Recently the National Institutes of Health–funded Aging, Demographics, and Memory Study (ADAMS) published representative data of dementia prevalence in the United States (Langa et al., 2005; Plassman et al., 2007). One of ADAMS enrollment criteria was age of 70 years or older, and approximately a third of participants were older than 85 years. The study included a detailed cognitive battery (Langa et al., 2005). Diagnosis of dementia was based on Diagnostic and Statistical Manual of Mental Disorders (DSM)-III-R and DSM-IV criteria. Dementia prevalence increased from 5% among 71- to 79-year-olds to 37.4% in those 90 years and older (Plassman et al., 2007). In addition to older age, the authors reported that African American race and lower educational achievement independently contributed to increased odds of dementia syndrome of any etiology.

Diagnostic Criteria

In 2001, the Practice Parameters Subcommittee of the American Academy of Neurology (AAN) published a summary of evidence-based guidelines for early detection, diagnosis, and management of dementia (Knopman et al., 2001). These guidelines offered basic recommendations for diagnosis and both pharmacological and nonpharmacological management of dementia. The experts recommend the use of the American Psychiatric Association’s DSM criteria (the most recent being the DSM-IV) for establishing a diagnosis of dementia syndrome (Box 66.1). It should be noted, however, that despite the requirement for memory impairment in some forms of dementia, memory impairment is not an early symptom.

Box 66.1

DSM-IV Criteria for Dementia

Adapted with permission from American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental Disorders, fourth ed., American Psychiatric Association, Washington, DC.

Increased focus on the earliest stages of dementia has led to the recognition of a state called mild cognitive impairment (MCI). MCI was initially considered a transition state between normal aging and dementia of the Alzheimer type, but the concept quickly evolved to denote an intermediate state between normal cognitive aging and any type of dementia. Practice parameters on MCI from the Quality Standards Subcommittee of the AAN have been published (Petersen et al., 2001b). MCI is an increasingly important focus of research and clinical attention. Identifying early disease—especially Alzheimer disease (AD) and vascular cognitive impairment—and initiating therapeutic interventions to prevent or slow decline are the goals. However, the concept of MCI as a clinical entity has undergone refinement and multiple modifications (Dubois and Albert, 2004; Petersen et al., 2001a, 2006; Petersen and Morris, 2005; Winblad et al., 2004) and has most recently been challenged as an artificial construct (Dubois et al., 2007). MCI is now recognized as a nonspecific state that may progress to AD or a non-AD dementia, may remain stable, or may revert to normal cognition. Many studies now attempt to define MCI of the AD type as the harbinger of AD. Because the onset of functional decline is typically late in the course of the disease, the combination of characteristic cognitive features in the presence of positive disease biomarkers has been proposed as sufficient for the diagnosis of the disease in the early predementia stages. These new criteria were proposed for use in AD (Dubois et al., 2007), but it is very likely this idea will be adopted for other dementias in the near future.

General Approach to Dementia Diagnosis

The essential feature of dementia is the acquired and persistent compromise in multiple cognitive domains that is severe enough to interfere with everyday functioning (see Box 66.1). This definition stands in contrast to delirium or acute confusional states (ACSs), which are distinguished primarily by prominent deficits or fluctuations in attentional processing. Although dementia syndromes tend to be chronic, progressive, and irreversible, and ACSs tend to be acute to subacute, fluctuating, and reversible, these distinctions are more relative than absolute. Toxic, metabolic, or other systemic physiological disturbances are more likely to be reversible which is not the case for neurodegenerative or cerebrovascular dementia syndromes. On the other hand, dementia renders the patient more vulnerable to delirium, highlighting the need for a comprehensive evaluation of potentially reversible etiological disorders in the context of even well-established dementia syndrome. Careful evaluation of persons referred for dementia evaluation can identify treatable or reversible disorders in up to 20% of cases (Hejl et al., 2002).

Distinguishing cognitive impairment associated with an underlying disorder from potentially reversible cognitive symptoms (e.g., depression, metabolic encephalopathy) often demands ongoing evaluation, including appropriate diagnostic tests and sometimes a therapeutic challenge with an antidepressant drug. Abrupt onset of thinking changes in temporal relation to a psychological stressor, poor effort on cognitive testing (particularly with demanding tasks), and prominent neurovegetative signs such as insomnia and anorexia are characteristic of depression-associated cognitive impairment. Drug-induced cognitive impairment occurring as a result of anticholinergic agents, sedative-hypnotic drugs (e.g., benzodiazepines), or opiate analgesics commonly occurs in the elderly and may cause or contribute to cognitive symptomatology. Any temporal association of cognitive decline with initiation or titration of certain drugs should prompt taper and withdrawal of the offending agent. Multifactorial etiology is common in the elderly, and the search for potentially reversible or modifiable conditions should be done with vigor and vigilance.

History

A detailed comprehensive history of the presenting illness is the single most important part of the evaluation of cognitively impaired subjects. In general, the history should be obtained from both the patient and a reliable and knowledgeable informant, because cognitive deficits, lack of insight, and lapses in judgment actively interfere with accurate self-reporting. A comprehensive history should determine the initial manifestations, mode of onset, and course over time (Box 66.2).

Box 66.2 Historical Components of the Dementia Evaluation

Temporal Factors

Mode of onset (acute, subacute, insidious)

Course (static progressive, improvement over time, fluctuating)

Individual Factors

Cultural background

Educational level

Social/occupational demands (and changes incurred by symptoms)

Life circumstances (social, financial, occupational, living arrangements)

Premorbid personality characteristics

Hereditary Factors

Familial risk factors (stroke, hypertension, diabetes mellitus)

Genetic: family history suggesting autosomal dominant inheritance or multiple cases in family suggesting non–mutation-associated familial disease (e.g., apolipoprotein E4–associated Alzheimer disease)

Medical/Neurological Conditions

General medical conditions (hypothyroidism, hypertension, diabetes mellitus, heart disease)

Neurological conditions (transient ischemic attacks, strokes, seizures, syncope, head trauma)

Associated motor features (tremor, gait difficulties, speech/swallowing disturbance, ataxia)

Sleep disturbances (sleep apnea, insomnia, sleep-associated movement disorder)

Insidious onset and gradual progression over time are characteristic for degenerative disorders as the cause of dementia, whereas abrupt or discrete temporal onset raises suspicion for an associated neurological condition such as a stroke or a tumor, extenuating life circumstances, or medication change. Age, cultural background, educational level, and prevailing social and occupational demands modulate the clinical expression of dementia and its clinical significance. Recognizing changes from a previous level of functioning or altered behavioral patterns may help discern subtle diagnostically relevant clinical deficits. AD typically manifests with recent memory impairment combined with variable degrees of language, visuospatial, and executive dysfunction. Initial isolated disturbances in any of the latter cognitive domains, especially in the settings of preserved episodic memory or prominent initial changes in behavior and composure, warrant considering alternative diagnoses, such as cerebrovascular disease, mass lesion, or a focal neurodegenerative condition. Early parkinsonian features prompt an evaluation for dementia with Lewy bodies (DLB), Parkinson disease dementia (PDD) or another parkinsonian dementing disorder, vascular dementia, and normal-pressure hydrocephalus. Concurrent medical conditions, history of neurological insults, and primary disturbances of sleep are other factors that have to be accounted for as possible causes of dementia or concomitant factors that exacerbate its expression.

Cognitive Assessment

Standardized assessments such as the Mini-Mental State Examination (MMSE) (Folstein et al., 1975) provide a brief focused survey of cognitive domains most often affected in AD. The MMSE is a crude assessment that offers very limited assessment of memory and executive functions, however. The Montreal Cognitive Assessment (MoCA) is another short focused test that can be easily adopted in the busy private neurologist’s practice (Nasreddine et al., 2005). In addition to the cognitive domains tapped into by the MMSE, the MoCA also includes drawing a cube, the clock-drawing test, and a shortened version of Trails B, a test of working memory, set shifting, and complex attention, as well as a more detailed verbal memory test. MoCA may be better suited for detection of cognitive changes in the predementia stage of AD and other types of dementia (Gagnon et al., 2010; Hoops et al., 2009; Nazem et al., 2009; Smith et al., 2007; Videnovic et al., 2010). Additional tests for attention, language, praxis, executive and visuomotor functioning, and abstract thinking should be used selectively at the bedside or during an office visit to augment the initial crude cognitive screen (Table 66.1). Finally, for more refined cognitive assessment, patients should be referred for an evaluation by a trained neuropsychologist. Assessment by the latter is especially informative in borderline cases when the physician is uncertain of the presence of cognitive decline beyond what is expected by aging alone or under special circumstances such as limited education.

Table 66.1 Clinical Cognitive Assessment * (Normal Ranges)

Cognitive Feature	Means of Assessment
Working memory	Digit span forward (7 ± 2)
Complex attention (require manipulation of items in working memory)	Digit span backward (6 ± 2)Months in reverse order (15-20 seconds)
Orientation	Time, location, autobiographical data
Language	Confrontational naming (high- and low-frequency items)
	Verbal fluency (e.g., animals, grocery items) (18 ± 6/1 minute)
	Repetition (sentences of varying length)
	Comprehension (yes-or-no questions; performing multistep tasks)
	Reading aloud/comprehension
	Sentence writing (spontaneous, to dictation)
Visuospatial	Figure copying (two- and three-dimensional figures)
	Visual scene analysis (describe whole and individual parts)
	Line bisection (place “X” in center of horizontal line)
	Clock drawing (draw clock face with hands set to “10 after 11”)
Verbal recent memory	Word list (3-5 words): immediate/delayed/recognition recall
	Paragraph recall
Nonverbal recent memory	Figure copy recall
Remote memory	Historical events (family milestones, recent presidents)
Abstract conceptualization	Similarities (rose/tulip, poem/statue); proverb interpretation
Praxis	Have patient demonstrate saluting a flag, hammering a nail
Sequencing	Graphomotor sequencing (have patient draw a simple alternating pattern)
	Luria gestures (sequentially alternating fist, side, palm)

* Further details regarding administration and interpretation can be found in Hodges, J.R., 1994. Cognitive Assessment for Clinicians. Oxford University Press, Oxford.

Neuropsychiatric Assessment

Neuropsychiatric symptoms are exceedingly common among the cognitively impaired. Mood state (depressed, euphoric), vegetative status (eating, sleeping), changes in personality (apathetic, disinhibited), and alterations in perception (hallucinations) or thought (delusions) are major areas to be probed in the course of diagnostic assessment. The presence or absence of certain neuropsychiatric symptoms may guide the physician in the differential diagnosis. For instance, while apathy is exceedingly common among all neurodegenerative dementias, euphoria is only rarely seen in AD or DLB but common in frontal-variant frontotemporal dementia (fvFTD). Early hallucinations, delusions, and fluctuations, which can also predate the onset of cognitive decline, point to DLB. Attention to these and other neuropsychiatric concomitants of dementia is important because they represent treatable components of excess morbidity and often have a profound impact on caregiver burden and the need for institutionalization (Kaufer et al., 1998).

Differentiating between primary neurodegenerative disorder with concomitant depressive symptoms and primary depressive disorder that can result in pseudodementia can be challenging. Good history taking is essential and sometimes a trial of antidepressants is warranted. It is also very important to distinguish apathy (loss of interest or motivation) from primary depression on the basis of neutral affect and the absence of vegetative signs.

Although structured interviews for neuropsychiatric assessment in dementia have been developed, their use is restricted primarily to research settings. Recently a brief questionnaire version of the Neuropsychiatric Inventory (NPI) (Cummings, 1997) was designed for the practice setting. The Neuropsychiatric Inventory Questionnaire (NPIQ) (Kaufer et al., 2000) is relatively easy to administer and can provide a clinical screening examination for common neuropsychiatric manifestations of dementia and their associated impact on caregivers.

Laboratory Studies

An electroencephalogram (EEG) has not proved useful in the routine evaluation of AD (Knopman et al., 2001).

Cerebrospinal fluid (CSF) studies are indicated when metastatic cancer, suspected central nervous system (CNS) infection, reactive serum syphilis serology, hydrocephalus, age younger than 55 years, rapidly progressive or unusual dementia, immunosuppression, or suspected CNS vasculitis are present. CSF opening pressure, cell count, and protein and glucose concentrations are normal in neurodegenerative dementias. Other CSF tests could be considered in diagnostically challenging cases where a diagnosis of AD is possible yet not certain; these include CSF β-amyloid and tau or phospho-tau.

Neuroimaging

Based on evidence-based methodology, the AAN guidelines for diagnostic evaluation of dementia (Knopman et al., 2001) recommended the use of neuroimaging to screen patients with cognitive impairment. These recommendations resulted from the findings of one class II study showing that 5% of all patients with cognitive complaints harbored a causative nondegenerative lesion such as a slow-growing brain neoplasm (most commonly of the frontal lobes), subdural hematoma (SDH), or normal-pressure hydrocephalus (NPH) (Chui and Zhang, 1997). Inasmuch as the vast majority of MCI patients harbor neurodegenerative pathology, it is sensible to extend the guidelines for diagnostic imaging to that patient cohort as well. Justification for a structural imaging evaluation of the cognitively impaired individual is twofold: it can help detect a potentially treatable disorder in need of urgent targeted intervention, and it can identify vascular ischemic comorbidity that could either be the source of or a contributing factor to cognitive decline. Magnetic resonance imaging (MRI) is preferred, but in those instances where MRI technology is not available or when an MRI is contraindicated (for instance in patients in pacemakers), computed tomography (CT) should be used.

Although several other neuroimaging techniques and analytic approaches are vigorously being developed as candidate diagnostic and prognostic biomarkers, none of them are yet ready for routine clinical use. The scientific advances based on these technologies will be discussed in detail in the specific disorders sections that follow.

Neurodegenerative Dementias

Alzheimer Disease

AD was originally described in 1907 by the German psychiatrist and neuropathologist, Alois Alzheimer. The index case was a 51-year-old woman with paranoid delusions, progressive memory impairment, and subsequent progressive aphasia. At autopsy, Alzheimer noted brain atrophy, and as he applied the newly available silver stains, he uncovered the senile plaques (now termed neuritic or amyloid plaques) consisting of dystrophic neurites clustered around what subsequently was revealed to be a central amyloid core. He also described a second deposition, the neurofibrillary tangle, consisting of intraneuronal staining in a fibrillar pattern. In honor of Alois Alzheimer, Kraepelin subsequently named the condition Alzheimer’s disease. What initially was considered a presenile condition was later appreciated to be far more common in the elderly after 65 years of age, and with the age distinction removed, AD became the most common neurodegenerative disorder and one of the most common diseases of the aging population. It is now listed as the fifth most common cause of death (Alzheimer’s Association, 2010).

Advancing knowledge of the clinical onset and progression, neuroimaging, genetics, neurochemistry and neuropathological cascades, and molecular biology of neuritic plaques and neurofibrillary tangles has set the stage for current and future therapeutic interventions. Autosomal dominant AD has been traced to three genetic mutations in the amyloid precursor protein (APP) gene and the presenilin-1 and presenilin-2 (PS1, PS2) genes. These cases typically affect the middle-aged population. The roles of genetic risk factors for the more common sporadic late-onset disease are the subject of intense research. Inheritance of the E4 allele of the apolipoprotein E (ApoE) gene remains the most established genetic risk factor for AD. Many other candidates are under study, consonant with advances in collecting and analyzing population samples and in human genome screening and gene identification. Improved understanding of the pathological mechanisms will be the basis for future therapies.

Diagnostic Criteria

The most commonly used and widely accepted criteria for dementia of the Alzheimer type are the National Institute of Aging and Stroke–Alzheimer’s Disease and Related Disorders Association (NINCDS-ADRDA) criteria (McKhann et al., 1984). While these criteria are very useful and well operationalized for the dementia state of AD, the accumulating evidence of a long prodromal (or latent) AD state has challenged our former understanding of AD and resulted in incredibly insightful research into the MCI state and more recently into prodromal AD. The AD field is now rapidly moving toward early and presymptomatic AD diagnosis that will rely heavily on disease biomarkers as outlined in the Dubois criteria for prodromal AD (Dubois et al., 2007) (Box 66.3). As such, AD has moved from being a “diagnosis of exclusion,” post documenting normal findings on a battery of laboratory studies, to a specific diagnosis based on clinical course and a characteristic pattern (or evolving pattern) of neuropsychological deficits in the presence of disease-associated biomarkers.

Box 66.3 Dubois Diagnostic Criteria for Alzheimer Disease

Probable Alzheimer Disease

Evidence satisfying the core diagnostic criteria (A) plus one or more supportive features (B-E)

Core diagnostic criteria:

A. Presence of an early and significant episodic memory impairment that includes the following features:

Gradual progressive change in memory function reported by patients or informants over more than 6 months

Objective evidence of significantly impaired episodic memory on testing

Episodic memory impairment can be isolated or associated with other cognitive changes at the onset of AD or as AD advances

Supportive features:

B. Presence of medial temporal lobe atrophy:

Volume loss of hippocampi, entorhinal cortex, and/or amygdala evidenced on MRI with qualitative ratings using visual scoring (referenced to well-characterized population with age norms) or quantitative volumetry of regions of interest (referenced to well-characterized population with age norms)

C. Abnormal cerebrospinal fluid biomarkers:

Low amyloid β_1-42, increased total tau or phospho-tau concentrations, or combinations of the three

D. Specific pattern on functional neuroimaging with PET:

Reduced glucose metabolism in bilateral temporoparietal regions

Other well-validated ligands, including those that foreseeably will emerge, such as Pittsburg Compound B or FDDNP

E. Proven autosomal dominant AD mutation within the immediate family

Criteria for Definite Alzheimer Disease

Both clinical and histopathological (brain biopsy or autopsy) evidence of the disease

Both clinical and genetic evidence (mutation on chromosome 1, 14, or 21) of AD

AD, Alzheimer disease; FLAIR, fluid-attenuated inversion recovery; MRI, magnetic resonance imaging; PET, positron emission tomography.

Epidemiology

AD is the most common cause of dementia worldwide. ADAMS recently published nationally representative data of dementia prevalence in the United States (Langa et al., 2005; Plassman et al., 2007). Diagnosis of dementia of the Alzheimer type was based on the NINCDS-ADRDA criteria. Overall dementia of the Alzheimer type accounted for 70% of dementia cases, ranging from 47% among those aged 71 to 75 years to 80% in the age 90+ group. The ADAMS data revealed that as of 2002, there were 3.4 million dementia cases in the United States (95% confidence interval [CI], 2.8-4.0), of whom 2.4 million (95% CI, 1.9-2.9) were due to AD. In addition to older age, the authors reported that African American race, lower educational achievement, and ApoE E4 genotype independently contributed to increased odds of dementia syndrome of any etiology (Plassman et al., 2007). In a following publication (Plassman et al., 2008) the ADAMS investigators reported that an estimated 22.2% (or 5.4 million Americans) 71 years or older have cognitive impairment in the absence of overt dementia. Of these, 2 million subjects were reported to have prodromal AD, defined as cognitive impairment without dementia but with a pattern of clinical symptoms or performance on neuropsychological testing suggestive of prodromal AD and no other medical or neuropsychiatric condition present to preclude an eventual diagnosis of AD. The annualized rate of progression to dementia of the Alzheimer type in prodromal AD subjects was substantially higher than the rate of progression to dementia among all cognitively impaired (17%-20% per year in prodromal AD versus 12% in cognitive impairment but no dementia from any cause). Among those who progressed to dementia, 83% were diagnosed with dementia of the Alzheimer type, 16.7% were diagnosed with vascular dementia, and 0.4% with dementia of undetermined cause (Plassman et al., 2008).

Based on our discussion thus far, it should not be surprising that advancing age is the most powerful risk factor for the development of AD. Other risk factors include the presence of one or more of the following: ApoE E4 allele in a given individual, lower education level, family history of AD, and cardiovascular risk factors (Bendlin et al., 2010; Rocchi et al., 2009; Steptoe et al., 2011). Prevalence but not incidence appears to be higher in women than in men, suggesting that susceptibility is similar, but duration of survival is greater in women. Midlife hypertension, elevated homocysteine, and elevated fat in the diet (the latter perhaps in combination with presence of an ApoE E4 allele) also are reported risk factors (Bendlin et al., 2010).

Clinical Presentation

AD is a progressive disorder of recent episodic memory, language, visuospatial function, and executive function associated with high frequency of neurobehavioral abnormalities at some point in the course. Onset of AD usually is in late life. If onset age is younger than 45 years, an autosomal dominant pedigree usually is involved. Most commonly, a PS1 mutation is the cause. In elderly individuals presenting with mild memory loss, the diagnosis of prodromal AD could be somewhat challenging owing to the presence of comorbid conditions that may alter cognition (e.g., congestive heart failure, cerebrovascular disease) or if medications for such conditions have anticholinergic side effects. Serial follow-up evaluations may be necessary to establish decline and a characteristic pattern of dementia. Nonetheless, if memory is not a prominent early complaint, other potential causes of dementia should be considered. In unselected series, diagnostic accuracy has been relatively low when tested against autopsy diagnosis. Using established diagnostic criteria for AD such as those of the NINCDS/ADRDA in conjunction with a standardized evaluation, clinical diagnostic accuracy of about 90% can be achieved (Knopman et al., 2001), accuracy approaching 90% can be anticipated, and few if any “treatable causes” are missed.

Memory Loss

Memory impairment is a sine qua non in the diagnosis of AD. Although AD can manifest with neuropsychological deficits other than memory (e.g., aphasia, abnormal executive function, apathy or other personality change), memory tasks in general are the earliest deficits noted. Memory is a complex cognitive function involving several anatomical systems underlying different types of memory. Terminology for memory is varied, and definitions are not always interchangeable.

AD is classically associated with episodic memory impairment. Episodic memory is the memory store for personal experiences that occur in a particular spatial and temporal context. Unlike working memory (see later discussion), episodic memory does not have to be actively maintained (or “rehearsed”) for successful retrieval. Episodic memory is dependent on a number of neocortical structures, but the medial temporal lobes, particularly the hippocampus, play a central role. AD even in its earliest stages prominently impairs episodic memory, particularly for recent events. Recent episodic memory is assessed by delayed recall (i.e., remembering material over intervals longer than 30 seconds to a few minutes, during which an intervening distracting task is given to prevent active rehearsal). Other tests of episodic memory include information about orientation (e.g., date and location) and current events. Deficits in episodic memory can only be definitively established in the context of sufficiently preserved attention and concentration. For instance, distractibility and inattention can prevent the delirious patient from successfully encoding any incoming information, and he or she will fail a later recall test. Such deficits should not be attributed to AD-type pathology until the patient is retested in a more lucid state.

Early in the disease course of AD, distant episodic memories (i.e., memories before disease onset) tend to be spared. This somewhat paradoxical phenomenon may be accounted for by neocortical consolidation of episodic memories over time, which renders old memories less dependent on hippocampal function. Although detailed examination of remote memory may reveal some impairment, especially later in the disease course, the brunt of the early symptomatology lies in recent episodic memory.

Working memory is a limited-capacity storage system in which internal representations are actively held “on line.” Working memory can be divided into two systems: one responsible for maintenance or storage of information, and a second central executive system responsible for coordinating information processing and manipulation. Tasks that involve simple rehearsal of items, such as keeping a phone number “on line” (or forward digit span) and word and block span, appear to require both inferior prefrontal cortex for maintenance and more posterior cortical regions for storage. Working memory tasks that require manipulation of items, also known as complex attention tasks (e.g., digit span backwards, letter/number sequencing), require a more diffuse network including the dorsolateral prefrontal cortex. Mild impairment in these complex attention tasks may be noted relatively early in the course of AD, whereas the more simple working memory tasks tend not to be impaired until later in the disease course.

A third type of memory, semantic memory (Tulving, 1987, 1992), is defined as our knowledge of facts about the world and is not associated with a spatiotemporal context of the learning event. For example, while we know that a tiger has stripes and that Washington DC is the capital of the United States, we do not recall the context in which we learned these pieces of information. Although episodic memory impairment is more prominent in AD, semantic memory is also impaired relatively early in the disease course. Some of the language impairment seen in AD relates to disintegration of semantic memory. A common test of semantic memory is the category fluency test, when the patient is asked to name as many items as possible from a given category such as animals or vegetables. Semantic memory can be reduced in the early stages of AD.

Finally, procedural memory, which depends on the implicit learning of tasks and actions, is usually preserved in AD. AD patients are capable of procedural or implicit learning in the absence of any apparent parallel declarative learning.

Apathy can be found in 42% of those with mild, 80% of those with moderate, and 92% of those with advanced AD (Mega et al., 1996). It presents with loss of interest in previously enjoyed activities (e.g., hobbies, social outings, spending time with beloved relatives), aloofness, diminished spontaneity and emotional behavior, and reduced motivation. It is thought to reflect disruption of the connections within the frontosubcortical–anterior cingulate circuitry and their connections with other cortical regions. Apathy and depression commonly co-occur, but they are not synonymous with each other (Cummings, 2003). Apathy can be distinguished from primary depression on the basis of its neutral affect and the absence of vegetative signs.

Depression is very common in AD, occurring in 10% of mild, 40% to 60% of moderate, and 60% or more of severe AD patients (Mega et al., 1996). The symptoms are rarely severe enough to merit diagnosis of major depressive disorder; more often they represent minor depression/dysphoria. Risk factors for developing depression are familial or personal history of depressive disorder, female gender, and younger age (Lyketsos and Olin, 2002).

Anxiety is another early feature of AD. In the early stages, anxiety may be a manifestation of the patient’s subjective awareness of his/her cognitive decline, his or her increased dependency on others, and fear of the disease and its progression. In the moderate stages, anxiety over abandonment and fear of being left alone are common. Changes in the daily routine and the environment can trigger anxiety in the demented patient and could easily escalate to agitation and aggression.

Agitation and irritability frequently co-occur. Agitation is more common in males, those with later onset of dementia, and those of more advanced age. It encompasses disruptive, aggressive, and/or resistive behaviors and is related to changes in frontal cortex on functional imaging studies and postmortem examination (Cummings, 2003). Common sources of resentment and irritability are the patient’s inability to successfully accomplish tasks that were accomplished with ease in the past, or a feeling of being mistreated or ignored.

AD patients may show a whole host of psychotic features such as hallucinations, delusions, or delusional misidentifications. These typically occur in the moderate to severe stages in AD, whereas in DLB they can occur early on and even be the first manifestation of the disorder. Most hallucinations are in the visual modality. Delusions also tend to favor the later stages of the disorder and occur in 30% to 50% of patients. Most common are delusions of infidelity, theft, and paranoia. Delusions often co-occur with aggression, anxiety, and aberrant motor behavior (Cummings, 2003).

Attention to the neuropsychiatric features of dementia is important because they represent treatable components of excess morbidity and often have a profound impact on caregiver burden and prompt institutionalization (Kaufer et al., 1998).

Neurological Examination

In the majority of cases, findings on the neurological examination are normal, outside of the mental status examination. Early on, mild abnormalities of tone (gegenhalten or, more commonly, mitgehen) may be present. Pathological reflexes such as grasp, root, and suck reflexes usually emerge in the overt dementia stages. At end stage, the patient is mute, incontinent, and bedridden, with flexion deformities of the limbs and impaired swallowing. Weight loss, which may begin in the midstages, is common at this stage. The presence of parkinsonian symptoms early in the course (within a year of onset of cognitive problems) suggests the presence of DLB. Usually the parkinsonism is of the akinetic rigid type without tremor. More severe abnormalities of gait and tone in the presence of cognitive abnormalities direct the diagnosis to other parkinsonian dementias.

Laboratory Studies

The evidence-based guidelines of the AAN for diagnostic evaluation of dementia (Knopman et al., 2001) recommend routine screening for vitamin B₁₂ deficiency and hypothyroidism. Other blood tests, such as screening for syphilis, are also justifiable if a clinical suspicion for neurosyphilis is present, either because of high-risk behavior or because of location in an endemic region.

Genetic testing for the ApoE genotype is not recommended on a routine basis. A large multicenter study demonstrated that the presence of the ApoE E4 allele increased the positive predictive value of diagnosing AD by only 4% over diagnoses made on clinical grounds alone (90% versus 94%) (Mayeux et al., 1998).

The AAN guidelines noted that CSF tests for β-amyloid, tau, and neuropil thread protein (AD7C-NTP) gave insufficient data, demonstrating values above and beyond the relatively high sensitivity and specificity of the clinical diagnosis of AD (Knopman et al., 2001). New studies assessing the diagnostic or predictive capabilities of β-amyloid and tau or phosphorylated tau (phosphotau) in the CSF, however, suggest the possibility that they may have a role to play in difficult cases, confirmation of diagnosis, or prediction of development of AD in patients in the predementia stages (Mattsson et al., 2009). Low CSF β-amyloid (Aβ1-42) coupled with elevated tau or phosphotau increases the likelihood that the patient has AD-type pathology (Mattsson et al., 2009; Sunderland et al., 2003). However, it should also be kept in mind that some pathologically confirmed AD patients have shown normal CSF Aβ1-42 and tau levels prior to death (Brunnstrom et al., 2010). In the predementia stages, a CSF pattern suggestive of AD relates to a very high probability of developing AD in the next 5 years (Hansson et al., 2006).

Concentrations of CSF neurotransmitters, neuropeptides, amino acids, and trace elements are of no diagnostic value. Ubiquitin levels in CSF are increased in AD, but the levels are similar to those found in other neurodegenerative disorders. In blood, acute-phase proteins may be elevated but are not helpful in specific diagnosis.

Genetics

A family history of AD is a major risk factor. Familial AD (FAD) has two forms: early-onset autosomal dominant and late-onset familial AD. The latter seems to show a complex polygenic pattern but has increased frequency in families. The former is associated with mutations in one of three genes: the amyloid precursor protein (APP), and the presenilin 1 (PSEN1) and 2 (PSEN2) genes. Early-onset autosomal dominant families in the world have been identified with PS1, accounting for the vast majority of cases. PS1 likely encodes a protein from the γ-secretase complex, an enzyme complex responsible for the cleavage of the APP molecule to β-amyloid. Consideration of genetic testing for mutations requires appropriate pretest advice and counseling, because positive results have deterministic implications for the family. All three genetic mutations—in APP, PSEN1, and PSEN2—increase brain and blood levels of β-amyloid.

Some but not all late-onset FAD pedigrees are associated with inheritance of the E4 allele of ApoE. In addition to the late-onset familial cases, ApoE also contributes to sporadic disease. Increased amyloid load and earlier age at onset are related to ApoE4 gene dosage (Mayeux et al., 1998). The apoE4 genotype is, however, only a risk factor, being neither sufficient nor necessary for disease development, and this is why it is not used routinely in clinical evaluations.

The search in the human genome for other gene variants that may affect risk of AD is ongoing. Genetic variations (polymorphisms) generally have been sought in proteins or lipoproteins related to the pathogenesis of AD (e.g., amyloid metabolism, inflammation, oxidative stress). Numerous genetic risk factors have been proposed, including allelic variants in sortilin-related receptor L (SORL1), clusterin (CLU, also known as ApoJ), phosphatidylinositol binding clathrin assembly protein (PICALM), and many others. To date, only the ApoE4 gene has been consistently confirmed as a “risk gene.” New searches for risk genes in AD use genome scans to look for candidate regions, and loci on several chromosomes are under investigation.

Neuroimaging

The AAN dementia practice parameter guidelines state that at least one unenhanced CT or MRI scan should be performed in patients with cognitive decline to rule out unexpected structural lesions and also to provide information about potential silent vascular injury (Knopman et al., 2001). MRI, with its improved resolution, allows better quantification of cerebral structures and better discrimination of normal from mildly affected patients with AD than is possible with CT. Noninvasive neuroimaging has greatly aided the accurate diagnosis of AD; structural lesions such as tumors, hydrocephalus, subdural hemorrhage, and strokes are identified easily. MRI and to a lesser extent CT aid in identification of vascular lesions that may be primary causes of dementia or contributory to cognitive decline in cases of mixed AD/vascular dementia.

Mesial temporal atrophy including the entorhinal cortex, hippocampus, and amygdala are considered typical for the prodromal AD stages (Apostolova et al., 2006, 2009; Jack et al., 2004) (Fig. 66.1). In the dementia stage, global brain atrophy—more striking in the temporoparietal than in the frontal regions—and ventricular enlargement are also pronounced (Apostolova et al., 2007; Thompson et al., 2003) (Fig. 66.2). A gradient-echo sequence on MRI could reveal cortico-subcortical microhemorrhages suggestive of the presence of vascular amyloidosis.

Fig. 66.1 Postmortem images (7T) of through corresponding slices of the hippocampal head (left column) and hippocampal body (right column) in a normal elderly person (NC) and an Alzheimer disease patient. Significant atrophy of the hippocampal structure and the entorhinal cortex (EC) is clearly evident.

(Courtesy Dr. L. Apostolova.)

Fig. 66.2 Brain atrophy in Alzheimer disease (AD) as seen on 1.5T T1-weighted MRI. In the prodromal stages (top row), only mild hippocampal atrophy, mild ventriculomegaly, but already noticeable parietal lobe atrophy is noted. In advanced AD (middle row), severe global brain atrophy is easily identified. Note significant atrophy of hippocampal formations on coronal view on the left. Finally, the bottom row images present a case with posterior cortical atrophy (visual variant of AD). Note highly preserved global cognitive performance comparable to that of the prodromal AD subject (Mini-Mental State Examination score = 29/30), preserved hippocampal formations, but striking posterior-predominant cortical atrophy as depicted on axial view on right.

New and advanced methodologies provide unique opportunities to study the earliest changes in the hippocampal structure. Some studies have documented subtle atrophy present as early as 3 years prior to MCI and 6 years prior to the dementia stages of AD (Apostolova et al., 2010b) (Fig. 66.3). Imaging biomarkers are presently being developed as diagnostic and prognostic biomarkers as well as surrogate biomarkers for clinical trials, with hippocampal atrophy, the most validated structural biomarker, already being accepted as a biomarker criterion for AD presence in the prodromal AD stages (Dubois et al., 2007).

Fig. 66.3 Three-dimensional hippocampal atrophy maps showing the amount of atrophy (in %) accumulated over 3-year period in cognitively normal elderly who remained cognitively normal for 6 years or longer since baseline (NL-NL), and cognitively normal elderly who were diagnosed with amnestic mild cognitive impairment (MCI) at 3 years and AD at 6 years (NL-MCI_AD).

Functional brain imaging with single-photon emission computed tomography (SPECT) and positron emission tomography (PET) can identify disease-specific patterns such as temporoparietal abnormalities in AD, frontal or anterior temporal abnormalities in FTD, and temporo-parieto-occipital abnormalities in DLB (O’Brien, 2007) (Fig. 66.4). SPECT findings of blood-flow abnormalities in a temporoparietal distribution may aid in confirmation of AD, especially in cases without significant atrophy. PET using [¹⁸F]fluorodeoxyglucose (FDG), a measure of energy utilization in the brain that predominantly marks synaptic activity, also shows temporoparietal deficits, is more sensitive than SPECT, and can confirm the diagnosis (see Fig. 66.4). Although validation studies of these methods generally have shown high sensitivity, relatively lower specificity poses the risk of false-positive diagnoses. To date, the sensitivity and specificity of PET have not been assessed outside of research clinics (Silverman, 2004; Silverman et al., 2003). It is nevertheless intriguing that similar although less marked deficits in FDG-PET in the cingulate, temporal, and parietal cortices have been demonstrated in presymptomatic persons homozygous for ApoE E4 (Reiman et al., 1996). Further work to more rigorously validate these methods is necessary before they can become more widely applicable (Cummings et al., 2007; Thal et al., 2006).

Fig. 66.4 Metabolic deficits in Alzheimer disease. [¹⁸F]-Fluorodeoxyglucose positron emission tomography (FDG-PET) scan showing typical biparietal (arrows, left image) and bitemporal (arrows, right image) hypometabolism, with characteristic sparing of sensorimotor cortex.

(Courtesy Dr. C. Meltzer.)

The recent development of PET ligands for imaging proteins contained in amyloid plaques—Pittsburgh Compound B (PIB) (Klunk et al., 2004; Mathis et al., 2007)—or concomitantly labeling amyloid plaques and neurofibrillary tangles—[¹⁸F]-FDDNP (Small et al., 2006)—offers the prospect of a preclinical or early diagnostic biomarker for AD (Mintun et al., 2006). PIB uptake is seen in the same cortical regions that show diminished FDG-PET activity (Fig. 66.5), but PIB imaging offers the advantage of greater spatial resolution and greater effect sizes relative to conventional FDG-PET imaging (Ziolko et al., 2006).

Fig. 66.5 Structural, functional, and amyloid imaging of the brain in Alzheimer disease (AD). Columns represent individual subjects. First column consists of PiB-PET, MRI, and FDG-PET scans obtained in a 71-year-old normal control subject (MMSE score = 30). PiB retention is not seen, and MRI and FDG-PET findings are normal. Center column (AD-1) contains scans from a 68-year-old man with mild AD (MMSE score = 26). PiB retention is marked, and both MRI and FDG-PET findings are normal. Third column (AD-2) shows scans from a more affected 69-year-old patient with AD (MMSE score = 21). Findings include significant PiB retention, unremarkable MRI appearance, and characteristic deficits in glucose metabolism for AD on FDG-PET scan. All PET values are corrected for atrophy. Scale bars indicate standardized uptake values (SUVs) for FDG and PiB. Note that the dynamic range of PiB is twice that of FDG. FDG-PET, Fluorodeoxyglucose positron emission tomography; MMSE, Mini-Mental State Examination; MRI, magnetic resonance imaging; PiB-PET, Pittsburgh compound B positron emission tomography.

(Courtesy University of Pittsburgh Amyloid Imaging Group.)

Amyloid PET imaging shows a bimodal uptake distribution in MCI, with some subjects showing AD-like and others normal control (NC)-like retention patterns (Kemppainen et al., 2007; Pike et al., 2007). Longitudinal amyloid imaging studies in the MCI and NC states has revealed that high PIB retention conveys substantially higher risk for progression from MCI to AD (Okello et al., 2009) and for cognitive decline in NC (Morris et al., 2009). PIB retention in amyloid-rich regions has also been documented in postmortem specimens (Ikonomovic et al., 2008; Thompson et al., 2009). Although PIB binding shows the expected correlation with cognitive function cross-sectionally (Jack et al., 2008, 2009; Mormino et al., 2009; Pike et al., 2007; Tolboom et al., 2009), longitudinal PIB studies have surprisingly uncovered that PIB retention levels off in the symptomatic stages, suggesting its greater usefulness as a preclinical rather than an overt clinical biomarker (Engler et al., 2006).

Pathology

Brain atrophy with regional neuronal and synaptic loss and the presence of amyloid and neuritic plaques, and neurofibrillary tangles (NFTs) are the primary pathological features of AD. Deposition of amyloid in the cerebrovascular wall is always present but to a variable degree.

Amyloid protein isolated from plaque cores is predominantly a 40- to 42-amino-acid peptide (Aβ1-42 or Aβ1-40) derived from its larger precursor protein, APP. APP is a transmembrane protein expressed in both neural and non-neural tissue. The APP gene located on the long arm of chromosome 21 has been linked to autosomal dominant early-onset AD. Elevation of Aβ1-42 concentrations with increased synaptic activity suggests it plays a role in synaptic function (Cirrito et al., 2005). A number of metabolic pathways for the processing of APP are recognized, and a shift in metabolism to the β-peptide is thought to be fundamental to amyloid deposition (Selkoe and Schenk, 2003). In the primary metabolic pathway, α-secretase cleaves APP just above the surface of the membrane, resulting in a large-fragment, soluble APP (sAPPα). This cleavage occurs within the β-amyloid domain, precluding Aβ production (Fig. 66.6). The β-amyloid fragment, either Aβ1-42 or Aβ1-40, is cleaved by two different proteases, the β- and γ-secretases at the N- and C-termini, respectively. Monomers aggregate into an oligomeric structure that is neurotoxic. Fibrillization leads to insoluble amyloid that is less neurotoxic. Aβ1-42 has a greater propensity to fibril formation and is the predominant form in early plaques. Vascular amyloid is predominantly Aβ1-40. Posttranslational modification of the fragments may play a role in transition to neuritic plaques. In AD, an apparent partial shift to the amyloidogenic pathway may occur, or there may be impaired clearance, resulting in greater β-amyloid production and subsequent deposition of diffuse in the brain tissue. Basic research studies have shown that the monomeric Aβ1-42 can lower the threshold for neuron death in cell culture and induce oxidative stress. Thus, β- or γ-secretase inhibition is a proposed therapeutic strategy, as are immunotherapies or other strategies to decrease amyloid content in the brain (Fig. 66.7). Markers of inflammation such as α₁-antichymotrypsin and other proteases and peptidases are found in neuritic plaques (NP), as is ApoE. ApoE is thought to play a role in amyloid removal. The ApoE E4 allele appears to be less effective in aiding the removal of β-amyloid from the brain, and this may be the reason that patients with AD who are ApoE4 carriers have more amyloid deposition in their brains and earlier onset of disease than non-ApoE E4 carriers.

Fig. 66.6 Mechanism of amyloid metabolism. Primary metabolism of amyloid precursor protein (APP) is action of α-secretase, resulting in soluble APPα (sAPPα). Normally, small amounts of the peptides Aβ1-42 and Aβ1-40 (β-amyloid) also are produced by action of the amyloidogenic pathway that uses β-secretase for the initial cut and γ-secretase (thought to be presenilin-1 [PSEN1] or a complex containing PSEN1) for the second cut. PSEN2, a protease homologous to PSEN1 but encoded by a gene located on chromosome 1, also appears capable of making the second cut inside the membrane. In Alzheimer disease, there is increased β-amyloid (arrow indicating more), which may (as monomers) induce oxidative stress and lower the threshold for neuronal death. The Aβ fragments tend to aggregate in the presence of divalent cations such as Zn²⁺ and Cu²⁺ and form dimers, oligomers, and subsequently fibrils that deposit in the neuropil as the core substance of the neuritic plaques. The increase in β-amyloid may be due to decreased clearance of the peptide.

Fig. 66.7 Proposed therapeutic strategies for amyloid and amyloid plaque reduction. Secretase inhibitors (for both β-secretase and γ -secretase) or enhancers of α-secretase would diminish production of β-amyloid fragments. Production of anti–β-amyloid antibodies by immunization or infusion of anti–β-amyloid antibodies might remove β-amyloid by allowing phagocytosis by microglia or altering equilibrium with the level of β-amyloid in the periphery and causing egress of β-amyloid to serum to be bound. Other compounds without the ability to cross the blood-brain barrier might bind peripheral β-amyloid, alter equilibrium between brain and blood, and lead to movement of β-amyloid out of brain to peripheral circulation. In transgenic mice that carry the mutated human APP gene and exhibit increased levels of β-amyloid and amyloid plaque deposition, plaque deposition can be prevented or decreased by these methods.

NP, also called senile plaques, are found predominantly in the cerebral cortex, especially in the association areas and the hippocampus (Fig. 66.8). Amyloid deposition does not follow the topographic sequence for the development of NFTs; it is seen initially in the cortex, then in the hippocampus, and then in other regions (Arnold et al., 1991). NPs range in diameter between 25 and 200 mm and consist of abnormal nerve processes often referred to as dystrophic neuritis, processes of activated microglia and astrocytes, and a central core of β-amyloid (see Fig. 66.8, A). β-Amyloid is a 40- or 42-amino-acid peptide that is cleaved from the large APP molecule. Diffuse or immature plaques are not associated with dystrophic neurites or with fibrillar amyloid; immunostaining reveals the presence of the β-amyloid protein, predominantly Aβ-1-42. It is believed that diffuse plaques precede classical neuritic plaques and that progressive fibril formation by the monomeric Aβ-1-42 into oligomers leads to initiation of an inflammatory process that results in protease and peptidase deposition and neuropil destruction. Some authors refer to “burnt-out plaques” in which a prominent amyloid core is surrounded by a very thin rim of dystrophic neurites.

Fig. 66.8 Neuropathology of Alzheimer disease (AD). A, Classic neuritic plaque (arrows) identified on a hematoxylin-eosin-stained section. B, Neuritic plaque immunostained for Aβ40. C, Section from an AD patient’s hippocampus, immunostained for tau. Several neurons within the granule cell layer running diagonally through image show prominent tau immunoreactivity. Arrow points to a tau-immunoreactive senile plaque and arrowheads point to intraneuronal neurofibrillary tangles in endplate region of hippocampus. D, Large hippocampal pyramidal neuron with granulovacuolar degeneration (arrow). E, Torpedo-like eosinophilic Hirano body in hippocampal neuropil of AD patient (arrow). F, Cerebral amyloid angiopathy. Note intense staining of vessel wall with Aβ40 antibody; smooth muscle cells of arteriolar media are essentially replaced by fibrillar Aβ.

(Courtesy Dr. H. Vinters.)

The primary component of NFTs is the microtubule-associated protein, tau. Tau protein is a cytoskeletal protein found predominantly in axons. In AD, tau is abnormally hyperphosphorylated and extensively cross-linked, forming an insoluble intracellular deposit (see Fig. 66.8, B). This altered form has reduced binding to microtubules, thereby disrupting the cytoskeleton and leading to interference with neuronal metabolism and subsequently to neuron death. Tau has six isoforms. These isofoms result from alternative splicing of the gene; three of the isoforms have three microtubule binding domain repeats (3R tau), andthree of the isoforms have four repeated sequences (4R). Variations seen in Western blots of tau isolated from brain in AD and Pick disease, corticobasal degeneration, and hereditary tauopathy indicate differential deposition of isoforms (Wszolek et al., 2006). Ubiquitin is frequently found in association with tangles, probably representing an attempt by the neuron to degrade the abnormal protein. Deposition of NFT appears to begin in the transentorhinal and entorhinal cortex, spreading from there to the hippocampus followed by the temporal neocortex and beyond (Delacourte et al., 1999). There is early laminar predilection is initially for layer 2 and then layer 5 of the entorhinal cortex followed by the CA1 and subicular region of the hippocampus, and the pyramidal cells of neocortical layers 3 and 5. Ultrastructurally, NFTs consist of paired helical filaments with an individual filament diameter of 10 nm wound in a double helix, with a total diameter of 200 nm and a periodicity of 160 nm. NFTs are not exclusive to AD and can be found in other conditions such as dementia pugilistica, prion disease, and Kufs disease. Neuropil threads (curly fibers) are also found and represent paired helical filament–containing neurites. Their number usually parallels the severity of neurofibrillary tangle formation.

NP and NFT burden correlate with disease severity, with the correlation being stronger with NFTs. Still, cases with high neuropathological burden can be seen in the absence of dementia. Synaptic and neuronal loss shows the highest correlations with global cognitive impairment. However, to date, neuropathological criteria for AD continue to be based on plaque and tangle numbers.

Other histological AD features include granulovacuolar degeneration (see Fig. 66.8, C) and Hirano bodies (see Fig. 66.8, D). Granulovacuolar degeneration is typically seen in hippocampal and consists of clear, round, intracytoplasmic basophilic vesicles 4 to 5 mm in diameter that immunolabel for tubulin, ubiquitin, and neurofilament. Hirano bodies are eosinophilic rodlike structures located almost exclusively in the Sommer sector of the hippocampal pyramidal layer, adjacent to and occasionally in neuronal bodies. Hirano bodies contain actin and actin-associated proteins, tau, and a C-terminal fragment of APP. Mild cortical spongiosis that is distinct from the spongiform changes of CJD can be found in the association cortices, sparing the primary sensory and motor cortex in some cases.

Cell loss preferentially involves the large neurons of the deeper layers of the cortex. Dysfunction of the cholinergic system in early stages (before loss of the synthetic enzyme ChAT) is suggested by the loss of basal forebrain neurons and receptors for the cholinergic neurotrophin nerve growth factor, as well as by the sensitivity of patients with AD to anticholinergics. Studies of patients with end-stage AD in the 1970s and 1980s uncovered extensive neuron loss in the nucleus basalis of Meynert and other cholinergic basal forebrain nuclei. These structures are the origin of the cholinergic projection to cerebral cortex and other forebrain regions. This loss was found to associate with loss of choline acetyltransferase (ChAT) in the projection fields of these nuclei. More recent work, however, revealed that (1) the loss of ChAT does not occur until much later in AD’s course (DeKosky et al., 2002), and (2) that cholinergic cells in the cholinergic basal forebrain neurons are preserved until very late in the course (Mufson et al., 2002). Neuron loss occurs to a lesser extent within the locus ceruleus and the nucleus raphe. As these studies were performed in late-stage cases, it is possible that these neurons, too, are preserved longer than was earlier thought.

The reason for the selective damage to the cholinergic system, with loss of ChAT and acetylcholinesterase, is not known. A number of other subcortical ascending neurotransmitter projections are involved, in particular the noradrenergic and serotonergic systems. Changes in the glutamatergic system also occur. Glutamatergic receptors decline is especially prominent in the pyramidal neurons (in cortex and in entorhinal cortex), which bear the brunt of cell loss and NFT deposition. The discovery of the cholinergic deficit has led to the first rational treatment for AD. The report of glutamatergic dysfunction led to the other currently approved therapeutic strategy.

Treatment

Acetylcholinesterase Inhibitors

Acetylcholinesterase inhibitors (AChEIs) are the first line of pharmacological treatment in AD. Treatment of AD with AChEIs was judged to be an accepted standard of care in the AAN practice guidelines for dementia (Doody et al., 2001). The following three agents are now available in the United States, Europe, and many other parts of the world: donepezil (Aricept), rivastigmine (Exelon), and galantamine (Reminyl). All three medications are effective and as shown in double-blind placebo-controlled trials can potentially stimulate or stabilize performance. Initial studies were 3 to 6 months in duration; double-blind placebo-controlled trials of up to 1 year (donepezil) show maintenance of medication effects, and open-label extension studies suggest continued efficacy for several years. Sudden improvement in cognitive function occurs in a small percentage of subjects. Slowing of decline or stabilization over time, difficult to determine in a brief exposure to the drug, also occurs with these medications. Gastrointestinal cholinergic-mediated side effects are commonly seen with AChEIs. Side effects are more common during the dose escalation phase of treatment; emphasizing this to patients and caregivers will enhance compliance. Gastrointestinal effects can be reduced by giving these drugs after meals or slowing the dose escalation rate. Beneficial effects of these agents also have been shown on neurobehavioral symptoms, notably apathy and visual hallucinations, and on stabilizing activities of daily living functions or slowing their decline. If therapy with one agent appears ineffective or causes intolerable side effects, another AChEI should be tried. Because of the nature of the neurochemical action of these medications, withdrawal effects are possible on switching drugs, so prolonged washout periods should be avoided.

Glutamate Receptor Modulators

The abnormalities in glutamate metabolism and receptors noted postmortem in brains of patients who had AD led to hypotheses that glutamate metabolic dysregulation leads to excitotoxicity and neuronal death in AD (Rogawski and Wenk, 2003). Memantine, an N-methyl-d-aspartate (NMDA) antagonist that had been available in Germany since the 1980s, was assessed as solitary therapy in two pivotal trials in moderate to severe AD (MMSE score of 5 to 15) in the United States (Reisberg et al., 2003) or as an add-on to ongoing AChEI therapy (Tariot et al., 2004). Both studies showed a significant benefit at the end of the trial in cognition and function, and memantine was approved for use in moderate to severe AD. No specific side effects were identified in the pivotal trials; side effects were significantly more frequent than in the placebo groups.

Other Cognitive-Enhancing or Disease-Modifying Drugs

With the vast increase in knowledge of the disease mechanisms operative in AD, many medications aimed at slowing the disease process and ameliorating cognitive decline are under development or evaluation. These include drugs currently approved for other indications, new medications, and complementary or alternative medications. Most proposed or unregulated preparations have been suggested by virtue of their potential ability to interfere with the known pathobiological cascades in AD, notably oxidative stress and inflammation. Vitamin E, an antioxidant, in a dose of 1000 IU twice daily had some effect on delaying time to nursing home placement, decline from moderate to severe AD, or death (Sano et al., 1997). However, recent data suggesting that high doses of vitamin E may cause cardiac toxicity, coupled with the failure of high-dose vitamin E to slow conversion of MCI to AD (Petersen et al., 2005) have resulted in abandonment of vitamin E as a treatment. Other free radical inhibitors and antioxidants are used, but without support from prospective data.

Estrogen or estrogen-progesterone replacement therapy was shown in several retrospective studies to delay or prevent emergence of AD, but two studies, one large and one small, showed that they have no effect on disease progression (Henderson et al., 2000; Mulnard et al., 2000). The Women’s Health Initiative Memory Study (WHIMS) showed no effect of these agents on incidence of either MCI or AD; in fact, a slight increase in dementia incidence was noted (Shumaker et al., 2003).

Anti-inflammatory agents also have been tried in symptomatic AD; neither low-dose prednisone nor nonsteroidal antiinflammatory drugs (NSAIDs) have been shown to slow progression of symptoms.

The use of Ginkgo biloba as a treatment for AD has now been credibly disputed by two separate studies. A carefully performed double-blind placebo-controlled 6-month-long study of ginkgo showed no effect at 6 months (Schneider et al., 2000). Another randomized double-blind placebo-controlled 7-year-long clinical trial with 3069 community-dwelling participants aged 72 to 96 demonstrated no benefit of gingko on cognitive performance in normal cognition or MCI (Snitz et al., 2009).

Latrepirdine (Dimebon) attracted interest after showing potential beneficial effects in AD animal models, a phase II human trial, and a phase III double-blind trial with mild- to moderate-stage patients (Doody et al., 2008). A second phase III monotherapy trial showed no benefit. An ongoing study is assessing the utility of latrepirdine in patients receiving donepezil.

Medications for Behavioral and Neuropsychiatric Symptoms

The behavioral and neuropsychiatric symptoms of AD are treated in several ways. First, if possible, are nonpharmacological techniques. Disorientation may be reduced by providing a quiet, familiar environment with clear labels on doors and other objects throughout the house. Efficient lighting to the bathroom or other amenities is important to reduce confusion at night.

Acute confusional states also may be produced by intercurrent medical problems. Even minor urinary tract or other infections may lead to marked behavioral disruption. Any patient with AD presenting with sudden deterioration of cognition should undergo careful clinical examination and investigation for signs of medical or surgical disease; depending on dementia severity, the patient may not be able to articulate the problem.

Aggressive behavior may pose considerable management problems and should always invoke thorough clinical assessment to determine the etiology of the symptoms. Psychological intervention includes specific training for professional and family caregivers in dealing with aggression, the need for alertness to the signs of an impending crisis, and the use of positive and clear language to reassure and distract the patient.

Depressive symptoms should be treated with selective serotonin uptake inhibitors (SSRIs) because they have the least anticholinergic effects. They also may ease anxiety, irritability, or other nonspecific symptoms that may accompany depression. The SSRI, citalopram, may be useful for agitation.

Agitation or disruptive behavior may require a neuroleptic. The newer atypical antipsychotic medications (quetiapine, risperidone, olanzapine) have been recommended in low doses with careful titration, allowing time for the medications to exert their effects. The Clinical Antipsychotic Trials of Intervention Effectiveness (Schneider et al., 2006), however, indicated that within the limitations inherent in attempts to do a carefully controlled outpatient trial on agitation, side effects, and limited efficacy, the atypical antipsychotics did not show clear advantages over the older antipsychotic in the study. Traditional neuroleptics are more likely to produce extrapyramidal symptoms, which may worsen cognitive function. The newer antipsychotic agents are associated with increased mortality in the elderly and have had a “black box” warning label added by the U.S. Food and Drug Administration (FDA). The older antipsychotics, however, also carry an increased risk of death (Schneeweiss et al., 2007). Thus, judicious use of all of these medications, with frequent reassessment of the need for such a drug (at least every month or two), is appropriate.

Short-acting benzodiazepines frequently are used for sleep but should not be used indefinitely; trazodone, an antidepressant with sedative properties, frequently is used to promote sleep. The newer sleep medications have not been assessed in controlled trials for their effects and possible adverse effects in AD, so no recommendation for or against their use can be made at this time.

Future Therapies

Although efforts will continue to develop better cognitive enhancers and effective neurobehavioral medications with favorable side-effect profiles, significant attention now is directed at therapies that would directly interdict the pathological cascades in AD. Much discussion centers on whether NFTs or amyloid may be “primary” in disease causality. While both pathological cascades are the target of intense research interest and study, and it must be noted that the relationship between NPs and NFTs is unknown. Studies to disrupt NFT deposition center on preventing them, because the extensive cross-linking and hyperphosphorylation make removing them a difficult if not impossible task. Glycogen synthase kinase 3 (GSK3) is thought to be involved in the phosphorylation of tau, and targeting GSK3 or finding other ways to interfere with pathological tau phosphorylation is under study.

Interruption of the amyloid cascade also has a number of strategies, including β-secretase inhibitors, γ-secretase inhibitors or modifiers, or α-secretase enhancers (see Fig. 66.7). Compounds that would insert themselves on the polymerizing β-amyloid, the so-called plaque busters, conceivably could aid in slowing β-amyloid accumulation. Accelerated removal of β-amyloid once it is formed, either by antibody infusion (passive immunization) or active immunization to produce endogenous antibodies, is effective in transgenic mouse models. Initial active immunization trials in humans were interrupted by neuroinflammation and encephalitis in a subset of the subjects, probably due to T-cell sensitization. Autopsy studies in several immunized individuals from that study showed the expected neuroinflammatory changes (Orgogozo et al., 2003), as well as a remarkable amyloid clearance in several areas of the cortex, suggesting that—as in mice—the amyloid in humans can be removed (Masliah et al., 2005; Nicoll et al., 2003). Future studies will determine the feasibility of removing amyloid safely, whether it results in slowing of cognitive decline, and whether it has any effects on NFT formation.

Cerebral Amyloid Angiopathy

β-Amyloid deposits can be seen in the cerebral blood vessels of all patients with AD (see Fig. 66.8, E). Severity of involvement varies, ranging from the presence of small amounts of amyloid to major deposits that distort artery architecture, causing cortical microinfarcts and microaneurysm with or without cerebral hemorrhages. Familial cases of AD and cerebral amyloid angiopathy (CAA) are associated with multiple hemorrhages in the brain, spinal cord, and leptomeninges. Although these hemorrhages could be clinically silent, patients may present with progressive dementia and spastic paraparesis. MRI using a gradient-echo sequence can show multiple small hemorrhages in the brain and spinal cord. The most common forms of familial CAA without AD manifest with spontaneous lobar hemorrhages rather than dementia (i.e., Icelandic and Dutch cerebral amyloid angiopathy). Lobar hemorrhage in late life in nonfamilial cases also is suggestive of CAA (Table 66.2). This diagnosis is suspect when the hemorrhage is “lobar,” located at the cortical gray/white-matter junction, and not in the typical distribution of hypertensive hemorrhages (i.e., basal ganglia, brainstem, or cerebellum). AD can be accompanied by severe amyloid angiopathy.

Table 66.2 Boston Criteria for Diagnosis of Cerebral Amyloid Angiopathy–Related Hemorrhage

Definite CAA	Full postmortem examination demonstrating CAA
	Absence of other diagnostic lesions
Probable CAA with supporting pathology in specimen	Clinical data and histopathological specimen (evacuated hematoma or biopsy) demonstrating some degree of CAA
	Absence of other diagnostic lesions
Probable CAA	Clinical data and MRI or CT demonstrating multiple hemorrhages restricted to lobar, cortical, or corticosubcortical regions (cerebellar hemorrhage allowed) in individuals age ≥55 years
Possible CAA	Clinical data and MRI or CT demonstrating single lobar, cortical, or corticosubcortical region hemorrhage in individuals age ≥55 years

CAA, Cerebral amyloid angiopathy; CT, Computed tomography; MRI, magnetic resonance imaging.

Adapted from Knudsen, K.A., Rosand, J., Karluk, D., et al., 2001. Clinical diagnosis of cerebral amyloid angiopathy: validation of the Boston Criteria. Neurology 56, 537-539.

More recently, an inflammatory presentation of amyloid angiopathy has been described. This syndrome is associated with subacute cognitive decline, seizures, and extensive white-matter abnormalities and microhemorrhages on imaging studies. Histologically, this presentation was found to be associated with giant-cell inflammatory response. The CAA-related inflammation appears to be relatively responsive to corticosteroids and cyclophosphamide (Eng et al., 2004).