Donepezil

Dementia is characterised by multiple cognitive deficits and is often accompanied by behavioural disorders and mood changes.[1,2] In industrialised countries the prevalence of dementia increases with age and is about 3 to 11% in people aged more than 65 years.[3–6]

Alzheimer’s disease is the most common type of senile dementia and accounts for 50 to 60% of dementia cases.[7] It is a progressive, neurodegenerative neurological disease which usually occurs in later life, although it may sometimes present in younger people.[8] The prevalence of Alzheimer’s disease doubles every 5 years after the age of 60 years.[6]

The initial stages of Alzheimer’s disease are most commonly characterised by the insidious development of memory impairment. Depression, anxiety or personality changes may also occur in the early stages of the disease. Signs of brain dysfunction (language difficulties, sensory deficits, impaired motor function and impaired executive function) are usually observed after a few years and may affect the patient’s ability to carry out daily activities. Psychotic symptoms and behavioural abnormalities (verbal or physical aggression, wandering) may develop as the disease progresses. Seizures and myoclonus may occur in the later stages of the disease,[1,8] and death usually occurs within 5 to 10 years of diagnosis.[3,9]

There is substantial variation in the rate of progression of the disease and the way in which symptoms manifest, thus suggesting a number of potential aetiologies and making diagnosis difficult in some patients.[3] A definite diagnosis of Alzheimer’s disease can only be made at autopsy or after a brain biopsy.[4,8]

The aetiology of Alzheimer’s disease has not yet been fully determined but risk factors for its development include age, a family history of Alzheimer’s disease, over-representation of apolipoprotein E ε4 allele, head trauma, presence of Down’s syndrome, neurotransmitter deficits and exposure to environmental toxins.[8–11]

Alzheimer’s disease is associated with deficits in a number of cerebral neurotransmitters such as acetylcholine, noradrenaline (norepinephrine) and serotonin (5-hydroxytryptamine; 5-HT).[5,8,12] However, the cognitive deficits associated with Alzheimer’s disease are thought to be primarily related to the degeneration of cholinergic neurons in the cortex and hippocampus, resulting in deficits of cholinergic transmission.[5,13] Deficiencies in the cholinergic system may also be involved in the formation of amyloid plaques and neurofibrillary tangles which are observed at autopsies of patients with this disease.[14,15] Thus, symptomatic treatment for Alzheimer’s disease primarily involves restoration of cholinergic function. Four potential options to increase central cholinergic function have been explored:[8,9]

• loading with acetylcholine precursors such as lecithin or choline

• augmentation of choline acetyltransferase (an enzyme involved in the synthesis of acetylcholine)

• use of cholinergic agonists such as bethanechol, arecoline or milameline

• inhibition of acetylcholinesterase (an enzyme involved in the degradation of acetylcholine).

The 2 former treatment options have not proved successful.[8] A number of muscarinic partial agonists are in development,[8] but, to date, the cholinesterase inhibitors have demonstrated the greatest clinical efficacy in Alzheimer’s disease.[9,16]

There are 2 types of cholinesterase enzyme, acetylcholinesterase and butyrylcholinesterase. Acetylcholinesterase is found predominantly in the brain [mainly as the globular (G)₄ and G₁ forms in mammalian brain[17]], striated muscle and erythrocytes whereas butyrylcholinesterase is mainly found in the periphery (cardiac and smooth muscle, skin and plasma).[14,15,17,18] Although preliminary evidence suggests that CNS butyrylcholinesterase may be involved in the neuronal damage associated with Alzheimer’s disease,[17,19,20] treatment strategies are aimed at acetylcholinesterase since it is this enzyme which is involved in synaptic function.[19]

Agents with putative efficacy in Alzheimer’s disease include NSAIDs and other agents with anti-inflammatory actions (such as corticosteroids), estrogens, melatonin, ginkgo biloba, antioxidants [such as selegiline (deprenyl) and tocopherol (vitamin E)] and chelating agents such as desferrioxamine.[1,5,21] Indeed, in a preliminary report of a retrospective analysis, significantly greater cognitive benefits were observed in postmenopausal women with mild to moderate Alzheimer’s disease (n = 603) when estrogens were combined with the cholinesterase inhibitor donepezil compared with donepezil alone. The increased beneficial effects of estrogens were partially negated, however, by the addition of a progesterone.[22]

Donepezil (E-2020) is a reversible, noncompetitive, piperidine-type cholinesterase inhibitor which is structurally dissimilar to other currently available cholinesterase inhibitors (fig. 1).[5,14,16,23] It is selective for acetylcholinesterase rather than butyrylcholinesterase,[14,23,24] thus its actions in the periphery should be minimal. It is indicated for the management of patients with mild to moderate Alzheimer’s disease.

Chemical structures of donepezil, rivastigmine, tacrine and galantamine.

Full size image

Most data presented in this section have been obtained from in vitro, ex vivo and animal studies. Some pharmacodynamic properties of donepezil have been evaluated in healthy volunteers and also in clinical trials in patients with mild to moderate Alzheimer’s disease (see section 4 for trial details).

2.1 Inhibition of Cholinesterases

2.1.1 Binding Characteristics

Donepezil exhibited a high affinity in vitro for acetylcholinesterase extracted from the electric eel and from human erythrocytes.[25–27] Donepezil is largely a noncompetitive inhibitor of acetylcholinesterase, although both competitive and noncompetitive inhibition of this enzyme is observed; thus donepezil is defined as a mixed inhibitor of acetylcholinesterase.[25–27] Estimated dissociation constants (K_i) for competitive and noncompetitive inhibition of acetylcholinesterase by donepezil were about 25 and 74nmol.[27] In contrast, corresponding K_i values for tacrine were higher than those for donepezil for electric eel (by about 5- and 4-fold)[26] and erythrocyte acetylcholinesterase (by about 11- and 15-fold),[27] thus showing that donepezil is the more potent inhibitor of acetylcholinesterase.

2.1.2 Selectivity for Different Cholinesterases

Effects on Brain Enzymes

Donepezil inhibited acetylcholinesterase originating from the frontal cortex of patients with moderate to severe Alzheimer’s disease less potently than that from the cortex of individuals with no neurological disease. In an in vitro study, the concentration of donepezil needed to produce 50% inhibition (IC₅₀) of acetylcholinesterase extracted from senile plaques was significantly greater (about 9- to 3000-fold) than that from other fractions of diseased brain or for acetylcholinesterase extracted from the autopsied brain of participants with no neurological disease (p < 0.05). A similar effect was observed with tacrine, rivastigmine and physostigmine; IC₅₀ values of these agents were generally 5- to 27-fold lower for acetylcholinesterase from the cortex of healthy individuals than those for acetylcholinesterase extracted from senile plaques.[28]

In an ex vivo study, donepezil (generally 16 µmol/kg orally or 8 and 16µg intracerebroventricularly) significantly inhibited acetylcholinesterase from the frontal cortex, hippocampus (the 2 areas of the brain which are most affected by Alzheimer’s disease) and hypothalamus of rodent models by about 65 to 90% of control values (p < 0.05 or 0.01). In comparison, tacrine inhibited acetylcholinesterase from the same regions of the brain (by about 70 to 90% of control values, p < 0.05 or 0.01) at a dose which was 4 to 8-fold greater than that of donepezil (128 µmol/kg orally or 64µg intracerebroventricularly), thus suggesting that it is a less potent inhibitor of this enzyme.[29]

Effects on Butyrylcholinesterase

The selectivity of donepezil for acetylcholinesterase versus butyrylcholinesterase was observed in in vitro studies using rodent brain acetylcholinesterase and plasma butyrylcholinesterase[30] or using enzymes extracted from human erythrocytes.[27] The ratio of the IC₅₀ of donepezil for butyrylcholinesterase to acetylcholinesterase from rodents or humans was 1252: 1 and 405: 1, respectively.[27,30] Furthermore, these IC₅₀ ratios were about 1400-[30] and 2000-fold[27] greater than that of tacrine and about 100-fold greater than that of physostigmine[30] suggesting that donepezil is more selective for acetylcholinesterase than these 2 agents.

Similarly, donepezil (16 µmol/kg orally) did not significantly inhibit serum butyrylcholinesterase from rodents compared with that observed in control animals, whereas this effect was observed with tacrine (100 µmol/kg orally).[29]

Effects on Isoenzymes

Donepezil inhibited acetylcholinesterase from rat skeletal muscle (A isoenzyme) more potently than enzymes from rat brain (G₄ isoenzyme) or human erythrocytes (G₂ isoenzyme); IC₅₀ values were 0.22 (p < 0.05 vs G₂ or G₄ isoforms), 0.82 and 1.03 nmol/L, respectively.[28] Tacrine, physostigmine and rivastigmine also more potently inhibited A isoenzyme, but the inhibitory effects of donepezil on all 3 isoenzymes were more potent than those of comparator agents (statistical analysis not reported).[28] However, in an ex vivo study presented as an abstract, donepezil (10 mg/kg orally) more potently inhibited cholinesterase from rat brain versus enzymes from the plasma, heart or pectoral muscle. Some effect was also apparent on enzyme from the small intestine. In contrast, tacrine less potently inhibited brain cholinesterase than that from other areas.[31]

2.1.3 Effects in Healthy Volunteers and in Patients with Alzheimer’s Disease

The inhibitory effects of donepezil on erythrocyte acetylcholinesterase appear to closely correspond to those observed in the rodent cortex.[19,32] Thus, this surrogate marker has been used to demonstrate this effect in trials with donepezil.

Donepezil (1 to 10 mg/day orally for 12 to 98 weeks) inhibited erythrocyte acetylcholinesterase in 133 to 462 patients with Alzheimer’s disease.[19,32–34] Mean percentage inhibition at 6 or 12 weeks was about 19 and 44% with donepezil 1 and 3mg, respectively,[19] 64% with donepezil 5mg and 75 to 77% with donepezil 10mg.[19,32,33] The level of inhibition plateaued at donepezil plasma concentrations above 50 µg/L (achieved with donepezil 5 and 10 mg/day) and was about 77 to 90%.[19,32] The concentration of donepezil needed to produce half the maximum inhibition (EC₅₀) was about 13 to 16 µg/L.[32–34] Mean erythrocyte acetylcholinesterase inhibition at steady state was about 65 and 78% for donepezil 5 and 10mg, respectively, in healthy volunteers.[35] In another study in healthy volunteers, acetylcholinesterase inhibition was rapidly reversible with the 2mg dose, lasting for less than 24 hours. However, inhibition persisted for 72 hours with donepezil 4 and 6mg.[36]

2.2 Effects on Central Neurotransmitters

Donepezil [0.5 and 2 mg/kg subcutaneously (SC) and 0.65 and 2 mg/kg intraperitoneally (IP)] significantly increased extracellular acetylcholine levels in the hippocampus and cortex of rats.[37,38] Predictably, donepezil-induced increases in acetylcholine levels correlated strongly with its inhibitory effects on cholinesterase enzymes.[38] Increases in acetylcholine levels, measured using a microdialysis probe, were dose-dependent with peak effects occurring at or within 1 hour after the dose (peak concentrations were 2 to 19 times that of baseline values, p < 0.05 or 0.01 vs controls; values estimated from graph for SC donepezil).[37,38] Effects persisted for 2 hours after the 0.65 mg/kg dose and for 4 or 6 hours after the 2 mg/kg dose.[37,38] In comparison, tacrine had a less potent effect on acetylcholine levels than donepezil as shown by the 2.5- to 8-fold higher tacrine dose (5 mg/kg IP) needed to produce a significant effect.[37]

Donepezil 0.5 and 2 mg/kg (SC) significantly increased extracellular levels of noradrenaline (by up to 50 and 100%, respectively) in the cortex of rats, and the 2 mg/kg dose also increased dopamine levels (by up to 80%). However, neither dose affected serotonin levels.[38]

2.3 Other Cholinergic Effects

Donepezil 3 to 10 mg/kg IP produced significant hypothermia and the 10 mg/kg dose induced tremor (attributable to central cholinergic effects) which persisted for 4 or 6 hours in rodent models. However, no significant peripheral effects such as miosis or secretory signs (lacrimation or salivation) were observed.[39] In contrast, tacrine induced both centrally- and peripherally-mediated effects (p < 0.05 vs vehicle-treated control animals) but at doses that were higher than those of donepezil (generally 10 and 30 mg/kg), and effects persisted for longer with tacrine (6 to 24 hours).[39] In a second study in rats, donepezil (2 mg/kg IP) induced peripherally- and some centrally-mediated effects (exophthalmos, lacrimation, salivation, twitches and tremor) lasting for 75 minutes, but no changes were observed in the level of consciousness, muscle tone or spontaneous motor activity. A similar effect was observed with tacrine (5 mg/kg IP) persisting for 2 hours, although exophthalmos, lacrimation and salivation were more prominent with tacrine than with donepezil.[37]

2.4 Animal Models

2.4.1 Effects on Memory

Donepezil ameliorated memory deficits in various models of cognition in a number of animal studies. The minimum effective dose (MED) of donepezil (administered via different routes) producing significant reversal of induced working and reference memory deficits or naturally occurring memory deficits was generally lower than that of tacrine (table I).[27,29,40–44] However, in 1 study the MED of donepezil needed to reverse induced working memory deficits was the same as that of tacrine (1 mg/kg), although donepezil was administered IP and tacrine SC (table I).[39] In studies which included physostigmine as a comparator, the MED of donepezil which significantly reversed deficits in reference memory was higher than that of physostigmine (0.1 to 3 mg/kg vs 0.03 to 0.1 mg/kg orally),[27,40,41] whereas that needed to reverse deficits in working memory was lower (0.18 to 0.25 mg/kg vs 0.5 mg/kg orally).[43] One study did not include any comparator cholinesterase inhibitors.[45]

Pharmacodynamic effects of donepezil on memory using various models of impaired cognition in animals; data for tacrine are presented for comparison where available

Full size image

2.4.2 Effects on Attention

Donepezil (1 mg/kg IP) significantly attenuated scopolamine-induced performance impairment in various measures of the 5-choice serial reaction time test task compared with scopolamine alone (p < 0.05 or 0.01). In this study in rodent models, the effective dose of tacrine was 3-fold higher than that of donepezil.[39]

The pharmacokinetic properties of single and multiple doses of oral donepezil have been assessed in volunteers,[35,36,46–49] elderly patients[47] and patients with impaired renal or hepatic function.[50,51] Distribution data, particularly that relating to the CNS, are not available for donepezil. The pharmacokinetic properties of the drug have not been assessed in patients with Alzheimer’s disease.

3.1 Volunteers

The pharmacokinetics of donepezil can be described by a 2 compartment, open, linear model.[46,47] The maximum plasma concentration (C_max) and the area under the plasma concentration-time curve (AUC) of donepezil were linearly dose-related up to a dose of 10mg.[36,46] After single dose donepezil (5 to 10mg) the C_max (7.2 to 25.6 µg/L) was achieved in 2.4 to 4.4 hours (t_max).[36,46,47,49] In 1 study t_max decreased as the dose increased.[46] The AUC ranged from about 500 to 1000 µg/L · h and the volume of distribution (Vd) ranged from about 755 to 837L, or 14 L/kg with donepezil 5mg (table II).[36,46,47,49] Donepezil is highly protein bound (about 93%).[46,47] The presence of food did not affect the absorption of single dose donepezil in volunteers; C_max, t_max and AUC values did not differ significantly when donepezil was administered during the fed or fasted states (table II).[46]

Pharmacokinetic parameters of single or multiple doses of oral donepezil (tablets unless otherwise stated): summary of single- or multiple-dose generally randomised, placebo-controlled studies performed in volunteers, and single dose nonblind, nonrandomised studies in elderly patients and patients with impaired renal or hepatic function. Mean values are presented (unless otherwise specified)

Full size image

Steady state was achieved 14 to 22 days after repeated administration of donepezil (5 or 10 mg/day for 21 or 28 days).[35,46,48] Predictably, accumulation of donepezil was apparent (with a 5- to 7-fold increase in plasma concentration according to accumulation ratios[35,48]) until steady state.[35,46,48] However, the extent of accumulation was not dose dependent.[35,48] As seen after single doses, donepezil was highly protein bound (about 96%).[35] In general, t_max and Vd of donepezil (5 and 10mg) appeared to be similar to those achieved after single doses but C_max appeared to be higher (table II). In 1 study, t_max was shorter at steady state than after the first dose (table II) which may have been due to the increase in digestive tract motility induced by donepezil.[46]

Donepezil is largely metabolised by the liver, although some of the dose is recovered in the urine as unchanged drug (about 11 to 17%) after both single and repeated administration.[46,47,49] Donepezil undergoes significant first-pass metabolism and is metabolised by the cytochrome P450 isoenzymes 3A4 and 2D6.[49] However, in vitro studies reported in the manufacturer’s prescribing information, suggest that the effects of donepezil on these enzymes are minimal considering the low rate of binding to these enzymes (inhibitory constant = about 50-130µmol) relative to the therapeutic plasma concentrations of donepezil (164 nmol/L).[52] The major metabolites are 6-O-desmethyl donepezil, the only metabolite with similar activity to that of the parent drug, donepezil-cis-N-oxide, 5-O-desmethyl donepezil and the glucuronide conjugate of 5-O-desmethyl donepezil.[53] Various unidentified metabolites were also present in the urine.[49] More of the dose was recovered in the urine (about 23 to 57%[46,47,49]) than in the faeces (about 9 or 15%[46,49]) after single or multiple doses of donepezil (2 to 10 mg/day) or single dose ¹⁴C-donepezil (5mg), suggesting that donepezil and its metabolites are eliminated via the renal rather than the biliary route.[49]

Clearance parameters of single dose donepezil (5 to 10mg) [t_½, renal clearance (CL_R), and total body clearance (CL)] were independent of dose (table II).[36,46,47,49]

Similarly, at steady-state clearance parameters were independent of dose.[35,48] CL values (0.11 to 0.13 L/kg/h)[48] appeared to be similar to those achieved with single doses,[36] but t_½ appeared to be longer than those observed for the corresponding single dose of tablet (about 67 to 73 vs 53 hours with donepezil 5mg and about 74 vs 48 hours with donepezil 10mg) [table II]. The pharmacokinetics of donepezil appeared to be stable over a 21- to 28-day period based on CL values, which were similar to those observed in single dose studies,[35,46] and the linearity of AUC values on day 1 compared with day 21.[48]

3.2 Elderly Patients

Compared with values in young volunteers, the t_max, t_½ (table II) and mean residence time (about 85 vs 144 hours) were significantly longer and the Vd was significantly greater (table II) [p < 0.05 for all parameters] after single dose donepezil 2mg in elderly patients without Alzheimer’s disease but with other medical conditions (generally hypertension and/or needing post-cerebrovascular accident care).[47]

3.3 Hepatic or Renal Impairment

The pharmacokinetics of single dose oral donepezil 5mg were generally unaffected by impaired hepatic or renal function. Most pharmacokinetic parameters did not differ significantly between patients with compensated liver cirrhosis or moderate to severe renal impairment and healthy volunteers (table II).[50,51] However, C_max was significantly higher in patients with impaired liver function (by 37.5%, p = 0.02) [table II]. Although this difference was in the range representing clinical significance (±20%), the higher C_max was attributed to variability associated with a small study population and was not considered to be clinically significant. Furthermore, this difference may be attenuated at steady state.[50]

3.4 Drug Interactions

The potential for donepezil to interact with other drugs has been evaluated in nonblind, randomised or nonrandomised, crossover studies in 12 or 18 male volunteers (aged 18 to 55 years).[54–58] Antipsychotic agents are commonly used for the treatment of psychosis and agitation associated with dementia, but the potential for these drugs to interact with donepezil has not yet been assessed in fully published trials.

Coadministration of donepezil (5 mg/day) with cimetidine (800 mg/day) for 7 days resulted in significantly higher donepezil C_max and AUC values on day 1 (by about 15 and 10%, respectively, p = 0.001 and 0.03) and on day 7 (by about 12% for both values, p = 0.0002 and 0.0001) compared with donepezil monotherapy. The C_max of cimetidine increased only on day 1 (by 24%, p = 0.001) and the t_max was significantly lower (by about 16%, p = 0.02) on day 7. None of these changes were considered clinically significant, and no significant changes were observed in any other pharmacokinetic parameters of either drug.[54]

Plasma concentrations of donepezil also significantly increased after concurrent administration of donepezil (5 mg/day) and ketoconazole (200 mg/day) for 7 days. The C_max and AUC of donepezil increased on day 1 (by about 13 and 14%, respectively, p < 0.01 or 0.001) and on day 7 (by about 37 and 36%, respectively, p < 0.0001 for both values) compared with donepezil alone. The accumulation ratio on day 7 also significantly increased (by about 19%, p < 0.0001) suggesting that donepezil clearance is affected by ketoconazole, but this was not considered clinically relevant. In contrast ketoconazole pharmacokinetics were not significantly affected by donepezil. When donepezil data were extrapolated to estimate the effects of ketoconazole on donepezil concentrations at steady state, a 23 to 30% increase was predicted. The increase in donepezil plasma levels was lower than that observed for other agents coadministered with ketoconazole.[55]

Donepezil (5 mg/day for 6 to 10 days or 10 mg/day for 12 days) did not significantly affect the pharmacokinetic parameters of digoxin (0.25 mg/day for 6 days),[57] theophylline (starting dose 200mg twice daily and titrated to achieve optimal plasma concentrations; given for 10 days)[58] or warfarin (25mg single dose)[56] compared with those observed after any drug alone. The effects of these 3 agents on the pharmacokinetics of donepezil were not evaluated.

The efficacy of donepezil in patients with mild to moderate Alzheimer’s disease has been assessed primarily in well-designed US[19,32,33] or multinational trials.[59] Three of these were large phase III trials[32,33,59] and one a smaller phase II study.[19] These trials evaluated 161 to 818 patients (aged 50 to 94 years) with a probable or established diagnosis of Alzheimer’s disease.[19,32,33,59] The long term efficacy of donepezil has been assessed in nonblind extensions of 2 of these trials,[19,32] and these are discussed separately in section 4.2.[34,60] Three preliminary reports of long term trials in patients with mild to moderate Alzheimer’s disease are also discussed in this section.[61–63]

Patients received oral donepezil 1 to 10mg once daily for 12 to 24 weeks followed by a 2- to 6-week single-blind placebo-washout period.[19,32,33,59] In those patients receiving 10 mg/day, treatment was initiated with the 5mg dosage and increased, after 1 week, to 10mg.[32,33,59]

Primary end-points used to assess efficacy included:

• Alzheimer’s Disease Assessment Scale-cognitive subscale (ADAS-cog) which is the main efficacy measure used by most trials in Alzheimer’s disease because it assesses the cognitive deficits most often observed and is specifically designed for these patients. It assesses changes in memory and other types of cognitive function. This scale uses 11 items to evaluate the severity of different aspects of memory, language, orientation, reasoning and praxis. Scores range from 0 (no impairment) to 70 (very severe impairment).[17,32,64]

• Clinician’s Interview-Based Impression of Change (CIBIC plus), which is a modified version of the Alzheimer’s Disease Cooperative Study — Clinical Global Impression of Change, is a global rating used to assess the severity and progression of the disease. It involves a semi-structured clinician interview with patients and caregivers and assesses general symptoms, cognition, behaviour and activities of daily living using a 7-point scale (1 indicates marked improvement, 4 no change and 7 marked deterioration).[32,33]

• Clinical Global Impression of Change (CGIC) is based on investigators’ experience with patients and assesses disease severity and/or global improvement. The rating scale for severity ranges from 1 (not ill) to 7 (extremely ill).[64]

A number of secondary efficacy parameters were also used in these trials:

• Mini-Mental Status Examination (MMSE) which is a patient-based assessment of cognition primarily evaluating memory, orientation and language rather than executive function. The assessment scale has a narrow range, thus subtle changes in cognitive function are difficult to identify.[19,32,64]

• Clinical Dementia Rating Scale-Sums of Boxes (CDR-SB) which monitors symptomatic progression of the disease. It is a global scale which is based on the sum of the 6 CDR domains (boxes) of cognition (memory, orientation, judgement and problem solving) and function (community affairs, home and hobbies, and personal care).[14,32,33,64] Scores on the scale used in the trials reviewed in this section can range from 0 to 18, with a higher score indicating more severe impairment (J. Ieni and R. Pratt, Eisai, Inc., personal communication).

• Uniform Activities of Daily Living (ADL) is a caregiver-determined assessment of the patient’s ability to carry out common activities.

• Modified Interview for Deterioration in Daily living activities in Dementia (IDDD) is a caregiver-based assessment evaluating the impairment of ADL. It rates self care tasks and complex tasks, and the severity of impairment is based on the sum of 33 items rated on a 7-point scale (1 to 2 indicates no or slight impairment, 7 indicates severe impairment; total scores range from 33 to 231).[59]

• Patient- and caregiver-based quality-of-life assessments which evaluate the effects of treatment on relationships, eating, sleeping and social/leisure activities domains. Each domain is rated from 0 (poorest quality) to 50 (best quality).[19,32] It should be noted that no quality-of-life instrument has been validated in patients with Alzheimer’s disease.[32]

Efficacy assessments were based on the intention-to-treat population defined as all patients who received at least 1 dose of the drug and had at least 1 postbaseline evaluation while receiving treatment.[19,32,33,59] Three trials also included an analysis of only those patients who were fully evaluable, but results were similar between the 2 analyses and thus were not discussed further.[32,33,59] Efficacy was assessed at end-point, defined as each patient’s last assessment during the double-blind phase (last observation carried forward).[19,32,33,59] The 2 large US trials were not powered to detect differences between donepezil 5 and 10mg.[32,33]

It is important to note that treatment efficacy may be difficult to assess in Alzheimer’s disease because of the large intrapatient variation in the way this disease impacts on symptoms.[4] Results from the large multinational trial were generally presented in graphical form, thus numerical values presented from this trial have been extrapolated from these graphs.[59]

Three additional double-blind Japanese trials were conducted in patients with mild to moderate Alzheimer’s disease.[65–67] Two short term trials (8 or 12 weeks’ duration) compared donepezil 2mg with 0.1mg once daily[65] and donepezil 3 and 5mg once daily with placebo in 65 and 187 patients, respectively.[67] Another small, longer term (24 to 48 weeks) placebo-controlled trial evaluated the efficacy of donepezil 2mg once daily in 39 patients.[66] Except for ADAS-Jcog (the Japanese version of ADAS-cog) and MMSE, outcome measures differed from those employed in other trials of Alzheimer’s disease and included the mental function impairment scale (MENFIS), Hasegawa’s dementia scale (HDS), final global improvement rating (FGIR) and global utility rating (GUR).

4.1 Effects on Cognition and Clinical Function

4.1.1 Worldwide Data

In patients with mild to moderate Alzheimer’s disease (defined as an MMSE of 10 to 26 and a CDR of 1 or 2), donepezil 5 and 10 mg/day significantly improved cognition according to ADAS-cog scores (generally decreased by about 0.7 to 2.7 points)[19,32,33,59] and global clinical function according to CIBIC plus (final scores 3.8 to 4.2) relative to placebo (table III).[32,33] In the phase II trial, donepezil did not significantly alter mean CGIC scores (no further data reported), but treatment failure (defined as an end-point CGIC score of 5 to 7) occurred in significantly fewer patients receiving donepezil 5mg than placebo (11 vs 20%, p < 0.05) [table III].[19] In the multinational trial, ADAS-cog scores increased with donepezil 5mg but this was significantly less than the increase observed with placebo recipients (by 0.2 vs 1.5 points, p < 0.005) [table III].[59]

Efficacy of once daily oral donepezil (DON) in the treatment of patients with mild to moderate Alzheimer’s disease: summary of short term, double-blind, randomised, multicentre placebo (PL)-controlled trials. Mean values are presented

Full size image

An improvement of ≥4 points in ADAS-cog scores is generally recognised as being clinically significant by regulatory authorities including the US Food and Drug Administration.[32,33] Clinically meaningful improvements were observed in more donepezil 5 and 10mg recipients than in patients receiving placebo according to 3 parameters (statistical analysis not reported): a change of ≥4 points from baseline in ADAS-cog (about 38% with donepezil 5mg and 54 or 60% with donepezil 10mg vs 27 or 30% with placebo),[32,33] no decline in ADAS-cog scores during treatment (about 80% with donepezil 5 and 10mg vs 58% with placebo[33] and an end-point score of ≤3 in CIBIC plus (21 to 32% with donepezil 5mg and 25 to 38% with donepezil 10mg vs 11 to 18% with placebo) [fig. 2].[32,33,59]

Efficacy of donepezil in Alzheimer’s disease. Percentage of patients showing clinically meaningful improvements in symptoms indicated by Alzheimer’s Disease Assessment Scale-cognitive subscale (ADAS-cog) scores ≥4, no decrease in ADAS-cog scores and Clinician’s Interview-Based Impression of Change (CIBIC plus) scores ≤3 after donepezil (DON) [5 or 10mg orally once daily] or placebo (PL) for 12 to 24 weeks. Results from 3 double-blind, randomised trials in 450 to 818 patients with mild to moderate Alzheimer’s disease (A,[32] B[33] and C[59]). Note for study A: percentage of patients with ADAS-cog scores ≥4 was not reported for donepezil 5mg.

Full size image

Cognition also significantly improved with donepezil (5 and 10mg) according to the MMSE in the 2 large US trials but not in the smaller study (table III).[19,32,33] These data are consistent with a 12-week study, presented as an abstract, in 894 community-based patients with mild to moderate Alzheimer’s disease who more accurately reflected those in whom anti-Alzheimer’s drugs are used in the community than patients enrolled in other studies. Mean changes in MMSE scores were 1.28 and 1.57 with donepezil 5 and 10 mg/day, respectively (p < 0.001 vs baseline for both values).[68] Similarly, in a preliminary report of a 24-week trial in 208 nursing home patients with mild to moderate Alzheimer’s disease, mean changes in MMSE scores were significantly greater with donepezil 10 mg/day than with placebo at weeks 8, 16 and 20 (p < 0.05). Furthermore, significantly more donepezil than placebo recipients had an improvement of ≥3 points in MMSE scores for at least 1 time-point (51 vs 36%, p < 0.05). Donepezil also significantly improved CDR-SB (modified for nursing home use) compared with placebo at week 24 (p < 0.05), but no significant differences between treatment groups were observed in the Neuropsychiatric Inventory modified for the nursing home (NPI-NH; NPI assesses symptoms in patients[62]).[69]

Donepezil 10mg significantly improved functioning associated with performing complex tasks relative to placebo in the largest trial (changes in IDDD scores were about 69 vs 71, p < 0.05) [table III].[59] As expected in patients with early stage Alzheimer’s disease the ability to perform self care tasks was not impaired, thus the effects of donepezil on the ability to perform these tasks could not be assessed in this trial.[59] In a retrospective analysis of another trial[33] a substantial loss of function according to a Kaplan-Meier analysis of ADL scores occurred later with donepezil 5mg (91 weeks) and 10mg (123 weeks, p < 0.009 vs placebo and p < 0.05 vs donepezil 5mg) than with placebo (68 weeks).[70] In a third trial, however, donepezil 5mg did not significantly improve function according to the uniform ADL (table III).[19]

In the 2 trials of longest duration, donepezil (5 and 10mg) significantly improved CDR-SB scores compared with placebo, suggesting that donepezil slows the symptomatic progression of the disease (table III).[33,59]

Donepezil produced significant improvements in efficacy parameters as early as 3 weeks after initiation of therapy. Significant improvements in ADAS-cog[19,32] and MMSE[32] relative to placebo were observed from week 3 with donepezil 5 and 10mg. In the 2 remaining trials (in which patients were assessed 6, as opposed to 3, weekly) significant improvements in ADAS-cog, CIBIC plus and MMSE scores were observed from week 6 or 12 after donepezil 5 and 10mg.[33,59]

At the end of the 2- to 6-week placebo-washout period after treatment, all parameters tended towards, or were worse than, baseline values,[19,32,33,59] thus representing a decline in clinical condition. Most values at this point did not differ significantly to those observed with placebo.[19,32,33,59] However, significant differences in favour of donepezil 5 or 10mg over placebo persisted to the end of the placebo-washout period in ADAS-cog[19,32] and MMSE[32] scores and between donepezil 5mg and placebo in CDR-SB scores according to some trials.[33] A rebound effect is not expected with donepezil on treatment withdrawal according to the manufacturers’ prescribing information.[52,53]

Changes in the ADAS-cog,[19,32] MMSE,[19] CIBIC plus[32] and patient-evaluated quality-of-life assessments[19] significantly correlated with plasma donepezil concentrations (p = 0.01 to 0.001). Furthermore, changes in ADAS-cog[19,32] and CIBIC plus[32] also significantly correlated with the extent of erythrocyte acetylcholinesterase inhibition (p ≤ 0.008 or < 0.001).

4.1.2 Japanese Data

The recommended doses of donepezil in Japan are lower than those used elsewhere (see section 6).[71] Trials conducted in Japan examined the efficacy of lower daily dosages than studies performed in the US (also see section 4.2.2).

The findings of 2 Japanese trials generally did not support those observed in studies conducted in the US.[65,67] No significant differences were observed between donepezil 2 and 0.1mg[65] or between donepezil 3 or 5mg and placebo[67] in any efficacy parameter. However, when only those patients with mild to moderate Alzheimer’s disease, with baseline ADAS-Jcog scores ≥15, were assessed, donepezil 5mg was significantly better than the 3mg dosage and placebo at improving ADAS-Jcog (final scores 3.89, 1.75 and 1.63, respectively, p = 0.03 for both comparisons).[67] Significantly more of these patients receiving donepezil 5mg than 3mg or placebo were also rated as ‘more than improved’ according to the FGIR (37, 14 and 16%, respectively, p = 0.007 for both comparisons).[67]

4.2 Long Term Efficacy

4.2.1 Worldwide Data

Donepezil maintained cognition, global function or ADL for about 21 to 81 weeks. Disease progression occurred around this time, as seen by a deterioration in cognition and global function in 1 trial, but at a slower rate than that expected in untreated patients.

In a fully published extension trial performed in the US (n = 133), ADAS-cog scores improved until week 38 with donepezil (10 mg/day in ≥95% of patients), but scores began to increase from week 50 (by a mean of 6.6 points/year based on the slope of the mean change from baseline in ADAS-cog score curve), reflective of disease progression (table IV). Similarly, CDR-SB scores remained around baseline levels up to week 26 at which point disease progression was apparent as seen by the increase in scores (by a mean of 1.8 points/year according to the mean change from baseline in CDR-SB curve) [table IV].[34]

Long term efficacy of donepezil (DON) in patients with mild to moderate Alzheimer’s disease; summary of nonblind, noncomparative extensions of short term (14- or 15-week) trials performed in the US

Full size image

Changes in ADAS-cog and CDR-SB observed in this trial were less than those likely to be seen in untreated patients with Alzheimer’s disease. According to assessments using a quadratic equation, the annual deterioration in ADAS-cog scores if patients in this trial had not received treatment was estimated to be about double (11.6 points) the observed value. Furthermore, a separate study in patients with mild Alzheimer’s disease found that CDR-SB scores increased by about 2.4 points/year without treatment.[72]

Similarly, in the other US extension study in 395 patients, presented as an abstract, improvements in ADAS-cog scores were maintained for up to 51 weeks with donepezil (5 or 10 mg/day) at which point most patients (>90%) were receiving the 10 mg/day dose (table IV). CDR-SB scores remained around baseline values until week 39 after which scores began to deteriorate (table IV).[60]

A preliminary analysis of a nonblind, noncomparative trial performed in routine clinical practice in the UK in 9 to 70 patients (depending on the assessment point) receiving donepezil (10 mg/day in most patients) included an interim analysis after about 81 weeks (18 months). ADAS-cog and MMSE scores improved by a mean of 1.07 (statistical analysis not reported) and 0.96 (p < 0.02) points, respectively, at about 13.5 weeks (3 months) and improvements were maintained for about 27 (6 months; ADAS-cog) or 81 (MMSE) weeks. ADAS-cog then deteriorated at about 40.5 and 54 weeks (9 and 12 months), but this result may have been influenced by data from nonresponders, and a benefit according to ADAS-cog scores was observed for up to about 68 weeks (15 months) in patients showing a response. Overall, ADAS-cog improved in 58% of patients at 3 months with 39% of patients having a ≥4-point increase from baseline and 17% of patients having a ≥7-point increase. Similarly, MMSE improved in 56% of patients at this time-point. NPI scores and NPI-D (NPI distress scale; assesses caregiver distress) scores also improved for up to 81 weeks.[62]

Results of the first double-blind, randomised, placebo-controlled trial to assess the long term (52 weeks) efficacy of donepezil have been presented in a preliminary report. Donepezil 10 mg/day significantly improved global function, cognition and ADL according to the Gottfries, Bråne and Steen (GBS) scale, MMSE and the Progressive Deterioration Scale (PDS) in this trial recruiting 192 evaluable Northern European patients. Changes from baseline on the GBS scale and MMSE were significantly greater with donepezil than with placebo at weeks 24, 36 and 52 (p < 0.05 or ≤ 0.02) and those on the PDS were significantly greater at week 52 (p < 0.05) [data presented in graphical form].[61,73]

In another double-blind, randomised 54-week US trial (n = 431), presented as an abstract, donepezil 10 mg/day delayed the median time to clinically significant loss of function by about 21 weeks (5 months) compared with placebo according to the Alzheimer’s disease functional assessment and change scale (ADFACS).[63]

4.2.2 Japanese Data

As mentioned in section 4.1.2 the recommended doses of donepezil in Japan are lower than those employed elsewhere (see section 6).[71]

In the small Japanese trial no significant differences were observed between donepezil 2mg and placebo in any efficacy parameter (MMSE, FGIR, GUR, HDS) during long term treatment.[66]

4.3 Effects on Quality of Life

In general, donepezil did not appear to improve the quality of life in patients with mild to moderate Alzheimer’s disease. However, as previously mentioned, no quality-of-life assessment has been validated in patients with this disease.

Quality-of-life assessments showed a large degree of inter- and intrapatient variability.[19,32,59] Generally, there were no significant differences in effects on quality of life between donepezil (5 and 10mg) and placebo,[19,32,33,59] although in 1 trial there was a trend towards improvement from week 12 with the 5 and 10mg doses, and donepezil 5mg was significantly better than placebo at week 24 (p = 0.05) [table III] but not at end-point.[33] In another trial, quality of life significantly deteriorated with donepezil 10mg (p = 0.02 vs placebo), but tended to improve with donepezil 5mg and placebo (table III).[32]

A preliminary report of a 12-week trial in 318 patients with mild Alzheimer’s disease assessed the effects of donepezil on the quality of life of caregivers. Donepezil 5 and 10 mg/day reduced caregiver stress as shown by the reduction in score on the Relative Stress Scale (a measure of caregiver distress). Mean scores decreased by 0.51 with donepezil 5 and 10 mg/day but increased by 1.65 with placebo (p < 0.01).[74] These results were consistent with an interim analysis of a 14-month trial in 109 patients with mild to moderate Alzheimer’s disease completing the first 4 months of treatment. Donepezil 5 and 10 mg/day significantly reduced caregiver burden as shown by improvements in total scores on the Zarit Caregiver Burden Scale (p < 0.03). This significant benefit was observed for up to 4 months.[75]

Data presented in this section have been obtained from 4 short term, placebo-controlled trials (treatment period = 12 to 24 weeks)[19,32,33,59] and a long term, noncomparative extension trial of 1 of these (treatment period = 98 weeks)[34] in 133 to 818 patients with mild to moderate Alzheimer’s disease (see section 4 for study details). Most trials reported treatment-emergent adverse events that were defined as signs and symptoms beginning during or after the first dose or a pre-existing sign or symptom becoming more severe during treatment.[19,32–34]

5.1 Systemic Events

In short term trials, donepezil (5 and 10 mg/day) was well tolerated, and most adverse events were mild and transient and generally related to the cholinergic system.[19,32,33,59] The overall incidence of adverse events appeared to be similar between donepezil (5mg) and placebo (67 and 79% vs 65 and 76%, respectively)[19,32] but numerically higher with the 10mg dose compared with placebo (78 vs 69% and 86 vs 76%) according to 2 trials presenting these data (statistical analysis not reported for any comparisons).[32,59] Treatment discontinuation rates because of adverse events, whether treatment-emergent or not, appeared to be similar between donepezil 5mg and placebo (about 4 to 9% vs 1 to 10%) but greater with donepezil 10mg (9 to 18%) compared with placebo (statistical analysis not reported for either comparison).[19,32,33,59]

Individual events occurring significantly more frequently with donepezil (primarily the 10mg dose) than with placebo (p < 0.05 or 0.001) were nausea (17 to 22% vs 4 to 8%), diarrhoea (13 to 17% vs 3 to 7%), insomnia (18vs 5%), vomiting (10 vs 2%), fatigue (8 vs 2%) and muscle cramps (8 vs 1%).[32,33] In the largest trial, adverse events relating to the digestive (nausea, diarrhoea, vomiting and anorexia) and nervous systems (dizziness, confusion and insomnia) occurred more frequently with donepezil (5 and 10mg) than with placebo (overall incidence 36 vs 24% and 38 vs 29%, respectively, p ≤ 0.05 for both comparisons).[59] The increased incidence of events associated with donepezil 10mg may have been due to the forced rapid increase (after 7 days) from 5 to 10 mg/day dictated by trial protocol.[32,33,59] In a nonblind trial, presented as an abstract, titrating the dose from 5 to 10 mg/day over a 6-week period reduced the incidence of adverse events associated with donepezil 10mg to a level similar to that observed with the 5mg dosage and placebo in previous trials[32,33] (fig. 3).[76] The manufacturers of donepezil suggest increasing the dose from 5 to 10 mg/day after a 4- to 6-week period (see section 6).[52,53]

Incidence of adverse events with various dose titration schedules of donepezil. The incidence of adverse events with donepezil (DON) 5 mg/day (no titration),[32,33] donepezil 10 mg/day (increased to this dose over 1[32,33] or 6[76] weeks) and placebo (PL)[32,33] in 269 to 315 patients with Alzheimer’s disease. Results from a nonblind trial[76] compared with combined results from 2 double-blind, randomised trials.[32,33]

Full size image

As with short term use, adverse events associated with long term donepezil administration (10 mg/day in ≥95% of patients) were generally mild and transient. The overall cumulative incidence of events was 83%, but many of these were related to the underlying disease. Individual events with an incidence of >5% included agitation, dizziness, pain, urinary tract infection, diarrhoea, common cold, upper respiratory tract infection, headache, nausea, confusion, insomnia and accidents (fig. 4). About 4% of patients withdrew from treatment because of nonserious adverse events.[34]

Tolerability profile of donepezil. Incidence of adverse events with donepezil (10mg orally once daily in most patients) after long term treatment (98 weeks) in 133 patients with mild to moderate Alzheimer’s disease. Results from a nonblind, noncomparative, nonrandomised trial.[34] URTI = upper respiratory tract infection.

Full size image

5.2 Serious Events

The incidence of serious adverse events, whether treatment-emergent or not, was numerically greater with donepezil 10mg (10%) than with the 5mg dose (5%) or placebo (6%) in 1 trial,[33] but similar between donepezil (5 or 10mg) and placebo in a second trial (4 vs 5%)[32] [statistical analysis not reported in either study] and the same between donepezil (5 or 10mg) and placebo in the largest trial (9 vs 9%).[59] Serious events considered to be possibly related to donepezil (5 or 10mg) included stomach ulcer with bleeding (1 patient), syncope and transient ischaemic attack (1 patient) and nausea, aphasia, tremor and diaphoresis (1 patient).[32] Treatment discontinuation rates because of serious events were similar between donepezil 5mg (0%) and 10mg (1%) and placebo (1%) [statistical analysis not reported].[32]

Serious adverse events after long term use were reported by 26% of 133 donepezil recipients (34 patients) resulting in treatment discontinuation by 7% (9 patients). Most events were not considered to be related to donepezil except for fall (assessed as being possibly related to treatment by both the investigator and the drug company), syncope, possible seizure and depression (assessed as being possibly related to treatment by the drug company but not the investigator) and cancer, atrial arrhythmia, anxiety and agitation (assessed as being possibly related to treatment by the investigator but not the drug company).[34]

5.3 Cardiac and Laboratory Events

Mean heart rate decreased from baseline values with donepezil 5 and 10mg (by 2.65 and 2.26 beats/min, respectively) in 1 trial. Although these reductions were significantly greater than those observed with placebo (0.09 beats/min, p < 0.03), they were considered to be clinically insignificant. Furthermore, the incidence of bradycardia (heart rate <50 beats/min) did not differ significantly between active treatment and placebo (no further information reported).[32]

ECG changes were observed in 2 patients receiving donepezil 5mg and 2 placebo recipients. However, ST abnormalities were already present at baseline in 1 patient receiving donepezil, and neither donepezil recipient reported cardiovascular-related adverse events.[32]

Donepezil had no clinically significant effects on vital signs or haematological or biochemical findings after both short and long term use, and importantly hepatotoxicity was not observed.[19,32–34,59]

Donepezil is indicated for use in patients with mild to moderate Alzheimer’s disease. The recommended dose is 5 to 10mg orally at night. In countries other than Japan, it is recommended that treatment should be initiated with the 5mg dose and continued for 4 to 6 weeks (to allow steady state to be achieved and to minimise adverse events) before increasing the dose to 10mg. Sudden withdrawal of treatment does not result in a rebound effect.[52,53]

In Japan the recommended initial donepezil dose is 3 mg/day increasing to 5 mg/day after 1 or 2 weeks.[71] The dosage used in Japan is lower than that used in other countries. This lower dosage was decided in order to reduce the occurrence of adverse events (Eisai Japan, personal communication).

Donepezil does not interfere with the metabolism of theophylline, warfarin, cimetidine or digoxin and, conversely, the metabolism of donepezil is not clinically affected by cimetidine (see section 3.4) or digoxin. No interactions were observed between donepezil and selective serotonin reuptake inhibitors, antipsychotics or anti-Parkinsonian treatment in long term studies.[53] No clinically significant interaction was observed between donepezil and ketoconazole in a short term trial in healthy volunteers (see section 3.4).[55] However, the potential for a pharmacodynamic interaction does exist between donepezil and drugs with anticholinergic activity, succinylcholine, other neuromuscular blocking agents, and cholinergic agonists.[53]

Cholinesterase inhibitors may exaggerate the activity of succinylcholine-type muscle relaxants, have vagotonic effects on heart rate (of importance in patients with ‘sick sinus syndrome’ or other conditions involving supraventricular cardiac conduction) and may have the potential to cause generalised seizures. However, as mentioned in section 1, seizures can be a manifestation of the disease itself. These agents should be prescribed with care to patients with a history of asthma or obstructive pulmonary disease, and monitoring is recommended when used in patients at risk of developing ulcers.[53]

Alzheimer’s disease has a major socioeconomic impact and results in substantial personal and emotional stress for patients and for their families and/or caregivers. The main economic burdens associated with this disease include institutionalisation, which comprises the largest cost, and the need for a caregiver to assist the patient with the activities of daily living in the community.[3] In many cases, however, the caregiver is an unpaid family member.[4]

The incidence of Alzheimer’s disease is predicted to increase in the future because of the growing proportion of elderly people in the population in countries such as the US and Japan, as well as in other developed countries.[4,64,77]

Treatment strategies in Alzheimer’s disease aim to optimise functional ability by improving cognition, mood and behaviour, improve quality of life, delay the need for institutionalisation, reduce patients’ and caregivers’ burden and, if possible, delay disease progression.[11,64,78]

As mentioned in section 1, Alzheimer’s disease is associated with deficits of cholinergic transmission. The cholinesterase inhibitors are currently the most promising group of drugs for the management of this disease. Importantly, these agents may not be effective in all patients since only 50% of patients have cholinergic deficits; 30% may also have noradrenaline deficits.[17] Furthermore, the response to cholinesterase inhibitors may decrease with time since the cholinergic system continues to deteriorate and a once effective treatment may, thus eventually become ineffective.[17]

Physostigmine was the first cholinesterase inhibitor available for patients with Alzheimer’s disease, but limitations such as its short duration of action, poor ability to cross the blood-brain barrier and its severe adverse events have precluded further use of this agent.[12,13,79] However, a slow-release form of physostigmine is currently under investigation.[12,13,79,80]

Donepezil is a more recently approved cholinesterase inhibitor. Other agents in the same class include tacrine and rivastigmine. The comparative properties of these 3 agents are presented in table V. A variety of other cholinesterase inhibitors are currently in various stages of clinical development.[80]

Comparative properties of donepezil (DON), rivastigmine (RIV) and tacrine (TAC)[5,11,16,17,19,23,32,33,53,81]

Full size image

A number of other agents are also available in Japan (such as the nootropic agents). However, these drugs tend to have been approved based more on their tolerability profile than on their efficacy.[17]

As mentioned in section 1, donepezil is a reversible, noncompetitive cholinesterase inhibitor. It is selective for neuronal acetylcholinesterase and does not appear to inhibit butyrylcholinesterase, thus peripheral adverse events should theoretically be minimal with this drug. Donepezil also appears to inhibit acetylcholinesterase, to some extent, in other tissues such as skeletal muscle, heart, muscle and small intestine according to in vitro and ex vivo studies. Additional data on the distribution of donepezil are not available.

Well-designed, short term (14 to 30 weeks) US and multinational studies indicate that donepezil (5 or 10mg) significantly improves cognition and global clinical function in patients with mild to moderate Alzheimer’s disease. Improvements in cognition were observed from as early as 3 weeks after the start of treatment. In 2 trials an effect on symptomatic disease progression with donepezil was observed according to improvements on the CDR-SB scale.[33,59] To date, no agent is able to alter the underlying neurodegenerative disease process (disease modification effect).[6] A modest level of improvement in symptoms or the ability to maintain symptoms at a stable level for 6 to 12 months is a realistic goal of treatment because of the progressive nature of this disease; dramatic improvements in patients’ conditions may be an unrealistic expectation. Although donepezil appears to have only modest short term efficacy, effects on cognition and global clinical function were clinically meaningful.

Donepezil 10 mg/day significantly improved the ability to perform complex tasks in the largest short term trial, and in a retrospective analysis of another study donepezil 10 mg/day significantly delayed loss of function according to ADL (by 55 weeks) compared with placebo. However, in 1 trial donepezil 5 mg/day did not significantly improve ADL.

Although donepezil does not appear to improve patients’ quality of life, no instrument for assessing this parameter has been validated in patients with Alzheimer’s disease. However, in 1 large study, quality of life significantly deteriorated with the 10mg dose. As mentioned previously, caregivers of patients with Alzheimer’s disease experience a great deal of stress, thus studies should ideally assess the effects of treatment on the quality of life of both patients and caregivers.[6] Preliminary reports of a 12-week study and after the first 4 months of a 14-month study suggest that donepezil (5 and 10 mg/day) reduces caregiver distress (see section 4.3). In another preliminary study the level of caregiver stress was dependent upon the severity of patients’ neuropsychiatric, but not cognitive, symptoms.[82]

In the long term, donepezil appears to maintain cognitive and global function for up to about 1 year before subsequent gradual deterioration occurs. In interim analyses of long term extensions (98 or 144 weeks) of phase II or III trials, cognition deteriorated after 38 to 51 weeks and global clinical function deteriorated after 26 to 39 weeks of donepezil (3 to 10 mg/day) treatment following an initial improvement or stabilisation of these symptoms. However, donepezil may delay the deterioration compared with that observed in untreated patients. Similarly, in a preliminary report of another trial, improvements in cognition (according to ADAS-cog) with donepezil 10mg were maintained for up to about 81 weeks in patients showing a response to treatment. Consistent with these data, preliminary reports of double-blind, randomised, long term trials showed that significant improvements in cognition and function compared with placebo were maintained for 21 to 52 weeks.

Donepezil (5 and 10mg) was well tolerated during both short and long term use, and most adverse events were mild and transient. The overall incidence of adverse events during short term use was similar between donepezil 5mg and placebo but appeared to be higher with the 10mg dose. Predictably, most adverse events were cholinergic in nature and were generally related to the gastrointestinal and nervous systems. These events occurred significantly more frequently with donepezil 10mg than with placebo, but this may have been due to the rapid dose increase used in early studies. It appears these can be minimised by increasing the dose after a longer period (6 weeks), although there are no data assessing the recommended 4-week titration interval. No clinically significant effects on laboratory parameters or vital signs were observed and, importantly, there was no evidence of hepatotoxicity even after long term use.

An important issue concerning the treatment of Alzheimer’s disease is deciding when the drug should be stopped. Treatment should be withdrawn if tolerance or compliance is poor, if the patient’s condition continues to deteriorate at the pretreatment rate after 3 to 6 months of treatment or if there is an accelerated deterioration.[83] It is thought that a temporary treatment-free period can help to assess whether the patient is continuing to benefit from the drug,[12,83] but when treatment is restarted the subsequent level of benefit may not be the same as that achieved with initial treatment. However, a preliminary report of a recent long term extension trial (n = 760) shows that temporary withdrawal of treatment may be detrimental to patients.[84] In this report donepezil (5 or 10 mg/day) was re-initiated after the placebo-washout periods (3 or 6 weeks) of 2 short term trials (see section 4).[32,33] Original levels of benefit (assessed using the ADAS-cog and CDR-SB) were achieved when treatment was restarted after the 3-week treatment-free period but not after the longer period (6 weeks).[84] In short term clinical trials, most efficacy parameters generally deteriorated during the 2- to 6-week placebo-only period.

Donepezil has a good tolerability profile and has shown significant efficacy in well-designed, placebo-controlled clinical trials, primarily improving cognition and global clinical function. Tacrine is associated with a number of limitations (peripheral cholinergic adverse events, poor gastrointestinal tolerance, hepatotoxicity, the need for 4 times daily dosing and the need to titrate the dose over several weeks[12,14]), thus donepezil should be used in preference to this agent. Indeed, donepezil has been recommended as first-line treatment ahead of tacrine by the American Psychiatric Association,[1] although other agents in this class are not yet available in the US.[1,12] Whether donepezil continues to be the preferred drug can only be determined once direct comparative studies with other newer cholinesterase inhibitors, such as rivastigmine, become available.

Additional information regarding donepezil which would prove useful include:

• fully published studies assessing the effects on ADL since this impacts directly on the ability of patients to maintain independence, the burden of caregivers and the need for admission to a nursing home

• effectiveness of a 1-month dose-increase schedule at reducing adverse events

• effects on isoenzymes originating from humans (as opposed to rodents)

• distribution data including the drug’s ability to penetrate into the CNS

• the likelihood of interactions with other drugs commonly used in patients with Alzheimer’s disease

• further assessments of the effects of treatment on the quality of life of caregivers.

Although Alzheimer’s disease is the most common type of senile dementia, mixed dementia disorders are also frequently observed.[85] Thus, whether donepezil shows efficacy in these patients would also be of interest.

The cost effectiveness of donepezil may depend largely on the ability of the drug to improve the patient’s daily functioning and behaviour (thereby reducing caregiver burden) and to delay institutionalisation. Present pharmacoeconomic data suggest that treatment with donepezil is effectively cost neutral in patients with mild to moderate Alzheimer’s disease. Appropriate prescribing of the drug will be important to optimise its cost effectiveness.[3] A full pharmacoeconomic evaluation of donepezil treatment has been presented in a previous review (see Foster and Plosker[3]).

In conclusion, donepezil (5 and 10mg) is an agent with a simple once-daily dosage schedule which improves cognition and global clinical function in the short (up to 24 weeks) and long term (for up to about 1 year) in patients with mild to moderate Alzheimer’s disease. Improvements in ADL were also observed with donepezil 10 mg/day in these trials. Adverse events associated with donepezil are mainly cholinergic. Donepezil has been extensively studied and should be considered a first-line treatment in patients with mild to moderate Alzheimer’s disease.