Acetylcholinesterase inhibitors: donepezil, rivastigmine, tacrine or galantamine for non-alzheimer s dementia

Transcription

1 STEER 2002; Vol 2: No.2 Acetylcholinesterase inhibitors: donepezil, rivastigmine, tacrine or galantamine for non-alzheimer s dementia Bunmi Fajemisin Evidence search date: November

2 What is the STEER Service? The STEER service is a rapid ondemand reviewing service, which: informs decisions by providing evidence-based answers to focused questions produces the reviews within a short period of time, usually 8-10 weeks is provided for policy makers by the Wessex Institute for Health Research and Development together with Bazian Ltd, an independent company that specialises in evidence-based medicine What is a STEER? STEER stands for Succinct and Timely Evaluated Evidence Review. A STEER is: a short, pragmatic review of major sources of published literature to answer focused questions designed to further decisions by quickly surveying and reporting on a large number of sources of evidence descriptive in nature, rarely employing meta-analysis conducted by reviewers using validated search strategies, data extraction, and peer review How is a STEER produced and quality-controlled? A structured STEER question is developed through liaison with the commissioner and with experts in the field. A systematic search of the published literature is performed by an experienced information scientist, using validated search strategies (available on request for each review). An initial check of study abstracts is performed to exclude irrelevant studies. Identified papers are then obtained. An initial appraisal of each paper is then performed by two experienced appraisers, using standard, validated critical appraisal techniques. Irrelevant or poor quality studies are excluded at this stage. Selected papers are sent to a reviewer to produce a draft STEER report. Reviewers are supported throughout the reviewing process by an experienced in-house team, advising on methods and providing guidance as needed. The draft report is independently and anonymously peer-reviewed by other members of the network of reviewers. The manuscript is then checked and edited by experienced inhouse editors (from the Wessex Institute and Bazian Ltd) familiar with review methods. The final proof is re-checked by the reviewer before dissemination. STEER 2002; Vol 2: No.2 p.1

3 How can I request a STEER? The STEER service originated to support the decisions of health service commissioners and policy makers in the South East Region of the NHS. A limited number of STEER reports are available to NHS staff based in the South East Region. Because of this limitation on the number of reports, topics should have been prioritised through discussion with Health Authorities or the Regional Office before a request is made. Call for contributors Please let us know if you would like to join the STEER College of Reviewers. The service provides an ideal opportunity for practitioners in all branches of healthcare to increase their reviewing experience, and improve their skills. Some previous reviewing experience is preferred, although we will provide assistance, coaching and feedback throughout the process. If you are not based in the South East Region of the NHS then please contact the Wessex Institute for details on how to purchase STEER reports (individually or in batches). Before requesting a STEER report please: Check the STEER website The site features a system called SIGNPOST, which will help you to find any existing reviews on your topic. The site also provides an index and the full text of previous STEERs. You may find that your question has already been answered. Contact the Wessex Institute or Bazian Ltd. to submit a request. Contact details STEER administrator Bazian Ltd Suites 1 and Upper Street London N1 1QP tel: fax: administrator@bazian.com website: p.2 STEER 2002; Vol 2: No.2

4 Acetylcholinesterase inhibitors: donepezil, rivastigmine, tacrine or galantamine for non-alzheimer s dementia Author name and details Bunmi Fajemisin Clinical Effectiveness Pharmacist Oxford Radcliffe Hospitals NHS Trust Headley Way, Headington Oxford OX3 7EP Conflict of Interest None NB: Users of this report should check for later evidence that may alter the STEER conclusions. This STEER report should be cited as: Fajemisin B. Acetylcholinesterase inhibitors: donepezil, rivastigmine, tacrine or galantamine for non-alzheimer s dementia. In Foxcroft DR, Muthu V (Eds) STEER: Succinct and Timely Evaluated Evidence Reviews 2002; 2(2). Wessex Institute for Health Research & Development, University of Southampton. [WWW document] URL Series Editors: David Foxcroft, Vivek Muthu STEER Editor: Vivek Muthu Assistant STEER Editors: Anna Donald, Debbie Singh Information specialist: Sam Vincent, Liz Payne, Alison Price STEER Reports Copyright 2001 Wessex Institute for Health Research and Development. All rights reserved STEER 2002; Vol 2: No.2 p.3

5 Acetylcholinesterase inhibitors: donepezil, rivastigmine, tacrine or galantamine for non- Alzheimer s dementia Question 1. What are the effects of acetylcholinesterase inhibitors (donepezil, rivastigmine, tacrine, galantamine) on functional outcomes in people with non-alzheimer s dementia? Summary We found insufficient evidence about functional effects of acetylcholinesterase in people with non-alzheimer s dementias. We found limited evidence from one RCT that acetylcholinesterase inhibitors may improve some cognitive and behavioural symptom scores in people with suspected Lewy body dementia. However, the clinical importance of the outcome measures was unclear and the study may have lacked power to detect clinically important effects for some of those outcomes. p.4 STEER 2002; Vol 2: No.2

6 Background The acetylcholinesterase inhibitors donepezil, rivastigmine, galantamine and tacrine are effective for delaying neurocognitive decline in people with Alzheimer s disease of mild to moderate severity. The UK National Institute of Clinical Excellence issued guidance for the use of these three agents in people with mild to moderate Alzheimer s disease in January However, guidance does not extend to people with non- Alzheimer s dementias, such as vascular dementia or Lewy body dementia. Vascular dementia is caused by a single or multiple strokes while dementia with Lewy bodies is a neurodegenerative disorder, which may be suspected on clinical grounds but is confirmed by brain biopsy. Both vascular and Lewy body dementia share clinical features with Alzheimer s dementia. Lewy body dementia is characterised by marked fluctuations in cognitive performance, visual hallucinations and the motor features of Parkinson s disease. Treatment is symptomatic, but is commonly complicated by a trade off between psychiatric and motor symptoms. Antiparkinsonian drugs may worsen psychiatric symptoms such as hallucinations, while neuroleptic drugs may worsen movement symptoms. Acetylcholinesterase inhibitors may offer an opportunity to treat cognitive and hallucinatory symptoms without worsening parkinsonian symptoms. Vascular dementia is characterised by abrupt and stepwise cognitive decline, with memory impairment, emotional lability, focal neurological signs and Parkinsonian motor symptoms. Judgement, personality and insight may be relatively well preserved. Pharmacotherapy is largely aimed at preventing further infarcts by reducing atherosclerotic risk factors. Acetylcholinesterases, however, may have direct neurocognitive effects in people with vascular dementia. Search Methods Primary sources: Medline 1966 to date; Embase 1980 to date; Cochrane 2001 issue 4; Search date: November Evidence found We found two trials examining effects of acetylcholinesterase inhibitors in people with non- Alzheimer s dementia. One study examined the effects in people with suspected Lewy body dementia and the other in people with vascular dementia. 1,2 We found one further trial comparing galantamine versus placebo in people with vascular dementia. However, results are yet to be published and the study has therefore been excluded from further discussion in the present review. 3 The first trial was a double blind RCT, which compared rivastigmine (up to 6mg daily) versus placebo in 120 people with Lewy body dementia diagnosed on clinical grounds. 1 Two primary efficacy measures were used: the neuropsychiatric inventory scores (NPI-10 and NPI-4) and a combined STEER 2002; Vol 2: No.2 p.5

7 score that indicated speed of response to selected tests from the cognitive drug research computerised cognitive assessment system. The original Neuropsychiatric Inventory scale (NPI-10) evaluates 10 behavioural domains (delusions, hallucinations, dysphoria, anxiety, agitation, euphoria, apathy, irritability, disinhibition and aberrant motor behaviour) based on an interview with the caregiver. 4 It was developed to help distinguish among different dementias. NPI-4 is a subscale of NPI-10 that was developed from baseline NPI analysis on the same sample of patients included in this study. The second primary efficacy measure, the computerised cognitive assessment system, consists of tests of attention, working memory, and episodic memory. The test has not been established as measure of change in dementia or of response to therapy. The second trial was unblinded and compared effects of rivastigmine versus aspirin on the mini-mental state examination (MMSE), ten point clock drawing test (TPC), relative stress scale and other outcome measures in 16 people with probable vascular dementia. 2 Quality of Evidence Found The first study was a well designed double blind RCT with concealed treatment allocation in 120 people with probable Lewy body dementia. 1 No hypotheses were made about between-group differences, so it is unclear whether the sample size was adequate to detect significant effects of active therapy over placebo. One of the primary efficacy measures used in this study (NPI-4) has not been validated as a measure of responsiveness to treatment for dementia. Evidence for responsiveness of the NPI-10 score to change is based on an open-label study. 5 NPI-4 is a subscore of NPI- 10 that was developed from baseline analysis of patients in this same study. This measure has not been subjected to separate reliability and validity studies. The clinical implications of changes on this scale are, therefore, unclear. Rivastigmine and placebo were presented in identical capsules and patients in each study were well matched for age, sex, mini-mental state examination score, comorbidity and other baseline characteristics. Patients were allocated to groups according to a randomised block design. Although results were analysed by intentionto-treat, a significant proportion of the patients withdrew or was lost to follow up (31% in active group v 16% in the placebo group). This may have exaggerated the results of the study. People with severe parkinsonian motor symptoms were excluded prior to randomisation. The trial in people with vascular dementia was an unblinded open study and included only 16 participants. 2 The study was published as a letter and it is unclear from this document whether patients were well matched between groups, particularly with respect to prior and concomitant therapy. No primary endpoint was specified. The authors listed several cognitive measurements taken at baseline, but only comment on results that p.6 STEER 2002; Vol 2: No.2

8 showed remarkable differences at the end of the study (see results). Aspirin was specified as the control intervention, but dose was not stated. Absolute values and absolute differences in outcome measures were not reported. Study Results The first trial found that rivastigmine (up to 6mg twice daily) significantly improved some cognitive symptoms, such as apathy, indifference, anxiety, delusion, hallucinations, and aberrant motor behaviour in people with probable Lewy body dementia. 1 Intention to treat analysis found no significant difference between rivastigmine and placebo (see table 1 below). NPI-4 scores, but not NPI- 10, attained significance only if data were analysed without intention to treat or on the basis of a last observation carried forward However, dropout rate was high (>25%), and non-intention to treat analyses may not be reliable. A significantly greater proportion of patients achieved 30% improvement in NPI-4 score (AR for 30% improvement in NPI-4 score at 5 months 48% with rivastigmine v 28% with placebo; NNT 5, 95%CI 7 to 38). However, the clinical importance of a 30% change in NPI- 4 is not clear since the validity and reliability of this score has not been established. The study reported that rivastigmine and placebo achieved similar rates of response (defined as 30% improvement) for NPI-10 score, but the absolute rates were not reported. At 12 and 20 weeks, the mean computerised cognitive assessment system speed score was significantly improved in the rivastigmine group compared with Table 2 Summary of outcome measurements in RCT of rivastigmine v placebo in DLBLewy body dementia (NPI-4) Measure Mean change from baseline in NPI-4 placebo rivastigmine Difference 95% CI Mean SD Mean SD to 20.6 Outcome 30% improvement in NPI-4 at 20 weeks AR ( placebo) AR (rivastigmine) 27.9% 47.5% RR (95%CI) ARR (95%CI) NNT (95%CI) 1.7 (1.05 to 2.76) 19.6% (2.6% to 36.6%) 5 (7 to 38) Table 2 computerised cognitive assessment system speed score: mean change (95%CI) from baseline Time to event rivastigmine* placebo* P Value 12 weeks weeks *Confidence intervals values not provided in text but observed on graph only; 95% CI crossed line of no-effect. STEER 2002; Vol 2: No.2 p.7

9 the placebo group. Results were imprecise (the confidence intervals at 20 weeks were wide for both groups) and were consistent with a lack of effect in either group (the confidence intervals spanned the null in both the active and the control groups, see table 2). Furthermore, the test has not been established as a valid measure of response to treatment in people with dementia. The trial found no significant difference between groups for clinical global change (CGC-plus) score or mini mental state examination scores. The significance or trends of change in any of several other secondary outcomes were not reported. Adverse effects, effects (predominantly nausea, vomiting, anorexia and somnolence) were more common in the rivastigmine group compared with placebo, but the proportion of patients withdrawn due to adverse events was similar in both groups. Extrapyramidal motor effects, measured by the unified Parkinson s disease rating scale (UPDRS) did not worsen with rivastigmine treatment, although emergent tremor was noted in four people treated with rivastigmine but not reported with placebo. People with severe extrapyramidal symptoms (Hoehn and Yahr score >3 or scores >3 for rigidity, tremor, or bradykinesia on UPDRS) were excluded from this study. The efficacy and safety data from this study might not apply to people falling in this category. The second trial found that rivastigmine (3 to 6mg daily) produces greater stability in performance of cognitive tasks as measured by ten point clock drawing test TPC (p < 0.05), a slightly better performance of executive functions as measured by NPI score (p < 0.01) and generally a better behavioural response as measured by the relative stress scale (p < 0.05) compared with aspirin. 2 However, the results are likely to be unreliable because of the small sample size, lack of randomisation and lack of blinding in this study. Furthermore, the authors listed several outcome measures but only reported results that favoured the active group; other measurements with negative outcomes or that were inconclusive may have gone unreported. Conclusions We found little reliable evidence about clinical efficacy and safety of acetylcholinesterase inhibitors in non-alzheimer s dementia. We found one small controlled trial in people with vascular dementia. 2 Results were inconclusive because of the small size and lack of proper randomisation. The results of a recently completed RCT in people with vascular dementia are awaited. 3 We found one RCT in people with probable Lewy body dementia. 3 It found a non-significant trend in favour of rivastigmine over placebo for improving cognitive and behavioural function measured by NPI-4 and NPI-10 score. A significantly higher proportion of people treated with rivastigmine achieved 30% improvement in NPI- 4 score compared with people given p.8 STEER 2002; Vol 2: No.2

10 placebo. However, the clinical importance of this finding is unclear, because NPI scores have not been validated as a sensitive measure of change in people with dementia, or as a measure of response to dementia treatment. The trial found that although pre-existing parkinsonian symptoms did not worsen with rivastigmine, emergent tremor was noted as an adverse event in four rivastigmine-treated participants. 1 In conclusion, the available evidence suggests that acetylcholinesterase inhibitors may prove useful for treating cognitive and behavioural symptoms in people with non- Alzheimer s dementia. However, we found insufficient evidence from studies to date of clinically important effects. Large RCTs are needed, examining outcomes that have been well validated as measures of response to therapy in dementia. Tolerability studies are needed, in particular to evaluate effects on parkinsonian symptoms in people with Lewy body dementia. 4. Cummings, J.L., et al., The Neuropsychiatric Inventory: comprehensive assessment of psychopathology in dementia. Neurology, (12): p Cummings, J.L., The Neuropsychiatric Inventory: assessing psychopathology in dementia patients. Neurology, (5 Suppl 6): p. S Roman, G., T. Tatemichi, and T. Erkinjutti, Vascular dementia: diagnostic criteria for research studies. Report NINDS-AIREN in Workshop Neurology, : p References 1. McKeith, I., et al., Efficacy of rivastigmine in dementia with Lewy bodies: a randomised, double-blind, placebo-controlled international study. Lancet, (9247): p Moretti, R., et al., Rivastigmine in subcortical vascular dementia: a comparison trial on efficacy and tolerability for 12 months follow-up. Eur J Neurol, (4): p Wilcock, G., The safety and efficacy of Galantamine in the treatment of vascular and mixed dementia, in Current Controlled Trials STEER 2002; Vol 2: No.2 p.9

11