Hcmh.ca2

human psychopharmacologyHum Psychopharmacol Clin Exp 2005; 20: 183–187.
Published online 7 February 2005 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/hup.676 A comparison of donepezil and galantamine in thetreatment of cognitive symptoms of Alzheimer’sdisease: a meta-analysis Robin D. J. Harry and Konstantine K. Zakzanis* University of Toronto at Scarborough, Toronto, Canada This review was conducted in order to determine the efficacy of donepezil and galantamine in the treatment of cognitivesymptoms of Alzheimer’s disease, and also to determine whether galantamine was a superior pharmacological intervention.
Meta-analytic methods were used to analyse the data from eight empirical studies which met the inclusion criteria specified.
By finding the mean effect sizes of the treatment on the outcome measures of cognition, it was determined that neither drugwas greatly efficacious. However, this result does not necessarily diminish the practical value of the drug. It was also foundthat galantamine was no better than donepezil at treating cognitive decline in AD. Copyright # 2005 John Wiley & Sons, Ltd.
key words — donepezil; galantamine; Alzheimer’s; pharmacology; meta-analysis injury and the presence of the apolipoprotein e4 allele(Davidson et al., 2002). Several rare, causative genetic Alzheimer’s disease (AD), also known sometimes as mutations have been identified in AD, including the dementia of the Alzheimer’s type, is the most preva- mutation of the amyloid precursor protein on chromo- lent cause of dementia in the aged population. It some 21, and presenilin genes 1 and 2 on chromo- is a progressive neurodegenerative disorder which somes 14 and 1, respectively (Cummings, 2003).
accounts for about 50% to 70% of dementia in older The most prominent neurobehavioural deficit of people. Although AD can only be conclusively diag- AD is memory impairment (Banich, 1997; Cummings, nosed at autopsy, the evolution of new techniques for 2003). Patients with AD also suffer visuospatial difficul- structural and functional imaging of the brain has ties, executive function impairment and language diffi- allowed more insight into the epidemiology of the dis- culties. The pathophysiology of AD includes amyloid ease. Diagnosis has also been improved by the provi- plaques and neurofibrillary tangles, both of which cause sion of more defined guidelines to diagnosing AD by atrophy of the cortex (Banich, 1997; Cummings, 2003).
the National Institute of Neurologic and Communica- There is also atrophy in the nucleus basalis, which tive Disorders and Stroke and the Alzheimer’s Disease causes the cholinergic deficit observed in patients with and Related Disorders Association (NINCDS— ADRDA) in 1984 (Ishii et al., 1991).
The loss of cholinergic transmission is believed to Age is the primary risk factor of AD. Gender is also be the cause of the deterioration of cognitive function a risk factor; women are more likely to die from AD seen in patients with AD (Francis et al., 1999). This than men. Other risk factors include a history of head hypothesis, known as the cholinergic hypothesis,spawned the use of pharmacological interventions thatincrease the activity of the system, and so retard theprogression of cognitive impairment. This was most * Correspondence to: Dr K. K. Zakzanis, Department of Life successfully done with acetylcholinesterase inhibi- Sciences, University of Toronto at Scarborough, 1265 Military tors. These drugs inhibit the activity of acetylcholines- Trail, Toronto, Ontario, MIC 1A4 Canada.
E-mail: [email protected] terase in the synaptic cleft, hence increasing the Copyright # 2005 John Wiley & Sons, Ltd.
amount of acetylcholine available for neurotransmis- with statistical significance) of the effect size in the sion, enhancing the activity of the cholinergic system.
Two of these drugs, donepezil (Aricept1) and galan- Meta-analyses control for biases that are potentially tamine (Reminyl1), are the focus of this review.
present in significance testing, such as subjective Donepezil was approved for treatment of mild-to- study selection and inaccurate interpretations of statis- moderate AD in the USA in 1996. In 1997, it was tical findings (Wolf, 1986). Also, Cohen’s d does not approved in the UK, and in Canada as the first drug necessitate the assumption of homogeneity of var- treatment for AD. It is highly selective for acetylcho- iance, as it controls for participant variability. Meta- linesterase, has few side-effects, and in the literature, analyses also account for the ‘file-drawer’ problem is said to have beneficial effects on cognition and daily by reporting the fail-safe N statistic. This statistic tells living in AD patients (Jones, 2000). Galantamine has the number of studies needed to overturn the mean also shown similar results. It was approved for treat- effect size that was obtained in the meta-analysis ment of mild-to-moderate AD in the USA and Canada (Zakzanis, 2001). This statistic can be a good indica- in 2001. However, galantamine has a modus operandi tion as to whether or not the results of the study can be somewhat different to that of donepezil: in addition to allosterically modulates the nicotinic receptors in the cholinergic system. This action increases the releaseof acetylcholine, and so enhances cholinergic activity The review of the available literature began with a search on Psyc-Info and PubMed. Using the advanced This difference in drug mechanisms begs the ques- search option, the search was done through the data- tion: which drug is more effective? Does galantami- base using key words as follows: galantamine, ne’s dual mechanism make it a superior intervention galanthamine, Reminyl, donepezil, Aricept, Alzhei- to donepezil in the treatment of AD? This review aims mer, dementia, efficacy, pharmacology. Relevant arti- to determine the efficacy of donepezil and galanta- cles were found in eight journals, both through the mine in the treatment of the cognitive deficits of electronic databases and through manual searches.
AD, and determine which drug is the superior pharma- The articles were retrieved through online access to the journals, and from Gerstein Science InformationCentre at the University of Toronto.
The method used to carry out this investigation was Criteria for articles to be included in this study are meta-analysis. Meta-analytic techniques provide a outlined in Table 1. Publication after 1984 ensures means of quantifying the magnitude of an effect or a that all participants were diagnosed according to the deficit. The magnitude of the effect was indexed in NINCDS—ADRDA criteria. The randomized, double- this study by the effect size statistic, Cohen’s d, which blind design reduced the risk of researcher biases in the is a measure of the degree to which a phenomenon is reporting and interpretation of the results. Effect sizes present, or in theoretical terms, the degree by which were computed from statistics such as means, standard the null hypothesis is false (Zakzanis, 1998). The deviations and standard errors of the mean.
meta-analysis reports the mean effect size, whichreflects the average of all the effect sizes of the studies included in the review. The mean effect size is theninterpreted in terms of the study.
Recorded variables in the studies included in this Cohen’s d can be converted to a corresponding investigation were mean age, gender, weight and eth- overlap statistic (OL%), which directly shows the per- nicity. The scores on the screening MMSE were also centage of overlap between patient and control sam-ples. Also, Cohen’s benchmarks provide a useful guide to interpreting the effect size. Cohen assigned the labels of a small effect to a d of 0.2, a medium (b) A randomized, double-blind, placebo controlled study design effect to a d of 0.5, and a large effect to a d of 0.8 (c) Participants with mild to moderate AD, and without current (Zakzanis, 2001). While this is a useful frame of refer- diagnosis of any other psychiatric or neurological disorder ence, it is not always practical, and care should be (d) Outcome measures of cognitive ability in AD patients(e) Study statistics convertible to the effect size statistic d taken to interpret the importance (not to be confused Copyright # 2005 John Wiley & Sons, Ltd.
Hum Psychopharmacol Clin Exp 2005; 20: 183–187.
Characteristics and participant demographics of donepezil studies Characteristics and participant demographics of galantamine studies recorded. Outcome measures of cognition were the of these studies used the ADAS-cog as a measure of cognition. As with the donepezil studies, there were (ADAS-cog), the Japanese version of the ADAS- considerably more female participants than male cog, and the mini-mental state examination (MMSE).
Table 4 includes the mean effect sizes for the mea- sures of cognition across all the studies. Effect sizes were computed according to Cohen’s d formula where Eight studies met the inclusion criteria for this inves- the mean change of the treatment group was sub- tigation; three on donepezil, and five on galantamine.
tracted from the mean change of the placebo group, In total, test results from 1174 AD patients treated with placebo, 668 treated with donepezil, and 1510 (Zakzanis, 2001). This table also includes the number treated with galantamine were recorded across meta- of effect sizes which contributed to each mean. Effect sizes were computed for each treatment dosage of the Table 2 includes a description of the study charac- drug. Also included in the table are the standard devia- teristics and demographics of the participants in the tion of the mean effect size, the overlap percentage studies on donepezil. The age of the participants did (OL %), and the minimum and maximum effect sizes not vary significantly across the studies. Notably, all that contributed to the mean. The fail-safe N at a cri- three studies had approximately twice the number of terion of 0.02 was also calculated for each outcome female participants than male participants. Two of the studies used the MMSE as well as the ADAS-cog as The ADAS-cog evaluation of donepezil yielded a d outcome measures of cognition, and Homma et al.
of 0.48, which translates to an overlap percentage of 67.8. The MMSE yielded a d of À0.36, which yielded Table 3 includes a description of the characteristics an overlap percentage of 75.6. It should be noted that and the demographics of the galantamine studies. All the ADAS-cog yields a positive d. This reflects the Copyright # 2005 John Wiley & Sons, Ltd.
Hum Psychopharmacol Clin Exp 2005; 20: 183–187.
placebo group having a higher score; a higher score on (Wolfson et al., 2000). A meta-analysis of the effects the ADAS-cog means greater dysfunction. The nega- of donepezil and galantamine on the functional per- tive d on the MMSE means that the placebo did worse formance and quality of life of AD patients may serve on the test than the treatment group; lower scores to give a better idea of the efficacy of the drugs, and mean greater dysfunction. Galantamine showed a d of 0.52 on the ADAS-cog, which means an overlap The second purpose of this review was to compare the efficacy of donepezil with that of galantamine.
Galantamine does not seem to have an advantage.
This seems counter-intuitive, because galantamine’s mechanisms, theoretically, are designed to increase This review was conducted on the clinical trials of the the amount of acetylcholinesterase in the synaptic drugs donepezil and galantamine to determine and cleft to a greater extent than donepezil. Donepezil compare their respective efficacies in treating the cog- only inhibits acetylcholinesterase. Galantamine inhi- nitive symptoms of AD. Using meta-analytic techni- bits acetylcholinesterase, modulates presynaptic nico- ques on the available literature, it was possible to tinic receptors so that they release more acetylcholine, determine how effective each of the drugs was.
and modulates postsynaptic nicotinic receptors so that While some of the AD patients treated with donepe- the neuron is activated. Despite this, it does not seem zil showed better cognitive function than the partici- to be more effective than donepezil. This could be pants in the placebo group, the number was far from because donepezil may just be a more powerful drug, the majority. As evident from the mean effect size, despite the relative simplicity of its action. Perhaps most patients showed the same decline in cognitive donepezil inhibits acetylcholinesterase more than function as did the placebo-treated patients. Galanta- mine showed the same result. The majority of people An alternative explanation is that the modulation of treated with galantamine in those studies showed the nicotinic receptors does not help the cholinergic sys- same cognitive deficits on the scales as did the patients tem as much as would be believed. Allosteric modula- tion of the presynaptic nicotinic receptors is capable This brings to question the practicality of the use of of increasing the amount of glutamate, as well as acet- either donepezil or galantamine as interventions of the ylcholine, released into the synaptic cleft (Maelicke, cognitive decline that befalls AD patients. The mean 2000). However, increased levels of glutamate in the effect sizes of donepezil and galantamine on the central nervous system have also been associated with ADAS-cog were 0.48 and 0.52, respectively. If inter- neuronal dysfunction and cell death (Danysz and preted according to Cohen’s benchmarks, these effect Parsons, 2003). By allosterically modulating the nico- sizes would fall just along the border of low to mod- tinic receptors, galantamine may be causing harm as erate efficacy of the drug. That does not paint a pro- mising picture for patients with AD.
This investigation is not without its limitations. The However, it is not entirely sensible to draw all con- number of effect sizes contributing to the mean is small, clusions about the practicality of the pharmacological and therefore these results cannot be generalized. As interventions based solely on their small effect. AD is seen by the small fail-safe N, these results can be over- a progressive neurodegenerative disorder, which is turned easily with a minimal number of studies with always terminal, and is still incurable. The effect that negligible effect sizes. Another limitation is the possibi- donepezil and galantamine have may be small, but it lity of moderator variables that this study did not exists nonetheless. Anything that could possibly slow account for. For example, the doses of the acetylcholi- the progression of AD, and give the patients a bit more nesterase inhibitor used, as well as the duration of the study, may have had a significant impact on the effect Perhaps a better analysis of the practicality of these sizes obtained. Correlational analyses would have to drugs would be a consideration of their cognitive be done to understand exactly how these moderator enhancements in conjunction with their effects on variables may have affected the outcome of this study.
the functional disability of patients with AD. It standsto reason that an improvement in cognitive function would also mean an improvement in the functionalperformance of AD patients. Also, the level of func- This meta-analytic review aimed to discover the effi- tional performance also contributes heavily to the cacy of two of the acetylcholinesterase inhibitors used quality of life that a person with AD experiences to reduce the rate of decline in patients with AD, Copyright # 2005 John Wiley & Sons, Ltd.
Hum Psychopharmacol Clin Exp 2005; 20: 183–187.
namely donepezil and galantamine. By evaluation of Ishii T, Allsop D, Selkoe DJ (eds). 1991. Frontiers of Alzheimer the mean effect size across a number of studies, it Research. Elsevier Science: Amsterdam.
Jones R. 2000. Drug Treatment in Dementia. Blackwell Science: was found that neither drug was greatly efficacious.
However, this does not mean they are worthless, for Maelicke A. 2000. Allosteric modulation of nicotinic receptors as a any treatment for a progressive incurable disorder treatment strategy for Alzheimer’s disease. Dement Geriatr Cogn It was also found that galantamine does not have an *Raskind MA, Peskind ER, Wessel T, et al. 2000. Galantamine in AD: a 6-month randomized, placebo-controlled trial with a advantage over donepezil, despite its dual mechan- 6-month extension. Neurology 54: 2261–2268.
isms of increasing cholinergic activity. This could be *Rockwood K, Mintzer J, Truyen L, Wessel T, Wilkinson D. 2001.
because the drug simply has inferior pharmacoki- Effects of a flexible galantamine dose in Alzheimer’s disease: a netics to donepezil. It could also be because galanta- randomized, controlled trial. J Neurol Neurosurg Psychiatry 71:589–595.
mine may be cancelling out the good that it does by *Rogers SL, Doody RS, Mohs RC, et al. 1998a. Donepezil harming the cholinergic system with increased levels improves cognition and global function in Alzheimer disease of glutamate. Further studies on the neurotransmission [Electronic version]. Arch Intern Med 158: 1021–1031.
of patients on galantamine are needed to rule out this *Rogers SL, Farlow MR, Doody RS, et al. 1998b. A 24-week, double-blind, placebo-controlled trial of donepezil in patientswith Alzheimer’s disease. Neurology 50: 136–145.
*Tariot PN, Solomon PR, Morris JC, et al. 2000. A 5-month, rando- mized, placebo-controlled trial of galantamine in AD. Neurology54: 2269–2276.
References marked with an asterisk indicate studies included in the *Wilcock GK, Lilienfeld S, Gaens, E. on behalf of the Galantamine International-1 Study Group. 2000. Efficacy and safety of galan-tamine in patients with mild to moderate Alzheimer’s disease: Banich M. 1997. Neuropsychology: The Neural Bases of Mental multicentre randomised controlled trial [Electronic version]. Br Cummings JL. 2003. Alzheimer’s disease. In The Neuropsychiatry *Wilkinson D, Murray J, Galantamine Research Group. 2001.
of Alzheimer’s Disease and Related Dementias. Taylor and Fran- Galantamine: a randomized, double-blind, dose comparison in patients with Alzheimer’s disease [Electronic version]. Int J Ger- Danysz W, Parsons CG. 2003. The NMDS receptor antagonist memantine as a symptomatological and neuroprotective treat- Wolf FM. 1986. Meta-analysis: Quantitative Methods for Research ment for Alzheimer’s disease: preclinical evidence [Electronic version]. Int J Geriatr Psychiatry 24(Suppl. 1): 23–32.
Wolfson C, Moride Y, Perrault A, Momoli F, Demers L, Oremus M.
Davidson GC, Neale JM, Blankstein KR, Flett GL. 2002. Aging and 2000. Drug Treatments for Alzheimer’s Disease: II: A Review psychological disorders. In Abnormal Psychology: Canadian Edition. John Wiley and Sons: Etobicoke, Ontario, Canada; Coordinating Office for Health Technology Assessment: Ottawa, Francis PT, Palmer AM, Snape M, Wilcock GK. 1999. The choliner- Zakzanis KK. 1998. Brain is related to behavior ( p < 0.05). J Clin gic hypothesis of Alzheimer’s disease: a review of progress.
J Neurol Neurosurg Psychiatry 66: 137–147.
Zakzanis KK. 2001. Statistics to tell the truth, the whole truth, *Homma A, Takeda M, Imai Y, et al. 2000. Clinical efficacy and and nothing but the truth: formulae, illustrative numerical exam- safety of donepezil on cognitive and global function in patients ples, and heuristic interpretation of effect size analyses for with Alzheimer’s disease. Dement Geriatr Cogn Disor 11: neuropsychological researchers. Arch Clin Neuropsychol 16: Copyright # 2005 John Wiley & Sons, Ltd.
Hum Psychopharmacol Clin Exp 2005; 20: 183–187.

Source: https://www.hcmh.ca/research_articles/Harry%20Zakzanis%20(2005)%20Hum%20Psychopharm%2020.pdf