Dementia: Specific Therapies

Specific Therapies

The psychiatric care of patients with dementia involves a broad range of psychosocial treatments for the patient and his or her family, as already described. In addition, some patients may benefit from more specific psychosocial interventions. These more specific psychosocial treatments for dementia can be divided into four broad groups: behavior oriented, emotion oriented, cognition oriented, and stimulation oriented.

Although few of these treatments have been subjected to rigorous double-blind, randomized controlled trials, some are supported by research findings and practice. The studies were generally small, and many of the reports failed to fully characterize the intervention, the nature of the subjects’ dementia or their baseline status, or the posttreatment outcome. Nonetheless, a review of the literature reveals modest efficacy of such treatments (although the limited available follow-up data suggest that the benefits of most do not persist beyond the duration of the interventions).

  1. Goals
    While these treatments differ in philosophy, focus, and methods, they have the broadly overlapping goals of improving quality of life and maximizing function in the context of existing deficits. Many have as an additional goal the improvement of cognitive skills, mood, or behavior. They are discussed together here because of their overlapping goals and apparent nonspecificity of action.
  2. Types of psychotherapies/treatments and their efficacy
    1. Behavior-oriented approaches.
      Although there are limited data from formal assessments of these treatments, there is widespread agreement that behavioral approaches (56, 57) can be effective in lessening or abolishing problem behaviors (e.g., aggression, screaming, incontinence) (58). The first step is a careful description of the behavior in question, including where it occurs, when it occurs, and how often it occurs. The next step is an assessment of the specific antecedents and consequences of each problem behavior, which will often suggest specific strategies for intervention. Activities that consistently precede the problem behavior may be acting as precipitants and should be avoided whenever possible. If the activity is a necessary one, such as bathing, it may be helpful to decrease its frequency or alter the environment so that the negative consequences are minimized (e.g., switch bath time to allow a home health aid to supervise, or change the location of baths to decrease the impact of aggressive outbursts on family members or other patients). When multistep activities, such as dressing and eating, precipitate problem behaviors, such as aggression, it often helps to simplify them or break them into parts (e.g., using clothing with Velcro closures, serving several simple snacks instead of a large meal). Whatever the intervention, it is critical to match the level of demand on the patient with his or her current capacities, avoiding both infantilization and frustration, and to modify the environment insofar as possible to compensate for deficits and capitalize on the patient’s strengths.
    2. Emotion-oriented approaches.
      These interventions include supportive psychotherapy (59), reminiscence therapy (reviewed by Burnside and Haight [60]), validation therapy (61, 62), sensory integration (63), and simulated presence therapy (64).

Reminiscence therapy, which aims to stimulate memory and mood in the context of the patient’s life history, has been shown in three studies of “confused” elderly persons (65-67) to be associated with modest short-lived gains in mood, behavior, and cognition. In a single small study (68), the effects of validation therapy, which aims to restore self-worth and reduce stress by validating emotional ties to the past, on cognitive, functional, and mood measures were not significantly different from the effects of reality orientation or no intervention. Preliminary evidence from one small study (64) suggests that simulated presence therapy may be helpful in diminishing problem behaviors associated with social isolation. In another small study (63), sensory integration showed no difference from no intervention in effects on cognitive, behavioral, and functional measures. Supportive psychotherapy also falls under this rubric. It has received little or no formal assessment, but some clinicians find it useful in helping mildly impaired patients adjust to their illness.

  1. Cognition-oriented approaches.
    These techniques include reality orientation (reviewed by Powell-Proctor and Miller [69]) and skills training (70). The aim of these treatments is to redress cognitive deficits, often in a classroom setting. In a number of studies of both institutionalized and noninstitutionalized patients, reality orientation has produced modest transient improvement in verbal orientation (65, 71-80). Some studies have also demonstrated slight transient improvement in other measures of cognition, function, behavior, and social interaction. Of note, there have been case reports of anger, frustration, and depression precipitated by reality orientation (81). There is also some evidence for transient benefit from cognitive remediation and from skills (or memory) training (70, 82-86) but there have been reports of frustration in patients and depression in caregivers associated with this type of intervention (47). The slight improvements observed with some of these treatments have not lasted beyond the treatment sessions and thus do not appear to warrant the risk of adverse effects.
  2. Stimulation-oriented approaches.
    These treatments include activities or recreational therapies (e.g., crafts, games, pets) and art therapies (e.g., music, dance, art). They provide stimulation and enrichment and thus mobilize the patient’s available cognitive resources. There is some evidence that, while they are in use, these interventions decrease behavioral problems and improve mood (79, 87, 88). Although the data supporting efficacy are limited either by small number of subjects (79, 87) or multiple interventions (87), there is anecdotal and common sense support for their inclusion as part of the humane care of patients with dementia. Additional support for this approach comes from the work of Teri et al. (89, 90), who have developed a behavioral protocol for managing Alzheimer’s disease that includes a number of interventions. The core of this protocol is identifying and increasing the number of pleasant activities, which has been shown in preliminary studies to improve the mood of patients and caregivers alike.


  1. Side effects
    Short-term adverse emotional consequences have been reported with psychosocial treatments. This is especially true of the cognitively oriented treatments, during which frustration, catastrophic reactions, agitation, and depression have been reported. Thus, treatment regimens must be tailored to the cognitive abilities and frustration tolerance of each patient.
  2. Implementation
    Behavioral interventions have strong support in clinical practice and deserve careful trials with patients who have behavioral problems that are difficult to manage. Many stimulation treatments provide the kind of environmental stimulation that is recognized as part of the humane care of patients, and thus such treatments are often included in the care of patients with dementia. Beyond this, the choice of therapy is generally based on patient characteristics and preference, availability, and cost. For instance, some approaches are available only in institutional settings, such as nursing homes or day care centers, while others can be used at home. In many cases, several modalities will be selected at the same time. Because these treatments generally do not provide lasting effects, those that can be offered regularly may be the most practical and beneficial. These treatments are generally delivered daily or weekly.

Rates of short-term response to emotion-oriented treatments are consistent with modest efficacy on a wide variety of outcome measures, and thus these treatments may be helpful for some patients. Cognition-oriented treatments are not supported by efficacy data and also have the potential to produce adverse effects.

D. Somatic Treatments

The sections that follow describe medications (and for depression, ECT) used for the purpose of treating the cognitive and functional losses associated with dementia; psychosis, anxiety, and agitation; depression and apathy; and sleep disturbances. Although the sections are organized by these target symptoms, many medications have broader impact in actual practice.

  1. Special considerations for elderly and demented populations
    Certain principles underlie the pharmacologic treatment of elderly and demented patients. They will be discussed in more detail in the APA practice guideline for geriatric psychiatry (in preparation) and can be summarized as follows. First, it must be remembered that elderly individuals may have decreased renal clearance and slowed hepatic metabolism, so lower starting doses, smaller increases in dose, and longer intervals between increments must be used (this practice is sometimes referred to as “start low and go slow”). Because elderly individuals are more likely to have a variety of general medical problems and take multiple medications, one must be alert to general medical conditions and medication interactions that may further alter the serum binding, metabolism, and excretion of the medication. In addition, certain medication side effects pose particular problems for elderly and demented patients, so medications with these effects must be used especially judiciously. Anticholinergic side effects may be more burdensome in the elderly owing to coexisting cardiovascular disease, prostate or bladder disease, or other general medical conditions. Especially in demented elderly persons, these medications may also lead to worsening cognitive impairment, confusion, or even delirium (91). Elderly individuals have decreased vascular tone, are more likely to be taking other medications that cause orthostasis, and, especially if they are demented, are also more prone to falls and associated injuries. Medications associated with CNS sedation may worsen cognition, increase the risk of falls, and put patients with sleep apnea at risk of additional respiratory depression. Last, the elderly, especially those with Alzheimer’s or Parkinson’s disease, are especially susceptible to extrapyramidal side effects.

For all these reasons, medications should be used with considerable care. The use of multiple agents (sometimes referred to as “polypharmacy”) should be avoided if possible. However, as elderly demented individuals frequently manifest multiple behavioral symptoms that do not respond to psychosocial interventions, and multiple general medical problems as well, some patients benefit from the use of several medications at once.

  1. Treatments for cognitive and functional losses
    1. Goals.
      There are a number of psychoactive medications that are used for the purposes of restoring cognitive abilities, preventing further decline, and increasing functional status in patients with dementia. These include cholinesterase inhibitors (tacrine and donepezil); a-tocopherol (vitamin E); selegiline (deprenyl), approved for Parkinson’s disease but studied and used in demented populations; and ergoloid mesylates (Hydergine), which are approved for nonspecific cognitive decline. In addition, a number of other medications have been proposed for the treatment of cognitive decline, including NSAIDs, estrogen supplementation, melatonin, botanical agents (e.g., ginkgo biloba), and chelating agents. Many additional agents are currently being tested; for patients who have access to academic medical centers, participation in clinical trials is another option. Interventions for specific medical conditions, such as blood pressure control and use of aspirin to prevent further strokes, and prescription of L-dopa as a general treatment of Parkinson’s disease, are beyond the purview of this guideline.
    2. Cholinesterase inhibitors.
      In 1993 tacrine became the first agent approved specifically for the treatment of cognitive symptoms in Alzheimer’s disease. Tacrine is a reversible cholinesterase inhibitor and is thought to work by increasing the availability of intrasynaptic acetylcholine in the brains of Alzheimer’s disease patients. The medication may also have other actions. Donepezil, another reversible cholinesterase inhibitor, is now available for treatment of Alzheimer’s disease. Additional agents that increase cholinergic function are in development.

      1. Efficacy.
        The efficacy of tacrine in mild to moderate Alzheimer’s disease has been extensively studied. At least five double-blind, placebo-controlled trials with parallel group comparisons including a total of over 2,000 patients have been reported (92-96). Overall, these clinical trials consistently demonstrated differences between tacrine and placebo: approximately 30 to 40% of patients taking tacrine who were able to complete the trials showed modest improvements in cognitive and functional measures over study periods ranging from 6 to 30 weeks, compared to up to 10% of those taking placebo. Modest improvement in these studies corresponds to maintaining or improving function by an amount typically lost over 6 months in untreated groups of similar Alzheimer’s disease patients. Response appeared to be related to dose, at least in the largest clinical trial (94), in which patients who could tolerate 120 to 160 mg/day were more likely to respond. Only approximately 60% of the patients overall completed the tacrine trials even at moderate doses; 30% of subjects were dropped from these trials prior to completion because of elevation in hepatic transaminases, as specified in the protocols (i.e., more than three times the upper limit of normal, a lower threshold than the current prescribing guidelines, described in the section on implementation), and another 10% had to leave because of other adverse effects, mainly cholinergic effects (e.g., nausea and vomiting). The benefits and adverse effects of administration beyond 30 weeks are unknown. However, one observational study suggests that continued use of tacrine at doses above 120 mg/day was associated with delay in nursing home placement compared with patients who used daily doses below 120 mg (97). Anecdotal reports suggest that individuals who respond to the medication and then stop taking it may have a significant decline. The effects of tacrine on individuals with more severe or very mild Alzheimer’s disease or with other dementing illnesses have not been assessed.

The efficacy of donepezil has been reported in three trials (98; Aricept package insert). A 12-week double-blind, randomized parallel group trial (99) included 160 patients with mild to moderate Alzheimer’s disease randomized to receive placebo or 1, 3, or 5 mg/day of donepezil. Modest improvements in neuropsychological test results and clinicians’ impressions were reported for higher doses. In another 12-week double-blind, parallel group trial (Aricept package insert), approximately 450 patients were randomized to receive placebo, 5 mg/day, or 10 mg/day of donepezil (after 5 mg/day for 1 week). A third trial (Aricept package insert), this one for 24 weeks, involved 473 patients similarly randomized to placebo, 5 mg, or 10 mg. In both of the latter two trials, the treatment groups showed modest improvement in neuropsychological test performance, clinician’s impression of change, and Mini-Mental State examination scores with a trend for a somewhat greater response for the 10-mg dose during the course of the trial. Consistent with observations of tacrine, patients discontinued from donepezil after 12 or 24 weeks of treatment returned to the cognitive level of the placebo treated patients within 3-6 weeks. In long-term observations over 2 years, patients continued on doses of 5 mg or greater of donepezil overall maintained their performance at or above baseline for an average of 40 weeks and deteriorated less when compared to a historical comparison group (100).

  1. Side effects and toxicity.
    As would be expected with cholinesterase inhibitors, side effects associated with cholinergic excess, particularly nausea and vomiting, are common, but tend to be mild to moderate for both agents. Observed rates are on the order of 10%-20% of patients. Additional cholinergic side effects include bradycardia, which can be dangerous in individuals with cardiac conduction problems, and increased gastrointestinal acid, a particular concern in those with a history of ulcer and those taking NSAIDs. However, these effects appear to occur infrequently with these agents. In general, cholinergic effects tend to wane within 2 to 4 days, so if patients can tolerate unpleasant effects in the early days of treatment, they may be more comfortable later on.

A unique property of tacrine is direct medication-induced hepatocellular injury. Approximately 30% of patients develop significant (e.g., three times the upper limit of normal) but reversible and asymptomatic elevations in liver enzyme levels, and for 5%-10% of patients the medication must be stopped owing to more marked elevations (e.g., 10 times the upper limit of normal). However, perhaps 80% of patients who initially develop elevations in liver enzyme levels can be successfully rechallenged with more-gradual increases in dose, as described in the following section. The hepatotoxicity is more common in women and tends to occur about 6 to 8 weeks into treatment. It has thus far proved to be reversible with discontinuation of the medication (101). No additional toxicities with donepezil have been reported, but the experience with this agent is limited.

  1. Implementation.
    Given the evidence for modest improvement in some patients and the lack of established alternatives, it is appropriate to consider a trial of one of these agents for mildly or moderately impaired patients with Alzheimer’s disease for whom the medication is not contraindicated (e.g., in the case of tacrine, because of liver disease). Patients and their families should be apprised of the limited potential benefits and potential costs (including, in the case of tacrine, risk of hepatotoxicity) (102). Although the currently available data do not allow a direct comparison, they suggest similar degrees of efficacy for the two medications. However, donepezil has the advantage of greater ease of use because it can be given once instead of four times per day and does not require regular liver function tests. Thus, donepezil may prove preferable as a first-line treatment. However, accumulated data from clinical practice with typical Alzheimer’s disease patients, along with the outcome of additional clinical trials, will be critical in developing a more complete picture of donepezil’s efficacy and adverse effects. Because the efficacy of these two agents is modest, it is also appropriate to discuss alternative options, including vitamin E or possibly selegiline, psychosocial interventions, participation in a trial of an experimental treatment (if available locally), or no treatment.

For tacrine, the starting dose is currently 10 mg q.i.d. The dose may be increased by 10 mg q.i.d. (40 mg/day) every 6 weeks up to a maximum dose of 40 mg q.i.d.. The highest tolerated dose (up to 160 mg/day) should be administered, since cognitive improvement is more likely to occur at higher doses. If there is no improvement in clinical status after 3-6 months, most clinicians would stop the medication.

Because of hepatotoxicity, the patient’s baseline level of alanine aminotransferase (ALT) should be measured before tacrine treatment is begun; patients with elevations should not receive the medication. Once the medication is begun, ALT should be measured every 2 weeks for approximately 3 months after each dose increase, and once the dose has been stable for 3 months, ALT may be measured every 3 months. If the ALT level is three to five times the upper limit of normal, the dose should be decreased to the prior dose. A later repeat trial at the higher dose can often be accomplished without a significant increase in ALT. If it is five to 10 times the upper limit of normal, the medication should be temporarily discontinued, with a rechallenge considered after the ALT level returns to baseline. A later repeat trial can sometimes be accomplished without a significant increase in ALT. If the ALT level is more than 10 times the upper limit of normal, the medication should be discontinued. These recommendations are currently under review and may change in the near future. In addition, a trial of sustained-release tacrine is underway, so the frequency of administration may change and four-times-per-day dosing may no longer be necessary.

For donepezil, the currently recommended starting dose is 5 mg/day. After 1 week, the dose may be increased to 10 mg/day. The higher dose should be used if tolerated, as it is associated with greater efficacy; however, it has a greater tendency to cause cholinergic side effects. It should be noted that dosing and other aspects of administration may evolve as clinical experience with this medication accumulates.

  1. Vitamin E.
    There has been considerable interest in vitamin E (a-tocopherol) as a treatment for Alzheimer’s disease and other dementias because of its antioxidant properties. Vitamin E has been shown to slow nerve cell damage and death in animal models and cell culture (including damage associated with amyloid deposition, and thus possibly relevant to the development and progression of Alzheimer’s disease) (103-106).

    1. Efficacy.
      A single clinical trial has been conducted concerning vitamin E in Alzheimer’s disease (106). This placebo-controlled,double-blind, multicenter trial included 341 moderately impaired patients randomized to receive either 1000 IU b.i.d. of vitamin E alone, 5 mg b.i.d. of selegiline alone, both, or placebo, and found that vitamin E alone and selegiline alone were equivalently helpful in delaying the advent of a poor outcome (defined as death, institutionalization, or significant functional decline). Combined treatment performed somewhat worse than either agent alone, but the difference was not statistically significant. The benefit observed among individuals treated with vitamin E or selegiline alone was equivalent to approximately 7 months delay in reaching any of the endpoints designated as a poor functional outcome. It should be noted that there was no evidence of improvement in function compared to baseline, but only of decreased rates of functional decline on active treatment compared to on placebo. Despite the evidence for a better functional outcome in the treatment groups compared to the placebo group, all groups showed similar rates of cognitive decline during the 2-year study period. There are no data concerning the role of vitamin E in Alzheimer’s disease with mild or severe impairment, or in other dementing illnesses. There are also no data concerning the effect of vitamin E in combination with medications other than selegiline.
    2. Side effects and toxicity.
      Vitamin E has been widely used clinically, has been observed in many clinical trials for other indications, and is considered to have low toxicity. Doses between 200-3000 IU/day have been shown to be safe and well-tolerated in many studies (107). At high doses, it has sometimes been noted to worsen blood coagulation defects in patients with vitamin K deficiency (107).

In the trial described above (106), the vitamin E group showed an elevated rate of falls and syncope compared to placebo, but the difference was not statistically significant and did not lead to attrition from the study. Vitamin E has not been associated with this effect in trials for other indications (107).

  1. Implementation.
    On the basis of these data, vitamin E may be used in moderately impaired patients with Alzheimer’s disease in order to delay the progression of disease. Vitamin E has not been studied in Alzheimer’s disease with mild or severe impairment, but, given its lack of toxicity (and possible other health benefits), some physicians might consider the medication for patients at these stages of disease as well. Vitamin E has not been studied in combination with cholinesterase inhibitors, and there is no clinical experience with this combination. However, given vitamin E’s lack of medication interactions, it might be considered for use in combination with a cholinesterase inhibitor. Available evidence suggests that there is no benefit to giving vitamin E in combination with selegiline, since the combination treatment performed no better than either agent alone in the large clinical trial described above (106).

The efficacy data reported here are for a fairly high dose of 2000 IU/day. There are no data concerning other doses. Because of the association of vitamin E and worsening of coagulation defects in patients with vitamin K deficiency, vitamin E should be limited to conventional doses (200-800 IU/day) in this population.

  1. Selegiline.
    Selegiline (also known as l-deprenyl) is a selective MAO-B inhibitor licensed in the United States for the treatment of Parkinson’s disease. It is approved as a dementia medication in some European countries and is used by some clinicians in the United States for this indication. It has been suggested that selegiline may act as an antioxidant or neuroprotective agent and slow the progression of Alzheimer’s disease, although, because of its effects on catecholamine metabolism, it could also act in a variety of other ways (106).

    1. Efficacy.
      Selegiline has been studied in six double-blind, randomized clinical trials involving over 500 patients with dementia followed over periods ranging from 1 month to 2 years (106, 108-115). The largest and most methodologically rigorous of these trials was the comparison of selegiline and vitamin E to placebo, described in section III.D.2.c.1., in which selegiline appeared similar to vitamin E and both were superior to placebo in delaying poor outcome, and combined treatment appeared similar but somewhat worse than either agent alone.

The remaining five double-blind, randomized trials (108-115) were generally smaller, briefer (1-3 months), and had methodologic limitations. Nonetheless, all but one (110, 111) showed statistically significant or nearly significant improvements in a variety of measures across multiple domains, both cognitive and noncognitive. In addition there were five small, within-subject crossover studies of similar durations that supported a beneficial effect of selegiline (116-120).

  1. Side effects and toxicity.
    Selegiline’s principal side effect is orthostatic hypotension, which has been reported to interfere with some patients’ tolerance of the medication. However, this may be more common in patients with Parkinson’s disease, since it was not observed in the large Alzheimer’s disease trial reported above (106). The investigators reported that the agent was well-tolerated, and, although a somewhat higher rate of falls and syncope were reported with selegiline than with placebo, the differences were not statistically significant and did not lead to drop-out from the study.

Beyond these effects, selegiline is reported to be quite activating, which is helpful for some patients but may lead to anxiety and/or irritability in others (118). The 5- to 10-mg/day dose used in the treatment of dementia is relatively selective for MAO-B and does not fully inhibit MAO-A, so a tyramine-free diet and avoidance of sympathomimetic agents are not required. However, patients and caregivers should be warned about the symptoms of hypertensive crisis and the critical nature of the 10-mg dose ceiling. More critical, adverse effects of medication interactions, including changes in mental status, seizures, and even death have been observed with meperidine, SSRIs, and tricyclic antidepressants, although there are also reports of patients who have tolerated these combinations. Selegiline is generally considered contraindicated for patients who are taking any of these agents.

  1. Implementation.
    The recent large study confirms the efficacy of selegiline in delaying the progression of Alzheimer’s disease with moderate impairment. However, in the same study, vitamin E, which is generally less expensive and has a more favorable side effect profile and less potential for medication interactions, had approximately equal efficacy, and thus might appear preferable to selegiline. On the other hand, there is also modest support for cognitive and functional improvement with selegiline, rather than simply the prevention of functional decline, so a trial of selegiline might be considered, especially for those patients who cannot take cholinesterase inhibitors. There are limited data concerning the role of selegiline at other stages of Alzheimer’s disease, but it is possible that it would offer a benefit to either milder or more severely impaired patients. Last, if the side effects and medication interactions do not pose a problem, selegiline may be continued in patients whose families report a benefit. The standard dose of selegiline for dementia is 5-10 mg/day. In the largest trial (106), which was reported above, 10 mg was used.
  1. Ergoloid mesylates.
    1. Efficacy.
      A mixture of ergoloid mesylates, known by the trade name Hydergine, is currently marketed for the treatment of nonspecific cognitive impairment. It has been available for at least 40 years and has been studied in at least 150 clinical trials. Of these, seven were double-blind, placebo-controlled, randomized trials with a parallel group design involving a total of 297 patients with diagnoses consistent with Alzheimer’s disease (84, 121-126). A recent meta-analysis (116) suggested that there might have been improvements in some neuropsychological and behavioral measures, but the overall effect sizes were not statistically significant. There was a general impression that any improvement observed was in behavioral rather than cognitive measures. In seven trials involving a total of 140 patients with vascular dementia (123, 125, 127-131), there was somewhat more compelling evidence of modest improvement on neuropsychological and behavioral measures.
    2. Side effects and toxicity.
      Ergoloid mesylates occasionally cause mild nausea or gastrointestinal distress, but no significant side effects or toxicity have emerged during long-term use. However, the medication is contraindicated for patients with psychosis.
    3. Implementation.
      The questionable efficacy of ergoloid mesylates suggested by extensive study argues against routine use of this medication in the treatment of dementia. However, under some circumstances it may be appropriate to offer a trial of this agent for vascular dementia. For Alzheimer’s disease, a trial of a cholinesterase inhibitor, vitamin E, or selegiline is probably preferable, and for some patients participation in a clinical trial at an academic medical center may also be preferable. However, when these options are inappropriate or unsuccessful, a trial of ergoloid mesylates may be appropriate for patients with a strong interest in pharmacologic therapy. In addition, use of the medication may be safely continued for patients whose families report a benefit. The manufacturer’s recommended dose is 3 mg/day, but studies using 4 or more (up to 9) mg/day were more likely to show significant improvements in patient outcomes.
  1. Other agents.
    A number of additional medications marketed for other indications have been proposed for the treatment of dementia on the basis of epidemiologic data or pilot studies, but they cannot be recommended for use at this time. Aspirin and other NSAIDs have been proposed because of epidemiologic data suggesting that they protect against the development of the disease (132-135) and because of hypotheses regarding the involvement of inflammatory mechanisms (136). In a single small treatment trial for patients with Alzheimer’s disease, patients receiving indomethacin, 100 to 150 mg/day, experienced less decline over 6 months than did a matched control group (134). Patients using or considering these agents for other indications (e.g., arthritis treatment) might consider this preliminary evidence when weighing the risks and benefits of nonsteroidal therapy.

Estrogen replacement therapy, which is known to affect cognitive function (137), has been shown to be beneficial in the treatment of dementia in at least two case series (138, 139), and it has been associated with later onset and/or decreased risk of cognitive decline in at least two observational studies of postmenopausal women (140, 141). A clinical trial of estrogen in the treatment of postmenopausal women with Alzheimer’s disease is in progress. In the meantime, postmenopausal women weighing the risks and benefits of estrogen replacement might consider this preliminary evidence (142).

There is also interest in the hormone melatonin and in botanical agents, such as ginkgo biloba, which are available without a prescription. Because some of these agents are quite popular, psychiatrists should routinely inquire about their use and should advise patients and their families that these agents are marketed with limited quality control and have not been subjected to efficacy evaluations.

The chelating agent desferrioxamine has also been studied as a possible treatment for Alzheimer’s disease on the basis of hypotheses regarding heavy metals in the pathogenesis of the disease. In one small single-blind trial, there was some evidence of a decrease in cognitive decline over 2 years (143). Study of another chelating agent has failed to confirm this finding (144). Because chelating agents are quite toxic and support for them is so weak, they cannot be recommended for the treatment of dementia.

  1. Treatments for psychosis and agitation
    1. Goals.
      Use of such treatments is intended to decrease psychotic symptoms (including paranoia, delusions, and hallucinations) and associated or independent agitation, screaming, combativeness, or violence and thereby increase the comfort and safety of patients and their families and caregivers. Although DSM-IV defines one subtype of Alzheimer’s disease (and other dementias) on the basis of delusions and another subtype by behavioral disturbances, this section covers both, along with hallucinations, paranoia, and suspiciousness. This section also briefly addresses the treatment of anxiety in demented individuals.

In the consideration of an intervention, it is critical to be specific in describing target symptoms, both to select the optimal treatment and to monitor the effect of that treatment (145-147). However, the treatments used for this broad group of symptoms overlap to a considerable extent, so they are discussed together here.

  1. General principles.
    Interventions for psychosis should be guided by the patient’s level of distress and the risk to the patient or caregivers. If there is little distress or danger, reassurance and distraction are often all that is required. If the patient is distressed or if accompanying agitation, combativeness, or violent behavior put the patient or others in danger, psychopharmacologic treatment is indicated. The principles for anxiety are similar, although it is less commonly associated with dangerous behavior.

“Agitation” is an umbrella term that can refer to a range of behavioral disturbances, including aggression, combativeness, hyperactivity, and disinhibition. The first priority in treating such conditions is a careful medical evaluation. Agitation can result from an occult general medical problem, untreated or undertreated pain, depression, sleep loss, or delirium. The agitation will often resolve with treatment of the underlying condition. The next step is an assessment of the patient’s overall situation: agitation can also result from physical discomfort, such as hunger, constipation, or sleep deprivation; an interpersonal issue, such as a change in living situation or a new caregiver or roommate; or an emotional difficulty, such as frustration, boredom, or loneliness. Attending to unmet needs, providing reassurance, or redirecting activities may resolve the problem. If the agitation occurs repeatedly, it is often helpful to institute the behavioral measures discussed in section III.C.2.a. If these measures are unsuccessful, then pharmacologic treatment should be considered, especially if the agitation puts the patient or others in danger.

If the psychosis and/or agitation is deemed dangerous to the patient and caregiver, the psychiatrist must undertake the measures necessary to assure safety, in addition to pharmacologic intervention. Such additional measures may include hospitalization, one-on-one care, or physical restraint.

Whatever agent is used in the treatment of behavioral disturbances, its continued use must be evaluated and justified on an ongoing basis. As a dementing illness evolves, psychosis and agitation may wax and wane or may change in character: more or less of a medication, a change in medications, or no medication at all may be indicated in response to these changes.

  1. Antipsychotics
    1. Efficacy.
      Antipsychotic medications have been extensively studied in the treatment of psychosis and agitation in demented individuals. For example, a 1990 review (148) identified seven double-blind, placebo-controlled, randomized parallel group clinical trials including 252 patients studied over 3 to 8 weeks (149-155). Despite some methodologic flaws, notably small numbers of subjects and a lack of diagnostic specificity, taken together these studies constitute solid evidence for a modest improvement in behavioral symptoms with antipsychotic treatment. A meta-analysisof these seven trials (148), using clinician assessment of improvement in a variety of behavioral symptoms as the primary outcome, showed improvement in 59% of the subjects taking antipsychotics and 41% of those taking placebo. The studies varied widely in dose, ranging from 66 to 267 mg/day in chlorpromazine equivalents, and efficacy for behavioral symptoms was not correlated with standardized dose. Adverse effects were common, but specific rates are not available. Dropout rates were also high, whether associated with side effects or poor efficacy. Of note, one study suggested that antipsychotic agents are most effective specifically for psychotic symptoms (149). The available studies comparing antipsychotics to one another are of limited power but suggest no difference in efficacy (147, 148). Of note, there are limited data on the efficacy of antipsychotic medications for demented individuals beyond 8 weeks of follow-up, although extensive clinical experience suggests that they are helpful for longer periods of time.

Newer agents, such as risperidone and clozapine, have not been studied for demented populations in well-controlled trials. However, risperidone was effective in several case series of geriatric patients, including those with dementia (156), and also has support from some geriatric clinicians, who report that it is effective against agitation and psychosis in geriatric patients even at very low doses (i.e., 0.5-2.0 mg/day), which may limit extrapyramidal side effects. Clozapine has been found to be useful in controlling psychotic symptoms in Parkinson’s disease (157) and Lewy body disease (158) and may also be useful for patients with Alzheimer’s disease who are sensitive to the extrapyramidal effects of conventional antipsychotic agents (159). Olanzapine was released in 1996, and sertindole and quetiapine are due to be released in the near future. Early data suggest that their efficacy is similar to that of conventional antipsychotic agents, but they have not yet been tested in geriatric or demented populations.

  1. Side effects and toxicity.
    Antipsychotic agents have a broad range of common side effects that tend to vary with medication potency, although any effect can be seen with any agent. High-potency agents (e.g., haloperidol, fluphenazine) are most strongly associated with akathisia (which can worsen the target behaviors) and parkinsonian symptoms. Low-potency agents (e.g., thioridazine, chlorpromazine) are associated with sedation (which can lead to worsening cognition or falls), confusion, delirium, postural hypotension (which can also lead to falls), and a variety of peripheral anticholinergic effects (e.g., dry mouth, constipation). Risperidone shares many features with high-potency antipsychotic agents, although some clinicians feel that it has a somewhat lower risk of extrapyramidal effects, especially if low doses (0.5-2.0 mg/day) are used. Unfortunately, there are few data from direct comparisons of risperidone and conventional antipsychotics using equivalent doses, especially for elderly patients. Clozapine is less commonly associated with extrapyramidal side effects but is associated with sedation, postural hypotension, and an elevated risk of seizures. Early trials of the newer agents-olanzapine, sertindole, and quetiapine-suggest that they carry little or no risk of extrapyramidal effects for general psychiatric patients, but additional research and clinical experience will be necessary to characterize their performance for elderly and demented patients in general practice.

All of these side effects can be minimized by using the lowest effective dose. This principle is particularly important in order to minimize sedation and akathisia, both of which can actually worsen target behaviors and may thus make antipsychotics less effective (160). It may also be helpful to select an agent with the side effect profile most suited to a given patient. Anticholinergic agents may be effective in the treatment of parkinsonian side effects, but the high risk of associated cognitive decline, delirium, and other anticholinergic effects suggests that they should be used only with extreme caution for elderly and demented patients.

In addition to these common side effects, antipsychotic agents are associated with a risk of more serious complications that must be considered in weighing the risks and benefits of antipsychotic treatment. The first is tardive dyskinesia, which is more likely with increasing dose and duration of treatment and occurs more commonly in women, demented individuals, and the elderly in general. The risk may be as high as 30% for elderly patients with significant exposure (161-163). The second additional possible complication is neuroleptic malignant syndrome, which is rare but potentially lethal. Both of these complications have been reported with risperidone, although they may occur at lower frequencies. Clozapine appears less likely to be associated with these two complications, although they have been described with clozapine treatment. Clozapine has a significant risk of agranulocytosis, which is more common in the elderly than in younger patients (159), and regular monitoring of blood counts is required. Olanzapine, sertindole, and quetiapine have not been associated with these complications, although additional research data and practice experience will be needed to identify and characterize their effects on demented and geriatric patients in general practice.

  1. Implementation.
    Antipsychotics are the only pharmacologic treatment available for psychotic symptoms in dementia. They are also the most commonly used and best-studied pharmacologic treatment for agitation, and there is considerable evidence for their efficacy. However, there are a number of nonpharmacologic interventions that can be used before a trial of an antipsychotic or other medication is begun, as already outlined. Use of nonpharmacologic treatments is particularly critical given the large number and potential severity of side effects and more serious complications associated with antipsychotic medications.

There are no efficacy data to guide the choice among antipsychotic agents. Instead, the choice is based on the side effect profile. Some clinicians recommend agents that fall between the extremes of side effect profiles (e.g., perphenazine), but there are no data to support the contention that these agents have fewer adverse effects. It generally makes sense to select an agent whose most typical side effects are least likely to cause problems for a given patient (for instance, a higher-potency agent, e.g., haloperidol, might be selected if the patient is likely to be sensitive to anticholinergic effects, or a lower-potency agent, e.g., thioridazine, might be chosen if the patient has parkinsonian symptoms) or might actually be beneficial (e.g., a more sedating medication, given at bedtime, for a patient with difficulty falling asleep). Some clinicians believe that risperidone also poses a lower risk of extrapyramidal symptoms when used at doses of 2 mg/day or lower (164), but others disagree (165). Clozapine may be a good choice for individuals with Parkinson’s disease (157) or Lewy body disease (158) and possibly for others who cannot tolerate extrapyramidal side effects (156). Olanzapine and the two novel agents due to be released soon (sertindole and quetiapine) may prove to be good choices for geriatric and demented patients, especially those who are sensitive to extrapyramidal side effects, but insufficient data are available to recommend them at this time.

Antipsychotics are most commonly administered in the evening, so that maximum blood levels occur when they will help foster sleep and treat behavioral problems that peak in the evening hours (sometimes called “sundowning”). Most of these medications have long half-lives, so once-a-day dosing is generally sufficient. However, morning doses or twice-a-day doses may be helpful for patients with different symptom patterns.

On the whole, antipsychotic agents are given as standing doses rather than as needed, although as-needed doses may be appropriate for symptoms that occur infrequently. Oral administration is generally preferred, although an intramuscular injection may sometimes be used in an emergency or when a patient is unable to take medications by mouth (e.g., for a surgical procedure). Very low doses of depot antipsychotic medications (e.g., 1.25-3.75 mg/month of fluphenazine decanoate) were shown in a small open study to be effective in managing chronic behavioral problems in this population (166).

Low starting doses are recommended, e.g., 0.5 mg/day of haloperidol, 10-25 mg/day of thioridazine, 2 mg/day of perphenazine, 1 mg/day of thiothixene, 0.5-1.0 mg/day of risperidone, 12.5 mg/day of clozapine. The dose can be increased on the basis of the response of the target symptom(s). The usual maximum doses of these agents for demented elderly patients are 2-5 mg/day of haloperidol, 10-15 mg/day of thiothixene, 10-15 mg/day of trifluoperazine, 16-24 mg/day of perphenazine, 50-100 mg/day of thioridazine, 4-6 mg/day of risperidone, and 75-100 mg/day of clozapine. Most patients with dementia do best with doses below these maxima, but younger and less frail individuals may tolerate and respond to somewhat higher doses.

Given their side effects and potential toxicity, the risks and benefits of antipsychotic agents must be reassessed on an ongoing basis. The lowest effective dose should be sought, and emergent side effects should first be treated by dose reduction. The routine prescribing of anticholinergic agents is to be avoided. In addition, periodic attempts (e.g., every several months) to reduce or withdraw antipsychotic medications should be considered for all patients in the context of the probability of a relapse and the dangerousness of the target behavior(s).

  1. Benzodiazepines.
    1. Efficacy.
      The use of benzodiazepines in the treatment of behavioral symptoms in dementia has been studied in at least seven randomized clinical trials. Five studies including a total of 825 patients compared benzodiazepines to antipsychotics (167-171), and two studies compared benzodiazepines to placebo (172, 173). These studies are limited by poorly specified diagnosis, a mixture of target symptoms, limited outcome measures, and, in most cases, high doses of long-acting agents. Nonetheless, they show fairly consistently that benzodiazepines perform better than placebo but not as well as antipsychotics in reducing behavior problems. However, it is somewhat difficult to extrapolate these results, most of which are based on substantial doses of long-acting agents (e.g., 12 mg/day of diazepam), to the lower doses or the shorter-acting agents more commonly used today. There are no data concerning the efficacy of benzodiazepines after 8 weeks or whether one benzodiazepine is more effective than another.
    2. Side effects and toxicity.
      The most commonly reported side effects are sedation, ataxia, amnesia, confusion (even delirium), and paradoxical anxiety. These can lead to worsening cognition and behavior and can also contribute to the risk of falls (174). They also carry a risk of respiratory suppression in patients with sleep-related breathing disorders. Because all of these effects are dose related, the minimum effective dose should be used. Agents with long half-lives and long-lived metabolites can take weeks to reach steady-state levels, especially in elderly patients, so these agents must be used with particular caution. There is some evidence that elderly patients taking long-acting benzodiazepines are more likely to fall, and to suffer hip fractures, than those taking short-acting agents (175), although it is possible that the total dose, and not the duration of action, is the culprit (176). Clinical experience suggests that, like alcohol, benzodiazepines may lead to disinhibition, although there are few data to support this. The risk of medication dependence (and withdrawal, if the medication is stopped abruptly) is also a concern for some patients.
    3. Implementation.
      Although benzodiazepines may have a higher likelihood of side effects and a lower likelihood of benefit than antipsychotics, they can be useful in treating agitation in some patients with dementia, particularly those in whom anxiety is prominent. They may be particularly useful on an as-needed basis for patients who have only rare episodes of agitation or those who need to be sedated for a particular procedure, such as a tooth extraction. However, given the risk of disinhibition (and thus worsening of the target behaviors), oversedation, falls (and associated injuries), and delirium, their use should be kept to a minimum.

Among the benzodiazepines, many clinicians favor agents such as oxazepam and lorazepam that do not require oxidative metabolism in the liver and have no active metabolites. Temazepam shares these characteristics but is more problematic because of its long half-life. Lorazepam may be given on an as-needed basis in doses from 0.5 to 1.0 mg every 4-6 hours. Standing doses of 0.5 to 1.0 mg may be given from one to four times per day. Oxazepam is absorbed more slowly, so it is less useful on an as-needed basis. Standing doses of 7.5 to 15.0 mg may be given one to four times per day. Some clinicians prefer long-acting agents, such as clonazepam (starting at 0.5 mg/day with increases up to 2 mg/day) (177). However, such agents must be used with caution: dose increases must be made very gradually, as the medication can continue to accumulate over a substantial period, and vigilance concerning the increased risk of falls must be exercised. If benzodiazepines are used for an extended period (e.g., a month), they should be tapered rather than stopped abruptly owing to the risk of withdrawal.

  1. Anticonvulsants.
    1. Efficacy.
      Use of carbamazepine has support from several case series (178), a small open trial (179), a double-blind, nonrandomized trial (180), and one randomized trial (181) in which it was associated with nonsignificant decreases in behavioral measures.

Several favorable case reports and open trials have been reported for the anticonvulsant valproate (182, 183).

  1. Side effects and toxicity.
    The principal side effects of carbamazepine include ataxia, sedation, and confusion, which are particular concerns for elderly and demented patients. In addition, in rare instances carbamazepine can lead to bone marrow suppression or hyponatremia.

Valproate’s principal side effects are gastrointestinal disturbances and ataxia. In addition, in rare instances it can lead to bone marrow suppression or hepatic toxicity.

  1. Implementation.
    Given the sparse data on these agents, they cannot be recommended with confidence for the treatment of agitation in demented patients. Nonetheless, a therapeutic trial of one of these agents (especially carbamazepine, for which the data are somewhat stronger at this point) may be appropriate for some nonpsychotic patients, especially those who are mildly agitated or are sensitive or unresponsive to antipsychotics. Given the potential toxicity of these anticonvulsant agents, it is particularly critical to identify and monitor target symptoms and to stop administering the medication if no improvement is observed.

Carbamazepine may be given in two to four doses per day, started at a dose of 100 mg/day, and increased gradually as warranted by behavioral response and side effects or until blood levels reach 8-12 ng/ml. Valproate is given in two or three doses per day and should be started at 125 to 250 mg/day, with gradual increases based on behavioral response and side effects or until blood levels reach 50-60 ng/ml (or, for rare patients in this population, 100 ng/ml).

Many clinicians recommend monitoring CBC and electrolyte levels in patients taking carbamazepine and monitoring CBC and liver function values in patients taking valproate, owing to the possibility of bone marrow suppression, hyponatremia, and liver toxicity. However, this practice is not uniform. For details concerning the assessment and monitoring necessary during use of these agents, along with their side effects and potential toxicities, the reader is referred to the American Psychiatric Association’s Practice Guideline for the Treatment of Patients With Bipolar Disorder (184). However, a particularly cautious approach is warranted when treating elderly and demented patients, who may be more vulnerable to adverse effects, particularly CNS effects, and yet less likely to be able to report warning symptoms.

  1. Other agents.
    A number of other agents have been proposed for the treatment of agitation in patients with dementia (reviewed in references 148, 185, and 186). Efficacy data for these agents generally come from case reports or small open trials, often of mixed populations.

Data on trazodone have been provided by a few favorable case reports and case series (187-189), one small open trial (190), and one small double-blind, randomized clinical trial (191). Postural hypotension, sedation, and dry mouth are the principal side effects. Trazodone is generally given before bed but can be given in two or three divided doses per day. It can be started at 25 to 50 mg/day and gradually increased up to a maximum dose of 150 to 250 mg/day. Preliminary data suggest that SSRIs may also be useful in the treatment of agitation (190, 192, 193).

There have been at least two case reports (194, 195) and one open trial (196) concerning buspirone as a treatment for agitation or anxiety in elderly patients with dementia. Buspirone is generally well tolerated by elderly individuals but is sometimes associated with nausea, headache, dizziness, light-headedness, and fatigue. It is not associated with psychomotor impairment or tolerance and dependence. It is given in up to four doses per day and can be started at 5 mg/day and increased up to a total daily dose of 60 mg. There have been case reports of serotonin syndrome when buspirone is combined with SSRIs, so it should be used with SSRIs or other serotonergic agents only with cautious monitoring.

Given the limited efficacy data, none of these less-well-studied agents can be recommended with any confidence for the treatment of agitation and psychosis in patients with dementia. Nonetheless, a therapeutic trial of trazodone, buspirone, or perhaps an SSRI may be appropriate for some nonpsychotic patients, especially those with relatively mild symptoms or those who are intolerant of or unresponsive to antipsychotics.

When male patients display intrusive disinhibited sexual behavior, a particular problem in patients with frontal lobe dementias, medroxyprogesterone and related hormonal agents are sometimes recommended (197-199), but only case series support this recommendation at present.

Lithium carbonate has also been suggested because of its occasional utility for mentally retarded patients, but support for it is quite limited and side effects (including a considerable risk of delirium) are common (148).

Beta blockers, notably propranolol, metoprolol, and pindolol, have also been reported to be helpful for some agitated patients with dementia (200). However, most of the patients included in the case reports had unusual illnesses. In addition, large doses (e.g., 200-300 mg/day of propranolol) were used, and such doses create a considerable risk of bradycardia, hypotension, and delirium for elderly patients.

  1. Treatments for depression
    1. Goals.
      Somatic treatments for depression in demented patients are used to improve mood, functional status, and quality of life. Even patients with depressed mood who do not meet the diagnostic criteria for major depression should be considered for treatment. This is in keeping with the subtypes of Alzheimer’s disease and other dementias included in DSM-IV, which are diagnosed when a patient with dementia has a depressed mood with or without the full depressive syndrome. However, patients should be carefully evaluated for neurovegetative signs, suicidal ideation, and other indicators of major depression, since these may indicate a need for safety measures (e.g., hospitalization for suicidality) or more vigorous and aggressive therapies (such as higher medication doses, multiple medication trials, or ECT).

One goal of treating depression in dementia is to improve cognitive symptoms. Sometimes cognitive deficits partially or even fully resolve with successful treatment of the depression. Individuals whose cognitive symptoms recover fully with treatment of the depression are considered not to have been demented (this condition is sometimes referred to as “pseudodementia”); however, as many as one-half of such persons develop dementia within 5 years (2). Thus, caution is urged in ruling out an underlying early dementia in patients with both mood and cognitive impairment.

A related goal in the treatment of patients with dementia is diminishing apathy, which is common in Alzheimer’s disease and many other dementing illnesses, especially those affecting the frontal lobes, and may occur even in the absence of depression. The treatments for apathy overlap those for depression, so they are reviewed here as well.

Before any treatment is instituted, patients should be evaluated for general medical problems (e.g., hypo- or hyperthyroidism, electrolyte imbalance), substance abuse, and medications (e.g., ß blockers, corticosteroids, benzodiazepines) that may be causing or contributing to the depression. Correctable general medical problems should be addressed, and potentially offending medications should be discontinued (with appropriate substitutions as necessary).

  1. Antidepressants.
    1. Efficacy.
      The evidence for the efficacy of antidepressants for demented patients is limited. There have been five small placebo-controlled studies (192, 193, 201-203) and an additional two studies using a within-subject design (204, 205). These studies are limited by small number of subjects, mixed and poorly characterized dementia syndromes, mixed and poorly characterized depressive symptoms (defined variously by mood alone, by diagnosis, or by symptom severity threshold), lack of randomization, and the use of uncommon agents (including many not available in the United States). However, the available evidence suggests that depressive symptoms (including depressed mood alone and with neurovegetative changes) in dementia are responsive to antidepressant therapy. Cognitive symptoms beyond the impairment induced by the depression do not appear to respond to antidepressant treatment. Indeed, one imipramine trial suggested that antidepressant treatment might exert a negative effect on cognition, but this observation might have been due to the medication’s anticholinergic properties (201).

Although the data from the best-designed trials concern unusual agents, they document that depression in patients with dementia is responsive to treatment. This, along with extensive clinical experience in using more-typical antidepressants for this population (206), suggests that the more substantial efficacy literature concerning pharmacologic treatment of depression may be cautiously applied to patients with dementia. The reader is referred to the Practice Guideline for Major Depressive Disorder in Adults (207) for a summary of this literature.

The literature concerning the treatment of apathy is much sparser. There is minimal evidence that dopaminergic agents, such as psychostimulants (d-amphetamine, methylphenidate), amantadine, bromocriptine, and bupropion, are helpful in the treatment of severe apathy, but promising case reports suggest that efficacy studies are warranted (208). Psychostimulants have also received some support for the treatment of depression in elderly individuals with severe general medical disorders (209-211).

  1. Side effects and toxicity.
    The reader is referred to the practice guideline on major depression (207) for a more detailed discussion of the side effects of antidepressant agents and to the forthcoming practice guideline on geriatric psychiatry for a discussion of the particular issues for the elderly. These effects, divided by medication class, are summarized briefly here.

SSRIs, including fluoxetine, paroxetine, and sertraline, tend to have a more favorable side effect profile than do cyclic agents. However, any SSRI can produce nausea and vomiting, agitation and akathisia, parkinsonian side effects, sexual dysfunction, and weight loss, although some of these effects are more common with one agent than another. In addition, physicians prescribing SSRIs should be aware of the many possible medication interactions.

The structurally unique agent bupropion is associated with a risk of seizures, especially at high doses. Venlafaxine is associated with elevations in blood pressure, which sometimes diminish over time.

Cyclic antidepressants generally have significant cardiovascular effects, including orthostatic hypotension and delays in cardiac conduction. Their effects on conduction make these agents dangerous in overdose, so they should be used only for patients who are adequately supervised to guard against accidental or purposeful overdose. Most cyclic antidepressants have anticholinergic properties to some degree, including blurred vision, tachycardia, dry mouth, urinary retention, constipation, sedation, impaired cognition, and delirium. These effects are most marked for amitriptyline and imipramine and least so for nortriptyline and desipramine, but there is considerable variation from patient to patient. Trazodone has minimal cardiac conduction or anticholinergic effects but is associated with postural hypotension, sedation, and a risk of priapism. Nefazodone is most commonly associated with sedation.

MAOIs, including tranylcypromine and phenelzine, can lead to postural hypotension, a particular concern with the elderly because of the risk of falls. In addition, they have complex medication interactions (sympathomimetic agents, narcotics, especially meperidine, and serotonergic agents must be avoided) and require dietary modifications (tyramine-containing foods, such as cheeses, preserved meat, and red wine, must be avoided), which may make them potentially dangerous for poorly supervised individuals with dementia.

Psychostimulants (d-amphetamine, methylphenidate) are associated with tachycardia, restlessness, agitation, sleep disturbances, and appetite suppression. Bromocriptine is associated with psychosis, confusion, and dyskinesias. Amantadine is sometimes associated with anticholinergic effects, including delirium.

  1. Implementation.
    There are no efficacy data on which to base the selection of one antidepressant over another. Thus, the choice of an antidepressant is generally based on the side effect profile and the general medical and psychiatric status of each patient. For example, if sedation is desired, trazodone or perhaps nortriptyline may be selected. If activation is desired, fluoxetine, bupropion, or desipramine may be selected. If the patient has urinary outflow obstruction, even agents with modest anticholinergic effects should be avoided. If the patient has a prolonged Q-T interval or A-V block, fascicular block, or significant coronary artery disease, cyclic antidepressants should be avoided if possible. These and other implementation issues are discussed in greater detail in the practice guidelines on depression and on geriatric psychiatry and are only briefly summarized here.

Many clinicians choose SSRIs as the initial treatment because of their better side effect profiles. Once-a-day dosing is appropriate. Fluoxetine should be started at 5-10 mg/day and increased at several-week intervals to a maximum of 40-60 mg/day. Paroxetine has the same dosing, but the dose can be increased every 1-2 weeks because of its shorter half-life. Sertraline may be started at 25 mg/day and increased at 1-2-week intervals up to a maximum dose of 150-200 mg/day.

Some clinicians favor bupropion. However, it appears to decrease the seizure threshold, so it should not be the first choice for individuals with a high risk of seizures. It is started at 37.5 mg b.i.d. and increased every 5 to 7 days as tolerated up to a maximum of 350-450 mg/day in divided doses. No more than 150 mg should be given within any 4-hour period because of the risk of seizures. Venlafaxine should be avoided for individuals with hypertension if good alternatives are available; if it is used, careful monitoring of blood pressure and adjustment of antihypertensive medication are required. It is started at 18.75 to 37.50 mg b.i.d. and may be increased at approximately weekly intervals up to a maximum dose of 300-375 mg/day. If elevations in blood pressure occur and do not diminish over time and venlafaxine is effective in treating depression in an individual who has not responded to trials of other agents, the medication may be continued and the hypertension may be treated.

Among the tricyclic and heterocyclic agents, theoretical reasoning and clinical experience suggest avoiding agents with prominent anticholinergic activity (e.g., amitriptyline, imipramine). Among the remaining agents, sample dosing strategies are given here for nortriptyline, desipramine, and trazodone. Nortriptyline may be started at 10 to 25 mg/day, with increases at 5-7-day intervals up to a maximum daily dose of 100-150 mg. Dosing is guided by clinical response and side effects. Blood levels, which should not exceed 100-150 ng/ml, may also be helpful. For desipramine, the starting dose is 25 to 50 mg/day, with increases at 5-7-day intervals up to a maximum daily dose of 200 mg. Blood levels should not exceed 150-250 ng/ml. For trazodone, the starting dose is 25-50 mg/day, with increases at 5-7-day intervals up to a maximum daily dose of 300-400 mg.

Because of their side effects and the extra monitoring required, MAOIs should be considered only for individuals who are unresponsive to or unable to take other agents. The MAOIs tranylcypromine and phenelzine may be used at starting doses of 10 mg/day and 15 mg/day, respectively, with monitoring of orthostatic blood pressure, and increased at weekly intervals to maximum doses of 40 and 60 mg/day (in divided doses), respectively. Patients and caregivers must be advised in detail about dietary and medication restrictions. They should also be educated about the symptoms of hypertensive crisis and advised to seek medical attention immediately if these symptoms arise. It is important to inform caregivers that dietary supervision is necessary, since demented patients are unlikely to remember dietary restrictions on their own.

Stimulants are sometimes used in the treatment of apathy or of depression in individuals with serious general medical illness. Dextroamphetamine and methylphenidate are started at 2.5 to 5.0 mg in the morning. They can be increased by 2.5 mg every 2 or 3 days to a maximum of 30 to 40 mg/day. As they are controlled substances, adequate steps to avoid abuse should be taken. Amantadine is sometimes used in the treatment of apathy as well. It may be started at 100 mg/day and increased to a maximum of 200 mg/day. Bromocriptine may be started at 1.25 mg b.i.d. and gradually increased; few patients tolerate more than 2.50 mg b.i.d.

Most patients with dementia will not tolerate the higher maxima given for antidepressant agents, but younger and less frail individuals may tolerate and respond to somewhat higher doses. When a rapid response is not critical, a still more gradual increase may increase the likelihood that a therapeutic dose will be tolerated.

  1. Electroconvulsive therapy.
    There is no substantial literature on the efficacy of ECT in the treatment of depression in dementia. However, considerable clinical experience suggests that ECT may be beneficial for patients with severe major depression who are ineligible for, cannot tolerate, or do not respond to other agents (212). Dementia increases the likelihood of delirium and of memory loss following ECT, but these effects are generally of short duration: delirium tends to resolve within days and memory loss within weeks. Twice-weekly rather than thrice-weekly and unilateral rather than bilateral ECT may decrease the risk of cognitive side effects after ECT.


  1. Treatments for sleep disturbance
    1. Goals.
      Treatment of sleep disturbance in dementia is aimed at decreasing the frequency and severity of insomnia, interrupted sleep, and nocturnal confusion in patients with dementia. The goals are to increase patient comfort and to decrease the disruption to families and caregivers. Sleep disorder is common in dementia (213, 214) and is not always so disruptive that the risk of medication side effects is outweighed by the need for a pharmacologic trial. Thus, the psychiatrist assessing a patient for a sleep disorder should first consider whether treatment is needed and then whether appropriate sleep hygiene-including regular sleep and waking times, limited daytime sleeping, avoidance of fluid intake in the evening, calming bedtime rituals, and adequate daytime physical and mental activities (215)-has been tried. If the patient lives in a setting that can provide adequate supervision without undue disruption to others, permitting daytime sleep and nocturnal awakening may provide an alternative to pharmacologic intervention. Pharmacologic treatment should be instituted only after other measures have been unsuccessful. In addition, the clinician should consider whether the sleep disorder could be due to an underlying condition. It is particularly important to be aware of sleep apnea (216), which is relatively common in elderly individuals and contraindicates the use of benzodiazepines or other agents that suppress respiratory drive.
    2. Efficacy.
      There are no available reports of studies that have assessed the efficacy of pharmacologic treatment for sleep disturbances specifically in individuals with dementia or that have compared pharmacologic to nonpharmacologic therapies. However, there are some data concerning use of various agents for mixed elderly populations. Reports of two small studies of chloral hydrate use with the elderly (217, 218) are available. Piccione et al. (217) found chloral hydrate to be better than placebo but not as good as triazolam in the short-term treatment of insomnia in elderly individuals. Linnoila et al. (218) found chloral hydrate to be superior to both tryptophan and placebo in the treatment of sleep disturbances in elderly psychiatric patients. Zolpidem was studied in 119 elderly psychiatric inpatients (219), of whom 50% suffered from dementia. In a double-blind, randomized parallel group clinical trial, zolpidem was superior to placebo on multiple sleep outcomes. A dose of 10 mg appeared to be superior to 20 mg: it was equally effective in promoting sleep and lacked the daytime sleepiness and ataxia sometimes observed at the higher dose. The impact of benzodiazepines and antipsychotics on sleep has not been studied systematically in demented elderly patients. Clinical experience suggests that low-dose antipsychotics (e.g., haloperidol, 0.5-1.0 mg) can be helpful in managing sleep problems in patients with dementia. Clinical experience with benzodiazepines is less favorable, although short- to medium-acting agents at low to moderate doses (e.g., lorazepam, 0.5-1.0 mg; oxazepam, 7.5-15.0 mg) are sometimes helpful for short-term disturbances (e.g., after a change in caregivers).

In addition to pharmacologic agents, there is preliminary evidence from three small open trials for elderly subjects with dementia (220-222) that early morning or evening bright light therapy may improve sleep (and possibly behavior as well). Others have reported preliminary evidence that the hormone melatonin may also be beneficial in the treatment of sleep disturbances in elderly individuals (223, 224), but the agent has not yet been subjected to controlled trials with demented individuals. In addition, it should be noted that such agents lack the quality controls of pharmaceutical agents.

  1. Implementation.
    Given the sparse efficacy data, the choice of pharmacologic agents is generally guided by the presence of other symptoms. For instance, if the patient has psychotic symptoms and sleep disturbance, antipsychotics will generally be given at bedtime, and a relatively sedating antipsychotic (e.g., thioridazine, mesoridazine) may be chosen if not otherwise contraindicated. If the patient is depressed and has a sleep disturbance, an antidepressant with sedative properties (e.g., trazodone, nortriptyline) will be given at bedtime. When anxiety is prominent, a benzodiazepine (e.g., lorazepam) may be selected.

When sleep disturbances occur without other psychiatric symptoms beyond the dementia itself, there is little to guide the choice among agents. Some clinicians prefer trazodone at 25-100 mg h.s. for sleep disturbances. Some prefer zolpidem, 5-10 mg h.s. Benzodiazepines (e.g., lorazepam, 0.5-1.0 mg; oxazepam, 7.5-15.0 mg) and chloral hydrate (250-500 mg) may be used but are generally not recommended for other than short-term sleep problems because of the possibility of tolerance, daytime sleepiness, rebound insomnia, worsening cognition, disinhibition, and even delirium. Triazolam in particular is not recommended for individuals with dementia because of its association with amnesia. Diphenhydramine, which is found in most over-the-counter sleep preparations, is used by some clinicians, but its anticholinergic properties make it suboptimal for the treatment of demented patients

Excerpted from “The American Psychiatric Association’s Practice Guideline for the Treatment of Patients With Alzheimer’s Disease and Other Dementias of Late Life” This practice guideline was approved in December 1996 and was published in May 1997.

© Copyright 2000 American Psychiatric Association