116 Vol. 2, No. 4 n August 2005

ABSTRACT

Alzheimer’s disease (AD) is the most commoncause of dementia. It is estimated that 4.5 millionAmericans have AD. Several risk factors for ADhave been identified and others are emerging. Byfar, age is the greatest risk factor, followed by fami-ly history. There are 2 forms of AD: familial (ie, earlyonset) and sporadic (ie, late onset). Familial casesusually occur before age 60 years and are causedby inheritance of mutations in 1 or more of 3 genes:PS1 (presenilin 1), PS2 (presenilin 2), and APP(amyloid precursor protein). APOE (apolipoproteinE) is the most studied gene that increases risk for ADin both familial and sporadic AD. The «4 allele ofAPOE is associated with the highest AD risk, but it isneither necessary nor sufficient for developing AD.It does appear to reduce the age of maximal preva-lence of AD in larger studies. This article reviews thecurrent data on other risk factors under investigation(ie, race, hormone therapy in women, and depres-sion), in addition to the role of APOE «4 in mild cog-nitive impairment (thought to be a possibleprecursor of the early stage of AD) and other typesof cognitive impairment. This article also discussesthe costs of AD from several perspectives (patient,caregiver, society, and the healthcare system), andthe limitations and challenges in interpreting costanalyses (in particular, the methods used to calculatethe cost of unpaid caregiver time). Attempts to per-form cost-effectiveness analyses are fraught withcomplications that do not currently per-

mit a comparison of drugs used to treat or a firmestimate of cost differences with or without drugtreatment. However, these medications, along withdrugs used to treat the neuropsychiatric symptoms,play a critical role in managing AD throughout thedisease process.(Adv Stud Pharm. 2005;2(4):116-125)

Alzheimer’s disease (AD) is gaining muchattention, in part because of severalfamous persons who have discussed theirdiagnosis publicly (eg, President Ronald

Reagan and Charlton Heston), but also because the“baby boomers” are now entering retirement age, whichis the age of risk for this disease. It is estimated that by2030, the entire baby-boom generation will be olderthan age 65 years, referred to as the “2030 problem.”1

Also, as AD is discussed more openly, the ravages of thisdisease are becoming more apparent to the general pub-lic, thus potential patients and family members have aclearer idea of the challenges they will face. TheAlzheimer’s Association, in a Gallup poll commissionedby them, indicates that AD is a familiar occurrence; 1 in10 Americans say they have a family member with ADand 1 in 3 know someone with the disease.2,3

INCIDENCE AND PREVALENCE

Alzheimer’s disease is the most common cause ofdementia, accounting for approximately 60% ofcases.4 The Sidebar, titled “Some of the MostCommon Forms of Dementia,” lists other commonforms of dementia. AD often coexists with vasculardementia (VaD), which can impede diagnosis, partic-ularly because the risk factors for VaD are so commonin older persons (dyslipidemia, hypertension, andheart disease). It is estimated that approximately 1 in

REVIEW

ALZHEIMER’S DISEASE: EPIDEMIOLOGY AND RISK FACTORS—

Manju Beier, PharmD, FASCP*

*President, Geriatric Consultant Resources LLC, ClinicalAssociate Professor of Pharmacy, University of MichiganSchool of Pharmacy, Ann Arbor, Michigan.

Address correspondence to: Manju Beier, PharmD,FASCP, President, Geriatric Consultant Resources LLC,Clinical Associate Professor of Pharmacy, University ofMichigan School of Pharmacy, 2001 CommonwealthBoulevard, Suite 205, Ann Arbor, MI 48105. E-mail: tanja@umich.edu.

Advanced Studies in Pharmacy n 117

REVIEW

10 individuals older than age 65 years and nearly 50%older than age 85 years have AD.3,5 In actual numbers,that translates to an estimated 4.5 million persons withAD, based on data from the 2000 US Census. Of these4.5 million, 7% are 65 to 74 years, 53% are 75 to 84years, and 40% are 85 years or older. The census dataalso suggest that by 2050, a 3-fold increase in the preva-lence of AD is expected (13.2 million persons).6

As will be discussed by Gary M. Levin, PharmD,BCPP, FCCP, later in this monograph, AD can be cate-gorized as mild, moderate, or severe. The current distri-bution is weighted toward mild disease, as shown inFigure 1, although all patients invariably progressthrough these stages, ultimately to death.6 Once diag-nosed, persons with AD can live from 8 to 20 years.3 Thesurvival time is determined by age at diagnosis and theseverity of other medical conditions.3,7 AD was theeighth leading cause of death in 2003, with a 6%increase in age-adjusted death rate from 2002.8

RISK FACTORS

Several risk factors for AD have clearly been iden-tified and others are emerging. By far, age is the great-est risk factor, followed by family history. Thoseindividuals with a parent or sibling with AD are 2 to 3times more likely to develop AD, and the likelihoodincreases with each additional affected family mem-ber.9 However, the exact role that genes play in ADdevelopment is not entirely straightforward.

There are, in fact, 2 forms of AD: familial (ie, earlyonset) and sporadic (ie, late onset). Familial cases usual-ly occur before age 60 years and are caused by inheri-tance of mutations in 1 or more of 3 genes: PS1(presenilin 1), PS2 (presenilin 2), and APP (amyloid pre-

cursor protein). As Jeanne Jackson-Siegal, MD, discuss-es in her article later in this monograph, all of these genesare involved with production of beta amyloid, a proteinthat forms plaques in the brain, which are the patho-physiologic hallmark of AD. Those individuals whoinherit one of these genes will almost certainly developAD; family members who do not have one of these geneshave the same risk of developing AD as the general pop-ulation.9 Familial AD is, in fact, rare. Only approximate-ly 200 family lines in the world carry these mutations.9

Sporadic AD usually occurs after age 60 years, most-ly in the seventh and eighth decade of life. Inheritance ofany genetic risk factors is more complex than with famil-ial AD. APOE (apolipoprotein E) is the most studiedgene that increases risk for AD. APOE transports choles-terol and fats throughout the body, thus suggesting thatreducing serum cholesterol levels can prevent AD. Infact, early studies appeared to show a reduced risk of ADwith statin use.10,11 However, subsequent analyses haveshown no effect on AD risk with statin use.11-14

GENETICS

The APOE gene has 3 alleles: «2, «3, and «4. «4 isassociated with the highest risk of AD, but it is neithernecessary nor sufficient for developing the disease.

Some of the Most Common Forms of Dementia

Alzheimer’s diseaseVascular (or multi-infarct) dementiaParkinson’s dementiaHuntington’s diseaseLewy body dementiaFrontotemporal dementiaHIV dementia complexCreutzfeld-Jacob diseaseDementia associated with syphilis or meningitis

HIV = human immunodeficiency virus.

Figure 1. Prevalence of Alzheimer’s Disease by AgeGroup and Severity

Data show the prevalence of severe, moderate, and mild AD in each of 3 agegroups, in a biracial community population from 3 adjacent neighborhoods inChicago, IL.AD = Alzheimer’s disease.Reprinted with permission from Hebert et al. Arch Neurol. 2003;60:1119-1122.6

118 Vol. 2, No. 4 n August 2005

REVIEW

Approximately 35% to 50% of people with AD have atleast one copy of the APOE-«4 allele. Those individualswho are homozygous for APOE «4 have an even higherrisk of AD, but again homozygosity does not guaranteethat AD will develop.9 Data from the Cache CountyStudy (a large Utah county with 5677 elderly) reveal that,of those identified with dementia, 69% had definite, pos-sible, or probable AD (see Dr Levin’s article later in thismonograph for definitions). The presence of «4 allelesreduced the age of maximum prevalence in those patientswith AD from 95 years in patients with no «4 alleles to87 years in heterozygotes and 73 years in homozygotes.15

New efforts are focusing on what may be the earlieststages of AD, or mild cognitive impairment (MCI). Therelationship between APOE «4 and MCI is not yet clear-ly defined. The Religious Orders Study is a collaborativestudy with Rush University Medical Center (and severalother US medical centers) and more than 1000 older,religious clergy (nuns, priests, and brothers) who haveagreed to medical and psychological evaluation each yearand brain donation after death.16 In this study, 181 par-ticipants identified as having MCI (and who had at least1 follow-up evaluation) were followed for a mean of 5.7years, during which 43.6% developed AD. The presenceof the APOE-«4 allele nearly doubled the risk of devel-oping AD as compared to age- and education-matched persons who did not have an APOE-«4allele (Table 1).17 However, the authors note thatthe low impact of education may be because ofthe overall high levels of education for this cohort,as compared to the levels of other cohorts.17 Bycontrast, analysis of APOE-«4 allele frequency in232 white non-Hispanic older adults showed noapparent relationship with subjective memorycomplaints.18 In a bi-ethnic sample, the impact ofAPOE «4 on age of AD onset was greater in whitenon-Hispanic individuals compared to whiteHispanic individuals, suggesting that the effectmay vary with ethnic group.19

Another group of cognitively impairedpatients have been studied. They are classified as“cognitive impairment, no dementia” (CIND),defined (in this study) as some degree of cogni-tive impairment on clinical examination andneuropsychological testing, which does notmeet the criteria for dementia according to theDiagnostic and Statistical Manual of MentalDisorders, Third Edition, Revised. In this study,APOE «4 was a significant risk factor for con-

verting from CIND to AD and was associated withdecreased age of AD onset. The age-adjusted oddsratios (OR) for age, sex, education, and presence of the«4 allele are shown in Table 2.20 Until the exact rela-

Table 1. Relative Risk of Incident AD Associatedwith Possession of an APOE-«4 Allele*

Model terms Relative risk 95% CI

Age 1.100 1.060–1.141

Education 0.992 0.924–1.066

APOE «4 1.948 1.196–3.173

Data are from 181 participants in the Religious Orders Study at RushUniversity Medical Center who met criteria for mild cognitive impairment. Aproportional hazards model included terms for age, education, and «4 allele.The results show that possession of the «4 allele nearly doubles the risk ofdeveloping AD, when controlling for age, sex, and education level. Older agewas associated with an increased risk of AD with a 10% increase in risk foreach year of age. Educational attainment showed less of an effect in this group,perhaps because of the overall high levels of education in this group overall.*Data from a proportional hazards model stratified for sex.AD = Alzheimer's disease; APOE = apolipoprotein E; CI = confidenceinterval.Reprinted with permission from Aggarwal et al. Neurocase. 2005;11:3-7.17

Table 2. Role of the APOE-«4 Allele in Progression of CINDto AD

Adjusted Odds Ratio (95% CI)

Education (for Presence vs Age (for every every year absence of

Progression year increase) Sex (F:M) increase) APOE-«4 allele

CIND → NS NS NS NSnormal (n = 29)

CIND → 1.05 NS 0.83 NSCIND (n = 85) (1.01–1.09) (0.77–0.89)

CIND → 1.13* 1.61 0.82 2.69*AD (n = 68) (1.08–1.18) (0.90–0.89) (0.76–0.89) (1.48–4.92)

CIND → VaD 1.11 NS 0.62 NS(1.01–1.23) (0.49–0.79)

Data from a nested case-control study design of the Canadian Study of Health and Aging, alongitudinal cohort (n = 10 263) of aging and cognition. From these data, the positive pre-dictive value of APOE «4 for predicting CIND conversion to AD was 0.48 and the negativepredictive value was 0.65.*Significant interaction was found between the presence of APOE-«4 allele and age, with thepresence of the «4 allele causing a decrease in age at onset of dementia. No significant first-degree interactions were found between other factors.AD = Alzheimer's disease; APOE = apolipoprotein E; CIND = cognitive impairment, nodementia; NS = not significant; VaD = vascular dementia.Reprinted with permission from Hsiung et al. CMAJ. 2004;171:863-867.20

Advanced Studies in Pharmacy n 119

tionship between possession of an «4 allele and AD isdescribed, genetic testing is not recommended for thegeneral population. Furthermore, for family membersof a patient with AD, testing is recommended onlyafter comprehensive genetic counseling.

GENDER AND HORMONE THERAPY

The role of estrogen and hormone replacementtherapy (HRT) in the risk or prevention of AD hasbecome controversial or at least unclear, given theresults of the Women’s Health Initiative MemoryStudy (WHIMS).21 Observational studies had suggest-ed that HRT may protect postmenopausal womenagainst AD, but randomized, controlled trials refutedthose ideas. Specifically, the WHIMS compared therates of dementia or cognitive decline in women tak-ing HRT as estrogen-plus-pro-gestin or estrogen-onlyreplacement therapy versus placebo.22 Neither hor-mone therapy protected against cognitive decline, andthey may have even increased the risk.23 Questionsarise as to whether the type of hormone therapy (eg,different doses and different sources of estrogen orprogestin) or the time at which the therapy is taken (ie,postmenopausal vs perimenopausal) may affect therisk of cognitive decline. Several authors have suggest-ed that both of these factors can affect cognitive out-comes.24-26 Thus, the decision to use anyperimenopausal hormone therapy should be based onclinical factors relating to menopause and any patient-specific considerations (eg, history of breast cancer),and not on any possible effect on future cognitivedecline.

SOCIAL AND MENTAL STIMULATION

One of the most interesting findings in recent yearsis the role of social and mental stimulation in predict-ing risk of future AD. Participants in the Alzheimer’sDisease Case Control Study at Case Western ReserveUniversity reported their participation in 26 nonoccu-pational activities (classified as passive, intellectual,and physical) throughout early (ages 20–39 years) andmiddle (ages 40–60 years) adulthood. Participants inthis analysis included 193 people with possible orprobable AD and 358 healthy controls. The resultsshowed that the controls (ie, those individuals not hav-ing AD) were more active during mid-life than thecases, in diversity (different types of activities) andintensity (number of hours per month), as shown inTables 3 and 4. If participation increased from early to

middle adulthood, the probability of AD decreasedsignificantly, and those individuals who performed lessthan the mean value of activities had a 3.85 higher riskof AD than those who performed at least the mean lev-els (95% confidence interval [CI], 2.65–5.58; P <.001). However, it was unclear if the reduction inactivities with time became a risk factor for AD, wasan early manifestation of the disease, or both.27

REVIEW

Table 4. Comparison Between Case- andControl-Group Members on Intensity Scores ofNonoccupational Activities

Adjusted Mean*Measure Case Control

Early adulthoodPassive intensity 62.23 68.43Intellectual intensity 57.45 67.94†

Physical intensity 31.18 34.50

Middle adulthoodPassive intensity 99.33 101.84Intellectual intensity 68.15 79.21†

Physical intensity 37.74 41.09

Activities are classified as passive, intellectual, or physical and are assessedbased on diversity of activities within each type and intensity (number of hoursper month). Cases are patients with possible or probable Alzheimer's disease.*Adjusted means take into account covariates (year of birth, education,gender, and income adequacy) in the statistical analyses.†P ≤.05.Adapted with permission from Friedland et al. Proc Natl Acad Sci U S A.2001;98:3440-3445.27

Table 3. Comparison Between Case- and Control-Group Members on Diversity Scores of Nonoccupational Activities

Adjusted Mean*Measure Case Control

Passive diversity 0.84 0.91

Intellectual diversity 0.44 0.54

Physical diversity 0.42 0.53

P ≤.001 for all comparisons.Activities are classified as passive, intellectual, or physical and are assessedbased on diversity of activities within each type and intensity (number of hoursper month). Cases are patients with possible or probable Alzheimer's disease.*Adjusted means take into account covariates (year of birth, education,gender, and income adequacy) in the statistical analyses.Adapted with permission from Friedland et al. Proc Natl Acad Sci U S A.2001;98:3440-3445.27

120 Vol. 2, No. 4 n August 2005

Lindstrom et al showed that for each additional dailyhour of middle-adulthood television viewing, the risk ofAD increased 1.3 times (controlling for year of birth,gender, income, and education).28 Hall et al showed thatrural residence during childhood can also have a mea-surable impact on risk for AD, even among those withlow education levels (≤6 years).29 The OR for AD forthose individuals with low education and rural residencewas 6.5 (95% CI, 2.6–16.7), 0.5 for those with low edu-cation and childhood urban residence (95% CI,0.1–2.9), and 1.5 for those with high education (≥7years) and childhood rural residence (95% CI, 0.4–5.2)compared to the high education and childhood urbanresidence group.29 Scarmeas and Stern reviewed thehypothesis that participation in mentally stimulatingactivities (eg, higher education) may create a cognitivereserve, which emerges from more efficient cognitivenetworks developed through mental stimulation.30,31

This cognitive reserve may delay onset of AD or permita greater tolerance for AD pathology.30,31

RACE

African Americans are known to have a higher riskof AD than whites, and also possess many of the riskfactors for vascular dementia. A report from theAlzheimer’s Association indicates that the genetic riskfactors for AD seem to be different for AfricanAmericans compared to whites, and that the APOEgenotype alone does not explain the increased fre-quency of AD in African Americans. The report alsonotes that African Americans tend to be diagnosed ata later stage, limiting the effectiveness of any treat-ments for the disease symptoms, and that the age-spe-cific prevalence of dementia is 14% to 100% higher inAfrican Americans than in whites.32 Taylor et al haverecorded a marked increase in the number of AfricanAmericans with AD, which has nearly doubled from1991 to 1999.33 The increase was particularly strong inAfrican-American women, in whom the rate increasedby 4.7-fold compared to 2.3-fold for white women. Infact, African Americans had a higher rate of identifiedAD in 1999 than whites (62.5/1000 vs 40.9/1000), areversal of rates from 1991 (13.7/1000 vs 16.5/1000),based on Medicare databases. The authors suggestedthat the causes of this increase are probably multifac-torial and may not reflect an increased risk per se ofAfrican Americans for AD. During this time period,African Americans had improved access to healthcareand there was increased sensitivity among healthcare

practitioners and the general public to AD in AfricanAmericans (ie, that its symptoms are not a normal partof aging). Also, changes in ICD-9 coding rules allowedfor identification of patients with AD when they alsohad risk factors for VaD, such as prior stroke, which ismore common among African Americans than whites.The authors were unable to estimate rates forHispanics or Asians because of the small number intheir database.33

OTHER POSSIBLE RISK FACTORS

More recently, depression and other factors are alsobeing evaluated for a role in AD onset. Depression is aknown, frequent comorbid condition with AD thatcan sometimes mask diagnosis of AD. The Multi-insti-tutional Research in Alzheimer’s Genetic Epidemiologystudy showed that the presence of depression in a fam-ily member of a patient with AD or of depressivesymptoms in a patient with AD before AD onset areassociated with the development of AD, even in fami-lies in which the first depressive symptoms occurredmore than 25 years earlier. The OR was highest ifdepression occurred within 1 year of AD onset (OR4.57; 95% CI, 2.87–7.31), lower if the depressionoccurred more than 1 year before AD onset (OR 1.38;95% CI, 1.03–1.85), but still substantial if the depres-sion first occurred more than 25 years before (OR1.71; 95% CI, 1.03–2.82).34

Two other studies have analyzed multiple risk factors for AD. Lindsay et al showed that use of non-steroidal anti-inflammatory drugs, wine consump-tion, coffee consumption, and regular physicalactivity were associated with a reduced risk of AD,whereas family history of dementia, gender, historyof depression, estrogen replacement therapy, headtrauma, antiperspirant or antacid use, smoking, highblood pressure, heart disease, or stroke were not asso-ciated with AD—conflicting with other reports onsome of these risk factors.35 Similarly, Tyas et alshowed that increased age, fewer years of education,history of migraines (especially in women), and self-reported memory loss at baseline increased the risk ofAD, whereas vaccinations and occupational exposureto excessive noise reduced the risk of AD.36 Clearly,the causes of AD are multifactorial and perhaps inter-connected, but healthy diet, regular exercise, normalhomocysteine level, and mental stimulation are mod-ifiable risk factors of which benefits extend wellbeyond reducing the risk of AD.37

REVIEW

Advanced Studies in Pharmacy n 121

REVIEW

THE COST OF ALZHEIMER’S DISEASE

Similar to risk factors, the exact cost of AD is difficultto quantify, but it is inherently obvious that AD is a dev-astating disease that exacts a large financial, emotional,and psychological toll on the patient’s family. The costsof caring for a patient with AD are numerous and varied.They include the cost of drugs to treat the symptoms,hospitalization for falls, specialized home care, ultimate-ly nursing home, hospice care, and cost to caregivers(time, stress, unemployment, and out-of-pocket expen-ditures). The types of costs change as the disease pro-gresses, as shown in Figure 2.38 Costs start as mostlyunpaid direct care and move quickly to long-term resi-dential care with increasing AD severity. Long-term carecosts ultimately constitute the major cost driver forAD.38,39 The extent to which the family pays directly forthese costs depends on their specific circumstances.

Several studies have been published, usingAmerican and other international cohorts, to estimatethe cost of AD. Creating an accurate cost-outcomeanalysis poses several challenges, in part because olderstudies may have used less accurate diagnostic methodsfor AD. However, the biggest challenge is to establisha cost or monetary value for unpaid caregiver time.The cost of unpaid care can be estimated by using areplacement wage (ie, the wage that a worker wouldhave to be paid to provide such care if the caregiverwas unable). These costs are considered to be directcosts. Caregiver costs can also be calculated using esti-mated lost wages when a caregiver has to stay home tobe with the patient with AD (an indirect cost), butmost caregivers are spouses of patients with AD andare usually of retirement age.40 Lost wages are moreapplicable to adult children of patients with AD whobecome the patient’s caregiver.

COSTS TO THE HEALTHCARE SYSTEM

For the US healthcare system, the total direct andindirect costs are estimated to be at least $100 bil-lion.3,41 For the person with AD, average lifetime costsare estimated at $174 000 (based on 1991 data).3,41 TheAlzheimer’s Association reports that Medicare costs forbeneficiaries with AD are expected to increase 75%,from $91 billion in 2005 to $160 billion in 2010.3,42 TheAlzheimer’s Association report also suggests thatMedicaid expenditures on residential dementia care willincrease 14% from $21 billion in 2005 to $24 billion in2010.3,42 Currently, the average cost for nursing homecare is $42 000 per year per patient with AD.3,43 O’Brien

and Caro also showed higher costs of nursing home carefor patients with AD compared to other types of demen-tia (mean annual costs $42 230 vs $40 470, respective-ly), based on data from the 1995 MassachusettsMedicaid nursing home database.44 Perhaps much ofthese costs are due to the severity of disease once patientsare placed in a nursing home. In this study, 80% of ADresidents were in the top 5 care levels compared to 71%of those individuals with other types of dementia (Figure3).44 This translates into an additional 103 hours of careper year for a patient with AD, based on the mediannumber of management minutes per care level. Theauthors concede that the dollar values may be unique toMassachusetts because the study used MassachusettsMedicaid 1997 per diem rates to determine costs (eachstate determines its own levels of reimbursement).However, the authors note that Medicaid is responsiblefor 72% of all nursing home resident days in the UnitedStates.44,45

Substantial costs also accrue even before nursinghome placement. In a study comparing 3 groups(patients with AD at an ambulatory AD center,patients with AD at an internal medicine practice, andcontrols, all older than age 65 years), the Medicarecosts were highest in the AD center group, but bothAD groups had much higher healthcare costs thancontrols. Mean per person annual Medicare costs were

Figure 2. Annual Healthcare Costs of Patients with ADAccording to Disease Severity

Data are from the Canadian Study of Health and Aging. Costs are expressed in 1996Canadian dollars. Caregiver time was assigned a value equal to the Ontario minimumwage. The Mini-Mental State Examination scores were used to grade severity: mild(21–26), mild to moderate (15–20), moderate (10–14), and severe (<10).Reprinted with permission from Hux et al. CMAJ. 1998;159:457-465.38

122 Vol. 2, No. 4 n August 2005

REVIEW

$19 418, $18 753, and $12 085, respectively. Of note,non-AD hospitalizations and length of stay by ADpopulations were the main cost drivers.46

COSTS TO THE PATIENTS AND CAREGIVERS

Patients with AD and their caregivers carry thelion’s share of the economic burden of AD, throughdirect medical costs and the indirect costs of emotion-al strain and lost social and professional opportunities.According to a report by the Alzheimer’s Association,AD costs American businesses $61 billion per year. Ofthat, $36.5 billion covers costs related to caregivers (ie,lost productivity, absenteeism, and worker replace-ment).3,47 More than 70% of patients with AD live athome, with family and friends providing 75% of theircare.3,48 The remaining expenditures are for paid care,costing an average of $19 000 per year, usually paidout of pocket by families.3,49 In a recent study ofpatients with AD in Massachusetts, the cost perpatient with AD was $23 436 for informal care and$8064 for formal services (with variations primarilyaccording to the level of assistance needed for instru-mental activities of daily living).50

COST OF ALZHEIMER’S DISEASE MEDICATIONS

There are currently 5 medications to treat thesymptoms of AD, which will be reviewed in detailin Part 2 of this monograph series on AD. Four ofthese medications (tacrine, donepezil, galanta-mine, and rivastigmine) are acetylcholinesteraseinhibitors, and are approved for treatment of mildto moderate AD. Memantine is a noncompetitiveNMDA receptor antagonist, approved for use inmoderate to severe AD. Tacrine is not used to anymeasurable extent because of several limitationsrelated to dosing and hepatotoxicity.

Several cost-outcome analyses have been pub-lished, in an effort to determine if the efficacy ofthese drugs translates into cost effectiveness.However, the study designs vary widely and are ofshort duration (from which long-term cost data areextrapolated, thus they do not consider the irregu-lar course of the disease), may not use empirical datafrom the efficacy studies, and do not consider dif-ferences or changes in reimbursement systems out-side the current system under study.51

For example, pooled data from 2 concurrentrandomized controlled trials compared galanta-mine to placebo (6 months of treatment) in 825

patients with mild to moderate AD. The analysis alsocompared the time spent assisting patients with theiractivities of daily living and the time patients could beleft unsupervised. In patients with mild AD, galanta-mine use resulted in 3.5 fewer hours per week requiredof the caregiver. In patients with moderate AD, care-givers spent 6 hours less each week with the patientstaking galantamine. However, nonresident caregiversof patients treated with galantamine were 3 times morelikely to say that the patient did not need supervisionversus resident caregivers, and there was no differencein the amount of time patients could be left unsuper-vised between treatment groups as reported by residentcaregivers. Thus, the ratio of resident to nonresidentcaregivers in each treatment group could have affectedthe results.52

In a study analyzing resource utilization data froma 28-week randomized, controlled trial of memantine(n = 252) versus placebo (n = 166) in patients withmoderate to severe AD, memantine use resulted in sig-nificantly less caregiver time (51.5 hours less permonth; 95% CI, -95.27 to -7.17; P = .02). Total costs,when considered from a societal perspective, decreasedby $1089.74/month (nonoverlapping 95% CI for

Figure 3. Proportion of Nursing Home Residents in EachManagement Category by Type of Dementia

Data are from a cost-outcome study of nearly 50 000 Massachusetts nursing home resi-dents covered primarily by Medicaid, of whom 8.8% had a diagnosis of AD and an addi-tional 17.6% had another type of dementia. Significantly more residents with AD wereassigned to the more resource-intensive levels of care.AD = Alzheimer’s disease.Adapted with permission from O’Brien and Caro. Int Psychogeriatr. 2001;13:347-358.44

Advanced Studies in Pharmacy n 123

REVIEW

treatment difference, -$1954.90 to -$224.58; P = .01),with the biggest differences seen as reduced total care-giver costs (-$823.77/month; P = .03) and direct non-medical costs (-$430.84/month; P = .07) withmemantine treatment. Patient direct medical costswere higher primarily because of the cost of the drug,with a trend toward lower direct nonmedical costs.53

It is also important to consider the perspective ofthe cost study—society, caregiver, managed care plan,or nursing home. For example, patients who are on theborder of being institutionalized may require muchmore home care to delay institutionalization, thusincreasing caregiver costs, but they may require theleast intensive care once admitted, thus lowering nurs-ing home costs.54 At this time, it is not yet possible tosay whether one drug is more cost effective thananother because of the specific conditions of costanalyses and the lack of direct, head-to-head compar-isons.55 A frequent quandary of cost-outcome studieswith AD treatments is that the drugs may preventinstitutionalization but as a result increase caregiverburden. However, experience tells us that drugs thatcan prolong patient functioning and cognition easenot only caregiver time requirements with the patientbut, more importantly, caregiver emotional and psy-chological burden and, for the patient, can increase thetime during which they can put their affairs in order.

As the disease progresses, patients with AD oftenexperience neuropsychiatric symptoms. These symp-toms will be discussed in more detail in Part 2 of thismonograph series, but they are present in mostpatients with AD.56-58 The most frequent neuropsychi-atric symptoms include apathy, depression, agita-tion/aggression, and wandering.56-58 These symptomscontribute greatly to caregiver stress and burden,diminished quality of life for patient and caregiver,excess patient morbidity, and cost of illness. One studysuggests that these symptoms account for 33% of theprimary caregiver’s hours of unpaid care, in addition totheir normal activities and duties as caregiver for apatient with AD.59 A recent study looked at costs of deal-ing with neuropsychiatric symptoms after accounting forseverity of cognitive impairment and comorbid medicalconditions in 128 patients with AD. Those patients withmore severe or frequent neuropsychiatric symptomsincurred higher costs than those patients with less severeor frequent symptoms, and the costs were caused pre-dominantly by significantly increased caregiver time andhigher long-term care costs. The authors caution that

these costs represent an average response for a cohort,not an individual patient, because the progression of dis-ease and its symptoms is not linear.39,60 Neuropsychiatricsymptoms can wax and wane throughout the course ofthe disease.61,62 Beeri et al found that management ofbehavioral and neuropsychiatric symptoms account forapproximately 30% of the total annual cost of AD, ofwhich more than 25% accounts for indirect annual costsand more than 35% for the total direct annual costs.59

As will be reviewed in Part 3 of this monographseries, drugs are not the only intervention for treatingAD symptoms, especially as the disease progresses.Treatments using other strategies, such as exercise, senso-ry stimulation, relaxation, and use of music or audiovi-sual stimulation, play an important role in optimizingquality of life for the patient with AD. The costs of thesetreatments must also be considered, as they are adminis-tered by nursing home medical staff or family members.Also, for the caregiver, other services are available, suchas respite care, foster care, support groups, transporta-tion, meal delivery, and part-time in-home care.39

CONCLUSIONS

Alzheimer’s disease is becoming a more frequenttopic of discussion among the general public, creatingan awareness for earlier detection and best strategies forloved ones once the disease manifests. With the largebaby-boomer population preparing to retire, the num-ber of patients with AD in the United States is expectedto increase substantially. Several risk factors have beenidentified for sporadic AD (the most common form),some of which are modifiable (eg, serum cholesterol lev-els, exercise, and mental stimulation activities). Thecosts of AD are difficult to quantify but several studiesare offering a general sense of cost. The direct medicalcare costs can be staggering, particularly because someportion of them often have to be paid out of pocket bythe patient or their family/caregivers. Studies are alsoshowing that the costs extend well beyond direct med-ical care for the patient with AD, inflicting extraordi-nary caregiver burden as disease progresses. Because thistype of care is unpaid, it is difficult to put a price tag onit. Attempts to perform cost-effectiveness analyses arefraught with complications that do not currently permita comparison of drugs or a firm estimate of cost differ-ences with or without drug treatment. However, thedrugs, along with drugs to treat the neuropsychiatricsymptoms, play a critical role in managing AD through-out the disease process.

124 Vol. 2, No. 4 n August 2005

REVIEW

REFERENCES

1. Knickman JR, Snell EK. The 2030 problem: caring foraging baby boomers-statistical data included. HealthServices Research. Available at: www.findarticles.com/p/articles/mi_m4149/is_4_37/ai_91568394. AccessedJune 9, 2005.

2. 1991 Gallup survey of 1015 individuals (as reported byreference 3).

3. Alzheimer’s Association Web site. Fact sheet: aboutAlzheimer’s disease statistics. Available at: www.alz.org.Accessed June 7, 2005.

4. Baldereschi M, Di Carlo A, Amaducci L. Epidemiology ofdementias. Drugs Today (Barc). 1998;34:747-758.

5. Evans DA, Funkenstein HH, Albert MS, et al. Prevalence ofAlzheimer’s disease in a community population of older per-sons: higher than previously reported. JAMA. 1989;262:2552-2556.

6. Hebert LE, Scherr PA, Bienias JL, et al. Alzheimer disease inthe US population. Prevalence estimates using the 2000census. Arch Neurol. 2003;60:1119-1122.

7. Larson EB, Shadlen MF, Wang L, et al. Survival after initialdiagnosis of Alzheimer disease. Ann Intern Med.2004;140:501-509.

8. Hoyert DL, Kung HC, Smith BL. Deaths: preliminary data for2003. Natl Vital Stat Rep. 2005;53:1-48.

9. Alzheimer’s Association Web site. Fact sheet: about genesand Alzheimer’s disease. Available at: www.alz.org.Accessed June 7, 2005.

10. Li G, Higdon R, Kukull WA, et al. Statin therapy and risk ofdementia in the elderly: a community-based prospectivecohort study. Neurology. 2004;63:1624-1628.

11. Zamrini E, McGwin G, Roseman JM. Association betweenstatin use and Alzheimer’s disease. Neuroepidemiology.2004;23:94-98.

12. Miller LJ, Chacko R. The role of cholesterol and statins inAlzheimer’s disease. Ann Pharmacother. 2004;38:91-98.

13. Doraiswamy PM, Steffens DC, McQuoid DR. Statin useand hippocampal volumes in elderly subjects at risk forAlzheimer’s disease: a pilot observational study. Am JAlzheimers Dis Other Demen. 2004;19:275-278.

14. Zandi PP, Sparks DL, Khachaturian AS, et al. Do statinsreduce risk of incident dementia and Alzheimer disease?The Cache County Study. Arch Gen Psychiatry.2005;62:217-224.

15. Breitner JC, Wyse BW, Anthony JC, et al. APOE-epsilon4count predicts age when prevalence of AD increases, thendeclines: the Cache County Study. Neurology.1999;53:321-331.

16. Rush University Medical Center Web site. The ReligiousOrders Study. Available at: http://www.rush.edu/rumc/page-1099611542043.html. Accessed June 8, 2005.

17. Aggarwal NT, Wilson RS, Beck TL, et al. The apolipopro-tein E epsilon4 allele and incident Alzheimer’s disease inpersons with mild cognitive impairment. Neurocase.2005;11:3-7.

18. Harwood DG, Barker WW, Ownby RL, et al. No associa-tion between subjective memory complaints andapolipoprotein E genotype in cognitively intact elderly. Int JGeriatr Psychiatry. 2004;19:1131-1139.

19. Harwood DG, Barker WW, Ownby RL, et al.Apolipoprotein E polymorphism and age of onset forAlzheimer’s disease in a bi-ethnic sample. Int Psychogeriatr.2004;16:317-326.

20. Hsiung GR, Sadovnick AD, Feldman H. Apolipoprotein E

epsilon4 genotype as a risk factor for cognitive decline anddementia: data from the Canadian Health Study of Healthand Aging. CMAJ. 2004;171:863-867.

21. Shumaker SA, Legault C, Rapp SR, et al. Estrogen plusprogestin and the incidence of dementia and mild cognitiveimpairment in postmenopausal women: the Women’sHealth Initiative Memory Study: a randomized controlledtrial. JAMA. 2003;289:2651-2662.

22. Espeland MA, Rapp SR, Shumaker SA, et al. Conjugatedequine estrogens and global cognitive function in post-menopausal women: Women’s Health Initiative MemoryStudy. JAMA. 2004;291:2959-2968.

23. Craig MC, Maki PM, Murph DG. The Women’s HealthInitiative Memory Study: findings and implications for treat-ment. Lancet Neurol. 2005;4:190-194.

24. Brinton RD. Impact of estrogen therapy on Alzheimer’s dis-ease: a fork in the road? CNS Drugs. 2004;18:405-422.

25. Henderson VW. Hormone therapy and Alzheimer’s disease:benefit or harm? Expert Opin Pharmacother. 2004;5:389-406.

26. Pinkerton JV, Henderson VW. Estrogen and cognition, witha focus on Alzheimer’s disease. Semin Reprod Med.2005;23:172-179.

27. Friedland RP, Fritsch T, Smyth KA, et al. Patients withAlzheimer’s disease have reduced activities in midlife com-pared with health control-group members. Proc Natl AcadSci U S A. 2001;98:3440-3445.

28. Lindstrom HA, Fritsch T, Petot G, et al. The relationshipsbetween television viewing in midlife and the developmentof Alzheimer’s disease in a case-control study. Brain Cogn.2005;58:157-165.

29. Hall KS, Gao S, Unverzagt FW, Hendrie HC. Low educa-tion and childhood rural residence: risk for Alzheimer’s dis-ease in African Americans. Neurology. 2000;54:95-99.

30. Scarmeas N, Stern Y. Cognitive reserve and lifestyle. J ClinExp Neuropsychol. 2003;25:625-633.

31. Scarmeas N, Stern Y. Cognitive reserve: implications fordiagnosis and prevention of Alzheimer’s disease. CurrNeurol Neurosci Rep. 2004;4:374-380.

32. Alzheimer’s Association. African Americans and Alzheimer’sdisease: the silent epidemic. Available at: http://www.alz.org/Media/newsreleases/2003/AA_ALZ.pdf. AccessedJune 8, 2005.

33. Taylor DH Jr, Sloan FA, Doraiswamy PM. Marked increasein Alzheimer’s disease identified in Medicare claims recordsbetween 1991 and 1999. J Gerontol A Biol Sci Med Sci.2004;59:762-766.

34. Green RC, Cupples LA, Kurz A, et al. Depression as a riskfactor for Alzheimer disease: the MIRAGE Study. ArchNeurol. 2003;60:753-759.

35. Lindsay J, Laurin D, Verreault R, et al. Risk factors forAlzheimer’s disease: a prospective analysis from theCanadian Study of Health and Aging. Am J Epidemiol.2002;156:445-453.

36. Tyas SL, Manfreda J, Strain LA, Montgomery PR. Risk factorsfor Alzheimer’s disease: a population-based, longitudinalstudy in Manitoba, Canada. Int J Epidemiol.2001;30:590-597.

37. Seshadri S, Beiser A, Selhub J, et al. Plasma homocysteineas a risk factor for dementia and Alzheimer’s disease. N Engl J Med. 2002;346:476-483.

38. Hux MJ, O’Brien BJ, Iskedjian M, et al. Relation betweenseverity of Alzheimer’s disease and costs of caring. CMAJ.1998;159:457-465.

39. Lamb HM, Goa KL. Rivastigmine: a pharmacoeconomicreview of its use in Alzheimer’s disease.Pharmacoeconomics. 2001;19:303-318.

Advanced Studies in Pharmacy n 125

REVIEW

40. Murman DL, Colenda CC. The economic impact of neuropsy-chiatric symptoms in Alzheimer’s disease. Can drugs ease theburden? Pharmacoeconomics. 2005;23:227-242.

41. Ernst RL, Hay JW. The US economic and social costs ofAlzheimer’s disease revisited. Am J Public Health.1994;84:1261-1264.

42. Saving Lives, Saving Money: Dividends for AmericansInvesting in Alzheimer Research. A report from the LewinGroup, commissioned by the Alzheimer’s Association.Washington, DC; 2004.

43. Rice DP, Fox PJ, Max W, et al. The economic burden ofAlzheimer’s disease care. Health Aff. 1993;12:164-176.

44. O’Brien JA, Caro JJ. Alzheimer’s disease and other demen-tia in nursing homes: levels of management and cost. IntPsychogeriatr. 2001;13:347-358.

45. HCIA, Inc., Arther Andersen LLP; 1998 (as indicated by ref-erence 41).

46. Bloom BS, Chhatre S, Jayadevappa R. Cost effects of a spe-cialized care center for people with Alzheimer’s disease. Am JAlzheimers Dis Other Demen. 2004;19:226-232.

47. Koppel R. Alzheimer’s Disease: The Costs to US Businessesin 2002. Washington, DC: Alzheimer’s Association; 2002.

48. Losing a Million Minds: Confronting the Tragedy ofAlzheimer’s Disease and Other Dementias. US CongressOffice of Technology Assessment; US Government PrintingOffice; 1987:14.

49. Fox PJ, Kohatsu N, Max W, Arnsberger P. Estimating thecosts of caring for people with Alzheimer disease inCalifornia: 2000-2040. J Public Health Policy.2001;22:88-97.

50. Harrow BS, Mahoney DF, Mendelsohn AB, et al. Variationin cost of informal caregiving and formal-service use forpeople with Alzheimer’s disease. Am J Alzheimers Dis OtherDemen. 2004;19:299-308.

51. Wimo A. Cost effectiveness of cholinesterase inhibitors inthe treatment of Alzheimer’s disease: a review with method-ological considerations. Drugs Aging. 2004;21:279-295.

52. Sano M, Wilcock GK, van Baelen B, Kavanaugh S. The effectsof galantamine treatment on caregiver time in Alzheimer’s dis-ease. Int J Geriatr Psychiatry. 2003;18:942-950.

53. Wimo A, Winblad B, Stoffler A, et al. Resource utilisationand cost analysis of memantine in patients with moderate tosevere Alzheimer’s disease. Pharmacoeconomics.2003;31:327-340.

54. Khang P, Weintraub N, Espinoza RT. The use, benefits, andcosts of cholinesterase inhibitors for Alzheimer’s dementia inlong-term care: are the data relevant and available? J AmMed Dir Assoc. 2004;5:249-255.

55. Fillit H, Hill J. The economic benefits of acetylcholinesteraseinhibitors for patients with Alzheimer’s disease and associat-ed dementias. Alzheimer Dis Assoc Disord. 2004;18(suppl1):S24-29.

56. Lyketsos CG, Lopez O, Jones B, et al. Prevalence of neu-ropsychiatric symptoms in dementia and mild cognitiveimpairment: results from the cardiovascular health study.JAMA. 2002;288:1475-1483.

57. Lyketsos CG, Sheppard JM, Steinberg M, et al.Neuropsychiatric disturbance in Alzheimer’s disease clustersinto three groups: the Cache County study. Int J GeriatrPsychiatry. 2001;16:1043-1053.

58. Lyketsos CG, Steinberg M, Tschanz JT, et al. Mental andbehavioral disturbances in dementia: findings from theCache County Study on Memory in Aging. Am J Psychiatry.2000;157:708-714.

59. Beeri MS, Werner P, Davidson M, Noy S. The cost ofbehavioral and psychological symptoms of dementia (BPSD)in community dwelling Alzheimer’s disease patients. Int JGeriatr Psychiatry. 2002;17:403-408.

60. Murman DL, Chen Q, Powell MC, et al. The incrementaldirect costs associated with behavioral symptoms in AD.Neurology. 2002;59:1721-1729.

61. Levy ML, Cummings JL, Fairbanks LA, et al. Longitudinalassessment of symptoms of depression, agitation, and psy-chosis in 181 patients with Alzheimer’s disease. Am JPsychiatry. 1996;153:1438-1443.

62. Devanand DP, Jacobs DM, Tang MX, et al. The course ofpsychopathologic features in mild to moderate Alzheimerdisease. Arch Gen Psychiatry. 1997;54:257-263.