Chapter 11 - Changes in Behavior

Section 2

1. "Are you feeling sad or depressed?" The more straightforward, the better. "Naw, I don't feel bad." He snaps his fingers. "I'm good." Had his mood (what he reports) or affect (what we observe) been depressed, we might explore this more fully.

2. "Do you have any scary thoughts? Like, people are out to get you?"An extensive series of questions may be necessary to probe for delusions - fixed false ideas - but even one simple query like this gives the patient a chance to ventilate about things that they may have been reluctant to bring up before. "What do ya' mean?" our patient says. No one wants to bother me."

3. "Do you see things that no one else can see? Or hear things like voices when there's no one around?"Again, we're only doing a quick screen for a subset of hallucinations. As blunt and awkward as these questions may sound, psychosis may be hidden until they're asked. "What, do you think I'm crazy?" It's the first time he bristles. You make some apology and carry on. He shrugs, quickly placated. But he turns to his daughter, "Now you take a pie to Fred, all right?" Spontaneous behaviors like this may be more revealing than any of our formal questions. There's no apparent reason for this skewing of his attention. Unless there is a subtext that we've missed, this man has just demonstrated again his tenuous connection to the circumstances he finds himself in.

A final issue to be considered—not so much as a specific point of examination but as a general question with every patient—is the possibility that the symptoms are psychosomatic, either unconscious somatoform problems (inappropriately called hysteria), or conscious malingering. Many articles have been written proposing tricks of history or examination to uncover such cases. However, as the late D. Frank Benson, M.D. of UCLA put it, "All hysterics die of organic disease." In fact, there is evidence that the signs relied upon to discriminate psychosomatic conditions are often positive in patients with organic illness. Also, a high proportion of patients discharged with an initial diagnosis of hysteria are soon re-examined and found to have an organic problem. We would much rather pursue organicity in the face of doubts than unwittingly dismiss a sick patient.

This is perhaps the essence of the neurobehavioral evaluation: we try to rapidly estimate the likelihood of delirium and to rule out severe mood or thought disorders. That's it. More extensive assessment will certainly improve our diagnostic accuracy, but the practical goal is to narrow the differential and assess the urgency of intervention. By this point we have asked about 20 questions and gotten a sense for the rate of onset and degree of change. We may not have a specific diagnosis, but we know our patient needs further attention today. We check with the nurse regarding our other patients, listen to our asthmatic's lungs, answer a phone call, look at some labs, and return to Mr. Johnson's room to proceed with a quick physical and elementary neurological examination.

FEATURES OF THE PHYSICAL AND ELEMENTARY NEUROLOGICAL EXAMINATION OF SPECIAL RELEVANCE TO THE ASSESSMENT OF BEHAVIOR

The components of the physical/neurological examination are well known and covered extensively elsewhere. In this section we will simply focus on several features of the examination that may speed diagnosis.

It is often remarked that an examination is focal or nonfocal. This is somewhat misleading, particularly when "nonfocal" is used as a substitute for "normal." Patients in status epilepticus or coma may have "nonfocal" examinations, so this designation hardly helps specify the diagnosis or the medical acuity. We suggest not using this designation, and instead cataloguing the presence or absence of certain distinguishing features, which help to construct a simple algorithm for narrowing the differential diagnosis. See Figure 3.

1. Is there any brainstem dysfunction? The presence of nystagmus, dysconjugate gaze, or altered pupillary responsiveness are easy-to-check hard signs that the patient's problem involves the CNS. Visual field deficits may be the only sign of temporal, parietal or occipital lobe lesions. We got an idea of hearing very early in the assessment, so formal testing is less urgent. We've also long since noted slurred speech or other dysarthria, which would strongly support a CNS localization, although we can't be sure that slurring specifically means brainstem dysfunction since cerebellar, basal gangliar, or even hemispheric injury may alter speech. Any brainstem signs strongly increase suspicion of a medical/ neurological cause of AMS.

2. Is there any motor dysfunction? Asymmetric strength is obviously suggestive of asymmetric CNS injury—whether ipsilateral or contralateral. But general slowing also narrows the diagnosis, suggesting decreased brainstem/arousal (as in toxic/metabolic encephalopathies), decreased midline frontal/motivation (e.g., frontal convexity meningioma, hydrocephalus, or depression), or impaired basal ganglia function (e.g., Parkinson's). Incoordination on the finger-to-nose test draws attention to the cerebellum, but may actually result from brainstem dysfunction due to toxic CNS depression, or increased intracranial pressure (IICP). Any motor deficit is a potent hint that there is a medical/neurological cause of AMS.

3. Is there any gait dysfunction? Even a ten second look at the patient's gait can be extremely revealing. Schizophrenics who do not have a superimposed neurological problem will walk well. Alzheimer's patients walk well until late in the course. In contrast, most processes that affect the brainstem, cerebellum, motor strip, basal ganglia, intracranial pressure, proprioceptive regions, or spinal cord may show up as slow or awkward gait.

4. Can the patient feel their feet? Peripheral neuropathy may be a quick hint at the presence of several problems causing AMS, such as diabetes or B12 deficiency. A detailed sensory examination can take a great deal of frustrating time in confused patients, but their awareness that you've moved a toe takes a few seconds and serves as an adequate screen.

Mr. Johnson has questionable bilateral gaze-evoked nystagmus. His gait is slightly stooped but he walks faster than we do, grinning at his own performance and not quite tipping on his excessively quick turn.

PART II: DIFFERENTIAL DIAGNOSIS OF CONFUSION AND DEMENTIA

At this point, we might review the spectrum of etiologies to be considered when a patient presents with a poorly defined, but clearly abnormal mental status. As the case we are considering illustrates, it is sometimes hard to distinguish between confusion (or delirium) and dementia. But, for the sake of organization, we will consider these as distinct, if overlapping, syndromes.

Dementias

Classifying dementias.Dementia is defined in the American Academy of Neurology Practice Handbook (1995) as "A clinical state characterized by significant loss of function in multiple cognitive domains, that is not due to an impaired level of arousal." The Handbook further states, "The presence of dementia does not necessarily imply irreversibility, a progressive course, or any specific cause." Note that it is difficult to provide a precise definition. The root meaning of dementia is simply a decline in mind. If we strictly applied the root meaning of the word, dementia would include any confusional state and any circumscribed deficit. For practical purposes, we exclude from the diagnosis of dementia individuals with either confusional state (delirium) or restricted deficits (e.g., aphasia). It is more controversial whether to exclude those with cognitive deficits accompanying psychiatric disorders. That is, some authorities regard the impaired cognition in severe depression as a "pseudodementia," others note that, whether or not it is reversible, it has all the features of dementia and should be called the "dementia of depression."

A straightforward way to classify dementias is as brain-related versus general medical disorder-related. For instance, the neurodegenerative dementias of Alzheimer's or Parkinson's disease, as well as other conditions such as brain tumors or normal pressure hydrocephalus are brain-related , relatively sparing other organ systems. In contrast, disorders such as hypothyroidism or vitamin B12 deficiency might be considered primarily systemic, due to a general medical disorder with significant effects beyond the cerebrum. This distinction breaks down a bit with vascular disease, since strokes can produce a dementia that is primarily brain related, but most stroke patients have widespread vascular disease and, in fact, most stroke survivors die of heart attacks.

Two other ways of classifying dementias used to be quite popular but have lost some luster with advancing research. One is the distinction between cortical versus subcortical dementias. Of course, some dementias exhibit most of their cell loss in the cortex of the brain (e.g., Alzheimer's disease) while others exhibit cell loss mostly from the deep subcortical structures, such as the midbrain in Parkinson's disease or the deep white matter in some vascular conditions. Theoretically, cortical dementias are more likely to produce symptoms of aphasia and amnesia; subcortical dementias are more likely to produce motor slowing with preserved ability to remember with cues. While motor signs are quite important since they will influence treatment, the so-called cortical and subcortical cognitive signs overlap so much that this distinction rarely offers practical help.

Similarly, there is an old tradition of classifying dementias as treatable versus untreatable ; for example, older textbooks often classified syphilis or B12 deficiency as treatable and Alzheimer's disease as untreatable. The doctor's express goal in working up dementia was to "rule out treatable causes." This is history. Today, there is meaningful treatment for cognitive and psychiatric features of all the common neurodegenerative dementias, and, in fact, we can probably produce a greater change in the natural history of Alzheimer's disease than we can in the two "treatable" disorders mentioned.

Mild cognitive impairment (MCI; formerly referred to as age-associated memory impairment (AAMI) or age-related cognitive decline) is a phrase used to refer to cognitive decline that is less global and severe than dementia, and less likely to interfere with daily functioning. Some authorities consider this as normal-for-age impairment; others believe that MCI is a transitional state en route to Alzheimer's. However, it is controversial whether these patients truly represent a homogenous group. When followed for several years, some MCI patients decline much more than others do. Therefore, it is most likely that some people who meet criteria for MCI will go on to exhibit some version of neurodegeneration, while others will remain in the category of "normal aging."

The most common dementias.The most common types of dementia, in order of frequency, are

• Dementia of the Alzheimer's Type (DAT)
• Dementia with Lewy Bodies (DLB)(described below)
• Vascular Dementia (VaD)
• Mixed dementias

Despite tremendous strides in unraveling the pathophysiology of dementias in the last decade, each of these syndromes and the boundaries between them remain controversial. (SeeTable 5 and Appendix B)

Table 5: Causes of Dementia and Confusion
(Please seeAppendix Bfor full list)

• Common Brain-related Syndromes
  e.g., Dementia of the Alzheimer's Type (DAT)
• Common General Medical Causes
  e.g., Hypothyroidism
• Less Common Brain-related Syndromes
  e.g., Frontotemporal lobar dementia
• Less Common General Medical Causes
  e.g., Hyperthyroidism
• II. Confusion/Delirium
• More Common Brain-related Causes
  e.g., Sleep deprivation
• More Common General Medical Causes
  e.g., Drug toxicity or withdrawal (seeTable 4)
• Less Common Brain-related Causes
  e.g., Epidural hematoma
• Less Common General Medical Causes
  e.g., Porphyria

The term "Alzheimer's disease" (AD), for instance, is often used to encompass two groups of conditions:

1. Several specific genetic diseases with related clinical and neuropathological features - chiefly the occurrence of excessive amyloid deposition, neuritic plaques and neurofibrillary tangles containing hyperphosphorylated microtubule-associated protein tau.

2. A broad spectrum of neurodegeneration without known genetic causes that is associated with some degree of decline in mind, but varies considerably in age of onset, rate of progression, and behavioral features, and overlaps to a poorly-defined extent with the clinical and pathological features of normal aging and the genetic Alzheimer's diseases.

In fact, it is unclear at this point what constitutes "normal aging," and whether to consider the various neurodegenerative conditions, including so-called Alzheimer's, as a collection of distinct "diseases" or as a spectrum of overlapping changes expected in the human aging process. Therefore, it is probably imprecise to label the above two spectrums as a single "disease". It may be more precise to reserve the term AD for the four related amyloid-excess genetic diseases (about 10 percent of cases), and to refer to the much larger second group as exhibiting Alzheimer's-type aging, Alzheimer's syndrome, ordementia of the Alzheimer's type (DAT)(about 90-95 percent of cases). The actual pathogenesis of DAT is unknown, and the relationship to so-called "normal aging" is unclear: there is probably a multifactorial spectrum between "normal" neurodegeneration (slower) and abnormal neurodegeneration (faster). The most common associated brain changes of DAT are 1) greater-than-expected for age neuronal loss, especially in the hippocampus and the basal forebrain; 2) greater-than-expected for age occurrence of neuritic plaques, excessive amyloid deposition in plaques and blood vessels, and neurofibrillary tangles with hyperphosphorylated tau protein; 3) greater-than-expected for age synaptic loss; and 4) greater-than-expected for age decline in cholinergic innervation.

A typical case of DAT will begin with gradual onset of forgetfulness, word finding difficulty, or both. However, there is great variation in initial symptomatology. Most patients will have a disproportionate problem with short term memory versus long-term, but others will have trouble even with old memories. Most patients have a subtle aphasia similar to a "transcortical sensory aphasia" with more trouble comprehending than speaking and preserved repetition, but any language problem may occur, and language may predominate (primary progressive aphasia). Most patients will have preserved perception, but a few will have a marked loss of ability to interpret what they see (posterior cortical dysfunction). Noncognitive behavior changes may be subtle, or may dominate the clinical picture, especially with depression early in the course and agitation or psychosis further along.

One popular set of criteria for DAT is offered in the Diagnostic and Statistical Manual of Mental Disorders (1994). Again, one of the major conundrums of neurobehavior is that we have yet to figure out how to draw a conceptually meaningful line between DAT and normal aging. But for practical purposes, we see a group of patients with somewhat greater-than-expected for age cognitive decline rapidly increasing in frequency after age 65: about 5 percent are affected at age 70, 15 percent at age 75, 25 percent at age 80, and 35–50 percent at age 85. DAT is the most common cause of dementia, accounting for about 40-70 percent of cases. Apart from the pure genetic forms, risk factors for DAT include age, Apolipoprotein e4 genotype, traumatic brain injury, and family history of DAT. About 6–8 years on average elapse between diagnosis and death. Death is usually from infectious complications, especially pulmonary disease. DAT is currently the fourth most common cause of adult death, after heart disease, cancer, and stroke. However, the prevalence of DAT is increasing rapidly with the aging of the U.S. population. About 4 million people have DAT now, and the disease is estimate to cost the U.S. 100 billion dollars a year. As many as 17 million may be diagnosable by the year 2030, at which point it may even become the leading cause of death.

We clinically diagnose DAT when there is a dementia without any other apparent cause. That is, despite the emerging availability of somewhat specific markers via functional neuroimaging and CSF (and soon serum), in practice DAT remains a diagnosis of exclusion. Note that apolipoprotein E4 genotype is not recommended as either a screen for risk of DAT or a diagnostic test, since it is neither sensitive nor specific.

There is a great deal of interest in prevention of DAT. Estrogen replacement therapy decreases the risk of DAT in postmenopausal women and there is preliminary evidence that estrogen-based oral contraceptives do the same. Nonsteroidal anti-inflammatories also cut the risk, but carry their own risk of death due to gastrointestinal bleeding. High dose vitamin E (2000 IU/D) has been shown to decrease DAT symptoms, but the long term hemorrhagic risk of this dose is currently unclear, and the more conventional 200–400 I.U./d may be sufficient for prophylaxis.

Management of DAT is three-pronged: 1) ameliorate the cognitive disorder, 2) treat the noncognitive behavioral features, and 3) address the distress of the caregiver and the impact on the social milieu. The cholinesterase inhibitor tacrine (Cognex®) definitely has some efficacy in slowing the rate of decline in mild to moderate DAT, but the newer agent Aricept® is easier to use because it is a once-a-day drug with minimal liver toxicity. A host of cholinergic and other agents are on the verge of FDA approval. The noncognitive behavioral features include depression in 10 to 30 percent, agitation in 10 to 40 percent, and psychosis in 5 to 25 percent. There is recent evidence that DAT-associated depression responds to antidepressants. Treatment of agitation is often a trial-and-error process, balancing cognitive and sedative side effects against efficacy. While haloperidol and similar high-potency, low-dose neuroleptics have been very commonly used with a slight to moderate benefit, there is encouraging evidence that carbamazepine may be equally helpful with less side effects. When cholinesterase inhibitor treatment is used for treatment of cognition, some patients exhibit more and some less agitation. Psychosis may respond to neuroleptics, although, again, drugs such as haloperidol are preferable to drugs such as chlorpromazine due to the lessened anticholinergic side effects.

The impact of DAT on the family is enormous. In fact, the patient may be increasingly oblivious to the situation, as the caregivers are increasingly distressed. Referral to educational and support groups, contact with the Alzheimer's association, social work consultation, legal advice regarding assets and durable powers of attorney, and careful discussion of either temporary respite care or lasting institutionalization are necessary as the disease progresses. The cost of institutional care, about $40,000/year, is double that of care at home, which is a compelling argument for therapies that can prolong the period of relative independence prior to placement, for expanding home-based medical support systems for patients and caregivers, and possibly for long-term care insurance.

Dementia with Lewy Bodies (DLB) has recently come to be considered a frequent cause of dementia because about 25 to 33 percent of DAT brains are found to have some cortical Lewy Bodies—oval intraneuronal inclusions. (Actually, using the term "Lewy bodies" is controversial because the inclusions seen in cortical neurons in DLB don't look exactly like the classic Lewy bodies found in the brainstem of Parkinson's patients). In an effort to shed light on the controversy concerning DAT, Parkinson's disease, and DLB, McKeith et al published a consensus report in 1996. DLB is thought to be clinically distinguishable from DAT by:

1. Earlier and more disabling parkinsonian extrapyramidal symptoms, such as rigidity, bradykinesia, and tremor.
2. More likelihood of fluctuations in clinical course.
3. High frequency of visual hallucinations.

DLB is clearly similar to Parkinson's disease with dementia, but Parkinson's patients usually exhibit extrapyramidal signs at least several years before dementia. DLB probably constitutes the second-most prevalent cause of dementia, accounting for about 10–25 percent of cases.

Treatment of DLB is controversial. When this condition is suspected, the best approach is probably a combination of symptomatic treatment of extrapyramidal features with anti-parkinsonian medications and cholinergic therapy for the cognitive dysfunction. However, since the risk of psychosis is high, and is increased with the very dopaminergic agents that might help the parkinsonian symptoms, there is a need for carefully balancing dopamine-blocking neuroleptics with the anti-parkinsonian medications.

Parkinson's disease (PD) involves a loss of dopamine-producing neurons in the substantia nigra in the midbrain. There is recent evidence that people born with a deficiency of mitochondrial complex I have a higher risk of developing PD. There is also evidence that exposure to pesticides may increase the risk of PD, and that drinking coffee is perhaps associated with a reduced risk. PD is conventionally expected to be associated with dementia in about 40 percent of cases, usually after at least several years of progressive motor changes. However, some authorities have suggested that dementia occurs in virtually all persons with Parkinson's disease, if followed through the advanced stages. The classical presentation is one of so-called subcortical dementias, with less aphasia and more psychomotor retardation. Yet, in practice, there is a good deal of overlap between the clinical features of dementia in PD and in DAT. Dopaminergic agents may somewhat delay or ameliorate the dementia, but pose the risks of 1) fluctuating mental status as the drugs cycle through peak and trough effects, and 2) delirium with psychotic features due to dopamine excess. New options to treat PD include several types of ablative surgery, or the transplantation of fetal dopamine-producing cells.

Vascular disease can affect the brain in many ways, from chronic reduced perfusion due to stenotic carotid arteries to microangiopathy of intracranial vessels related to hypertension or diabetes. The definition for vascular dementia (VaD) is debatable; AMS may be associated with large or small strokes, one or many strokes, subcortical white matter changes attributed to ischemia, or a combination of these conditions. AMS may occur acutely in SAH, or may evolve over decades with the cumulative impact of ischemia. Further complicating the differentiation of vascular dementia from DAT: 1) changes in blood vessels significantly contribute to the clinical picture of DAT, and 2) DAT-type pathological changes correlate with (and may determine) the occurrence of dementia in cerebrovascular disease. Thus, there is an important overlap between so-called DAT and so-called VaD, and the current definitions of VaD are operational. For practical purposes, VaD is diagnosed when:

1. Dementia is present
2. Cerebrovascular pathology is present in areas of the brain thought to be important to cognition (e.g., the hemispheres of the brain, the basal ganglia, and the thalamus), as detected by history, neurological signs, and/or laboratory tests.
3. There is evidence of a relationship between the onset or progression of dementia and the occurrence of cerebrovascular pathology.

This last is the most difficult criterion to fulfill. We are often confronted with a well-established dementia and signs of past strokes. It is of course speculative to attribute the dementia to the past strokes, since we cannot rule out the co-occurrence of DAT-type brain pathology, or determine how much of the AMS is caused by one or the other. A "step-wise progression" has sometimes been considered a hallmark of VaD, but DAT also frequently progresses with drops in cognition followed by apparent plateaus. Also, the introduction of MRI scanning has revealed a high frequency of subcortical white matter changes in elderly patients, with or without dementia. It was formerly popular to use such instruments as the Hachinski scale to diagnose VaD (a scale assigning a point value to various risk factors), reasoning that increased stroke risk factors could be diagnostic. But there is only a rough correlation between such risk factors and a definitive accounting for the etiology of cognitive loss in a given patient, so that many patients have probably been inappropriately diagnosed with VaD using such scales. There seems to be a threshold effect, such that cumulative microvascular changes actually have little impact on cognition until they are quite extensive, whereas a single small stroke in a strategic place can produce immediate dementia. In practical terms, therefore, VaD should be diagnosed when cerebrovascular disease predominates in the clinical picture and can be meaningfully temporally related to the course of the dementia. Mixed dementia (DAT/VaD) is diagnosed when cerebrovascular disease is apparently related to AMS, but does not fully account for the AMS. (For example, a patient has a stroke followed by marked decline, then some recovery, then a much slower, gradual decline, suggesting the superposition of DAT on VaD).

Management of VaD is, at present, largely directed to preventing further cerebrovascular disease by controlling blood pressure and glucose intolerance, judicious choice of antiplatelet or anticoagulant medications, and selective endarterectomy for high grade carotid stenosis. It is important to note that an element of DAT may underlie the deterioration in some patients diagnosed with VaD, and since new treatments for DAT are usually well tolerated, it may be appropriate to offer this option to VaD patients. There is preliminary evidence that cell transplant therapy may be beneficial after stroke, which raises the possibility that such therapy might also reduce the symptoms of VaD.

Less Common Dementias

Other neurodegenerative diseases. Pick's disease is a focal atrophy of the prefrontal and temporal cortex with disproportionate behavioral changes, especially a depressed, apathetic or disinhibited mental status. Diagnosis is sometimes aided by the presence on brain scanning of gyri so thin they resemble the blade of a knife ("knife-edge atrophy"). Frontotemporal lobar dementia (FTLD) is coming to be recognized as a better term that encompasses the spectrum of neurodegenerative conditions (including Pick's) in which the main problem is a loss of structure and/or function of the frontal and temporal lobes (unlike Alzheimer's, in which the major problem is temporoparietal dysfunction). Although there remains some controversy about whether FTLD truly represents a distinct entity, new consensus criteria were published in 1998. These criteria recognize three types of FTLD: 1) "frontotemporal dementia," marked by a deterioration in social conduct, 2) "progressive nonfluent aphasia," with a decline in verbal production, and 3) "semantic aphasia with associative agnosia," marked by fluent but empty speech and impaired recognition of objects. One hint FTLD represents a unique entity is the discovery of familial forms linked to mutations on Chromosome 17, and perhaps to Chromosome 3. Huntington's disease (HD) occurs both as a autosomal dominant inherited condition and as a sporadic mutation. HD is one form of a large group of problems due to abnormal lengthening of repeating sequences of trinucleotides in the genome. HD is specifically due to expansion of CAG repeats on Chromosome 4. Clinically we expected the writhing movement disorder of choreoathetosis mixed with various degrees of cognitive impairment, irritability, depression and psychosis. Neuroleptics may provide limited relief.

Metabolic disorders. Hypothyroidism is most common cause of progressive metabolic dementia, and may not be accompanied by obvious changes on the physical examination such as goiter, skin changes, or deep tendon reflex abnormalities. However, there is almost always some psychomotor retardation. Diagnosis is usually made by the thyroid function tests, but borderline abnormalities are of uncertain significance. Restoration of the euthyroid state often reduces cognitive impairments. Hypoglycemia is usually considered in the acute setting as the most common metabolic cause of delirium, but there is evidence that hypoglycemic episodes contribute to long-term cognitive loss—yet another reason for careful glucose control.

Vitamin abnormalities. B12 deficiency can appear as the full syndrome of subacute combined degeneration with both pyramidal signs and loss of posterior column function expressed as decreased proprioception and vibration sense in the feet. It can also appear as isolated dementia with minimal neurologic changes. We cannot assume that megaloblastic anemia will occur in all cases. Treatment is 1000 microgram IM per month of B12, and ameliorates the dementia in some cases. Folate deficiency is drawing increased attention as a possible cause of dementia. Thiamine deficiency causing a persistent Korsakoff's syndrome with marked amnesia, agitation, and sometimes psychosis is a possible result of alcohol abuse. The increased popularity of megadose vitamins has led to increased frequency of vitamin A toxicity , with AMS with or without an accompanying syndrome of "benign" intracranial hypertension.

Normal pressure hydrocephalus (NPH), consists of disproportionate ventriculomegaly as compared with hemispheric atrophy, and the clinical triad of dementia, incontinence, and slowed gait. NPH is something of a misnomer, since there are in fact waves of IICP. diagnosis is pursued when the classic clinical pattern is accompanied by a persuasive CT or MR scans, and usually consists of introducing radio-opaque dye into the lumbar subarachnoid space and watch to see over 48 hours whether it is taken up over the convexity (normal) or hung up in the ventrical (abnormal). A trial of removing CSF may also help diagnostically. Management may consist of ventriculoperitoneal shunting, although the selection criteria for surgical cases is controversial. The use of acetozolamide to reduce CSF production is also unproven.

Traumatic brain injury, especially involving subdural hematoma, can present insidiously with no report of head trauma. Diagnosis is usually made by neuroimaging. Neurosurgical consultation may help decide the risk/benefit ratio of evacuation. Slow spontaneous resorption over months may lead to reduced symptoms.

Obstructive sleep apnea occurs in about 4 percent of men and 2 percent of women between the age of 30 and 60. Nighttime snoring, morning headache, excessive daytime drowsiness, and subtle cognitive decline are markers. Newly approved minor surgical techniques may relive the symptoms.

The AIDS-dementia complex can occur with or without other sings of opportunistic infectious or neoplastic CNS involvement. Varying degrees of progressive cognitive impairment, sometimes admixed with a mild confusional state, may be recognized. There is some controversy about the efficacy of triple drug therapy, including protease inhibitors, for reducing the symptoms once they have become overt.

Neurosyphilis has increased in frequency with the emergence of HIV infections, as an opportunistic infection. The serum VDRL and RPR may produce more false positive results than the MHATP. Any positive serology mandates CSF examination for VDRL—a sign of active infection. Penicillin treatment is necessary to arrest neurovascular progression, but may or may not help the cognitive state.

Creutzfeldt-Jakob disease is apparently due to an infectious protein called a prion, and is related to or perhaps equivalent to bovine encephalitis (mad cow disease). It was recently found that an abnormal prion protein may start a cascade in the brain, serving as a template to convert other prion proteins to the abnormal form that makes neurons vulnerable to oxidative damage. Transmission is usually associated with direct contact with the virus, i.e., in neuropathologists or recipients of corneal transplants or infected growth hormone, but the agent apparently can also be transmitted by ingestion of infected meat. This rare dementia is associated with rapidly progressive AMS, myoclonic jerks, periodic complexes on EEG, with death in less than two years. There is no treatment.

The Diagnostic Work-Up for Dementia
Diagnostic work-up for dementia includes neurologic history and examination, complete blood count, serum electrolytes (including calcium), glucose, BUN/creatinine, liver function tests, thyroid function tests (FTI and TSH), serum vitamin B12 level and syphilis serology. In some clinical circumstances it may also be helpful to order a serum folate level, HIV testing, chest x-ray, urinalysis, 24-hour urine for heavy metals, toxicology screen, or a neuroimaging study.

Structural neuroimaging using either a noncontrast CT or MR scan is recommended in the routine initial evaluation of dementia patients according to the current AAN practice parameters.There is evidence (Chui and Zhang, 1997)that without a scan, a certain number of cases will be misdiagnosed as DAT who have subdural hematomas, tumors, infections, hydrocephalus, or stroke. This obliges the physician to develop a threshold for when to order these tests. In general, the presence of atypical symptoms, headache, fluctuating or rapidly progressive course, elementary neurological signs, or risk factors for tumor, stroke, or infection should lower the threshold for ordering such a scan. MRI scans have an advantage, since 1) they are more sensitive to any cerebrovascular process (excepting SAH), 2) they will show brainstem lesions invisible on CT, and 3) many MR scanners can get a simultaneous MR angiogram at a modest additional cost. However, patients with AMS may fail to tolerate the 35 minutes supine in a tunnel, giving CT's some advantage. As fast MR's become more widely used, this will be less of an issue.

Neuropsychological testing is optional. It is mostly useful when attempting to detect subtle changes, to discriminate between depression and dementia, and in cases with a preexisting cognitive deficit, e.g., developmental delay. It may also help for ascertaining level of disability, confirming capacity to participate in decisions, or designing rehabilitation for reversible dementias.

Lumbar puncture is not necessary as a routine part of the dementia work-up. However, it should be performed when new AMS accompanies fever, or (as per AAN practice parameters) in cases of metastatic cancer, suspicion of CNS infection, reactive syphilis serology, communicating hydrocephalus, dementia under age 55, rapid or atypical progression, immunosuppression or suspicion of CNS vasculitis.

The principle advantages of EEG are to help distinguish confusion from dementia in equivocal cases, to assist when seizure is known or suspected, and to help rule out the uncommon Creutzfeldt-Jakob disease. EEG is more useful in the diagnosis of confusion than dementia, and will be considered below.

DELIRIUM: DIFFERENTIAL DIAGNOSIS

Approaching the diagnosis of an acute or subacute confusional state (delirium) is often a matter of familiarity with the odds. Drugs effects and metabolic disorders - including hypoxia - are most common. Traumatic brain injury, systemic or neuroinfections, space-occupying lesions, cardio- or cerebrovascular disorders, and epilepsy are also frequent.

Drugs and Confusion

Simply in terms of frequency, drug intoxication and drug withdrawal may be the most frequent cause of acute or subacute AMS, followed closely by metabolic disorders. The number of prescribed and nonprescribed drugs that can cause confusion, either in therapeutic or toxic amounts, is dazzling (seeTable 4 for a partial list). We will discuss the most commonly offending agents roughly in order of frequency, and mention several interesting and possibly insufficiently recognized syndromes.

Allsedative hypnotics and opioid analgesics can produce delirium, with the most common offenders being benzodiazepines prescribed for anxiety or insomnia; benzodiazepine withdrawal can also produce an encephalopathy related, in part, to sleep disturbance. Benadryl (and other sedating antihistamines) is very commonly prescribed as a sleep aid, and may be a frequent cause of delirium or exacerbation of DAT due to its anticholinergic effects. Haloperidol is the most commonly prescribed neuroleptic, but it may also be the most commonly prescribed sedative in the elderly. This drug is a very common cause of delirium as well as parkinsonian motor impairments and, less commonly, tardive dyskinesia (to which elderly women seem especially vulnerable.) Similarly, compazine or metochlopramide may be prescribed for GI indications, but their neuroleptic and anticholinergic impacts make them common culprits producing AMS. Phenergan® (promethazine) is a weaker dopamine blocker, but is still a frequent cause of confusion (see neuroleptic malignant syndrome, below). A particular concern among the analgesics is meperidine HCl (Demerol®), which is associated with convulsions. This drug should never be combined with monoamine oxidase inhibitors since the combination greatly increases the chance of convulsions or death.

It is obvious that anticholinergic agents directly mimic and exacerbate the cholinergic deficit of DAT, but is less obvious that the tricyclic antidepressants imipramine (Tofranil®) and amitriptyline (Elavil®) are notoriously common causes of confusion due to their potent anticholinergic side effects. In this era when so many less toxic, equally effective agents are available, there is little reason to prescribe the most toxic agent, amitriptyline, to any patient. Nortriptyline is perhaps the best tricyclic choice, but the serotonin reuptake inhibitors are even less toxic and are probably better choices for treating depression, especially in the elderly.

Cimetidine- one of the most commonly prescribed gastrointestinal agents - causes confusion in roughly 10 percent of elderly patients. Other H2 receptor antagonists may have less CNS toxicity.

Among the cardiovascular agents, perhaps the most common offenders are the lipophilic beta-blockers such as propranolol, which produce depression - sometimes quite profound—in about 5 to 10 percent of patients. Amiodarone (Cordarone®) is well known for its potentially fatal pulmonary toxicities, but it also very commonly produces neurological changes (in up to 40 percent of patients) including confusion, involuntary movements, and gait disturbances. Calcium channel blockers , especially nifedipine, produce nervousness in about 5 to 8 percent of patients, and abnormal dreams or psychosis in about 1 percent. Digitalis preparations, of course, may cause confusion, psychosis, or apathy even at borderline elevated levels; this syndrome is sometimes anticipated by a yellowing of vision. The alpha-1-selective adrenergic blocker doxazosin can produce somnolence. Less well known are the anticholinergic effects producing confusion in patients taking the type 1 antiarrythmic disopyramide phosphate (Norpace® Methyldopate HCl (Aldomet®), is an infrequently used antihypertensive these days, in part because it can produce depression, psychosis, agitation, or confusion, sometimes associated with parkinsonism.

Among the most prescribed drugs are selective serotonin reuptake inhibitors (SSRIs). SRIs, especially those with shorter half-lives such as sertraline, but also the long-half life fluoxetine, have recently been shown to produce a withdrawal syndrome if stopped abruptly, with anxiety, agitation, or confusion as well as dizziness. SRIs have also recently been associated with the syndrome of inappropriate ADH secretion (SIADH), in which hyponatremia often leads to AMS and seizures.

Anti-inflammatory drugs, especially corticosteroids, are well known to produce AMS, but the common expectation of mania is incorrect. Depression with various degrees of delirium and psychosis is more common. With the increasing over-the-counter use of nonsteroidal anti-inflammatories such as ibuprofen, there have been increasing reports of AMS, although it is not clear to what extent these are idiosyncratic reactions to therapeutic dose as opposed to toxic encephalopathies. Paradoxically, NSAIDs may help to prevent DAT and many people might be self-prescribing them for this reason, which could increase the frequency of reported AMS. Aspirin toxicity, of course, can cause tinnitus and delirium.

Drugs with stimulant or sympathomimetic properties such as aminophylline and theophylline notoriously case confusion at toxic levels, less frequently in the therapeutic range. Ephedrine, phenylephrine and phenylpropanolamine are popular components of over-the-counter cold remedies that also can cause marked behavioral change in susceptible people. Ephedra-containing Chinese herbal preparations are an increasing cause of concern for the same reason.

Themuscle relaxant baclofen (Lioresal®) is a frequent cause of severe encephalopathy. Either by increasing the dose too rapidly, or by precipitous withdrawal, this agent can produce confusion or even coma, accompanied by autonomic instability that sometimes requires ICU monitoring. Carisprodol (found in Soma Compound® is another muscle relaxant causing confusion.

Among frequently prescribed antimicrobials, sulfamethoxasole occasionally produces delirium or even a picture of aseptic meningitis; change in mental status also accompanies the occasionally induced toxic epidermal necrolysis or hepatic necrosis. In addition to their well-known ototoxicity, aminoglycosides can produce an encephalopathy. Tetracycline occasionally causes symptomatic increased intracranial pressure. Ticarcillin can produce convulsions, especially in patients with reduced renal clearance.

Antineoplastic drugs can cause various syndromes of AMS. Aminoglutethimide causes drowsiness in about one third of patients. Asparginase can cause not only depression or confusion directly, but can produce a hyperglycemic hyperosmolar state with attendant encephalopathy. 5-Flurouracil is often associated with an acute syndrome of cerebellar dysfunction that may masquerade as intoxication. Methotrexate, even when used systemically for rheumatoid arthritis, can produce either a transient delirium or a toxic and persistent leukoencephalopthy (white matter degeneration). While the vinca alkaloids are best known for producing neuropathy, they can also cause depression.

A problem of unclear dimensions that is likely to increase in the near future is the psychological impact of cholesterol-lowering agents. Treatment with these drugs has been tentatively associated with increased aggression and possibly with suicide.

The neuroleptic malignant syndrome , consisting of subacute delirium associated with fever, tremor, and increased CPK, is most frequently associated with high dose neuroleptic prescription —or rapid increase in dose. However, this problem has actually been reported in patients after a single low dose. The diagnosis is frequently delayed because the syndrome goes unrecognized, or is misinterpreted as increased agitation in a psychotic patient, leading to the counterproductive escalation in neuroleptic dose.

Drugs of abuse are also obvious causes of encephalopathy, either as a toxic or withdrawal effect. The risk of delirium tremens probably peaks about 48 hours after alcohol withdrawal, and may be preceded by the more benign minor withdrawal syndrome of nervousness, tremor, and confusion. Stimulants such as cocaine, amphetamines, and even (rarely) caffeine can cause irritability, agitation, mania, psychosis, or delirium. Hallucinogens may cause not only hallucinations, but also a general delirium. Phencyclidine is known for the dramatic agitated psychoses it sometimes produces, but it's more likely to simply produce delirium, possibly with autonomic instability. Inhalants including glues and nitrous oxide can produce profound confusion and may not be identified by a toxicology screen.

General Medical Causes of Confusion

In the general medical realm, fever by itself is perhaps one of the most common causes of confusion.Sepsis invariably alters mental status, even with no evidence of CNS infection, although the reason for this is unclear. Delirium occurs with any process that decreases delivery of oxygenated blood to the brain. A wide spectrum of conditions may produce this effect. Hypoxemia or hypoperfusion, low-output cardiac failure or valvular abnormalities, vertebrobasilar insufficiency, anemia sufficient to lower the oxygen carriage capacity (Hct<24), COPD, emphysema, pneumonia or reactive airway disease that lowers the PO2, or carbon monoxide exposure - all can cause delirium. Hypoglycemia is one of the most common causes of delirium, especially among diabetics, who are also vulnerable to hyperosmolar hyperglycemic encephalopathy. Specific neurotoxicity is produced in uremic and hepatic encephalopathies, which sometimes present with less than markedly abnormal blood chemistries. Even dehydration itself, particularly in the elderly, can cause AMS apparently out of proportion to its severity as measured by the BUN/creatinine ratio. Hyponatremia causes encephalopathy depending to some extent on the rate of change, but almost all patients with NA of 120 or less will have some cognitive changes. SIADH is one well known cause of hyponatremia, but water intoxication- simply due to excessive drinking of water in demented or psychotic patients - is another cause that sometimes adds delirium to the underlying process. Endocrinopathies other than diabetes may also cause delirium. The effects of hypothyroidism (and less often hyperthyroidism) may appear as a gradual onset dementia-like syndrome, but may also cause acute delirium, especially on the case of thyrotoxicosis. Either adrenal insufficiency or hyperactivity may produce delirium.

Less common medical causes of delirium include hypokalemia, hypercalcemia, and hypomagnesemia. Parathyroid abnormalities, especially elevated PTH, may be one cause of so-called dialysis dementia. Another cause, now rare, is aluminum intoxication. More often, dialysis dementia is a multifactorial syndrome related to rapid shifts of both fluid and many electrolytes. Extracranial cancers may produce paraneoplastic encephalopathies by generating antibodies to cerebral tissue, such as anti-Purkinje cell antibodies, by producing carcinomatous meningitis (most often associated with small cell lung cancer) or via limbic encephalitis—a rare inflammation of the brainstem and medial temporal tissue.
Neuropsychiatric Causes of Confusion

Specific neuropsychiatric syndromes that commonly cause confusion include simple sleep deprivation (with or without sleep apnea), as well as other REM and non-REM sleep disorders. These are described in more detail in the chapter on sleep. Migraine headache, thiamine deficiency causing Wernicke's encephalopathy,subdural hematoma, andneuroinfection are other frequent causes of confusion. Diagnosis of subdural hematoma is often unfortunately delayed because the head trauma was not reported—or even recalled—by the patient, and no neuroimage was obtained. Similarly, Wernicke's encephalopathy is sometimes missed unless the full syndrome appears: delirium, gait disorder, and abnormal eye movements.

While bacterial meningitis is often quickly suspected due to fever or characteristic skin changes, perhaps the most frequently delayed diagnosis of a treatable condition among neuroinfections is herpetic encephalitis. There may be no fever, no history of exposure, and a benign neuroimage. This diagnosis requires a high index of suspicion in an otherwise unexplained confusion. The EEG may show distinctive periodic complexes, but it is much better to go right to the CSF examination and initiate acyclovir treatment as soon as this condition is clinically suspected. Other infectious encephalitides include tuberculosis, especially to be suspected in inner-city dwellers and recent immigrants from endemic areas, and fungal and protozoal infections. All of these may escape detection by not causing fever, and, again, CSF examination is more or less essential if confusion has not been diagnosed by blood tests or a neuroimage.

Cerebrovascular disease may appear first as confusion. Hypertensive encephalopathy may be suspected early on the basis of vital signs. But the patient with SAH may have already lost the ability to report the headache that preceded their current agitated, staggering state. Hypertensive hemorrhage in the pons may show obvious brainstem signs, but in the thalamus or basal ganglia may be tough to localize on examination, particularly because the patient's confusion complicates careful testing. Stroke-in-evolution, or completed strokes in the right frontal lobe or occipital-parietal-temporal junction area may show little on neurologic examination but much on mental status testing. As we noted previously, given the rapid improvement in immediate post-stroke care, the faster we suspect this diagnosis and scan the patient, the better the outcome may be.

Traumatic brain injury is an obvious explanation for delirium when there has been a known loss of consciousness or superficial signs of injury are apparent. However, even mild trauma without loss of consciousness and indirect trauma, as in whiplash, can produce confusion for variable periods of time, and the related brain injury may be undetectable by CT or MR. (Note that 85 percent of those cases in which MR reveals abnormalities immediately after head trauma show no abnormality on CT). Children are often traumatized in bicycle accidents; the elderly often hit their heads in falls. Neither group will necessarily report the injury. Perhaps the best measure of traumatic brain injury severity is not the structural injury, or even the duration of unconsciousness, but the length of time it takes the patient to recover their continuous stream of consciousness, with no difficulty with orientation. This is called the duration of the post-traumatic amnesia (PTA), and requires careful sequential testing of mental status, even when the patient superficially seems recovered. Even minor trauma may cause deposition of beta-amyloid, so the transient delirium seen in traumatic brain injury may be followed, many years later, by an increased risk of dementia .

Any cause of increased intracranial pressure, in particular space-occupying lesions or hydrocephalus, may cause delirium. While tumors may be suspected due to asymmetry in the elementary examination, masses in certain areas (especially the right frontal lobe) may be clinically silent apart from confusion. Hydrocephalus may develop insidiously, e.g., following SAH when the patient seems to be doing well, and during the following week develops AMS because the re-uptake of CSF has been impaired by the blood remaining in the cranium. Similarly, a patient with neurocysticercosis, rapidly increasing in the US due to the influx of central and South American residents, may develop a ball-valve effect with intermittent AMS due to hydrocephalus above a cyst that obstructs the third ventricle.

Rarely, unrecognized complex partial or petit mal status epilepticus can cause delirium. While many cases will exhibit suggestive signs, such as purposeless movements or staring, others seem to be conscious. More commonly, an unrecognized postictal state after an unwitnessed seizure produces lasting confusion. In either case, an EEG is extremely helpful diagnostically.

Primary psychiatric conditions can produce a disorganization of thinking that appears indistinguishable from delirium. In fact, functional neuroimaging supports the hypothesis that severe psychiatric disturbances in fact disrupt brain blood flow and metabolism just as medical encephalopathies might. Depression with psychomotor retardation, mania with psychosis, and exacerbations of schizophrenia may give this appearance. Sometimes, the simple procedure of asking the patient to copy a figure can help distinguish these conditions; delirious patients often fail. Psychotic patients often succeed.

The Laboratory Work-Up of Confusion

The laboratory tests that may quickly help to diagnose a delirium are about the same as those proposed for the dementia evaluation. In fact, the major reason for most of the blood tests previously listed is to rule out toxic and metabolic causes of confusion. Given the higher stakes of prompt diagnosis, there are several tests we may be more eager to order in a confusional state. For instance, if the general medical examination hints at hypoxia or conditions that might alter acid/base balance, we are more likely to obtain a chest x-ray and an arterial blood gas. With the slightest suspicion, or especially when there is no suspicion, a toxicology screen should be ordered. In the same way, because the consequences of delayed diagnosis are grave, we are much more likely to perform a lumbar puncture. However, concern about precipitating herniation in a patient with IICP necessitates a CT scan before the LP. (Note: the risk of this complication is higher in children). EEG has limited utility in dementia assessment; however, it can be key in proving the presence of delirium. Most alterations in level of responsiveness will appear as slowing and disorganization of the EEG background; almost all acute neuroinfections and most drug-induced confusion alter the EEG. In contrast, most dementias show little EEG change until advanced stages, and most primary psychiatric conditions, including somatoform illness, will produce a largely normal EEG.

Returning to the Examining Room

Our examination of Mr. Johnson is done. We don't have a diagnosis, but we have a plan. Our 83-year-old patient, and any patient with hints of an acute or subacute onset of confusion (with or without evidence of CNS dysfunction on the elementary examination) needs a laboratory work-up right away. Looking atTable 5(see expanded version in Appendix B), we might check off the processes that are more or less likely to produce this syndrome (e.g., large hemispheric stroke is less likely given the absence of motor asymmetry, neurodegenerative disease is more likely given the global cognitive dysfunction). But we need not feel forced to commit to a diagnosis at this point, since the differential remains huge in cases that tread the border between delirium and dementia, and most clinical settings offer access to extraordinarily valuable and relatively cost-efficient laboratory screening. We order some stat labs.

Forty minutes later, we have written prescriptions for our young asthmatic, seen two other patients briefly, and are somewhat more confident that we know what is going on: Mr. Johnson's sodium came back at 114 mEq/l. All the other tests are normal.

We ask his daughter, "By chance, have you noticed whether he's been drinking a lot of water?" She hasn't noticed. But the patient overhears the question, "Yeah, so what's it to you?" He's not really irritable; he says it with a smile. "You gotta', you know, you gotta wash out the poisons, know what I mean?"

We still aren't certain. But here's one scenario that could account for this clinical picture: an insidious loss of cognitive ability perhaps led Mr. Johnson to abandon his favorite pastime of gardening several years ago. A garrulous fellow, competent in his own environment, he can talk a good game in casual conversation. So, like most early Alzheimer's patients, he hasn't seemed demented to his family on their weekly visits. Not until his Alzheimer's progressed to the point where he exaggerated his lifelong habit of drinking two glasses of water a day—to "wash out the poisons"—has his dementia become overt: forgetting that he already drank his quota, he's been repeating the procedure six times a day. His resulting water intoxication tips him over the edge into overt cognitive impairment with some features of dementia and some of delirium - a typical "beclouded" dementia. We'll admit him. It's 5:30. We also call home, to slightly rearrange the evening.

Appendix A: Mini-Mental State Examination (MMSE)

Orientation = 10 points

Time: Ask for the date, then ask for parts omitted:

1. What year is this?
2. What day of week?
3. What month?
4. What date?
5. What season is it? (correct if within 1 mo.)

Place: Ask in turn:

6. Can you tell me the name of this place?
7. What City is this?
8. What State is this?
9. What County is this?
10. What floor are we on?

Memory = 6 points

Registration: Ask the patient if you may test his memory; then say the names of three unrelated objects. After you have said all three, ask him to repeat them. This first repetition determines the score, one point for each word correctly repeated, but keep saying them (up to six trials) until he can say all three; e.g.,

11. Apple,
12. Table,
13. and Penny

Recall (Note: the original MMSE does not specify a delay, but we advise a three minute delay): Ask the patient if he can recall the three words you previously asked him to remember; e.g.,

14. Apple
15. Table
16. and Penny

Attention/Concentration = 5 points
Serial 7's, subtracted back from 100. Stop after 5 answers. (One point for each correct subtraction.) Alternatively, spell WORLD backwards (One point for each letter in correct sequence)

= 17-21

Language = 8 points

Naming to confrontation

Show the patient each object and ask him what it is.

23. Watch
24. Pencil

Repetition:

22. Repeat the following: "No ifs, ands, or buts"

Following three-stage command
(Which should be presented before subject touches the paper):

27. Please take this paper in your right hand
28. Fold it in half
29. And put it in your lap
    (Note: the original MMSE specifies the right hand; however, the authors have found that reaching for a proffered object with the dominant hand is an overlearned response; a truer test of comprehension may be to specify the left hand)

Reading

25. Subject is shown a card that states in large print: "Close your eyes;" Subject is asked to "read obey the following."

Writing

26. Write a sentence, (1 point for a complete sentence, even if handwriting and spelling are poor.)

Construction = 1 point

30. Correctly copies a drawing of two interlocking pentagons

Total = 30 points
Typical cut-off for "dementia" = < 24 points, but this is highly education-dependent.

Folstein M, Folstein S, McHugh PR. Mini-mental state: a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12:189-198.

Appendix B: Expanded Table 5. Causes of Dementia and Confusion