CHAPTER 19 Dementia

Christopher I. Wright, MD, PhD, Nhi-Ha Trinh, MD, MPH, Deborah Blacker, MD, ScD, William E. Falk, MD

• Dementia is a syndrome with multiple etiologies. It is characterized by a disabling decline in memory and impairment in at least one other higher cortical activity (e.g., with aphasia, apraxia, agnosia, or executive dysfunction).

• An evaluation of a suspected dementia must include a complete history; physical, neurological, and psychiatric examinations; evaluation of cognitive function; and appropriate laboratory testing.

• Alzheimer’s disease accounts for 50% to 70% of all dementias and afflicts at least 4.5 million Americans; however, other etiologies, both common (e.g., vascular dementia and Lewy body disease) and rare (e.g., frontotemporal dementia and progressive supranuclear palsy), must be considered during a comprehensive evaluation.

• Longevity is the greatest risk factor for Alzheimer’s disease, as well as most other dementing disorders, but additional factors, such as genetic propensity (particularly in the rare cases of early-onset Alzheimer’s disease) and vascular pathology, can contribute to its prevalence.

• Animal models, structural and functional imaging studies, and postmortem neuropathological assessments are adding to our understanding of many dementing disorders. With Alzheimer’s disease in particular, current palliative treatments and potential curative approaches are based on this expanding knowledge.

OVERVIEW

The Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV)¹ criteria for dementia require a decline in memory (see Key Points), as well as an impairment of at least one other domain of higher cognitive function (e.g., aphasia [a difficulty with any aspect of language]; apraxia [the impaired ability to perform motor tasks despite intact motor function]; agnosia [an impairment in object recognition despite intact sensory function]; or executive dysfunction [such as difficulty in planning, organizing, sequencing, or abstracting]). Several important qualifiers (i.e., the condition must represent a change from baseline, social or occupational function must be significantly interfered with, the impairment does not occur exclusively during an episode of delirium, or cannot be accounted for by another Axis I disorder, such as major depression) are included in the definition.¹

Many elderly adults complain of difficulties with their memory, often with learning new information or names, or finding words. In most circumstances, such lapses are normal. The term mild cognitive impairment (MCI) was coined to recognize an intermediate category between the normal cognitive losses that are associated with aging and with dementia. MCI is characterized by a notable decline in memory or other cognitive functions as compared with age-matched controls. MCI is common among the elderly, although estimates vary widely depending on the diagnostic criteria and assessment methods employed; some, but not all, individuals with MCI progress to dementia. The identification of those at risk for progression is an active area of research.^2,³

EPIDEMIOLOGY OF DEMENTIA

The most common type of dementia (accounting for 50% to 70%) is Alzheimer’s disease (AD). Among the neurodegenerative dementias, Lewy body dementia is the next most common, followed by frontotemporal dementia (FTD). Vascular dementia (formerly known as multiinfarct dementia) can have a number of different etiologies; it can exist separately from AD, but the two frequently co-occur. Dementias associated with Parkinson’s disease and Creutzfeld-Jacob disease (CJD) are much less common.⁴

THE ROLE OF AGE OF ONSET

The onset of dementia is most common in the seventies and eighties, and is quite rare before age 40. Both the incidence (the number of new cases per year in the population) and the prevalence (the fraction of the population that has the disorder) rise steeply with age. This general pattern has been observed both for dementia overall, and for AD in particular. Estimates of the prevalence of dementia vary to some extent depending on which diagnostic criteria are used. The incidence of dementia of almost all types increases with age such that it may affect 15% of all individuals over age 65, and up to 45% of those over age 80.⁵ However, the peak age of onset varies somewhat among the dementias, with FTD and vascular dementia tending to begin earlier (e.g., in the sixties) and AD somewhat later.

The economic burden of dementia is enormous, and it is expected to rise steeply with anticipated demographic shifts. A substantial fraction of the increased burden will fall on developing countries, as a larger fraction of the population survives to old age.

Because AD and many other neurodegenerative disorders have an insidious onset, the precise age of onset is indeterminate. However, the approximate age of onset (i.e., within 2 to 5 years) is critical from both the personal and the public health points of view. It may offer an important leverage on disease risk and morbidity. A sufficient delay in onset can be equivalent to prevention with a late-onset disorder. Later onset can also decrease the burden of disease by pushing it later in the life span. From the research perspective, age of onset shows a robust relationship to genetic risk factors, and may have value as an outcome in genetic and other causal models.

EVALUATION OF THE PATIENT WITH SUSPECTED DEMENTIA

The diagnostic criteria for dementia require that a number of confounding disorders have been ruled out, particularly delirium and depression. Of note, an underlying diagnosis of dementia may predispose a patient to delirium; thus, further evaluation of cognitive deficits is required once delirium has been treated successfully. Delirium may be differentiated from dementia in several important ways; the onset of delirium is typically acute or subacute, the course often has marked fluctuations, and level of consciousness and attention are usually impaired.

Depression also may be difficult to distinguish from dementia, particularly as the two diagnoses are often co-morbid. Further, depression alone may cause some cognitive impairment; in such cases it may precede AD by several years. Certain features (such as a more acute onset, poor motivation, prominent negativity, and a strong family history of a mood disorder) favor a diagnosis of depression.

Establishing a precise etiology of dementia (Table 19-1) whenever possible allows for more focused treatment and for an accurate assessment of prognosis. Although reversible causes will be found in less than 15% of new cases, a diagnosis may help a patient and his or her family to understand what the future holds for them and to make appropriate personal, medical, and financial plans.

Table 19-1 Etiologies of Dementia

Vascular

Stroke, chronic subdural hemorrhages, postanoxic injury, diffuse white matter disease

Infectious

Human immunodeficiency virus (HIV) infection, neurosyphilis, progressive multifocal leukoencephalopathy (PMLE), Creutzfeldt-Jakob disease (CJD), tuberculosis (TB), sarcoidosis, Whipple’s disease

Neoplastic

Primary versus metastatic carcinoma, paraneoplastic syndrome

Degenerative

Alzheimer’s disease (AD), frontotemporal dementia (FTD), dementia with Lewy bodies (DLB), Parkinson’s disease, progressive supranuclear palsy (PSP), multisystem degeneration, amyotrophic lateral sclerosis (ALS), corticobasal degeneration (CBD), multiple sclerosis (MS)

Inflammatory

Vasculitis

Endocrine

Hypothyroidism, adrenal insufficiency, Cushing’s syndrome, hypoparathyroidism/hyperparathyroidism, renal failure, liver failure

Metabolic

Thiamine deficiency (Wernicke’s encephalopathy), vitamin B₁₂ deficiency, inherited enzyme defects

Toxins

Chronic alcoholism, drugs/medication effects, heavy metals, dialysis dementia (aluminum)

Trauma

Dementia pugilistica

Other

Normal pressure hydrocephalus (NPH), obstructive hydrocephalus

History should never be obtained from the patient alone; family members or others who have observed the patient are essential informants for an accurate history. The patient often fails to report deficits, usually because the patient is unaware of them (anosognosia). Ideally, family members should be interviewed separately from the patient so that they can be as candid as possible.

When obtaining a history one should determine the nature of the initial presentation, the course of the illness, and the associated signs and symptoms (including those that are psychiatric or behavioral). These areas are important in determining disease etiology (e.g., recognizing the characteristic abrupt onset of vascular dementia as opposed to the subtle onset of AD).

Review of systems should be extensive and must include inquiry about gait and sleep disturbance, falls, sensory deficits, head trauma, and incontinence. Obtaining the full medical history may reveal risk factors for stroke or for other general medical or neurological causes of the cognitive difficulties. The psychiatric history may suggest co-morbid illness (such as depression or alcohol abuse), particularly if prior episodes of psychiatric illness are elicited.

Up to one-third of cases of cognitive impairment may be at least partially caused by medication effects; common offenders include anticholinergics, antihypertensives, various psychotropics, sedative-hypnotics, and narcotics. Any drug is suspect if its first prescription and onset of symptoms are temporally related.

Family history can be helpful in determining the type of dementia, as discussed later. A social and occupational history is useful when assessing premorbid intelligence and education, changes in the patient’s level of function, and any environmental risk factors.

A general physical examination is essential with particular focus paid to the cardiovascular system. Physical findings can also suggest endocrine, inflammatory, and infectious causes. A complete neurological examination (including assessment of cranial nerves, sensory and motor functions, deep tendon reflexes, and cerebellar function) may reveal focal findings that suggest vascular dementia or another degenerative disorder, such as Parkinson’s disease. Screening for acuity of vision and hearing is important because it may reveal losses that can masquerade as, or exacerbate, cognitive decline.

Psychiatric examination may reveal evidence of delirium, depression, or psychosis. On formal mental status testing, such as with the Folstein Mini-Mental State Examination (MMSE; see Table 2-8),⁶ and other cognitive tests (Table 19-2), documentation of particular findings, in addition to the overall score, allows cognitive functions to be followed over time.

Table 19-2 Supplemental Mental State Testing for Patients with Dementia

Area	Test
Memory	Recall name and address: “John Brown, 42 Market Street, Chicago”
	Recall three unusual words: “tulip, umbrella, fear”
Language	Naming parts: “lab coat: lapel, sleeve, cuff; watch: band, face, crystal”
	Complex commands: “Before pointing to the door, point to the ceiling”
	Word-list: “In one minute, name all the animals you can think of”
Praxis	“Show me how you would slice a loaf of bread”
	“Show me how you brush your teeth”
Visuospatial	“Draw a clock face with numbers, and mark the hands to ten after eleven”
Abstraction	“How is an apple like a banana?”
	“How is a canal different from a river?”
	Proverb interpretation

Laboratory evaluation should include a complete blood count (CBC) and levels of vitamin B₁₂ and folate, a sedimentation rate, electrolytes, glucose, homocysteine, blood urea nitrogen (BUN), and creatinine, as well as tests of thyroid function and liver function (see Chapter 3). Screening for cholesterol and triglyceride levels is also useful. Syphilis serology is often included in an initial panel, but evidence of active infection is very rare. A computed tomography (CT) scan of the brain, without contrast, can also be useful in identifying a subdural hematoma, hydrocephalus, stroke, or tumor.

Additional studies are indicated if the initial workup is uninformative, if a particular diagnosis is suspected, or if the presentation is atypical (Table 19-3). Such investigations are particularly important in young patients with rapid progression of dementia or an unusual presentation. Neuropsychological testing may be quite useful, and is essential in cases where a patient’s deficits are mild or difficult to characterize. Briefer screening tools, such as the Folstein MMSE, have poor sensitivity and specificity for dementia, particularly in highly educated or intelligent patients. Such tests also generally fail to assess executive function and praxis.

Table 19-3 Supplemental Laboratory Investigations

What	When	Why
Neuropsychological testing	Patient’s deficits are mild or difficult to characterize	The sensitivity of the MMSE for dementia is poor, particularly in highly educated or intelligent patients (who can compensate for deficits)
Lumbar puncture (including routine studies and cytology)	Known or suspected cancer, immunosuppression, suspected CNS infection or vasculitis, hydrocephalus by CT, rapid or atypical courses	Look for infection, elevated pressure, abnormal proteins
MRI with gadolinium; EEG	Any atypical findings on neurological examination; suspected toxic-metabolic encephalopathy, complex partial seizures, Creutzfeldt-Jakob disease (CJD)	More sensitive than CT for tumor, stroke Look for diffuse slowing (encephalopathy) vs. focal seizure activity
HIV testing	Risk factors or opportunistic infections	Up to 20% of patients with HIV infection develop dementia, although it is unusual for dementia to be the initial sign
Heavy metal screening, screening for Wilson’s disease or autoimmune disease	Suggested by history, physical examination, laboratory findings	May be reversible

CNS, Central nervous system; CT, computed tomography; EEG, electroencephalogram; MMSE, Folstein Mini-Mental State Examination; MRI, magnetic resonance imaging.

Other diagnostic tests can supplement initial assessment when indicated. Magnetic resonance imaging (MRI) of the brain is more sensitive for recent stroke and should be considered when focal findings are detected on the neurological examination; in addition, MRI scans can identify lesions not apparent on physical examination, so it should always be considered when the diagnosis is unclear. An electroencephalogram (EEG) may be used to identify toxic-metabolic encephalopathy, complex partial seizures, or CJD. Lumbar puncture (LP) may be indicated when cancer, infection of the central nervous system (CNS), hydrocephalus, or vasculitis is suspected. Testing for the human immunodeficiency virus (HIV) is indicated in a patient with appropriate risk factors, because up to 20% of patients with HIV infection develop dementia, although dementia is uncommon as an initial sign. Heavy metal screening, as well as tests for Lyme disease, Wilson’s disease, and autoimmune diseases, should be reserved for patients in whom these etiologies are suspected.

ALZHEIMER’S DISEASE

Brief Description

Alzheimer’s disease (AD) is a progressive, irreversible brain disorder that robs those who have it of memory and overall mental and physical function; it eventually leads to death.⁷

Epidemiology of Alzheimer’s Disease

AD is the most common cause of dementia, affecting at least 4.5 million Americans. The incidence of the disorder is steadily increasing as the population ages, and by 2050 the prevalence may reach 14 million cases in the United States alone.⁸ Currently, AD is the seventh leading cause of death for adults, and the cost of care for patients with AD is now approximately $100 billion per year.⁹ Female gender carries an increased risk, even when accounting for differences in longevity, but some of the difference is offset by greater risk of vascular dementias in men. Vascular risk factors (except for male gender), including diabetes, atherosclerosis, hypertension, and elevated cholesterol, all increase the risk of AD, as well as of vascular dementia, although the mechanisms are unclear. A history of head trauma also increases the risk of AD. Putative protective environmental factors include greater educational attainment (a finding that holds even when possible biases associated with screening procedures are taken into account). Several pharmacological agents, notably postmenopausal estrogen supplementation, nonsteroidal antiinflammatory drugs (NSAIDs), and antioxidants, are associated with a decreased risk in epidemiological studies, but clinical trials of these agents have thus far been disappointing. More preliminary evidence links statins and H₂ blockers to protection against AD. Mental and physical exercise have also been associated with prevention of AD, but there is a concern that these findings may be related to the impact of prodromal symptoms at baseline. However, as part of a comprehensive vascular risk reduction program, physical exercise has many benefits that may include dementia prevention.¹⁰

Pathophysiology

In his initial 1907 case, Alois Alzheimer identified abnormal nerve cells and fiber clusters in the cerebral cortex using then new silver-staining methods at autopsy. These findings, considered to be the hallmark neuropathological lesions of AD, are known as neurofibrillary tangles (NFTs) and neuritic plaques (NPs) (Figure 19-1). Beta-amyloid protein, present in soluble form in the brain but also a major component of plaques, is thought to play a central pathophysiological role in the disease, perhaps via direct neurotoxicity.¹¹ NFTs are found in neurons and are primarily composed of anomalous cytoskeletal proteins (such as hyperphosphorylated tau), which may also be toxic to neurons and contribute to AD pathophysiology.¹²^–¹⁵

Figure 19-1 Plaques and tangles in AD. NP, neuritic plaques; *, neurofibrillary tangles.

Genetics of Alzheimer’s Disease

Before trying to identify specific genes, epidemiological approaches are used to look for evidence that genetic factors are involved in disease etiology. For AD, multiple lines of evidence support a role for genetic factors. First-degree relatives of AD patients have a twofold to threefold increased risk of developing AD, and often at a similar age, which can be reassuring for middle-aged individuals who care for elderly parents with AD. Examination of AD families generally reveals no clear pattern of inheritance, although some rare families—typically with an early onset—have autosomal dominant inheritance, in which the disease is observed in approximately half the children in each generation, and an affected child typically has an affected parent. Twin studies are used to examine whether the increased risk in families is due to genetic factors. For AD, monozygotic twins (who share all their genes) typically show greater concordance than dizygotic twins (who share only half, like ordinary siblings). However, monozygotic twins do not uniformly share an AD diagnosis, and their age of onset may differ by 10 or more years.

In the context of genetic research, early-onset AD and late-onset AD are usually considered separately, divided by age of onset at 60 or sometimes 65 years. Early-onset AD is more likely to be familial; most autosomal dominant families have an early onset. However, late-onset AD also runs in families, and family and twin studies also support a role for genes in late-onset AD.¹⁶ These family and twin findings hold, despite controlling for the known late-onset AD gene APOE.

Four genes are recognized as being involved in AD. Three lead to early-onset AD in an autosomal dominant fashion. Although these genes have limited public health impact, they are devastating for affected families. These early-onset genes probably account for half of the cases of AD that occur before age 60. More critically, they have made a large contribution to our understanding of the pathophysiology of AD, and to the development of promising therapeutic strategies. Amyloid precursor protein (APP) on chromosome 21 was recognized first. To date, there are 26 mutations that affect 72 families. The age of onset for these APP mutations varies, and it is modified by APOE genotype. Next discovered was presenilin 1 (PSEN1) on chromosome 14. There are 156 PSEN1 mutations affecting 342 families to date; many have been found in a single family, often referred to as “genetically private.” PSEN1 mutations are associated with an age of onset in the forties and fifties; these mutations are not modified by the APOE genotype. Although overall quite rare, PSEN1 accounts for the great majority of autosomal dominant early-onset AD. PSEN1 mutations have also been observed in nonfamilial early-onset AD cases. The last early-onset gene is presenilin 2 (PSEN2) on chromosome 1. PSEN2 has 10 reported mutations affecting 18 families to date. The age of onset is quite variable, extending into the late-onset AD range, and it is modified by APOE genotype.¹⁷

The final AD gene is apolipoprotein E (APOE). Rather than a deterministic gene like the other three, APOE is a susceptibility gene that increases the risk for AD without causing the disease. APOE has three alleles, 2, 3, and 4, which have a complex relationship to risk for both AD and cardiovascular disease, with the 2 allele decreasing risk of both disorders and increasing longevity, and the 4 allele increasing risk and decreasing longevity. The effect of APOE-4 varies with age; it is most marked in the sixties, and falls substantially beyond age 80 or 90 years. APOE seems to act principally by modifying the age of onset, which is lowest in those with two copies of the risk allele, and intermediate in those with one. The APOE effect appears to be stronger in women and in Caucasians, which may relate to their lower risk of cardiovascular disease. Investigators believe that there are more AD genes to be discovered, and one analysis predicts 4 to 7 additional AD genes. However, these genes have been quite difficult to identify, and APOE remains the only documented late-onset AD gene.¹⁸

Patients frequently ask about their risk of AD based on their family history. Those with an autosomal dominant history are best referred for genetic counseling, ideally from an Alzheimer’s Disease Research Center or a local genetic counselor. Genetic testing is commercially available for PSEN1, which is likely to be involved when there is an autosomal dominant family history and the age of onset is 50 or lower. It can be used both for confirmation of diagnosis and for the prediction of disease onset, but there are complex logistical and ethical issues.¹⁹ Currently, genetic testing is only available for the remaining early-onset genes in research settings. Patients without such a history can be advised that there is an increased risk of AD in first-degree relatives. However, they should be made aware that this increase is modest and that age of onset tends to be correlated in families. Genetic testing for APOE can be used as an adjunct to diagnosis, but it contributes minimally. It is not recommended for the assessment of future risk because it lacks sufficient predictive value at the individual level. Many normal elderly carry an APOE-4 allele, and many AD patients do not.²⁰

Clinical Features and Diagnosis

The typical clinical profile of AD is one of progressive memory loss. Other common cognitive clinical features include impairment of language, visuospatial ability, and executive function. Patients may be unaware of their cognitive deficits, but this is not uniformly the case. There may be evidence of forgetting conversations, having difficulty with household finances, being disoriented to time and place, and misplacing items frequently. At least two domains of cognitive impairment, including progressive memory decline (that affects functional ability), are required for a clinical diagnosis of AD.²¹ In addition to its cognitive features, a number of neuropsychiatric symptoms are common in AD, even in its mildest clinical phases.²² In particular, irritability, apathy, and depression are common early features, with psychosis (delusions and hallucinations) occurring more frequently later in the course of the disease.

A definitive diagnosis of AD rests on postmortem findings of a specific distribution and number of its characteristic brain lesions (NFTs and NPs).²³ Detailed clinical assessments (by psychiatry, neurology, and neuropsychology) in combination with structural and functional neuroimaging methods have a high concordance rate with autopsy-proven disease. Structural neuroimaging studies (such as MRI or CT) typically show atrophy in the medial temporal lobes, as well as in the parietal convexities bilaterally (Figure 19-2). Functional imaging studies of resting brain function or blood flow (i.e., positron emission tomography [PET] or single-photon emission computed tomography [SPECT]) display parietotemporal hypoperfusion or hypoactivity (Figure 19-3).