An estimated 22% of adults 65 years of age or older have Alzheimer’s disease (AD).1 Because the incidence of AD doubles every 5 years after age 65, its prevalence is expected to increase as the United States population ages.2 The impact of this neurodegenerative disease encompasses a range of symptoms affecting cognition, function, and behavior.

Like many other diseases, there is a spectrum of continuum for AD, and there are different features at each stage. Artificially, the spectrum has been divided into mild, moderate, and severe stages. In the mild stage (Mini-Mental State Examination [MMSE] 21-26), cognitive and memory deficits are the most apparent symptoms. Behavioral changes, such as depression, begin to appear but may remain subtle.3 As the disease progresses to the moderate stage (MMSE 10-20), cognitive losses are striking. The ability to perform instrumental activities of daily living (ADLs), such as cooking or handling finances, declines rapidly, and behavioral symptoms increase in number and severity.3 In the severe stage (MMSE < 10), the ability to use expressive language declines and basic ADLs, such as feeding or dressing oneself, are lost (Figure 1).3 Behavioral disturbances may decrease in number; however, agitation, dysphoria, anxiety, apathy, and aberrant motor behavior tend to increase in the severe stage.4

Although AD is a complex disorder, much of its treatment can be managed successfully by primary care physicians, who can improve the lives of patients with AD with the appropriate use of treatments that maximize cognition and function while minimizing behavioral symptoms.5 Increased cognitive and functional impairment has been associated with a higher incidence of medical comorbidities.6 Effective treatment of AD may reduce the risk of comorbidities by ameliorating the deficits that mask other conditions. Clinical practice guidelines exist for the screening and diagnosis, as well as for the nonpharmacologic management, of AD.7,8 However, practical guidelines for the pharmacologic management of the disease have yet to be developed. This review focuses on the efficacy data for currently prescribed AD treatments and factors that may affect the choice of pharmacologic agent in the primary care setting.

CLINICAL ASSESSMENTS AND EXPECTATIONS
The symptoms of AD can be classified into those that affect cognition, those that affect function, and those that affect behavior. Hence, tools were developed to assess these three domains to help in the diagnosis and staging of the disease. These same tools can help to assess potential treatment effects of our therapies. Although these tools are primitive, they are often sufficient to detect a treatment signal. Relatively easy-to-administer scales that measure cognition,9 function,10 and behavioral symptoms11 have been developed, and they provide reliable indications of disease severity and symptom progression (Table).9-16 The MMSE,9 for example, can be administered readily by a trained assistant.17 This test relies on simple questions and thought exercises to measure the severity of cognitive impairment. Questions such as “What is the year? The season? The date?” measure the patient’s orientation, while attention and calculation are measured by asking the patient to spell world backwards. Similarly, functional dependence can be measured by asking a family member or other caregiver simple questions about the patient’s ability to perform tasks. The Functional Activities Questionnaire10 rates patient independence based on observer responses to questions such as “Is the patient able to keep track of current events?” These tools can also provide physicians with practical ways to measure the effects of therapy. If a patient regains or retains the ability to dress himself/herself or to work on his/her hobby, that outcome has real meaning for the patient. As all of these tools have numeric values that allow them to be used in clinical trials to look for statistical patterns in treatment effects, even relatively small changes of a point or two on these scales are useful indications that the treatment is having an effect. Clinically, the stabilization of symptoms or the ability to slow the decline are also desirable treatment outcomes and are treatment successes for patients with AD. In the absence of treatment, patients are likely to experience progressive debilitation and decline.

RELATIVE EFFICACY OF PHARMACOLOGIC TREATMENTS
The most commonly prescribed pharmacologic treatments for AD are the cholinesterase inhibitors (ChEIs), which include donepezil, galantamine, and rivastigmine. A fourth option, the N-methyl-d-aspartate (NMDA) antagonist memantine, was granted Food and Drug Administration (FDA) approval in 2003, although it has been available in Europe since 1982, originally for the treatment of organic brain syndrome. As with all other areas of medicine, physicians should examine the breadth and depth of evidence for each therapy before choosing a specific agent.

Donepezil
Donepezil hydrochloride is a reversible, noncompetitive antagonist of acetylcholinesterase. It is a once-daily preparation with an effective dose of 5 mg per day and a target dose of 10 mg per day, with a one-step titration at 4-6 weeks. It received FDA approval for the treatment of mild-to-moderate AD in 1996. Its studies span the spectrum of disease (mild, moderate, and severe AD) and provide data for cognitive, functional, and behavioral parameters, as well as mild cognitive im-pairment.18 This review will focus on the different stages of AD and not on the mild cognitive realm. Donepezil is available in 5-mg and 10-mg tablets and in orally disintegrating tablet formulations.

Mild-to-moderate AD. The cognitive benefits of donepezil treatment have been demonstrated in placebo-controlled, clinical trials in which the data were interpreted conservatively by intent-to-treat (ITT) or last observation carried forward (LOCF) analysis. In a 6-month study, patients treated with donepezil showed significant improvement on the Alzheimer’s Disease Assessment Scale-cognitive subscale (ADAS-cog) as early as week 6 and throughout the study compared with placebo-treated patients (P < 0.0001).19 In a 1-year study, donepezil-treated patients maintained their cognition throughout the study and were significantly improved compared with the MMSE scores of patients who received placebo, whose scores declined throughout the study (P < 0.001) (Figure 2).20 Long-term, open-label studies that followed patients for 321 and 4.922 years, respectively, have confirmed that the cognitive benefits of persistent treatment with donepezil are maintained.

The functional benefits of donepezil treatment have been demonstrated in placebo-controlled, clinical trials that enrolled patients with mild-to-moderate AD. In a 1-year study of the preservation of global function and ADL performance, Kaplan-Meier survival estimates showed that the time to clinically evident functional decline was extended 72% longer for patients treated with donepezil compared with placebo-treated patients (log-rank, P = 0.0019; Wilcoxon, P = 0.0051).23 Therefore, patients treated with donepezil were able to perform ADLs 5 months longer than patients receiving placebo. In the recently published AD2000 trial, donepezil treatment significantly improved functional ability over 2 years, as measured by the Bristol Activities of Daily Living Scale (BrADLS) (P < 0.0001).24 It should be noted that this trial was sufficiently powered to confirm the cognitive and functional benefits of donepezil, despite being drastically underpowered for its primary endpoints of rate to institutionalization and progression of disability (only 565 patients randomized instead of 3000).

Moderate-to-severe AD. Donepezil treatment also has been shown to improve cognition in advanced disease (baseline MMSE 5-12). In a 6-month, placebo-controlled study, donepezil-treated patients had MMSE scores that were above baseline throughout the study and were significantly improved compared with placebo-treated patients (P < 0.0001). The Severe Impairment Battery (SIB) has been shown to provide a more accurate measure of cognition in patients with advanced AD.25 The SIB scores for these patients were above baseline as early as week 4 and remained above baseline throughout the study. The SIB scores were significantly improved in patients treated with donepezil compared with the scores of placebo-treated patients, which declined throughout the study.25

In the same study (baseline MMSE 5-12), global function as measured by the Clinician’s Interview-Based Impression of Change with caregiver input (CIBIC-plus) remained above baseline until week 24, whereas scores from the placebo-treated patients declined throughout the study. Significant treatment differences favoring donepezil were observed at each evaluation and at the study endpoint (P < 0.0001). The ability to perform ADLs as measured by the Disability Assessment for Dementia (DAD) remained stable throughout the study for the donepezil-treated group, in which scores declined 0.74 point, while the placebo-treated group declined nearly 9 points by the end of 6 months.25

Donepezil treatment has been shown to improve behavioral symptoms in advanced AD—the disease stage in which the majority of behavioral symptoms emerge.3 Neuropsychiatric Inventory (NPI) 12-item total scores were used to assess the behavioral component. For donepezil-treated patients, the NPI score improved 4.6 points, while total scores for the placebo-treated patients worsened by 1 point (P < 0.001). On individual components of the NPI, all of the scores were improved for donepezil-treated patients compared with placebo-treated patients.26 These benefits translate into improved mood for the patient and less burden for the caregiver.

Summary. The data from clinical trials of donepezil show that it is an effective treatment across the spectrum of AD, ranging from the mild-to-moderate to severe stages. There were benefits in the domains of cognition, function, and behavior. Donepezil is the only ChEI with 1- and 2-year, placebo-controlled trials in patients with mild-to-moderate AD.20,23,24 These placebo-controlled trials demonstrate the long-term benefits of donepezil treatment on cognition and function, helping patients retain their everyday abilities. Long-term, open-label studies show that persistent treatment with donepezil maximizes cognition and function in patients with mild-to-moderate AD.21,22 In addition, placebo-controlled trials have shown that donepezil treatment benefits continue in moderate-to-severe AD, improving patients’ cognition and behavioral symptoms and stabilizing function after 6 months of treatment.

Galantamine
Galantamine hydrobromide is a reversible, competitive inhibitor of acetylcholinesterase with nicotinic receptor modulating effects. It was approved by the FDA for the treatment of mild-to-moderate AD in 2001. Originally a twice-daily-dosing regimen with multiple titrations, a once-per-day extended-release preparation recently became available in 8-mg, 16-mg, and 24-mg dosages. The oral solution at 4 mg/mL needs to be administered twice per day. The recommended starting dose for galantamine is 8 mg per day, titrated at 4-week intervals to 16-24 mg per day.

Mild-to-moderate AD. In a 6-month, clinical trial, observed case data showed a treatment difference favoring galantamine over placebo.27 Patients who received the recommended maximum dose of galantamine 24 mg28 had ADAS-cog scores that were improved compared with placebo (P < 0.001). The ADAS-cog data from a 5-month study showed a similar cognitive benefit favoring galantamine over placebo (Figure 3).29

Observed case analysis of an open-label, extension study showed that patients who received galantamine 24 mg for 12 months had ADAS-cog scores that were stable relative to baseline even at the study end. Patients who received galantamine for the entire study had better ADAS-cog scores than those who received placebo for the first 6 months.30

A 1-year, head-to-head comparison of galantamine and donepezil in patients with mild-to-moderate AD showed no difference between the drugs for the primary efficacy measure, the BrADLS,31 which is a functional measure. Of the secondary endpoints, the MMSE showed no difference between the two treatments for the whole study population. However, if some of the patients were removed and only the moderate patients were extracted, then there seemed to be a difference between the groups. Clearly, when the primary endpoint is negative and the secondary endpoint is negative, it brings into question the validity of extracting a subgroup of patients for analysis. Head-to-head analysis will require studies with larger numbers of patients that are powered sufficiently to detect true differences between treatments.

Functional measures of galantamine efficacy in patients with mild-to-moderate AD have produced mixed results. The CIBIC-plus scores in 5- and 6-month studies indicate that a larger percentage of patients treated with galantamine remain stable compared with placebo.27,29,30 The DAD scores from two trials, however, do not show treatment benefit for patients receiving galantamine 24 mg compared with those receiving placebo.27,30 A second ADL measure, the Alzheimer’s Disease Cooperative Study-Activities of Daily Living inventory (ADCS-ADL), showed that galantamine-treated patients declined significantly compared with baseline after 5 months; however, they did not decline as much as the placebo group.29

A single study of behavioral symptoms showed that total NPI scores for patients with mild-to-moderate AD who received galantamine 24 mg were not significantly improved compared with baseline at the study end, although there was a treatment difference compared with the declining placebo group.29

Moderate AD. Post hoc analyses of two 6-month, clinical trials,32,33 one with a 6-month, open-label extension,32 suggested that galantamine offered significant cognitive benefit to patients with “advanced-moderate” AD. This advanced-moderate population was not consistently defined across the two post hoc studies (MMSE ≤ 12 vs MMSE ≤ 14), and the magnitude of the treatment effect was measured relative to placebo groups that declined precipitously over the course of the study. For the 6-month extension phase of the one study, historical placebo data were used from two 12-month studies evaluating sabeluzole (a nerve growth factor stimulant) in a small number of patients with mild-to-moderate AD (MMSE 14-22). Prospective, placebo-controlled trials are needed to confirm the efficacy of galantamine in more advanced AD.

Summary. Observed case analyses in clinical trials of up to 6 months have shown that galantamine treatment offers more cognitive benefit than placebo. A single, open-label study suggests treatment benefit at 12 months as well. However, the long-term cognitive benefits of galantamine in all disease stages need to be confirmed in randomized, placebo-controlled trials. Similarly, the cognitive benefits in patients with more advanced disease need to be confirmed. Galantamine may help stabilize global function in patients with AD. Likewise, patients treated with galantamine may derive short-term maintenance in behavioral symptoms.

Rivastigmine
Rivastigmine tartrate, a reversible acetylcholinesterase inhibitor, was approved by the FDA for the treatment of mild-to-moderate AD in 2000. It is a twice-daily dose regimen with multiple titrations. Rivastigmine is available in capsules of 1.5 mg, 3 mg, 4.5 mg, 6 mg, and in an oral solution of 2 mg/mL. Titrations start from 1.5 mg twice per day and should be increased at 2-week intervals as tolerated to 6 mg twice per day.

Mild-to-moderate AD. In two 6-month, clinical trials,34,35 the ADAS-cog scores were not significantly different from placebo for patients receiving rivastigmine 1-4 mg per day in one trial (Figure 4)35, but were statistically significant in the other trial.34 Patients receiving 6-12 mg had ADAS-cog scores that were improved compared with the placebo group in both trials by observed case analysis, while the ITT analysis yielded mixed results. These 6-month trials report very different levels of efficacy. The treatment difference reported by Corey-Bloom et al34 was perhaps overstated because efficacy was measured against a placebo group that declined much more rapidly than placebo groups in similar trials.19,27,35

In a 6-month, open-label extension of the Corey-Bloom study,34 ADAS-cog scores were compared with those of a “projected placebo group,” whose decline was modeled after the original, fast-declining placebo group. A large treatment difference was reported in this study36; however, the authors did not account for potential flaws in their placebo model. Long-term, placebo-controlled trials are needed to confirm these results.

Measures of rivastigmine efficacy on patient function have yielded consistent results in two 6-month studies. The CIBIC-plus scores in both studies fell below baseline by week 26, but they remained different from the declining placebo group. Similarly, scores on the Progressive Deterioration Scale (PDS), a measure of the ability to perform ADLs, declined from baseline by week 26. Although PDS scores were significantly different from placebo by observed case analysis, ITT analysis showed mixed results.34,35

The effects of rivastigmine on the behavioral symptoms of AD have not been assessed by a behavior-specific measure, such as the NPI. No long-term, head-to-head trials of rivastigmine and the other ChEIs have been completed.

Moderate AD. A post hoc analysis of the aforementioned extension study showed similar decline for a group of patients who were reported to have moderate AD.37 A modeling procedure was used to predict the ADAS-cog score of the original placebo group had these patients continued to receive placebo for 1 year, and a robust treatment difference was reported. However, the study was not statistically powered for this subanalysis, nor did the study account for large baseline differences between the placebo and rivastigmine groups. These methodologic flaws may limit the ability of the study to detect a true treatment difference.

Summary. The clinical trial data for rivastigmine show that cognitive benefits are improved over placebo after 6 months of treatment at 6-12-mg dosing. Rivastigmine may offer benefits over placebo in terms of global function and ADL performance. However, those functional benefits are below baseline function after 6 months of treatment. In AD, behavioral benefits from rivastigmine remains to be proven. Efficacy in mild-to-moderate AD needs to be confirmed in long-term, randomized, placebo-controlled trials in which the placebo group does not decline precipitously. Efficacy in moderate AD needs to be confirmed in robust, randomized, placebo-controlled trials.

Memantine
Memantine is a 20-year-old drug that was approved by the FDA in 2003 for the treatment of moderate-to-severe AD. Memantine is available in 5-mg and 10-mg tablets, as well as an oral solution of 2 mg/mL. It is a twice-daily dosing regimen with multiple titration steps. The starting dose of 5 mg per day should be titrated up to 20 mg per day in 5-mg increments in weekly intervals.

Moderate-to-severe AD. In a 6-month study, LOCF analysis of SIB scores showed a treatment difference favoring memantine over placebo (P < 0.001), although patients receiving memantine declined after week 12 and thereafter.38 This treatment difference was not confirmed by MMSE scores, perhaps because the MMSE is a less sensitive measure of cognition in severe AD.

In a 6-month trial, the results of functional assessments were mixed. A treatment difference was seen for the CIBIC-plus by observed case, but not LOCF, analysis. Assessment by the Global Deterioration Scale showed no between-group treatment difference, while the ADCS-ADL (modified for severe dementia) and Functional Assessment Staging scale showed treatment differences favoring memantine.38 Although some measures of function showed treatment benefits compared with placebo, all patients declined continuously after week 12.

The NPI assessment over 6 months showed no treatment difference between memantine and placebo for the behavioral symptoms of AD.38 A trial of the combination of memantine and donepezil in patients who had been receiving long-term (nearly 2.5 years) donepezil treatment showed significant improvements in cognition, ADLs, global function, behavior, and care dependence.39

Severe AD. In a 3-month, clinical trial of care-dependent patients with severe dementia (MMSE < 10), measures of patient function showed a treatment difference on the Clinical Global Impression of Change scale favoring memantine.40

Summary. Memantine provides minimal cognitive and functional benefits to patients with advanced AD because they tend to have less than expected decline. However, those benefits have not been consistently demonstrated in trials where memantine was used as monotherapy. There were statistically significant benefits in cognition and function but not in behavior. Yet, in combination with donepezil, there were statistical benefits across all the domains. Hence, at present, the evidence suggests that memantine may offer some benefit to patients when it is added to ChEI therapy in advanced AD or in patients who cannot tolerate ChEI.

RELATIVE SAFETY AND TOLERABILITY
The adverse event profiles of pharmacologic agents often dictate patient compliance, length of therapy, and the patient’s ability to tolerate the effective dose. Therefore, the risk of an adverse event is an important consideration when choosing a therapeutic agent. The meta-analysis of clinical trial data performed by Lanctôt et al41 shows that ChEIs provide significant therapeutic benefits with a relatively small risk of an adverse event.41 When discontinuation rates due to adverse events were examined by drug, however, there were clear differences: discontinuations in excess of placebo were 2% for the donepezil group, 9% for the rivastigmine group, and 14% for the galantamine group.41

A 12-week, head-to-head study of donepezil and rivastigmine reported that twice the percentage of patients in the rivastigmine group discontinued because of adverse events than did patients in the donepezil group. Reports of nausea were four times higher in patients treated with rivastigmine than in patients treated with donepezil, and reports of vomiting were three times higher in patients treated with rivastigmine than in those treated with donepezil.42

Rivastigmine is associated with significant gastrointestinal side effects,43 which may be a consequence of peripheral butyrylcholinesterase inhibition or due to its nonlinear pharmacokinetics. When dosages are increased twofold from 3 to 6 mg twice daily, blood concentrations of the drug increase by threefold.43 In one case, a single dose of rivastigmine 4.5 mg taken after being off the medication for 8 weeks caused severe vomiting with rupture of the esophagus.44

A head-to-head study of donepezil and galantamine reported similar adverse events between comparator groups—although larger percentages of galantamine-treated patients reported adverse events, including drug-related adverse events.31

Although direct comparisons of memantine and ChEIs have yet to be made, clinical trials of memantine report no difference in the incidence of adverse events between drug and placebo groups.38,40 However, in a study of combined ChEI-memantine therapy, confusion, headache, urinary incontinence, and urinary tract infection occurred more frequently in the group treated with memantine than in the group given ChEI alone.39

RELATIVE DOSING AND COMPLIANCE
Because many caregivers of community-dwelling patients with AD are family members who provide care during daily visits, a complex dosing schedule may affect patient compliance and result in the loss of clinical benefits. Donepezil is administered once daily,45 whereas rivastigmine and galantamine require twice-daily administration.28 The newly available once-per-day galantamine extended-release should help with compliance.43 Therefore, caregivers for patients who are taking donepezil or galantamine extended-release need to make only one visit per day to ensure the patient receives a full dose of medication. Donepezil also has fewer dose-titration steps, and because the initial dose of donepezil is clinically effective,45 patients receive benefits sooner with this treatment. By contrast, galantamine, rivastigmine, and memantine require multistep dosing regimens that begin with subtherapeutic doses,28,43 which may delay the onset of clinical benefits.

A retrospective analysis of prescription claims data46 showed that ease of use may translate into practical clinical benefits. Of 6635 patients, those who took donepezil were more compliant and maintained therapy significantly longer than did those who took rivastigmine. At 1 month, 80% of donepezil-treated patients continued to refill prescriptions versus only 40% of rivastigmine-treated patients, and donepezil-treated patients were almost four times more likely than rivastigmine-treated patients to remain on therapy after 3 months (57.7% vs 15.4%, respectively). Patients taking donepezil remained on therapy significantly longer than those taking rivastigmine (170.3 days vs 75.2 days; P < 0.0001). This disparity may be explained by the fact that only 57% of rivastigmine-treated patients reached an effective dose during the study, compared with 99.5% of donepezil-treated patients. This real-life data compliments the scientific knowledge that has been accumulated with regard to effective therapies for patients with AD.

SUMMARY
The loss of cholinergic neurotransmission is the primary biological problem in AD.47 Currently, there are no definitive cures, and therefore the aim of dementia treatment in primary care should be to start therapies that have been shown to improve the symptoms of AD, and to initiate this therapy promptly after diagnosis. Evidence-based clinical practice guidelines support the use of ChEIs as first-line treatments in the management of AD symptoms.48

The hallmarks of AD are gradual yet persistent losses in cognition and function and the emergence of behavioral symptoms. Although there is no cure, primary care physicians can provide effective treatments for these symptoms. The choice of therapeutic agent should be based on the available evidence. All ChEIs have been approved for the treatment of mild-to-moderate AD by demonstrating cognitive and global benefits for patients. However, donepezil is currently the only ChEI with prospective, randomized, placebo-controlled trials that show efficacy from mild to severe AD and across symptom domains. Long-term trials of donepezil have shown that persistent ChEI treatment provides symptomatic relief. Given that patients have been shown to underreport the symptoms of comorbidities as their dementia becomes more severe,6 alleviating the symptoms of AD may help physicians detect and manage other comorbidities. Therefore, treating patients with agents that have broad, proven efficacy for the symptoms of AD should improve patient outcomes.

Dr. Lin has indicated that he is a speaker for Pfizer Inc., Novartis AG, and Janssen, L.P.