memantine

MemantineA Review of its Use in Moderate to Severe Alzheimer’s Disease

Kate McKeage

Adis, a Wolters Kluwer Business, Auckland, New Zealand

Various sections of the manuscript reviewed by:D. Adamis, Research and Academic Institute of Athens, Athens, Greece; A. Atri, Department of Neurologyand Massachusetts Alzheimer’s Disease Research Center, Massachusetts General Hospital, Boston,Massachusetts, USA; S. Gauthier, McGill Center for Studies in Aging, Douglas Hospital, Verdun, Quebec,Canada; H. Kavirajan, Department of Psychiatry and Biobehavioral Sciences, University of California, LosAngeles, California, USA; M.C. Woodward, Aged and Residential Care Services, Heidelberg RepatriationHospital, Heidelberg, Victoria, Australia.

Data Selection

Sources:Medical literature published in any language since 1980 on ‘memantine’, identified usingMEDLINE and EMBASE, supplementedby AdisBase (a proprietary database). Additional references were identified from the reference lists of published articles. Bibliographicalinformation, including contributory unpublished data, was also requested from the company developing the drug.

Search strategy: MEDLINE, EMBASE and AdisBase search terms were ‘memantine’ and ‘Alzheimer’s disease’. Searches were lastupdated 21 August 2009.

Selection: Studies in patients with moderate to severe Alzheimer’s disease who received memantine. Inclusion of studies was basedmainly on the methods section of the trials. When available, large, well controlled trials with appropriate statistical methodology werepreferred. Relevant pharmacodynamic and pharmacokinetic data are also included.

Index terms: Memantine, Alzheimer’s disease, dementia, pharmacodynamics, pharmacokinetics, therapeutic use, tolerability.

Contents

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8821. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8832. Pharmacodynamic Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 885

2.1 Mechanism of Action and Preclinical Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8852.2 Studies in Patients with Alzheimer’s Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 885

3. Pharmacokinetic Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8863.1 Absorption and Distribution. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8863.2 Metabolism and Elimination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8863.3 Special Populations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8863.4 Drug Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 886

4. Therapeutic Efficacy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8874.1 Monotherapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 887

4.1.1 Double-Blind Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8874.1.2 Extension Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 888

4.2 In Combination with an Acetylcholinesterase Inhibitor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8884.3 Meta-Analyses and Pooled Analyses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8894.4 Clinical Practice Setting. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 891

5. Pharmacoeconomic Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8916. Tolerability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8927. Dosage and Administration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8938. Place of Memantine in the Management of Moderate to Severe Alzheimer’s Disease . . . . . . . . . . . 894

ADIS DRUG EVALUATIONCNS Drugs 2009; 23 (10): 881-897

1172-7047/09/0010-0881/$49.95/0

ª 2009 Adis Data Information BV. All rights reserved.

SummaryAbstract Memantine is an uncompetitive, moderate-affinity NMDA receptor antagonist

that is indicated for the treatment of moderate to severe Alzheimer’s disease. In

well designed trials in patients with moderate to severe Alzheimer’s disease, oral

memantine monotherapy improved outcomes in the area of functional ability

more than placebo in one trial, but in a second trial, treatment differences did not

reach significance. Memantine has a distinct mode of action compared with that

of acetylcholinesterase (AChE) inhibitors, and in a well designed study, combi-

nation therapy with memantine plus donepezil improved outcomes more than

donepezil plus placebo in all four domains (function, cognition, behaviour and

global change).

Memantine is generally well tolerated, with adverse events occurring with a

similar incidence to that reported with placebo. In modelled cost-effectiveness

analyses, memantine was dominant to no therapy in regard to cost per quality-

adjusted life-year (QALY) gained, and the combination of memantine plus

donepezil was dominant to donepezil therapy alone in regard to QALYs

gained when treatment periods exceeded 1 year in patients with moderate to

severe disease. Thus, in the management of patients with moderate to severe

Alzheimer’s disease, memantine provides an effective treatment option. To

date, clinical trial support is greater for memantine use in combination with

an AChE inhibitor, while more data are needed to confirm its efficacy as mono-

therapy.

PharmacologicalProperties

Memantine is an uncompetitive (open-channel), moderate-affinity NMDA

receptor antagonist, with strong voltage dependency and rapid blocking and

unblocking kinetics. The blocking of the NMDA channel modulates the effects

of pathologically elevated tonic levels of the neurotransmitter glutamate that

may lead to neuronal dysfunction. Memantine reduces NMDA-induced excitoto-

xicity, but still allows receptor signalling for physiological activation.

Oral memantine is completely absorbed, with an absolute bioavailability of

about 100%. Linear pharmacokinetics were observed over the therapeutic dose

range. At steady state, plasma memantine concentrations ranged from 70 to

150 ng/mL, with large interindividual differences. The mean volume of distribu-

tion is about 10L/kg, and about 45% of the drug is bound to plasma proteins.

Memantine is converted primarily to three polar metabolites, which have minimal

NMDA receptor antagonist activity.Memantine and its metabolites are primarily

excreted by the kidneys. Total renal clearance in healthy volunteers was

170mL/min/1.73m2. The terminal elimination half-life of memantine is about

60–100 hours.

Therapeutic Efficacy In one well designed study of memantine 20mg/day monotherapy for 6 months,

the outcome of one primary efficacy measure (function) was significantly better

with memantine than placebo, but the difference between treatment groups did

not reach significance for the other primary endpoint (global change). Secondary

endpoints of cognitive performance, dementia progression and the need for

caregiver time showed significantly less decline with memantine than with pla-

cebo. In a 6-month, nonblind extension phase, patients switched to memantine

882 McKeage

ª 2009 Adis Data Information BV. All rights reserved. CNS Drugs 2009; 23 (10)

had a significantly slower rate of decline than that recorded when they received

placebo. In a second well designed, 6-month trial, the difference in efficacy

between memantine and placebo did not reach statistical significance for either

co-primary endpoint.

The efficacy of memantine combined with the AChE inhibitor donepezil in a

well designed, 6-month trial was significantly better than that of donepezil plus

placebo for both primary endpoints (domains of function and cognitive perfor-

mance), as well as in secondary global, behavioural and care dependency assess-

ments. Better outcomes were also associated with combination therapy than with

AChE inhibitor therapy or standard care alone over 30 months’ treatment.

Meta-analyses of randomized controlled trials of about 6 months duration in-

dicate that memantine alone or in combination with stable AChE inhibitor treat-

ment generally improves outcomes in patients with moderate to severe Alzheimer’s

disease across all four domains (global change, cognition, function and behaviour).

Cost-effectiveness analyses indicate that memantine is dominant to no phar-

macotherapy in patients with moderate to severe Alzheimer’s disease for treat-

ment periods of 2–5 years in regard to the cost per QALY gained. Compared with

monotherapy with donepezil, combination therapy with memantine plus done-

pezil was associated with a incremental cost-effectiveness ratio of $US382 per

QALY gained over 1 year, but for periods greater than 1 year, combination

therapy was dominant to donepezil alone.

Tolerability Memantine was generally well tolerated when used alone or in combination with

AChE inhibitors in patients with moderate to severe Alzheimer’s disease. In

clinical trials, most adverse events were mild to moderate in severity and were not

related (or not likely to be related) to study medication. The incidence of adverse

events with memantine (74–84%) was not different from that with placebo

(73–87%) in 6-month trials, and the most common event was agitation, which

occurred less with memantine than with placebo.

The tolerability profile of memantine during long-term (up to 1 year) therapy

was similar to that observed over 6 months. There were no clinically relevant

changes in laboratory tests, vital signs and electrocardiogram measurements

comparedwith baseline associated withmemantine during 6- and 12-month studies.

1. Introduction

Alzheimer’s disease is the most common causeof dementia and is estimated to affect about 6% ofthe population aged ‡65 years, with the incidenceincreasing with age.[1] The cause of the disease isunknown, but some of the key pathological find-ings include extracellular b-amyloid plaques andthe accumulation of intraneuronal neurofibrill-ary tangles of a hyperphosphorylated form of thetau protein.[2] The severity of dementia correlateswith the number of neurofibrillary tangles present.

Other causative factors include the disruption ofthe glutamate and acetylcholine neurotransmittersystem and cerebral atrophy.[2]

Alzheimer’s disease usually presents withcognitive dysfunction characterized by memoryloss. Gradually, other symptoms emerge, includ-ing psychiatric symptoms (e.g. depression),behavioural disturbances (e.g. agitation) anddifficulties in performing activities of daily living(e.g. dressing and eating).[3] The clinical diagnosisis based on medical history, physical examinationand a mental state assessment, with laboratory

Memantine: A Review 883


and imaging studies providing supporting evi-dence.[3,4] An important issue at the time of di-agnosis is to exclude other causes of dementia.[4]

The severity of dementia is defined according tothe DSM-IV of the American Psychiatric Asso-ciation and ICD-10 of the World Health Orga-nization.[4] A variety of specific and global ratinginstruments are available (table I) to assesssymptoms and determine severity levels. Mildcognitive impairment (MCI) has recently beenestablished as a diagnosis in patients with signsof pre-dementia; however, this term is not yetconsidered an homogenous clinical classification,

and more work is needed to establish validateddiagnostic criteria.[4] Further discussion ofMCI and mild Alzheimer’s disease (generally aMMSE score of 21–26)[5] is outside the scope ofthis review.

Memantine (Axura�, Ebixa�, Namenda�)[reviewed previously in Drugs][6] is a moderateaffinity, uncompetitive NMDA receptor antago-nist, which modulates glutamate signal trans-mission. This article reviews the pharmacologicalproperties of the drug and focuses on its clinicaluse in the treatment of patients with moderate tosevere Alzheimer’s disease.

Table I. Description of assessment scales used in clinical trials of patients with Alzheimer’s disease

Endpoint Definition Description

ADAScog Alzheimer’s Disease Assessment Scale –

cognitive subscale

Standard instrument for measuring cognitive ability on an 11-item scale

in patients with mild to moderate disease. Scores range from 0 to 70,

with higher scores indicating more impairment

ADCS-ADL19 or 23 Alzheimer’s Disease Cooperative Study –

Activities for Daily Living

Assesses function on a 19-item, 54-point scale modified for moderate to

severe dementia (23-item scale is used for patients with mild to moderate

disease). A higher score indicates better function

BDS Blessed Dementia Scale Evaluates cognitive impairment on a 37-point scale. Higher scores

indicate worse performance

BDS-IMC Blessed Dementia Scale – Information

Memory Concentration

Subscale of the BDS. Evaluates the degree of intellectual and personality

deterioration

BGP Behavioural rating scale for Geriatric

Patients

Evaluates 35 observable aspects of cognition, function and behaviour,

each scored 0, 1 or 2 (a higher score indicates worse function)

CIBIC-Plus Clinician’s Interview-Based Impression of

Change Plus Caregiver Input

Evaluates overall global change relative to baseline in cognition,

function and behaviour on a scale ranging from 1 (markedly improved) to

7 (markedly worse)

EMD Explorations modul Demenz (Dementia

Exploration Module)

Psychometric questionnaire consisting of 13 simple closed questions

evaluating cognition, everyday behaviour and affectivity, and disease-

related self-awareness

FAST Functional Assessment Staging Evaluates dementia progression from stage 1 (normal) to 7 (severe

dementia)

GDS Global Deterioration Scale Evaluates overall cognitive and functional capacity on a 7-stage scale,

based on patient and caregiver assessment. Higher stages indicate

greater impairment

MMSE Mini-Mental State Examination Evaluates cognitive function on a 30-point scale. A higher score indicates

better function

NOSGER Nurses’ Observation Scale for Geriatric

Patients

Evaluates level of impairment in memory, activities of daily living, mood,

social behaviour and disturbing behaviour

NPI Neuropsychiatric Inventory Caregiver-rated assessment that evaluates behaviour on a 12-item,

144-point scale. A lower score indicates better behaviour

RUD Resource Utilization in Dementia Assesses the burden of dementia on the caregiver

SIB Severe Impairment Battery Evaluates cognitive dysfunction on a 40-item, 100-point scale in patients

with moderate to severe Alzheimer’s disease. A higher score indicates

better cognitive function

Weintraub ADL Weintraub Activities of Daily Living Assesses function on a 31-item questionnaire. Scores range from 0%(normal) to 100% dependency

884 McKeage


2. Pharmacodynamic Properties

The pharmacodynamic properties of meman-tine have been reviewed in detail previously.[6-9]

This section provides a brief summary of theseproperties, with particular emphasis on recent(2006 onwards) pharmacodynamic studies inpatients with Alzheimer’s disease.

2.1 Mechanism of Action andPreclinical Studies

Memantine is an uncompetitive (open-chan-nel), moderate-affinity NMDA receptor antago-nist, which binds at, or near, the magnesiumbinding site, with strong voltage dependency andrapid blocking and unblocking kinetics.[7,8] Theblocking of the NMDA channel modulates theeffects of elevated tonic levels of the neurotrans-mitter glutamate.[10] The strong voltage depen-dence and fast agonist-concentration-dependentunblocking kinetics of memantine allow neuronaltransmission at the NMDA receptor site whenglutamate levels are transiently high allowing fornormal physiological function.[11]

Possible neuroprotective effects resulting fromthe prevention of abnormal glutamate neuro-transmission have been attributed to memantineand other NMDA receptor antagonists.[12] It isthought that excess glutamate levels and chronicdepolarization lead to NMDA receptor dysregu-lation and excessive calcium influx causingneuronal death. Memantine and other NMDAreceptor antagonists provide sufficient channelblockade to attenuate this process.[12] While theprimary mechanism of action of therapeutic re-levance for memantine is NMDA receptor an-tagonism, the drug also acts as an antagonist ofserotonin 5-HT3 receptors and neuronal nicotinicreceptors.[7]

In a series of in vitro and in vivo models, me-mantine reduced NMDA-induced excitotoxicity,but still allowed signalling of the receptor forphysiological activation.[13]

Memantine improved spatial learning intransgenic mice models of Alzheimer’s disease, asdetermined by Morris water maze tests, sugges-tive of disease-modifying efficacy.[14,15] In one

study,[15] after 2 months’ treatment with subcuta-neous memantine 7.2 or 14.4mg/kg/day followedby a wash-out period to eliminate bias from sus-tained symptomatic effects, acquisition perfor-mance (path length) and spatial accuracy duringprobe trial in the water maze were improved in adose-dependent manner. Furthermore, aged ratsreceiving intraperitoneal memantine 20mg/kg/dayfor 21 days exhibited improvements in age-induced long-term recognition memory deficitsand reduced oxidative damage to proteins in thecortex and hippocampus (areas involved withmemory) [p < 0.05 for all].[16]

Memantine did not affect the reversible inhi-bition of acetylcholinesterase (AChE) when co-administered with donepezil or galantamine.[17]

2.2 Studies in Patients withAlzheimer’s Disease

In patients with Alzheimer’s disease (n = 11),memantine 20mg/day for 1 year was associatedwith significantly reduced levels of CSF phos-phorylated tau (p = 0.018), a pathological featureof Alzheimer’s disease, compared with baseline.[2]

Possible evidence of the antiexcitotoxic effect ofmemantine was observed when the glutamate :creatinine ratio in the left hippocampus was re-duced significantly more in healthy individuals andcognitively impaired patients (n= 10) who receivedmemantine 20mg/day for 6 months than in un-treated controls (n = 6) [p < 0.01], as measured bymagnetic resonance spectroscopy; similar effectswere not demonstrated in the right hippo-campus.[18] This result may be attributable to agreater pathological involvement of the left thanthe right hippocampus in Alzheimer’s disease.

Adding memantine to stable dosages of anAChE inhibitor has been evaluated in two recentpharmacodynamic studies in patients withAlzheimer’s disease, both reported in abstracts.[19,20]

Combination therapy with memantine 20mg/dayplus an AChE inhibitor for 6 months was asso-ciated with a significantly slower rate of righthippocampal volume atrophy than treatmentwith an AChE inhibitor alone during the pre-vious 6-month run-in period (secondary end-point; n= 47) [p = 0.04]; as measured by MRI.[19]



The primary endpoint, which measured totalbrain volume atrophy, was not significantly af-fected by dual therapy.

A second dual-therapy study demonstratedincreased metabolic activity in the parietal andtemporal cortex regions after 10 weeks of treat-ment with memantine 20mg/day plus an AChE in-hibitor (n= 15), compared with baseline (p£ 0.001),as measured by PET imaging.[20] Metabolic in-creases in the right parietotemporal cortex wereassociated with improved symptom scores on theNeurobehavioural Rating Scale (p < 0.001), butnot the Dementia Rating Scale.

3. Pharmacokinetic Properties

The pharmacokinetic profile of memantinehas been reviewed previously.[6,21] This sectionsummarizes data taken primarily from the EU[22]

and US[17] prescribing information.

3.1 Absorption and Distribution

Memantine is completely absorbed after oraladministration, with an absolute bioavailabilityof about 100%.[22] Maximum plasma concentra-tions are reached in 3–8 hours, and memantine20mg/day achieves steady-state plasma con-centrations ranging from 70 to 150 ng/mL, withlarge interindividual differences. Food con-sumption does not affect bioavailability.[22]

In a population pharmacokinetic analysis, theimportant determinants of memantine apparentclearance included total bodyweight, the presenceof concomitant medication eliminated via tubu-lar secretion and the type of memantine for-mulation (i.e. tablets or solution).[23]

A mean CSF : serum ratio of 0.52 has beenestimated.[22] The mean volume of distribution ofmemantine is about 10L/kg, and about 45% ofthe drug is bound to plasma proteins.[22]

3.2 Metabolism and Elimination

Memantine is partially metabolized by the li-ver and is converted primarily to three polarmetabolites, which have minimal NMDAreceptor antagonist activity.[17] No cytochromeP450 (CYP) metabolism has been detected

in vitro.[17,22] Memantine and its metabolites areprimarily excreted by the kidneys, and activetubular secretion moderated by pH-dependenttubular reabsorption are involved in renal clear-ance.[17,22] The total renal clearance of meman-tine was 170mL/min/1.73m2 in healthy volun-teers with normal renal function.[22] The terminalelimination half-life (t½b) of memantine is about60–100 hours.[22]

3.3 Special Populations

In patients with mild (creatinine clearance[CLCR] 50–80mL/min), moderate (CLCR

30–49mL/min) or severe (CLCR 5–29mL/min)renal impairment, mean memantine area underthe plasma concentration-time curve increased by4%, 60% and 115%, and mean t½b increased by18%, 41% and 95%, respectively, compared withhealthy volunteers.[17] No dosage adjustment isnecessary in patients with mild or moderate renalimpairment, but in patients with severe renalimpairment, a target dosage of 5mg twice daily isrecommended.[24]

Memantine exposure in patients with moderatehepatic impairment was not increased comparedwith healthy volunteers, but t½b increased by about16%.[17] Dosage adjustment is not necessary in pa-tients withmild tomoderate hepatic impairment (seealso section 7).[17,22] The pharmacokineticsof memantine in the elderly (aged >65 years) aresimilar to those in younger individuals,[17] anddosage adjustments are not required.[22]

3.4 Drug Interactions

No pharmacokinetic drug interactions areexpected with memantine and coadministereddrugs metabolized by CYP enzymes (section3.2).[17] Drugs that alkalinize the urine (e.g.sodium bicarbonate, carbonic anhydrase inhibi-tors) could potentially reduce the renal eliminationof memantine (section 3.2). The coadministra-tion of memantine and donepezil did not affectthe pharmacokinetics of either agent,[17] andthe coadministration of memantine and rivas-tigmine did not affect the pharmacokinetics ofrivastigmine.[25]

886 McKeage


4. Therapeutic Efficacy

The efficacy of oral memantine 20mg/day inmoderate to severe Alzheimer’s disease has beenevaluated in three large (n > 250), randomized,double-blind, placebo-controlled, multicentre,6-month trials.[26-28] Two trials evaluated theagent as monotherapy (section 4.1.1),[26,28] and athird study evaluated memantine in combinationwith stable dosages of the AChE inhibitor done-pezil (section 4.2);[27] donepezil is the only AChEinhibitor that has been compared with meman-tine in well controlled trials. The longer-term(>6 months) efficacy of the drug has also beenevaluated in a 6-month, non-blind extensionphase[29] of a monotherapy trial[26] (section 4.1.2),and in various non-randomized, observationalstudies (section 4.4).[30-32]

Although providing a lower level of evidencethan randomized controlled trials, data fromformal meta-analyses (section 4.3)[33-35] and stu-dies conducted in a clinical practice setting (sec-tion 4.4)[30-32,36] are included in this review. Inaddition, data from various pooled analyses[37-44]

and post hoc subgroup analyses[45-52] of rando-mized controlled trials are included; however,they should be considered as purely exploratoryand hypothesis generating.

Patients enrolled in the double-blind studieshad a diagnosis of probable Alzheimer’s diseaseaccording to the criteria of the National Instituteof Neurological and Communicative Disordersand Stroke/Alzheimer Disease and Related Dis-orders Association[27,28] and the DSM-IV.[26] Inaddition, patients were required to have hadbrain-imaging evaluation consistent with prob-able Alzheimer’s disease within the previous12 months.[26-28] Eligible patients were aged‡50 years (average age ranged from 75.5 to78.2 years). Other inclusion criteria included thepresence of a reliable caregiver to ensure thatstudy medication was administered,[26-28] and aMMSE (see table I) score of 3–14[26] or 5–14.[27,28]

Generally, MMSE scores of 11–20 and 0–10are considered to indicate moderate and severeAlzheimer’s disease, respectively.[53] In the stu-dies conducted in a clinical practice setting (sec-tion 4.4), patients had a mean age of about

75 years and were likely to have moderate to se-vere Alzheimer’s disease consistent with the ap-proved indication of memantine.[30-32,36]

All three double-blind studies had two co-primary endpoints.[26-28] In two studies,[27,28]

these included the change from baseline on theADCS-ADL19 and the SIB (see table I). Theprimary endpoints in the third study[26] were theCIBIC-Plus global score, and the change frombaseline in the modified ADCS-ADL19.

Secondary endpoints included the NPI,[26-28]

SIB,[26] CIBIC-Plus,[27,28] BGP,[27,28] GDS,[26]

MMSE,[26] FAST[26,28] and RUD.[26] Primaryefficacy analyses were performed on the intent-to-treat population, which included patients whohad completed at least one post-baseline assess-ment, and last observation carried forward(LOCF) analyses were used.[26-28]

4.1 Monotherapy

4.1.1 Double-Blind Studies

In one monotherapy study (MRZ-9605),patients receiving memantine generally hadbetter outcomes than those receiving placebo,as determined by some primary and secon-dary endpoints,[26] and in the other study (MEM-MD-01), differences between memantineand placebo groups were not significant for anyof the endpoints assessed.[28]

In the MRZ-9605 study, there was significantlyless decline with memantine than with placebo infunction (p < 0.05), as determined by ADCS-ADL19 (primary endpoint), but the difference inthe other primary endpoint CIBIC-Plus was notsignificant between groups (table II).[26] Somesecondary endpoints in this study also showed atreatment benefit with memantine compared withplacebo, including measures of cognitive perfor-mance (SIB) [p < 0.001; table II], dementia pro-gression (FAST score was 0.2 with memantine vs0.6 with placebo; p< 0.05), and a greater reduc-tion in the need for caregiver time (derived fromthe RUD instrument) [difference between groupswas 45.8hours per month; p= 0.01], but scores onthe MMSE (-0.5 vs -1.2), GDS (0.1 vs 0.2) andNPI (table II) were not significantly differentbetween treatment groups.[26] When analyses



were performed with observed cases rather thanLOCF, similar results were observed, as well as asignificant difference in favour of memantine forthe primary endpoint of CIBIC-Plus (4.4 vs 4.7,respectively; p < 0.05).[26]

In a post hoc evaluation of the effects of me-mantine on function, as determined from ADCS-ADL19 data from the MRZ-9605 trial, patientstreated withmemantine were more likely to remainautonomous than placebo recipients (odds ratio3.03; 95%CI 1.38, 6.66).[46] Furthermore, althoughdifferences in behavioural changes (assessed viathe NPI) between treatment groups in this trialwere not significant (table II), post hoc analysessuggest that, in memantine versus placebo recipi-ents, the NPI domain scores of agitation/aggres-sion and delusion improved significantly more inthose experiencing these symptoms at baseline,and agitation/aggression emerged in significantlyfewer of those who were asymptomatic at base-line (p < 0.05 for all comparisons).[45]

In a post hoc number-needed-to-treat (NNT)analysis of theMRZ-9605 trial, the NNTwas 6 toachieve a response (predefined as an improve-ment or no deterioration in the CIBIC-Plus andeither ADCS-ADL19 or SIB), and significantlymore memantine than placebo recipientsachieved a response (29% vs 10%; p= 0.004).[47]

4.1.2 Extension Study

In a nonblind, 24-week extension phase[29]

(n= 175) of the MRZ-9605 trial,[26] all patientscontinued, or were switched to, memantine therapy(titrated up to 20mg/day). The favourable treatmenteffect observed with memantine compared with pla-cebo in the double-blind phase was continued in theextension phase, with a significantly slower rateof decline in the CIBIC-Plus (p < 0.001), ADCS-ADL19 (p < 0.05) and SIB (p < 0.05) during theextension phase than during the double-blindphase in patients initially randomized to placebothen switched to memantine.[29]

In those who received memantine throughout52 weeks, in the extension phase compared withthe double-blind phase, the rate of decline inCIBIC-Plus was significantly slower (p < 0.001),the rate of decline in the ADCS-ADL19 was sig-nificantly faster (p < 0.05), and the rate of declinein the SIB did not differ.[29]

4.2 In Combination with anAcetylcholinesterase Inhibitor

When memantine was combined with stabledonepezil therapy in a randomized, double-blind,24-week study (MEM-MD-02), combinationtherapy was associated with better outcomes than

Table II. Efficacy of memantine in patients (pts) with moderate to severe Alzheimer’s disease. Results of large (n > 100), randomized, double-

blind, placebo-controlled studies evaluating the efficacy of memantine (MEM) 20mg/day or placebo (PL) over 24[27,28] or 28[26] weeks. In one

study,[27] pts also received the acetylcholinesterase inhibitor donepezil (DON) [mean dosage 9.4mg/day]. Primary efficacy analyses were

performed on the intent-to-treat populations using last observation carried forward methodology

Study (acronym) Regimena No. of pts CIBIC-Plus ADCS-ADL19 (function)b SIB (cognition)b NPI (behaviour)b

(global

change)bBL change BL change BL change

Monotherapy

Reisberg et al.[26] MEM 126 4.5c 26.8 -3.1*c 65.9 -4.0*** 21.4 -0.5

(MRZ-9605) PL 126 4.8c 27.4 -5.2c 68.3 -10.1 19.5 -3.8

van Dyck et al.[28] MEM 178 4.3 33.1 -2.0c 77.2 -2.0c 20.3 -1.0

(MEM-MD-01) PL 172 4.6 33.6 -2.7c 75.6 -2.5c 17.5 -1.1

In combination with DON

Tariot et al.[27] DON +MEM 202 4.4* 35.5 -2.0*c 78.0 0.9***c 13.4 -0.1**

(MEM-MD-02) DON +PL 201 4.7 35.8 -3.4c 80.0 -2.5c 13.4 -3.7a MEM titrated upwards from 5mg/day at wk 1, increasing by 5mg/wk to final dosage of 20mg/day.

b Refer to table I for definition of efficacy endpoints.

c Primary endpoint.

BL =baseline; * p< 0.05, ** p< 0.01, *** p <0.001 vs PL –DON.

888 McKeage


donepezil plus placebo.[27] In the primary eva-luation, functional (ADCS-ADL19) and cognitive(SIB) outcomes were significantly better withcombination therapy than with single-agent do-nepezil (table II). The results of secondary ana-lyses also favoured dual therapy, including global(CIBIC-Plus) and behavioural (NPI) endpoints(table II), and the care dependency subscale of theBGP (0.8 with memantine vs 2.3 with placebo;p= 0.001 [baseline scores were 9.5 vs 9.8, respec-tively]).[27] The beneficial effect of combinationtherapy was evident at week 4, at which timethere was a significant difference between thedual-therapy versus donepezil-plus-placebogroups in the ADCS-ADL19 (p < 0.05).[27]

Several post hoc evaluations[45,48-52] have fur-ther analysed data from the trial of memantineplus donepezil compared with donepezil alone(MEM-MD-02).[27] A responder analysis sug-gested that the addition of memantine treatmentin patients maintained on stable donepezil ther-apy improved and stabilized symptoms acrossmultiple outcomes, including individual mea-sures of the ADCS-ADL19, SIB, CIBIC-Plus andNPI, and combined measures of these endpoints(p < 0.05 for all).[52] When treatment response wasdefined as stabilization of individual outcomes,combination therapy with memantine and done-pezil was associated with significantly higher re-sponse rates than donepezil plus placebo, and theNNT ranged from 8 to 10.[52]

Further analysis of behavioural (NPI) datasuggests that patients receiving combinationtherapy with donepezil plus memantine mayexperience less agitation/aggression (p = 0.001),irritability/lability (p = 0.005) and appetite chan-ges (p = 0.045) than those receiving donepezilalone.[45,48] Furthermore, in patients without NPIagitation/aggression domain scores at baseline,the emergence of this symptom/s occurred infewer memantine plus donepezil-treated patientsthan donepezil plus placebo-treated patients over24 weeks (15% vs 27%; p= 0.016).[48]

Post hoc analyses using mixed model repeatedmeasures of ADCS-ADL19 data from the MEM-MD-02 trial suggest that combination therapyis associated with beneficial effects on function,including toileting, grooming, conversing and

being left alone, compared with donepezil treat-ment alone (p < 0.05 for all).[49] The ADCS-ADL19 subscales of higher-level functions andconnectness/autonomy were improved more inthe combination-therapy group than in thedonepezil-alone group (p < 0.05 for both).[49]

Furthermore, a recent analysis (reported in anabstract) of SIB and ADCS-ADL19 data fromthis same study, indicates that combinationtherapy improves verbal and functional commu-nication significantly more than donepezil plusplacebo (p < 0.05 for both).[51]

A post hoc exploratory re-analysis of the effect ofmemantine on cognitive response in patients re-ceiving donepezil, as determined by individual itemson the SIB, suggest that combination therapy withmemantine plus donepezil significantly improvesmemory, language and praxis more than donepezilplus placebo (p< 0.05 for all comparisons).[50]

4.3 Meta-Analyses and Pooled Analyses

Overall, several meta-analyses[33-35] and pooledanalyses[37-44] of randomized controlled trials ofabout 6 months duration have indicated that mem-antine alone or in combination with stable AChEinhibitor treatment generally improves outcomes inpatients with moderate to severe Alzheimer’sdisease across all four domains (global change,cognition, function and behaviour).

Results of three meta-analyses, which calcu-lated treatment effect sizes for memantine versusplacebo, are summarized in table III.[33-35] Inthese meta-analyses, memantine was associatedwith significant improvements compared withplacebo in all efficacy domains (p < 0.05), exceptfor behaviour (NPI) in one study, which demon-strated borderline significance (p = 0.05) for thisdomain.[34]

In two pooled analyses (one was reported in anabstract),[37] outcomes in all four efficacy domains(SIB, CIBIC-Plus, ADCS-ADL19 and NPI) weresignificantly better with memantine than with pla-cebo (p< 0.05 for all) at 6 months.[37,38] In one ofthese analyses that was conducted only in patients(n= 593) with specific behavioural symptoms(i.e. agitation/aggression and/or psychosis), themean improvement from baseline in the NPI



cluster score for agitation/aggression, delusionsand hallucinations was greater in the memantinegroup than in the placebo group at week 12 (-0.8vs 0.5; p< 0.01) and at week 24/28 (-0.7 vs 0.7;p< 0.001). Furthermore, the NPI cluster score im-proved in more patients in the memantine groupthan in the placebo group at week 12 (56% vs 44%)and 24/28 (58% vs 45%) [p< 0.001 for both].[38]

Memantine-treated patients also had sig-nificantly better NPI total scores (p = 0.008) andless deterioration in mean change from baselinein 3 of 12 single item scores, including delusions(-0.06 vs 0.23; p = 0.001), agitation/aggression(0.05 vs 0.42; p = 0.001) and irritability/lability(0.04 vs 0.38; p= 0.005) at week 24 or 28 thanplacebo (item scores were estimated from a fig-ure) in another pooled analysis of Alzheimer’spatients (n = 1788).[40] Similar results were evidentat 12 weeks, although the difference betweentreatment groups for irritability/lability had notreached significance.[40] When only those whowere symptomatic for NPI items at baseline wereassessed, symptomatic improvement occurred insignificantly more patients receiving memantinethan placebo in the items of delusions (62% vs52%), agitation/aggression (56% vs 47%) anddisinhibition (68% vs 57%) at 6 months (p < 0.05for all) and, in those who were asymptomatic atbaseline, more patients in the memantine than theplacebo group remained asymptomatic in the itemsof agitation/aggression (82% vs 74%; p = 0.002)and irritability/lability (84% vs 77%; p = 0.004)[item scores were estimated from a figure].[40]

When all domains except behaviour wereevaluated in a pooled analysis (n = 1826) of6-month data, fewer memantine-treated patients

(some of whom were already receiving AChE in-hibitors) than placebo-treated patients deterio-rated clinically (18% vs 28%; p< 0.001), asdefined by the CIBIC-Plus, ADCS-ADL19, SIB andADAScog.

[43] In an analysis of specific areas offunctional ability, as measured by ADCS-ADL19

data, memantine treatment was associated withsignificantly less decline than placebo (p < 0.05for all) in 5 of 19 items, including getting aroundoutside and paying attention to conversation.[41]

In two pooled analyses (n= 1826 in each) of theeffects on memantine on cognition using SIB andADAScog data, after 6 months of treatment, mem-antine was associated with significantly greaterbenefits on overall cognitive ability than place-bo.[39,44] In the observed cases population, sig-nificantly greater improvements from baselinewere reported for the ADAScog single items ofcommands (p < 0.001), ideation praxis (p < 0.05),orientation (p < 0.01), comprehension (p < 0.05)and remembering test instructions (p < 0.05), andfor the SIB single items of language (p < 0.05),memory (p < 0.05), orientation (p < 0.01), praxis(p < 0.001) and visuospatial ability (p < 0.01).[44]Similarly, in the second analysis, significantlymore patients receiving memantine than placeboimproved in language (40% vs 31%; p< 0.01),memory (58% vs 40%; p < 0.001) and praxis (34%vs 28%; p < 0.01) [percentages were estimatedfrom a figure].[39]

A recent analysis (reported in an abstract) ofthe effects of memantine 20mg/day on languageusing pooled data from four studies (n = 257)supports these results.[42] A language scale con-structed from the SIB scale (SIB-L)[54] indicatedthat after 6 months, the mean change from

Table III. Meta-analyses of 24- and 28-week trials of memantine (MEM) in patients (pts) with moderate to severe Alzheimer’s disease.

Analyses were performed on observed cases[34] or last observation carried forward[33,35] in the intent-to-treat population

Analysis No. of pts Effect sizes for MEM vs placebo (95% CI)

global change

(CIBIC-Plus)

cognition

(SIB, ADAScog)

function

(ADCS-ADL19 or 23)

behaviour

(NPI)

Doody et al.[33] 1006 0.27 (0.39, 0.14)*** 0.29 (0.54, 0.03)* 0.19 (0.32, 0.06)** 0.17 (0.30, 0.04)**

McShane et al.[34]a £1992 0.21 (0.3, 0.11) 0.21 (0.3, 0.11) 0.15 (0.26, 0.03) 0.09 (0.18, 0)

Winblad et al.[35] 1826 0.22*** 0.26*** 0.18*** 0.12*

a Reported in an abstract. p-Values were not reported.

* p <0.05, ** p £0.01, *** p £0.001 vs placebo.

890 McKeage


baseline in the SIB-L was significantly less inmemantine than placebo recipients (p < 0.05). Inaddition, more patients receiving memantinethan placebo were considered SIB-L responders(18.7% vs 6.5%; p = 0.011).[42]

4.4 Clinical Practice Setting

In studies conducted in a clinical practice set-ting, memantine alone or combined with an AChEinhibitor was associated with improved symptomsin patients with Alzheimer’s disease.[30-32,36]

A prospective cohort of patients in a MemoryDisorders Unit with a diagnosis of probable Alz-heimer’s disease were allocated, in a non-blind,non-randomized manner, to receive memantineplus an AChE inhibitor (n= 116), an AChE in-hibitor alone (n= 122) or no treatment (n= 144)based on consistency with previous care, and fol-lowed up for a mean period of 30 months.[30]

Following adjustments for baseline variationsbetween groups, patients receiving combinationtherapy had significantly lower mean annualizedrates of deterioration in BDS-IMC (baseline scoreof 11) and Weintraub ADL (baseline score of31.2) scores than those receiving an AChE in-hibitor alone or no treatment, with effect sizesincreasing with increasing treatment duration.[30]

The effect size of combination therapy versusAChE inhibitor alone was 0.10 (not significant) inyear 1 increasing to 0.49 (p< 0.001) in year 4, andfor the comparison of combination therapy versusno treatment, the effect sizes were 0.56 (p< 0.001)at year 1 and 0.77 (p< 0.001) by year 4.[30]

Similarly, in a recent retrospective, observationalstudy in 943 patients with probable Alzheimer’sdisease with at least 1-year of follow-up data, thetime to nursing home admission was significantlydelayed in those receiving memantine plus anAChE inhibitor compared with those receivingAChE inhibitor therapy alone (hazard ratio: 0.29;95% CI 0.11, 0.72).[31] During follow-up, 5% ofpatients in the combination group versus 18% ofpatients in theAChE inhibitor groupwere admittedto a nursing home (p< 0.001). There was no asso-ciation between treatment and time to death.

After a mean of 6 months’ treatment withmemantine 20mg/day (22% also received another

antidementia drug) in patients (n = 1580) withpredominantly moderate to severe Alzheimer’sdisease (MMSE score of 15.4), cognitive function(determined by MMSE and EMD; see table I),ability to perform daily activities (NOSGER andEMD) and physician-assessed global perfor-mance improved compared with baseline.[36]

The MMSE (n = 1199) improved by 2.5 points(p < 0.0001 vs baseline score of 15.4), all items ofthe NOSGER (including memory, activities ofdaily living, mood, social and disruptive beha-viour) improved significantly at 3 and 6 months(p £ 0.0001 vs baseline for all comparisons at bothtimepoints), mean total patient- and caregiver-assessed EMD scores improved at 6 months(p < 0.0001 vs baseline) and 78.8% of patientsexhibited improved or stabilized symptoms ac-cording to the physician’s global assessment.[36]

Results from a 2-year follow-up cohort study of4600memantine-treated patients randomly extractedfrom the French National Health Care Databasesuggest that the initiation of memantine therapy isassociated with a lower use of psychotropic agentscompared with pre-memantine use.[32] Data wereanalysed using interrupted time series. In the12 months before memantine therapy, the propor-tion of patients receiving psychotropic therapy(antidepressants, antipsychotics, hypnotics andanxiolytics) increased from 39% to 50%, but fol-lowing memantine initiation, the increasing trendstopped and stabilized at about 53% 11 monthsafter commencing memantine.[32] The differencein the trends for the use of psychotropics pre- andpost-memantine was significant (p < 0.0001).

5. Pharmacoeconomic Considerations

This section summarizes cost-effectivenessstudies with a year of costing ‡2003.[55-59] Allstudies used similar Markov models, which si-mulated patient progression through a series ofhealth states and dependency based on clinicaland epidemiological data. Resource use and spe-cific costs varied depending on the study per-spective and local unit costs.

Several cost-effectiveness analyses have in-dicated that memantine is dominant (i.e. moreeffective and less costly) to no pharmacotherapy



in patients with moderate to severe Alzheimer’sdisease for treatment periods of 2–5 years in re-gard to cost per quality-adjusted life-year (QALY)gained (table IV).[55-58] Compared with mono-therapy with donepezil, combination therapy withmemantine plus donepezil was associated withan incremental cost-effectiveness ratio (ICER) of$US382 per QALY gained over 1 year (table IV),but for periods greater than 1 year, combinationtherapy was dominant over donepezil alone.[59]

Sensitivity analyses generally confirmed thatbase-case results were robust.[55-59]

Pharmacoeconomic analyses of memantine, aswith all such modelled analyses, are subject to anumber of limitations, with the potential for in-put data to differ from real-life situations. How-ever, modelled cost-effectiveness analyses withmemantine in Alzheimer’s disease were generallywell conducted, included appropriate para-meters, and model designs were applicable toAlzheimer’s disease.

6. Tolerability

Tolerability data on memantine are taken fromclinical trials and pooled analyses discussed insection 4. Memantine 20mg/day in patients with

moderate to severe Alzheimer’s disease was gen-erally well tolerated when used alone or in combi-nation with AChE inhibitors for periods of up to 1year. Most adverse events were mild to moderatein severity and not related (or not likely to berelated) to study medication.[26-29]

Not surprisingly in this patient group, adverseevents were common.[26-29] In placebo-controlled,6-month trials, the incidence of adverse events wasnot different between memantine and placebogroups, and ranged from 74% to 84% with mem-antine (alone orwith donepezil) versus from 73% to87% with placebo.[26-28] Agitation was the mostfrequently reported event in all double-blind stu-dies, and it occurred more often in placebo groups(32%,[26] 14%[28] and 12%[27]) than in memantinegroups (18%,[26] 9%[28] and 9%[27]). In the studycomparing memantine plus donepezil with done-pezil alone, the incidence of diarrhoea was 4.5%versus 8.5%, respectively, and the incidence of fae-cal incontinence was 2% versus 5%, respectively.[27]

Treatment discontinuation due to adverseevents occurred in a numerically higher propor-tion of patients receiving placebo than memantinein two studies (17% vs 10%[26] and 12% vs 7%[27]),and in a similar proportion of both groups in athird study (13% vs 12%).[28] Serious adverse

Table IV. Summary of base-case results of modelled, cost-effectiveness analyses of memantine (MEM) in patients with moderate to severe

Alzheimer’s disease. The studies were performed from the perspective of society[55,59] or a public healthcare provider,[56-58] and were modelled

on published pharmacoeconomic data

Reference (country) Year of costing

(annual discount

rate %)

Treatment QALY

gained

Cost per patient over time ICER

MEM compared with no pharmacotherapy (NT)

Antonanzas et al.[55] (Spain) 2005 (6) MEM 0.71a h24 055 over 2 y MEM dominant over NT

NT 0.51a h24 721 over 2 y

Gagnon et al.[56] (Canada) 2005 (5) MEM 0.71 $Can75 191 over 2 y MEM dominant over NT

NT 0.679 $Can76 467 over 2 y

Jones et al.[57] (UK) 2003 (3.5) MEM 0.61 d60 516 over 2 y MEM dominant over NT

NT 0.57 d62 479 over 2 y

Jonsson[58] (Sweden) 2004 (3) MEM 1.667 SEK920276 over 5 y MEM dominant over NT

NT 1.519 SEK1020804 over 5 y

MEM plus donepezil (DON) compared with DON alone

Weycker et al.[59] (USA) 2005 (3) DON +MEM 0.1990 $US29 866 over 1 y $U382 per QALY gained

vs DON alone

DON 0.1715 $US29 856 over 1 y

a Time in an independent state.

ICER = incremental cost-effectiveness ratio; QALY = quality-adjusted life year; SEK =Swedish kroner.

892 McKeage


events occurred in 13% of patients receivingmemantine compared with 18% receiving placeboin one monotherapy trial,[26] and 15% versus 17%,respectively, in the other monotherapy trial.[28]

The tolerability profile of memantine duringlong-term (up to 1 year) therapy[29] was consistentwith that of 6-month, double-blind studies. Therewere no clinically relevant changes in laboratorytests, vital signs and electrocardiogram measure-ments compared with baseline in memantine or pla-cebo groups during 6-[26-28] and 12-month[29] studies.

The results of meta-analyses[33,35] and an anal-ysis of pooled tolerability data[60] support those ofindividual studies. When data were pooled fromthe three 6-month, double-blind trials[26-28] and the6-month, non-blind, extension study,[29] memantinewas found to have a similar tolerability profile tothat of placebo (figure 1), with an overall incidenceof at least one adverse event occurring in 68% ofpatients in both groups.[60] Although statisticalanalyses were not performed, the incidence of

confusion and hypertension was numericallyhigher in the memantine than the placebo group.

7. Dosage and Administration

Memantine is approved for use in patientswith moderate to severe Alzheimer’s disease inthe EU,[22] and moderate to severe dementia ofthe Alzheimer’s type in the US.[17]

The recommended maintenance dosage is20mg/day, and memantine can be taken with orwithout food.[17,22] Initially, the dosage shouldbe titrated, starting with 5mg once daily for thefirst week, then increasing by 5mg/day eachweek until the maintenance dosage is reached byweek 4.[17,22] It is recommended that the main-tenance dose of memantine is administered oncedaily (at the same time each day) in the EU[22] andin two divided doses in the US.[17] Once-dailyadministration is supported by results of a recent

0

2

4

6

8

10

12

14

16

18

20

Agitat

ion Fall

Dizzine

ss

Diarrh

oea

Accide

ntal

injur

y

Confu

sion

Urinar

y

incon

tinen

ce

Urinar

y tra

ct

infec

tion

Influ

enza

-like

sym

ptom

s

Hyper

tens

ion

Inci

denc

e (%

of p

atie

nts)

MemantinePlacebo

Fig. 1. Tolerability profile of memantine 20mg/day in patients with moderate to severe Alzheimer’s disease. Results are from a pooledanalysis of three double-blind, 6-month studies[26-28] and one non-blind, 6-month, extension study[29] (n =1180). Treatment-emergent adverseevents that occurred in ‡5% of patients are shown.[60]



small trial in Alzheimer’s patients (n= 78), in whichsimilar efficacy and tolerability was achieved withonce- or twice-daily administration.[61]

In patients with severe hepatic impairment,memantine is not recommended,[22] or should beused with caution,[17] and dosage reduction is re-commended in patients with severe renal im-pairment (section 3.3).[17,22] Local prescribinginformation should be consulted for further de-tails regarding precautions, contraindicationsand warnings.

8. Place of Memantine in theManagement of Moderate to SevereAlzheimer’s Disease

As disease-modifying agents are not yet avail-able for Alzheimer’s disease, the primary goals oftherapy for moderate to severe disease are toimprove symptoms and slow symptom progres-sion.[4] Therapeutic interventions target cogni-tive, behavioural and functional symptoms aswell as the reduction of caregiver burden. Ap-proved drugs include the AChE inhibitors (e.g.donepezil, rivastigmine, galantamine), which in-crease levels of the neurotransmitter acetylcho-line by inhibiting the action of cholinesterase, andmemantine, which modulates glutaminergic-mediated calcium transmission as a voltage-dependent moderate affinity NMDA receptorantagonist (section 2). AChE inhibitors have amod-erate but generally effective symptom-modifyingeffect in many patients with Alzheimer’s disease(all severity levels), improving measures of cogni-tion and global assessment in clinical trials.[62]

Memantine has also demonstrated efficacy inpatients with moderate to severe Alzheimer’sdisease, but not all monotherapy trials have beenconclusive (section 4.1.1).

The distinct mode of action of the AChE in-hibitors and memantine (i.e. the modulation ofthe cholinergic and glutaminergic neurotrans-mitter systems) suggest a possible additive effectwhen these agents are coadministered. In a6-month trial of dual therapy with memantineplus donepezil in patients with moderate to severeAlzheimer’s disease, better outcomes were

achieved in the dual-therapy arm than in thedonepezil-monotherapy arm (section 4.2). Un-fortunately, this trial did not include a meman-tine-only arm, and no data are available onwhether dual treatment is more effective thanmemantine monotherapy. Post hoc analyses(section 4.2) and a 30-month study in a clinicalpractice setting (section 4.4) support the favour-able results achieved with dual therapy.

Primary data from clinical trials of memantinehave shown improvements in the domains offunction (monotherapy and combination therapywith donepezil) and cognition (combinationtherapy with donepezil). However, post hoc datasuggest that memantine improves outcomes inpatients with moderate to severe Alzheimer’sdisease across all four domains (global change,cognition, function and behaviour) [sections4.1.1, 4.3 and 4.4], although the treatment effecton behaviour appears to be less clinically detect-able (section 4.3). As yet, no randomized trial hasevaluated memantine using NPI behaviouraldata as a primary efficacy endpoint.

Treatment guidelines in the US recommend thatthe decision to commence therapy with either anAChE inhibitor or memantine is based on anindividual assessment of benefits and risks for eachpatient, including tolerability issues, easeof use and drug costs.[63] In the UK, The NationalInstitute for Health and Clinical Excellence (NICE)recommends AChE inhibitors as an option for themanagement of moderate Alzheimer’s disease, butrecommends that memantine use is restricted toAlzheimer’s disease patients involved in clinicaltrials of the drug, based on their assessments ofclinical data and cost effectiveness.[5] Widespreaddebate has followed this recommendation, raisingvarious questions including NICE’s reliance onQALYs as a measure of effectiveness in de-mentia.[64] In cost-effectiveness analyses summar-ized in this review (section 5), memantine wasdominant to no pharmacotherapy in patients withmoderate to severe Alzheimer’s disease for treat-ment periods of 2–5 years, primarily as a result ofreduced caregiver costs, and combination therapywith memantine plus donepezil was dominant todonepezil alone when treatment periods weregreater than 1 year.

894 McKeage


Debate has also surrounded the questionconcerning whether the significant improvementsin symptom measures in dementia trials canbe considered clinically significant.[63] Hopefully,as more efficacy data become available, and waysof measuring the value of symptom-modifyingtreatment, as opposed to disease-modifyingtreatment, in neurodegenerative diseases im-proves, a more definitive treatment guidancewill result.

Ongoing research should also provide in-formation on the optimal duration of therapy,which has not yet been determined for meman-tine or for AChE inhibitors.[63] Guidelines statethat once slowing symptom progression ceases tobe a goal (i.e. once a patient fails to show aclinically important response), treatment shouldbe stopped.[63] The development of agents thatare able to modify the disease process and lead tomore dramatic clinical improvements than exist-ing agents is also a priority for further research.Currently, immunotherapy targeting b-amyloidis showing some promise, with the monoclonalantibody bapineuzumab reaching phase IIItrials.[65] Another anti-amyloid agent in a late-phase clinical trial is the gamma secretase in-hibitor semagacestat (LY450139), which worksby reducing production of b-amyloid.[65] Agentsthat target tau-related pathology are also beinginvestigated.[65] Dimebon, which was originallyused as an antihistamine, has also reached latephase clinical trials in Alzheimer’s disease basedon promising initial results. The supposed me-chanism of action includes NMDA receptor an-tagonism, but it is also thought to involve stabi-lization of the mitochondria.[65]

The tolerability profile is an important con-sideration when deciding whether to initiate orcontinue anti-dementia therapy.[63] Memantinewas generally well tolerated in clinical trials whenused alone or combined with donepezil, and inpatients with moderate to severe disease, the in-cidence of adverse events, and the overall toler-ability profile, was not different to that reportedwith placebo (section 6).

In conclusion, in well designed trials in pa-tients with moderate to severe Alzheimer’s dis-ease, oral memantine monotherapy improved

outcomes in one (functional ability) of two pri-mary endpoints more than placebo in one trial,but in a second trial, treatment differences did notreach significance. Memantine has a distinctmode of action to that of AChE inhibitors, and ina well designed study, combination therapy withmemantine plus donepezil improved outcomesmore than donepezil plus placebo in all four do-mains (function, cognition, behaviour and globalchange). Memantine is generally well tolerated,with adverse events occurring with a similar in-cidence to that reported with placebo. In mod-elled cost-effectiveness analyses, in regard to costper QALY gained, memantine was dominant tono therapy, and the combination of memantineplus donepezil was dominant to donepezil ther-apy alone when treatment periods exceeded1 year. Thus, in the management of patients withmoderate to severe Alzheimer’s disease, meman-tine provides an effective treatment option. Todate, clinical trial support is greater for meman-tine use in combination with an AChE inhibitor,while more data are needed to confirm its efficacyas monotherapy.

Disclosure

The preparation of this review was not supported by anyexternal funding. During the peer review process, the manu-facturer of the agent under review was offered an opportunityto comment on this article. Changes resulting from commentsreceived were made on the basis of scientific and editorialmerit.

References1. Kukull WA, Higdon R, Bowen JD, et al. Dementia and

Alzheimer disease incidence: a prospective cohort study.Arch Neurol 2002 Nov; 59: 1737-46

2. Degerman Gunnarsson M, Kilander L, Basun H, et al. Re-duction of phosphorylated tau during memantine treat-ment of Alzheimer’s disease. Dement Geriatr Cogn Disord2007; 24 (4): 247-52

3. Burns A, Iliffe S. Alzheimer’s disease. BMJ 2009 Feb; 338:467-71

4. European Medicines Agency Committee for MedicinalProducts for Human Use. Guideline on medicinal productsfor the treatment of Alzheimer’s disease and other de-mentias [online]. Available from URL: http://www.emea.europa.eu [Accessed 2009 Feb 23]

5. National Institute for Health and Clinical Excellence. NICEtechnology appraisal guidance 111 (amended). Denepezil,galantamine, rivastigmine (review) and memantine for the



treatment of Alzheimer’s disease (amended) [online].Available from URL: http://www.nice.org.uk/Guidance/TA111 [Accessed 2009 Jan 19]

6. Robinson DM, Keating GM.Memantine: a review of its usein Alzheimer’s disease. Drugs 2006; 66 (11): 1515-34

7. RammesG,DanyszW,ParsonsCG.Pharmacodynamics ofme-mantine: an update. Curr Neuropharmacol 2008; 6 (1): 55-78

8. Parsons CG, Danysz W, Quack G. Memantine is a clinicallywell tolerated N-methyl-D-aspartate (NMDA) receptorantagonist: a review of preclinical data. Neuropharmacol-ogy 1999 Jun; 38 (6): 735-67

9. Parsons CG, Stoffler A, Danysz W. Memantine: a NMDAreceptor antagonist that improves memory by restorationof homeostasis in the glutamatergic system: too little acti-vation is bad, too much is even worse. Neuropharmacology2007 Nov; 53 (6): 699-723

10. Danysz W, Parsons CG, Mobius HJ, et al. Neuroprotectiveand symptomatological action of memantine relevant forAlzheimer’s disease: a unified glutamatergic hypothesis onthe mechanism of action. Neurotox Res 2000; 2 (2-3): 85-97

11. Gilling KE, Jatzke C, Parsons CG. Agonist concentrationdependency of blocking kinetics but not equilibrium blockof N-methyl-D-aspartate receptors by memantine. Neuro-pharmacology 2007; 53 (3): 415-20

12. Wenk GL, Parsons CG, Danysz W. Potential role ofN-methyl-D-aspartate receptors as executors of neurode-generation resulting from diverse insults: focus on mem-antine. Behav Pharmacol 2006 Sep; 17 (5-6): 411-24

13. Volbracht C, van Beek J, Zhu C, et al. Neuroprotective prop-erties of memantine in different in vitro and in vivo models ofexcitotoxicity. Eur J Neurosci 2006 May; 23 (10): 2611-22

14. Minkeviciene R, Banerjee P, Tanila H.Memantine improvesspatial learning in a transgenic mouse model of Alzheimer’sdisease. J Pharmacol Exp Ther 2004 Jun 10; 311 (2): 677-82

15. Van Dam D, De Deyn PP. Cognitive evaluation of disease-modifying efficacy of galantamine and memantine in theAPP23model. Eur Neuropsychopharmacol 2006; 16: 59-69

16. Pieta Dias C, Martins de Lima MN, Presti-Torres J, et al.Memantine reduces oxidative damage and enhances long-term recognition memory in aged rats. Neuroscience 2007Jun 8; 146 (4): 1719-25

17. Forest Pharmaceuticals Inc. Namenda (memantine HCL)tablets/oral solution: full prescribing information [online].Available from URL: http://www.frx.com/pi/namenda_pi.pdf [Accessed 2009 Jun 12]

18. Glodzik L, King KG, Gonen O, et al. Memantine decreaseshippocampal glutamate levels: a magnetic resonance spec-troscopy study. Prog Neuropsychopharmacol Biol Psy-chiatry 2008 May 15; 32 (4): 1005-12

19. Weiner MW, Graham SM, Hofbauer RK, et al. Memantinetreatment in patients with Alzheimer’s disease is associatedwith a slower right hippocampal volume loss: an open-label, multi-center trial [abstract no. P4-472]. 12th Interna-tional Conference on Alzheimer’s Disease and RelatedDisorders; 2008 Jul 26-31; Chicago (IL)

20. Sultzer DL, Melrose RJ, Harwood DG, et al. Effect ofmemantine treatment on regional cortical metabolism inAlzheimer’s disease [abstract no. P2-413]. 12th Interna-tional Conference on Alzheimer’s Disease and RelatedDisorders; 2008 Jul 26-31; Chicago (IL)

21. Mobius HJ, Stoffler A, Graham SM. Memantine hydro-chloride: pharmacological and clinical profile. Drugs To-day 2004; 40 (8): 686-95

22. European Medicines Agency. Axura film-coated tablets:summary of product characteristics [online]. Availablefrom URL: http://www.emea.eu/humandocs/PDFs/EPAR/axura/H-378-PI-en.pdf [Accessed 2009 Jun 12]

23. Kornhuber J, Kennepohl EM, Bleich S, et al. Memantine phar-macotherapy: a naturalistic study using a population pharma-cokinetic approach. Clin Pharmacokinet 2007; 46 (7): 599-612

24. Periclou A, Ventura D, Rao N, et al. Pharmacokinetic studyof memantine in healthy and renally impaired subjects.Clin Pharmacol Ther 2006 Jan; 79 (1): 134-43

25. Shua-Haim J, Smith J, Picard F, et al. Steady-state phar-macokinetics of rivastigmine in patients with mild tomoderate Alzheimer’s disease not affected by co-adminis-tration of memantine: an open-label, crossover, single-centre study. Clin Drug Investig 2008; 28 (6): 361-74

26. Reisberg B, Doody R, Stoffler A, et al. Memantine inmoderate-to-severe Alzheimer’s disease. N Engl J Med2003 Apr 3; 348 (14): 1333-41

27. Tariot PN, Farlow MR, Grossberg GT, et al. Memantinetreatment in patients with moderate to severe Alzheimerdisease already receiving donepezil: a randomized con-trolled trial. JAMA 2004 Jan 21; 291 (3): 317-24

28. van Dyck CH, Tariot PN, Meyers B, et al. A 24-week ran-domized, controlled trial of memantine in patients withmoderate-to-severe Alzheimer disease. Alzheimer Dis As-soc Disord 2007 Apr; 21 (2): 136-43

29. Reisberg B, DoodyR, Stoffler A, et al. A 24-week open-labelextension study of memantine in moderate to severe Alz-heimer disease. Arch Neurol 2006 Jan; 63 (1): 49-54

30. Atri A, Shaughnessy LW, Locascio JJ, et al. Long-term courseand effectiveness of combination therapy in Alzheimer dis-ease. Alzheimer Dis Assoc Disord 2008; 22 (3): 209-21

31. LopezOL, Becker JT,WahedAS, et al. Long-term effects of theconcomitant use of memantine with cholinesterase inhibitionin Alzheimer’s disease. J Neurol Neurosurg Psychiatry2009 Jun; 80 (6): 600-7

32. Vidal JS, Lacombe JM, Dartigues JF, et al. Evaluation ofthe impact of memantine treatment initiation on psycho-tropics use: a study from the French National Health Caredatabase. Neuroepidemiology 2008 Sep; 31 (3): 193-200

33. Doody RS, Tariot PN, Pfeiffer E, et al. Meta-analysis of six-month memantine trials in Alzheimer’s disease. AlzheimersDement 2007; 3 (1): 7-17

34. McShane R, Sastre AA, Minakaran N. Memantine formoderate and severe Alzheimer’s disease: systematic reviewand meta-analysis of 9 trials [abstract no. 02-04-04]. 12thInternational Conference on Alzheimer’s Disease and Re-lated Disorders; 2008 Jul 26-31; Chicago (IL)

35. Winblad B, Jones RW,Wirth Y, et al. Memantine in moderateto severe Alzheimer’s disease: a meta-analysis of randomisedclinical trials. Dement Geriatr Cogn Disord 2007; 24 (1): 20-7

36. Calabrese P, Essner U, Forstl H. Memantine (Ebixa) inclinical practice: results of an observational study. DementGeriatr Cogn Disord 2007; 24 (2): 111-7

37. Tocco M, Graham SM. Memantine treatment preventsclinical worsening: a pooled analysis of patients withmoderate to severe Alzheimer’s disease [abstract no. P4-397]. 12th International Conference on Alzheimer’s Dis-ease and Related Disorders; 2008 Jul 26-31; Chicago (IL)

38. Wilcock GK, Ballard CG, Cooper JA, et al. Memantine foragitation/aggression and psychosis in moderately severe tosevere Alzheimer’s disease: a pooled analysis of 3 studies.J Clin Psychiatry 2008 Mar; 69 (3): 341-8

896 McKeage


39. Emre M, Mecocci P, Stender K. Pooled analyses on cognitiveeffects of memantine in patients with moderate to severeAlzheimer’s disease. J Alzheimers Dis 2008 Jun; 14 (2): 193-9

40. Gauthier S, Loft H, Cummings J. Improvement in beha-vioural symptoms in patients with moderate to severeAlzheimer’s disease by memantine: a pooled data analysis.Int J Geriatr Psychiatry 2008 May; 23 (5): 537-45

41. Doody R, Wirth Y, Schmitt F, et al. Specific functional ef-fects of memantine treatment in patients with moderate tosevere Alzheimer’s disease. Dement Geriatr Cogn Disord2004; 18: 227-32

42. Fokkens J, Ferris S, Robert P, et al. Assessment of mem-antine treatment effects on language [abstract no. P4-365].12th International Conference on Alzheimer’s Disease andRelated Disorders; 2008 Jul 26-31; Chicago (IL)

43. Wilkinson D, Andersen HF. Analysis of the effect of mem-antine in reducing the worsening of clinical symptoms inpatients with moderate to severe Alzheimer’s disease.Dement Geriatr Cogn Disord 2007; 24 (2): 138-45

44. Mecocci P, Bladstrom A, Stender K. Effects of memantine oncognition in patients with moderate to severe Alzheimer’sdisease: post-hoc analyses of ADAS-cog and SIB total andsingle-item scores from six randomized, double-blind, placebo-controlled studies. Int J Geriatr Psychiatry 2009; 24: 532-8

45. Gauthier S, Wirth Y, Mobius HJ. Effects of memantine onbehavioural symptoms in Alzheimer’s disease patients: ananalysis of the Neuropsychiatric Inventory (NPI) data oftwo randomised, controlled studies. Int J Geriatr Psy-chiatry 2005 Apr 25; 20 (5): 459-64

46. Rive B, Vercelletto M, Damier FD, et al. Memantine en-hances autonomy in moderate to severe Alzheimer’s dis-ease. Int J Geriatr Psychiatry 2004 May; 19 (5): 458-64

47. LivingstonG, Katona C. The place of memantine in the treat-ment of Alzheimer’s disease: a number needed to treat ana-lysis. Int J Geriatr Psychiatry 2004 Sep 27; 19 (10): 919-25

48. Cummings JL, Schneider E, Tariot PN, et al. Behaviouraleffects of memantine in Alzheimer disease patients receiv-ing donepezil treatment. Neurology 2006 Jul 11; 67: 57-63

49. Feldman HH, Schmitt FA, Olin JT. Activities of daily livingin moderate-to-severe Alzheimer disease: an analysis of thetreatment effects of memantine in patients receiving stabledonepezil treatment. Alzheimer Dis Assoc Disord 2006; 20(4): 263-8

50. Schmitt FA, van Dyck CH, Wichems CH, et al. Cognitiveresponse to memantine in moderate to severe Alzheimerdisease patients already receiving donepezil: an exploratoryreanalysis. Alzheimer Dis Assoc Disord 2006; 20 (4): 255-62

51. Tariot PN, Saxton J, Tocco M, et al. Memantine preventsworsening of communication abilities in patients withmoderate to severe Alzheimer’s disease: mean change- andresponder analysis [abstract no. NR5-057]. 161st AnnualMeeting of the American Psychiatric Association; 2008May 3-8; Washington, DC, 238

52. van Dyck CH, Schmitt FA, Olin JT. A responder analysis ofmemantine treatment in patients with Alzheimer diseasemaintained on donepezil. Am J Geriatr Psychiatry 2006May; 14 (5): 428-37

53. PerneczkyR,Wagenpfeil S, KomossaK, et al.Mapping scoresonto stages: mini-mental state examination and clinical de-mentia rating. Am J Geriatr Psychiatry 2006; 14 (2): 139-44

54. Fokkens J, Ferris S, Robert P, et al. Development of a languageassessment tool for patients with Alzheimer’s disease [abstractno. P3-009]. 12th International Conference on Alzheimer’sDisease and Related Disorders; 2008 Jul 26-31; Chicago (IL)

55. Antonanzas F, Rive B, Badenas JM, et al. Cost-effectivenessof memantine in community-based Alzheimer’s diseasepatients: an adaptation in Spain. Eur J Health Econ 2006Jun; 7 (2): 137-44

56. Gagnon M, Rive B, Hux M, et al. Cost-effectiveness ofmemantine compared with standard care in moderate-to-severe Alzheimer disease in Canada. Can J Psychiatry 2007Aug; 52 (8): 519-26

57. Jones RW, McCrone P, Guilhaume C. Cost effectiveness ofmemantine in Alzheimer’s disease: an analysis based on aprobabilistic Markov model from a UK perspective. DrugsAging 2004; 21 (9): 607-20

58. Jonsson L. Cost-effectiveness of memantine for moderate tosevere Alzheimer’s disease in Sweden. Am J Geriatr Phar-macother 2005 Jun; 3 (2): 77-86

59. Weycker D, Taneja C, Edelsberg J, et al. Cost-effectivenessof memantine in moderate-to-severe Alzheimer’s diseasepatients receiving donepezil. Curr Med Res Opin 2007May; 23 (5): 1187-97

60. Farlow MR, Graham SM, Alva G. Memantine for thetreatment of Alzheimer’s disease: tolerability and safetydata from clinical trials. Drug Saf 2008; 31 (7): 577-85

61. Jones RW, Bayer A, Inglis F, et al. Safety and tolerabilityof once-daily versus twice-daily memantine: a rando-mised, double-blind study in moderate to severe Alzhei-mer’s disease. Int J Geriatr Psychiatry 2007 Mar; 22 (3):258-62

62. Raina P, Santaguida P, Ismaila A, et al. Effectiveness ofcholinesterase inhibitors and memantine for treating de-mentia: evidence review for a clinical practice guideline.Ann Intern Med 2008 Mar 4; 148 (5): 379-97

63. Qaseem A, Snow V, Cross Jr JT, et al. Current pharmaco-logic treatment of dementia: a clinical practice guidelinefrom the American College of Physicians and the AmericanAcademy of Family Physicians. Ann Intern Med 2008Mar; 148 (5): 370-8

64. Iliffe S. The National Institute for Health and ClinicalExcellence (NICE) and drug treatment for Alzheimer’sdisease. CNS Drugs 2007; 21 (3): 177-84

65. Rafii MS, Aisen PS. Recent developments in Alzheimer’sdisease therapeutics. BMC Medicine [online]. Availablefrom URL: http://www.biomedcentral.com/content/pdf/1741-7015-7-7.pdf [Accessed 2009 Mar 12]

Correspondence: Kate McKeage, Adis, a Wolters KluwerBusiness, 41 Centorian Drive, Private Bag 65901, MairangiBay, North Shore 0754, Auckland, New Zealand.E-mail: [email protected]



memantine

Documents