Delirium Predicts 12-Month Mortality - JAMA Network

Among patients with dementia, there was a weak, nonsignificant effect of delirium on survival. After adjustment for covariates, ... DeliriumPredicts12-MonthMortality|DementiaandCognitiveImpairment|JAMAInternalMedicine|JAMANetwork Ourwebsiteusescookiestoenhanceyourexperience.Bycontinuingtouseoursite,orclicking"Continue,"youareagreeingtoourCookiePolicy | Continue [SkiptoNavigation] fulltexticon FullText contentsicon Contents figureicon Figures/Tables multimediaicon Multimedia attachicon SupplementalContent referencesicon References relatedicon Related commentsicon Comments DownloadPDF TopofArticle Abstract Methods Results Comment ArticleInformation References Figure1. ViewLargeDownloadUnadjustedKaplan-Meiersurvivalcurvesofthe12-monthmortalityratebystudygroup.Figure2. ViewLargeDownloadAssessingchangesovertimeintheadjustedeffectofdeliriumonmortality.Thesolidlineshowsthelogarithmofthehazardratio(logHR)expressingrelativerisksinthedeliriumcohortcomparedwithcontrolsadjustedforallcovariatesincludedinthemultivariablemodelsummarizedinTable2asafunctionoffollow-uptime.Thedottedcurvesrepresentpointwise95%confidenceintervalsaroundtheestimatedlogHR.Thefactthattheestimateisanalmostconstantfunctionoftimeindicatesthattheprognosticabilityofthebaselinedeliriumremainsstableover12monthsoffollow-up. Table1. ViewLargeDownloadSelectedCharacteristicsofCohortsatEnrollment*Table2. ViewLargeDownloadResultsofProportionalHazardsAnalysesof1-YearMortality*Table3. ViewLargeDownloadResultsofProportionalHazardsAnalysisof1-YearMortalityintheDeliriumCohort* 1.Cole  MPrimeau  F Prognosisofdeliriuminelderlyhospitalpatients.  CMAJ.1993;14941- 46GoogleScholar2.Rabins  PVFolstein  MF Deliriumanddementia:diagnosticcriteriaandfatalityrates.  BrJPsychiatry.1982;140149- 153GoogleScholarCrossref3.Pompei  PForeman  MRudberg  MAInouye  SKBraund  VCassel  CK Deliriuminhospitalizedolderpersons:outcomesandpredictors.  JAmGeriatrSoc.1994;42809- 815GoogleScholar4.Inouye  SRushing  JForeman  MPalmer  RPompei  P Doesdeliriumcontributetopoorhospitaloutcomes?athree-siteepidemiologicstudy.  JGenInternMed.1998;13234- 242GoogleScholarCrossref5.Francis  JMartin  DKapoor  WN Aprospectivestudyofdeliriuminhospitalizedelderly.  JAMA.1990;2631097- 1101GoogleScholarCrossref6.Francis  JKapoor  WN Prognosisafterhospitaldischargeofoldermedicalpatientswithdelirium.  JAmGeriatrSoc.1992;40601- 606GoogleScholar7.O'Keeffe  SLavan  J Theprognosticsignificanceofdeliriuminolderhospitalpatients.  JAmGeriatrSoc.1997;45174- 178GoogleScholar8.Rockwood  KCosway  SCarver  DJarrett  PStadnyk  KFisk  J Theriskofdementiaanddeathafterdelirium.  AgeAgeing.1999;28551- 556GoogleScholarCrossref9.Pfeiffer  E Ashortportablementalstatusquestionnairefortheassessmentoforganicbraindeficitinelderlypatients.  JAmGeriatrSoc.1975;23433- 441GoogleScholar10.Inouye  SVanDyck  CAlessi  CBalkin  SSiegal  AHorwitz  R Clarifyingconfusion:theconfusionassessmentmethod—anewmethodfordetectionofdelirium.  AnnInternMed.1990;113941- 948GoogleScholarCrossref11.Lewis  LMMiller  DKMorley  JENork  MJLasater  LC UnrecognizeddeliriuminEDgeriatricpatients.  AmJEmergMed.1995;13142- 145GoogleScholarCrossref12.Jorm  A AshortformoftheInformantQuestionnaireonCognitiveDeclineintheElderly(IQCODE):developmentandcross-validation.  PsycholMed.1994;24145- 153GoogleScholarCrossref13.Jorm  FABroe  AGCreasey  H  etal.  FurtherdataonthevalidityoftheInformantQuestionnaireonCognitiveDeclineintheElderly(IQCODE).  IntJGeriatrPsychiatry.1996;11131- 139GoogleScholarCrossref14.Jorm  AChristensen  HHenderson  AJacomb  PAKorten  AMackinnon  A Informantratingsofcognitivedeclineofelderlypeople:relationshiptolongitudinalchangeoncognitivetests.  AgeAgeing.1996;25125- 129GoogleScholarCrossref15.Law  SWolfson  C ValidationofaFrenchversionofaninformant-basedquestionnaireasascreeningtestforAlzheimer'sdisease.  BrJPsychiatry.1995;167541- 544GoogleScholarCrossref16.McCusker  JCole  MBellavance  FPrimeau  F Reliabilityandvalidityofanewmeasureofseverityofdelirium.  IntPsychogeriatr.1998;10421- 433GoogleScholarCrossref17.Fillenbaum  G MultidimensionalFunctionalAssessment:TheOARSMethodology—AManual. Durham,NCDukeUniversityCenterfortheStudyofAgingandHumanDevelopment1978;18.Charlson  MPompei  PAles  KMacKenzie  R Anewmethodofclassifyingprognosticcomorbidityinlongitudinalstudies:developmentandvalidation.  JChronicDis.1987;40373- 383GoogleScholarCrossref19.Knaus  WDraper  EWagner  DZimmerman  J APACHEII:aseverityofdiseaseclassificationsystem.  CritCareMed.1985;13818- 829GoogleScholarCrossref20.Charlson  MSax  FMacKenzie  RFields  SBraham  RDouglas  R Assessingillnessseverity:doesclinicaljudgmentwork?  JChronicDis.1986;39439- 452GoogleScholarCrossref21.Inouye  SViscoli  CHorwitz  RHurst  LTinetti  M Apredictivemodelfordeliriuminhospitalizedelderlymedicalpatientsbasedonadmissioncharacteristics.  AnnInternMed.1993;119474- 481GoogleScholarCrossref22.Cox  D Regressionmodelsandlifetables(withdiscussion).  JRStatSoc.1972;34187- 220GoogleScholar23.Abrahamowicz  MMacKenzie  TEsdaile  J Time-dependentHR:modelingandhypothesistestingwithapplicationinlupusarthritis.  JAmStatAssoc.1996;811432- 1439GoogleScholarCrossref24.Pompei  PForeman  MRudberg  MAInouye  SKBraund  VCassel  CK Deliriuminhospitalizedolderpersons:outcomesandpredictors.  JAmGeriatrSoc.1994;42809- 815GoogleScholar25.Thomas  RICameron  DJFahs  MC Aprospectivestudyofdeliriumandprolongedhospitalstay:exploratorystudy.  ArchGenPsychiatry.1988;45937- 940GoogleScholarCrossref26.Cole  MMcCusker  JBellavance  F  etal.  Randomizedtrialofsystematicdetectionandtreatmentofdeliriuminolderhospitalizedmedicalpatients.  CMAJ.GoogleScholar SeeMoreAbout DementiaandCognitiveImpairmentGeriatricsNeurology SelectYourInterests SelectYourInterests CustomizeyourJAMANetworkexperiencebyselectingoneormoretopicsfromthelistbelow. AcidBase,Electrolytes,Fluids AddictionMedicine AllergyandClinicalImmunology Anesthesiology Anticoagulation ArtandImagesinPsychiatry BleedingandTransfusion Cardiology CaringfortheCriticallyIllPatient ChallengesinClinicalElectrocardiography ClinicalChallenge ClinicalDecisionSupport ClinicalImplicationsofBasicNeuroscience ClinicalPharmacyandPharmacology ComplementaryandAlternativeMedicine ConsensusStatements Coronavirus(COVID-19) CriticalCareMedicine CulturalCompetency DentalMedicine Dermatology DiabetesandEndocrinology DiagnosticTestInterpretation Diversity,Equity,andInclusion DrugDevelopment ElectronicHealthRecords EmergencyMedicine EndofLife EnvironmentalHealth Ethics FacialPlasticSurgery GastroenterologyandHepatology GeneticsandGenomics GenomicsandPrecisionHealth Geriatrics GlobalHealth GuidetoStatisticsandMedicine Guidelines HairDisorders HealthCareDeliveryModels HealthCareEconomics,Insurance,Payment HealthCareQuality HealthCareReform HealthCareSafety HealthCareWorkforce HealthDisparities HealthInequities HealthInformatics HealthPolicy Hematology HistoryofMedicine Humanities Hypertension ImagesinNeurology ImplementationScience InfectiousDiseases InnovationsinHealthCareDelivery JAMAInfographic LawandMedicine LeadingChange LessisMore LGBTQ LifestyleBehaviors MedicalCoding MedicalDevicesandEquipment MedicalEducation MedicalEducationandTraining MedicalJournalsandPublishing Melanoma MobileHealthandTelemedicine NarrativeMedicine Nephrology Neurology NeuroscienceandPsychiatry NotableNotes Nursing Nutrition Nutrition,Obesity,Exercise Obesity ObstetricsandGynecology OccupationalHealth Oncology OphthalmicImages Ophthalmology Orthopedics Otolaryngology PainMedicine PathologyandLaboratoryMedicine PatientCare PatientInformation Pediatrics PerformanceImprovement PerformanceMeasures PerioperativeCareandConsultation Pharmacoeconomics Pharmacoepidemiology Pharmacogenetics PharmacyandClinicalPharmacology PhysicalMedicineandRehabilitation PhysicalTherapy PhysicianLeadership Poetry PopulationHealth PreventiveMedicine ProfessionalWell-being Professionalism PsychiatryandBehavioralHealth PublicHealth PulmonaryMedicine Radiology RegulatoryAgencies Research,Methods,Statistics Resuscitation Rheumatology RiskManagement ScientificDiscoveryandtheFutureofMedicine SharedDecisionMakingandCommunication SleepMedicine SportsMedicine StemCellTransplantation Surgery SurgicalInnovation SurgicalPearls TeachableMoment TechnologyandFinance TheArtofJAMA TheArtsandMedicine TheRationalClinicalExamination Tobaccoande-Cigarettes Toxicology TraumaandInjury TreatmentAdherence Ultrasonography Urology Users'GuidetotheMedicalLiterature Vaccination VenousThromboembolism VeteransHealth Violence Women'sHealth WorkflowandProcess WoundCare,Infection,Healing SavePreferences PrivacyPolicy|TermsofUse OthersAlsoLiked ThisIssue Citations 419 ViewMetrics DownloadPDF Twitter Facebook More LinkedIn Cite This Citation McCuskerJ,ColeM,AbrahamowiczM,PrimeauF,BelzileE.DeliriumPredicts12-MonthMortality.ArchInternMed.2002;162(4):457–463.doi:10.1001/archinte.162.4.457 Downloadcitationfile: Ris(Zotero) EndNote BibTex Medlars ProCite RefWorks ReferenceManager Mendeley ©2021 Permissions OriginalInvestigation February 25,2002 DeliriumPredicts12-MonthMortality JaneMcCusker,MD,DrPH;MartinCole,MD;MichalAbrahamowicz,PhD;etal FrancoisPrimeau,MD;EricBelzile,MSc AuthorAffiliations ArticleInformation FromtheDepartmentsofClinicalEpidemiologyandCommunityStudies(DrMcCuskerandMrBelzile)andPsychiatry(DrsColeandPrimeau),StMary'sHospitalCenter;EpidemiologyandBiostatistics(DrsMcCuskerandAbrahamowicz)andPsychiatry(DrsColeandPrimeau),McGillUniversity;andtheDivisionofClinicalEpidemiology,MontrealGeneralHospital(DrAbrahamowicz),Montreal,Quebec. ArchInternMed.2002;162(4):457-463.doi:10.1001/archinte.162.4.457 visualabstracticon VisualAbstract editorialcommenticon EditorialComment relatedarticlesicon RelatedArticles authorinterviewicon Interviews multimediaicon Multimedia Abstract Background  Deliriumhasnotbeenfoundtobeasignificantpredictorofpostdischargemortality,butpreviousresearchhasmethodologiclimitationsincludingsmallsamplesizesandinadequatecontrolofconfounding.Thisstudyaimedtodeterminetheindependenteffectsofpresenceofdelirium,typeofdelirium(incidentvsprevalent),andseverityofdeliriumsymptomson12-monthmortalityamongoldermedicalinpatients.Methods  Aprospective,observationalstudyof2cohortsofmedicalinpatientswasconductedwithpatients65yearsorolder:243patientshadprevalentorincidentdelirium,and118controlshadnodelirium.Baselinemeasuresincludedpresenceofdeliriumand/ordementia,severityofdeliriumsymptoms,physicalfunction,comorbidity,andphysiologicalandclinicalseverityofillness.Mortalityduringthe12monthsafterenrollmentwasanalyzedwiththeCoxproportionalhazardsmodelwithadjustmentforcovariates.Results  Theunadjustedhazardratioofdeliriumwithmortalitywas3.44(95%confidenceinterval,2.05-5.75);theadjustedhazardratiowas2.11(95%confidenceinterval,1.18-3.77).Theeffectofdeliriumwassustainedovertheentire12-monthperiodafteradjustmentforcovariatesandwasstrongeramongpatientswithoutdementia.Amongpatientswithdementia,therewasaweak,nonsignificanteffectofdeliriumonsurvival.Afteradjustmentforcovariates,mortalitydidnotdifferbetweenpatientswithincidentandprevalentdelirium,butamongpatientswithdeliriumwithoutdementia,greaterseverityofdeliriumsymptomswasassociatedwithhighermortality.Conclusions  Deliriumisanindependentmarkerforincreasedmortalityamongoldermedicalinpatientsduringthe12monthsafterhospitaladmission.Itisaparticularlyimportantprognosticmarkeramongpatientswithoutdementia. DELIRIUMisafrequentphenomenonamongolderhospitalizedpatientsandhasbeenfoundtoberelatedtoseveraladverseoutcomes,includingalongermeanlengthofhospitalstay,poorfunctionalstatusandneedforinstitutionalcare,andmortality.1Withrespecttomortality,theevidenceisnotconsistent2;controlledstudieshavereportedthatdeliriumisassociatedwithincreasedin-hospitalmortality.2,3However,Inouyeetal4controlledforage,sex,dementia,illnessseverity,andfunctionalstatusandfoundnosignificantelevationinin-hospitalor3-monthmortality.Severalstudieswithupto2yearsoffollow-upreportednosignificantincreaseinpostdischargemortality.3,5-7Arecentstudywithamedianfollow-upof32.5monthsreportedahazardratio(HR)of1.71(95%confidenceinterval[CI],1.02-2.87)afteradjustmentforcomorbidity,dementia,frailty,age,sex,maritalstatus,andinstitutionalresidence.8Nevertheless,thesestudieshaveanumberofmethodologiclimitations,includingsmallsamplesizes,often-limitedfollow-up,andinadequatecontrolofconfoundingfactorssuchasdementia,comorbidity,andseverityofillness.Furthermore,itisnotknownwhethersurvivaldependsontheseverityofthedelirium,oronwhetherthedeliriumisdiagnosedatadmission(prevalentdelirium)orafteradmission(incidentdelirium).Theformerisrelatedtofactorsprecedinghospitaladmission,whereasthelattermaybeduetoaspectsofthecarereceivedinthehospital.Also,littleisknownaboutwhethertheadverseconsequencesofdeliriumaresimilaramongdementedandnondementedpatients. Weundertookthisstudytodeterminetheprognosticeffectofdeliriumontheoutcomeofolderhospitalmedicalinpatientsduringthe12monthsafteradmission.Inthisarticle,wereporttheindependenteffectofdelirium,adjustedforimportantconfoundingvariables,on12-monthmortalityandexaminetheeffectsonmortalityoftypeofdelirium(incidentvsprevalent),severityofdeliriumsymptoms,andpresenceofdementia. Methods Thisprospective,observational,cohortstudywasconductedata400-bed,university-affiliated,primaryacutecarehospitalinMontreal,Quebec.Wecompared6-and12-monthoutcomesin2cohorts:adeliriumcohortwithprevalentorincidentdeliriumdetectedduringthefirstweekofhospitalization,andacontrolcohortwithoutdelirium.Thestudywasconductedsimultaneouslywitharandomizedcontrolledtrialofthedetectionandtreatmentofdelirium,andasubgroupofthedeliriumcohortalsoparticipatedinthetrial. Enrollmentofsubjects Astudynursewasresponsibleforpatientscreeningandenrollmentinthe2studies.Onlypatients65yearsorolderwhowereadmittedfromtheemergencydepartmenttothemedicalserviceswereincludedinthestudies.Weexcludedpatientswithaprimarydiagnosisofstroke,thoseadmittedtotheoncologyunit,thosewhospokeneitherEnglishnorFrench,andthoseadmittedtotheintensivecareunitorcardiacmonitoringunitunlesstheyweretransferredtoamedicalwardwithin48hoursofadmission.Atenrollmentandduringthefirstweekofhospitalization,thenursescreenedeligiblepatientsfordeliriumusingtheShortPortableMentalStatusQuestionnaire(SPMSQ),a10-itemquestionnairethatevaluatesorientation,memory,andconcentration,9andreviewofthenursingnotes.ThenurseconductedtheConfusionAssessmentMethod(CAM)10interviewwithsubjectswhomade3ormoreerrorsontheinitialSPMSQ(indicatingmoderatetoseverecognitiveimpairment),subjectswhoseSPMSQscoresincreasedbyatleast1errorfromthefirstassessment,andsubjectswhosenursingnotesindicatedpossiblesymptomsofdelirium.(TheCAMisastructuredinterviewthatassesses9symptomdomainsofdeliriumspecifiedintheDiagnosticandStatisticalManualofMentalDisorders,RevisedThirdEdition.10) Prevalentdeliriumwasdiagnosedifthecriteriaforprobableordefinitedelirium11weremetatenrollment;incidentdeliriumwasdiagnosedifthecriteriawerenotmetuntilafterenrollment.Controlswereselectedfrompatientswhowerescreenedfordeliriumandfoundnottohaveit.Tobalancethedistributionsofageandpriorcognitiveimpairmentamongpatientswithdeliriumandcontrols,thesamplingmethodtookintoaccounteachpatient'sageandinitialSPMSQscore.Thus,controlswereselectedfrompatients70yearsandolder,andpatientswithSPMSQscoresof3ormorewereoversampled. Subjectswithfewerthan5errorsontheSPMSQgaveinformedconsenttoparticipateinthestudy;thosewith5ormoreerrorsassentedtoparticipation,andarelativeprovidedwrittenconsent.Thestudieswereapprovedbythehospital'sresearchethicscommittee. Datacollectionandmeasures Patientswereassessedatenrollmentbyaresearchassistant,blindtostudygroup,whoalsointerviewedafamilymember.Dateofdeathduringfollow-upwasascertainedbytheresearchassistant,whoobservedpatientsatleastweeklyduringtheirhospitalstay,at8weeksafterdischarge,andat6and12monthsafterenrollment.Otherbaselinedatawerecollectedbychartreviewbyanurseabstracter,blindtostudygroup. Dementiawasassessedfromthe16-itemInformantQuestionnaireonCognitiveDeclineintheElderly(IQCODE),12whichhashighinternalconsistencyandtest-retestreliability12-14inbothitsoriginal32-itemformandinits16-itemshort-form.Cutoffpointsof3.38,12,13and3.6orhigher15havebeenused;weusedanintermediatecutoffofhigherthan3.5todefinedementia. WeassessedtheseverityofdeliriumsymptomswiththeDeliriumIndex(DI)16basedsolelyonpatientobservation,withoutadditionalinformationfromfamilymembers,nursingstaff,orthepatient'smedicalchart.Only7of9symptomdomainsassessedontheCAM(disordersofattention,thought,consciousness,orientation,memory,perception,andpsychomotoractivity;acuteonsetandsleep-wakedisturbancewereexcluded)wereratedona4-pointscale(0,absent;1,mild;2,moderate;3,severe);theminimumandmaximumpossiblescores,therefore,were0(nosymptoms)and21(maximumseverity),respectively.TheDIhassatisfactoryinterraterreliabilityandconcurrentcriterionvalidity.17 WeadministeredtheInstrumentalActivitiesofDailyLiving(IADL)questionnairefromtheOlderAmericanResourcesandServices(OARS)project17toaninformantandusedittoassesspremorbidfunction(priortothecurrentillnessbutnotmorethan1monthbeforehospitaladmission).Thescalescorerangeis0(completelydependent)to16(completelyindependent). Threemeasuresofillnessburdenandseveritywereused.ComorbidityatadmissionwasassessedbychartreviewusingtheCharlsonComorbidityIndex,aweightedindexthattakesintoaccountthenumberandseverityofcomorbidconditions.18AcutephysiologicseverityofillnesswasassessedwiththeAcutePhysiologyScore,derivedfromtheAPACHEIIscale.19Theindexwascodedbychartreviewbasedonlaboratoryandclinicalmeasuresmadeonorbeforethedateofenrollment.Clinicalseverityofillnesswasassessedbytheresearchnurseatenrollment.20,21Thescoresrangedfrom1(minimal)to9(mostsevere). TheInternationalClassificationofDiseases,NinthRevision,codesforprimarydischargediagnoseswereobtainedfromthehospitaladministrativedatabase.Demographicvariables(age,sex,maritalstatus,andresidence)wererecordedinstudybaselineforms. Statisticalmethods Analysesofmortalityfocusedon2mainquestions:(1)theroleofdeliriumasanindependentprognosticfactorfordeath;and(2)identificationofprognosticfactorsformortalityinthedeliriumcohort.First,the2cohortswerecomparedwithrespecttothebaselinedistributionofvariousprognosticriskfactors,usingtheindependentgroupsttestandχ2testforquantitativeandcategoricalvariables,respectively.Toassesstheimpactofdeliriumonmortality,survivalanalyticaltechniqueswereusedtocomparesurvivalratesindeliriumandcontrolcohorts.Time0wasdefinedasthestudyenrollment,andsubjectswerecensoredatthetimeoflosstofollow-uporattheendofthe12-monthfollow-upperiod,whicheveroccurredearlier.Weusedtheexponentialmodelforthesurvivaltimedistributiontoestimatetheyearlymortalityratesseparatelyforeachofthe2cohorts.UnadjustedanalysisreliedonthecomparisonoftheKaplan-MeiersurvivalcurvesandonthescoretestintheunivariateCoxproportionalhazardsmodel,whichisequivalenttothelog-ranktest.22 Toadjusttheestimatedeffectofdeliriumonmortalityforthepossibledifferencesinthedistributionofotherriskfactorsinthe2cohorts,weusedthemultivariableCoxproportionalhazardsmodelwiththefollowingcovariatesselectedapriori:dementia,comorbidity,clinicalseverity,AcutePhysiologyScore,admittingservice(medicinevsgeriatrics),anddemographicvariables. InourprimaryanalysiswedidnotadjustforpremorbidIADLbecausethismeasureisaffectedbythepresenceofdementia,avariableofinterestinthisstudy.Inoursecondaryanalysis(notreportedhere)adjustingalsoforIADL,wefoundthat,asexpected,theeffectofdeliriumwasessentiallyunchanged,whereastheeffectofdementiawasslightlysmaller.Wealsoconductedasecondaryanalysis(notreportedhere)inwhichweevaluatedtheeffectofconsideringprimarydischargediagnosisasanadditionalcovariate,groupedintothe13categoriesshowninTable1.Theinclusionofdiagnosishadnoeffectonthemagnitudeoftheeffectofdeliriumonsurvival. Theproportionalhazardsassumptionwasverifiedusingtheregressionsplinemodel–basedlikelihoodratiotest.23Thisallowedustoformallytestwhethertheprognosticvalueofaninitialdiagnosisofdeliriumchangedduringthe12-monthfollow-up.TheimportanceofsuchpotentialchangeswasthenassessedbasedonagraphrepresentingthevariationofthelogarithmoftheHRforthedeliriumvscontrolgroupasafunctionofthefollow-upduration.Inaddition,tofurtherassesswhethertheassociationbetweendeliriumdiagnosisatbaselineandmortalitychangedwithincreasingfollow-upduration,separateanalyseswerecarriedoutfor3timeintervals:frominceptiontotheendofthefirstmonth;fromthesecondmonththroughthesixthmonth;andfromtheseventhmonththroughthe12thmonth.Ineachcase,theanalysiswasrestrictedtosubjectsaliveatthebeginningoftherespectiveinterval;andsubjectswhodidnotdieuntiltheendoftheintervalwerecensoredatthattime.Usingasimilarapproach,separateanalyseswerealsocarriedoutforwithin-hospitalandpostdischargemortality. Toassesswhethertheimpactofdeliriumonmortalitydependedonsomeotherpatientcharacteristic(s),weevaluatedfirst-orderinteractionsbetweendeliriumandeachofthecovariatesbyforcingallthecovariatesintothemultivariableCoxmodelandthenselectingstatisticallysignificantinteractionsthroughforwardselection.CutoffforentryintothemodelwasP<.10. finally similarmethodswereusedtoidentifyprognosticfactorsformortalityinthedeliriumcohort.themainfocusofthean alysiswasontheassessmentoftheroleofdeliriumtypeandseverity.specifically results duringthestudyenrollmentperiod atenrollment thecohortsdidnotdifferwithrespecttoage impactofdeliriumonmortality duringthe12-monthfollow-up table2liststheunivariateandmultivariateproportionalhazardsmodelsfor12-monthmortality.theunadjustedas sociationofdeliriumwithmortalitywasverystrong thepresenceofdementia mortalitywasexaminedbytimeperiodusing2differenttimegroupings.first theformaltestofthechangesovertimeintheeffectofdeliriumonmortalityyieldedadefinitelynonsignificantres ult someothervariablesdifferedintheirassociationswithin-hospitalvspostdischargemortality.thecharlsoncomo rbidityindexwasasignificantpredictorofmortalityinbothtimeperiods survivalanalysiswithinthedeliriumcohort weexaminedthefollowingprognosticfactorsformortalitywithinthedeliriumcohort:definitevsprobabledeliriu m comment theresultsofthisstudyindicateasignificantlyhigher12-monthmortalityrateamongmedicalinpatientsdiagnose dwithdeliriumthanforcontrolswithoutdelirium ourstudyprovidesnewevidenceoftheimportanceofdeliriumasaprognosticindicatorformortality.previousresea rchhasfoundanassociationonlywithin-hospitalmortality ofinterestistheobservationthatall3measuresofdiseaseburdenandseverityusedinthisstudyweresignificantin dependentpredictorsofmortality.allthesemeasuresareknowntopredictmortality deliriumduringhospitalizationseemstobeastrong threereasonsforthisdiscrepancycanbeconsidered.first theresultsofthisstudyhaveimplicationsforthecareofoldermedicalinpatientsandforresearchinthispopulatio n.first acceptedforpublicationjuly2 thisresearchwassupportedbygrantsfromthemedicalresearchcouncilofcanada correspondingauthorandreprints:janemccusker references x . accessyoursubscriptions signin accessthroughyourinstitution addorchangeinstitution freeaccesstonewlypublishedarticles createafreepersonalaccount toregisterforemailalertswithlinkstofreefull-textarticles purchaseaccess subscribetojournal getfulljournalaccessfor1year buyarticle getunlimitedaccessandaprintablepdf rentarticle rentthisarticlefromdeepdyve accesstofreearticlepdfdownloads saveyoursearch subscribenow customizeyourinterests createapersonalaccountorsigninto: registerforemailalertswithlinkstofreefull-textarticles accesspdfsoffreearticles manageyourinterests savesearchesandreceivesearchalerts privacypolicy makeacomment>