Appraisal Consultation Document: Alzheimer's disease - donepezil, rivastigmine, galantamine and memantine (review)
NATIONAL INSTITUTE FOR CLINICAL EXCELLENCE
Appraisal Consultation Document
Donepezil, rivastigmine, galantamine and memantine for the treatment of Alzheimer’s Disease
|
The Department of Health and the Welsh Assembly Government have asked the National Institute for Clinical Excellence (NICE or the Institute) to conduct an appraisal of donepezil, rivastigmine, galantamine and memantine for the treatment of Alzheimer’s disease and provide guidance on its use to the NHS in England and Wales. The Appraisal Committee has had its first meeting to consider both the evidence submitted and the views put forward by the representatives nominated for this appraisal by professional organisations and patient/carer and service user organisations. The Committee has developed preliminary recommendations on the use of donepezil, rivastigmine, galantamine and memantine. This document has been prepared for consultation with the formal consultees. It summarises the evidence and views that have been considered and sets out the preliminary recommendations developed by the Committee. The Institute is now inviting comments from the formal consultees in the appraisal process (the consultees for this appraisal are listed on the NICE website, http://www.nice.org.uk/). Note that this document does not constitute the Institute's formal guidance on this technology. The recommendations made in Section 1 are preliminary and may change after consultation. The process the Institute will follow after the consultation period is summarised below. For further details, see the Guide to the Technology Appraisal Process (this document is available on the Institute's website, http://www.nice.org.uk/).
The key dates for this appraisal are: Details of membership of the Appraisal Committee are given in Appendix A and a list of the sources of evidence used in the preparation of this document is given in Appendix B. |
|
Note that this document does not constitute the Institute's formal guidance on this technology. The recommendations made in Section 1 are preliminary and may change after consultation. |
| 1 | Appraisal Committee's preliminary recommendations |
| 1.1 | Donepezil, rivastigmine and galantamine are not recommended for use in the treatment of mild to moderate Alzheimer’s disease (AD). |
| 1.2 | Memantine is not recommended for the treatment of moderately severe to severe AD, except as part of ongoing or new clinical studies that are designed to generate robust and relevant data on long-term outcomes, disease progression, quality of life and costs. |
| 1.3 | People currently receiving donepezil, rivastigmine, galantamine and memantine, whether as routine therapy or as part of a clinical trial, may be continued on therapy (including at the conclusion of a clinical trial) until it is considered appropriate to stop. |
| 2 | Clinical need and practice |
| 2.1 | Dementia is a chronic progressive mental disorder that adversely affects higher cortical functions including memory, thinking and orientation. AD is the most common form of dementia. AD is a primary degenerative cerebral disease with characteristic neuropathological and neurochemical features. |
| 2.2 | AD is usually insidious in onset and develops slowly but steadily over a period of several years. AD predominantly affects the elderly. Progression is characterised by a deterioration in cognition (thinking, conceiving, reasoning), functional ability (activities of daily living), and behaviour and mood. Changes in one or more of these domains and their effects on the person and their carer’s wellbeing provide the basis for diagnosis, by assessing the severity and progression of the condition. AD progresses from diagnosis to death in about 5-7 years. |
| 2.3 | People with AD lose the ability to carry out routine daily activities like dressing, undressing, toileting, travelling and handling money and, as a result, many of them require a high level of care. Often, this is provided by an elderly relative, whose own health and quality of life can be seriously affected by the burden of care provision. Behavioural changes in the person such as aggression are particularly disturbing for carers. |
| 2.4 | Non-cognitive symptoms in dementia include agitation, behavioural disturbances (for example, wandering or aggression), depression, delusions and hallucinations. These features are common, often difficult to treat, and are a much stronger predictor of both carer stress and entry into institutional care than cognitive impairment; they are therefore important targets for therapeutic intervention. |
| 2.5 |
Several different methods are used to assess the severity of Alzheimer’s disease. These include: the Clinician's Interview-based Impression of Change (CIBIC) and CIBIC-plus for global outcomes, the Progressive Deterioration Scale (PDS) for functional/quality of life scales and the 70 point, Alzheimer's Disease Assessment Scale - cognitive subscale (ADAS-cog) or the 30 point, Mini Mental State Examination (MMSE) for cognitive outcomes. MMSE score, for example, denotes the severity of cognitive impairment as follows:
|
| 2.6 | Population data (2002) for England and Wales show an estimated prevalence of 290,000 people with AD. On the basis of these figures a Primary Care Trust (PCT) with a population of 200,000 might expect to have approximately 1100 cases of AD. The incidence rate for AD in those over 65 years old has been estimated at 4.9 per 1000 person-years in the UK. The incidence rate appears to have been stable over the past two decades and is found to be related to age (rising with increasing age) and gender (women have higher risk than men). In people with AD, 50-64% are estimated to have mild to moderate disease while approximately 50% of patients have moderately severe to severe AD. |
| 2.7 | An estimated 22% of people with dementia live alone, 36% with carers and 29% in nursing homes. The total costs of care for people with dementia are estimated by the Audit Commission to be £6 billion for dementia, with half of this amount attributed to Health and Social Services. |
| 2.8 | People with dementia usually present to their general practitioner with memory problems, and an estimated 39% present to specialist clinics. The role of memory clinics has been further clarified by the National Service Framework for Older People. This states that referral to specialist mental health services should be considered for those with suspected dementia in a number of circumstances, not only for consideration of treatment but also, for example, if the diagnosis is uncertain, if certain behavioural and psychological symptoms are present, or if there are safety concerns with anti-dementia drugs, in accordance with local protocols. |
| 3 | The technologies |
| Acetylcholinesterase inhibitors: donepezil, rivastigmine, galantamine | |
| 3.1 | Acetylcholinesterase (AChE) inhibitors raise the concentration of acetylcholine at sites of neurotransmission. Since the original NICE guidance of 2001 the number of prescribed defined daily doses for the AChE inhibitors, especially donepezil, has increased markedly. Substantial regional variation in spending is seen across Strategic Health Authorities in England and Wales. |
| 3.2 | Donepezil (Aricept®, Eisai) is a specific and reversible inhibitor of AChE licensed in the UK at a dose of 5 mg/day and 10 mg/day for the symptomatic treatment of mild to moderately severe Alzheimer’s dementia. Prices are £63.54 for 28 tablets of 5 mg and £89.06 for 28 tablets of 10 mg (excluding VAT; British National Formulary [BNF] 48th edition and January 2005 communication from Eisai on price adjustment). This equates to £828.30 and £1160.96 per year of treatment respectively. Costs may vary in different settings because of negotiated procurement discounts. In 2003, 77% of prescriptions for AChE inhibitors were for donepezil. |
| 3.3 | Rivastigmine (Exelon®, Novartis) is an acetyl- and butyrylcholinesterase licensed in the UK for the treatment of mild to moderately severe AD. The usual dose range is 3–6 mg twice daily, with a maximum of 6 mg twice daily (BNF 48th edition). Prices are £68.04 for 56 capsules of 1.5 mg, 3 mg, 4.5 mg and 6 mg (excluding VAT; BNF 48th edition). This equates to approximately £886.95 per year of treatment. Costs may vary in different settings because of negotiated procurement discounts. |
| 3.4 | Galantamine (Reminyl®, Shire) is a selective, competitive and reversible inhibitor of AChE licensed in the UK for symptomatic treatment of mild to moderately severe dementia of the AD type. Galantamine also appears to enhance the activity of both pre- and post-synaptic nicotinic acetylcholine receptors, but the clinical relevance of this is unclear. The maintenance dose is between 8 and12 mg twice daily (BNF 48th edition). Prices are £68.32 for 56 tablets of 8 mg and £84.00 for 56 tablets of 12 mg (excluding VAT; BNF 48th edition). This equates to approximately £890.60 and £1095 per year of treatment, respectively. Costs may vary in different settings because of negotiated procurement discounts. |
| 3.5 | Side effects of donepezil, rivastigmine and galantamine are typically related to the gastrointestinal tract. These side effects are usually produced acutely, and tend to be dose related. For full details of side effects and contraindications, see the Summary of Product Characteristics. |
| Memantine | |
| 3.6 | Memantine (Ebixa®, Lundbeck) is a voltage-dependent, moderate-affinity uncompetitive NMDA-receptor antagonist that blocks the effects of pathologically elevated tonic levels of glutamate that may lead to neuronal dysfunction. Memantine is licensed for the treatment of people with moderately severe to severe AD. The recommended maintenance dose is 10 mg twice daily. Prices are £69.01 for 56 tablets of 10 mg (excluding VAT; BNF 48th edition and January 2005 communication of Lundbeck on price adjustment). This equates to approximately £899.54 per year of treatment. Costs may vary in different settings because of negotiated procurement discounts. |
| 3.7 | The mild to moderately severe side effects associated with the use of memantine include hallucinations, dizziness, confusion, headache and tiredness. For full details of side effects and contraindications, see the Summary of Product Characteristics. |
| 4 | Evidence and interpretation |
| The Appraisal Committee considered evidence from a number of sources (see Appendix B). | |
| 4.1 | Clinical effectiveness |
| Mild to moderate AD | |
| 4.1.1 | The quality of the reporting and methodology of the included published randomised controlled trials (RCTs) of the cholinesterase inhibitors was generally mixed. The Assessment Group suspected selection bias, measurement bias and attrition bias in a number of the reviewed studies. |
| 4.1.2 | Donepezil |
| 4.1.2.1 | Thirteen RCTs (aggregate number of people randomised 4200), one unpublished RCT and two systematic reviews met the inclusion criteria set by the Assessment Group for the systematic review of clinical effectiveness of donepezil (the original guidance included five RCTs, four studies from manufacturers and three systematic reviews). Three of the new trials followed participants for a period longer than 6 months. |
| 4.1.2.2 | Six RCTs reviewed by the Assessment Group showed, using the ADAS-cog scale, that donepezil appears to confer a statistically significant benefit to participants’ cognition when compared with placebo. Higher doses of donepezil were associated with increasing benefit. Three RCTs with a duration of 12 to 24 weeks contained data in a form that could be combined by the Assessment Group in a meta-analysis. A weighted mean difference of -2.51 (95% confidence interval [CI] -3.26 to -1.76) in terms of a change from baseline was found for the 5 mg daily dose and a weighted mean difference of -3.01 (95% CI -3.91 to -2.10) was found for the 10 mg daily dose when compared with placebo. An analysis based on the trial of 24 weeks duration produced a mean difference in ADAS-cog change from baseline at 24 weeks of -2.88 (95% CI -4.27 to -1.49). |
| 4.1.2.3 | Eight RCTs showed trends towards improved MMSE scores in the donepezil-treated groups when compared with placebo. The AD2000 study differs from the trials sponsored by the drug industry in that the participants were typical UK patients, referred from memory clinics and managed by local doctors who had minimum study-related tasks. Participants were treated in a placebo-controlled randomised trial with indefinitely long follow-up at doses of 5 mg or 10 mg donepezil a day. No statistically significant effect of donepezil was seen on the primary outcomes of the study (entry into institutional care and progression of disability). Cognition, measured by MMSE, was included as a secondary outcome. Over the 2-year study period the MMSE scores of the donepezil group were an average of 0.8 points higher than those of the placebo group (95% CI 0.5 to 1.2, p < 0.0001). AD2000 was not included in the Assessment Group’s meta-analysis of effect on MMSE for donepezil. |
| 4.1.2.4 | Seven RCTs (aggregate number of people randomised 2500) assessed the effect of donepezil on global outcomes compared with placebo using the CGIC or CIBIC-plus. Participants using donepezil showed a significantly greater change in CGIC/CIBIC-plus scores from baseline when compared with placebo. |
| 4.1.2.5 | Studies reporting on the short-term effects of donepezil using a variety of measures of activities of daily living generally found better, or less deterioration in, functional ability than placebo, although these findings were not statistically significant in all trials. No significant difference was apparent between donepezil and placebo in rates of institutionalisation in the AD2000 trial (9% donepezil versus 14% placebo at 1 year, p = 0.15; 42% versus 44%, respectively, at 3 years, p = 0.4). Furthermore, similar proportions of participants had progression of disability (13% donepezil versus 19% placebo at 1 year, p = 0.3; 55% versus 53%, respectively, at 3 years, p = 0.9). |
| 4.1.2.6 | Quality of life estimates associated with the use of donepezil showed varied results, and only three studies reported on this outcome. Over the three studies, the impact of donepezil derived from this set of health measurements is unclear. Moreover, all trials used non-validated scales for dementia. The respective results showed improvements, no change and worsening of quality of life, with the impact of the dose used for donepezil being unclear. |
| 4.1.2.7 | Behavioural symptoms were measured using the neuropsychiatric inventory (NPI) in four RCTs of donepezil. The results varied but generally suggested that donepezil may have some effect in improving or limiting further deterioration on the NPI scale compared with placebo, at least in the short term. |
| 4.1.2.8 | Adverse events were more frequently recorded in participants receiving donepezil than in those on placebo, and higher doses of donepezil increased the frequency of adverse events. Withdrawals due to adverse events were associated with similar losses in the groups on low-dose donepezil and those on placebo. However, higher doses of donepezil tended to lead to more withdrawals. |
| 4.1.2.9 | The manufacturer’s submission included a 24-week RCT that evaluated the safety and efficacy of donepezil in people with moderately severe AD (baseline MMSE score 5–17) and compared them with placebo. People receiving donepezil scored statistically significantly better on global, cognitive, functional and behavioural outcomes. A number of open-label and observational studies were also included in the manufacturer’s submission. The effect size of donepezil on cognitive and global outcomes in these studies was similar to those recorded by other RCTs. The use of donepezil also appeared to show a benefit on outcomes such as ‘delayed time to nursing home placement’ and improvements in social behaviour (assessed by the care giver diary) when valued in the context of these observational studies. |
| 4.1.2.10 | The manufacturer’s submission and the Assessment Report included a study that aimed to establish the effect of continuation of treatment with donepezil for 153 people who had not shown a response (‘no apparent clinical benefit’) after 24 weeks of open-label donepezil treatment. Randomisation of those who showed no response and who were subsequently treated for 12 weeks (double-blind) with donepezil (10 mg/day) was associated with a significantly higher MMSE score and a significantly lower NPI score than those receiving placebo. |
| 4.1.2.11 | In summary, evidence from studies using cognitive and global outcome measurement scales suggests that donepezil appears to be beneficial in treating AD. The effect of donepezil on functional outcomes, quality of life and behavioural symptoms in AD is less conclusive. The AD2000 trial substantiates these general findings but suggests a lower effect size on cognition than was found in the Assessment Group meta-analysis. |
| 4.1.3 | Rivastigmine |
| 4.1.3.1 | Four published RCTs (aggregate number of people randomised 1940), two unpublished RCTs (aggregate number of people randomised 1380) and three systematic reviews met the inclusion criteria set by the Assessment Group for the systematic review of the clinical effectiveness of rivastigmine. All published comparisons were versus placebo with trials reporting doses of between 1 and 12 mg/day and durations of 26 weeks or less. The original NICE guidance was based on three systematic reviews, five RCTs and two unpublished studies from manufacturers. |
| 4.1.3.2 | Four RCTs reviewed by the Assessment Group showed that the higher dose of rivastigmine (6–12 mg/day, mean dose approximately 10 mg/day) appeared to confer a statistically significant benefit to participants when compared with placebo, as measured using the ADAS-cog scale. One RCT found no significant differences. A meta-analysis by the Assessment Group of two RCTs, both with a duration of 26 weeks, was associated with a weighted mean difference of –3.08 (95% CI –3.78, –2.38) for 6–12 mg of rivastigmine per day when compared with placebo. Statistically significant heterogeneity was found when pooling the two studies for meta-analysis, which led the Assessment Group to conclude that the statistically significant treatment effect seen for rivastigmine (6–12mg) in the fixed-effect model should be treated with caution. |
| 4.1.3.3 | Three RCTs showed significantly better MMSE scores in the groups treated with rivastigmine in the higher dose regimen compared with placebo. None of the studies using a low-dose regimen found statistically significant differences for rivastigmine versus placebo. |
| 4.1.3.4 | Four RCTs assessed the effect of rivastigmine compared with placebo on the CIBIC-plus scale. In the two published RCTs participants demonstrated mean changes from baseline that were statistically significantly better with rivastigmine in the high-dose regimen only. The percentage of improvers or responders on the CIBIC-plus was also calculated in the two published studies. Clinical improvement was defined as a score of one, two or three on the CIBIC-plus scale. Between the two trials, 16% to 20% of participants treated with placebo were judged to have responded versus 30% to 57% for users of rivastigmine. A statistically significant difference was found for the high-dose regimen only. |
| 4.1.3.5 | Generally, participants treated with 6–12 mg/day rivastigmine demonstrated statistically significant better functional outcomes than participants on placebo. One (700 people randomised) of the four studies (aggregate number of people randomised 2800) using the PDS showed that there was no statistically significant difference from placebo for either the low- or high-dose regimen when compared with placebo. |
| 4.1.3.6 | The Nurses Observation Scale for Geriatric Participants (NOSGER) was used in two rivastigmine RCTs. No statistically significant benefit was demonstrated on measures of mood and behaviour in the groups treated with rivastigmine compared with the placebo groups. |
| 4.1.3.7 | Levels of nausea and vomiting were particularly high among participants treated with the higher dose of rivastigmine. Withdrawals due to adverse events were reported in all studies and these varied considerably. The manufacturer noted that better adherence to the up-titration schedule should improve gastrointestinal tolerability. |
| 4.1.3.8 | A number of open-label and observational studies were included in the manufacturer’s submission. The duration of these trials was between 26 weeks and 5 years. The effect size of rivastigmine on cognitive and behavioural outcomes was similar to those seen in the RCTs. Other open-label and observational studies, and experience with rivastigmine in a ‘real-world’ setting, appeared to show some benefit in outcomes such as ‘delayed time to nursing home placement’ and care giver burden. |
| 4.1.3.9 | The manufacturer’s submission included a prospective, open-label study that evaluated the efficacy and safety/tolerability of rivastigmine in people who had failed to benefit from treatment with donepezil (because of a lack of efficacy [80%] or tolerability [11%], or both [9%]). After 26 weeks 56.2% of patients had responded to rivastigmine (defined as improvement/stabilisation of symptoms using the CGIC). |
| 4.1.3.10 | In summary, a range of fixed and flexible dosing regimes of rivastigmine were investigated across studies, which makes interpretation of the evidence more difficult. Evidence from studies using cognitive and global outcome measurement scales suggests that rivastigmine may be beneficial in AD at higher doses (6–12 mg). Evidence for an effect on functional outcomes was less conclusive and no benefit of rivastigmine on measures of behaviour and mood was reported. Side effects of the use of rivastigmine were considerable at higher doses and led to withdrawals in studies. The results of the meta-analysis on cognition should be treated with caution because of statistically significant heterogeneity in the fixed-effect model. |
| 4.1.4 | Galantamine |
| 4.1.4.1 | Six published RCTs, one unpublished RCT (aggregate number of people randomised 4300) and one systematic review met the inclusion criteria set by the Assessment Group for the systematic review of clinical effectiveness of galantamine. All comparisons were versus placebo, with trials reporting dosing regimens between 8 mg/day and 36 mg/day and durations of 3 to 6 months. One systematic review, three RCTs and three studies from manufacturers were identified in the literature at the time of the original appraisal. |
| 4.1.4.2 | All six published RCTs and the unpublished RCT assessed the clinical effectiveness of galantamine compared with placebo using the ADAS-cog scale. In all studies galantamine appeared to confer a statistically significant benefit to participants when compared with placebo. The benefit varied depending on the dose of galantamine. Three RCTs of treatment with galantamine at a dose of 24 mg were combined by the Assessment Group in a meta-analysis. The fixed-effect model showed a weighted mean difference of –3.37 (95% CI –4.11 to –2.63) favouring galantamine over placebo. |
| 4.1.4.3 | Six RCTs assessed the effect of galantamine compared with placebo on the CIBIC-plus. They showed that in individual studies more participants on galantamine improved than on placebo (0% to 6.5% more), whereas more participants on placebo than on galantamine deteriorated (4% to 18% more). When the studies were pooled by the Assessment Group (aggregate number of people randomised 3300) no statistical significance was noted between treatment groups and placebo. |
| 4.1.4.4 | Participants receiving galantamine appeared to undergo statistically significantly less deterioration than those on placebo for doses between 16 mg/day and 32 mg/day on scales that measure activities of daily living. |
| 4.1.4.5 | In one trial higher doses of galantamine (16 mg/day or over) were associated with a statistically significant slowing in the deterioration of participants’ condition on NPI compared with placebo. In two trials the slowing of deterioration was not statistically significantly different between treatment and control groups. |
| 4.1.4.6 | Across RCTs, between 2% and 27% more participants on galantamine suffered from an adverse event compared with placebo. Withdrawals due to adverse events were associated with a loss of between 6% and 44% of galantamine participants, with differences following a dose-response relationship. |
| 4.1.4.7 | The manufacturer’s submission presented a pooled analysis of two 6-month RCTs in people with mild to moderate AD that suggested a statistically significant decrease in the overall mean amount of time care givers spent assisting people with activities of daily living. The decrease was 32 minutes/day for patients treated with galantamine at a dose of 24 mg/day when compared with placebo. A number of open-label studies included in the manufacturer’s submission suggested a slightly reduced long-term decline in the cognition of people treated with galantamine. |
| 4.1.4.8 | In 6-week follow-on studies of two RCTs (aggregate number of people randomised 570) included in the manufacturer’s submission, patients who were switched from galantamine to placebo experienced a greater decline on measures of cognition than those who remained on galantamine. This difference reached statistical significance only in the study where no randomisation was involved in deciding whose treatment should be stopped (number of participants 500). |
| 4.1.4.9 | Evidence from studies using cognitive and functional outcome measurement scales suggest that galantamine may be beneficial in AD. Improved benefits in cognition tended to be related to higher doses. Improvements in measurements of function were also demonstrated at higher doses. On global outcome measures, individual studies showed higher proportions of participants improving with galantamine, but this was not reflected in the meta-analysis. In some studies considerably more participants withdrew because of adverse events. |
| 4.1.5 | Head-to-head comparisons |
| 4.1.5.1 | Three RCTs met the inclusion criteria for the review by the Assessment Group. Two compared donepezil with rivastigmine and one compared donepezil with galantamine. The Assessment Group regarded the quality of the studies as generally poor. The manufacturer’s submission for galantamine included a study comparing galantamine with donepezil, but this study was excluded by the Assessment Group because the study did not describe its population as mild to moderate AD by any definition and the MMSE scores fell outside the range of 10–26. |
| 4.1.5.2 | When directly compared with donepezil, treatment with rivastigmine showed no statistically significant differences on measures of cognition or function. Rates of adverse events tended to be higher in participants in the rivastigmine groups but the manufacturer’s submission for rivastigmine argued that slower titration is recommended for clinical practice than was used in this trial. |
| 4.1.5.3 | In the comparison between galantamine and donepezil that was sponsored by the manufacturer of donepezil, participants on galantamine showed improvement on measures of cognition and function but the improvement in participants on donepezil was greater. However, in the comparison that was funded by the manufacturer of galantamine this effect seemed to be reversed and it appeared that galantamine exerted a more sustained effect than donepezil. |
| Moderately-severe to severe AD | |
| 4.1.6 | Memantine |
| 4.1.6.1 | Two RCTs (aggregate number of people randomised 650) met the inclusion criteria set by the Assessment Group for the systematic review of the clinical effectiveness of memantine. Both studies reported on participants with moderately severe to severe AD, as measured by the MMSE, and compared 20 mg/day of memantine with placebo; one study made the comparison over a period of 24 weeks and the other over 28 weeks. In the first study participants were included on the basis that they had already been receiving donepezil for more than 6 months before entering the trial, and had been at a stable dose (5–10 mg/day) for at least 3 months. These participants were maintained on stable donepezil for the duration of the study. The quality of reporting and methodology of the two trials was generally good. |
| 4.1.6.2 | People treated with memantine in both studies showed statistically significantly less deterioration of cognitive function measured by the Severe Impairment Battery (SIB), but deterioration as measured by the MMSE was not statistically significantly different between the treatment and control groups. Memantine was statistically significantly more effective than control in improving global function as measured by CIBIC-plus. Both trials used the Alzheimer’s Disease Cooperative Study – Activities of Daily Living 19 (ADCS-ADL19/sev) scale (which focuses on the assessment of later stages of dementia) to measure functional outcome, and reported statistically significantly less deterioration following the use of memantine compared with control. Participants receiving memantine who were also on a course of donepezil had a statistically significantly lower NPI score for behavioural symptoms compared with the control group of donepezil alone. In the separate trial of memantine alone (250 people randomised), however, no statistically significant difference compared with placebo was observed. The frequency of adverse effects was similar for both the memantine and control groups. Despite this, the rate of withdrawal due to adverse events was relatively high in the memantine group. |
| 4.1.6.3 | During the course of the review and after the Assessment Report had been produced, the manufacturer of memantine provided summary results of a further trial of memantine against placebo, the results of which are commercial-in-confidence. In other material provided by the manufacturer, care givers, as assessed by the Resource Utilization in Dementia score, were estimated to spend less time with participants receiving memantine (difference between treatment groups, 45.8 hours per month; 95% CI, 10.37 to 81.27; p = 0.01). This amounts to between 20 minutes and 2 hours 45 minutes per day, with a most probable difference of 1 hour 30 minutes a day. The results of an open-label extension study of one of the pivotal trials suggest that effects of memantine can be sustained over a period of 12 months. |
| 4.1.6.4 | Memantine showed beneficial effects in participants with moderately severe to severe AD in terms of functional and global measurements, where participants in the treatment arms showed less deterioration than those in the placebo arm. The effect of memantine on cognitive outcome measurements was also favourable, although this was not always statistically significant (notably on MMSE). On measures of behaviour and mood memantine was beneficial only in the groups already receiving donepezil. |
| 4.2 | Cost effectiveness |
| Mild to moderate AD | |
| 4.2.1 | Twenty-one published economic evaluations of the three AChE inhibitors and memantine were available to the Appraisal Committee. All four manufacturers also submitted their own economic evaluations. The Assessment Group re ran each of the manufacturer’s economic models using their preferred assumptions and also performed their own economic evaluation of the three AChE inhibitors, but not of memantine. |
| 4.2.2 | Donepezil |
| 4.2.2.1 | Eleven economic evaluations for donepezil were found. Three related to the UK. One of the eleven studies was of treatment for people with mild AD; the other ten were of mild to moderate AD. In five (of 11) studies donepezil was found to be cost saving. |
| 4.2.2.2 | Of the three UK-based studies, an early independent study, based on drug costs only, estimated a cost per quality-adjusted life-year gained (CQG) for the 5 mg dose per day of £21,000 (2-year model) to £86,000 (10-year model) for a gain of 0.08 QALYs per person, and of £35,000 to £139,000 when the QALY gain was only 0.05. |
| 4.2.2.3 | In a UK study associated with the manufacturer, the cost of gaining an additional year in a non-severe state was measured. The cost was estimated to have ranged from £4500 to £7000, depending on dose and starting point (mild or moderate AD). |
| 4.2.2.4 | The AD2000 study found that the drug was not cost effective, mainly because there were no apparent benefits of the drug in delaying progression of the disease or of time to institutionalisation. The authors doubted that the drug would prove to be cost effective; even if changes in carer costs were included, the drug would not prove to be cost effective. |
| 4.2.2.5 | The manufacturer’s model used a transition state modelling approach in which disease progression is modelled across different levels of AD severity to estimate the incremental cost effectiveness of donepezil compared with placebo. Transition probabilities were derived from trial data, with the drug efficacy rate persisting for the initial 12-month cycle of the model. For the remainder of the 5-year model the transition probabilities for the treated group were proportional to those of the placebo group. Cost estimates were taken from the literature in which they were calculated for different severity levels of AD by MMSE score. The submission reported that for the base case of people with a MMSE score of 13–26, treatment with donepezil 10 mg daily resulted in an estimated £1200 per year outside of the severe AD state. Inclusion of the people with a MMSE of 10–12 increased this to £4000 per year outside of the severe state. The manufacturer’s model allowed for estimates of CQG to be calculated but did not report results in terms of CQG either in the base case analysis or in the sensitivity analyses. |
| 4.2.2.6 | The Assessment Group noted that the use of cognitive function alone to model disease progression is likely to misrepresent this disease progression over time. Where the Assessment Group incorporated alternative cost estimates as well as an increased mortality risk and a half-cycle correction, the manufacturer’s model estimated an incremental cost effectiveness of £14,000 per year outside of the severe state. When the Assessment Group used an incremental utility of 0.3 to represent the transition between severe and non-severe AD, this incremental cost effectiveness translated to a CQG estimated to be £45,000. |
| 4.2.3 | Rivastigmine |
| 4.2.3.1 | Five economic evaluations for rivastigmine were found, one of them in abstract form only. Two related to the UK. All were of people with mild to moderate AD. Four, including all three industry-associated studies, were found to be cost saving. |
| 4.2.3.2 | Of the two UK-based studies, an independent study estimated a range of incremental cost-effectiveness ratios; the estimates varied depending on the time duration used by the models, which ranged from 1 year (more cost effective) to 5 years (less cost effective) and on the number of QALYs gained (0.05 or 0.08). These models resulted in CQG estimates ranging from £16,000 to £46,000. Separate estimates were provided when non-drug treatment costs were included, and these ranged from £15,000 to £89,000. |
| 4.2.3.3 | In a study supported by the manufacturer, estimated cost savings (but not including the cost of rivastigmine) after 2 years were £1300 for people with mild AD and £800 for those with moderate AD. |