Stahl's Essential Psychopharmacology: Neuroscientific Basis and Practical Applications, 4th Ed.

Chapter 13. Dementia and its treatment

   Causes, pathology, and clinical features of dementia

    Alzheimer’s disease: β-amyloid plaques and neurofibrillary tangles

    The amyloid cascade hypothesis

    ApoE and risk of Alzheimer’s disease

   Three stages of Alzheimer’s disease

    First stage of Alzheimer’s disease: preclinical (asymptomatic amyloidosis)

    Second stage of Alzheimer’s disease: mild cognitive impairment (MCI) (symptomatic, predementia stage of amyloidosis plus some neurodegeneration)

    Third and final stage of Alzheimer’s disease: dementia (amyloidosis with neurodegeneration plus cognitive decline)

   Targeting amyloid as a future disease-modifying treatment of Alzheimer’s disease

    Vaccines and immunotherapy

    Gamma-secretase inhibitors

    Beta-secretase inhibitors

   Targeting acetylcholine as a current symptomatic treatment of Alzheimer’s disease

    Acetylcholine and the pharmacologic basis of cholinesterase treatments for dementia

    The cholinergic deficiency hypothesis of amnesia in Alzheimer’s disease and other dementias

    Cholinesterase inhibitors

   Targeting glutamate

   Treatments for psychiatric and behavioral symptoms in dementia

   Other proposed targets for dementia


This chapter will provide a brief overview of the various causes of dementias and their pathologies, including the most recent diagnostic criteria and the emerging integration of biomarkers into clinical practice for Alzheimer’s disease. Full clinical descriptions and formal criteria for how to diagnose the numerous known dementias should be obtained by consulting standard reference sources. The discussion here will emphasize the links between various pathological mechanisms, brain circuits, and neurotransmitters and the various symptoms of dementia, with an emphasis on Alzheimer’s disease. The goal of this chapter is to acquaint the reader with ideas about the clinical and biological aspects of dementia and its currently approved treatments as well as new treatments that are on the horizon. The emphasis here is on the biological basis of symptoms of dementia and of their relief by psychopharmacologic agents, as well as on the mechanism of action of drugs that treat these symptoms. For details of doses, side effects, drug interactions, and other issues relevant to the prescribing of these drugs in clinical practice, the reader should consult standard drug handbooks (such as Stahl’s Essential Psychopharmacology: the Prescriber’s Guide).

Causes, pathology, and clinical features of dementia

Dementia consists of memory impairment (amnesia) plus deficits in either language (aphasia), motor function (apraxia), recognition (agnosia), or executive function such as working memory and problem solving. Personality changes can also be present, sometimes even before memory impairment begins. There are many causes of dementia (Tables 13-1 through 13-3), and the unique pathologies associated with some of the major dementias are listed in Table 13-1. Knowing the pathology does not mean that a treatment is available, as it is often not evident how to translate information about brain pathology into pharmacological treatments. The best hope currently is in the area of amyloid pathology, where new treatments under investigation are attempting to interfere with amyloid processing in Alzheimer’s disease, as will be discussed later in this chapter.

Table 13-1 Pathological features of selected degenerative dementias



Alzheimer’s disease

Amyloid/tau pathology

Dementia with Lewy bodies

Parkinson’s dementia

Multisystem atrophy

Alpha-synuclein pathology

Frontotemporal dementia

Progressive supranuclear palsy

Corticobasilar degeneration

Tau pathology

Huntington’s disease

Spinocerebellar ataxia

Trinucleotide repeat

Wilson’s disease (copper)

Hallervorden–Spatz disease (iron)


Metachromatic leukodystrophy


Creutzfeldt–Jakob disease

Variant Creutzfeldt–Jakob disease (bovine spongiform encephalopathy)

Gerstmann–Sträussler–Scheinker disease

Fatal familiar insomnia (thalamic dementia)

Prion-related dementias

Just because a patient develops memory disturbance does not mean it is Alzheimer’s disease (Table 13-2). Alzheimer’s dementia is perhaps the best-known and commonest dementia, but it is often the other symptoms associated with memory loss that help make the diagnosis clinically (Table 13-2). Just to complicate things, many patients have mixed types of dementia, particularly Alzheimer’s dementia plus dementia with Lewy bodies, or Alzheimer’s dementia plus vascular dementia (Figure 13-1). Such cases are complicated to diagnose clinically, and definitive diagnosis sometimes must await autopsy. Most dementias are really pathological diagnoses, not clinical diagnoses.

Table 13-2 Not all memory disturbance is Alzheimer’s disease: clinical features of selected degenerative dementias


Clinical features

Alzheimer’s disease

Memory deficit




Dementia with Lewy bodies

Memory deficit

Fluctuating attention

Extrapyramidal signs

Psychosis (hallucinations)

Frontotemporal dementia

Memory deficit

Speech/language disorders



Huntington’s disease

Memory deficit

Executive dysfunction


Creutzfeldt–Jakob disease

Memory deficit



Language disturbance

Figure 13-1. Mixed dementia. There are several types of dementia, of which Alzheimer’s disease is the most common. They are distinguished by their underlying pathologies. It is possible to have more than one dementia, and in fact many patients have both Alzheimer’s disease and either dementia with Lewy bodies or vascular dementia.

A wide variety of dementias are considered nondegenerative, and these are listed in Table 13-3. Many of these are treatable upon discovering the underlying cause, but others are not. Extensive clinical evaluation and laboratory testing must rule out these causes prior to concluding that a case of dementia is due to Alzheimer’s disease.

Table 13-3 Nondegenerative dementias


Multi-infarct dementia

Strategic single-infarct dementia

Small vessel disease

Watershed area hypoperfusion


HIV dementia


Whipple’s disease

Progressive multifocal leukoencephalopathy





Multiple sclerosis



Cushing’s syndrome

Adrenal insufficiency



Brain injuries



Chronic subdural hematoma

Vitamin deficiency

B12, B1, folate, niacin


Lupus erythematosus

Sjögren’s disease


Heavy metal (storage) disorders (arsenic, mercury, lead)

Industrial/environmental toxins (fertilizers, pesticides)


Chronic alcohol/drug abuse

Wernicke–Korsakoff syndrome

Marchiafava–Bignami disease

Organ failure

Hepatic encephalopathy

Uremic encephalopathy

Pulmonary insufficiency

Other causes

Dementia syndrome of depression

Normal pressure hydrocephalus

Nonconvulsive status epilepticus

Acute intermittent poyphyria

Alzheimer’s disease: β-amyloid plaques and neurofibrillary tangles

Without the introduction of disease-modifying treatments, Alzheimer’s disease is poised for an exponential increase throughout the world, with projections that it will quadruple over the next 40 years to affect 1 in every 85 people on earth: over 100 million people by 2050. Fortunately, new treatments are being designed to interfere with various known pathological processes, particularly the formation of amyloid plaques, in an attempt to halt or slow disease progression in Alzheimer’s disease before neurons are irretrievably lost. To understand the current diagnostic criteria for Alzheimer’s disease, how and why biomarkers are being integrated into the diagnosis of this disorder, and the rationale behind the hot pursuit of new therapeutics, it is necessary to understand how the two hallmarks of this disorder, amyloid plaques and neurofibrillary tangles, are thought to be formed in the brain in Alzheimer’s disease.

The amyloid cascade hypothesis

The leading contemporary theory for the biological basis of Alzheimer’s disease centers around the formation of toxic amyloid plaques from peptides due to the abnormal processing of amyloid precursor protein (APP) into toxic forms of Abeta (Aβ) peptides (Figures 13-2 through 13-9). Why do we make Aβ in the first place? Although this is not fully understood, nontoxic Aβ peptides have antioxidant properties, can chelate metal ions, regulate cholesterol transport, and may be involved in blood vessel repair, as a sealant at sites of injury or leakage, possibly protecting from acute brain injury. Hypothetically, Alzheimer’s disease is a disorder in which toxic Aβ peptides are formed, leading to deposition of amyloid plaque in the brain, with the ultimate destruction of neurons diffusely throughout the brain, somewhat analogous to how the abnormal deposition of cholesterol in blood vessels causes atherosclerosis.

Thus, Alzheimer’s disease may be essentially a problem of too much formation of Aβ amyloid-forming peptides, or too little removal of them. One idea is that neurons in some patients destined to have Alzheimer’s disease have abnormalities either in genes that code for a protein called amyloid precursor protein (APP), or in the enzymes that cut this precursor into smaller peptides, or in the mechanisms of removal of these peptides from the brain and from the body. APP is a transmembrane protein with the C-terminal inside the neuron and the N-terminal outside the neuron. One pathway for APP processing does not produce toxic peptides and involves the enzyme α-secretase (Figure 13-2). Alpha-secretase cuts APP close to the area where the protein comes out of the membrane, forming two peptides: a soluble fragment known as α-APP and a smaller 83-amino-acid peptide that remains embedded in the membrane until it is further cleaved by a second enzyme acting within the neuronal membrane, called γ-secretase (Figure 13-2). That enzyme produces two smaller peptides, p7 and p3, which are apparently not “amyloidogenic” and therefore not toxic (Figure 13-2).

Figure 13-2. Processing of amyloid precursor protein into soluble peptides. The way in which amyloid precursor protein (APP) is processed may help determine whether an individual develops Alzheimer’s disease or not. A nontoxic pathway for APP processing is shown here. APP is a transmembrane protein with the C-terminal inside the neuron and the N-terminal outside the neuron. The enzyme α-secretase cuts APP close to where it comes out of the membrane to form two peptides: α-APP, which is soluble, and an 83-amino-acid peptide that remains in the membrane. A second enzyme, γ-secretase, cuts the embedded peptide into two smaller peptides, p7 and p3, which are not “amyloidogenic” and thus are not toxic.

Another pathway for APP processing can produce toxic peptides that form amyloid plaques (i.e., “amyloidogenic” peptides). In this case a different enzyme, β-secretase, cuts APP a little bit further away from the area where APP comes out of the membrane, forming two peptides: a soluble fragment known as β-APP and a smaller 91-amino-acid peptide that remains embedded in the membrane until it is further cleaved by γ-secretase within the membrane (Figure 13-3). This releases Aβ peptides of 40, 42 or 43 amino acids that are “amyloidogenic,” especially Aβ42 (Figure 13-3).

Figure 13-3. Processing of amyloid precursor protein into Aβ peptides. The way in which amyloid precursor protein (APP) is processed may help determine whether an individual develops Alzheimer’s disease or not. A toxic pathway for APP processing is shown here. APP is a transmembrane protein with the C-terminal inside the neuron and the N-terminal outside the neuron. The enzyme β-secretase cuts APP at a spot outside the membrane to form two peptides: β-APP, which is soluble, and a 91-amino-acid peptide that remains in the membrane. Gamma-secretase then cuts the embedded peptide; this releases Aβ peptides of 40, 42, or 43 amino acids. These toxic (amyloidogenic) peptides form amyloid plaques.

In Alzheimer’s disease, genetic abnormalities may produce an altered APP that, when processed by this second pathway involving β-secretase, produces smaller peptides that are especially toxic. Individuals who do not get Alzheimer’s disease may produce peptides that are not very toxic, or may have highly efficient removal mechanisms that prevent neuronal toxicity from developing. The amyloid cascade hypothesis of Alzheimer’s disease therefore begins with an APP that is hypothetically genetically abnormal, or genetically or environmentally abnormal in the way it is processed, so that when it is cut into smaller peptide fragments too many toxic peptides are made, accumulate, and form neuron-destroying amyloid plaques, i.e., amyloidosis, and neurofibrillary tangles. Hypothetically, this process triggers a lethal chemical cascade that ultimately results in Alzheimer’s disease (Figures 13-3through 13-8).

Figure 13-4. Amyloid cascade hypothesis, part 1: increased production of Aβ42. One theory for the pathophysiology of Alzheimer’s disease is that there are genetic abnormalities in amyloid precursor protein (APP), so that when it is processed by the pathway involving β-secretase, it produces smaller, toxic peptides (especially Aβ42, as shown here).

Figure 13-5. Amyloid cascade hypothesis, part 2: Aβ42 oligomers form and interfere with synaptic function. Aβ42 peptides assemble together to form oligomers, which interfere with synaptic functioning and neurotransmitter actions but are not necessarily lethal to neurons.

Figure 13-6. Amyloid cascade hypothesis, part 3: formation of amyloid plaques causing inflammation. Aβ42 oligomers clump together along with other molecules to form amyloid plaques. These plaques can cause inflammatory responses, activation of microglia and astrocytes, and release of toxic chemicals such as cytokines and free radicals.

Figure 13-7. Amyloid cascade hypothesis, part 4: amyloid plaque induces formation of tangles. Amyloid plaques and the chemical events they cause activate kinases, cause phosphorylation of tau proteins, and convert microtubules into tangles within neurons.

Figure 13-8. Amyloid cascade hypothesis, part 5: neuronal dysfunction and loss. The effects of amyloid plaques and the build-up of neurofibrillary tangles can ultimately lead to neuronal dysfunction and death.

Specifically, abnormal genes or other influences cause the formation of an altered APP, or altered processing into too many toxic Aβ42 peptides (Figure 13-4). Next, the Aβ42 peptides form oligomers (a collection of a few copies of Aβ42 assembled together: Figure 13-5). These oligomers can interfere with synaptic functioning and neurotransmitter actions such as those of acetylcholine, but they are not necessarily lethal to the neurons at first. Eventually, Aβ42 oligomers form amyloid plaques, which are even larger clumps of Aβ42 peptides stuck together with a number of other molecules (Figure 13-6). A number of nasty biochemical events then occur, including inflammatory responses, activation of microglia and astrocytes, and release of toxic chemicals including cytokines and free radicals (Figure 13-6). These chemical events then hypothetically trigger the formation of neurofibrillary tangles within neurons by altering the activities of various kinases and phosphatases, causing hyperphosphorylation of tau proteins, and converting neuronal microtubules into tangles (Figure 13-7). Finally, widespread synaptic dysfunction from Aβ42 oligomers, neuronal dysfunction and death from formation of amyloid plaques outside of neurons and neurofibrillary tangles within neurons leads to diffuse neuronal death (Figure 13-8) and regional expansion of neuronal destruction in the cortex, causing the relentless progression of Alzheimer’s symptoms of amnesia, aphasia, agnosia, apraxia, and executive dysfunction. Some investigators believe that Alzheimer’s disease may spread from neuron to neuron, with pathological phosphorylated tau transported down axons, released at synapses and then taken up by neighboring cells. Pathological tau possibly then latches onto normal tau in the connected neurons, triggering the formation of new pathological mis-folded tau, from one affected neuron to the next.

Support for the amyloid cascade hypothesis comes from genetic studies of those relatively rare inherited autosomal dominant forms of Alzheimer’s disease. Sporadic (i.e., noninherited) cases account for the vast majority of Alzheimer’s disease cases, but inherited cases can provide clues for what is wrong in the usual sporadic cases of Alzheimer’s disease. Rare familial cases of Alzheimer’s disease have an early onset (i.e., before age 65) and have been linked to mutations in at least three different chromosomes: 21, 14, and 1. The mutation on chromosome 21 codes for a defect in APP, leading to increased deposition of β-amyloid. Recall that Down’s syndrome is also a disorder of this same chromosome (i.e., trisomy 21), and virtually all such persons develop Alzheimer’s disease if they live past age 50. A different mutation on chromosome 14 codes for an altered form of a protein called presenilin 1, a component of the γ-secretase enzyme complex. A third mutation, on chromosome 1, codes for an altered form of presenilin 2, a component of a different form of γ-secretase. It is not yet clear what if anything these three mutations in the rare familial cases tell us about the pathophysiology of the usual sporadic, nonfamilial, and late-onset cases of Alzheimer’s disease. However, they all point to abnormal processing of APP into amyloidogenic β-amyloid peptides as a cause for the dementia, consistent with the amyloid cascade hypothesis. Theoretically, different abnormalities in amyloid processing may occur in sporadic Alzheimer’s disease from those identified in inherited cases, and there may even be multiple abnormalities that could be responsible for sporadic Alzheimer’s disease as a final common pathway, but the evidence nevertheless implicates something in the amyloid cascade that goes wrong in Alzheimer’s disease. If so, this implies that preventing the formation of amyloidogenic peptides could prevent Alzheimer’s disease.

ApoE and risk of Alzheimer’s disease

A corollary to the amyloid cascade hypothesis is the possibility that something may be wrong with a protein that binds to amyloid peptides in order to remove them (Figure 13-9). This protein is called apolipoprotein E (ApoE). In the case of “good” ApoE, it binds to β-amyloid peptides and removes them, hypothetically preventing the formation of Alzheimer’s disease and dementia (Figure 13-9A). In the case of “bad” ApoE, a genetic abnormality in the formation of ApoE causes it to be ineffective in how it binds to β-amyloid peptides. This causes amyloid plaques to be formed and deposited around neurons, which goes on to damage neurons and cause Alzheimer’s disease (Figure 13-9B).

Figure 13-9. ApoE and Alzheimer’s disease. Another version of the amyloid cascade hypothesis is the possibility that something is wrong with the protein apolipoprotein E (ApoE). (A) Properly functioning (“good”) ApoE binds to β-amyloid and removes it, thus preventing development of Alzheimer’s disease and dementia. (B) An abnormality in DNA could lead to the formation of a defective or “bad” version of the ApoE protein, such that it cannot effectively bind to amyloid. This would prevent removal of amyloid, allowing it to accumulate and damage neurons, so that Alzheimer’s disease develops.

Genes coding for ApoE are associated with different risks for Alzheimer’s disease. There are three alleles (or variants) of this gene coding for this apolipoprotein called E2, E3, and E4, and everyone has two alleles. The E4 variant on chromosome 19 (“bad” ApoE) is linked to many cases of late-onset Alzheimer’s disease, the usual form of this illness. ApoE is associated with cholesterol transport and involved with other neuronal functions including repair, growth, and maintenance of myelin sheaths and cell membranes. Having one or two copies of E4 increases the risk of getting Alzheimer’s disease. In fact, some studies show that you have a 50–90% chance of developing Alzheimer’s disease by age 85 if you are an E4 homozygote (i.e., you have two copies of E4); a 45% chance if you are a heterozygote for E4, versus the risk in the general population at 20%. Alzheimer’s patients with the E4 gene also have more amyloid deposits and progress more rapidly to dementia than those without the E4 gene. The E2 variant may actually be somewhat protective.

Three stages of Alzheimer’s disease

The old way of diagnosing Alzheimer’s dementia was neurological and neuropsychological testing for the tentative clinical diagnosis of possible or probable Alzheimer’s disease and then postmortem evaluation for confirmation of an actual diagnosis of Alzheimer’s disease. In 2011, the diagnostic criteria were revised in two major ways: firstly, they expanded the notion of Alzheimer’s disease into three stages to reflect the current dynamic sequence model of structural and functional brain changes over time in elderly people who are first cognitively normal, then have mild cognitive changes, and finally develop Alzheimer’s disease (Figure 13-10). Secondly, the new diagnostic criteria have incorporated biomarkers. The five biomarkers in the new Alzheimer’s dementia criteria include both biomarkers of amyloidosis/amyloid accumulation and biomarkers of neurodegeneration (Table 13-4).

Figure 13-10. The three stages of Alzheimer’s disease. During stage I of Alzheimer’s disease (presymptomatic), cognition is intact despite elevated levels of brain amyloid as evidenced by both positive amyloid positron emission tomography (PET) and reduced levels of Aβ toxic peptides in cerebrospinal fluid (CSF). Clinical signs of cognitive impairment in the form of episodic memory deficits begin to manifest during stage II (mild cognitive impairment, MCI). The onset of clinical symptoms in stage II appears to be correlated with neurodegeneration, as evidenced by elevated CSF tau, brain glucose hypometabolism on fluorodeoxyglucose PET (FDG-PET) scans and volume loss in key brain regions on magnetic resonance imaging (MRI) scans. During stage III of Alzheimer’s disease (dementia), cognitive deficits can be severe. Currently, treatment of Alzheimer’s symptoms does not typically begin until stage III, long after the actual disease onset.

Table 13-4 Biomarkers integrated into the diagnostic criteria for Alzheimer’s disease

Biomarkers of amyloid accumulation

Biomarkers of neurodegeneration

Abnormal radioactive tracer retention on amyloid PET scans

Elevated CSF tau (total and phosphorylated tau)

Low CSF amyloid levels of Aβ42

Decreased FDG uptake on PET


Atrophy on structural magnetic resonance imaging

·        hippocampal atrophy

·        ventricular enlargement

·        cortical thinning

CSF, cerebrospinal fluid; FDG, fluorodeoxyglucose; PET, positron emission tomography.

First stage of Alzheimer’s disease: preclinical (asymptomatic amyloidosis)

It seems obvious that the dementia of Alzheimer’s disease does not occur as soon as the first amyloid plaque arrives in the brain. Early plaques in fact seem to be relatively asymptomatic, but somewhere along the way, sufficient accumulation of them seems to trigger the neurodegeneration, or at least is associated with the neurodegeneration, that leads to dementia. It is not clear whether amyloid is an epiphenomenon of the neurodegenerative process, or whether amyloid drives the neurodegenerative process. Amyloid biomarkers are certainly assisting in the diagnostic process for identifying early stages of Alzheimer’s disease. Therapeutics, discussed below, includes many agents in clinical testing that interfere with amyloid accumulation, hypothesizing that amyloid drives neurodegeneration, and that interfering with amyloidosis will halt or delay the progress of Alzheimer’s disease. However, the lack of convincing evidence of clinical benefit of this approach to date suggests that another process may be the culprit causing neurodegeneration while amyloid is accumulating.

Here we will discuss how amyloid imaging is enhancing the diagnostic accuracy of Alzheimer’s disease in its early stages. The first stage of Alzheimer’s disease is now considered to be preclinical and silent, but trouble is brewing (Figures 13-10 and 13-11). That trouble is the slow, relentless deposition of Aβ peptides into the brain rather than their elimination via the CSF, plasma, and liver. This presymptomatic stage can now be identified with biomarkers (Table 13-4Figures 13-10 and 13-11): for example, CSF levels of Aβ are low because Aβ is being deposited in the brain instead of leaving the brain. Furthermore, amyloidosis is detectable with PET scans at the presymptomatic stage using radioactive neuroimaging tracers that label amyloid plaques (Figure 13-11). Tracers bind to the fibrillar form of amyloid and thus label mature neuritic plaques which can be seen on PET scans after administering a radioactive chemical that binds to amyloid.

Figure 13-11. Amyloid PET imaging. Positron emission tomography (PET) using amyloid tracers can be used to detect the presence of amyloid during the progression of Alzheimer’s disease. In cognitively normal controls (A), amyloid PET imaging shows the absence of amyloid. Individuals who are cognitively normal but have moderate accumulation of amyloid (B) are likely in the presymptomatic first stage of Alzheimer’s disease. Although mild cognitive impairment (MCI) is often present in the prodromal second stage of Alzheimer’s disease, not all patients with MCI have brain amyloid deposition (C). In such cases, the clinical presence of cognitive impairments is likely attributable to a cause other than Alzheimer’s disease. Unfortunately, MCI is often a harbinger of impending Alzheimer’s dementia. In these cases (D), amyloid deposition accompanies cognitive impairments and both amyloid accumulation and clinical symptoms of MCI worsen as Alzheimer’s disease progresses (E). In the third and final stage of Alzheimer’s disease, when full-blown dementia is clinically evident, a large accumulation of brain amyloid can readily be seen (F).

Interestingly, amyloid is rarely detected in the brains of individuals under the age of 50, even those with the high-risk E4 genotype. Although most cognitively normal healthy elderly people show no evidence of amyloid deposition (13-11A), about a quarter of cognitively normal elderly controls are amyloid positive (Figure 13-11B), and are thus considered to have presymptomatic Alzheimer’s disease (Figure 13-10). About half of patients with MCI show no evidence of amyloid deposition (Figure 13-11C), but the other half do show either moderate (Figure 13-11D) or severe amyloid deposition (Figure 13-11E). Almost 100% of patients with clinically probable Alzheimer’s disease show heavy amyloid deposition (Figure 13-11F).

Thus, Aβ amyloid pathology is not specific for the dementia phase of Alzheimer’s disease, but may mean that the fuse is already lit in the presymptomatic phase. Serial amyloid scans show an annual increase of up to 4% of amyloid in patients with probable Alzheimer’s disease. Although amyloid can be seen in the presymptomatic stage of Alzheimer’s disease, by definition, clinical changes are not detectable at this stage, presumably because there is not much neurodegeneration yet. It is not the amyloid plaques per se, but the neurodegeneration with which they are later associated, that seems to correlate with the onset of symptoms in the second and third stages of Alzheimer’s disease.

Most worrisome for the eventual progression of presymptomatic Alzheimer’s disease to the MCI symptomatic stage of illness is that some studies suggest that Aβ deposition in the preclinical stage is already associated with some degree of gray-matter atrophy in the hippocampus and the posterior cingulate gyrus that can be demonstrated with structural MRI scanning (Figure 13-12). Cognitively normal elderly adults with the E4 genotype have greater hippocampus volume loss than do cognitively normal adults without E4. Furthermore, those cognitively normal elderly adults with the E4 genotype exhibit faster atrophy, so both elevated Aβ levels and the E4 genotype are associated with gray-matter atrophy in subjects even without cognitive impairment. Reliable atrophy and cortical thinning are identifiable in the hippocampus and in entorhinal, temporal, and parietal cortices in asymptomatic individuals nearly a decade before the onset of dementia.

Figure 13-12. Structural MRI. Although hippocampal atrophy (A) and ventricular enlargement (B) are seen with normal aging, the progression of this volume loss is significantly more rapid in patients with Alzheimer’s disease. The brains of patients with Alzheimer’s disease also show a progressive thinning of the cortex (C).

Now that biomarkers are clarifying this first stage of presymptomatic Alzheimer’s disease, it has come to be viewed as the leading edge of a continuum of a process of formation of plaques and tangles causing a relentless march towards dementia. Since some patients may develop amyloid plaques but do not progress to neurodegeneration or dementia, the notion of a preclinical stage is intended for research purposes at this time in order to sort out more reliably who is destined to progress (and thus possibly who to treat at this stage with anti-amyloid therapies on the horizon) and who is not. Early brain changes at the preclinical stage that herald clinical progression to Alzheimer’s dementia is consistent with the current amyloid cascade hypothesis, but at this point it is not proven for whom this is true, nor how to apply an individual patient’s biomarker results to make an accurate prognosis for that individual. A major research question remains as to whether there is a threshold level of Aβ deposition at which brain atrophy occurs, or whether there is another process occurring later and then in parallel that causes the brain atrophy. The exact definition of pathological levels of Aβ deposition (Figure 13-11) remains open, which is why this presymptomatic stage of Alzheimer’s disease with amyloidosis but no symptoms is a research diagnosis now.

Risk factors at this presymptomatic stage of illness that may hasten the pace or increase the likelihood of progressing to dementia include depression, type 2 diabetes, ApoE4 genotype, and vascular disease, particularly cerebral emboli. Some experts even wonder whether the effects of type 2 diabetes in the brain can be called “type 3 diabetes,” since toxic amyloid peptides are overexpressed in the fat cells of obese individuals, and in the brains of Alzheimer’s patients with dementia insulin concentrations, insulin-like growth factor, and insulin receptors are decreased by up to 80%. Since insulin modulates neurotransmitter release, tubulin activity, neuronal survival, and synaptic plasticity, losing your insulin from diabetes does not look like a good idea if you want to prevent Alzheimer’s dementia. Thus factors that prevent type 2 diabetes may also help prevent the progression of preclinical Alzheimer’s disease to dementia, but that is still a matter of investigation – although it does seem like a matter of common sense. Other common-sense factors that may promote healthy brain aging but are not yet proven to slow the progression of preclinical Alzheimer’s disease to dementia include:

·        a healthy diet

·        adequate sleep

·        daily exercise

·        active, socially integrated lifestyle

·        leisure activities

·        cognitive stimulation

·        optimized treatment of depression and other mental illnesses

·        meditation and other mindfulness strategies (e.g., yoga)

·        spiritual activities

·        controlling vascular risk factors (hypertension, diabetes, dyslipidemia, obesity)

Second stage of Alzheimer’s disease: mild cognitive impairment (MCI) (symptomatic, predementia stage of amyloidosis plus some neurodegeneration)

Patients with mild cognitive impairment (MCI) have mild cognitive symptoms but not dementia. Some call this stage of disease “predementia Alzheimer’s disease,” or “MCI due to Alzheimer’s disease” or even “prodromal Alzheimer’s disease.” However, this diagnosis of MCI does not yet mean that Alzheimer’s disease pathology has necessarily caused the symptoms or even that MCI patients will inevitably progress to dementia. In fact, there is great debate about what is MCI versus what is “normal aging.” Hopefully, the study of biomarkers will be able to settle this in the future. From a purely clinical perspective, over half of elderly residents living in the community complain of memory impairment. They have four common complaints: compared to their functioning of 5–10 years ago, they experience diminished ability (1) to remember names, (2) to find the correct word, (3) to remember where objects are located, and (4) to concentrate. When such complaints occur in the absence of overt dementia, depression, anxiety disorder, sleep/wake disorder, pain disorder, or ADHD (attention deficit hyperactivity disorder) it is called MCI. As already mentioned, the new diagnostic criteria, coupled with biomarkers, are attempting to make the distinction among those with normal aging, those with reversible conditions, and those with MCI destined to progress to the dementia stage of Alzheimer’s disease. On clinical grounds alone and without biomarkers, studies show that between 6% and 15% of MCI patients convert to a diagnosis of dementia every year; after 5 years about half meet the criteria for dementia; after 10 years or autopsy, up to 80% will prove to have Alzheimer’s disease. Also, MCI patients who have depression exhibit more neurodegeneration and brain atrophy than MCI patients without depression.

Biomarker studies seek to determine who among these MCI patients are destined to progress inevitably to Alzheimer’s dementia and who are the lucky ones with a benign and nonprogressive condition. Already foreseen is the need to identify the high-risk group of MCI patients in order to be able to treat them to prevent dementia rather than to treat them as we are doing now, after the brain has already degenerated and the third stage of Alzheimer’s disease has been reached. Also, there is the need to identify those with benign conditions or conditions other than those linked to neurodegeneration associated with amyloidosis, so that they are not needlessly treated with amyloid-targeted therapies which may be both expensive and have side effects.

Only a subset of MCI patients display measurable amyloidosis (Figure 13-11CD, and E) and only a small proportion (about 10%) of MCI patients without amyloid progress to dementia. However, in those MCI patients with both cognitive symptoms and amyloidosis, the assumption is that they have progressed beyond a presumed earlier state of presymptomatic and silent amyloidosis without damage to the brain (Table 13-4: positive amyloid PET scans and low CSF Aβ levels) to early neurodegeneration (Table 13-4: high CSF tau levels plus structural neuroimaging abnormalities) (Figure 13-10). About half of amyloid-positive MCI patients progress to dementia within a year, and 80% may progress to dementia within 3 years, with MCI patients having the E4 genotype progressing even more rapidly. Those MCI patients who convert to dementia have higher amyloid loads, and the utility of amyloid PET to identify Alzheimer’s pathology in the setting of clinical MCI is becoming increasingly convincing. Impairment in episodic memory (the ability to learn and retain new information) is the cognitive symptom most commonly seen in those MCI patients who eventually progress to Alzheimer’s dementia. Those MCI patients with the E4 genotype have acceleration of the time to progression from MCI to dementia, so that amyloid imaging, plus testing of episodic memory, plus genetic determination of E4 may currently be the most valuable way to predict higher risk of progression to dementia from the MCI stage.

However, as already mentioned, brain amyloidosis alone does not appear sufficient to produce cognitive decline. Rather, neurodegeneration must occur as the probable direct substrate of cognitive impairment, with the rate of cognitive decline being driven by the rate of neurodegeneration, not by the rate of amyloid deposition. Neurodegenerative atrophy on structural magnetic resonance imaging (MRI) scans both precedes and parallels cognitive decline. The new diagnostic criteria for Alzheimer’s disease suggest that it is important to determine whether an MCI patient with impairment of episodic memory has neurodegeneration (as well as amyloidosis and the E4 genotype), to help distinguish MCI that is destined to progress to dementia from nonprogressive MCI and normal aging. Thus, the MCI stage of Alzheimer’s disease by definition not only has amyloidosis and cognitive symptoms, but also biomarker evidence of neurodegeneration (Table 13-4Figures 13-10 through 13-13).

One biomarker for neurodegeneration is the presence of elevated cerebrospinal fluid (CSF) tau (including phospho-tau), thought to be associated with neuronal loss in the brains of Alzheimer’s disease patients because it is also elevated in other neurodegenerative diseases such as stroke and Creutzfeldt–Jakob disease (Table 13-4). Numerous neuroimaging biomarkers of neurodegeneration also are available, from single photon emission computerized tomography (SPECT) and fluorodeoxyglucose positron emission tomography (FDG-PET) scans (Figure 13-13), to structural magnetic resonance imaging (MRI) of hippocampal atrophy and of cortical thinning (Figure 13-12), to functional MRI and beyond.

Figure 13-13. Glucose hypometabolism in Alzheimer’s disease. The brains of normal, healthy controls show robust glucose metabolism throughout the brain using fluorodeoxyglucose positron emission tomography (FDG-PET). During the early, prodromal stage of Alzheimer’s disease, when mild cognitive impairment (MCI) is present, there is a reduction in brain glucose metabolism in more posterior brain regions such as temporoparietal cortex. As the disease progresses to full-blown Alzheimer’s dementia, brain glucose hypometabolism becomes evident on FDG-PET. The worsening of glucose metabolism with the progression of Alzheimer’s disease is believed to reflect accumulating neurodegeneration, especially in key brain areas such as temporoparietal cortices.

Neurodegeneration and MRI

MRI is more widely available than PET scanning and has numerous structural techniques including volumetric MRI, diffusion-weighted MRI, diffusion tensor imaging (DTI), and magnetization transfer ratio (MTR), and functional techniques such as perfusion MRI, arterial spin labeling (ASL), and fMRI (functional MRI). Some fMRI studies show decreased activation in the hippocampus during episodic memory tasks in dementia.

Among all the MRI techniques, volumetric MRI is generally the method of choice as a biomarker for staging Alzheimer’s disease, for measuring disease progression serially over time, and for clinical trials attempting to detect disease-modifying treatments (Figure 13-12). MRI machines are widely available and have good test/retest reliability, and there is a good correlation between MRI measures of atrophy and neuronal cell loss.

Hippocampal atrophy (Figure 13-12A) identifies patients with Alzheimer’s dementia, and atrophy in this region progresses more rapidly in patients with Alzheimer’s dementia (about 5% per year) compared to healthy elderly controls without cognitive symptoms (about 1.5% per year). Ventricular enlargement (Figure 13-12B) is about 1.3 cm3 per year in healthy elderly patients without cognitive problems, about 2.5 cm3 per year in MCI patients, and about 7.7 cm3 per year in those with Alzheimer’s dementia. Alzheimer’s disease is also associated with progressive cortical thinning (Figure 13-12C), reflecting loss of brain substance in the cortex, and this cortical thinning can distinguish between Alzheimer’s dementia and normal healthy elderly people without cognitive impairment.

The signature pattern of brain atrophy is medial temporal cortex (14% cortical thinning), temporal cortex (11%), parietal cortex (9.6%), and frontal cortex (7.8%) in mild Alzheimer’s dementia compared to normal healthy controls. This same topographical pattern is detected in patients with MCI and in normal elderly people with positive amyloid PET scans. Interestingly, autopsy studies of Alzheimer’s patients with whole-brain atrophy measured by MRI show that atrophy is related both to cognitive decline and to the amount of neurofibrillary tangles, but not to the amount of amyloid plaques. Since increased brain amyloid and decreased CSF Aβ levels are associated with hippocampal and other brain area atrophy in MCI by volumetric MRI in some but not all MCI patients, as previously mentioned, the working hypothesis is that those MCI patients with increased brain amyloid and decreased CSF Aβ levels (i.e., those who have amyloidosis) (Figure 13-11) and who also have brain atrophy on structural MRI scans (Figure 13-12, i.e., those with documented neurodegeneration) are defined not only as those who are at the MCI stage of Alzheimer’s disease (Figure 13-10), but also as those at the highest risk for progressing to the dementia stage of Alzheimer’s disease.

Neurodegeneration and FDG-PET

FDG-PET measures synaptic activity, so low amounts of FDG uptake, called hypometabolism, indicate synaptic dysfunction. SPECT imaging provides data similar to FDG-PET, if less spatial resolution, with evidence of reduced activity in temporoparietal cortex in Alzheimer’s dementia. Hypometabolism on FDG-PET is also seen mostly in temporoparietal region of the cortex (Figure 13-13), and the lower the metabolism in these areas on FDG-PET scans, the greater the amount of amyloid deposition seen in these same brain areas on amyloid PET scans in patients with Alzheimer’s dementia. However, FDG-PET is not abnormal in cognitively normal subjects with amyloidosis on amyloid PET scans, suggesting no neurodegeneration has yet occurred in these subjects. On the other hand, some studies show hypometabolism on FDG-PET scans in normal cognitively functioning elderly if they have the E4 genotype. The hypometabolic FDG-PET pattern is also seen in MCI subjects, with some studies predicting progression from MCI to dementia of 80–90% within 1–1.5 years, a rate that is even faster in subjects who have the E4 genotype. Amyloid PET scans and hippocampal MRI volume provide complementary information for diagnosis of dementia along with FDG-PET scans, and the best predictors of progression of MCI to dementia might include right entorhinal cortical thickness and right hippocampal volume. However, MCI patients who have the combination of gray-matter atrophy on MRI plus hypometabolism of posterior cingulate on FDG-PET also have a higher risk of progression to dementia than those MCI patients with either finding alone. There is a fourfold risk of conversion of MCI to dementia within 2 years if the patient has abnormal episodic memory alone, but there is a 12-fold risk if the patient is abnormal on both FDG and episodic memory in MCI. Thus, combinations of abnormal biomarkers in MCI enhance the odds that such a patient may progress to dementia. The findings suggest that it is really neurodegeneration and not amyloidosis that drives the onset of symptoms at the MCI stage of Alzheimer’s disease as well as the progression of symptoms from the MCI stage to the dementia stage of this illness.

When is depression a major depressive episode and when is it the MCI prodrome of Alzheimer’s dementia?

Depression can not only be mistaken for dementia, but it can also precede the onset of dementia and be associated with a twofold increase in the risk of developing cognitive impairment or dementia. When depression occurs in late life, whether it is a recurrent episode in a patient with a lifetime of episodes, or a first episode in late life, a major depressive episode can actually present with prominent cognitive symptoms, especially apathy, lack of interest, and slowing of information processing, rather than depressed mood and sadness. Depression with lack of interest or sadness can also occur in patients with established dementia, in patients whose depression may represent the MCI prodrome to dementia, and even in patients who ultimately prove to have reversible cognitive impairments from “pseudodementia” or the “dementia of depression.” It remains controversial whether depression reflects a causative factor for MCI or dementia, is part of MCI, or shares neuropathological features with the dementia stage of Alzheimer’s disease. Hopefully, studies of biomarkers in elderly patients with depression with or without the MCI prodrome will help sort this out. CSF Aβ peptides are low in cognitively intact elderly with major depressive disorder (MDD) similar to individuals with MCI or Alzheimer’s dementia. Other studies suggest that there are increased amyloid plaques in cognitively intact elderly patients with major depression (i.e., similar to the findings at the preclinical asymptomatic stage of Alzheimer’s disease), and that depressive symptoms in the symptomatic MCI stage without dementia predict greater brain atrophy in those patients with both depression and MCI. Some experts believe that depressive symptoms associated with MCI are an ominous combination, with depression being a prodromal manifestation of dementia. Thus, depression that begins in late life may possibly represent an Alzheimer’s disease symptomatic prodrome, whereas recurrent depression with another episode in late life may be related either to vascular dementia or to no dementia at all.

Third and final stage of Alzheimer’s disease: dementia (amyloidosis with neurodegeneration plus cognitive decline)

The final stage of Alzheimer’s disease is dementia, which applies to those who develop cognitive or behavioral problems that interfere with function at work or in everyday activities. Similar to the old guidelines, the new criteria classify patients into “probable” and “possible” Alzheimer’s dementia, with no change in those with probable Alzheimer’s dementia. However, the new criteria include two new categories: probable and possible Alzheimer’s dementia with evidence of the Alzheimer’s pathophysiological process. These new criteria are for research purposes.

To diagnose probable Alzheimer’s dementia, one must first diagnose dementia itself (see Table 13-5 for the core clinical criteria for “all-cause” dementia). Patients who meet these criteria for all-cause dementia have probable Alzheimer’s dementia when they also meet the core clinical criteria outlined in Table 13-6. Briefly, patients with probable Alzheimer’s disease have dementia which is insidious in onset, clearly has demonstrated worsening of cognition over time, and has either an amnestic (problems with learning and recall) or a nonamnestic presentation (language, visuospatial, or executive dysfunction). Probable Alzheimer’s dementia with increased level of certaintycan be diagnosed pre-mortem when the patient meets the core criteria (Table 13-6) and also has formal documented cognitive decline on neuropsychological testing or has been proven to be a carrier of a causative Alzheimer’s disease genetic mutation (in the genes for APP, presenilin 1, or presenilin 2). Despite many ominious associations with neuroimaging biomarkers, the E4 allele of ApoE is not sufficiently specific to be considered in this category.

Table 13-5 Core clinical criteria for all-cause dementia

Dementia is diagnosed when there are cognitive or behavioral symptoms that:

1. Interfere with the ability to function at work or at usual activities; and

2. Represent a decline from previous levels of functioning and performing; and

3. Are not explained by delirium or major psychiatric disorder.

4. Cognitive impairment is detected and diagnosed through a combination of (1) history-taking from the patient and a knowledgeable informant and (2) an objective cognitive assessment, either a “bedside” mental status examination or neuropsychological testing. Neuropsychological testing should be performed when the routine history and bedside mental status examination cannot provide a confident diagnosis.

5. The cognitive or behavioral impairment involves a minimum of two of the following domains:

a. Impaired ability to acquire and remember new information – symptoms include: repetitive questions or conversations, misplacing personal belongings, forgetting events or appointments, getting lost on a familiar route.

b. Impaired reasoning and handling of complex tasks, poor judgment – symptoms include: poor understanding of safety risks, inability to manage finances, poor decision-making ability, inability to plan complex or sequential activity.

c. Impaired visuospatial ability – symptoms include: inability to recognize faces or common objects or to find objects in direct view despite good acuity, inability to operate simple implements or orient clothing to the body.

d. Impaired language functions (speaking, reading, writing) – symptoms include: difficulty thinking of common words while speaking; hesitation; speech, spelling, and writing errors.

e. Changes in personality, behavior or comportment – symptoms include: uncharacteristic mood fluctuations such as agitation, impaired motivation or initiative, apathy, loss of drive, social withdrawal, decreased interest in previous activities, loss of empathy, compulsive or obsessive behaviors, socially unacceptable behaviors.

Table 13-6 Core clinical criteria for probable Alzheimer’s dementia

Probable Alzheimer’s dementia is diagnosed when the patient:

Meets criteria for dementia (Table 13-5) and in addition, has the following characteristics:

1. Insidious onset. Symptoms have a gradual onset over months to years, not sudden over hours or days

2. Clear-cut history of worsening of cognition by report or observations

3. The initial and most prominent cognitive deficits are evident on history and examination in one of the following categories:

i. Amnestic presentation. It is the most common syndromic presentation of Alzheimer’s dementia. The deficits should include impairment in learning and recall of recently learned information. There should also be evidence of cognitive dysfunction in at least one other cognitive domain, as defined earlier in the text.

ii. Nonamnestic presentations:

1. Language presentation: the most prominent deficits are in word-finding, but deficits in other cognitive domains should be present.

2. Visuospatial presentation: the most prominent deficits are in spatial cognition, including object agnosia, impaired face recognition, simultanagnosia, and alexia. Deficits in other cognitive domains should be present.

3. Executive dysfunction: the most prominent deficits are impaired reasoning, judgment, and problem solving. Deficits in other cognitive domains should be present.

4. The diagnosis of probable Alzheimer’s dementia should not be applied when there is evidence of:

i. Substantial concomitant cerebrovascular disease defined by a history of a stroke temporally related to the onset or worsening of cognitive impairment; or the presence of multiple or extensive infarcts or severe white-matter hyperintensity burden; or

ii. Core features of dementia with Lewy bodies other than dementia itself; or

iii. Prominent features of semantic variant primary progressive aphasia or nonfluent/agrammatic variant primary progressive aphasia; or

iv. Evidence for another concurrent, active neurological disease, or a non-neurological medical comorbidity or use of medication that could have a substantial effect on cognition.

The new research category of probable Alzheimer’s dementia with evidence of the Alzheimer’s pathophysiological process includes patients with probable Alzheimer’s disease (Table 13-6) who have clearly positivebiomarker evidence either of brain amyloid deposition/amyloidosis (Figure 13-11) or of downstream neuronal degeneration (Figures 13-12 and 13-13). (In these cases, results from biomarker studies can be judged to be clearly positive, clearly negative, or indeterminate.) The new research category of possible Alzheimer’s dementia with evidence of the Alzheimer’s pathophysiological process is for persons who meet clinical criteria for a dementia other than Alzheimer’s disease but who have clearly positive biomarker evidence (pre-mortem) or neuropathological evidence (postmortem) of the Alzheimer’s pathophysiological process, including both evidence of amyloidosis and evidence of neuronal degeneration. This does not preclude that a second pathophysiological condition may also be present.

Targeting amyloid as a future disease-modifying treatment of Alzheimer’s disease

The likely fate of subjects with asymptomatic amyloidosis (stage I presymptomatic Alzheimer’s disease) or cognitive changes with early neurodegeneration (stage II Alzheimer’s disease, MCI) means it is becoming ever more urgent to intervene in this disorder at an earlier and earlier time, when brain changes are present without any overt cognitive decline, or certainly before dementia sets in. Since nearly all current Alzheimer’s drug development candidates target some aspect of the amyloid cascade, biomarkers have the potential to work hand-in-glove not only with making an early diagnosis potentially in the first or second stage of this disease to therefore identify which patients to treat with a specific agent of a given mechanism of action, but also to demonstrate objectively whether disease progression is slowed, halted, or even reversed by novel treatments that interfere with amyloidosis. From the point of view of current clinical practice, the value of the information that biomarkers can provide has to be balanced against their costs, the side effects of radioactivity, the invasiveness of lumbar puncture, the availability of specialized technology, and of course the psychological costs of learning about Alzheimer’s brain pathology and possible clinical prognosis at a time when there is no cure or even any therapy to halt or slow the progression of this disorder. Thus the main current utility of biomarker-enhanced early detection of Alzheimer’s disease is to identify those with high risk of progression to dementia, for participation in clinical trials of new drug testing, and especially of new drug testing of various anti-amyloid therapies.

Vaccines and immunotherapy

The quest for an Alzheimer’s vaccine has great appeal, but its clinical development has had its ups and downs. Immunizing the body to β-amyloid could in concept not only slow or stop progression of cognitive decline but, by removal of plaques already formed, potentially improve cognitive function. Positive tests of amyloid vaccines in animals led to early clinical trials that showed evidence not only of stabilization of memory in Alzheimer’s patients but, perhaps more importantly, that amyloid plaques were removed (Figure 13-14). However, the first vaccine to the Aβ peptide (AN1792) caused brain inflammation (meningoencephalitis) in 6% of cases in phase II, and the trials had to be stopped. Other immunotherapy trials include passive immunization with antibodies against Aβ peptide. However, results with bapineuzumab (humanized mouse monoclonal antibody against terminal portion of Aβ), solanezumab (humanized mouse monoclonal antibody against the midportion of Aβ), and others (crenezumab) have so far yielded disappointing results in clinical trials. There are also clinical trials of passive immunization with intravenous immunoglobulin (IVIG) in the hopes that it might contain naturally occurring antibodies against β-amyloid and promote the clearance of β-amyloid from the brain. Some surprisingly positive results in stopping decline in cognitive function with IVIG have been reported, with further testing required to follow this up.

Figure 13-14. Future treatments: β-amyloid immunizations. One potential future treatment for Alzheimer’s disease is a vaccine that immunizes against β-amyloid, which could not only slow cognitive decline but also perhaps remove already formed plaques.

Gamma-secretase inhibitors

Another strategy to block amyloid plaque formation is to inhibit the enzyme γ-secretase (Figure 13-15). Several γ-secretase inhibitors (GSIs) are in clinical development. Notably, however, semagacestat (LY450139) was terminated from clinical trials for safety reasons, namely that this agent actually impaired cognition and function more than placebo did, and also increased the occurrence of skin cancer. The future of this approach is now in doubt. What is confusing about the semagacestat findings is that it was shown to successfully target the enzyme γ-secretase and to reduce Aβ production in a dose-dependent manner, yet this did not translate into a clinical benefit. One mechanism of toxicity for semagacestat may be the fact that many GSIs also inhibit other proteases, especially one called Notch, which is involved in cell fate pathways in rapidly dividing cells, which may have caused the skin-cancer side effect. Future GSIs selective for γ-secretase and not for Notch may be necessary to move this target forward in Alzheimer’s disease.

Figure 13-15. Gamma-secretase inhibitors and modulators. The way in which amyloid precursor protein (APP) is processed may help determine whether an individual develops Alzheimer’s disease. Thus a drug that affects this process could prevent or treat Alzheimer’s disease. The enzyme γ-secretase cleaves embedded peptides, which in some cases leads to release of toxic peptides (particularly Aβ42). Inhibition of this enzyme could therefore prevent formation of toxic peptides, and so could modulation of this enzyme with selective amyloid-lowering agents (SALAs).

Beta-secretase inhibitors

Inhibitors of the β-secretase enzyme have been difficult to synthesize, but compounds such as SCH 1381252, CTS21666, and others are moving forward in clinical development, and their results are eagerly awaited because of their theoretical promise as a mechanism of preventing β-amyloid formation (Figure 13-16).

Figure 13-16. Beta-secretase inhibitors. The way in which amyloid precursor protein (APP) is processed may help determine whether an individual develops Alzheimer’s disease. Thus a drug that affects this process could prevent or treat Alzheimer’s disease. The enzyme β-secretase cuts APP at a spot outside the membrane to form two peptides: β-APP, which is soluble, and a 91-amino-acid peptide that remains in the membrane. Gamma-secretase then cuts the embedded peptide; this releases Aβ peptides of 40, 42, or 43 amino acids, which are toxic. Thus inhibition of β-secretase could prevent the formation of toxic peptides.

Targeting acetylcholine as a current symptomatic treatment of Alzheimer’s disease

Acetylcholine and the pharmacologic basis of cholinesterase treatments for dementia

Many of the current approved agents used to treat symptoms of dementia in Alzheimer’s disease are based upon boosting the availability of the neurotransmitter acetylcholine. Prior to discussing these treatments, we will review the pharmacology of acetylcholine.

Acetylcholine is formed in cholinergic neurons from two precursors: choline and acetyl coenzyme A (AcCoA) (Figure 13-17). Choline is derived from dietary and intraneuronal sources, and AcCoA is made from glucose in the mitochondria of the neuron. These two substrates interact with the synthetic enzyme choline acetyl-transferase (CAT) to produce the neurotransmitter acetylcholine (ACh).

Figure 13-17. Acetylcholine is produced. Acetylcholine is formed when two precursors – choline and acetyl coenzyme A (AcCoA) – interact with the synthetic enzyme choline acetyl-transferase (CAT). Choline is derived from dietary and intraneuronal sources, and AcCoA is made from glucose in the mitochondria of the neuron.

ACh’s actions are terminated by one of two enzymes, either acetylcholinesterase (AChE) or butyrylcholinesterase (BuChE), sometimes also called “pseudocholinesterase” or “nonspecific cholinesterase” (Figure 13-18). Both enzymes convert ACh into choline, which is then transported back into the presynaptic cholinergic neuron for resynthesis into ACh (Figure 13-18). Although both AChE and BuChE can metabolize ACh, they are quite different in that they are encoded by separate genes and have different tissue distributions and substrate patterns. There may be different clinical effects of inhibiting these two enzymes as well. High levels of AChE are present in brain, especially in neurons that receive ACh input (Figure 13-18). BuChE is also present in brain, especially in glial cells (Figure 13-18). As will be discussed below, some cholinesterase inhibitors specifically inhibit AChE, whereas others inhibit both enzymes. It is AChE that is thought to be the key enzyme for inactivating ACh at cholinergic synapses, although BuChE can take on this activity if ACh diffuses to nearby glia. AChE is also present in the gut, skeletal muscle, red blood cells, lymphocytes, and platelets. BuChE is also present in the gut, plasma, skeletal muscle, placenta, and liver. BuChE may be present in some specific neurons, and it may also be present in amyloid plaques.

Figure 13-18. Acetylcholine’s action is terminated. Acetylcholine’s action can be terminated by two different enzymes: acetylcholinesterase (AChE), which is present both intra- and extracellularly, and butyrylcholinesterase (BuChE), which is particularly present in glial cells. Both enzymes convert acetylcholine into choline, which is then transported out of the synaptic cleft and back into the presynaptic neuron via the choline transporter. Once inside the presynaptic neuron, choline can be recycled into acetylcholine and then packaged into vesicles by the vesicular transporter for acetylcholine (VAChT).

ACh released from CNS neurons is destroyed too quickly and too completely by AChE to be available for transport back into the presynaptic neuron, but the choline that is formed by the breakdown of ACh is readily transported back into the presynaptic cholinergic nerve terminal by a transporter similar to the transporters for other neurotransmitters already discussed earlier in relation to norepinephrine, dopamine, and serotonin neurons. Once back in the presynaptic nerve terminal, it can be recycled into new ACh synthesis (Figure 13-18). Once synthesized in the presynaptic neuron, ACh is stored in synaptic vesicles after being transported into these vesicles by the vesicular transporter for ACh (VAChT), analogous to the vesicular transporters for the monoamines and other neurotransmitters.

There are numerous receptors for ACh (Figures 13-19 and 13-20). The major subtypes are nicotinic and muscarinic subtypes of cholinergic receptors. Classically, muscarinic receptors are stimulated by the mushroom alkaloid muscarine and nicotinic receptors by the tobacco alkaloid nicotine. Nictotinic receptors are all ligand-gated, rapid-onset, and excitatory ion channels blocked by curare. Muscarinic receptors, by contrast, are G-protein-linked, can be excitatory or inhibitory, and many are blocked by atropine, scopolamine, and other well-known so-called “anticholinergics” discussed throughout this text. Both nicotinic and muscarinic receptors have been further subdivided into numerous receptor subtypes.

Figure 13-19. Muscarinic acetylcholine receptors. Acetylcholine neurotransmission can be regulated by G-protein-linked muscarinic acetylcholine receptors, shown here. Muscarinic 1 (M1) receptors are postsynaptic and important for regulation of memory. Muscarinic 2 (M2) receptors exist both presynaptically as autoreceptors and postsynaptically. Other postsynaptic muscarinic receptors include M3, M4, and M5.

Figure 13-20. Nicotinic acetylcholine receptors. Acetylcholine neurotransmission can be regulated by ligand-gated excitatory ion channels known as nicotinic acetylcholine receptors, shown here. There are multiple subtypes of these receptors, defined by the subunits they contain. Two of the most important are those that contain all α7 subunits and those that contain α4 and β2 subunits. The α7 receptors can exist presynaptically, where they facilitate acetylcholine release, or postsynaptically, where they are important for regulating cognitive function. The α4β2 receptors are postsynaptic and regulate release of dopamine in the nucleus accumbens.

Subtypes of muscarinic receptors include the well-known postsynaptic M1 subtype, which appears to be key to the regulation of some of the memory functions of ACh acting at cholinergic synapses (Figure 13-19). The M2subtype is presynaptic, and serves as an autoreceptor, blocking the further release of ACh when it is activated by the build-up of synaptic levels of ACh (Figure 13-19). The functions of other muscarinic receptor subtypes are still under investigation, including some, such as the M3 subtype, which are also expressed outside of the brain and may mediate some of the peripheral side effects of some anticholinergics.

A number of nicotinic receptor subtypes also exist in the brain, with different subtypes outside the brain in skeletal muscle and ganglia. Two of the most important CNS nicotinic cholinergic receptors are the subtype with all α7 subunits, and the subtype with α4 and β2 subunits (Figure 13-20). The α4β2 subtype is postsynaptic and plays an important role in regulating dopamine release in the nucleus accumbens. It is thought to be a primary target of nicotine in cigarettes, and to contribute to the reinforcing and addicting properties of tobacco. The α4β2 subtypes of nicotinic cholinergic receptors are discussed in further detail in Chapter 14 on drug abuse.

Alpha-7-nicotinic cholinergic receptors can be either presynaptic or postsynaptic (Figures 13-20 and 13-21). When they are postsynaptic, they may be important mediators of cognitive functioning in the prefrontal cortex. When they are presynaptic and on cholinergic neurons, they appear to mediate a “feed-forward” release process where ACh can facilitate its own release by occupying presynaptic α7-nicotinic receptors (Figure 13-20). Furthermore, α7-nicotinic receptors are present on neurons that release other neurotransmitters, such as dopamine and glutamate neurons (Figure 13-21). When ACh diffuses away from its synapse to occupy these presynaptic heteroreceptors, it facilitates the release of the neurotransmitter there (e.g., dopamine or glutamate) (Figure 13-21).

Figure 13-21. Presynaptic nicotinic heteroreceptors facilitate dopamine and glutamate release. Acetylcholine (ACh) that diffuses away from the synapse can bind to presynaptic α7-nicotinic receptors on dopamine and glutamate neurons, where it stimulates release of these neurotransmitters.

Just as described for other ligand-gated ion channels such as the GABAA receptor and the NMDA receptor, it appears that ligand-gated nicotinic cholinergic receptors are also regulated by allosteric modulators (Figure 13-22). Positive allosteric modulators (PAMs) have been identified for nicotinic receptors in brain; indeed, the cholinesterase inhibitor galantamine has a second therapeutic mechanism as a PAM for nicotinic receptors, as described for this agent below.

Figure 13-22. Allosteric modulation of nicotinic receptors. Nicotinic receptors can be regulated by allosteric modulators. These ligand-gated ion channels control the flow of calcium into the neuron (top panel). When acetylcholine (ACh) is bound to these receptors, it allows calcium to pass into the neuron (middle panel). A positive allosteric modulator bound in the presence of acetylcholine increases the frequency of opening of the channel and thus can allow for more calcium to pass into the neuron (bottom panel).

The principal cholinergic pathways are illustrated in Figures 13-23 and 13-24. Cell bodies of some cholinergic pathways arise from the brainstem and project to many brain regions, including prefrontal cortex, basal forebrain, thalamus, hypothalamus, amygdala, and hippocampus (Figure 13-23). Other cholinergic pathways have their cell bodies in the basal forebrain, project to prefrontal cortex, amygdala, and hippocampus, and are thought to be particularly important for memory (Figure 13-24). Additional cholinergic fibers in the basal ganglia are not illustrated.

Figure 13-23. Cholinergic projections from the brainstem. The cell bodies of cholinergic neurons can be found in the brainstem and project to many different brain areas including prefrontal cortex (PFC), basal forebrain (BF), thalamus (T), hypothalamus (Hy), amygdala (A), and hippocampus (H).

Figure 13-24. Cholinergic projections from the basal forebrain. Other cholinergic neurons project from the basal forebrain (BF) to prefrontal cortex (PFC), amygdala (A), and hippocampus (H). They are thought to be important for memory.

The cholinergic deficiency hypothesis of amnesia in Alzheimer’s disease and other dementias

Numerous investigators have shown that a deficiency in cholinergic functioning is linked to a disruption in memory, particularly short-term memory. For example, blockers of muscarinic cholinergic receptors (such as scopolamine) can produce a memory disturbance in normal human volunteers that has similarities to the memory disturbance in Alzheimer’s disease. Boosting cholinergic neurotransmission with cholinesterase inhibitors not only reverses scopolamine-induced memory impairments in normal human volunteers, but also enhances memory functioning in patients with Alzheimer’s disease. Both animal and human studies have demonstrated that the nucleus basalis of Meynert in the basal forebrain is the major brain center for cholinergic neurons that project throughout the cortex (Figure 13-24). These neurons have the principal role in mediating memory formation. It is suspected that the short-term memory disturbance of Alzheimer’s patients is due to degeneration of these particular cholinergic neurons. Other cholinergic neurons, such as those in the striatum and those projecting from the lateral tegmental area, are not involved in the memory disorder of Alzheimer’s disease.

Cholinesterase inhibitors


The most successful approach to boosting cholinergic functioning in patients with Alzheimer’s disease and improving memory has been to inhibit ACh destruction by blocking the enzyme acetylcholinesterase (Figure 13-18). This causes the build-up of ACh because it can no longer be destroyed by acetylcholinesterase. The enhanced availability of acetylcholine can impact the clinical outcome in Alzheimer’s disease, from enhancing memory in some patients, to slowing the decline in function of Alzheimer’s patients for several months, rather than improving their memory. Since cholinergic agents require postsynaptic cholinergic receptors to mediate the benefits of the enhanced cholinergic input, they may be most effective in the early stages of Alzheimer’s disease, while postsynaptic cholinergic targets are still present. However, late in the illness, degeneration of neurons that have postsynaptic ACh receptors means that the drug may lose its benefits.


Donepezil is a reversible, long-acting, selective inhibitor of acetylcholinesterase (AChE) without inhibition of butyrylcholinesterase (BuChE) (Figure 13-25). Donepezil inhibits AChE in pre- and postsynaptic cholinergic neurons, and in other areas of the CNS outside of cholinergic neurons where this enzyme is widespread (Figure 13-25A). Its CNS actions boost the availability of ACh at the remaining sites normally innervated by cholinergic neurons, but which are now suffering from a deficiency of ACh as cholinergic neurons die off (Figure 13-25A). Donepezil also inhibits AChE in the periphery, where its actions in the gastrointestinal (GI) tract can produce GI side effects (Figure 13-25B). Donezezil is easy to dose and has mostly gastrointestinal side effects, which are mostly transient.

Figure 13-25. Donepezil actions. Donepezil inhibits the enzyme acetylcholinesterase (AChE), which is present both in the central nervous system (CNS) and peripherally. (A) Central cholinergic neurons are important for regulation of memory; thus in the CNS, the boost of acetylcholine caused by AChE blockade contributes to improved cognitive functioning. (B) Peripheral cholinergic neurons in the gut are involved in gastrointestinal effects; thus the boost in peripheral acetylcholine caused by AChE blockade may contribute to gastrointestinal side effects. Donepezil is represented here by a straitjacket icon with the American and Japanese flags (the countries of its manufacturers).


Rivastigmine (Figure 13-26) is “pseudoirreversible” (which means it reverses itself over hours), intermediate-acting, and not only selective for AChE over BuChE, but perhaps for AChE in the cortex and hippocampus over AChE in other areas of brain (Figure 13-26A). Rivastigmine also inhibits BuChE within glia, which may contribute somewhat to the enhancement of ACh levels within the CNS (Figure 13-26A). Inhibition of BuChE within glia may be even more important in patients with Alzheimer’s disease as they develop gliosis when cortical neurons die, because these glia contain BuChE, and inhibition of this increased enzyme activity may have a favorable action on increasing the availability of ACh to cholinergic receptors via this second mechanism. Rivastigmine appears to have comparable safety and efficacy to donepezil, although it may have more gastrointestinal side effects when given orally (Figure 13-26B), perhaps due to its pharmacokinetic profile, and perhaps due to inhibition of both AChE and BuChE in the periphery (Figure 13-26C). However, there is now a transdermal formulation of rivastigmine available that greatly reduces the peripheral side effects of oral rivastigmine, probably by optimizing drug delivery and reducing peak drug concentrations.

A. Rivastigmine actions, part 1. Rivastigmine inhibits the enzymes acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE), which are present both in the central nervous system (CNS) and peripherally. Central cholinergic neurons are important for regulation of memory; thus in the CNS the boost of acetylcholine caused by AChE blockade contributes to improved cognitive functioning. In particular, rivastigmine appears to be somewhat selective for AChE in the cortex and hippocampus – two regions important for memory – over other areas of the brain. Rivastigmine’s blockade of BuChE in glia may also contribute to enhanced acetylcholine levels. Rivastigmine is represented here by two straitjacket icons end-to-end, one for acetylcholinesterase and another for butyrylcholinesterase, with the Swiss flag indicating the country of its manufacturer.

B. Rivastigmine actions, part 2. Rivastigmine inhibits the enzymes acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE), which are present both in the central nervous system (CNS) and peripherally. Inhibition of BuChE may be more important in later stages of disease, because as more cholinergic neurons die and gliosis occurs, BuChE activity increases.

C. Rivastigmine actions, part 3. Rivastigmine inhibits the enzymes acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE), which are present both in the central nervous system (CNS) and peripherally. Peripheral cholinergic neurons in the gut are involved in gastrointestinal effects; thus the boost in peripheral acetylcholine caused by AChE and BuChE blockade may contribute to gastrointestinal side effects.

Figure 13-26


Galantamine is a very interesting cholinesterase inhibitor found in snowdrops and daffodils! It has a dual mechanism of action, matching AChE inhibition with positive allosteric modulation (PAM) of nicotinic cholinergic receptors (Figure 13-27). Theoretically, the inhibition of AChE (Figure 13-27A) could be enhanced when joined by the second action of galantamine at nicotinic receptors (Figure 13-27B). Thus, raising ACh levels at nicotinic cholinergic receptors by AChE inhibition could be boosted by the PAM actions of galantamine (Figure 13-27B). However, it has not been proven that this theoretically advantageous second action as a nicotinic PAM translates into clinical advantages.

A. Galantamine actions, part 1. Galantamine inhibits the enzyme acetylcholinesterase (AChE). Central cholinergic neurons are important for regulation of memory, and thus in the CNS the boost of acetylcholine caused by AChE blockade contributes to improved cognitive functioning. Galantamine is represented here by a straitjacket icon with a lightbulb on top. The straitjacket has a daffodil on front, since galantamine was originally extracted from daffodils; the light bulb represents a second mechanism of action of galantamine, namely positive allosteric modulation of nicotinic receptors.

B. Galantamine actions, part 2. Galantamine is unique among cholinesterase inhibitors in that it is also a positive allosteric modulator (PAM) at nicotinic cholinergic receptors, which means it can boost the effects of acetylcholine at these receptors. Thus galantamine’s second action as a PAM at nicotinic receptors could theoretically enhance its primary action as a cholinesterase inhibitor.

Figure 13-27

Targeting glutamate

The glutamate hypothesis of cognitive deficiency in Alzheimer’s disease

Glutamate has been hypothesized to be released in excess during Alzheimer’s disease, perhaps in part triggered by neurotoxic amyloid plaques and neurofibrillary tangles (Figure 13-28). In the resting state, glutamate is normally quiet, and the NMDA receptor is physiologically blocked by magnesium ions (Figure 13-28A). When normal excitatory neurotransmission comes along, a flurry of glutamate is released (Figure 13-28B). The postsynaptic NMDA receptor is a “coincidence detector” and allows inflow of ions if three things happen at the same time: neuronal depolarization, often from activation of nearby AMPA receptors; glutamate occupying its binding site on the NMDA receptor; and the cotransmitter glycine occupying its site on the NMDA receptor (Figure 13-28B). If plaques and tangles cause a steady “leak” of glutamate, this would theoretically interfere with the fine-tuning of glutamate neurotransmission, and possibly interfere with memory and learning but not necessarily damaging neurons (Figure 13-28C). Hypothetically, as the disease progresses, glutamate release could be increased to a level that is tonically bombarding the postsynaptic receptor, eventually killing off dendrites and then killing off full neurons due to excitotoxic cell death (Figure 13-28C).

Figure 13-28. Amyloid plaques and glutamate excitotoxicity. (A) In the resting state glutamate is quiet and N-methyl-D-aspartate (NMDA) receptors are blocked by magnesium. (B) With normal neurotransmission, glutamate binds to NMDA receptors and, if the postsynaptic receptor is depolarized and glycine is simultaneously bound to the NMDA receptors, the channel opens and allows ion influx. (C) If amyloid’s synaptic effects include downregulating the glutamate transporter, inhibiting glutamate reuptake, or enhancing glutamate release, this could cause a steady leak of glutamate and result in excessive calcium influx in postsynaptic neurons, which in the short term may cause memory problems and in the long term may cause accumulation of free radicals and thus destruction of neurons.


The rationale for the use of memantine, a type of NMDA antagonist, is to reduce abnormal activation of glutamate neurotransmission and thus interfere with the pathophysiology of Alzheimer’s disease, improve cognitive function, and slow the rate of decline over time. Blocking NMDA receptors chronically would interfere with memory formation and neuroplasticity. So what do you do to decrease the excessive and sustained but low level of excitotoxic activation of NMDA receptors yet not interfere with learning, memory, and neuroplasticity, and without inducing a schizophrenia-like state?

The answer seems to be that you interfere with NMDA-mediated glutamatergic neurotransmission with a weak (low-affinity) NMDA antagonist that works at the same site plugging the ion channel where the magnesium ion normally blocks this channel at rest. That is, memantine is an uncompetitive open-channel NMDA receptor antagonist with low to moderate affinity, voltage dependence, and fast blocking and unblocking kinetics. That is a fancy way of saying that it only blocks the ion channel of the NMDA receptor when it is open. This is why it is called an open-channel antagonist and why it is dependent upon voltage: namely, to open the channel. It is also a fancy way of saying that memantine blocks the open channel quickly, but is readily and quickly reversible if a barrage of glutamate comes along from normal neurotransmssion.

This concept is illustrated in Figure 13-29. Firstly, the hypothetical state of the glutamate neuron during Alzheimer’s excitotoxicity is shown in Figure 13-29A. Here, steady, tonic and excessive amounts of glutamate are continuously released in a manner that interferes with the normal resting state of the glutamate neuron (Figure 13-28A), and in a manner that interferes with established memory functions, new learning, and normal neuronal plasticity. Eventually, this leads to the activation of intracellular enzymes that produce toxic free radicals that damage the membranes of the postsynaptic dendrite and eventually destroy the entire neuron (Figure 13-29A). When memantine is given, it blocks this tonic glutamate release from having downstream effects, thus returning the glutamate neuron to a new resting state despite the continuous release of glutamate (Figure 13-29B). Hypothetically, this stops the excessive glutamate from interfering with the resting glutamate neuron’s physiological activity, therefore improving memory; it also hypothetically stops the excessive glutamate from causing neurotoxicity, therefore slowing the rate of neuronal death and also the associated cognitive decline that this causes in Alzheimer’s disease (Figure 13-29B).

Figure 13-29. Memantine actions. Memantine is a noncompetitive low-affinity N-methyl-D-aspartate (NMDA) receptor antagonist that binds to the magnesium site when the channel is open. (A) If amyloid’s synaptic effects lead to a steady (tonic) leak of glutamate and result in excessive calcium influx in postsynaptic neurons, this could cause memory problems and, in the long term, accumulation of free radicals and thus destruction of neurons. (B) Memantine blocks the downstream effects of tonic glutamate release by “plugging” the NMDA ion channel and thus may improve memory and prevent neurodegeneration. (C) Because memantine has low affinity, when there is a phasic burst of glutamate and depolarization occurs, this is enough to remove memantine from the ion channel and thus allow normal neurotransmission.

However, at the same time, memantine is not so powerful a blocker of NMDA receptors that it stops all neurotransmission at glutamate synapses (Figure 13-29C). That is, when a phasic burst of glutamate is transiently released during normal glutamatergic neurotransmission, this causes a depolarization that is capable of reversing the memantine block, until the depolarization goes away (Figure 13-29C). For this reason, memantine does not have the psychotomimetic actions of other more powerful NMDA antagonists such as PCP (phencyclidine) and ketamine, and does not shut down new learning or the ability of normal neurotransmission to occur when necessary (Figure 13-29C). The blockade of NMDA receptors by memantine can be seen as a kind of “artificial magnesium,” more effective than physiological blockade by magnesium, which is overwhelmed by excitotoxic glutamate release, but less effective than PCP or ketamine so that the glutamate system is not entirely shut down. Sort of like having your cake and eating it, too.

Memantine also has σ antagonist properties and weak 5HT3 antagonist properties, but it is not clear what these contribute to the actions of this agent in Alzheimer’s disease. Since its mechanism of action in Alzheimer’s disease is so different from cholinesterase inhibition, memantine is usually given concomitantly with a cholinesterase inhibitor to exploit the potential of both of these approaches and to get additive results in patients.

Treatments for psychiatric and behavioral symptoms in dementia

Dementia is not just a disturbance of memory, as many patients have a variety of behavioral and emotional symptoms as well. Treatment of agitation and aggression in dementia is a very controversial area, due to the potential for misuse of antipsychotics as “chemical straightjackets” to over-tranquilize patients, and also safety concerns about cardiovascular events and death from these drugs. Antipsychotics are thus not recommended for use for agitation and behavioral symptoms of Alzheimer’s disease, because there is little evidence of efficacy from controlled trials and also because there are demonstrated safety concerns that antipsychotics cause increased cardiovascular events and increased mortality in elderly patients with dementia. At this time, no antipsychotic is FDA-approved for this use and all carry warnings about the risk of cardiovascular events and increased mortality in this population. Because, in the real world, there are also risks of nontreatment, including early institutionalization and the dangers of agitated and psychotic behaviors to the patient and others around them, some patients will nevertheless require treatment with an atypical antipsychotic. In this case, risperidone is often a preferred agent at very low doses. Clinicians should be alerted to the need to distinguish Alzheimer’s disease from dementia with Lewy bodies prior to prescribing an antipsychotic. Patients with dementia with Lewy bodies can look psychotic, with their prominent behavioral symptoms, dramatic fluctuations, and visual hallucinations, but are exquisitely sensitive to extrapyramidal side effects even of the atypical antipsychotics, which can result in very severe and potentially life-threatening reactions to such drugs. An agent approved for treatment of the behavioral symptoms of dementia would be a welcome solution to a huge unmet need for these patients.

Before using medications at all, reversible precipitants of agitation in dementia should be managed: pain, nicotine withdrawal, medication side effects, undiagnosed medical and neurological illnesses, and provocative environments that are either too stimulating or not stimulating enough. When use of medications is necessary, cholinesterase inhibitors may be effective in some patients and are a first-line consideration in Alzheimer’s disease, but might work better for prevention of these symptoms than for their treatment once they have emerged. Also, frontotemporal dementia patients may be more likely to benefit from SSRIs (e.g., citalopram or escitalopram) or SNRIs. In general, first-line treatment of agitation and aggression in dementia is SSRI/SNRI therapy. Second-line treatments that may help to avoid use of atypical antipsychotics can also include β blockers, valproate, gabapentin, pregabalin, and selegilene. Others may respond to carbamazepine, oxcarbazepine, benzodiazepines, buspirone, or trazodone.

Other proposed targets for dementia

A number of psychopharmacological agents have been tested for their potential as treatments in Alzheimer’s disease, but none has yet proven effective. This includes various antioxidants, anti-inflammatory agents, statins, vitamin E, estrogen, the MAO inhibitor selegiline, the antidiabetic agent rosiglitazone and other peroxisome proliferator-activated receptor gamma (PPARγ) agonists, lithium and other glycogen synthase kinase (GSK) inhibitors, agents that attempt to block tau phosphorylation, and phosphodiesterase inhibitors.

Many of the same agents proposed as pro-cognitive in ADHD and discussed in Chapter 12, and as pro-cognitive in schizophrenia and discussed in Chapter 5, have also been studied in Alzheimer’s disease as potential symptomatic treatments, from H3 histamine antagonists, to AMPAkines, to nicotinic cholinergic agonists, 5HT6 antagonists, phosphodiesterase inhibitors, metabotropic glutamate receptor agents, and others, but not with robust promise at the present time.


The most common dementia is Alzheimer’s disease, and the leading theory for its etiology is the amyloid cascade hypothesis. Other dementias are briefly discussed as well, as are their differing pathologies. New diagnostic criteria now propose that there are three stages of Alzheimer’s disease. The first stage is preclinical, asymptomatic but with amyloid accumulation; the second stage is mild cognitive impairment, with both amyloid accumulation and biomarker evidence of neurodegeneration in the presence of memory problems; and the last stage, dementia. Major research efforts are attempting to find disease-modifying treatments that could halt or even reverse the course of this illness by interfering with amyloid accumulation in the brain. Leading treatments for Alzheimer’s disease today include the cholinesterase inhibitors, based upon the cholinergic hypothesis of amnesia, and memantine, an NMDA antagonist, based upon the glutamate hypothesis of cognitive decline.