Memantine for dementia

/Home/Alzheimer's Treatments/Prescription Medicines for Dementia Patients/About Alzheimer's Prescription Medicines/Glutamate Blockers for Treating Alzheimer's

Background: Memantine, a low affinity antagonist to glutamate NMDA receptors, may prevent excitatory neurotoxicity in dementia.

Objectives: To determine efficacy and safety of memantine for people with Alzheimer's disease (AD), vascular (VD) and mixed dementia.

Search strategy: The Specialized Register of the Cochrane Dementia and Cognitive Improvement Group was searched on 8 February 2006. This register contains references from all major healthcare databases and many ongoing trial databases and is updated regularly. In addition, the search engines Copernic and Google were used to identify unpublished trials through inspection of the websites of licensing bodies like the FDA , EMEA and NICE and of companies' websites (Lundbeck, Merz, Forest, Suntori etc) and clinical trials registries.

Selection criteria: Double-blind, parallel group, placebo-controlled, randomized trials of memantine in people with dementia.

Data collection and analysis: Data were pooled where possible. Intention-to-treat (ITT) and observed case (OC) analyses are reported.

Main results

1. Moderate to severe AD. Two out of three six month studies show a small beneficial effect of memantine. Pooled data indicate a beneficial effect at six months on cognition (2.97 points on the 100 point SIB, 95% CI 1.68 to 4.26, P < 0.00001), activities of daily living (1.27 points on the 54 point ADCS-ADLsev, 95% CI 0.44 to 2.09, P = 0.003) and behaviour (2.76 points on the 144 point NPI, 95% CI 0.88 to 4.63, P=0.004), supported by clinical impression of change (0.28 points on the 7 point CIBIC+, 95% CI 0.15 to 0.41, P < 0.0001).

2. Mild to moderate AD. Pooled data from three unpublished studies indicate a marginal benefical effect at six months on ITT cognition (0.99 points on the 70 point ADAS-Cog, 95% CI 0.21 to 1.78, P = 0.01) which was barely detectable clinically (0.13 CIBIC+ points, 95% CI 0.01 to 0.25, P = 0.03) but no effect on behaviour, activities of daily living or OC analysis of cognition.

3. Mild to moderate vascular dementia. Pooled data from two six month studies indicated a small beneficial effect of memantine on cognition (1.85 ADAS-Cog points, 95% CI 0.88 to 2.83, P = 0.0002), and behaviour (0.84 95% CI 0.06 to 0.91, P = 0.03) but this was not supported by clinical global measures.

4. Patients taking memantine were slightly less likely to develop agitation (134/1739, 7.7% versus 175/1873, 9.3% OR 0.78, 95% CI 0.61 to 0.99, P = 0.04). This effect was slightly larger, but still small, in moderate to severe AD (58/506 [12%] vs 88/499 [18%]; OR = 0.6, 95% CI 0.42 to 0.86, P = 0.005). There is no evidence either way about whether it has an effect on agitation which is already present.

5. Memantine is well tolerated.

Authors' conclusions: Memantine has a small beneficial effect at six months in moderate to severe AD. In patients with mild to moderate dementia, the small beneficial effect on cognition was not clinically detectable in those with vascular dementia and was detectable in those with AD. Memantine is well tolerated.


Plain language summary

Some evidence of efficacy of memantine for dementia

Memantine has a small beneficial, clinically detectable effect on cognitive function and functional decline measured at 6 months in patients with moderate to severe Alzheimer's Disease (AD). In patients with mild to moderate dementia, the small beneficial effect on cognition was not clinically detectable in those with vascular dementia and barely detectable in those with AD. It is well tolerated. Slightly fewer patients with moderate to severe AD taking memantine develop agitation, but there is no evidence either way about whether it has an effect on agitation which is already present.


McShane R, Areosa Sastre A, Minakaran N. Memantine for dementia. Cochrane Database of Systematic Reviews 2006, Issue 2. Art. No.: CD003154.

Background

Alzheimer's Disease is the commonest cause of dementia, and is found in at least 50% of autopsies of demented patients. The prevalence of the disease is approximately 1 to 2% at the age of 65, but doubles every 5 years to at least the age of 90. The disease is progressive. Memory impairment is the cardinal feature and is usually one of the first characteristics observed. As Alzheimer's disease progresses, memory and other cognitive deficits become increasingly evident and interfere with usual activities. Substantial progress has been made in understanding the basic neurobiology of this disease and new drugs have become available for symptomatic treatment. The focus of symptomatic treatment of Alzheimer's disease has been the enhancement of cholinergic transmission. However, there are other possible therapeutic approaches based on neurotransmitter substitution or modulation including serotoninergic, noradrenergic substances or neuropeptides, and those acting on excitatory amino acid receptors, such as for glutamic acid (Emre 2001).

Vascular dementia is the second most common cause of dementia in western societies. It affects 1 to 20% of people aged 65 years or older. The cognitive decline of vascular dementia is secondary to some form of vascular injury to the brain. It is a heterogeneous condition and clinical manifestations differ depending on the size and location of the cerebrovascular lesions. There is no therapy proven to reverse the neurological damage. At present, interventions focus on the prevention of the disease by better control of cardiovascular risk factors.

Mixed dementia, due to a combination of both Alzheimer type pathology and ischaemia, is being increasingly recognised. In autopsy studies mixed dementia has been reported as accounting for between 0% and 55% of cases of dementia. In addition to co-occurrence due simply to chance, Alzheimer's disease and vascular dementia may have aetiological or pathogenetic factors in common (Kalaria 1999). In comparison with sufferers from Alzheimer's disease, people with mixed dementia show higher frequencies of depressed mood, focal motor or sensory findings and gait disturbances, but the neuropsychological pattern is not distinctive. There have been few studies relevant to the prevention and treatment of mixed dementia (Zeckry 2002).

L-glutamate is the main excitatory neurotransmitter in the central nervous system, implicated in neural transmission, learning, memory processes and neuronal plasticity (Sucher 1996). There is evidence that enhancement of the excitatory action of this amino acid plays a role in the pathogenesis of Alzheimer's disease and in the damage due to an ischaemic stroke (Cacabelos 1999). However, physiological glutamate activity is required for normal brain activity and so cannot be abolished completely (Kornhuber 1997). Low affinity N-methyl-D-aspartate (NMDA) type receptor antagonists, such as memantine, might prevent excitatory amino acid neurotoxicity without interfering with the physiological actions of glutamate necessary for learning and memory.

Memantine was first synthesized at Eli Lilly as an agent to lower elevated blood sugar level, but was ineffective. In 1972, Merz applied for a German patent for memantine as a potential treatment for various neurological diseases, citing evidence for a beneficial effect on central nervous system activity. In 1975 and 1978, patents were granted in Germany and the USA respectively (Parsons 1999). Since then the drug has been tested in animal models of cognition and found to reverse deficits in learning and synaptic plasticity (Möbius 1999). It has been tested in several phase II and III clinical studies in the last fifteen years. These trials involved people with Alzheimer's, vascular, and mixed dementia at different stages.

Memantine (Ebixa® and Axura® (Europe) and Namenda® (USA)) was approved in February 2002 by the European Agency for the Evaluation of Medical Products (EMEA) (EMEA 2004) for the treatment of "moderately severe to severe Alzheimer's disease" and in 2003 by the USA Food and Drug Administration (FDA) (Anonymous 2003; Forest 2003) for the treatment of moderate to severe Alzheimer's disease. Memantine has not been approved for vascular dementia or earlier stages of Alzheimer disease in either jurisdiction Applications have been unsuccessfully filed with the FDA and EMEA for licenses for the treatment of mild to moderate Alzheimer's disease (Forest 2005b; Lundbeck 2005).

The UK National Institute for Clinical Excellence published consultation documents on the cost-effectiveness of drugs for Alzheimer's disease, including memantine in March 2005 and January 2006 (NICE 2005; NICE 2006). The consultation documents recommend that memantine should not be publicly funded for the treatment of moderately severe to severe AD, except as part of studies designed to generate data on long-term outcomes, disease progression, quality of life and costs. A final decision is expected in July 2006.


Objectives

The primary aim of the review was to assess the efficacy and safety of memantine for the treatment of dementia, as revealed in clinical trials involving people with Alzheimer's, vascular, or mixed dementia.


Methods

Criteria for considering studies for this review

Types of studies
Studies were selected for this review if they fulfilled the following criteria:
1-Clinical trials in Alzheimer's disease, vascular dementia, mixed dementia, or unspecified dementia
2-Double-blind, parallel-group, placebo-controlled, with randomized and unconfounded treatment assignment to placebo or memantine
3-Sample selection criteria were specified
4-Outcome instruments were specified
5-Duration was specified

Types of participants
People with Alzheimer's, vascular, mixed or unspecified dementia of all degrees of severity, treated as in- or out-patients.

Types of interventions
Treatment with memantine at any dose and by any route of administration in an acceptable clinical trial.

Types of outcome measures
The primary outcomes of interest are:

1- Clinical Global Impression
2- Cognitive function
3- Functional performance
4- Behavioural disturbance
5- Effect on carers
6- Quality of Life
7- Incidence and severity of adverse effects
8- Institutionalization
9- Costs


Search methods for identification of studies

The trials were identified from an updated search of the trial-based Specialized Register of the Cochrane Dementia and Cognitive Improvement Group on 8 February 2006 using the terms: memantin*, D-145, DMAA, DRG-0267, ebixa, axura and namenda

The Specialized Register at that time contained records from the following databases:

• CENTRAL: July 2005 (issue 3);
• MEDLINE: 1966 to 2005/08, week 2;
• EMBASE: 1980 to 2005/08, week 2;
• PsycINFO: 1887 to 2005/07;
• CINAHL: 1982 to 2004/07;
• SIGLE (Grey Literature in Europe): 1980 to 2004/06;
• ISTP (Index to Scientific and Technical Proceedings): to May 2000;
• INSIDE (BL database of Conference Proceedings and Journals): to June 2000;
• Aslib Index to Theses (UK and Ireland theses): 1970 to March 2003;
• Dissertation Abstract (USA): 1861 to March 2003;
• http://clinicalstudies.info.nih.gov/;
• National Research Register (issue 3/2005)
• ClinicalTrials.gov: last searched 1 January 2006;
• LILACS (Latin American and Caribbean Health Science Literature): last searched April 2003
• http://www.forestclinicaltrials.com/: last searched 1 September 2005
• ClinicalStudyResults.org: last searched 1 February 2006
• http://www.lillytrials.com/index.shtml: last searched 28 August 2005
• ISRCTN Register: last searched 1 September 2005
• IPFMA Clinical trials Register: www.ifpma.org/clinicaltrials.html: last searched September 2005


The search strategies used to identify relevant records in MEDLINE, EMBASE, PsycINFO, CINAHL and LILACS can be found in the Group's module.

Press releases from Merz, Lundbeck and Forest Laboratories were searched (February 2006) and all releases pertaining to Memantine were examined. Enquiries were made of Forest Laboratories, Merz and Lundbeck (11/08/04) as to whether there are any further completed unpublished Phase 3 studies beyond these two. A further request in January 2006 to Lundbeck for copies of all conference posters was referred to Merz. None have yet been received (February 9th 2006)

The Forest and Lundbeck clinical trials registry was examined in February 2006 (Forest 2005a).

We used Copernic to search the Internet for the terms of the form ' memantine MD-01', ' memantine MD-02' etc on the basis that this appeared to be the numbering system of trials. We searched for Ebixa, Namenda and Axura as well.


Data collection and analysis

SELECTION OF STUDIES
Abstracts of the references retrieved by the search were read by one reviewer (AAS) who discarded those that were clearly not eligible for inclusion. One reviewer (AAS) studied the full text of the remaining references and selected studies for inclusion. This selection was checked by Jacqueline Birks and Rupert McShane (Coordinating Editors CDCIG). Any disparity in the final lists was resolved by discussion in order to arrive at the final list of included studies. Abstracts from an updated search of the Cochrane Specialist Register in October 2005 and February 2006 were reviewed for new RCTs by RM.

QUALITY ASSESSMENT
The reviewer assessed the methodological quality of each trial using the Cochrane Collaboration guidelines (Mulrow 1997). In Category A (adequate), the report describes allocation of treatment by: (i) some form of centralized randomized scheme, such as having to provide details of an enrolled participant to an office by phone to receive the treatment group allocation; (ii) some form of randomization scheme controlled by a pharmacy; (iii) numbered or coded containers, such as in a pharmaceutical trial in which capsules from identical-looking numbered bottles are administrated sequentially to enrolled participants; (iv) an on-site or coded computer system, given that the allocations were in a locked, unreadable file that could be accessed only after inputting the characteristics of an enrolled participant; or (v) if assignment envelopes were used, the report should at least specify that they were sequentially numbered, sealed, and opaque; (vi) other combinations of described elements of the process that provide assurance of adequate concealment. Category B (intermediate) is where the report describes allocation of treatment by: (i) use of a "list" or "table" to allocate assignments; (ii) use of "envelopes" or "sealed envelopes"; (iii) stating the study as "randomized" without further detail. Category C (inadequate) is where the report describes allocation of treatment by: (i) alternation; (ii) reference to case record numbers, dates of birth, day of week, or any other such approach; (iii) any allocation procedure that is entirely transparent before assignment, such as an open list of random numbers or assignments. Empirical research has shown that lack of adequate allocation concealment is associated with bias. Trials with unclear concealment measures have been shown liable to yield more pronounced estimates of treatment effects than trials that have taken adequate measures to conceal allocation schedules, but the effect is less pronounced than in inadequately concealed trials (Chalmers 1983; Schulz 1995). Trials were to be considered if they conformed to categories A or B, but those falling into category C were excluded. Other aspects of trial quality were not assessed by a scoring system although details were noted of blinding, appropriateness of methods and the number of patients lost to follow-up.

DATA COLLECTION
Data for the meta-analyses were based on reported summary statistics for each study. The summary statistics required for each trial and each outcome for continuous data were the mean change from baseline, the standard error of the mean change, and the number of patients for each treatment group at each assessment. Where changes from baseline were not reported, the mean, standard deviation and the number of patients for each treatment group at each time point were extracted.

For binary data the numbers in each treatment group and the numbers experiencing the outcome of interest were sought. For the global impression of change, the endpoint itself is of clinical relevance as all patients are by definition at the same baseline score.

The baseline assessment is defined as the latest available assessment prior to randomization, but no longer than two months prior.

For the intention-to-treat analyses data were sought for each outcome measure on every patient randomized, irrespective of compliance. Analyses of completers was restricted to data on every patient who completed the study on treatment.

DATA ANALYSIS
The outcomes measured in clinical trials of dementia and cognitive impairment often arise from ordinal rating scales. Where the rating scales used in the trials have a reasonably large number of categories (more than 10) the data were treated as continuous variables arising from a normal distribution.

Summary statistics (n, mean and standard deviation) were required for each rating scale at each assessment time for each treatment group in each trial for change from baseline. When change from baseline results are not reported, the required summary statistics were calculated from the baseline and assessment time treatment group means and standard deviations. In this case a zero correlation between the measurements at baseline and assessment time was assumed. This method overestimates the standard deviation of the change from baseline, but this conservative approach is considered to be preferable in a meta-analysis.

The meta-analyses sometimes required the combination of data from trials that did not use the same rating scale to assess an outcome. The measure of the treatment difference for any outcome is the weighted mean difference (WMD) when the pooled trials used the same rating scale or test, and the standardized mean difference (SMD) - the absolute mean difference divided by the standard deviation - when they used different rating scales or tests.

The duration of the trials varied from 4 to 28 weeks. The range was considered too great to combine all trials into one meta-analysis. Separate analyses are therefore presented for various time periods.

For binary outcomes, such as dead or alive, clinical improvement or no clinical improvement, the odds ratio was used to measure treatment effect.

Overall estimates of the treatment difference are presented. In all cases the overall estimate from a fixed effects model is presented and a test for heterogeneity, using a standard chi-square statistic, has been performed. If a test of heterogeneity was negative then a weighted estimate of the typical treatment effect across trials, the 'typical odds ratio' (i.e. the odds of an unfavourable outcome amongst treatment-allocated patients compared to the corresponding odds amongst controls) was calculated using Peto's log-rank test adapted for ordinal data.

Data from patients who were treated 'per protocol' were used in 'observed case' analyses when the latter were not available. Observed case data were substituted for last observation carried forward data in two published studies (9408/Orgogozo 2002; 9202/Wilcock 2002) where the latter were not available for some measures.


Results

Description of studies
See: Characteristics of included studies; Characteristics of excluded studies; Characteristics of ongoing studies.

Twelve trials were identified that met inclusion criteria. They studied the efficacy and tolerability of various dosages of memantine in different types of dementia and at different stages of the disease (Ditzler 1991; Gortelmeyer 1992; 9408/Orgogozo 2002; Pantev 1993; MD-10/Peskind 2004; 9605/Reisberg 2003; MD-02/Tariot 2004; 9202/Wilcock 2002; 9403/Winblad 1999; MD-01; 99679(Lundbeck); MD-12). These trials were sponsored by Merz Pharma KGaA, Frankfurt, Germany (Ditzler 1991; Gortelmeyer 1992; 9408/Orgogozo 2002; Pantev 1993; 9605/Reisberg 2003; 9202/Wilcock 2002; 9403/Winblad 1999), and Forest Laboratories (MD-10/Peskind 2004; MD-02/Tariot 2004; MD-01; MD-12).

MD-01, MD-10/Peskind 2004, MD-12, 99679(Lundbeck) have yet to be published in peer reviewed journals. Reference to 99679(Lundbeck) in the Lundbeck clinical trial registry states 'publication process ongoing' (February 9th 2006). The study was completed before January 2004.

Three previously unidentified completed studies are referred to on the Lundbeck Clinical Trial Registry as 'publication process ongoing': 10116 (Lundbeck) is a double blind placebo controlled parallel group studies in Alzheimer's disease which evaluates efficacy, safety and tolerability and was started in January 2002; Studies 10113 and 10114 are double blind placebo controlled parallel group evaluations of safety and tolerabiliy in Alzheimer's disease. Although we were told that two of these have been reported as conference posters and that Merz would forward these, we had not received them by the time of submission of this review.

MRZ-9105 was a 12 week study in Portugal of 27 patients with 'mild to moderate severe stages of primary dementia'. MRZ-9104 was a 13 week study in France of 56 patients with AD of severity that we were unable to establish. MRZ-9206 was a 14 week study in Sweden of 56 patients with 'moderately severe VD'. We have not been able to identify the results or any associated publications or announcements belonging to these three studies.

METHODS
All included trials were of parallel-group design. There were nine phase III studies that lasted between 12 and 28 weeks (9408/Orgogozo 2002; 9605/Reisberg 2003; 9202/Wilcock 2002; 9403/Winblad 1999; MD-02/Tariot 2004; MD-10/Peskind 2004; MD-01; MD-12; 99679(Lundbeck)). The other three included studies were phase II trials that lasted four or six weeks (Ditzler 1991; Gortelmeyer 1992; Pantev 1993).

PARTICIPANTS
The number of participants ranged from 60 to 579.

The diagnosis of dementia was established using the last versions of the Diagnostic and Statistical Manual of Mental Disorders (DSM III-R; DSM IV) (Gortelmeyer 1992; 9408/Orgogozo 2002; Pantev 1993; 9605/Reisberg 2003; 9403/Winblad 1999). Two studies involved only people with vascular dementia defined by the National Institute of Neurological Disorders and Stroke and the Association International pour la Recherche et l'Enseignement en Neurosciences (NINDS-AIREN) (9408/Orgogozo 2002; 9202/Wilcock 2002). Six studies were restricted to people with Alzheimer's disease diagnosed according to the criteria of the National Institute of Neurologic, Communicative Disorders and Stroke and Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA: McKhann 1984) (MD-10/Peskind 2004; MD-12; 99679(Lundbeck); 9605/Reisberg 2003; MD-02/Tariot 2004; MD-01). Three studies included both types of dementia in various proportions (Ditzler 1991; Gortelmeyer 1992; 9403/Winblad 1999). In these studies, the Hachinski score was used to differentiate between Alzheimer's disease and vascular dementia (Ditzler 1991; Gortelmeyer 1992; 9403/Winblad 1999). In one trial there is no record of an attempt to distinguish different types of dementia (Pantev 1993).

Seven studies included participants with mild to moderate impairment as defined by scores on the Mini Mental State Examination (MMSE) of Folstein 1975 (9408/Orgogozo 2002; MD-10/Peskind 2004; 99679(Lundbeck); MD-12; 9202/Wilcock 2002) or the Sandoz Clinical Assessment Geriatric Scale (SCAG) (Ditzler 1991; Gortelmeyer 1992). The participants in Pantev 1993 were equally divided between mild, moderate and severe disease as assessed by the SCAG. Three studies enrolled severely impaired patients according to the MMSE and the Global Deterioration Scale (GDS) (9605/Reisberg 2003; 9403/Winblad 1999) or MMSE score alone (MD-02/Tariot 2004).

The details of the participants at baseline are reported in Table 1. 9408/Orgogozo 2002 and MD-02/Tariot 2004 described the characteristics of the participants more comprehensively than the other studies. Most of the studies specified race, duration of the diagnosis or the number of dependent patients at baseline, but only three trials reported concomitant diseases (Ditzler 1991; Pantev 1993; MD-02/Tariot 2004). One study reported previous use of drugs for dementia by the participants but without specifying types (9403/Winblad 1999). Two studies (MD-02/Tariot 2004; MD-12) enrolled only patients with a diagnosis of AD and ongoing cholinesterase inhibitors therapy with donepezil for more than 6 months before entrance into the study and at a stable dose (5 to 10 mg/dl) for at least 3 months. The exclusion criteria are not specified in one study which has only been reported as a poster at conferences (MD-10/Peskind 2004)

INTERVENTIONS
The trials studied different dosages of memantine with placebo. The doses used ranged from 10 to 30 mg/day. Most of the trials started with low doses progressively increasing to 20 mg/day to target levels. MD-02/Tariot 2004 and MD-12 compared the efficacy and safety of memantine in patients receiving stable treatment with donepezil.

OUTCOME MEASURES
The following range of outcome measures was used in the trials. Additional Table 2 summarises their use in the included studies.

1- Global rating scales:
* Clinician's Interview-Based Impression of Change scale (CIBIC-Plus) provides a global rating of function in four areas, general, cognitive, behaviour and activities of daily living. All participants are scored as 4 at baseline and subsequent assessments on a scale of 1 to 7 are relative to baseline, with 1 showing marked improvement and 7 marked worsening. Information is obtained from the caregiver and the patient. There are different versions: the Alzheimer's Disease Cooperative Study format (Schneider 1997) and the New York version (Reisberg 1997)
* The Clinical Global Impression of Change (CGIC) (Guy 1976) is a global rating of all domains of a patient's current condition in comparison with baseline. It is a seven point scale, from 1 (very much improved) to 7 (very much worse), 4 indicating no change. The assessment is conducted by the same clinician at both time points with input from relatives or carers.
* Physician's global impression. This unvalidated four point rating of the dementia syndrome and patient's general health status was used in Ditzler 1991.
*Clinical Global Impression (CGI). This seven point rating of severity of illness was used in Gortelmeyer 1992 to assess whether patients improved, remained unchanged or worsened.
* The Sandoz Clinical Assessment Geriatric Scale (SCAG) (Shader 1974) is a physician rating. It consists of 18 items and an overall impression (item 19), all rated on a seven-point format. There are five sub-scores: cognitive disturbances, disturbances in social behaviour, lack of drive, affective disturbances, somatic disturbances. Item 19 was used in Pantev 1993.

2- Cognitive Tests:
* The cognitive part of the Alzheimer's Disease Assessment Scale (ADAS-Cog) (Rosen 1984) comprises 11 individual tests, spoken language ability (0 to 5), comprehension of spoken language (0 to 5), recall of test instructions (0 to 5), word finding difficulty (0 to 5), following commands (0 to 5), naming object (0 to 5), construction drawing (0 to 5), ideational praxis (0 to 5), orientation (0 to 8), word recall (0 to 10) and word recognition (0 to 12). The total score ranges from 0 to 70, the high score indicating greater impairment.
* Syndrom-Kurztest determines patient attention and memory disturbances (Kim 1993).
* Severe Impairment Battery (SIB) (Schmitt 1997) evaluates cognitive performance in advanced Alzheimer's Disease. It is a 51 item scale which assesses social interaction, memory, language, visuospatial ability, attention, praxis and construction. The scores range from 0 (greatest impairment) to 100.

3- Activities of Daily Living:
* Alzheimer's Disease Cooperative Study-Activities of Daily Living (ADCS-ADL) (Galasko 1997) was specifically designed to assess functional capacity over a broad range of severity in patients with Alzheimer's disease. The 19 item ADCS-ADLsev19 has 54 points. The 23 item ADCS-ADL23 has 78 points.
* Activities of daily living test. This scale evaluates patients' abilities to cope with five instrumental task under the guidance of a psychologist.
* Behavioural Rating Scale for Geriatric Patients (BGP) (Van der Kam 1989) is a 45 item observer-rated scale for the assessment of functional and behavioural disturbances of geriatric patients, performed by nursing staff. The BGP contains several sub scales: care dependence, aggressiveness, physical, mental, disability and depressiveness, and inactivity. The "care dependence" scale consists of 23 items, measures activities of daily living and has the highest reliability and validity.
* Nurse's Observational Scale for Geriatric Patients (NOSGER) (Spiegel 1991) contains 30 items of behaviour, each rated in a 5-point scale according to frequency of occurrence. Item scores are summarized into 6 dimension scores (memory, instrumental activities of daily life, self-care activities of daily living, mood, social behaviour, and disturbance behaviour).

4- Mood and Behavioural measures:
* Neuropsychiatric Inventory (NPI) (Cummings 1994) assesses the frequency and the severity of behavioral and neuropsychiatric symptoms in patients with dementia based on an interview with the caregiver. There are 12 items with a total score ranging from 0 to 144.
* Nurses Observation Scale for Inpatient Evaluation (NOSIE) (Honigfeld 1974) assesses behaviour of psychiatric patients. It comprises 30 items of behaviour and the frequency of their occurrence. NOSIE subscale scores are for social competence, social interest, personal neatness, irritability, manifest psychosis, retardation, depression. Increasing values of the NOSIE index are indicative of improvement.
* Sandoz Clinical Assessment Geriatric Scale (SCAG) (Shader 1974) is a physician rating scale. It consists of 18 items and an overall impression (item 19), all rated on a seven-point format. There are also five sub-scores: cognitive disturbances, disturbances in social behaviour, lack of drive, affective disturbances, somatic disturbances.
* Behavioural Rating Scale for Geriatric Patients (BGP) (Van der Kam 1989) is a 45 item observer-rated scale for the assessment of functional and behavioural disturbances of geriatric patients, performed by nursing staff. The BGP contains several sub scales: care dependence, aggressiveness, physical, mental, disability and depressiveness, and inactivity. The "care dependence" scale consists of 23 items, measures activities of daily living and has the highest reliability and validity.
* Nurse's Observational Scale for Geriatric Patients (NOSGER) (Spiegel 1991) contains 30 items of behaviour, each rated in a 5-point scale according to frequency of occurrence. Item scores are summarized into 6 dimension scores (memory, instrumental activities of daily life, self-care activities of daily living, mood, social behaviour, and disturbance behaviour).

5- Combination Scales
Several scales used in earlier studies combine elements that are now more commonly assessed using separate instruments. Such scales have been used as a different method from 'overall clinical impression' for the systematic global assessment of dementia. These combination scales typically have several subscales, the results of which are sometimes presented separately and can be included in meta-analyses.
* Gottfries-Brane-Steen Scale (GBS) (Gottfries 1982) is a 26 item, physician assessed observer scale based on caregiver's information and an interview with the patient. It comprises three subscales: motor performance, intellectual and emotional capacity, and a group of six symptoms commonly observed in dementia.
Other combination scales used in studies of memantine are the SCAG, NOSIE, NOSGER and BGP which are detailed above.

6- Cost of resource utilisation.
* Resource Utilization in Dementia questionnaire (RUD) (Wimo 1988) is a structured interview of the patient's caregiver consisting of a baseline questionnaire (assessing basic patient demographic data and specific events that occurred during the past 1 month) and follow-up questionnaire (assessing specific events that occurred during the past 3 months). The specific events included time spent caring for the patient, changes in the caregiver's work status, healthcare resource utilisation by the caregiver, healthcare resource utilisation by the patient and the patient's residential status.

7- Safety and tolerability were assessed by the frequency and severity of reported adverse effects.

Risk of bias in included studies
9202/Wilcock 2002, 9403/Winblad 1999 and 9408/ Orgogozo 2000 studies reported that they were conducted to GCP standard. Since 9605/Reisberg 2003, MD-02/Tariot 2004, MD-01 and MD-10/Peskind 2004, 99679(Lundbeck) and MD-12 have been submitted for licensing approval it is highly likely that they too were conducted to GCP standard. All the studies are described as double-blind and randomized, but only three described the method of concealment of allocation (Gortelmeyer 1992; 9605/Reisberg 2003; MD-02/Tariot 2004). 9605/Reisberg 2003 and 9408/ Orgogozo 2000 reported the use a Ran Code program for this purpose and MD-02/Tariot 2004 used a randomization list generated by the Department of Biostatistics at Forest Laboratories.

Effects of interventions
Presentation of the results is divided according to the aetiology and severity of the dementia of the trial participants and duration of studies (sections 1 to 5). In sections 6 and 7 the results of pooled analyses are presented.

1- Effect of memantine (20 mg/day) on patients with moderate to severe Alzheimer disease at 24 to 28 weeks:
Three trials were identified in which compared the effect of six months of memantine and placebo in patients with moderate to severe Alzheimer's disease. This is the group for which memantine is licensed. 9605/Reisberg 2003 and MD-01 studied patients with an MMSE of 3-14 and 5-14 respectively. MD-02/Tariot 2004 studied patients with an MMSE of 5-14 who were also receiving stable treatment with donepezil. The same main outcome measures were used in all three studies.

* Global Impression: After six months there was a significant difference in the clinical impression of change, as measured by the 7 point CIBIC-Plus (ITT- LOCF), in favour of memantine (0.28 CIBIC+points, 95% Confidence Interval (CI) 0.15 to 0.41, P < 0.0001).
* Cognition: There was a significant difference in cognitive function, as measured by the 100 point SIB (ITT-LOCF), in favour of memantine (2.97 SIB points, 95% CI 1.68 to 4.26, P < 0.00001). There was substantial inconsistency between the three studies (I2 = 74%). MD-01 found no statistically significant effect. The smallest study, 9605/Reisberg 2003, found the largest effect.
* Activities of daily living: There was a significant difference in ability to perform activities of daily living, as assessed using the 54 point ADCS-ADLsev19 scale (ITT-LOCF), in favour of memantine (1.27 ADCS-ADLsev points, 95% CI 0.44 to 2.09, P = 0.003).
* Mood and behaviour: Patients taking memantine had significantly less worsening of mood and behaviour, as assessed on the 144 point NPI (ITT-LOCF), by six months (2.76 NPI points, 95% CI 0.88 to 4.63, P=0.004). Agitation occurred less commonly as an adverse event in those taking memantine (58/506 [12%] compared to those taking placebo (88/499 [18%]; Peto odds ratio (OR) = 0.6, 95% CI 0.42 to 0.86, P = 0.005). No evidence was presented to suggest that mood and behaviour problems which were apparent at the time of study entry were more likely to resolve in those taking memantine.
*Dropouts: The drop-out rate was significantly lower in the group of patients receiving memantine (103/507, 20.3% versus 139/499, 27.9% OR 0.66, 95% CI 0.49 to 0.88, P = 0.005).
* Adverse events: Memantine was well tolerated. There was no significant difference in the number of patients suffering at least one adverse event. Agitation occurred less commonly as an adverse event in all three studies in those taking memantine (58/506 [12%]) compared to those taking placebo (88/499 [18%]; OR = 0.6, 95% CI 0.42 to 0.86, P = 0.005). There were no other consistent differences in adverse events and no significant differences between memantine and placebo.
* Resource utilisation and cost effectiveness: Resource utilisation data and a cost analysis are available only for the 9605/Reisberg 2003 study (Wimo 2003). Based on data from the third and sixth month of the study, caregivers of patients taking memantine spent fewer hours/month in caregiving tasks and supervision. The unadjusted difference of 42 hours in the treated per protocol population did not reach significance (95% CI 113 to -29, P = 0.2). However, the difference was statistically significant (52 hours, 95%CI 95 to 7, P = 0.02) following adjustment for baseline caregiver time, residential status, caregiver gender, patient gender and caregiver-patient relationship, and in the LOCF-ITT population (46 hours 95%CI 10 to 81, P = 0.01). Of patients living in the community at the start of the study, 5/66 patients taking placebo and 1/84 patients taking memantine were admitted to an institution. The statistical significance of this depends on the test used. Overall costs to carers (in 1999 $US) were lower in the memantine, compared with the placebo group (824$US, 95% CI 906 to 742, P < 0.00001). Direct medical costs for the patient were greater in the memantine group (164 $US, 95% CI 161 to 169, P < 0.00001). Total societal costs were $1090 per month less in the group taking memantine (95%CI 1955 to 225, P = 0.01).

2-Effect of memantine (20 mg/day) on patients with mild to moderate Alzheimer disease at 24 weeks:
Three trials were identified that studied the effect of memantine on patients with mild to moderate Alzheimer's disease (MD-10/Peskind 2004; 99679(Lundbeck); MD-12).

*Global Impression: There was a significant difference in the clinical impression of change measured by the CIBIC-Plus (ITT) after 24 weeks (0.13 points, 95% CI 0.01 to 0.25, P = 0.03)
* Cognition: In the intention-to-treat analysis there was a significant difference in favour of memantine on the 70 point ADAS-Cog (0.99 ADAS-Cog points, 95% CI 0.21 to 1.78, P = 0.01). Conference presentation of observed case data from MD-10/Peskind 2004 showed no significant difference: 1.10 ADAS-Cog points, 95% CI 2.52 to -0.32. Observed case data were not available for the other two trials.
* Function: There was no significant difference in ADCS-ADL23 scores at 24 weeks (0.20 ADCS-ADL23 points, 95% CI -0.87 to 21.27, P = 0.72)
* Mood and Behaviour: There was no significant difference in NPI scores at 24 weeks (-0.25 NPI points, 95% CI -1.48 to 0.98, P = 0.69)
*Dropouts: There was no significant difference in drop-out rate (106/736, 14.4% versus 74/570, 13.0% OR 1.16, 95% CI 0.83 to 1.60, P = 0.38).
*Adverse events: There was no significant difference in the number of patients experiencing at least one adverse event (493/736, 67.0% versus 397/570, 69.6% OR 1.04, 95% CI 0.81 to 1.33, P = 0.76). In contrast to the finding in moderate-to-severe AD, there was no difference in the number of patients experiencing agitation as an adverse event (37/736, 5.0% versus 36/570, 6.3% OR 0.91, 95% CI 0.57 to 1.46, P = 0.70). Data from other individual adverse events are not available for all three trials. Significantly more people on memantine suffered somnolence by the end of the treatment (OR: 7.49, 95% CI 1.68 to 33.38, P = 0.008) in MD-10/Peskind 2004.

3-Effect of memantine (20 mg/day) on patients with mild to moderate vascular dementia at 28 weeks:
Two trials studied the effect of memantine on patients with mild to moderate vascular dementia (9408/Orgogozo 2002; 9202/Wilcock 2002).

* Global Impression: There was no difference in the clinical global for memantine (20 mg/day) compared with placebo at 28 weeks (WMD 0.03, 95% CI -0.13 to 0.19, P = 0.72)
* Cognition: Both studies provided ADAS-Cog data (9408/Orgogozo 2002 (ITT-LOCF), 9202/Wilcock 2002 (OC)). The change from baseline at the 28 weeks analysis gave statistically significant results in favour of memantine for 20 mg/day (1.85 ADASCog points, 95% CI 0.88 to 2.83, P = 0.0002).
* Activities of daily living: On the NOSGER self care subscale there was no difference between memantine and placebo groups (0.12, 95% CI -0.43 to 0.67, P = 0.66).
* Combination scales: Both studies reported data for memantine 20 mg/day using the GBS change from baseline at 28 weeks and found no statistically significant differences from placebo (WMD -1.81, 95% CI -4.21 to 0.58, P = 0.14) There were no differences in the NOSGER (complete scale) at 28 weeks in the same studies ( WMD -0.92, 95% CI -2.90 to 1.05, P = 0.4).
* Behaviour: There was significantly less agitation in the memantine group compared with placebo (20/460 [4.3%] versus 34/440 [7.7%], OR 0.54, 95%CI 0.31 to 0.96, P = 0.04) . On the NOSGER disturbing behaviour subscale, there was less disturbed behaviour in the memantine group (0.48 points 95%CI 0.06 to 0.91, P = 0.03)
* The drop-out rates and the numbers of people with any adverse event, dizziness or confusion were similar in treatment and placebo groups.

4-Effect of memantine (10 mg/day) in patients with severe vascular dementia, Alzheimer's disease and mixed dementia at 12 weeks:
9403/Winblad 1999 explored the effect of memantine (10 mg/day) in patients in long stay facilities with severe Alzheimer's disease, vascular dementia or mixed dementia at 12 weeks.
* Global Impression: There was a large benefit in favour of memantine (10 mg/day) compared with placebo at 12 weeks, for the numbers improved on the CGI (60/82 [73%] compared with 38/84 [45%]: OR 3.30, 95% CI 1.72 to 6.33, P = 0.0003)
* Function: 9403/Winblad 1999 provided data on the change from baseline at 12 weeks using the BGP care dependence subscore. Using the Wilcoxon Stratified test, the authors found a significant difference in favour of memantine. Using odds ratios there was no statistically significant difference between memantine 10 mg/day and placebo.

5- Effect of memantine (20 or 30 mg/day) in patients with mild to moderate vascular dementia, Alzheimer's disease or mixed dementia at 4 to 6 weeks:
* Global Impression: There was a benefit in favour of memantine (20 to 30 mg/day) compared with placebo at 4 to 6 weeks, as measured by various scales (SMD 0.76, 95% CI 1.04 to 0.48, P < 0.0001 ) (Ditzler 1991; Gortelmeyer 1992; Pantev 1993).
* Cognition: One study measured cognition using the SKT (Ditzler 1991). The change from baseline at 6 weeks for 30 mg/day gave statistically significant results in favour of memantine (3.04 SKT points, 95% CI 5.68 to 0.40, P = 0.02).
* Activities of daily living: One study provided data using a non-validated scale for measuring five simple instrumental tasks under the guidance of an investigator (Ditzler 1991). When pooled with the results from Pantev 1993 (BGP, need of help subscore), the OC analysis gave statistically significant results in favour of memantine for 30 mg/day in 6 weeks (SMD 1.34, 95% CI 1.73 to 0.94, P < 0.00001). However, there are discrepancies in the presentation of the results in different tables of the report of Pantev 1993.
* Mood and behaviour: One trial provided data using the NOSIE for memantine 30 mg/day at 6 weeks (Pantev 1993). The SCAG totals from Gortelmeyer 1992 and Ditzler 1991 were pooled with this on the grounds that many of the items from the SCAG are neuropsychiatric symptoms. Memantine had a significant beneficial effect (SMD 1.16, 95%CI 1.46 to 0.86, P < 0.0001).
* Adverse effects: Memantine seemed to be well tolerated in these studies (Ditzler 1991; Gortelmeyer 1992; Pantev 1993). There were no significant differences between memantine and placebo in the total number of adverse events. In one study the incidence of restlessness by the end of the treatment at 6 weeks was statistically significantly lower in the placebo group than in the group taking memantine 30 mg/day (Ditzler 1991) (15/30 compared with 2/29: OR 13.50, 95% CI 2.71 to 67.19, P = 0.001).

6- Pooled analyses: Effect of memantine in mild to severe Alzheimer's disease and vascular dementia at 24 to 28 weeks:
The results of eight, six-month trials were pooled (9605/Reisberg 2003; MD-02/Tariot 2004;MD-01; 9202/Wilcock 2002; 9408/Orgogozo 2002; MD-10/Peskind 2004; MD-12; 99679(Lundbeck)). There was a significant benefit of memantine at six months on all four main outcome measures.

* Global Impression: There was a significant benefit of memantine apparent on 7-point clinical global measures of change (WMD 0.15, 95% CI 0.07 to 0.23, P = 0.0001).
* Cognition: There was a significant benefit of memantine on cognitive function (SMD 0.24, 95% CI 0.17 to 0.30, P < 0.00001).
*Function: There was a significant benefit of memantine on patients' ability to perform activities of daily living (SMD 0.08, 95% CI 0.01 to 0.15, P = 0.03).
* Mood and behaviour: There was a significant benefit of memantine on patient's mood and behaviour as assessed by the NPI or the 'disturbed behaviour' subscale of the NOSGER (SMD 0.11, 95% CI 0.04 to 0.19, P = 0.003).
* Drop-outs: There was no significant difference in the dropout rate (315/1703, 18.5% versus 309/1509, 20.5% OR 0.91, 95% CI 0.76 to 1.09, P = 0.32)
* Adverse effects: There was no significant difference in the number of patients experiencing at least one adverse event (1239/1702, 72.8% versus 1103/1509, 73.1% OR 1.09, 95% CI 0.93 to 1.27, P = 0.31). Overall, a significantly lower number of patients taking memantine experienced agitation as an adverse event (134/1739, 7.7% versus 175/1873, 9.3% OR 0.78, 95% CI 0.61 to 0.99, P = 0.04). Data from other individual adverse events are not available for all trials.

Within the eight studies, one reported significantly more somnolence (OR 7.49, 95% CI 1.68 to 33.38) (MD-10/Peskind 2004), one reported more constipation (OR 2.57, 95% CI 1.29 to 5.12, P = 0.007) (9202/Wilcock 2002), and one reported more hypertension (OR 3.59, 95%CI 1.16 to 11.12, P = 0.03) (MD-01) in patients on memantine.

7- Pooled analyses - (10 to 30 mg/day) in patients with mild to moderate vascular dementia, Alzheimer's disease or mixed dementia at 4 to 12 weeks:
The only type of outcome measure that was available in all four shorter trials (9403/Winblad 1999; Ditzler 1991; Gortelmeyer 1992; Pantev 1993) was for clinical global.
* Global Impression: There was a significant benefit of memantine (SMD 0.62, 95% CI 0.41 to 0.82, P < 0.00001).


Discussion

The pooled results of three trials in patients with moderate to severe AD (9605/Reisberg 2003; MD-02/Tariot 2004; MD-01) suggest that taking 20 mg memantine daily has a positive effect on cognition, mood, behaviour and the ability to perform activities of daily living. These results are supported by a positive effect in the clinical impression of change which suggests that the effect is clinically detectable. The effect sizes were of 0.3 CIBIC+ points, 3.0 SIB points, 2.8 NPI points, and 1.3 ADCS-ADLsev points. There was substantial heterogeneity between studies in the effect on cognitive function. This was driven by the different effect sizes of the MD-01 and 9605/Reisberg 2003 studies. The moderate heterogeneity found in the effect on behaviour and mood was due to the different results of MD-01 compared to the other two studies. The heterogeneity did not seem to have arisen because the results of MD-02/Tariot 2004 were very different from the other two. This is important because it suggests that co-treatment with donepezil does not influence the effect of memantine. It is not clear why the results of MD-01 were different to those of the other two studies. Baseline cognitive function was broadly comparable in the three studies.

The results in patients with mild to moderate vascular dementia suggest a beneficial effect of 20 mg/day of memantine on cognitive function measured at 28 weeks (1.9 ADAS-Cog points). However, these results are not supported by an effect on the clinical impression of change. This suggests that, in patients with mild to moderate vascular dementia, the effect on cognitive function is not translated into clinically detectable changes. Although there was significant heterogeneity in the results of 9408/Orgogozo 2002 and 9202/Wilcock 2002, which remains to be adequately explained, the 9408/Orgogozo 2002 study was not greatly different in result from the studies in AD.

The results of trials in patients with mild to moderate AD (MD-10/Peskind 2004; MD-12; 99679(Lundbeck)) suggest a beneficial effect of 20 mg/day of memantine on cognitive function measured at 24 weeks (1.0 ADAS-Cog point), supported by a small positive effect in the clinical impression of change (0.1 CIBIC+ point). Observed case data from one study (MD-10/Peskind 2004) did not support the benefit on cognitive function seen on LOCF data. Observed case data on cognitive function have not been published for any study of mild to moderate AD.

The shorter, smaller studies suggest that there is a possible early benefit of the treatment. The standardised effect sizes on clinical global (SMD 0.6) and mood and behaviour (SMD 1.2) are larger than for the longer studies. However, there are many reasons to be wary of drawing such an inference. First, insufficient information is available from phase III studies to be able to examine any early beneficial effect. The graphs in 9605/Reisberg 2003 and MD-02/Tariot 2004 do not suggest any waning of effect after 12 weeks, which would argue against using the drug as a short term treatment. Second, trials conducted early in the development of a drug are typically more positive than later ones and shorter trials are typically more positive than longer ones. Thirdly, data from three short studies (MRZ-9104; MRZ-9105; MRZ-9206) were not available.

The adverse effects profile and tolerability are good and drop-out rates were generally low and similar in treatment and placebo groups.

Agitation
A consistent effect throughout the studies of moderate-to-severe dementia was a small reduction in the incidence of agitation. This is a common problem in late dementia for which improved treatments and prevention strategies are necessary. The finding is therefore potentially important. However, the effect size is small: 17 patients would need to be treated for six months to prevent one occurrence of agitation. It was not a pre-specified hypothesis and therefore needs to be treated cautiously.

Importantly, there was no available evidence addressing the question as to whether prevalent agitation can be usefully treated with memantine. MD-23 is listed as an ongoing trial on the Forest Clinical Trials Registry as a study of agitated, non-institutionalised patients with moderate to severe AD which started recruitment in October 2004. A recent Appraisal Consultation Document by NICE makes the following comment on the basis of data submitted in confidence by Lundbeck about "analyses ...undertaken for patients who were defined as 'behaviourally disturbed': For the analyses containing all three RCTs, less deterioration in cognitive function for patients receiving memantine as measured by the SIB (mean change from baseline for memantine [± donepezil] + versus placebo [± donepezil] -1.59 and -6.69, respectively, p < 0.001), ADCS-ADL (mean changes from baseline: -2.87 and -4.76, p = 0.001), NPI-cluster (mean changes from baseline: -0.65 and 0.74, p < 0.001) and CIBIC-plus (mean changes from baseline: 4.54 and 4.88, p < 0.001) was observed." The difference in change in clinical global in this subgroup (0.34 CIBIC+ points) is similar to that in the overall group. It is not possible to comment on how far the CIBIC+ change in the overall group is driven by the more disturbed subgroup. The difference in change in behaviour (1.39 NPI-cluster points) cannot be compared since data on this 'NPI-cluster' score is not available in the published studies.

Limitations
A limitation of this analysis is the small number of studies of each different severity and subtype of dementia which have been conducted. This reduces the statistical power to identify heterogeneity. Although the reporting of the studies is generally adequate, there are some details which are not addressed. For example, participants studied are not well characterized in terms of comorbidity or in the level of control of vascular risk factors, facts that could have influenced the results (Teresi 1997). The duration of the recent phase III studies is sufficient for licensing purposes. It would have been appropriate to include measures of executive function, as these are particularly impaired in vascular dementia and are not adequately dealt with by the ADAS-Cog (Roman 1999). No 6 month studies have been conducted in nursing home populations. MEM-MD22 is listed as an ongoing trial on the Forest Clinical Trials Registry as a study of nursing home patients with moderate to severe AD which started recrtuiment in October 2004.

A source of bias in the analysis of these studies is likely to lead to an under-estimation of the effect size. Since patients taking placebo tended to drop out earlier than those taking memantine, the practice of carrying forward the last observation will tend to reduce the difference between placebo and memantine.

The indications for treatment of severe dementia raise ethical issues. The benefits of slowing Alzheimer's disease progression in the later stages can be controversial (Post 1997). It is possible that the drug only extends the total time of deterioration without reducing the personal or social burden of the disease (Dresser 2000). The major phase III studies were too short and small to be expected to show any effect of memantine on life expectancy. In future studies it will also be important to assess patients at shorter intervals than 6 months, to include measures of patient quality of life, caregiver burden and costs, and to assess effect on mortality over a longer period than 6 months.

It has been suggested that memantine may exert a neuroprotective effect in early stages of dementia (Cacabelos 1999). There is no evidence for this from these clinical studies but an open label volumetric MRI study (MEM-MD15) started recruiting in June 2005.

Publication bias
The results of MD-01, which showed no effect of memantine, were posted on its clinical trials registry by Forest Laboratories on 2nd May 2005, more than 2 years after completion (23rd April 2003). They were not included in Lundbeck's June 2004 submission concerning moderate to severe AD to the UK National Clinical Institute for Excellence. Data from three short studies (MRZ-9104; MRZ-9105; MRZ-9206) were not available. It is possible that 10116 (Lundbeck) is a further unpublished, unannounced trial with important efficacy data.

Cost-effectiveness
An important question is whether the effect size is great enough to merit its purchase by health authorities, not least because the overall effect sizes are small. However, it is possible that small benefits at the moderate and severe stages have a greater impact on measures of quality of life and on cost-effectiveness than early in the course of dementia. Only one published analysis of costs incurred during an RCT has been conducted. This suggested a trade-off between costs which are borne by the caregiver (lower on memantine) and the direct medical costs (i.e. the drug) borne by the patient. These results should be treated with caution because they are only available in one trial and need to be confirmed.

The January 2006 NICE Appraisal Consultation Document gives an estimate of cost per QALY gained for patients with moderate to severe AD as approximately £35,000. However, it would appear from this Consultation Document that insufficiently detailed material was submitted by Lundbeck to allow the committee to evaluate various assumptions and definitions. The submission was, in part, based on presentation of data from selected studies which greatly increased the liklihood of bias. On the other hand, the data presented above, which were available to NICE, do not support the committee's conclusion that "the evidence for the clinical effectiveness of memantine was currently insufficient". The weight given in the NICE evaluation to an analysis based on MMSE changes may be inappropriate at lower MMSE levels.


Authors' conclusions

Implications for practice
Memantine 20 mg/day caused a clinically noticeable reduction in deterioration over 28 weeks compared with placebo in patients with moderate to severe Alzheimer's disease. This was supported by less functional and cognitive deterioration. Patients taking memantine were less likely to become agitated. Patients with mild to moderate dementia receiving memantine 20 mg/day had less cognitive deterioration at 28 weeks but the effects were not clinically discernible in those with vascular dementia. They were discernible in those with Alzheimer's disease. The drug is well tolerated in general and the incidence of adverse effects is low. The published data would only therefore justify the prescription of memantine for patients with moderate to severe AD, typically defined as those with an MMSE less than 15.


Implications for research
Studies assessing the neuroprotection hypothesis need to be longer than 6 months. Whether prescriptions of memantine for moderate-to-severe AD are publicly funded should depend on the results from longer studies of the effect of memantine on quality and duration of life, and how it performs in trials in more typical populations than are usual in licensing studies. Since the drug seems to be effective for moderate-to-severe AD, and is the only licensed drug available for this indication, robust data on the cost and effectiveness of other options for moderate-to-severe dementia are needed to inform the funding decision. Pragmatic studies should include comparisons of memantine with care-as-usual, since this is what would be available to most people if memantine is not publicly funded.


Acknowledgements

We thank Corinne Cavender for her contributions as Consumer Editor.
We thank Merz for providing an unpublished breakdown of global ratings of change and NOSGER scores from 9202/Wilcock 2002.

Data and analyses

Download statistical data


Comparison 1. Memantine vs placebo for moderate-to-severe Alzheimer's disease. 6 month studies. ITT-LOCF data.
Outcome or subgroup title
No. of studies
No. of participants
Statistical method
Effect size
1 Clinical Global: CIBIC+ (24-28 weeks)
3
964
Mean Difference (IV, Fixed, 95% CI)
0.28 [0.15, 0.41]
2 Cognitive function: SIB (change from baseline at 24-28 weeks)
3
976
Mean Difference (IV, Fixed, 95% CI)
2.97 [1.68, 4.26]
3 Activities of daily living: ADCS-ADLsev19 (change from baseline at 24-28 weeks)
3
978
Mean Difference (IV, Fixed, 95% CI)
1.27 [0.44, 2.09]
4 Behaviour and mood: NPI total (change from baseline at 24-28 weeks)
3
936
Mean Difference (IV, Fixed, 95% CI)
2.76 [0.88, 4.63]
5 Number of dropouts
3
1006
Odds Ratio (M-H, Fixed, 95% CI)
0.66 [0.49, 0.88]
6 Number suffering at least one adverse event
3
1005
Odds Ratio (M-H, Fixed, 95% CI)
1.13 [0.84, 1.52]
7 Number suffering agitation as an adverse event
3
1005
Odds Ratio (M-H, Fixed, 95% CI)
0.60 [0.42, 0.86]





Comparison 2. Memantine vs placebo for mild-to-moderate Alzheimer's disease. Published, 6 month studies. ITT-LOCF data
Outcome or subgroup title
No. of studies
No. of participants
Statistical method
Effect size
1 Clinical global: CIBIC+ (at 24 weeks)
3
1281
Mean Difference (IV, Fixed, 95% CI)
0.13 [0.01, 0.25]
2 Cognitive function: ADAS-Cog (change from baseline at 24 weeks)
3
1279
Mean Difference (IV, Fixed, 95% CI)
0.99 [0.21, 1.78]
3 Activities of daily living: ADCS-ADL23 (change from baseline at 24 weeks)
3
1271
Mean Difference (IV, Fixed, 95% CI)
0.20 [-0.87, 1.27]
4 Mood and behaviour: NPI total (change from baseline at 24 weeks)
3
1252
Mean Difference (IV, Fixed, 95% CI)
-0.25 [-1.48, 0.98]
5 Number of dropouts
3
1306
Odds Ratio (M-H, Fixed, 95% CI)
1.16 [0.83, 1.60]
6 Number suffering agitation as an adverse event
3
1306
Odds Ratio (M-H, Fixed, 95% CI)
0.91 [0.57, 1.46]
7 Number suffering fall as an adverse event
2
836
Odds Ratio (M-H, Fixed, 95% CI)
0.83 [0.41, 1.67]
8 Number suffering somnolence as an adverse event
1
403
Odds Ratio (M-H, Fixed, 95% CI)
7.49 [1.68, 33.38]
9 Number suffering confusion as an adverse event
2
836
Odds Ratio (M-H, Fixed, 95% CI)
1.40 [0.72, 2.70]
10 Number suffering at least one adverse event
3
1306
Odds Ratio (M-H, Fixed, 95% CI)
1.04 [0.81, 1.33]
11 Number suffering depression as an adverse event
2
836
Odds Ratio (M-H, Fixed, 95% CI)
0.71 [0.38, 1.32]





Comparison 3. Memantine vs placebo for mild-to-moderate vascular dementia. 6 month studies. LOCF or OC data
Outcome or subgroup title
No. of studies
No. of participants
Statistical method
Effect size
1 Clinical Global: CGI (at 28 weeks) ITT-LOCF or OC
2
775
Mean Difference (IV, Fixed, 95% CI)
0.03 [-0.13, 0.19]
2 Cognitive function ADAS-Cog (change from baseline at 28 weeks) ITT-LOCF
2
815
Mean Difference (IV, Fixed, 95% CI)
1.85 [0.88, 2.83]
3 Activities of daily living : NOSGER self care subscale (change from baseline at 28 weeks) OC
2
542
Mean Difference (IV, Fixed, 95% CI)
0.12 [-0.43, 0.67]
4 Behaviour: NOSGER disturbing behaviour subscale (change from baseline at 28 weeks) OC
2
541
Mean Difference (IV, Fixed, 95% CI)
0.48 [0.06, 0.91]





Comparison 4. Memantine vs placebo for severe Alzheimer disease, vascular and mixed dementia (12 weeks)
Outcome or subgroup title
No. of studies
No. of participants
Statistical method
Effect size
1 Activities of daily living (change from baseline at 12 weeks)
1
166
Mean Difference (IV, Fixed, 95% CI)
-2.01 [-5.65, 1.65]
1.1 Dose 10 mg/day, BGP subscore care dependence (change from baseline at 12 weeks)
1
166
Mean Difference (IV, Fixed, 95% CI)
-2.01 [-5.65, 1.65]
2 CGIC (numbers improved at 12 weeks)
1
166
Odds Ratio (M-H, Fixed, 95% CI)
3.30 [1.72, 6.33]
2.1 Dose 10mg/day
1
166
Odds Ratio (M-H, Fixed, 95% CI)
3.30 [1.72, 6.33]
3 Number of drop-outs
1
166
Odds Ratio (M-H, Fixed, 95% CI)
1.03 [0.25, 4.25]
4 Number suffering at least one adverse event
1
166
Odds Ratio (M-H, Fixed, 95% CI)
1.03 [0.49, 2.16]





Comparison 5. Memantine vs placebo for dementia (cause not specified) (4-6 weeks)
Outcome or subgroup title
No. of studies
No. of participants
Statistical method
Effect size
1 Global
3
213
Std. Mean Difference (IV, Fixed, 95% CI)
-0.76 [-1.04, -0.48]
1.1 Dose 30mg/day, Physicians global impression (6 weeks)
1
66
Std. Mean Difference (IV, Fixed, 95% CI)
-0.82 [-1.33, -0.32]
1.2 Dose 30mg/day, SCAG Clinical global impression of disturbances (4 weeks)
1
59
Std. Mean Difference (IV, Fixed, 95% CI)
-1.04 [-1.58, -0.49]
1.3 Dose 20mg/day, CGI (6 weeks)
1
88
Std. Mean Difference (IV, Fixed, 95% CI)
-0.55 [-0.98, -0.13]
2 Cognition
1
59
Mean Difference (IV, Fixed, 95% CI)
-3.05 [-5.68, -0.40]
2.1 Dose 30mg/day, SKT (change from baseline at 6 weeks)
1
59
Mean Difference (IV, Fixed, 95% CI)
-3.05 [-5.68, -0.40]
3 Activities of daily living
2
126
Std. Mean Difference (IV, Fixed, 95% CI)
-1.34 [-1.73, -0.94]
3.1 Dose 30mg/day, BGP care dependence subsscale (change from baseline at 6 weeks)
1
60
Std. Mean Difference (IV, Fixed, 95% CI)
-2.05 [-2.68, -1.42]
3.2 Dose 30mg/day, ADL-test total time (6 weeks)
1
66
Std. Mean Difference (IV, Fixed, 95% CI)
-0.88 [-1.39, -0.37]
4 Mood and behaviour
3
208
Std. Mean Difference (IV, Fixed, 95% CI)
-1.16 [-1.46, -0.86]
4.1 Dose 30mg/day, NOSIE (change from baseline at 4 weeks)
1
60
Std. Mean Difference (IV, Fixed, 95% CI)
-2.02 [-2.65, -1.39]
4.2 Dose 30mg/day, SCAG total (change from baseline at 6 weeks)
2
148
Std. Mean Difference (IV, Fixed, 95% CI)
-0.91 [-1.25, -0.57]
5 Number of dropouts
2
154
Odds Ratio (M-H, Fixed, 95% CI)
0.86 [0.27, 2.67]
6 Number suffering at least one adverse event
1
82
Odds Ratio (M-H, Fixed, 95% CI)
2.00 [0.83, 4.81]
7 Number suffering agitation as an adverse event
1
59
Odds Ratio (M-H, Fixed, 95% CI)
2.08 [0.48, 15.19]
8 Number suffering restlessness as an adverse event
1
59
Odds Ratio (M-H, Fixed, 95% CI)
13.06 [2.71, 67.18]





Comparison 6. Memantine vs placebo for mild to severe dementia.