thepartofshort‐termmemorywhichisconcernedwithimme‐diateconsciousperceptualandlinguisticprocessing.

77

2Searchstrategy

OvidMEDLINE(R)In‐Process&OtherNon‐IndexedCitations,OvidMEDLINE(R)Daily,OvidMEDLINE(R)

# Searches Results

1 expDementia/ 128933

2 LewyBodies/ 1591

3 alzheim$.mp. 111287

4 binswanger$.mp. 555

5 cadasil.mp. 996

6 cerad.tw. 501

7 dement$.mp. 98538

8 (ftldorftd$).tw. 3680

9 ((fronto?temporalorcortico?basalorfrontotemporalorcorticobasalorfrontallobe)adj5(degenerat$4ordysfunction$)).tw.

4266

10 (kluveradj5(bucyorbusy)).mp. 249

11 ((lew$2adj5bod$3)ordlbd).mp. 7187

12 (lobaradj5atroph$3adj5(brainorcerebr$2)).tw. 17

13 (mesulamadj5syndrome$).tw. 3

14 (pick$2adj5(disease$1orcomplex)).mp. 3987

15 posteriorcortic$atroph$.tw. 236

16 ((primaryorprogressive)adj5aphasi$).tw. 1311

17 sdat.tw. 648

18 sivd.tw. 118

19 ((subcortic$3orsub?cortic$3)adj5(encephalopath$3orleukoencephalo‐path$3)).tw.

988

20 (dement$oralzheim$).jw. 11541

21 (mesulamorhuntington*).tw. 13555

22 mesulamm.au. 86

23 amentia.tw. 72

24 or/1‐23[demens] 201139

25 Cognition/ 70348

26 expCognitionDisorders/ 68645

27 expMemoryDisorders/ 24068

28 ((cognit$ormemor$ormental$)adj5(disabil$ordisabl$ordeclin$ordefect$orimpair$orlos$ordeteriorat$)).tw.

93040

29 (memoryadj(deficitordisorder$)).tw. 3200

30 ((cognit$orbehavio?r$)adj5symptom$).tw. 20537

78

31 (cognit$adj2(abnormal$ordisorder$)).tw. 5387

32 (mci$1orcind$1).tw. 12363

33 or/25‐32[mildkognitivsvikt] 227065

34 PrimaryPrevention/orSecondaryPrevention/ 16433

35 pc.fs. 1077475

36 prevent*.ti,ab. 1016676

37 (earlyadj3(intervention*ortherap*ortreatment*)).ti,ab. 80321

38 or/34‐37[forebygging] 1855764

39 (24or33)and38 34224

40 systematicreview.kw. 1529

41 (((systematic*orliteratureorintegrative)adj3(overvieworreview*orsearch*))ormeta‐analys*).ti,ab.

312951

42 meta‐analysis.pt. 54768

43 or/40‐42 324678

44 39and43 1198

45 limit44toyr="2009‐Current" 687Embase1974to2014November20

# Searches Results

1 expdementia/ 234963

2 expprimaryprogressiveaphasia/ 1783

3 Binswangerencephalopathy/ 411

4 corticobasaldegeneration/ 1655

5 Lewybody/ 5429

6 alzheim$.mp. 152189

7 binswanger$.mp. 845

8 cadasil.mp. 1640

9 cerad.tw. 861

10 cerad.tw. 861

11 dement$.mp. 127933

12 (ftldorftd$).tw. 5250

13 ((fronto?temporalorcortico?basalorfrontotemporalorcorticobasalorfrontallobe)adj5(degenerat$4ordysfunction$)).tw.

5337

14 (kluveradj5(bucyorbusy)).mp. 311

15 ((lew$2adj5bod$3)ordlbd).mp. 10861

16 (lobaradj5atroph$3adj5(brainorcerebr$2)).tw. 24

17 (mesulamadj5syndrome$).tw. 2

18 (pick$2adj5(disease$1orcomplex)).mp. 5211

19 posteriorcortic$atroph$.tw. 348

79

20 ((primaryorprogressive)adj5aphasi$).tw. 1799

21 sdat.tw. 745

22 sivd.tw. 128

23 ((subcortic$3orsub?cortic$3)adj5(encephalopath$3orleukoencephalo‐path$3)).tw.

1202

24 (dement$oralzheim$).jw. 23003

25 (mesulamorhuntington*).tw. 15654

26 mesulamm.au. 85

27 amentia.tw. 81

28 or/1‐27[demens] 280278

29 cognition/ 161153

30 cognitivedefect/ 99835

31 expmemorydisorder/ 52627

32 ((cognit$ormemor$ormental$)adj5(disabil$ordisable$ordeclin$ordefect$orimpair$orlos$ordeteriorat$)).tw.

119465

33 ((cognit$orbehavio?r$)adj5symptom$).tw. 27113

34 (memoryadj(deficitordisorder$)).tw. 4021

35 (cognit$adj2(abnormal$ordisorder$)).tw. 7696

36 (mci$1orcind$1).tw. 19492

37 or/29‐36[mildkognitivsvikt] 341074

38 primaryprevention/ 28363

39 prophylaxis/ 71282

40 secondaryprevention/ 17234

41 pc.fs. 977503

42 prevent*.ti,ab. 1207917

43 (earlyadj3(intervention*ortherap*ortreatment*)).ti,ab. 107582

44 or/38‐43[forebygging] 2005544

45 (28or37)and44 51079

46 systematicreview.kw. 8906

47 (((systematic*orliteratureorintegrative)adj3(overvieworreview*orsearch*))ormeta‐analys*).ti,ab.

363376

48 “systematicreview”/ 81429

49 metaanalysis/ 84604

50 or/46‐49[SR] 404522

51 45and50 1960

52 limit51toyr="2009‐Current" 1161

53 limit52toembase 1157PsycINFO1806toNovemberWeek32014

# Searches Results

80

1 expdementia/ 54112

2 alzheimer'sdisease/ 32737

3 corticobasaldegeneration/ 230

4 kluverbucysyndrome/ 52

5 alzheim$.mp. 43337

6 binswanger$.mp. 425

7 cadasil.mp. 216

8 cerad.tw. 300

9 dement$.mp. 51846

10 (ftldorftd$).tw. 1988

11 ((fronto?temporalorcortico?basalorfrontotemporalorcorticobasalorfrontallobe)adj5(degenerat$4ordysfunction$)).tw.

2368

12 (kluveradj5(bucyorbusy)).mp. 127

13 ((lew$2adj5bod$3)ordlbd).mp. 2583

14 (lobaradj5atroph$3adj5(brainorcerebr$2)).tw. 5

15 (mesulamadj5syndrome$).tw. 2

16 (pick$2adj5(disease$1orcomplex)).mp. 651

17 posteriorcortic$atroph$.tw. 162

18 ((primaryorprogressive)adj5aphasi$).tw. 915

19 sdat.tw. 325

20 sivd.tw. 55

21 ((subcortic$3orsub?cortic$3)adj5(encephalopath$3orleukoencephalo‐path$3)).tw.

231

22 (dement$oralzheim$).jw. 9968

23 (mesulamorhuntington*).tw. 3665

24 mesulamm.au. 3

25 amentia.tw. 194

26 or/1‐25[demens] 79799

27 cognition/ 21672

28 cognitiveimpairment/ 23622

29 expmemorydisorders/ 8217

30 ((cognit$ormemor$ormental$)adj5(disabil$ordisable$ordeclin$ordefect$orimpair$orlos$ordeteriorat$)).ti,ab.

65797

31 ((cognit$orbehavio?r$)adj5symptom$).ti,ab. 20101

32 (memoryadj(deficitordisorder$)).tw. 2843

33 (cognit$adj2(abnormal$ordisorder$)).ti,ab. 4073

34 (mci$1orcind$1).ti,ab. 4157

35 or/27‐34[mildkognitivsvikt] 118087

36 Prevention/ 21979

37 preventivemedicine/ 1763

81

38 prevent*.ti,ab. 150008

39 earlyintervention/ 8704

40 (earlyadj3(intervention*ortherap*ortreatment*)).ti,ab. 18624

41 or/36‐40[forebygging] 169892

42 (26or35)and41 9685

43 "literaturereview"/ 21991

44 metaanalysis/ 3473

45 (((systematic*orliteratureorintegrative)adj3(overvieworreview*orsearch*))ormeta‐analys*).ti,ab.

80629

46 or/43‐45[SR] 96129

47 42and46 429

48 limit47toyr="2009‐Current" 236CinahlS48

S41ANDS45 Limiters‐ExcludeMEDLINErecords;PublishedDate:20100101‐20141231

2

S47

S41ANDS45 Limiters‐PublishedDate:20090101‐20141231 14

S46

S41ANDS45 16

S45

S42ANDS43ANDS44 3,411

S44

TI((((systematic*orliteratureorintegrative)N3(review*orovervieworsearch*))ormeta‐analys*))ORAB((((systematic*orliteratureorintegra‐tive)N3(review*orovervieworsearch*))ormeta‐analys*))

66,155

S43

(MH"MetaAnalysis") 15,067

S42

(MH"SystematicReview") 19,946

S41

S34ANDS40 12,594

S40

S35ORS36ORS37ORS38ORS39 462,399

S3

(MH"PreventiveHealthCare+") 139,635

S38

(MH"EarlyIntervention") 5,908

S37

TI((earlyN3(intervention*ortherap*ortreatment*)))ORAB((earlyN3(in‐tervention*ortherap*ortreatment*)))

14,242

S36

TIprevent*ORABprevent* 136,404

S35

MWpcORMJpc 283,361

82

S34

S24ORS33 90,993

S33

S25ORS26ORS27ORS28ORS29ORS30ORS31ORS32 57,280

S32

TI(mci*orcind*)ORAB(mci*orcind*) 1,848

S31

TI((cognit*N2(abnormal*ordisorder*)))ORAB((cognit*N2(abnormal*ordisorder*)))

876

S30

TI(((cognit*orbehavio?r*)N5symptom*))ORAB(((cognit*orbehavio?r*)N5symptom*))

2,665

S29

TI(((cognit*ormemor*ormental*)N5(disabil*ordisable*ordeclin*orde‐fect*orimpair*orlos*ordeteriorat*))or(memoryW0(deficitordisorder*)))ORAB(((cognit*ormemor*ormental*)N5(disabil*ordisabel*ordeclin*ordefect*orimpair*orlos*ordeteriorat*))or(memoryW0(deficitordisor‐der*)))

16,642

S28

(MH"Memory+") 12,749

S27

(MH"MemoryDisorders+") Searchmodes‐Boolean/Phrase Interface‐EBSCOhostResearchDatabases

3,484

S26

(MH"CognitionDisorders+") 13,114

S25

(MH"Cognition+") 24,111

S24

S1ORS2ORS3ORS4ORS5ORS6ORS7ORS8ORS9ORS10ORS11ORS12ORS13ORS14ORS15ORS16ORS17ORS18ORS19ORS20ORS21ORS22ORS23

43,984

S23

TIamentiaORABamentia 1

S22

AUmesulam 40

S21

TI((mesulamorhuntington*))ORAB((mesulamorhuntington*)) 719

S20

SOdement*oralzheim* 5,022

S19

TI(((subcortic*orsub?cortic*)N5(encephalopath*orleukoencephalopath*)))ORAB(((subcortic*orsub?cortic*)N5(encephalopath*orleukoencephalo‐path*)))

102

S18

TIsivdORABsivd 19

S17

TIsdatORABsdat 17

S16

TI(((primaryorprogressive)N5aphasi*))ORAB(((primaryorprogressive)N5aphasi*))

331

83

S15

TIposteriorcortic*atroph*ORABposteriorcortic*atroph* 51

S14

TI((pick*N5(diseasesorcomplex)))ORAB((pick*N5(diseasesorcom‐plex)))

10

S13

TI(mesulamN5syndrome*)ORAB(mesulamN5syndrome*) 0

S12

TI((lobarN5atroph*N5(brainorcerebr*)))ORAB((lobarN5atroph*N5(brainorcerebr*)))

1

S11

TI(((lew*N5(bodyorbodies))ordlbd))ORAB(((lew*N5(bodyorbodies))ordlbd))

615

S10

TI((kluverN5(bucyorbusy)))ORAB((kluverN5(bucyorbusy))) 6

S9

TI(((frontotemporalorcorticobasalorfrontotemporalorcorticobasalorfrontallobe)N5(degenerat*ordysfunction*)))ORAB(((frontotemporalorcorticobasalorfrontotemporalorcorticobasalorfrontallobe)N5(degenerat*ordysfunction*)))

398

S8 TI(ftldorftd*)ORAB(ftldorftd*) 380S7 TIdement*ORABdement* 21,186S6 TIceradORABcerad 78S5 TIcadasilORABcadasil 120S4 TIbinswanger*ORABbinswanger* 26S3 TIalzheim*ORABalzheim* 12,136S2 (MH"Aphasia+") 3,300S1 (MH"Dementia+") 35,097CDSR#1 MeSHdescriptor:[Dementia]explodealltrees 3863#2 MeSHdescriptor:[LewyBodies]thistermonly 6#3 alzheim*:ti,ab,kw 5076#4 binswanger*:ti,ab,kw 6#5 cadasil:ti,ab,kw 15#6 cerad:ti,ab,kw 16#7 dement*:ti,ab,kw 5327#8 ftldorftd*:ti,ab,kw 53#9 ((frontotemporalorcorticobasalorfrontotemporalorcorticobasalorfrontal

lobe)near/5(degenerat*ordysfunction$*)):ti,ab,kw50

#10 (kluvernear/5(bucyorbusy)):ti,ab,kw 1#11 ((lew*near/5bod*)ordlbd):ti,ab,kw 128#12 (lobarnear/5atroph*near/5(brainorcerebr*)):ti,ab,kw 0#13 mesulamnear/5syndrome*:ti,ab,kw 0#14 pick*near/5(disease*orcomplex):ti,ab,kw 38#15 posteriornextcortic*nextatroph*:ti,ab,kw 1#16 ((primaryorprogressive)near/5aphasi*):ti,ab,kw 11

84

#17 sdat:ti,ab,kw 42#18 sivd:ti,ab,kw 5#19 (subcortic*near/5(encephalopath*orleukoencephalopath*)):ti,ab,kw 15#20 mesulamorhuntington*:ti,ab,kw 184#21 amentia:ti,ab,kw 1#22 #1or#2or#3or#4or#5or#6or#7or#8or#9or#10or#11or#12or

#13or#14or#15or#16or#17or#18or#19or#20or#218991

#23 MeSHdescriptor:[Cognition]thistermonly 4956#24 MeSHdescriptor:[CognitionDisorders]explodealltrees 2823#25 MeSHdescriptor:[MemoryDisorders]explodealltrees 909#26 ((cognit*ormemor*ormental*)near/5(disabil*ordisable*ordeclin*orde‐

fect*orimpair*orlos*ordeteriorat*))or(memorynext(deficitordisor‐der*)):ti,ab,kw

7836

#27 ((cognit*orbehavio*)near/5symptom*):ti,ab,kw 2359#28 (cognit*near/2(abnormal*ordisorder*)):ti,ab,kw 3315#29 mci*orcind*:ti,ab,kw 2920#30 #23or#24or#25or#26or#27or#28or#29 20294#31 MeSHdescriptor:[PrimaryPrevention]explodealltrees 3506#32 MeSHdescriptor:[SecondaryPrevention]explodealltrees 257#33 AnyMeSHdescriptorwithqualifier(s):[Prevention&control‐PC] 77444#34 prevent*:ti,ab,kw 73266#35 (earlynear/3(intervention*ortherap*ortreatment*)):ti,ab,kw 8571#36 #31or#32or#33or#34or#35 131591#37 (#22or#30)and#36PublicationYearfrom2009to2014,inCochraneRe‐

views(ReviewsandProtocols),OtherReviewsandTechnologyAssessments410

CRD

Line Searchfor Hits

1 MeSHDESCRIPTORDementiaEXPLODEALLTREES 612

2 MeSHDESCRIPTORLewyBodiesEXPLODEALLTREES 0

3 (alzheim*) 523

4 (binswanger*) 4

5 (cadasil) 1

6 (cerad) 6

7 (dement*) 825

8 (ftldorftd*) 4

9 ((frontotemporalorcorticobasalorfrontotemporalorcorticobasalorfrontallobe)near5(degenerat*ordysfunction$*))

4

10 (kluvernear5(bucyorbusy)) 0

11 ((lew*near5bod*)ordlbd) 19

12 (lobarnear5atroph*NEAR5(brainorcerebr*)) 0

13 (mesulamnear5syndrome*) 0

85

Line Searchfor Hits

14 (pick*near5(disease*orcomplex)) 5

15 (posteriornextcortic*nextatroph*) 0

16 ((primaryorprogressive)near5aphasi*) 0

17 (sdat) 2

18 (sivd) 2

19 (subcortic*near5(encephalopath*orleukoencephalopath*)) 1

20 (mesulamorhuntington*) 24

21 (amentia) 0

22 #1OR#2OR#3OR#4OR#5OR#6OR#7OR#8OR#9OR#10OR#11OR#12OR#13OR#14OR#15OR#16OR#17OR#18OR#19OR#20OR#21

1083

23 MeSHDESCRIPTORCognitionEXPLODEALLTREES 221

24 MeSHDESCRIPTORCognitionDisordersEXPLODEALLTREES 254

25 MeSHDESCRIPTORMemoryDisordersEXPLODEALLTREES 36

26 ((cognit*ormemor*ormental*)near5(disabil*irdisable*ordeclin*ordefect*orim‐pair*orlos*ordeteriorat*))or((memorynear0(deficitordisorder*))

1392

27 ((cognit*orbehavio*)near5symptom*) 171

28 (cognit*near2(abnormal*ordisorder*)) 297

29 (mci*orcind*) 327

30 #23OR#24OR#25OR#26OR#27OR#28OR#29 2122

31 MeSHDESCRIPTORPrimaryPreventionEXPLODEALLTREES 812

32 MeSHDESCRIPTORSecondaryPreventionEXPLODEALLTREES 89

33 (prevent*) 16401

34 (earlynear3(intervention*ortherap*ortreatment*)) 0

35 #31OR#32OR#33OR#34 16515

36 #22OR#30 2756

37 #35AND#36 649

38 *INDARE,HTAFROM2009TO2014 34131

39 #37AND#38 186

Pubmed241114Search((((((prevent*OR(earlyAND(intervention*ORtherap*ORtreatment*)))))AND((((sys‐tematic*ORintegrativeORliterature)AND(review*ORoverview*ORsearch*))ORmeta‐analys*ORmedlineORpubmedORembase)))AND(((alzheim*ORdement*OR"lewybody"OR"lewybodies"ORbinswanger*ORcadasilORcerdaORftldORftd*OR((frontotemporalORcor‐ticobasal)AND(degenerat*ORdysfunction*))OR(kluverAND(bucyORbusy))ORdlbdOR(lo‐barANDatroph*AND(brainORcerebral*))ORmesulam*OR(pick*AND(disease*ORcom‐plex))OR(posteriorANDcortic*ANDatroph*)ORaphasi*ORsdatORsivdOR(subcortic*AND(encephalopath*ORleukoencephalopath*))ORhuntington*ORamentia))OR((((cognit*or

86

memor*ormental*)and(disable*ordisabil*ordeclin*ordefect*orimpair*orlos*ordeterio‐rat*))OR(memorydisorder*ormemorydeficit)or((cognit*orbehavio*)ANDsymptom*)OR(cognit*AND(abnormal*ordisorder*))ORmci*ORcind*)))))ANDpublisher[sb]

SearchHistory:WebofScienceTMCoreCollection

Set

Re‐sults

EditSets

CombineSets

AND OR

DeleteSets

#7

1,686 #6AND#5Indexes=SCI‐EXPANDED,SSCITimespan=2009‐2014

Edit Selecttocom‐

binesets.

Selecttode‐letethisset.

#6170,236

TOPIC:(((systematic*ORintegrativeORlitera‐ture)AND(review*ORoverview*ORsearch*)))Indexes=SCI‐EXPANDED,SSCITimespan=2009‐2014

Edit Selecttocom‐

binesets.

Selecttode‐letethisset.

#526,796 #4AND#3

Indexes=SCI‐EXPANDED,SSCITimespan=2009‐2014Edit Selecttocom‐

binesets.

Selecttode‐letethisset.

#4509,052

TOPIC:((prevent*OR(earlyAND(interven‐tion*ORtherap*ORtreatment*))))Indexes=SCI‐EXPANDED,SSCITimespan=2009‐2014

Edit Selecttocom‐

binesets.

Selecttode‐letethisset.

#3213,833

#2OR#1Indexes=SCI‐EXPANDED,SSCITimespan=2009‐2014

Edit Selecttocom‐

binesets.

Selecttode‐letethisset.

#2151,467

TOPIC:((((cognit*ormemor*ormental*)and(disabil*ordisable*ordeclin*ordefect*orim‐pair*orlos*ordeteriorat*))OR(memoryand(disorder*ordeficit))or((cognit*orbehavio*)ANDsymptom*)OR(cognit*AND(abnormal*ordisorder*))ORmci*ORcind*))Indexes=SCI‐EXPANDED,SSCITimespan=2009‐2014

Edit Selecttocom‐

binesets.

Selecttode‐letethisset.

#194,246 TOPIC:((alzheim*ORdement*or"lewybody"

OR"lewybodies"ORbinswanger*ORcadasilORceradORftldORftd*or((frontotemporalORcorticobasal)AND(degenerat*ORdysfunc‐tion*))OR(kluvyAND(bucyORbusy))ORdlbdOR(lobarANDatroph*AND(brainORcerebral*))ORmesulam*OR(pick*AND(dis‐ease*ORcomplex))OR(posteriorANDcortic*ANDatroph*)ORaphasi*ORsdatORsivdOR

Edit Selecttocom‐

binesets.

Selecttode‐letethisset.

87

(subcortic*AND(encephalopath*orleu‐koencephalopath*))ORhuntington*ORamen‐tia))Indexes=SCI‐EXPANDED,SSCITimespan=2009‐2014

AND O

R

88

3Excludedreviewsandreasonsforexclusion

Table18.Excludedreviewsandreasonsforexclusion.

Firstauthor(refno.)

Reasonforexclusion

Aarsland2010(111) Notasystematicreview(noassessmentofstudies’methodologicalquality)

Barnes2012(112) Notapopulationwithoutdementiadiagnosis

Blondell2014(53) Notasystematicreviewofhighmethodologicalquality(moderate)

Cheng2015(54) Notasystematicreviewofhighquality(moderate)

Cooper2013(55) Notasystematicreviewofhighmethodologicalquality(moderate)

Dangour2010(56) Notasystematicreviewofhighmethodologicalquality(moderate)

Etgen2012(113) Notasystematicreview(noassessmentofstudies’methodologicalquality)

Faucounau2010(114) Notasystematicreview(noassessmentofstudies’methodologicalquality)

Forbes2015(58) Notasystematicreviewofhighmethodologicalquality(lowtomoderatequality).Alsocognitivelyhealthypeopleandcognitivelyimpairedpeoplearenotanalysedseparately.

Fotuhi2009(115) Notasystematicreview(noassessmentofstudies’methodologicalquality)

Franco‐Martin2013(116)

Notasystematicreview(noassessmentofstudies’methodologicalquality)

Gizachew2012(117) Notasystematicreview(noassessmentofstudies’methodologicalquality)

Jiao2014(63) Mixedgroups.Datafrompeoplewithmildtomoderatecognitiveimpairment,werenotseparatelyanalysed.

Kim2014(118) Thereisnofulldescriptionofthesystematicreview(abstractonly)

Kueider2012(119) Notasystematicreview(noassessmentofstudies’methodologicalquality)

Kuiper2015(64) Notaninterventionreview

Law2014(66) Mixedgroups.Datafrompeoplewithcognitiveimpairmentanddementiawerenotseparatelyanalysed

Leung2014(67) Notapopulationwithoutdementiadiagnosis

Li2014(68) Notasystematicreviewofhighmthodologicalquality(moderate).

Ligthart2010(120) Notasystematicreview(noassessmentofstudies’methodologicalquality)

Loef2012(69) Notasystematicreviewofhighmethodologicalquality(moderate)

Lopez‐Leon2013(121)

Thereisnofulldescriptionofthesystematicreview(abstractonly)

89

Lourida2013(70) Notaninterventionreview

Maidment2013(122) Notaninterventiontopreventdementia

Martin2011(87) Notasystematicreviewofhighmethodologicalquality(moderatetolowquality)

Mauer2014(123) Notasystematicreview(noassessmentofstudies’methodologicalquality)

Muangpaisan2010(74)

Notasystematicreviewofhighmethodologicalquality(moderate)

Naqvi2013(124) Notasystematicreview(noassessmentofstudies’methodologicalquality)

OoiCheow2011(125) Notapreventioninterventionfordementia

Opie2013(75) Notasystematicreviewofhighmethodologicalquality(moderate)

Ott2015(126) Notapreventioninterventionfordementia

Regan2013(76) Notasystematicreviewofhighmethodologicalquality(low)

Rouch2015(77) Notasystematicreviewofhighmethodologicalquality(moderatequality).Also,cognitivelyhealthyandcognitivelyimpaired(andpeoplewithdementia)arenotreportedseparately

RuizAragon2010(127)

Notapopulationwithoutdementiadiagnosis

Santos2010(128) Notasystematicreview(noassessmentofstudies’methodologicalquality)

Simon2012(129) Notasystematicreview(noassessmentofstudies’methodologicalquality)

Sofi2011(130) Notasystematicreview(noassessmentofstudies’methodologicalquality)

Stern2010(79) Notasystematicreviewofhighmethodologicalquality(moderate)

Stern2009(80) Notasystematicreviewofhighmethodologicalquality(moderate)

Swiger2013(81) Notasystematicreviewofhighmethodologicalquality(moderate)

Tseng2011(82) Notasystematicreviewofhighmethodologicalquality(moderate)

Valenzuela2009(83,131)

Notasystematicreviewofhighmethodologicalquality(moderate)

VanderSchaft2013(132)

Notasystematicreview(noassessmentofstudies’methodologicalquality)

Wald2010(133) Notasystematicreview(noassessmentofstudies’methodologicalquality)

Weih2010(134) Notasystematicreview(noassessmentofstudies’methodologicalquality)

Wong2013(32) Notasystematicreview(noassessmentofstudies’methodologicalquality)

Yang2014(135) Notapopulationwithoutdementiadiagnosis

Zheng2015(86) Notasystematicreviewofhighmethodologicalquality(moderate)

90

4Assessmentofmethodologicalquality

WeusedtheKnowledgeCentre’schecklistforsystematicreviewsforassessingthesys‐tematicreviewsincludedinoutoverviewofreviews.Thechecklistconsistsof9questionswhichallanswerswitha‘yes’,‘unclear’or‘no’.1.Didtheauthorsofthereviewclearlydescribewhatmethodstheyusedtoidentifyprimarystudies?2.Wasacomprehensiveliteraturesearchconducted?3.Didthereviewauthorsdescribewhatcriteriatheyusedtoassesstheeligibilityofstudiesforinclusion(studydesign,participants,intervention,andend‐points)?4.Didtheauthorsattempttopreventbiaswhenselectingstudiesbyusingexpliciteligi‐bilitycriteria,andbyindependenteligibilityassessmentofstudiesbyatleasttwopeo‐ple?5.Wasasetofcriteriatoassessinternvalidityoforiginalstudiesdescribed?6.Wasthevalidityofincludedstudiesassessed(eitheratinclusionofprimarystudiesorintheanalysisofprimarystudies)usingrelevantcriteria?7.Werethemethodsusedwhensummarizingtheresultsclearlydescribed?8.Weretheresultssummarizedinanappropriateway?9.Aretheauthors’conclusionssupportedbythedataand/ortheanalysisreportedinthereview?Overallassessmentofeachreview(Question10.Howwillyourankthemethodologicalqualityofthisreview?)isitjudgedasbeing‘high’,‘moderate’or‘unclear’onthebasisofthefollowing: Highquality:Usedifallormostofthechecklistcriteriaareansweredwitha

‘yes’(adequate).Ifsomeofthecriteriaarenotfulfilled,itmustbeclearthatthiswillnothaveaneffectontheconclusionofthereview.

Moderatequality:Usedifsomeofthechecklistcriteriaarenotfulfilledornotsatisfactorydescribed.Theoverallassessmentindicatesthatthelikelihoodthatthiswouldhaveaneffectonthereviewconclusionsisverysmall.

Unclearquality:Usediffewornocriteriafromthechecklistarefulfilled,and/ornotadequatelydescribed.Overallassessmentindicatesthatitislikelythatthereview'sconclusionmaychangeduetothis.

91

Table19.Resultsoftheassessmentofthemethodologicalqualityofincludedreviews(n=8).Review 1 2 3 4 5 6 7 8 9

Farina2009 Yes Yes Yes Yes Yes Yes Yes Yes Yes

Lampit2009 Yes Yes Yes Yes Yes Yes Yes Un‐clear

Yes

Mazereeuw2012

Yes Yes Yes Un‐clear

Yes Yes Yes Yes Yes

McGuinness2009

Yes Yes Yes Yes Yes Yes Yes Yes Yes

McGuinness2016

Yes Yes Yes Yes Yes Un‐clear

Yes Yes Yes

Russ2012 Yes Un‐clear

Yes Yes Yes Yes Yes Yes Yes

Sydenham2012

Yes Yes Yes Yes Yes Yes Yes Yes Yes

Young2015 Yes Yes Yes Yes Yes Yes Yes Un‐clear

Yes

Note:NoneoftheincludedreviewshadusedtheGRADEtooltogradethecertaintyoftheevidence,orreportedtheirmainresultsinaSummaryoffindingstable.

92

5Ongoingreviews

Table20.Ongoingreviewsandoverviewsofreviews.AuthorYear ProtocoltitleAbraham2015(136) Vitaminandmineralsupplementation

forpreventionofdementiaordelayingcognitivedeclineinpeoplewithmildcognitiveimpairment

Forbes2015(137) Exerciseinterventionsformaintainingcognitivefunctionincognitivelyhealthypeopleinlatelife

Forbes2015(138) Exerciseinterventionsforpreventingde‐mentiaordelayingcognitivedeclineinpeoplewithmildcognitiveimpairment

Forbes2015 Exerciseinterventionsformaintainingcognitivefunctionincognitivelyhealthypeopleinmidlife

Harrison2015(139) Dietaryinterventionsformaintainingcognitivefunctionincognitivelyhealthypeopleinmidlife

Jordan2015(140) Aspirinandanti‐inflammatorydrugsforthepreventionofdementia

Krause2015(141) Dietaryinterventionsasaneuroprotec‐tivetherapyforthedelayoftheonsetofcognitivedeclineinolderadults:anum‐brellareviewprotocol

Siervo2015(142) Dietaryinterventionsformaintainingcognitivefunctionincognitivelyhealthypeopleinlatelife.

Tang2015(143) Dietaryinterventionsforpreventionofdementiainpeoplewithmildcognitiveimpairment

93

6Descriptionofincludedreviews

Table21.Descriptionofincludedreviews(listedinalphabeticalorder;n=8).

Farina2009(42) DescriptionofreviewPopulation: People(n=569)withmildcognitiveimpairmentfromtheUSA

andCanadaIntervention: VitaminEsupplements(2000IUperday)for3yearsComparator: PlaceboOutcomes: TimetoprogressiontopossibleorprobableADIncludedstudies: 3RCTs(Lloret2009,Petersen2005*,Sano1996)

*Includedinthisoverviewofreviews(onestudyoutof3). Lampit2014(65) DescriptionofreviewPopulation: Cognitivelyhealthyolderpeople(n=4,885)largelyfromthe

USAorEurope,butalsofromCanada,Australia,Israel,China,TaiwanSpecialAdministrativeRegion,RepublicofKorea,Ja‐panandAustralia

Intervention: >4hourscomputerizedcognitivetrainingComparator: In50%ofstudiesthecomparatorwaspassive(nointerven‐

tion)andintheremainingstudiesthecontrolconditionwasanotheractiveintervention.

Outcomes: Performanceinneuropsychologicaltests(notfurtherde‐scribed).

Includedstudies: 52trials(Ackerman2010;Anderson2013;Anguera2013;Ball2002;Barnes2013;Basak2008;Belchierstudy12013;Belchiorstudy22013;Berry2010;Boot2013;Bottiroli2009;Bozoki2013;Brehmer2012;Burki2014;Buschkuehl2008;Casutt2014;Colzato2011;Dahlin2008;Dustman1992;Ed‐wards2002;Edwards2005,Edwards2013;Garcia‐Campu‐zano2013;Goldstein1997;Heinzel2013;Lampit2014;Lee2012;Legault2011;Li2010;Lussier2012;Mahncke2006;Maillot2012;Mayas2014;McAvinue2013;Miller2013;Nou‐chi2012;O’Brien2013;Peng2012;Peretz2011;Rasmusson1999;Richmond2011;Sandberg2014;Shatil2013;Shatil2014;Simpson2012;Smith2009;Stern2011;Vance2007;VanMuijden2012;vonBastian2013;Wang2011,Wolinski2011).All52trialswereincludedinouroverviewofreviews.

Mazereeuw2012(71)

Descriptionofreview

Population: Cognitivelyhealthypeople(n=1,218),andcognitivelyIm‐pairedpeople(n=670),withnodementiadiagnosisfromthe

94

Netherlands,England,Wales,Japan,Australia,IsraelandtheUSA

Intervention: Omega‐3FattyAcidsComparator: PlaceboOutcomes: Changeincognitivetestscores(Immediaterecall;delayedre‐

call,recognition,workingmemoryandexecutivefunction,at‐tentionandprocessingspeed).

Includedstudies: 10trials(3trialsthatincludedcognitivelyhealthypeople:Dangour2010*;Johnson2008*;VandeRest2008*;4RCTsthatincludedpeoplewithCIND(Kotani2006*;Sinn2011*;Vakhapora2010*;Yurko‐Mauro2010*)andthreestudiestar‐getingpeoplewithAD(Chiu2008;Freund‐Levi2006;Quinn2010).Includedinthisoverview(4outof10trials,i.e.thetri‐alsofcognitivelyimpairedpeople).

McGuinness2016(72)

Descriptionofreview

Population: Cognitivelyhealthypeople(n=26,340)aged40to82yearsofwhom11,610wereaged70orolder,andwithevidenceofcer‐ebrovasculardiseaseorathighriskofcerebrovasculardis‐easefromUK,Scotland,IrelandandtheNetherlands

Intervention: Cholesterolloweringdrugsi.e.anymemberofthestatinfam‐ily.Dose:40mgperday.

Comparator: Placebo

Outcomes: Primaryoutcomes:objectivediagnosisofdementia,ADorVADaccordingtostandardcriteria,changeincognitivetestscores(MMSE,ADAS‐cogorotheracceptedobjectiveandstandardizedtests).

Secondaryoutcomes:cholesterollevel;incidenceandseverityofadverseeffects,changeincognitivestatusinpatientsatriskofdementiaontreatmentwithstatinsaccountingforpriorcholesterollevel,APOEgenotypeandcognitivelevel,qualityoflife,changeinActivitiesofDailyLiving(ADL),andchangeinbehaviour

Includedstudies: 2RCTs(HPS2002andPROSPER2002)ofwhichbothwererelevantforthisoverviewofreviews

McGuinness2009(73)

Descriptionofreview

Population: Cognitivelyhealthypeople(n=15,936)withadiagnosisofhy‐pertensionandanaverageageof75.4years(range60to89),

95

fromEurope(EastandWest),NorthAmerica,China,andtoalesserextentfromAustralasiaandTunisia.

Intervention: Antihypertensivedrugs

Note:Thereviewauthorsaimedtoincludealsostudiesevalu‐atingnon‐pharmacologicalhypertensiontreatmentinthere‐view,butfoundonlystudiesevaluatingpharmacologicalther‐apy

Comparator: Placebo

Outcomes: Primaryoutcomes:Incidenceofdementia(diagnosedaccord‐ingtostandarddiagnosticcriteriaorthoseappropriateatthetime),changeincognitivetestscores

Secondaryoutcomes:Bloodpressurelevel,incidenceandse‐verityofadverseeffects,andqualityoflife

Includedstudies: 4RCTs(HYVET2008;SCOPE2003;SHEP1991,SystEur1997)Allstudieswereeligibleforinclusioninouroverviewofre‐views.

Russ2012(78) Descriptionofreview

Population: Peoplewithmildcognitiveimpairment(n=5,149)fromtheUSA,Canada,Germany,Singapore,andfromanumberofcountriesthatwerenotnamed

Intervention: Cholinesteraseinhibitors(ofanydosebutatleastonemonths’duration)

Comparator: Placebo

Outcomes: Progressiontodementia,eitheringeneralorspecificsub‐types,sideeffects,

Secondaryoutcomes:changeincognitivetestscores

Includedstudies: 8RCTs(Doody2009,Petersen2005a;Salloway2004;Feld‐man2007,Koontz2005;Narasimhalu2010,Winblad2008(combined)).All8studieswereincludedinthisoverviewofreviews.

Sydenham2012(40)

Descriptionofreview

Population: Cognitivelyhealthypeople(n=4,080),60yearsandolderfromtheNetherlands,EnglandandWales.Inonestudythelocationwasunknown.

Intervention: VitaminEsupplements(capsulesorenrichedmargarine)

Comparator: Placeboorusualmargarine

96

Outcomes: Measuresofcognitivefunction,adverseeffectsandadherencetotherapy.

Includedstudies: 3RCTs,ofwhichonetrialincludedtwointerventions(Dangour2010;Geleijnse2011;VandeRest2008highandmediumdose).Allstudieswererelevantforthisoverviewofreviews.

Young2015(85) Descriptionofreview

Population: Cognitivelyhealthypeople(N=754);meanagerangedfrom61to91yearsacrossthestudieswhichwerefromUSA,FranceandCanada

Intervention: Aerobicexercisetraining

Comparator: Nointerventionorotheractiveintervention

Outcomes: Measuresofcognitivefunction,adverseeffectsanddrop‐out.Inaddition,anobjectivemeasureofcardiorespiratoryfitnesswasrequired.

Includedstudies: 11RCTS(Bakken2001,Blumenthal1989,Emery1990a,Fa‐bre2002,Kramer2001,Langlois2012,Legault2011,Madden1989,Moul1995,Oken2006, Panton1990,Whitehurst1991)ofwhichallwererelevantforouroverviewofreviews.

97

7GRADEevidenceprofiles

GRADEevidenceprofilesfortheeightincludedreviewsarefoundbeloworganisedinalphabeticalorder.

98

Author(s): Flodgren G, Berg R Date: 11.12.15 Question: Vitamin E compared to placebo for prevention of Alzheimer's disease in people with mild cognitive impairment Setting: USA and Canada Bibliography: Farina N, Isaac MG, Clark AR, Rusted J, Tabet N. Vitamin E for Alzheimer’s dementia and mild cognitive impairment. Cochrane Database of Systematic Reviews 2012, Issue 11. Art. No.: CD002854. DOI:10.1002/14651858.CD002854.pub3.

Quality assessment № of patients Effect

Quality

№ of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Vitamin E Placebo Relative (95% CI)

Absolute (95% CI)

Possible or probable Alzheimer's disease (follow up: mean 3 years; assessed with: no information)

1 randomised trials not serious not serious not serious serious 1,2 none 76/257 (29.6%) 73/259 (28.2%) HR 1.02 (0.74 to 1.41)

5 more per 1000 (from 65 fewer to 91 more) ⨁⨁⨁◯

MODERATE

HR – hazard ratio; CI: confidence interval 1. Only one study. Relatively small groups. 2. Wide CI overlapping no effec

99

Author(s): Flodgren G, Berg R Date: 17.12.15 Question: Computerised cognitive training compared to active or passive intervention for prevention of age-related cognitive decline in cognitively healthy older people Setting: the participants’ homes (unsupervised training) and in other facilities (centre-based group training). Trials mainly from the USA or Europe, with the addition of studies from Canada, Australia, Israel, China, Taiwan Special Administrative region, Republic of Korea and Japan Bibliography: Lampit A, Hallock H, Valenzuela M (2014) Computerized Cognitive Training in Cognitively Healthy Older Adults: A Systematic Review and Meta-Analysis of Effect Modifiers. PLoS Med 11(11): e1001756. doi:10.1371/journal.pmed.1001756

Quality assessment № of patients Effect

Quality

№ of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Computerised cognitive training Active or passive intervention Relative (95% CI)

Absolute (95% CI)

Cognitive test scores (assessed with: no information)

52 randomised trials serious 1 not serious not serious not serious none 2527 2358 - MD 0.22 higher (0.15 higher to 0.29 higher) ⨁⨁⨁◯

MODERATE

CI: Confidence interval; MD: Mean difference 1. Thirty-three of 52 studies at high risk of bias. Eighteen studies were at low risk.

100

Author(s): Flodgren G, Berg R Date: 17.12.15 Question: Omega-3 fatty acids to improve cognitive function as compared to placebo for people with cognitive impairment no dementia (CIND) Setting: the Netherlands, England, Wales, japan, Israel and the US Bibliography: Mazereeuw G, Lanctot KL, Chau SA, Swardfager W, Herrmann N. Effects of omega-3 fatty acids on cognitive performance: a meta-analysis. Neurobiol Aging 2012. Jul;33 (7):1482.e17-29.

Quality assessment № of patients Effect

Quality

№ of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Omega-3 fatty acids Placebo Relative (95% CI)

Absolute (95% CI)

Cognitive test scores - Composite memory (follow up: median 14.5 weeks; assessed with: unclear)

4 randomised trials not serious not serious not serious not serious none 349 327 - MD 0.1 higher (0.06 lower to 0.25 higher)

⨁⨁⨁⨁

HIGH

Cognitive test scores - Immediate recall (follow up: median 19.5 weeks; assessed with: unclear)

4 randomised trials not serious not serious not serious not serious none 349 327 - MD 0.16 higher (0.01 higher to 0.32 higher)

⨁⨁⨁⨁

HIGH

Cognitive test scores - Delayed recall (follow up: median 19.5 weeks; assessed with: unclear)

4 randomised trials not serious not serious not serious not serious none 349 327 - MD 0.03 higher (0.12 lower to 0.18 higher)

⨁⨁⨁⨁ HIGH

Cognitive test scores - Recognition (assessed with: unclear)

3 randomised trials not serious not serious not serious not serious none 337 318 - MD 0.03 lower (0.18 lower to 0.13 higher)

⨁⨁⨁⨁

HIGH

Cognitive test scores - Attention and processing speed (assessed with: unclear)

3 randomised trials not serious not serious not serious not serious none 107 86 - MD 0.32 higher (0.03 higher to 0.61 higher)

⨁⨁⨁⨁ HIGH

101

Quality assessment № of patients Effect

Quality

№ of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Omega-3 fatty acids Placebo Relative (95% CI)

Absolute (95% CI)

Cognitive test scores - Working memory and executive function (assessed with: unclear)

2 randomised trials not serious not serious not serious not serious none 277 256 - MD 0.04 higher (0.13 lower to 0.21 higher)

⨁⨁⨁⨁ HIGH

Cognitive test scores - MMSE (follow up: mean 24 weeks; assessed with: MMSE)

1 randomised trials not serious not serious not serious not serious none 242 241 - MD 0.06 lower (0.23 lower to 0.12 higher)

⨁⨁⨁⨁

HIGH

CI: Confidence interval; MD: Mean difference

102

Author(s): Flodgren G, Berg R Date: 11.12.15 Question: Blood pressure lowering drugs compared to placebo for the prevention of dementia in cognitively healthy people with hypertension Setting: Western and Eastern Europe, North America, China and to a lesser extent Australasia and Tunisia Bibliography: McGuinness B, Todd S, Passmore P, Bullock R. Blood pressure lowering in patients without prior cerebrovascular disease for prevention of cognitive impairment and dementia. Cochrane Database of Systematic Reviews 2009, Issue 4. Art. No.: CD004034. DOI: 10.1002/14651858.CD004034.pub3.

Quality assessment № of patients Effect

Quality

№ of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Antihypertensive drugs Placebo Relative (95% CI)

Absolute (95% CI)

Incidence of dementia (follow up: median 2.8 years; assessed with: DSM III-R, DSM-IV, ICD 10, MMSE)

4 randomised trials not serious not serious not serious serious 1 none 236/7767 (3.0%) 259/7660 (3.4%) OR 0.89 (0.74 to 1.07)

4 fewer per 1000 (from 2 more to 9 fewer)

⨁⨁◯◯ LOW

Change in cognitive test scores from baseline (MMSE) (follow up: range 1.8 years to 4.5 years; Scale from: 0 to 30)

3 randomised trials not serious serious 2 not serious not serious none 5402 5238 - MD 0.42 higher (0.3 higher to 0.53 higher) ⨁⨁⨁◯

MODERATE

Adverse events (follow up: range 2 years to 4.5 years)

3 randomised trials not serious Serious2 not serious not serious none 1138/6080 (18.7%) 1117/6011 (18.6%) OR 1.01 (0.92 to 1.11)

2 more per 1000 (from 12 fewer to 16 more) ⨁⨁⨁◯

MODERATE

MD – mean difference, OR – odds ratio 1. Wide CI. 2. High heterogeneity: I2=98%.

103

Author(s): Flodgren G, berg R Date: 17.12.15 Question: Statins compared to placebo for prevention of dementia in cognitively healthy people Setting: UK, Ireland and the Netherlands Bibliography: McGuinness B, Craig D, Bullock R, Passmore P. Statins for the prevention of dementia. Cochrane Database of Systematic Reviews 2016, Issue 1. Art. No.: CD003160. DOI: 10.1002/14651858.CD003160.pub3

Quality assessment № of patients Effect

Quality

№ of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Statins Placebo Relative (95% CI)

Absolute (95% CI)

Dementia incidence (follow up: range 3.2 years to 5 years)

1 randomised trials not serious not serious not serious serious 1 none 31/10269 (0.3%) 31/10267 (0.3%) OR 1.00 (0.61 to 1.65)

0 fewer per 1000 (from 1 fewer to 2 more) ⨁⨁⨁◯1

MODERATE

Change in cognitive test scores from baseline (MMSE) (follow up: range 3.2 years to 5 years)

1 randomised trials not serious not serious not serious not serious none 2891 2913 - MD 0.06 higher (0.04 lower to 0.16 higher)

⨁⨁⨁⨁ HIGH

Change in cognitive test scores from baseline (Stroop) (follow up: range 3.2 years to 5 years)

1 randomised trials not serious not serious not serious serious 1 none 2891 2913 - MD 0.8 higher (0.38 lower to 1.98 higher) ⨁⨁⨁◯1

MODERATE

Change in cognitive test scores from baseline (Picture world) (follow up: range 3.2 years to 5 years)

1 randomised trials not serious not serious not serious not serious none 2891 2913 - MD 0.02 higher (0.12 lower to 0.16 higher)

⨁⨁⨁⨁

HIGH

Change in cognitive test scores from baseline (Letter digit test) (follow up: range 3.2 years to 5 years)

1 randomised trials not serious not serious not serious not serious none 2891 2913 - MD 0.01 lower (0.25 lower to 0.23 higher)

⨁⨁⨁⨁ HIGH

104

Quality assessment № of patients Effect

Quality

№ of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Statins Placebo Relative (95% CI)

Absolute (95% CI)

Cognitive test scores- TICS-m at final visit (follow up: range 3.2 years to 5 years)

1 randomised trials not serious not serious not serious not serious none 10269 10267 - MD 0.02 higher (0.12 lower to 0.16 higher)

⨁⨁⨁⨁ HIGH

Adverse effects requiring discontinuation of treatment (follow up: range 3.2 years to 5 years)

2 randomised trials not serious not serious not serious not serious none 13160 13180 - OR 0.94 higher (0.83 higher to 1.05 higher)

⨁⨁⨁⨁

HIGH

CI: Confidence interval; OR: Odds ratio; MD: Mean difference 1. Wide CI.

105

Author(s): Flodgren G, Berg R Date: 11.12.15 Question: Cholinesterase inhibitors compared to placebo for people with mild cognitive impairment no dementia Setting: USA, Canada, Singapore and a number of other countries Bibliography: Russ TC, Morling JR. Cholinesterase inhibitors for mild cognitive impairment. Cochrane Database Syst Rev [Internet]. 2012 [cited CDSR 211114; (9).

Quality assessment № of patients Effect

Quality № of stu-dies Study design Risk of

bias Inconsistency Indirectness Imprecision Other considerations Cholinesterase inhibi-tors Placebo Relative

(95% CI) Absolute (95% CI)

Conversion to dementia (follow up: mean 3 years; assessed with: NINCDS-ADRDA with or without DSM-IV, and without explicit criteria in one study)

2 randomised tri-als

not serious not serious 2 not serious serious none 151/761 (19.8%) 182/769 (23.7%) RR 0.84 (0.70 to 1.02)

38 fewer per 1000 (from 5 more to 71 fewer)

⨁⨁⨁◯ MODERATE

Cognitive test scores - ADAS-Cog (follow up: mean 2 years; assessed with: ADAS-Cog (modified))

4 randomised tri-als

not serious serious 1 not serious serious 2 none 1321 1354 - MD 0.78 lower (1.92 lower to 0.35 higher)

⨁⨁◯◯ LOW

Any adverse event

5 randomised tri-als

not serious serious 3 not serious not serious none 1849/2101 (88.0%) 1737/2106 (82.5%)

RR 1.09 (1.02 to 1.16)

74 more per 1000 (from 16 more to 132 more)

⨁⨁⨁◯ MODERATE

Cognitive test scores - CDR Sum of boxes (follow up: mean 1 years)

2 randomised tri-als

not serious not serious not serious not serious none 632 637 - MD 0.1 lower (0.11 lower to 0.09 lower)

⨁⨁⨁⨁ HIGH

Cognitive test scores- Symbol digit modalities (follow up: mean 6 months)

2 randomised tri-als

not serious serious 4 not serious very serious 5

none 155 157 - MD 0.17 higher (2.87 lower to 3.21 higher)

⨁◯◯◯ VERY LOW

Cognitive test scores -Mini Mental State Examination (follow up: mean 1 years; Scale from: 0 to 30)

2 randomised tri-als

not serious serious 3 not serious serious 6 none 632 637 - MD 0.24 higher (0.13 lower to 0.61 higher)

⨁⨁◯◯ LOW

Cognitive test scores - ADCS-ADL (follow up: mean 1 years; Scale from: 0 to 78)

3 randomised tri-als

not serious not serious not serious serious 5 none 1191 1217 - MD 0.15 higher (0.27 lower to 0.57 higher)

⨁⨁⨁◯ MODERATE

Serious adverse events

106

Quality assessment № of patients Effect

Quality № of stu-dies Study design Risk of

bias Inconsistency Indirectness Imprecision Other considerations Cholinesterase inhibi-tors Placebo Relative

(95% CI) Absolute (95% CI)

5 randomised tri-als

not serious not serious not serious serious 2 none 391/2101 (18.6%) 401/2106 (19.0%) RR 0.97 (0.86 to 1.10)

6 fewer per 1000 (from 19 more to 27 fewer)

⨁⨁⨁◯ MODERATE

MD – mean difference, RR – risk ratio

1. High heterogeneity: I2=98%. 2. Wide CI overlapping no effect. 3. High heterogeneity: I2=79%. 4. High heterogeneity: I2=99% 5. Wide CI overlapping no effect. 6. Wide CI overlapping no effect.

107

Author(s): Flodgren G, Berg R Date: 17.12.15 Question: Omega-3 fatty acids compared to placebo for prevention of cognitive decline and dementia in cognitively healthy older people Setting: England and Wales Bibliography: Sydenham E, Dangour AD, Lim WS. Omega 3 fatty acid for the prevention of cognitive decline and dementia. Cochrane Database Syst Rev 2012;6:CD005379. (NB! Also published in Sydenham E, Dangour AD, Lim WS. Omega 3 fatty acid for the prevention of cognitive decline and dementia. Sao Paulo Med J 2012;130(6):419.)

Quality assessment № of patients Effect

Quality

№ of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Omega-3 fatty acids placebo Relative (95% CI)

Absolute (95% CI)

Cognitive test scores - MMSE (follow up: range 24 months to 40 months)

2 randomised trials not serious serious 1 not serious not serious none 990 2231 - MD 0.07 lower (0.25 lower to 0.1 higher) ⨁⨁⨁◯

MODERATE

Cognitive test scores -Immediate recall (follow up: range 6 months to 24 months)

3 randomised trials not serious not serious not serious not serious none 472 571 - SMD 0.01 higher (0.11 lower to 0.14 higher)

⨁⨁⨁⨁ HIGH

Cognitive test scores - Delayed recall (follow up: range 6 months to 24 months)

3 randomised trials not serious not serious not serious not serious none 472 571 - SMD 0.04 lower (0.16 lower to 0.09 higher)

⨁⨁⨁⨁

HIGH

Cognitive test scores - Word recognition (follow up: range 6 months to 24 months)

3 randomised trials not serious not serious not serious not serious none 472 571 - SMD 0.04 higher (0.08 lower to 0.16 higher)

⨁⨁⨁⨁ HIGH

Cognitive test scores- Number of animals named (follow up: range 6 months to 24 months)

108

Quality assessment № of patients Effect

Quality

№ of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Omega-3 fatty acids placebo Relative (95% CI)

Absolute (95% CI)

3 randomised trials not serious not serious not serious not serious none 472 570 - SMD 0.06 higher (0.06 lower to 0.18 higher)

⨁⨁⨁⨁ HIGH

Cognitive test scores - Digit span forward (follow up: range 6 months to 24 months)

3 randomised trials not serious not serious not serious not serious none 458 560 - MD 0.03 higher (0.25 lower to 0.31 higher)

⨁⨁⨁⨁

HIGH

Cognitive test scores - Digit span backwards (follow up: range 6 months to 24 months)

3 randomised trials not serious not serious not serious not serious none 458 557 - MD 0.12 higher (0.12 lower to 0.36 higher)

⨁⨁⨁⨁ HIGH

CI: Confidence interval; MD: Mean difference; SMD: Standardised mean difference 1. High I2.

109

Author(s): Flodgren GM, Berg R Date: 27.04.16 Question: Aerobic exercise compared to active or passive control for cognitive decline Setting: USA, France and Canada Bibliography: Young J, Angevaren M, Rusted J, Tabet N. Aerobic exercise to improve cognitive function in older people without known cognitive impairment. Cochrane Database of Systematic Reviews 2015, Issue 4. Art. No.: CD005381. DOI: 10.1002/14651858.CD005381.pub4.

Quality assessment № of patients Effect

Quality

№ of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations primary and secondary prevention interventions control Relative (95% CI)

Absolute (95% CI)

Cognitive test scortes

6 randomised trials serious 1 not serious not serious serious 2 none 389 - - SMD 0.12 SD higher (0.08 lower to 0.33 higher) ⨁⨁◯◯

LOW

Verbal memory function (immediate)

2 randomised trials serious 1 serious not serious serious 2 none 299 - - SMD 0.08 SD higher (0.38 lower to 0.55 higher) ⨁◯◯◯

VERY LOW

Visual memory functions (immediate)

2 randomised trials serious 1 serious not serious serious 3 none 89 - - SMD 0.26 SD lower (0.97 lower to 0.44 higher) ⨁◯◯◯

VERY LOW

Woorking memory

3 randomised trials serious 1 not serious not serious serious 2 none 238 - - SMD 0.1 SD higher (0.16 lower to 0.36 higher) ⨁⨁◯◯

LOW

Memory function (delayed)

110

Quality assessment № of patients Effect

Quality

№ of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations primary and secondary prevention interventions control Relative (95% CI)

Absolute (95% CI)

3 randomised trials serious 1 not serious not serious serious 2 none 249 - - SMD 0.1 SD higher (0.16 lower to 0.35 higher) ⨁⨁◯◯

LOW

Executive functions

6 randomised trials serious 1 serious 4 not serious serious 2 none 367 - - SMD 0.38 SD higher (0.14 lower to 0.9 higher) ⨁◯◯◯

VERY LOW

Perception

3 randomised trials serious 1 not serious not serious serious 2 none 178 - - SMD 0.01 SD lower (0.5 lower to 0.48 higher) ⨁⨁◯◯

LOW

Cognitive inhibition

4 randomised trials serious 1 not serious not serious serious 2 none 314 - - SMD 0.06 SD lower (0.28 lower to 0.17 higher) ⨁⨁◯◯

LOW

Visual attention

3 randomised trials serious 1 not serious not serious serious 2 none 265 - - SMD 0.22 SD higher (0.03 lower to 0.46 higher) ⨁⨁◯◯

LOW

Auditory attention

111

Quality assessment № of patients Effect

Quality

№ of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations primary and secondary prevention interventions control Relative (95% CI)

Absolute (95% CI)

4 randomised trials serious 1 not serious not serious serious 2 none 251 - MD 0.15 higher (0.38 lower to 0.69 higher) ⨁⨁◯◯

LOW

Motor function

2 randomised trials serious 1 not serious not serious serious 2 none 189 - - SMD 0.08 SD higher (0.2 lower to 0.37 higher) ⨁⨁◯◯

LOW

CI: Confidence interval; SMD: Standardised mean difference; MD: Mean difference 1. High to moderate risk of bias in one or more domains., 2. Fewer than 400 particpants., 3.No explanation was provided, 4. High heterogeneity I2=80%

112

Quality assessment № of patients Effect

Quality

№ of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations primary and secondary prevention interventions control Relative (95% CI)

Absolute (95% CI)

4 randomised trials serious 1 not serious not serious serious 2 none 251 - MD 0.15 higher (0.38 lower to 0.69 higher) ⨁⨁◯◯

LOW

Motor function

2 randomised trials serious 1 not serious not serious serious 2 none 189 - - SMD 0.08 SD higher (0.2 lower to 0.37 higher) ⨁⨁◯◯

LOW

CI: Confidence interval; SMD: Standardised mean difference; MD: Mean difference 1. High to moderate risk of bias in one or more domains., 2. Fewer than 400 particpants., 3.No explanation was provided, 4. High heterogeneity I2=80%

www.fhi.no

Utgitt av FolkehelseinstituttetSeptember 2016Postboks 4404 NydalenNO-0403 OsloTelefon: 21 07 70 00Rapporten lastes ned gratis fra Folkehelseinstituttets nettsider www.fhi.no