Adherence with psychopharmacotherapy can be enhanced by fully informing the older patient about what can reasonably be expected from the prescribed drug and by enlisting relatives or caregivers as informed, knowledgeable helpers. Thus, the explanation made to the patient and family is key to rational and effective treatment, influencing expectations, response, compliance, and the ability to tolerate adverse effects. The ideal discussion covers

Since many older patients leave their physician’s office without sufficient information about the drugs prescribed for them, oral instructions should be accompanied by the same information in writing whenever possible. This is because patients or their elderly caregivers may seem to understand in the office but forget once they get home or when confronted with a particular situation, such as a missed dose or a possible adverse effect.

Adverse effects can be explained in a matter of fact way without arousing alarm. The precise nature of the explanation should be tempered to match the patient’s clinical state, but it should never be entirely overlooked. An acutely disturbed schizophrenic may comprehend little but will still cooperate more readily if some rapport is attained. Depressed, anxious, or obsessive patients may require reassurance that they will not become addicted to medication, and it may also be necessary to explain that, although therapeutic effect is somewhat delayed, adverse effects can occur earlier.

For some schizophrenic, suicidal, or cognitively impaired patients, explanations should also be given to relatives who may need to assume responsibility, but in other instances, it may be important for the patient to assume complete control. The latter may be particularly true of patients who tend to place the onus for a “cure” entirely on the physician.

Finally, it is important to caution the patient about possible drug interactions and/or interference with activities such as driving.

Diagnosis. A related problem is delirium, which is characterized by alterations in mental status and ability to function. One report estimates that as many as 38% of hospitalized older adults will experience this phenomenon (15). The Diagnostic and Statistical Manual of Mental Disorders, 4th ed., Text Revision (DSM-IV-TR) (16) requires four features for the diagnosis:

Treatment. The Hospital Elder Life Program (HELP) found that lower rates of delirium were achieved in comparison to standard care when six risk factors were addressed (17):

Avoidance of drugs with anticholinergic properties is also important.

Thus, the first step in managing this condition is early recognition of predisposing risk factors. When delirium does develop, the next step is the search for and elimination of a potential underlying cause or causes. Behavioral approaches are often beneficial and can include maintaining regular schedules (including sleep-wake cycle), having the patient interact frequently with the same staff and/or family members, posting daily schedules, preventing sensory deprivation (e.g., glasses, hearing aids), and creating an environment that is consistent and uncluttered (17). One meta-analysis assessed strategies to prevent delirium in surgical patients and found support for (18)

At times, physical restraints may be required to ensure the safety of the patient and others, but they should be coupled with close, in-room monitoring when possible. Such close observation may even obviate the need for restraints.

When treatment of the underlying medical cause and behavioral interventions are not sufficient, medications may be necessary (19). Because the dopaminergic, cholinergic, and γ-aminobutyric acid (GABA)ergic systems have all been implicated, drug treatments typically impact these systems. Antipsychotics such as haloperidol (HPDL) are frequently used and most useful in controlling aggression (20). Higher doses of HPDL (e.g., >35 mg/day),
however, may predispose to cardiac conduction problems such as torsade de pointes (21). Other potential risks include acute extrapyramidal symptoms (EPSs) and neuroleptic malignant syndrome. Although low-dose SGAs (e.g., RISP, OLZ, or QTP) may be a better tolerated approach, a recent Cochrane review concluded that low-dose HPDL (<3 mg/day) was comparable in efficacy to RISP and OLZ in treating delirium without a greater frequency of adverse effects (22,23). Benzodiazepines (BZDs) and cholinesterase inhibitors (CIs) have also been considered in specific situations, but an acceptable risk-to-benefit ratio has not been clearly established (24).

In contrast to the data available for younger adults, one meta-analysis indicates a paucity of acute, randomized, and relapse-prevention clinical trials in the depressed elderly (43). Given this caveat, the review reports that 72% of the 18 acute, placebo-controlled trials included in their analysis demonstrated significantly greater efficacy with active drug versus placebo. A subsequent placebo-controlled trial in the very old (i.e., ≥75 years), however, found no difference between citalopram and placebo (44).

In general, the most important issues in choosing a drug for the elderly patient are similar to those outlined in Chapter 7. Additionally increased emphasis should be given to the adverse effect profile of a drug and the value of TDM to ensure adequate versus toxic or subtherapeutic dosing. For example, in a patient with cardiac conduction delay, the tricyclic antidepressants (TCAs) would not be the ideal choice. In a physically healthy elderly patient, however, the very cautious use of secondary amine TCAs, such as nortriptyline or desipramine, may be appropriate because of their clearly defined therapeutic plasma levels, proven efficacy, known adverse-effect profile, and low cost.

Selective serotonin reuptake inhibitors (SSRIs) avoid some of the more serious adverse effects seen with the TCAs (45). Thus, the SSRIs are considered an appropriate first choice across a range of severity of symptoms (46,47). Trazodone and bupropion are also appealing because of their milder anticholinergic and cardiovascular effects (48). With milder mood disturbances, a brief trial with psychostimulants, such as methylphenidate, might also be contemplated (see Chapter 7). Limited data suggest a possible role for testosterone therapy, particularly in males with lateonset depression (49).

Monoamine oxidase inhibitors (MAOIs) are usually not recommended because of uncertainty about adherence to dietary restrictions and the very real problem of hypotension. When used in selected patients, however, phenelzine has been found to be safe and effective. As noted in Chapter 7, however, the eventual introduction of agents such as selegiline TS or reversible inhibitors of MAO-A may lead to a greater use of this class of antidepressants in all age groups.

Electroconvulsive therapy (ECT) may be the most appropriate alternative for more severe forms of depression, characterized by a rapidly deteriorating course or nonresponsiveness to drug intervention or in patients with serious concurrent medical disorders (see Chapter 8) (50).

Dose of Antidepressant. Whatever agent is chosen, a general rule is to start at half the usual adult dose (and even lower if organicity is involved) and to increase the drug at a slower rate (12). Thus, for a drug such as desipramine or nortriptyline, 10 to 25 mg/day may be the best initial strategy. Fluoxetine can be initially given in doses of 5 or 10 mg/day or 20 mg every second or third day. Sertraline (initiated at 25 mg/day) is also an appropriate alternative strategy (51). If sedation is required, an agent such as trazodone (starting dose 25 mg/day at bedtime) or mirtazapine (7.5 mg/day at bedtime) may be used as the primary antidepressant or cautiously in combination with one of the other compounds discussed.

Adequate length of treatment may be longer (e.g., 6 to 12 weeks) in this population as a result of age-related pharmacokinetic and pharmacodynamic changes. TDM may be helpful to ensure levels that are therapeutic and nontoxic. Maintenance doses comparable with acute levels for 6 to 12 months after remission are usually appropriate.

In their review of the literature, Arean and Cook concluded that (52)

Further, results from Prevention of Suicide in Primary Care Elderly: Collaborative Trial (PROSPECT) indicate that diagnostic and treatment guidelines tailored for older patients plus care management were superior to usual care in reducing suicidal ideation regardless of depression severity (53,54). The three major elements of this approach included

A related issue is the modest body of evidence indicating that women with or without a prior history may be at greater risk of developing depression during this period of their lives (55). Perimenopause is defined as amenorrhea or irregular menses for less than 12 months. The absence of menses for more than 12 months defines the postmenopausal period. Estrogen levels decline during this time, and this hormone can modulate neurotransmitter activity implicated in the regulation of normal mood states.

Although SSRIs are commonly used to manage depression in this population, estrogen replacement therapy (ERT) has also been studied to clarify its potential benefits. For example, Schmidt et al. (56) reported significant improvement in mood scores after 3 weeks of ERT compared to placebo in perimenopausal women with major or minor depression. Subsequently, Rasgon et al. (57) reported benefit in an open study for ERT as a monotherapy or when augmenting fluoxetine in a small sample (n = 16) of perimenopausal women with major depressive disorder (MDD). Further, Soares et al. (58) studied 50 perimenopausal depressed (MDD, dysthymic disorder; minor depressive disorder) women in a double-blind, placebo-controlled trial of transdermal patches of 17β-estradiol (100 μg) over a 12-week period. Sixty-eight percent of those on estradiol patches met criteria for remission versus 20% on placebo (p < 0.001). Active treatment was well tolerated. By contrast, a 20week, randomized, controlled trial failed to separate estradiol (2 mg/day) from placebo in 115 older (≥70 years) women on measures of depression (Beck Depression Inventory, BDI), quality of life, or cognitive function (59). Further, longterm use of ERT may increase the risk of breast cancer and possibly adversely impact cognition.

Prevention strategies after an initial response to acute treatment are important given the high risk for recurrence, disability, and death in this age group. In this context, Reynolds et al. reported that elderly (≥70 years) depressed patients who responded to the combination of paroxetine and psychotherapy had a lower recurrence rate over 2years if they received maintenance paroxetine versus monthly maintenance psychotherapy (60). An alternative approach to prevent relapse in older patients with resistant depression is augmentation with low-dose SGAs such as RISP (e.g., 0.25 to 1 mg/day) (61).

The decision to use these agents should be made with considerable caution and only after possible underlying causes of the patient’s symptoms have been explored and treated
appropriately. Further, Salzman (78) has pointed out that relatively few research studies, most of which are seriously flawed, have examined the therapeutic effect of these agents in elderly patients. Thus, recommendations for the use of BZDs in elderly patients are derived primarily from studies of young adult patients, studies of pharmacokinetics and toxicity in elderly patients, and clinical and anecdotal experience.

As emphasized earlier in this chapter, age can significantly alter the pharmacokinetics of BZDs metabolized by oxidation, tending to reduce their clearance (see also Chapters 2 and 12). This change may lead to drug accumulation and possible toxicity during chronic administration, particularly in elderly men (79). BZDs transformed by conjugation have relatively short half-lives and accumulate less during multiple dosing, in part because the conjugation (as opposed to the oxidation) process appears to be less influenced by age. Age also appears to enhance pharmacodynamic sensitivity to BZDs so that at any given plasma or brain concentration, the elderly patient experiences an increased intensity of drug effect compared with the younger patient (79). In addition, physical illness (e.g., stroke; Parkinson disease; dementia; disorders that impair protein binding, hepatic metabolism, or renal clearance) may increase sensitivity. Finally, concomitant use of other drugs with CNS effects (e.g., antidepressants, stimulants, steroids, alcohol) may exacerbate BZD toxicity (78).

Adverse Effects. Use of both short and long halflife BZDs in elderly patients has been associated with a number of potentially serious unwanted effects, of which the most common are excessive sedation and cerebellar, psychomotor, and cognitive impairment. Table 15-3 lists common adverse effects of BZDs, which are often more pronounced in older patients. Further, BZDinduced amnesia, confusion, depression, and oversedation in elderly patients may be misdiagnosed as dementia (80).

Short half-life BZDs are usually recommended for the elderly patient because they are less likely to accumulate and are rapidly eliminated. Few comparisons, however, have been made of their effects on performance in the elderly population. In one study of nonanxious elderly volunteers, both diazepam (long half-life, oxidation) and oxazepam (short half-life, conjugation) produced comparable self-rated sedation and fatigue during long-term administration (72). These effects persisted in diazepam subjects during a 2-week washout period but rapidly returned to baseline in the oxazepam subjects.

Long half-life BZDs may increase the risk of daytime sedation, lethargy, cognitive impairment, and delirium, as well as falls and hip fractures (81,82 and 83). Long-term use of flurazepam (30 mg/day) has been associated with an increased incidence of ataxia and hallucinations (84). However, short half-life BZDs may also cause serious adverse effects. Ataxia, depression,
confusion, amnesic syndromes, and oversedation have been reported in elderly lorazepam users, and there is some evidence that short-acting BZDs may also increase the risk of falls (85,86,87 and 88).

Several issues should be carefully considered before drug treatment of dyssomnia is undertaken in an elderly patient. The first issue is whether a sleep disorder can be explained by another psychiatric or medical condition, which should then be addressed (e.g., an antidepressant for sleep disruption experienced during a major depressive episode or an analgesic in a patient with disabling pain). The second issue is whether any prescribed or nonprescribed drugs could explain the disorder. OTC agents and excessive caffeine ingestion may go unappreciated unless inquired about. Finally, the clinician must keep in mind that nonpharmacological interventions often suffice. In particular, a detailed review of sleep hygiene issues combined with instruction about stimulus control and sleep restriction may be most useful.

Although surveys indicate that BZDs are commonly prescribed for elderly patients, the NIH Consensus Development Conference stated that the efficacy and safety of sedatives and hypnotics have not been established for older people, nor has the extent to which they contribute to or alleviate sleep problems (89,90). In addition, sedatives and hypnotics as a group, and BZDs in particular, are often implicated in drug-related hospital admissions in the elderly due to adverse effects (91,92). For example, one recent randomized trial found temazepam (15 mg) superior to diphenhydramine (50 mg) and placebo but cautioned that falls associated with this BZD may mute this apparent advantage (93). This group is also at particular risk for abrupt drug discontinuation when hospitalized, with resulting withdrawal symptoms that may be unrecognized as such and attributed to other health problems (85,94,95). BZD hypnotics should not be routinely prescribed in the hospital unless the patient has a demonstrated sleep disorder. Even then, reassurance that restless sleep is normal in such a situation may obviate the need for a hypnotic (96).

Some evidence indicates that BZDs may also exacerbate breathing difficulties in patients with a chronic lung disorder (97). Because flurazepam has been reported to exacerbate sleep apneas in middle-aged and elderly normal volunteers, the possibility of undiagnosed sleep apnea should always be considered before a BZD is prescribed (98).

Triazolam. In 1983, the FDA-approved triazolam, 0.5, 0.25, and 0.125 mg, based on data suggesting that these doses would be safe and effective. Five years later, the FDA withdrew the 0.5-mg dose from the US market but continued approval of the 0.125- and the 0.25-mg dose (currently the recommended starting dose for nonelderly individuals) without new studies to demonstrate their efficacy. These actions were in response to scientific data indicating that the risk of adverse reactions to triazolam is dose dependent, a fact not emphasized by either the manufacturer or the FDA.

Between 1980 and 1991, only a few published studies examined the safety and the hypnotic efficacy of the 0.125-mg dose of triazolam in elderly patients (99,100,101 and 102). These studies and case
reports indicated that with continued use, the initial efficacy of the 0.125-mg dose of triazolam in elderly patients begins to wane after 1 week and progressively diminishes to ineffectiveness, usually by the sixth week of continuous administration. During the same period, the risk of potentially serious adverse reactions increases.

In the early 1990s, the Committee on Safety of Medicines of the United Kingdom concluded that the risks of treatment with triazolam at the licensed doses (0.25 and 0.125 mg) outweighed the benefits. The United Kingdom and a few other countries banned triazolam primarily because of persistent reports of adverse reactions. In 1992, France, Spain, and New Zealand suspended the 0.25-mg dose of triazolam but allowed continued marketing of the 0.125-mg dose, whereas Canada and Japan lowered the recommended starting dose for nonelderly insomniacs to 0.125 mg.

Since 1992, the FDA-approved labeling on triazolam has recommended that prescriptions be written only for short-term (7 to 10 days) treatment of insomnia; that use for more than 2 to 3 weeks requires a complete reevaluation of the patient; and that for geriatric or debilitated patients, a dose of 0.25 mg should not be exceeded, with higher doses being reserved for exceptional cases.

Triazolam may cause sedation, learning and memory impairment, confusion, irritability, paranoid delusions, aggression and irrational behavior, reversible delirium, anterograde amnesia, and automatic movements (86,103). In a placebo-controlled study of the pharmacokinetic and pharmacodynamic effects of single doses (0.125 and 0.25 mg) of triazolam in healthy young and elderly subjects, triazolam caused a greater degree of sedation and greater impairment of psychomotor performance in the elderly than in the young (102). Triazolam also has a narrow therapeutic range in elderly patients, and overdosage may occur with as little as 2 mg, an amount only four to eight times the recommended dose (96). Cases of fatal triazolam overdose have been reported in elderly patients who were debilitated or had taken other psychotropic drugs and alcohol (104).

In summary, if a sedative-hypnotic is judged appropriate, we suggest an initial trial with low doses of zolpidem or zaleplon (e.g., 2.5 to 5 mg at bedtime) or a short- to intermediate-acting BZD hypnotic (e.g., estazolam, 0.5 to 1 mg at bedtime). When drug therapy is initiated, it is extremely important to monitor older patients for cumulative effects, given their heightened organ system sensitivity (e.g., CNS) and decreased clearance rates (e.g., due to hepatic compromise). If sleep problems persist beyond 2weeks, a careful reassessment of diagnosis should be undertaken before represcribing a sedative-hypnotic.

Evidence indicates that abrupt discontinuation of long-term BZD use may be associated with severe withdrawal symptoms, including confusion, disorientation, and hallucinations, in elderly patients (80,95). Gradual discontinuation, however, appears to be tolerated as well by elderly patients as by younger patients (105). The cognitive impairment associated with BZD administration is reversible with drug discontinuation, often improving memory and concentration (78,90). For example, one controlled trial considered the impact of three approaches to tapering BZDs in 76 older adults with chronic insomnia (106). They included

More subjects (85%) who received the combined approach were BZD free after the initial intervention than were those who received medication taper only (48%) or CBT alone (54%). Further, the two groups that received CBT experienced greater subjective sleep improvements than those in the medication taper only group.

BZD hypnotics such as midazolam and triazolam are metabolized primarily via the cytochrome P450 (CYP) 3A3/4 microenzyme system. Other BZDs often used as hypnotics, such as diazepam, can also be metabolized by CYP 3A3/4 as well as by CYP 2C19. Any drugs that act as inhibitors or inducers of these isoenzymes could increase or decrease BZD levels, respectively (107). Thus, ketoconazole, macrolide antibiotics (e.g., erythromycin), SSRIs (e.g., fluoxetine-norfluoxetine and fluvoxamine), and other antidepressants (especially nefazodone) may decrease clearance and increase BZD levels to potentially toxic ranges. Conversely, rifampicin, CBZ, and dexamethasone may increase clearance and
decrease BZD levels to potentially subtherapeutic ranges.