ARTICLE OPEN ACCESS CLASS OF EVIDENCE

Long-term Effects of Cholinesterase Inhibitorson Cognitive Decline and MortalityHong Xu MD PhD Sara Garcia-Ptacek MD PhD Linus Jonsson PhD Anders Wimo MD PhD

Peter Nordstrom MD PhD and Maria Eriksdotter MD PhD

Neurologyreg 202196e2220-e2230 doi101212WNL0000000000011832

Correspondence

Dr Xu

hongxu2kise

AbstractObjectiveTo investigate whether cholinesterase inhibitors (ChEIs) are associated with slower cognitivedecline in Alzheimer dementia and decreased risk of severe dementia or death

MethodsPatients with Alzheimer dementia from the Swedish Dementia Registry starting on ChEIswithin 3 months of the dementia diagnosis were included and compared to nontreated patientswith Alzheimer dementia In a propensity scorendashmatched cohort the association between ChEIuse and cognitive trajectories assessed by Mini-Mental State Examination (MMSE) scores wasexamined with a mixedmodel and severe dementia (MMSE score lt10) or death as an outcomewas assessed with Cox proportional hazards models

ResultsThe matched cohort included 11652 ChEI users and 5826 nonusers During an average of 5years of follow-up 255 cases developed severe dementia and 6055 (35) died ChEI use wasassociated with higher MMSE score at each visit (013 MMSE points per year 95 confidenceinterval [CI] 006ndash020) ChEI users had a 27 lower risk of death (073 95 CI 069ndash077)compared with nonusers Galantamine was associated with lower risk of death (071 95 CI065ndash076) and lower risk of severe dementia (069 95 CI 047ndash100) and had the strongesteffect on cognitive decline of all the ChEIs (018 MMSE points per year 95 CI 007ndash028)

ConclusionsChEIs are associated with cognitive benefits that are modest but persist over time and withreducedmortality risk which could be explained partly by their cognitive effects Galantaminewasthe only ChEI demonstrating a significant reduction in the risk of developing severe dementia

Classification of EvidenceThis study provides Class III evidence that for patients with Alzheimer dementia ChEIsdecrease long-term cognitive decline and risk of death and that galantamine decreases the risk ofsevere dementia

RELATED ARTICLE

Patient PageThe Effects of Long-termMedication Use inAlzheimer Disease

Page e2247

MORE ONLINE

Class of EvidenceCriteria for ratingtherapeutic and diagnosticstudies

NPuborgcoe

PodcastDr Jeff Burns speaks withProf Maria Eriksdotterabout long-term effects ofcholinesterase inhibitors oncognitive decline andmortality

NPuborgfqxxkt

These authors contributed equally to this work

From the Division of Clinical Geriatrics (HX SG-P ME) Division of Neurogeriatrics (LJ AW) Department of Neurobiology Care Sciences and Society and Department of MedicalEpidemiology and Biostatistics (HX) Karolinska Institutet Department of Internal Medicine (SG-P) Neurology Section Sodersjukhuset Stockholm Sweden H Lundbeck AS (LJ)Copenhagen Denmark Department of Community Medicine and Rehabilitation (PN) Geriatric Medicine Umearing University and Theme Aging (SG-P MA) Karolinska UniversityHospital Stockholm Sweden

Go to NeurologyorgN for full disclosures Funding information and disclosures deemed relevant by the authors if any are provided at the end of the article

The Article Processing Charge was funded by FORTE

This is an open access article distributed under the terms of the Creative Commons Attribution License 40 (CC BY) which permits unrestricted use distribution and reproduction in anymedium provided the original work is properly cited

e2220 Copyright copy 2021 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the American Academy of Neurology

The acetylcholinesterase inhibitors (ChEIs) and the NMDAreceptor antagonist memantine are hitherto the only specificpharmacologic treatments approved for Alzheimer dementiathe most common type of dementia1 Although their benefitappears to be modest23 a significant body of evidence sup-ports their effectiveness for improving cognition and theircost-effectiveness4-12

Degeneration of basal forebrain cholinergic neurons is one ofthe earliest findings in Alzheimer dementia and precedesdementia development1314 Progression of Alzheimer de-mentia is better correlated with cholinergic system dysfunc-tion than amyloid plaque load15 Reduction of the volume ofthe basal forebrain precedes changes of hippocampal volumeand predicts the cortical spread of Alzheimer pathology16

ChEIs work by maximizing the availability of endogenous ace-tylcholine in the brain17However few randomized clinical trials(RCTs) have examined the effectiveness of ChEIs in Alzheimerdementia after 1 year of treatment18-22 or followed up patientsbeyond this point20 Studies of long-term cognitive decline aredifficult due to high attrition and loss to follow-up20 Althoughnot RCTs follow-up of ChEI-treated Alzheimer dementia co-horts has shown small cognitive benefits at 2 3 and gt10years23-25 Moreover a positive short-term response to ChEIscan delay nursing home placement26 Other studies have shownassociations between ChEI use and decreased risk of myocardialinfarction stroke and death in patients with dementia27-30

We conducted a longitudinal cohort study on the SwedishDementia Registry (SveDem) to investigate whether thecognitive benefit of ChEIs in routine settings persists over thelong term and whether ChEI use is associated with decreasedrisk of severe dementia and death

MethodsStudy Design and Data SourceThe longitudinal cohort study includes patients with incidentdiagnosed dementia registered in the SveDem (svedemse)SveDem is a web-based registry established in 2007 with theaim of registering all patients with incident dementia inSweden and following them up annually123132 The baselineregistration in SveDem is initiated at the time of the dementiadiagnosis The majority of patients are diagnosed and thusregistered in a mild stage of dementia but some patients dohave an advanced dementia by the time they seek care and arethus diagnosed in a more advanced stage The registry storesdata on demographics cognitive evaluation by Mini-Mental

State Examination (MMSE) the type of dementia andpharmacologic management SveDem was merged with theNational Patient Registry to include diagnoses made in spe-cialist clinics and hospitals the Prescribed Drug Registry andthe Total Population and Causes of Death Registry

Standard Protocol Approvals Registrationsand Patient ConsentsThe regional human ethics committee in Stockholm approvedthe study (dnr 2017501-31) Patients are informed aboutregistration in SveDem at the time of their dementia di-agnosis The aim of SveDem is to improve dementia treat-ment and care Patients can refuse registration obtaininformation on their registration any time and withdrawconsent at a later date Any research project on SveDem datamust be approved by the ethics committee Signed consent forresearch however was not required for this study in accor-dance with the protocol submitted and approved by the ethicscommittee Data were deidentified by Swedish authoritiesbefore delivery to the research team

Study PopulationFrom 2007 to 2017 78346 patients with dementia were reg-istered in SveDem in these patients the most common de-mentia types were Alzheimer dementia (31) and mixedAlzheimer dementia (19) followed by unspecified dementia(23) vascular dementia (19) Lewy body dementia (2)frontotemporal dementia (2) Parkinson disease with de-mentia (2) and other (2) In the present study we includedall patients with incident diagnosed Alzheimer dementia ormixed Alzheimer dementia (n = 39196) We defined the studyinclusion date as the date of the dementia diagnosis in SveDemthe date when the patient started ChEI or the first date onwhich a dementia diagnosis appeared in the National PatientRegistry (whichever came first) We excluded patients if datawere missing on age sex diagnosis or MMSE score at baseline(n = 1533) We excluded patients with baseline MMSE scorelt10 (n = 789) because the indication for initiating ChEIs inSweden is mild to moderate Alzheimer dementia (MMSEscore ge10) Finally we excluded patients with a first pre-scription date of ChEI treatment gt3 months from the baselineMMSE (n = 5808) to ensure that the MMSE score was rep-resentative of the cognitive status at the start of ChEI treatmentand to avoid increasing confounding by indication (figure 1)The majority (934) started treatment with ChEIs on thesame day of the MMSE measurement or after while 765(66) patients were on ChEI treatment before the baselineMMSE test was performed In Sweden MMSE or a similarscreening test is recommended when diagnosing dementia butis not required for ChEI prescription33 A total 31054 patients

GlossaryChEI = cholinesterase inhibitor CI = confidence interval DDD = defined daily dose DOMINO-AD = Donepezil andMemantine in Moderate to Severe Alzheimerrsquos DiseaseHR = hazard ratio ICD-10 = International Classification of Diseases 10revision MMSE = Mini-Mental State Examination RCT = randomized clinical trial SveDem = Swedish Dementia Registry

NeurologyorgN Neurology | Volume 96 Number 17 | April 27 2021 e2221

with Alzheimer dementia were eligible comprising 21826ChEI users and 9228 nonusers

ChEI ExposureChEI treatment (donepezil rivastigmine and galantamine) wasdefined as ChEI initiation within 3 months of the dementia di-agnosis (ChEI use) (table e-1 doiorg105061dryad2jm63xsmz) Patients who started ChEI after this initial 3-monthperiod were excluded from analyses (figure 1) Nonusers weredefined as never being treated with ChEIs during the duration ofthe follow-up period In our main analysis the exposure wasassumed to be constant This was a conservative designattempting to limit the confounding that could occur if the pa-tientrsquos rate of cognitive decline influenced treatment status afterinitiation imitating the intention-to-treat design of clinical trials

We collected information on the doses of each dispensation ofChEIs over the initial 3-month period ChEI doses were expressedas the number of defined daily doses (DDDs) present in each

package or dispensation The DDD for each drug is set by theWorld Health Organization and is by definition the ldquoassumedaverage maintenance dose per day for a drug used for its mainindication in adultsrdquo34 To simplify comparisons we summed thetotal number of DDDs dispensed during the initial 3-month pe-riod and averaged this number out over the number of days toobtain the average DDD per day For reference the DDD ofdonepezil is 75 mg of rivastigmine is 9 mg (oral) or 95 mg(transdermal) and of galantamine is 16mg IfmultipleChEIswereused their DDDs were summated For example in clinical prac-tice wemay start donepezil at a low dose of 5 mg after 4 weeks iftolerated the dose is increased to 10 mg For such a patient theaverageDDDwould be 83mg after 3months ([5mg times 7days times 4weeks] + [10mg times 7days times 8 weeks])(7 days times 12 weeks)

CovariatesCovariates were defined at the date of study entry and in-cluded age sex comorbid conditions (hypertension diabeteshistory of myocardial infarction congestive heart failure

Figure 1 Flowchart of Included Patients

ChEI = cholinesterase inhibitors MMSE = Mini-Mental State Examination

e2222 Neurology | Volume 96 Number 17 | April 27 2021 NeurologyorgN

peripheral vascular disease cerebrovascular disease chronicpulmonary disease chronic kidney disease cancer atrial fi-brillation) and medications (angiotensin-converting enzymeinhibitors angiotensin receptor blockers β-blockers calciumchannel blocker statins antipsychotics antidepressants) Thedefinition of comorbid conditions was based on ICD-10code35 Diabetes and hypertension were in addition enrichedwith information on purchase of related medication up to 3years before study entry date through the use of the Ana-tomical Therapeutic Chemical code A10 for antidiabeticsThe ICD and Anatomical Therapeutic Chemical codes aredetailed in table e-1 (doiorg105061dryad2jm63xsmz)

OutcomesBaseline and follow-up MMSE scores were obtained from Sve-Dem Severe dementia was defined as MMSE score lt10 duringfollow-up36 The occurrence of death was obtained from theTotal Population and Causes of Death Registry Patients werefollowed up from study entry until the event of interest death orend of follow-up (October 16 2018) whichever occurred first

Data AnalysisPropensity score matching was performed to balance con-founders between ChEI users and nonusers Using logisticregression models we estimated the propensity score to re-ceive ChEI treatment based on age sex baseline MMSEscore comorbid conditions and medications We performed21 propensity score matching to pair each ChEI user to anonuser without replacement by the nearest number match-ing and with a caliper of 001

The cognition trajectories (MMSE score change) between ChEIusers and nonusers were estimated with mixed-effects repeated-measures models of unstructured-variance-covariance matrixwhich included data from all visits during the 5 years of follow-up The model adjusted for baseline cognition ChEI treatment(yesno) visit time (year) and a product of ChEI treatment andvisit time Visit time had a strong nonlinear (approximatelyquadratic) relation with MMSE score (p lt 005) so we alsoincluded a quadratic of time in the adjusted model to consider anonlinear trend in MMSE score across time Missing data werehandled with the use of multiple imputation with chainedequations in further sensitivity analysis We in addition consid-ered the potential effects of general attrition from those lost tofollow-up due to dropout or to the presence of a competing riskbefore the end of follow-up such as death We compared thecognitive trajectories between ChEI users and nonusers with 3approaches (1) mixed-effect model in real data (2) multipleimputations leveraging auxiliary variables associated withMMSE score missingness in conjunction with mixed-effectmodels and (3) inverse probability of censoring weighting inconjunction with multiple imputations and mixed-effect modelThis last method was developed specifically to manage attritionin the SveDem cohort37

Incidence rates per 1000 person-years with 95 confidenceintervals (CIs) were calculated for severe dementia risk

(MMSE score lt10) in patients who had gt1 MMSE mea-surements and death (all patients) Cox proportional hazardsregression was used to calculate hazard ratios (HRs) associ-ated with ChEI and each outcome In addition we performedsubgroup analyses by sex age categories type of dementiaand different comorbid conditions Furthermore we per-formed a competing-risk analysis for the outcomes of severedementia with death as the competing risk Wemodeled ChEIuse as a continuous exposure for increasing doses in a cubicspline with each outcome The analyses were also run sepa-rately on donepezil rivastigmine and galantamine

All analyses were performed with R (r-projectorg R Foun-dation for Statistical Computing Vienna Austria) and Stataversion 160 (StataCorp College Station TX)

Data AvailabilityRequests for access to the SveDem data should be addressed tothe registry holder and the steering committee (svedemse)

This study provides Class III evidence that for patients withAlzheimer dementia ChEIs decrease long-term cognitivedecline and death Galantamine in addition decreases the riskfor severe dementia

Supplementary data are available from Dryad (tables e-1ndashe-5and figures e-1 and e-2) doiorg105061dryad2jm63xsmz

ResultsBaseline CharacteristicsFigure 1 shows how the propensity scorendashmatched cohort wasassembled Table 1 and table e-2 show the baseline character-istics by ChEI treatment in patients with Alzheimer dementiabefore and after matching Patients who were not on ChEIswere older had lower MMSE score and more comorbid con-ditions such as cardiovascular disorders and took more medi-cations than treated patients (table e-2) Matching removedmany of the significant imbalances especially in age (stan-dardized difference decreased from 67 to 1) baselineMMSE score (46 to 0) hypertension (29 to 0) con-gestive heart failure (28 to 2) history of atrial fibrillation(28 to 0) and prescription of β-blocker (26 to 1)

The final propensity scorendashmatched cohort included 11652ChEI users and 5826 nonusers 62 were women with amean age of 812 plusmn 63 years Mean baseline MMSE score was212 plusmn 42 points and the most common comorbid conditionwas hypertension (74) followed by cerebrovascular disease(17) diabetes (16) atrial fibrillation (16) and cancer(16) Antihypertensive and lipid-modifying agents werehighly prescribed (table 1) The median time between thedementia diagnosis and ChEI prescription was 2 (inter-quartile range 0ndash10 range 0ndash90) days Among ChEI usersdonepezil accounted for 62 prescriptions followed by gal-antamine (21) and rivastigmine (17)

NeurologyorgN Neurology | Volume 96 Number 17 | April 27 2021 e2223

ChEI Use and Long-term Cognitive DeclineIn total 27199 measures of MMSE were available for analysisThe number ofMMSEmeasurements for each patient was 16 plusmn09 (range 1ndash7) 6802 (40) had gt1 MMSE measurementtaken during up to 5 years of follow-up At baseline the meanMMSE scores were 220 points in ChEI users and 219 points innonusers ChEI users presented with betterMMSE scores at anyvisit compared to nonusers (013 MMSE points change slope95 CI 006ndash020 for real cohort) The average yearly reductionin MMSE score was minus162 (95 CI minus170 to minus154) points for

users and nonusers combined These associations were consis-tent throughout several sensitivity analyses (tables 2 and 3)when estimating from raw data and when applying a multipleimputation on missing MMSE measurement and adjusting forthe inverse probability weighting for dropout during follow-upIndividual ChEI drugs showed an association with higher cog-nition at follow-up compared to nonusers with galantaminepresenting the largest effect size (018 MMSE points change95 CI 007ndash028) (table 2) There were no significant differ-ences among different ChEIs effects on cognition (ptrend gt 005)

Table 1 Baseline Characteristics of Patients With Alzheimer Dementia in the Propensity ScorendashMatched Cohort

No ChEIa

(n = 5826)ChEIa

(n = 11652)StandardizedDifferenceb p Value

Diagnosis 1 049

Alzheimer dementia n () 3355 (576) 6774 (581)

Mixed dementia n () 2471 (424) 4878 (419)

Age mean (SD) y 811 plusmn 67 812 plusmn 61 1 038

Female n () 3593 (617) 7191 (617) 0 096

MMSE baseline score mean (SD) 212 plusmn 43 212 plusmn 41 0 087

Comorbid conditions n ()

Hypertension 4306 (739) 8599 (738) 0 087

Diabetes 913 (157) 1818 (156) 0 091

Myocardial infarction 584 (100) 1140 (98) 1 062

Congestive heart failure 551 (95) 1036 (89) 2 022

Peripheral vascular disease 288 (49) 538 (46) 2 034

Cerebrovascular disease 981 (168) 2024 (174) 1 038

Chronic obstructive pulmonary disease 500 (86) 1001 (86) 0 098

Renal disease 151 (26) 266 (23) 2 021

Cancer 941 (162) 1810 (155) 2 029

Atrial fibrillation 933 (160) 1865 (160) 0 099

Medication n ()

ACEisARBs 2613 (449) 5230 (449) 0 097

β-Blocking agents 2470 (424) 4884 (419) 1 054

Calcium channel blockers 1756 (301) 3477 (298) 1 068

Lipid-modifying agents 2270 (390) 4561 (391) 0 082

Antipsychotics 254 (44) 541 (46) 1 040

Antidepressants 1899 (326) 3839 (329) 1 064

Propensity score probability receivingChEIs treatment mean (SD)

07 plusmn 01 07 plusmn 01 0 100

Abbreviations ACEi = angiotensin-converting enzyme inhibitor ARB = angiotensin receptor blocker ChEI = cholinesterase inhibitors MMSE = Mini-MentalState Examinationa Cohort propensity score matched for dementia diagnosis age sex MMSE baseline measurement comorbidity (hypertension diabetes myocardial in-farction congestive heart failure peripheral vascular disease cerebrovascular disease chronic obstructive pulmonary disease renal disease cancer andatrial fibrillation) and medications (use of ACEiARB β-blocking agents calcium channel blockers lipid-modifying agents antipsychotics andantidepressants)b Standardized difference is calculated by dividing the mean by the SD of the difference between treated and untreated groups

e2224 Neurology | Volume 96 Number 17 | April 27 2021 NeurologyorgN

Analyses stratified by Alzheimer dementia and mixed Alz-heimer dementia diagnosis did not find any marked differencein the effects of ChEI between the Alzheimer dementia groupand the mixed dementia group (014 [95 CI 005ndash024] vs012 [95 CI 001ndash022] MMSE points change in the Alz-heimer dementia and mixed Alzheimerrsquos dementia grouprespectively) (table e-3 doiorg105061dryad2jm63xsmz)Patients who had lower cognition at the time of diagnosis(MMSE score lt20) had benefits similar to those with higherMMSE scores (017 [95 CI 006ndash028] vs 010 [95 CI001ndash018] MMSE points change in the groups with MMSEscore lt20 and ge20 respectively) (table e-3)

Figure 2A shows the dose-response effects of ChEI on cognitionHigher dispensed doses of ChEI were associated with higherMMSE measurements during follow-up This association wasgenerally observed throughout the whole range of doses con-sidered with a modest but significant effect size When analyseson separate ChEIs were conducted dose response of donepezilyielded results similar to the overall ChEI exposure (figure e-1Adoiorg105061dryad2jm63xsmz) Patients taking galant-amine showed improved MMSE scores at follow-up and thiswas significant when dispensed doses were ge16 mgd (DDDge1) Although rivastigmine use was significantly associated withimproved cognition (shown in table 2) a dose-response effectcould not be demonstrated (figure e-1B and e-1C)

ChEI and Severe Dementia RiskWhen severe dementia (MMSE score lt10) was considered asthe outcome there were 6802 patients (40 of the initialpopulation) who had at least 2 MMSE measurements Theincidence rates and the proportion of patients developing se-vere dementia were higher among the nonusers (incidence rate1221000 person-years 40 of all patients for nonusers vs1021000 person-years 37 of all patients for ChEI usersseparately) When stratified for separate ChEIs only galant-amine users had a statistically significant lower risk of severedementia (HR 069 95 CI 047ndash100) which was not sig-nificant for users of rivastigmine or donepezil or for ChEI usersas a whole (table 3) Stratified analyses for the risk of severedementia are presented in figure e-2A (doiorg105061dryad2jm63xsmz) Similar associations were observed for severedementia with the competing-risk model (table e-4)

ChEI and MortalityDuring an average of 5 years of follow-up corresponding to52042 person-years 6055 (35) patients died The overallmortality rate was asymp2 times higher (115021000 person-years) compared with the age- and year-matched generalSwedish population (60591000 person-years ) (table e-5doiorg105061dryad2jm63xsmz) Lower mortality ratewas noted for ChEI users compared with nonusers (105781000 person-years vs 136931000 person-years) ChEI

Table 2 Mixed Model Output of Estimated Cognition Trajectories Among ChEI Users and Nonusers

MMSE Changes Unit

PS-Matched CohortPS-Matched Cohort + MultipleImputation

PS-Matched Cohort + MultipleImputation + Inverse ProbabilityWeighting

Coefficient 95 CI Coefficient 95 CI Coefficient 95 CI

ChEI user compared to nonusers

ChEI 013c 006 to 020 015c 008 to 023 021c 014 to 027

Follow-up time minus162c minus170 to minus154 minus160c minus166 to minus153 minus168c minus187 to minus149

Stratification by different ChEI

No ChEI Ref Ref Ref

Donepezil 014c 006 to 021 014c 006 to 021 018c 011 to 024

Rivastigmine 006 minus005 to 017 005 minus006 to 016 010a 001 to 019

Galantamine 018c 007 to 028 017b 007 to 027 024c 013 to 034

p Trend for different ChEIs 019 020 008

MMSE change per 1 y of follow-up minus163c minus171 to minus155 minus163c minus171 to minus155 minus168c minus186 to minus150

Abbreviations ChEI = cholinesterase inhibitor CI = confidence interval MMSE = Mini-Mental State Examination PS = propensity score Ref = referentCohort matched for PS with dementia diagnosis age sex MMSE baseline measurement comorbidity (hypertension diabetes myocardial infarctioncongestive heart failure peripheral vascular disease cerebrovascular disease chronic obstructive pulmonary disease renal disease cancer and atrialfibrillation) and medications (use of angiotensin-converting enzyme inhibitorsangiotensin receptor blockers β-blocking agents calcium channel blockerslipid-modifying agents antipsychotics and antidepressants) The mixed model included ChEI treatment visit time (week as class effect) ChEI treatment byweek and the baselineMMSE score as a covariate with an unstructured covariancematrix within ChEIs treatment group for a repeated-measures covariancestructurea p lt 005b p lt 001c p lt 0001

NeurologyorgN Neurology | Volume 96 Number 17 | April 27 2021 e2225

users had a 27 lower risk of death (HR 073 95 CI069ndash077) compared with nonusers We observed significantdifferences among different ChEIs in regard to mortality risk(ptrend lt 005) with an HR for galantamine of 071 (95 CI065ndash076) for donepezil of 078 (95 CI 074ndash083) and forrivastigmine of 086 (95 CI 080ndash093) (table 4)

There was a graded association between dispensed averageDDD of ChEIs and risk of death (figure 2B) The HRs as-sociated with increasing average DDDs of donepezil riva-stigmine and galantamine are graphically represented infigure e-1D through e-1F (doiorg105061dryad2jm63xsmz) Patients who took higher dose of ChEI had

lower mortality risk in a dose-dependent response Patientstaking galantamine had a lower risk of death compared withuntreated patients (ratio lt1) (figure e-1F)

Patients taking galantamine at any dose gt75 mg donepezil(figure e-1D) or gt9 to 95 mg rivastigmine (figure e-1E) had alower risk of death compared with untreated patients (ratio lt1)(figure e-1F)

The associations between ChEI and death were consistentthroughout all subgroup analyses (figure e-2B) albeit ham-pered in patients with peripheral vascular disease by lack ofpower No interactions were observed between subgroup andChEI in predicting mortality (all p gt 005)

DiscussionIn this large longitudinal national dementia cohort ChEI usewas associated with a reduction in cognitive decline over timeand this effect was modest but persisted over the long termChEI use also was associated with reduced risk for mortalitywhich is in line with previous results from our group2728 Adose-response effect was observed for both of these outcomesOnly galantamine demonstrated a reduction in the risk forsevere dementia (MMSE score lt10)

Little is known about the long-term effects of ChEI on cog-nitive decline in Alzheimer dementia In a 2018 Cochranesystematic review which included 30 studies on donepezil forAlzheimer dementia only 3 studies18-20 had a follow-up of 1year and only 1 study18 could be included in the meta-analysisA recent meta-analysis (n = 16576 patients with Alzheimerdementia 63 RCTs) with an average duration of 8 monthsshowed that although ChEIs had a benefit for cognition theeffect did not reach significant improvement38 In the BritishAlzheimerrsquos Disease 2000 study patients with Alzheimer de-mentia were randomized to donepezil or placebo for 2 yearsDonepezil users had 08 points higher MMSE scores at 2 yearsbut the study was underpowered and hard to interpret20 Inanother study was performed in 5 Northern European coun-tries a significant advantage of donepezil treatment over pla-cebo was observed at 52 weeks in cognition activities of dailyliving and the Progressive Deterioration Scale18

The findings of our study showing significantly slower cog-nitive decline in patients with ChEI use are in line with resultsfrom other clinical trials However the magnitude of the effectappears to be somewhat smaller in our study which is prob-ably related to the characteristics of our cohort the particulardesign and limitations First we defined medication exposurewithin the 3 months after the dementia diagnosis a conser-vative design that intended to mimic the intention-to-treatdesign of clinical trials and to avoid reverse causation in whichthe rate of cognitive decline would cause changes in medi-cation status Patients who were defined as treated could havestopped taking ChEIs and would still be analyzed within the

Figure 2 Dose Response of ChEIs Using Cubic Splines With(A) MMSE Change and (B) All-Cause Death RiskCompared With Nonuse

Dose-response effect of increasing average cumulative daily dose of cho-linesterase inhibitor (ChEI) compared to nonuse of ChEI average (solid) and95 confidence interval (dash lines) Horizontal axis represents the numberof standard defined daily doses (DDDs) that patients took per day Forclarity the DDD for eachmedication is shown with a vertical dotted line Forexample a patient taking galantamine 8mgd would be taking half the DDDof galantamine which would be represented at the 05 point of the hori-zontal axis In panel A y-axis represents Mini-Mental State Examination(MMSE) score change In panel B y-axis represents adjusted hazard ratio fordeath Model included ChEI treatment visit time (class effect) interaction ofChEI treatment by visit and baselineMMSE score Referencewas set at DDD0 Cohort was matched for dementia diagnosis age sex MMSE baselinescore comorbidity (hypertension diabetes myocardial infarction conges-tive heart failure peripheral vascular disease cerebrovascular diseasechronic obstructive pulmonary disease renal disease cancer and atrial fi-brillation) and medications (angiotensin-converting enzyme inhibitorsan-giotensin receptor blockers β-blocking agents calcium channel blockerslipid-modifying agents antipsychotics and antidepressants)

e2226 Neurology | Volume 96 Number 17 | April 27 2021 NeurologyorgN

treatment group which could naturally attenuate the differ-ence in effect This could not be the case for nonusers pa-tients who were initially nonusers who started treatment after3 months were excluded from analyses and did not contributefollow-up time in either group In general patients are regis-tered in SveDem early in the disease process31 In our studypatients with a baseline MMSE score gt20 declined faster thanpatients who started with lowerMMSE scores Second ours isa cohort study using patients from the SveDem and natural-istic follow-ups Attrition to follow-up was high at asymp50Third we defined study inclusion at the date of dementiadiagnosis or the date when the patient started ChEI (which-ever came first) for this reason a small proportion of patients

(66 in our study) were already treated when theMMSEwasperformed which would contribute to reduce the magnitudeof the difference between users and nonusers

Patients with faster cognitive decline may have higher likeli-hood of dropout because institutionalization mortality andmanagement of the social aspects of advanced dementiaprobably dominate the care efforts and disrupt follow-upMissing follow-ups would then be more frequent among se-verely impaired patients This creates a situation in whichpatients who are followed up are more likely to start withbetter cognition and to decline less over time The magnitudeof the yearly decline observed in our study was minus162 points

Table 3 Estimated of Cognition Trajectories Among ChEI Users and Nonusers

Time of Measurement

No ChEI ChEI Use Difference

Estimated MMSEScore (95 CI)

ReductionFrom Baseline

Estimated MMSEScore (95 CI)

ReductionFrom Baseline

AbsoluteDifference (95 CI)

Estimation from raw data

Baseline 2190 (2196ndash2238) mdash 2203 (2188ndash2219) mdash 013 (006 to 020)

1 y 2038 (2020ndash2055) minus152 2057 (2041ndash2074) minus146 019 (011 to 027)

2 y 1885 (1864ndash1906) minus305 1911 (1893ndash1929) minus292 026 (012 to 039)

3 y 1733 (1706ndash1759) minus457 1765 (1744ndash1786) minus438 032 (012 to 052)

4 y 1580 (1548ndash1612) minus610 1618 (1594ndash1643) minus585 038 (012 to 065)

5 y 1427 (1389ndash1466) minus763 1472 (1444ndash1500) minus731 045 (011 to 078)

Estimation from imputation

Baseline 2190 (2196ndash2239) mdash 2205 (2188ndash2222) mdash 015 (008 to 023)

1 y 2038 (2019ndash2056) minus152 2057 (2040ndash2075) minus148 020 (011 to 028)

2 y 1885 (1864ndash1906) minus305 1909 (1890ndash1928) minus296 024 (012 to 036)

3 y 1733 (1708ndash1758) minus457 1761 (1740ndash1781) minus444 028 (011 to 045)

4 y 1580 (1551ndash1610) minus610 1612 (1589ndash1635) minus593 032 (009 to 055)

5 y 1428 (1393ndash1463) minus762 1464 (1438ndash1490) minus741 036 (008 to 065)

Estimation from imputation + IPCW

Baseline 2190 (2196ndash2237) mdash 2211 (2194ndash2228) mdash 021 (014 to 027)

1 y 2031 (2007ndash2055) minus159 2049 (2029ndash2069) minus162 018 (003 to 033)

2 y 1872 (1834ndash1910) minus318 1887 (1861ndash1913) minus324 015 (minus017 to 047)

3 y 1713 (1659ndash1767) minus477 1725 (1691ndash1759) minus486 012 (minus038 to 061)

4 y 1554 (1484ndash1625) minus636 1563 (1520ndash1606) minus648 009 (minus058 to 075)

5 y 1395 (1308ndash1483) minus795 1401 (1349ndash1453) minus810 006 (minus079 to 090)

Abbreviations ChEI = cholinesterase inhibitor CI = confidence interval IPCW = inverse probability of censoring weighting MMSE = Mini-Mental StateExaminationCohort matched for propensity score with dementia diagnosis age sex MMSE baseline measurement comorbidity (hypertension diabetes myocardialinfarction congestive heart failure peripheral vascular disease cerebrovascular disease chronic obstructive pulmonary disease renal disease cancer andatrial fibrillation) and medications (use of angiotensin-converting enzyme inhibitorsangiotensin receptor blockers β-blocking agents calcium channelblockers lipid-modifying agents antipsychotics and antidepressants) The mixed model included ChEI treatment visit time (week as class effect) ChEItreatment by week and the baseline MMSE score as a covariate with an unstructured covariance matrix within ChEI treatment group for a repeated-measures covariance structure

NeurologyorgN Neurology | Volume 96 Number 17 | April 27 2021 e2227

per year which is in line with previous clinical trials In theNational Alzheimerrsquos Coordinating Center study the averagedecline was 19 points in the first year and 15 in the secondyear of follow-up39 In the 1-year randomized placebo-con-trolled study of patients with mild to moderate AD18 theplacebo group declined asymp2 points per year compared to adecline of asymp05 points per year for the donepezil usersMeanwhile in our study the benefit observed with ChEIs wassmaller although our conservative definition of medicationexposure makes it hard to directly compare these results

ChEIs have been proven to have symptomatic effects in Alz-heimer dementia beyond those detected by standard measuresof cognition As previously shown by our group ChEIs havebeen associated with reductions in myocardial infarction28

stroke27 and mortality27 while anticholinergic medicationshave been associated with increases in stroke and mortalityrisk40 Better cognition may in itself be protective for mortalitybut ChEIs also may have beneficial systemic effects2728 Strokeand mortality prevention in mild to moderate dementia stagesis desirable and stroke prevention could theoretically prolongindependent functioning in dementia41-44 Meanwhile a recentstudy showed that initiation of antipsychotic treatment was

reduced in those treated with ChEIs45 In our study the benefitof ChEI was similar in those diagnosed withMMSE scores lt20and those with higherMMSE scores and the cognitive benefitspersisted over time In the Donepezil and Memantine inModerate to Severe Alzheimerrsquos Disease (DOMINO-AD)clinical trial withdrawal of donepezil in patients with moderateto severe Alzheimer dementia increased the risk of nursinghome placement during 12 months after treatment46 Con-tinuation was associated with better cognition47 In additiongalantamine use was associated with reduced risk for severedementia and mortality and had the largest effect size for theassociationwith cognitive decline in our study Galantamine is arapidly reversible ChEI and the only ChEI that acts as anallosteric nicotinic modulator4849 This dual effect as an ace-tylcholinesterase inhibitor and nicotinic receptors modulatormay explain its enhanced effects

The strengths of our study are the large sample size and longcognitive follow-up In addition data were obtained throughstandard patient registration and thus reflect real-world dataWe acknowledge some limitations First regarding the obser-vational study design we cannot infer causality and we ac-knowledge the possibility of residual and unknown

Table 4 Incidence Rate and Hazard Ratios for Severe Dementia and Mortality During Follow-up

Events n Person-Time y IR per 1000 py HR (95 CI) p Value

Severe dementia

No ChEIa (n = 1433) 58 4754 1220 (943ndash1578) Ref

ChEIa (n = 5369) 197 19383 1016 (884ndash1169) 084 (063ndash113) 024

Separate ChEIs

Donepezil (n = 3131) 107 10692 1001 (828ndash1209) 079 (059ndash107) 013

Rivastigmine (n = 883) 45 3244 1387 (1036ndash1858) 124 (087ndash177) 024

Galantamine (n = 1355) 45 5446 826 (617ndash1107) 069 (047ndash100) 005

p Trend for different ChEIs 001

Death

No ChEI (n = 5826) 2139 15621 13693 (13125ndash14286) Ref

ChEI (n = 11652) 3916 37021 10578 (10252ndash10914) 073 (069ndash077) lt0001

Separate ChEIs

Donepezil (n = 7181) 2238 21252 10531 (10103ndash10976) 078 (074ndash083) lt0001

Rivastigmine (n = 1997) 746 6623 11263 (10483ndash12101) 086 (080ndash093) lt0001

Galantamine (n = 2474) 932 9145 10191 (9558ndash10867) 071 (065ndash076) lt0001

p Trend for different ChEIs 001

Abbreviations ChEI = cholinesterase inhibitor CI = confidence interval HR = hazard ratio IR = incidence rate py =person-year Ref = referentCohort matched for dementia diagnosis age sex Mini-Mental State Examination baseline measurement comorbidity (hypertension diabetes myocardialinfarction congestive heart failure peripheral vascular disease cerebrovascular disease chronic obstructive pulmonary disease renal disease cancer andatrial fibrillation) and medications (use of angiotensin-converting enzyme inhibitorsangiotensin receptor blockers β-blocking agents calcium channelblockers lipid-modifying agents antipsychotics and antidepressants)a Selected those with at least 2 MMSE measurements from matched cohort

e2228 Neurology | Volume 96 Number 17 | April 27 2021 NeurologyorgN

confounding However we controlled for the unbalancedconfounders in our propensity scorendashmatching cohort Secondpatients were considered exposed throughout the wholefollow-up period according to treatment status at study entryWe attempted to mimic the intention-to-treat design of clinicaltrials to ensure a conservative estimate of the effects of ChEI oncognition and because of the fear of reverse causality in whichthe speed of cognitive decline could influence decisions to startor withdraw treatment Third we acknowledge that the vas-cular pathology in the group with mixed Alzheimer dementiamay have affected the response to ChEI on the cognition tra-jectories If anything this may have contributed to an un-derestimation of the effects of the ChEI presented in this studyIn addition individual patient information on reasons forprescription or side effects of ChEI and information on thedispensation form of the ChEI were not available Fourth thenational coverage of SveDem is not absolute and there is noclear count of how many patients develop dementia each yearin Sweden On the basis of different approximations of de-mentia incidence and prevalence the coverage of SveDem fornew dementia cases is estimated to be between 30 and 43depending on different estimations of incident cases regardlessof whether they receive a diagnosis50 The dementia diagnosticworkup follows standard clinical practice and few patients havea changed dementia diagnosis at follow-up31 which suggestsadequate diagnostic accuracy Fifth although the majority ofpatients are diagnosed early in the dementia disease processsome patients are diagnosed in a later stage resulting in avariation of cognitive functioning at the time of the initiation ofChEI treatment Finally because our data were collected inreal-world clinical practice there were differences in thenumber of MMSE measurements performed between indi-viduals and these MMSE measurements were not missing atrandom however we attempted to address the potentialconcern of dropout by adjusting the estimates using inverseprobability of censoring weighting Results with and withoutadjusted weighting were similar and robust

ChEIs are associated with cognitive benefits that are modestbut persist over the long term ChEIs are associated withreduced mortality risk which may be partly explained by themodest cognitive effects Galantamine was the only ChEI thatdemonstrated a significant reduction in the risk of developingsevere dementia in addition to presenting the strongest effecton cognition

AcknowledgmentThe authors are grateful to the SveDem (svedemse) forproviding data for this study They thank all patientscaregivers reporting units and coordinators in SveDem aswell as the SveDem steering committee

Study FundingSveDem is supported financially by the Swedish Associationsof Local Authorities and Regions This study was supportedby the regional agreement on medical training and clinicalresearch between the Stockholm county council and the

Karolinska Institutet (ALF) Swedish medical researchcouncil grant 2016-02317 FORTE grant 2017-01646Johanniterorden i SverigeSwedish Order of St John andSwedish Society for Medical Research H Xu is supported bya postdoctoral grant from StratNeuro (the Strategic ResearchArea NeurosciencendashKarolinska Institutet Umearing Universityand KTH) S Garcia-Ptacek is supported by a postdoctoralfellowship from the Swedish Society for Medical Research

DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgN for full disclosures

Publication HistoryReceived by Neurology April 8 2020 Accepted in final formJanuary 4 2021

References1 Nichols E Szoeke CEI Vollset SE et al Global regional and national burden of

Alzheimerrsquos disease and other dementias 1990-2016 a systematic analysis for theGlobal Burden of Disease Study 2016 Lancet Neurol 20191888ndash106

2 Birks JS Harvey RJ Donepezil for dementia due to Alzheimerrsquos disease CochraneDatabase Syst Rev 20186CD001190

3 Birks JS Grimley Evans J Rivastigmine for Alzheimerrsquos disease Cochrane databaseSyst Rev 2015Cd001191

4 Jelic V Winblad B Alzheimer disease Donepezil and nursing home placementbenefits and costs Nat Rev Neurol 20161211ndash13

5 Winblad B Black SE Homma A et al Donepezil treatment in severe Alzheimerrsquosdisease a pooled analysis of three clinical trials Curr Med Res Opin 2009252577ndash2587

6 Winblad B Kilander L Eriksson S et al Donepezil in patients with severe Alzheimerrsquosdisease double-blind parallel-group placebo-controlled study Lancet 20063671057ndash1065

Appendix Authors

Name Location Contribution

Hong Xu MDPhD

KarolinskaInstitutetStockholmSweden

Design and conceptualized studyanalyzed the data performed thestatistical analysis drafted themanuscript for intellectualcontent

Sara Garcia-Ptacek MDPhD

KarolinskaInstitutetStockholmSweden

Design and conceptualized studymajor role in the acquisition ofdata performed the statisticalanalysis contributed to draftingthe manuscript revised themanuscript for intellectualcontent

LinusJonsson MDPhD

KarolinskaInstitutetStockholmSweden

Interpreted the data revised themanuscript for intellectualcontent

AndersWimo MDPhD

KarolinskaInstitutetStockholmSweden

Interpreted the data revised themanuscript for intellectualcontent

PeterNordstromMD PhD

Umearing UniversitySweden

Interpreted the data revised themanuscript for intellectualcontent

MariaEriksdotterMD PhD

KarolinskaInstitutetStockholmSweden

Design and conceptualized studyacquisition of data fundingrevised the manuscript forintellectual content director ofSveDem

NeurologyorgN Neurology | Volume 96 Number 17 | April 27 2021 e2229

7 Schmidt R Hofer E Bouwman FH et al EFNS-ENSEAN guideline on concomitantuse of cholinesterase inhibitors and memantine in moderate to severe Alzheimerrsquosdisease Eur J Neurol 201522889ndash898

8 Tariot PN Farlow MR Grossberg GT et al Memantine treatment in patients withmoderate to severe Alzheimer disease already receiving donepezil a randomizedcontrolled trial JAMA 2004291317ndash324

9 Porsteinsson AP Grossberg GT Mintzer J Olin JT Memantine MEM-MD-12 StudyGroup Memantine treatment in patients with mild to moderate Alzheimerrsquos diseasealready receiving a cholinesterase inhibitor a randomized double-blind placebo-controlled trial Curr Alzheimer Res 2008583ndash89

10 Howard R McShane R Lindesay J et al Donepezil and memantine for moderate-to-severe Alzheimerrsquos disease N Engl J Med 2012366893ndash903

11 Grossberg GT Manes F Allegri RF et al The safety tolerability and efficacy ofonce-daily memantine (28 mg) a multinational randomized double-blind placebo-controlled trial in patients with moderate-to-severe Alzheimerrsquos disease taking cho-linesterase inhibitors CNS Drugs 201327469ndash478

12 Religa D Fereshtehnejad SM Cermakova P et al SveDem the Swedish DementiaRegistry a tool for improving the quality of diagnostics treatment and care of de-mentia patients in clinical practice PLoS One 201510e0116538

13 Mesulam M The cholinergic lesion of Alzheimerrsquos disease pivotal factor or sideshow Learn Mem 20041143ndash49

14 Bartus RT On neurodegenerative diseases models and treatment strategies lessonslearned and lessons forgotten a generation following the cholinergic hypothesis ExpNeurol 2000163495ndash529

15 Turnbull MT Boskovic Z Coulson EJ Acute down-regulation of BDNF signalingdoes not replicate exacerbated amyloid-beta levels and cognitive impairment inducedby cholinergic basal forebrain lesion Front Mol Neurosci 20181151

16 Teipel SJ Cavedo E Hampel H Grothe MJ Basal forebrain volume but not hip-pocampal volume is a predictor of global cognitive decline in patients with Alz-heimerrsquos disease treated with cholinesterase inhibitors Front Neurol 20189642

17 Li DD Zhang YH ZhangW Zhao P Meta-analysis of randomized controlled trials onthe efficacy and safety of donepezil galantamine rivastigmine and memantine for thetreatment of Alzheimerrsquos disease Front Neurosci 201913472

18 Winblad B Engedal K SoininenH et al A 1-year randomized placebo-controlled studyof donepezil in patients with mild to moderate AD Neurology 200157489ndash495

19 Mohs RC Doody RS Morris JC et al A 1-year placebo-controlled preservation offunction survival study of donepezil in AD patients Neurology 200157481ndash488

20 Courtney C Farrell D Gray R et al Long-term donepezil treatment in 565 patients withAlzheimerrsquos disease (AD2000) randomised double-blind trial Lancet 20043632105ndash2115

21 Karaman Y Erdogan F Koseoglu E Turan T Ersoy AO A 12-month study of theefficacy of rivastigmine in patients with advanced moderate Alzheimerrsquos diseaseDemen Geriatr Cogn Disord 20051951ndash56

22 FarlowM Anand R Messina J Jr Hartman R Veach J A 52-week study of the efficacyof rivastigmine in patients with mild to moderately severe Alzheimerrsquos disease EurNeurol 200044236-241

23 Vellas B Hausner L Frolich L et al Progression of Alzheimer disease in Europe datafrom the European ICTUS study Curr Alzheimer Res 20129902ndash912

24 Wallin AK Andreasen N Eriksson S et al Donepezil in Alzheimerrsquos disease what toexpect after 3 years of treatment in a routine clinical setting Demen Geriatr CognDisord 200723150ndash160

25 Wattmo C Londos E Minthon L Longitudinal associations between survival inAlzheimerrsquos disease and cholinesterase inhibitor use progression and community-based services Demen Geriatr Cogn Disord 201540297ndash310

26 Wattmo C Londos E Minthon L Short-term response to cholinesterase inhibitors inAlzheimerrsquos disease delays time to nursing home placement Curr Alzheimer Res201815905ndash916

27 Tan ECK Johnell K Garcia-Ptacek S et al Acetylcholinesterase inhibitors and risk ofstroke and death in people with dementia Alzheimers Dement 201814944ndash951

28 Nordstrom P Religa DWimo AWinblad B Eriksdotter M The use of cholinesteraseinhibitors and the risk of myocardial infarction and death a nationwide cohort studyin subjects with Alzheimerrsquos disease Eur Heart J 2013342585ndash2591

29 Lin YT Wu PH Chen CS Yang YH Yang YH Association between acetylcholin-esterase inhibitors and risk of stroke in patients with dementia Sci Rep 2016629266

30 Wattmo C Londos E Minthon L Response to cholinesterase inhibitors affects life-span in Alzheimerrsquos disease BMC Neurol 201414173

31 Garcia-Ptacek S Farahmand B Kareholt I Religa D Cuadrado ML Eriksdotter MMortality risk after dementia diagnosis by dementia type and underlying factors acohort of 15209 patients based on the Swedish Dementia Registry J Alzheimers Dis201441467ndash477

32 Garcia-Ptacek S Modeer IN Kareholt I et al Differences in diagnostic processtreatment and social Support for Alzheimerrsquos dementia between primary andspecialist care results from the Swedish Dementia Registry Age Ageing 201746314ndash319

33 National Board of Health and Welfare Nationella riktlinjer for varingrd och omsorg viddemenssjukdom stod for styrning och ledning (National guidelines for health andsocial care of dementia diseases support for organization and governance) Aringtta45Tryckeri AB 2017

34 World Health Organization Defined Daily Dose Essential Medicines and HealthProducts ATCDDD Toolkit[online] Available at whointmedicinesregulationmedicines-safetytoolkit_ddden Accessed April 8 2020

35 Ludvigsson JF Andersson E Ekbom A et al External review and validation of theSwedish national inpatient register BMC Public Health 201111450

36 National Institute for Health and Care Excellence National Institute for Health andCare Excellence Clinical Guidelines Dementia Assessment Management andSupport for People Living With Dementia and Their Carers National Institute forHealth and Care Excellence 2018

37 Handels R Jonsson L Garcia-Ptacek S Eriksdotter M Wimo A Controlling forselective dropout in longitudinal dementia data application to the SveDem registryAlzheimers Dement 202016789ndash796

38 Blanco-Silvente L Castells X Garre-Olmo J et al Study of the strength of theevidence and the redundancy of the research on pharmacological treatment for Alz-heimerrsquos disease a cumulative meta-analysis and trial sequential analysis Eur J ClinPharmacol 2019751659ndash1667

39 Haaksma ML Calderon-Larranaga A Olde Rikkert MGM Melis RJF Leoutsakos JSCognitive and functional progression in Alzheimer disease a prediction model oflatent classes Int J Geriatr Psychiatry 2018331057ndash1064

40 Tan ECK Eriksdotter M Garcia-Ptacek S Fastbom J Johnell K Anticholinergicburden and risk of stroke and death in people with different types of dementiaJ Alzheimers Dis 201865589ndash596

41 Subic A Cermakova P Norrving B et al Management of acute ischaemic stroke inpatients with dementia J Intern Med 2017281348ndash364

42 Subic A Zupanic E von Euler M et al Stroke as a cause of death in death certificatesof patients with dementia a cohort study from the Swedish Dementia Registry CurrAlzheimer Res 2018151322ndash1330

43 Garcia-Ptacek S Contreras Escamez B Zupanic E et al Prestrokemobility and dementiaas predictors of stroke outcomes in patients over 65 years of age a cohort study from theSwedish dementia and stroke registries J Am Med Dir Assoc 201819154ndash161

44 Subic A Cermakova P Religa D et al Treatment of atrial fibrillation in patients withdementia a cohort study from the Swedish Dementia Registry J Alzheimers Dis 2018611119ndash1128

45 Tan ECK Johnell K Bell JS et al Do acetylcholinesterase inhibitors prevent or delaypsychotropic prescribing in people with dementia Analyses of the Swedish DementiaRegistry Am J Geriatr Psychiatry 202028108ndash117

46 Howard R McShane R Lindesay J et al Nursing home placement in the Donepeziland Memantine in Moderate to Severe Alzheimerrsquos Disease (DOMINO-AD) trialsecondary and post-hoc analyses Lancet Neurol 2015141171ndash1181

47 Knapp M King D Romeo R et al Cost-effectiveness of Donepezil and Memantine inModerate to Severe Alzheimerrsquos disease (the DOMINO-AD trial) Int J GeriatrPsychiatry 2017321205ndash1216

48 Darreh-Shori T Soininen H Effects of cholinesterase inhibitors on the activities andprotein levels of cholinesterases in the cerebrospinal fluid of patients with Alzheimerrsquosdisease a review of recent clinical studies Curr Alzheimer Res 2010767ndash73

49 Ago Y Koda K Takuma K Matsuda T Pharmacological aspects of the acetylcho-linesterase inhibitor galantamine J Pharmacol Sci 20111166ndash17

50 SveDem Svenska Demensregistret-Aringrsrapport 2018 (Swedish Dementia Registry-Yearly Report 2018) SveDem 2018

e2230 Neurology | Volume 96 Number 17 | April 27 2021 NeurologyorgN

DOI 101212WNL0000000000011832202196e2220-e2230 Published Online before print March 19 2021Neurology

Hong Xu Sara Garcia-Ptacek Linus Joumlnsson et al Long-term Effects of Cholinesterase Inhibitors on Cognitive Decline and Mortality

This information is current as of March 19 2021

ServicesUpdated Information amp

httpnneurologyorgcontent9617e2220fullincluding high resolution figures can be found at

References httpnneurologyorgcontent9617e2220fullref-list-1

This article cites 45 articles 3 of which you can access for free at

Citations httpnneurologyorgcontent9617e2220fullotherarticles

This article has been cited by 1 HighWire-hosted articles

Subspecialty Collections

httpnneurologyorgcgicollectioncohort_studiesCohort studies

httpnneurologyorgcgicollectionclass_iiiClass III

httpnneurologyorgcgicollectionalzheimers_diseaseAlzheimers diseasefollowing collection(s) This article along with others on similar topics appears in the

Permissions amp Licensing

httpwwwneurologyorgaboutabout_the_journalpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in

Reprints

httpnneurologyorgsubscribersadvertiseInformation about ordering reprints can be found online

ISSN 0028-3878 Online ISSN 1526-632XWolters Kluwer Health Inc on behalf of the American Academy of Neurology All rights reserved Print1951 it is now a weekly with 48 issues per year Copyright Copyright copy 2021 The Author(s) Published by

reg is the official journal of the American Academy of Neurology Published continuously sinceNeurology