Availableonlineatwww.sciencedirect.com
journalhomepage:www.elsevier.com/locate/psyneuen
SHORTCOMMUNICATION
Thyroidstimulatinghormone,freethyroxineandcognitiveabilityinoldage:TheLothianbirthcohortstudy1936
TomBootha,b,*,IanJ.Dearya,b,JohnM.Starra,caCentreforCognitiveAgeingandCognitiveEpidemiology,TheUniversityofEdinburgh,EH89JZEdinburgh,UKDepartmentofPsychology,UniversityofEdinburgh,EH89JZEdinburgh,UKcAlzheimerScotlandDementiaResearchCentre,TheUniversityofEdinburgh,EH89JZEdinburgh,UK
bReceived17May2012;receivedinrevisedform30July2012;accepted31July2012
KEYWORDSThyroidstimulatinghormone;TSH;Freethyroxine;T4;Cognitiveability;Memory;ProcessingspeedSummaryThecurrentstudyinvestigatedtheassociationsbetweenthyroidstimulatinghor-mone(TSH),freethyroxine(T4)andcognitiveability(generalability,memoryandprocessingspeed),inalargeagehomogenoussample(n=659)ofgenerallyhealthyeuthyroidolderadults.Associationswereconsideredbothatbaseline(meanagewave1=69.5years;SD=0.8years)andapproximately3yearslater(meanagewave2=72.5years;SD=0.7years).ResultsindicatedmeanleveldecreasesacrosswavesinbothTSH(t=10.99,p<0.001)andT4(t=34.55,p<0.001).TherewerenosignificantassociationsbetweenTSHandT4withanyofthecognitivevariablesateitherwave.Therewasnosuggestionofnon-linearassociations.Thelackofassociationssupportssuggestionsthattheeffectsofthyroidhormonesoncognitionmayberestrictedtoolderindividualsaboveagiventhreshold,and/orthosewithlevelsofthyroidhormoneswithintherangedefiningclinicalthyroiddisorder.#2012ElsevierLtd.Allrightsreserved.1.IntroductionThemaintenanceofcognitiveabilitiesintolaterlifeisakeydeterminantofawidevarietyofpositivelifeoutcomesandsuccessfulageing(Deary,2012).Thyroidfunctionismootedtobeimportantforthedevelopmentandretentionofcognitive*Correspondingauthorat:CentreforCognitiveAgeingandCogni-tiveEpidemiology,DepartmentofPsychology,TheUniversityofEdinburgh,7GeorgeSquare,EH89JZ,UK.Tel.:+4401316508405.E-mailaddress:tom.booth@ed.ac.uk(T.Booth).
function(Beginetal.,2008),withthyroiddysfunction,mostcommonlyclinicalandsubclinicalhypothyroidism,inlaterlifebeingassociatedwithbothdementiaand,morespecifically,Alzheimerdisease(Smithetal.,2002;Liesbethetal.,2004).Asaresult,thereisgrowinginterestinwhetherthyroidhormonesatbothnormalandclinicallevels,areassociatedwithcogni-tiveperformanceacrossthelifespan.However,researchhasfailedtoshowconsistentassociationsbetweencognitiveabil-itiesandthyroidfunctioninsamplesofgenerallyhealthyelderlysubjectswithsubclinicalthyroiddysfunction.Forexample,St.Johnetal.(2009)foundnoassociationbetweenTSHlevelsandcognitiveabilityinasampleof489menandwomen(meanage=60.5years).Robertsetal.
0306-4530/$—seefrontmatter#2012ElsevierLtd.Allrightsreserved.http://dx.doi.org/10.1016/j.psyneuen.2012.07.018
598
(2006)foundnosignificantdifferencesinMini-MentalStateExam(MMSE)orMiddlesexElderlyAssessmentofMentalState(MEAMS)scoresbetweenindividualswithsubclinicalthyroiddisordersandeuthyroidfromalargecross-sectionalsample(BirminghamElderlyThyroidStudy;n=5685,age=>65years).HigherTSHwasassociatedwithhigherMMSEscores,butnottoMEAMSscore.Usingasubsampleofthesamecohortinarandomizedcontrolledtrial,Parleetal.(2010)foundnosignificantdifferencesbetweentheplacebogroup(n=42,meanage74.2years)andagroupreceivingT4hormonetreatment(n=52,meanage73.5years)inanycognitivemeasures.Similarly,inafollowupstudyof599olderindi-viduals(85—89years),Gusseklooetal.(2004)foundnoassociationsbetweenTSH,T3orT4andcognitivefunctioning.Hogervorstetal.(2008)foundthataftercontrollingforthepresencethyroiddisease,strokeanddementia,highernor-malT4wasassociatedwithlowerMMSEscoresatbaseline,andagreaterriskofdeclineacross2years,inasampleof1047olderadults(meanageatbaseline73.6years).Con-versely,usingtheNHANESIIIsample,Beydounetal.(2012)foundthathigherlevelsofthyroxine,TSHwereassociatedwithimprovedcognitiveperformanceonamathtest,andastoryrecalltest,inthe60—90yearsagegroup(n=5989and5878respectively).
Beginetal.(2008)reportingonfivepreviouslypublishedstudiesofeuthyroidsamplesconcludedthat,fromthemixedresults;thatitisunclearwhetherthyroidstimulatinghor-mone(TSH),tri-iodothyronine(T3)orfreethyroxine(T4)isthemostrelevantassociateofcognitivefunction;andthattheonlyconsistentcognitiveassociationwaswithmemoryimpairmentacrossallbiomarkers.Asimilarconclu-sionsisreachedfromtheninestudiesreviewedbyHogervorstetal.(2008;seeTable1,p.1014),inwhichfourstudiesfoundsomedegreeofsignificantassociationbetweenmemoryandthyroidhormonelevelsinoldersamples(agerangeacrossstudies49—89).
Intotality,resultsfrompriorstudieshavebeeninconsis-tent,andprovidelittleevidenceforassociationsbetweenthyroidhormonesandcognitiveabilityingenerallyhealthyolderadults.Samuels(2010),inreviewingthegenerallackofassociationsfoundinstudiesofsubclinicalhypothyroidism,notedthat:(a)fewstudiesreportonthelargelymoresensitiveT4measurements(althoughforimpairedthyroidfunctionTSHmightbepreferred);and(b)moststudiesrelyoncognitiveassessmentsdesignedtoidentifygrossimpair-ment(e.g.MMSEorMEAMS).Suchmeasuresmaybeinsensi-tivetochangesincognitivefunctionwithinthenormalrangeofabilities.
InthecurrentstudyweinvestigatetheassociationsbetweenTSH,T4andmajordomainsofcognitiveabilityassessedusingwidelyvalidatedpsychometrictools,inalargeage-homogeneoussampleofgenerallyhealthyolderadults.
2.Methods
2.1.Sample
ThecurrentsampleisdrawnfromtheLothianBirthCohortStudy1936(LBC1936),alongitudinalstudyofageing.Fullinformationontherecruitmentandtestingofparticipantshasbeenpreviouslypublishedinstudyprotocolsandprofile
T.Boothetal.
(Dearyetal.,2007,2011).Inshort,survivingmembersoftheScottishMentalHealthSurvey1947(SMS1947)residentinEdinburghandthesurroundingLothianarea,wererecruitedbetweenJune2004andNovember2006.
Intotal,1091participantstookpartinwave1oftestingwheretheycompletedawidearrayofpsychometrictests,aphysicalexaminationentailinganumberoftestsofphysicalfunctioning(6mwalk,stand-sittest,gripstrength,bloodpressureetc.).Bloodsampleswerealsotakenaspartofthephysicalexamination.Atwave2,866participantsreturnedapproximatelythreeyearslaterandrepeatedlargelythesamearrayofcognitive,psychometric,andphysicaltests.Inthepresentstudy,weincludedparticipantswho(nindicatingthenumberswhomeetinclusioncriteriaateachstep);
1.Completedbothwaves1and2(n=866).2.HadcompletedataforTSHandT4(n=810).
3.Didnotreportanythyroiddiseaseatbothwaves1and2(n=711).
4.Scored!24ontheMini-MentalStateExam(Folsteinetal.,1975),aslowerscoresareoftenconsideredtobeamarkerofcognitiveimpairment(n=701).
5.HadTSHlevelswithintherangeof0.2—4.5mU/latbothwaves1and2(n=660).
6.HadT4levelswiththerangeof9—21pmol/Lwaves1and2(n=659).
Applyingthesecriteriaresultedinafinalsampleof659(male=371;female=288;meanagewave1=69.5years,SD=0.8years;meanagewave2=72.5years,SD=0.7years)participantsforthecurrentanalyses.
2.2.Ethicalconsiderations
EthicalpermissionfortheLBC1936studyprotocolwasobtainedfromtheMulti-CentreResearchEthicsCommitteeforScotlandandtheLothianResearchEthicsCommittee.AllresearchwascarriedoutincompliancewiththeHelsinkiDeclaration.
2.3.TSHandT4Bloodsamplesweretakenatbothwaves1and2.Allanalyseswereconductedinthesamelaboratory(NHSLothian—ClinicalBiochemistry&Haematology,WesternGeneralHos-pital,UK)usingatwo-stepimmunoassay.TargetprecisionofboththeTSHandT4assayswerecoefficientofvariability<10%.
ForTSH,thelaboratoryreferencerangewas0.2—4.5mU/l,withcoefficientsofvariabilityrangingfrom3.0%to3.5%.ForT4,thelaboratoryreferencerangewas9—21pmol/L,withcoefficientsofvariabilityrangingfrom5.1%to8.9%.Inthecurrentstudyweusethelaboratoryreferencerangeasaselectioncriterion.Thisreferencerangehasbeenusedinotherpublishedworkondifferentsamples(Pattersonetal.,2010).
2.4.Cognitiveabilitytests
Inthecurrentstudyweused14cognitiveabilitysubscalescoresfrom12individualcognitivetests.Generalcognitive
Thyroidstimulatinghormone,freethyroxineandcognitiveabilityinoldage:TheLothianbirthcohortstudy1936
599
ability(g)wasmeasuredusing6subtestscoresoftheWAIS-IIIUK(Wechsler,1998a;BlockDesign,MatrixReasoning,DigitSpanBackward,DigitSymbol,SymbolSearchandLetter-NumberSequencing).ProcessingSpeedwasmeasuredusingtwotimedWAIS-IIIUKsubtestscores(DigitSymbolandSymbolSearch),asimpleandchoicereactiontimetask,andaninspectiontimetask.Lastly,MemorywasmeasuredusingfivesubtestscoresfromtheWMS-IIIUK(Wechsler,1998b;LogicalMemoryImmediateanddelayedrecall,VerbalPairedAssociatesimmediateanddelayedrecallandSpatialSpantotalscore)andtwoscoresfromtheWAIS-IIIUK,DigitSpanBackwardandLetter-NumberSequencing.
2.6.Covariates
2.5.Statisticalanalysis
Means,standarddeviations,skewandkurtosiswerecom-putedforallrawdata.Pearson’scorrelationsbetweenindi-vidualsubtestscores,TSHandT4atbothwaves1and2werecomputed.Next,weusedstructuralequationmodelling(SEM)toestimatethecross-laggedcorrelationsbetweenthreelatentcognitiveabilityvariables(g,processingspeed,andmemory)andTSHandT4measuresatbothwaves1and2.Thelatentfactorsrepresentthesharedassociationsbetweenindividualcognitiveabilitysubtestscoreshypothesizedtomeasurethesamemajordomainofcognitiveability.SEMprovideshighlyreliableestimatesofbroadcognitivedomainsbytakingaccountofmeasurementerrorpresentineachsubtest.
AllmodelswereestimatedinMplus6.0usingfullinforma-tionmaximumlikelihoodestimation.FulldetailsofthestructuralmodelscanbefoundinSupplementarymaterialA.Lastly,weproducedscatterplotswithfittedloesscurvesoflatentvariableestimatesinordertoconsidernon-lineareffects(seeSupplementarymaterialB).
Allmodelswereestimatedbothwithandwithouttheinclu-sionofhealthcovariates.Participants’smokingstatus(for-mer/current/never),alcoholconsumption(drink/notdrink),history(yes/no)ofdiabetes,hypertension,highcholesterol,cardiovasculardisease,circulatoryproblemsandstroke,weregatheredfromamedicalinterview.Inadditionweincludedthedepressionsubscalescorefromtheself-reportHospitalAnxietyandDepressionscale(HADS:ZigmondandSnaith,1983).
InordertoassesspotentialeffectsofBMIandmedicationsknowntoeffectthyroidhormonelevelswere-ranallmodelsexcludingfirstlyparticipantswithBMIlessthan20(n=18),andsecondlythosetakinganyofthefollowingmedications;hydrocortisone,prednisolone,growthhormone,octreotide,somatostatin,morphine,dihydrocodeine,tramadol,dopa-mine,L-dopa,madopar,sinemet,pimozide,phenotolamine,thioridazine,methysergide,cyproheptadine,iodine,lithium,ropinerole,pramipexole,cabergoline,iodideoramiodarone(n=24).
3.Results
3.1.Descriptivestatistics
FulldescriptivestatisticsarepresentedinTable1.Allvari-ableswereapproximatelynormallydistributedwithnovaluesforskewoutsideofÆ1.99.Acrossthewaves,thereweresignificantmeanleveldecreasesinbothTSH(t=10.99,p<0.001)andT4(t=34.55,p<0.001).
Univariatecorrelationsbetweenindividualcognitivesubt-estsandthyroidhormonesatwaves1and2yieldedfewsignificantresults(forfulltableseeSupplementarymaterial
Table1
Means,standarddeviations,skewandkurtosisforallvariables.
Wave1
Wave2
Mean
WAIS-IIIUKDigitSymbol
DigitSpanBackwardsBlockDesign
Letter-NumberSequencingMatrixReasoningSymbolSearchWMS-IIIUKLogicalMemoryImmediateLogicalMemoryDelayed
VerbalPairedAssociatesImmediateVerbalPairedAssociatesDelayedSpatialSpanSpeedtests
SimpleReactionTimeChoiceReactionTimeInspectionTimeThyroidhormonesTSHT4SD
Skew
Kurt.
Mean
SD
Skew
Kurt.À0.19À0.090.090.45À1.020.900.140.30À0.550.99À0.033.061.352.581.18À0.12
57.877.8634.8711.1514.0225.2345.0528.172.606.3114.880.270.63112.991.8215.29
12.832.2810.093.065.056.339.907.632.171.932.820.050.0811.230.882.04
0.100.450.190.12À0.13À0.04À0.34À0.360.82À1.030.051.991.02À0.900.820.12
À0.02À0.24À0.21À0.07À0.960.86À0.11À0.07À0.170.18À0.236.572.812.700.260.05
56.747.8634.1511.0213.3624.9746.1229.172.776.3914.770.270.65111.911.5712.65
12.172.2310.143.025.005.9510.067.762.232.022.700.050.0911.760.761.58
0.080.330.430.35À0.10À0.32À0.38À0.530.65À1.33À0.101.500.82À1.061.020.22
600
T.Boothetal.
Table2Cross-laggedcorrelationsbetweencognitiveabilitylatentvariables,TSHandT4.
1
1.TSHwave12.TSHwave23.T4wave14.T4wave25.gwave16.gwave2
7.Processingspeedwave18.Processingspeedwave29.Memorywave110.Memorywave2
2
3
4
5
6
7
8
9
10
—0.76——À0.06À0.05À0.05À0.07À0.07À0.07
0.76———À0.04À0.01À0.06À0.07À0.05À0.03
———0.430.070.040.05À0.040.040.09
——0.47—À0.05À0.060.05À0.030.020.02
À0.06À0.020.02À0.01—0.98————
À0.060.000.01À0.010.99————-À0.05À0.040.05À0.05———0.98——
À0.07À0.060.05À0.02——0.98———
À0.08À0.03À0.010.04—————0.86
À0.07À0.010.070.02——-—0.89—
Note:Allestimatesshowninboldtypearethewave1towave2stabilitycoefficients,allsignificantatp<0.001.AllotherestimatesaretheassociationsbetweenTSH,T4andcognition,allnon-significant(p>0.05).Estimatesbelowthediagonalareuncorrected.Estimatesabovethediagonalarecorrectedforsex,age,healthvariablesanddepressionscores.
C).DigitSpanBackwards(À0.09),Letter-NumberSequencing(À0.08)andSimpleReactionTime(0.12)weresignificantlyassociatedwithTSHlevelsatwave1,butnotatwave2.Conversely,BlockDesign(À0.10)andSimpleReactionTime(0.10)weresignificantlyassociatedwithT4levelsatwave2,butnotwave1.Intotality,fewsystematicassociationswerepresentinthecurrentsample.
EstimatesoftheassociationsbetweenTSHorT4levelsandg,processingspeedandmemoryderivedfromcross-laggedlatentvariablestructuralmodelsarepresentedinTable2.Allstructuralmodelsshowedexcellentfittothedata(seeSupplementarymaterialA).NosignificantassociationswerefoundbetweeneitherTSHorT4andanyofthecognitivemeasures.
AllchangestoparameterestimatesremovingcaseswithlowBMIwereattheseconddecimalplace,withallvaluesremainingnon-significant.Similarly,allchangesinparameterestimateswereattheseconddecimalplaceinthemodelsexcludingparticipantscurrentlytakingpotentiallyconfound-ingmedications(seeSupplementarymaterialDforrawresults).Inthismodel,theassociationbetweenT4atwave1andMemoryatwave2wassignificant(0.10,p<0.05).However,giventhelargenumberofparametersestimatedwesuggestthisissimplyaType1error.
ScatterplotswereproducedplottingthelatentfactorscoresofeachcognitiveabilityfactoragainstTSHandT4atbothwaves.Thescatterplotsprovidednoindicationofnon-linearassociations(seeSupplementarymaterialB).
4.Discussion
Inthecurrenteuthyroidsample,wefoundnosignificantassociationsbetweenthyroidfunctionasmeasuredbyTSHandT4andthreemajordomainsofcognitiveability(g,memoryandprocessingspeed).Further,wefoundnoevidenceofnon-linearassociations.Theseconclusionsholdtruebothwithandwithoutcontrollingforalargenumberofhealthcovariates,aswellaswhenanalyseswerere-runexcludingthoseonmedicationsandwithlowBMI.Notably,bothTSHandT4levelsdecreasedoverthethree-yearfollow-upperiodconsistentwithafallinthyrotropin-releasinghormone(TRH),althoughwedidnotmeasurethisdirectly.Althoughthereareseveralcross-sectionalstudiesthathave
foundlowerTRHlevelsinolderparticipants,thereisapaucityoflongitudinaldata(Leitoletal.,2002).
Thesefindingssupportsuggestionsfrompreviousstudiesinwhichtheauthorshavesuggestedthatthyroid-cognitionassociationsmayonlybepresentatclinicallevelsofthyroiddysfunction(Beginetal.,2008;St.Johnetal.,2009).Theassociationbetweenthyroidhormonelevelsandmemoryabilitywasidentifiedasperhapsthemostconsistentfindingfrompastresearch(e.g.Beginetal.,2008;Hogervorstetal.,2008).Againinthecurrentstudy,wefoundnosupportforthisassociation.
However,thecurrentstudyoffersanumberofimpor-tantextensionstopriorstudies.Firstly,weuseamultiplecognitiveteststoidentifylatentcognitiveabilityfactorsofg,processingspeed,andmemory,andthus,providerobustestimatesfromstructuralequationmodels.Sec-ondly,wewereabletoestimatetheassociationsinthesamelargesample(n=659)atbaselineandwithanapproximatethreeyearfollowup.Giventhissamplesize,thecurrentstudywaspoweredtoidentifycorrelationsofapproximately0.11(n=659;80%power;p<.05),whichisasmalleffectsize.Asaresult,weconsiderthecurrentfindingstobehighlyrobust.
Lastly,thecohortdesignandnarrowagerangeofthecurrentsampleprovidedalmosteliminatedtheeffectofchronologicalage,whichotherwisecouldhavebeenamajorconfounderofanyeffectsofthyroidhormonelevelsoncognition.Anumberofauthorshavesuggestedthatthemajorimpactofthyroidlevelsoncognitionoccursaboveakeythresholdof$80yearsofage(vandenBeldetal.,2005;Beginetal.,2008).Ifsuchathresholdexists,thenthecurrentsamplewouldbetooyoungforthemajoreffectsofthyroidfunctiontobeimpactingoncognitiveperfor-mance.However,giventheconsistencyofthecurrentresultswithpriorfindings,itissuggestedtheyareindicativeofalackofassociationwithinthenormalrangesofTSHandT4.
Roleoffundingsource
TheLBC1936isfundedbyAgeUK.TheCentreforCognitiveAgeingandCognitiveEpidemiologyreceivesfundingfromtheBBSRC,EPSRC,ESRC,andMRC.TBisfundedbyanMRCgrant.
Thyroidstimulatinghormone,freethyroxineandcognitiveabilityinoldage:TheLothianbirthcohortstudy1936
601
Conflictofinterest
Therearenoconflictsofinterest.
Ethicalconsiderations
EthicalpermissionfortheLBC1936studyprotocolwasobtainedfromtheMulti-CentreResearchEthicsCommitteeforScotlandandtheLothianResearchEthicsCommittee.AllresearchwascarriedoutincompliancewiththeHelsinkiDeclaration.
Acknowledgements
WethanktheScottishCouncilforResearchinEducationforallowingaccesstotheSMS1947.WethanktheLBC1936participants.Fordatacollectionandcollation,wethankAlanGow,CatherineMurray,JanieCorley,PaulRedmond,AlisonPattie,RossHenderson,CarolineBrett,MarthaWhite-man,MichelleTaylor,CarolineGrahamandPaulaDavies.TheLBC1936datawerecollectedbyaResearchIntoAgeingprogrammegrantandtheAgeUK-fundedDisconnectedMindproject.ThecurrentanalysiswasundertakenwithinTheUniversityofEdinburghCentreforCognitiveAgeingandCognitiveEpidemiology,partofthecrosscouncilLifelongHealthandWellbeingInitiative(G0700704/84698).FundingfromtheBBSRC,EPSRC,ESRC,andMRCisgratefullyacknowledged.
AppendixA.Supplementarydata
Supplementarydataassociatedwiththisarticlecanbefound,intheonlineversion,athttp://dx.doi.org/10.1016/j.psyneuen.2012.07.018.
References
Begin,M.E.,Langlois,M.F.,Lorrain,D.,Cunnane,S.C.,2008.Thyroid
functionandcognitionduringaging.Curr.Gerontol.Geriatr.Res.doi:10.1155/2008/474868.
Beydoun,M.A.,Beydoun,H.A.,Shroff,M.R.,Kitner-Triolo,M.H.,
Zonderman,A.B.,2012.Serumleptin,thyroxineandthyroid-stimulatinghormonelevelsinteracttoaffectcognitivefunctionamongUSadults:evidencefromalargerepresentativesurvey.Neurobiol.Aging33,1730—1743.
Deary,I.J.,2012.Humanintelligence.Annu.Rev.Psychol.63,453—482.Deary,I.J.,Gow,A.J.,Taylor,M.D.,etal.,2007.TheLothianbirth
cohort1936:astudytoexamineinfluencesoncognitiveageingfromage11toage70andbeyond.BMCGeriatr.7(December),28doi:10.1186/1471-2318-7-28.
Deary,I.J.,Gow,A.J.,Pattie,A.,Starr,J.M.,2011.Cohortprofile:
TheLothianBirthCohortsof1921and1936.Int.J.Epidemiol.doi:10.1093/ije/dyr197.
Folstein,M.F.,Folstein,S.E.,McHugh,P.R.,1975.Mini-mentalstate.
Apracticalmethodforgradingcognitivestateofpatientsforclinicians.J.Psychiatr.Res.12,189—198.
Gussekloo,J.,vanExel,E.,deCraen,A.J.M.,etal.,2004.Thyrid
status,disabilityandcognitivefunction,andsurvivalinoldage.JAMA292,2591—2599.
Hogervorst,E.,Huppert,F.,Matthews,F.E.,Brayne,C.,2008.Thy-roidfunctionandcognitivedeclineintheMRCCognitiveFunctionandAgeingStudy.Psychoneuroendocrinology33,1013—1022.Leitol,H.,Behrends,J.,Brabant,G.,2002.Thethyroidaxisin
ageing.NovartisFound.Symp.242,193—201.
Liesbeth,A.D.M.,vanOschHorgervorst,E.,Combrinck,M.,Smith,
A.D.,2004.Lowthyroid-stimulatinghormoneasanindependentriskfactorforAlzheimerdisease.Neurology62,1967—1971.Parle,J.,Roberts,L.,Wilson,S.,etal.,2010.Arandomized
controlledtrialoftheeffectofthyroxinereplacementoncogni-tivefunctionincommunity-livingelderlysubjectswithsubclinicalhypothyroidism:TheBirminghamElderlyThyroidStudy.J.Clin.Endocrinol.Metab.95,3623—3632.
Patterson,M.,Lonie,J.A.,Starr,J.M.,2010.Thyroidfunction,
cognition,functionalindependenceandbehaviouralandpsycho-logicalsymptomsofdementiainAlzheimer’sdisease.Int.J.Geriatr.Psychiatry25,1196—1199.
Roberts,L.M.,Pattison,H.,Roalfe,A.,etal.,2006.Issubclinical
thyroiddysfunctionintheelderlyassociatedwithdepressionorcognitivedysfunction?Ann.Intern.Med.145,573—581.
Samuels,M.H.,2010.Cognitivefunctioninsubclinicalhypothyroid-ism.J.Clin.Endocrinol.Metab.95,3611—3613.
Smith,J.W.,Evans,A.T.,Costall,B.,Smythe,J.W.,2002.Thyroid
hormones,brainfunctionandcognition:abriefreview.Neurosci.Biobehav.Rev.26,45—60.
St.John,J.A.,Henderson,V.W.,Gatto,N.M.,McCleary,C.A.,Spen-cer,C.A.,Hodis,H.N.,Mack,W.J.,2009.MildlyelevatedTSHandcognitioninmiddle-agedandolderadults.Thyroid19,111—117.vandenBeld,A.W.,Visser,T.J.,Feelders,R.A.,Grobbee,D.E.,
Lamberts,S.W.J.,2005.Thyroidhormoneconcentrations,dis-ease,physicalfunction,andmortalityinelderlymen.J.Clin.Endocrinol.Metab.90,6403—6409.
Wechsler,D.,1998a.WAIS-IIIUKAdministrationandScoringManual.
PsychologicalCorporation,London.
Wechsler,D.,1998b.WMS-IIIUKAdministrationandScoringManual.
PsychologicalCorporation,London.
Zigmond,A.S.,Snaith,R.P.,1983.Thehospitalanxietyanddepres-sionscale.ActaPsychiatr.Scand.67,361—370.
因篇幅问题不能全部显示，请点此查看更多更全内容