The "approved" agents for the treatment of insomnia will be discussed in detail below. However, a tremendous amount of "off-label" treatment of insomnia occurs in this country, encompassing a broad range of therapeutic agents. These include antidepressants, antihistamines, antipsychotics, and benzodiazepines. All of these compounds share sedating properties, which explains some of the rationale for their use (11—15).
However, the principal factors driving the use of these agents have nothing to do with efficacy, because they have not been demonstrated to be effective, nor with safety, because arguably, the risks associated with the use of these agents are likely to be substantially higher than those seen with approved agents. Among the factors driving prescriptions for these medications are cost and (for most of them) perceived lack of abuse potential.
Is the use, for example, of diazepam as a sleep-promoting agent irrational or reflective of poor practice? The answer is "not necessarily": if a patient is already using a sedating benzodiazepine compound in treatment of anxiety, it may be appropriate to give a bedtime dose as part of a plan to control anxiety and promote sleep. However, the agents that have been approved for the treatment of insomnia, whether they are classic benzodiazepines such as flurazepam, temazepam, or triazolam or benzodiazepine receptor agonists (BZRAs) such as zolpidem, zaleplon, or eszopiclone, have been selected for development to treat insomnia on the basis of specific sleep-promoting effects and relative absence of anxiolytic and musculoskeletal effects, in contrast with agents such as alprazolam, diazepam, or clonazepam.
Half-life is another important consideration. Although half-life is not the sole determinant of probability of residual sedation, it is clear that long half-life hypnotic agents such as flurazepam and sedatives such as clonazepam and the sedating antidepressants promote increased risk of residual sedation, an effect rarely seen with newer hypnotic agents with half-lives of about 8 hours or less.
Benzodiazepine receptor agonists
γ-Aminobutyric acid (GABA) is the predominant inhibitory neurotransmitter in the CNS. All approved hypnotic medications except ramelteon have their effects through impact on the GABAA receptor complex, typically at the benzodiazepine recognition site of this complex.
Two types of compounds approved for the treatment of insomnia are included within the BZRA category, with differentiation between them often made on somewhat artificial grounds. Agents with a classical benzodiazepine chemical structure are functionally BZRAs on the basis of their agonist properties at this receptor site. Five benzodiazepine hypnotic compounds are currently approved for use in the United States: flurazepam, quazepam, estazolam, temazepam, and triazolam. Quazepam and flurazepam have very long elimination half-lives of up to 120 hours. As a result, high plasma levels of these agents are maintained throughout the night. On the basis of these high levels, they are effective in promoting sleep over the night but have also been shown to lead to next-day cognitive and motor impairment as a consequence of residual sedation.
Estazolam and temazepam are often characterized as intermediate half-life agents. However, each has a half-life of greater than 8 hours and may promote residual sedation, especially at high doses or if taken without the opportunity to sleep for a full 8 hours.
Triazolam is a very effective and very potent hypnotic agent, with a relatively short half-life and short duration of action. It was widely used in the 1970s and 1980s, but during the late 1980s it was withdrawn from the market in several countries after questions were raised about its safety. It was associated with reported serious side effects, mostly psychological, usually occurring at high dosage levels. Triazolam remains available in the United States as a generic agent and has been used by many patients on a long-standing basis without toxicity or adverse reactions.
Benzodiazepine hypnotics have been demonstrated to be safe and effective in the treatment of insomnia but are restricted to short-term use. The recent State of the Science Conference observed that these agents have a higher frequency and severity of adverse effects than is seen with the BZRAs.
Several terms have been used to describe the most recently released hypnotic agents (zolpidem, zaleplon, and eszopiclone). The term "nonbenzodiazepine" has been used to describe these compounds, and this term is technically correct, because they do not have a benzodiazepine chemical structure, as do agents such as temazepam, flurazepam, and others.
However, these agents, like the benzodiazepines, promote sleep on the basis of their agonist effects at the benzodiazepine recognition site of the GABAA receptor complex. As such, they are functioning as BZRAs, and, thus, the acronym BZRA seems appropriate on the basis of the therapeutic effects of these agents.
Like the classic benzodiazepines, the BZRAs are classified as schedule IV agents by the Drug Enforcement Administration. This classification reflects an assessment of a low risk of abuse potential, and during their many years of widespread use in this country the BZRAs have demonstrated little to no actual risk of abuse, addiction, or diversion.
The three BZRA medications that have been available in this country in recent years have half-lives and durations of action ranging from short to intermediate. Zaleplon, with an extremely short half-life of about 1 hour, is effective for treatment of sleep initiation problems, and has the distinction of only requiring 4 hours after administration before normal activities can be resumed. However, probably because of its short half-life, zaleplon does not have the capacity to increase total sleep time or improve sleep maintenance.
Zolpidem, with a half-life of about 2.5 hours, has demonstrated efficacy in sleep promotion and the capacity to increase total sleep time. It has been extremely successful in the United States and global marketplace. Although it does increase total sleep time, it does not have an indication for sleep maintenance.
Eszopiclone, with a half-life of about 6 hours, has a longer duration of action than either zaleplon or zolpidem and, as a consequence, is approved for treatment of both sleep initiation and sleep maintenance difficulties. However, its longer half-life and duration of action increase the risk that some patients will experience residual sedation with this medication.
As a group, these agents share properties of safety, efficacy, and low abuse potential. They have also demonstrated safety and efficacy in association with treatment over periods of up to 6 months, without development of tolerance or rebound insomnia when discontinued. For example, Krystal et al. (2) evaluated use of eszopiclone compared with placebo in a 6-month, double-blind study of close to 800 patients with primary insomnia. The results were dramatic, with eszopiclone providing immediate and sustained improvements in multiple sleep parameters compared with placebo. The results were sustained for the 6-month period of evaluation without development of tolerance.
Another study of great interest to psychiatrists involving eszopiclone has recently been published (3). This study evaluated the effect of adding eszopiclone to fluoxetine for patients who met DSM-IV criteria for both major depressive disorder and insomnia. All subjects received morning fluoxetine and were randomly assigned to either nightly eszopiclone 3 mg or a placebo treatment arm. Treatment was given for 8 weeks.
Not surprisingly, patients in the active treatment group had significantly decreased sleep latency, decreased wake time after sleep onset, increased total sleep time, and improved sleep quality and depth of sleep at all time points compared with those receiving placebo. Eszopiclone cotherapy also resulted in significantly greater changes in Hamilton Depression Rating Scale scores at week 4 with progressive improvement at week 8 and significantly more responders to depression treatment and more remitters at week 8.
How should these data be interpreted? Prior work with zolpidem had demonstrated that use of this agent in combination with antidepressants improved sleep and did not interfere with the therapeutic effects of the antidepressants used (fluoxetine, sertraline, or paroxetine), when zolpidem was added to a treatment regimen already established to treat depression (4).
The study of Fava et al. (3) was prospective and adequately powered to demonstrate benefits, if they exist. Although this study demands replication, the fact that the improvements seen persisted for several weeks after eszopiclone was discontinued argues that synergy may exist between treatment of insomnia and of depression. This should not be too surprising to clinicians as the relationships between sleep and mood are myriad and recognized by both our patients and us. Poor sleep may be a harbinger of impending depression or mania; mania may be provoked by sleep deprivation. Can an impending depressive episode be averted by improved sleep quality and/or increased sleep duration? Although this has not yet been established on the basis of prospective research, few of us would not advise patients "at risk" to try to obtain as much good-quality sleep as possible in hopes of averting a depressive, manic, or psychotic episode.
The efficacy of the non-nightly use of zolpidem 10 mg has been demonstrated by Perlis et al. (5) in a double-blind study lasting 12 weeks. Subjects were instructed to take medication (zolpidem 10 mg or placebo) between three and five nights per week and rated sleep from various perspectives on nights when medication was used or was not used. In these studies, ratings of sleep latency, total sleep time, number of awakenings, and sleep quality were all improved on nights when zolpidem was taken, compared with results of subjects taking placebo.
Until recently, hypnotic agents were available only as immediate-release formulations. The only way to lengthen the hypnotic duration of action was to select an agent with a longer half-life or to increase the dose of the drug being used. Use of a hypnotic agent with a longer half-life will typically promote greater sedation through the night but with increased risk of next-day residual effects.
Dose escalation of hypnotic agents can increase the duration that a medication is above a minimum effective concentration threshold, promoting better sleep maintenance. However, this may increase the risk of side effects as a consequence of increases in maximal concentrations.
Modified-release formulations have been developed to provide altered pharmacokinetic profiles for hypnotic agents. In its most extreme form, such an agent would have an immediate onset, sustained effects, and stable concentrations for the desired therapeutic period, and a rapid offset of effects.
Such an agent is a practical impossibility, but use of techniques to modify the release of and availability of a hypnotic agent has led to the development of zolpidem CR. This compound is a combination of immediate-release and modified-release components designed to achieve initial plasma levels and peak concentrations similar to those seen with immediate-release formulations while maintaining higher plasma levels for a longer of time, extending the duration of efficacy. This technology allowed zolpidem CR to be approved for treatment of both sleep initiation and sleep maintenance complaints.
Is there any substantive benefit that can be obtained from the use of such extended-release formulations? Data from the research conducted as part of the approval process for this agent demonstrated that it is capable of promoting better sleep continuity and better capacity to return to sleep after mid-nocturnal arousal, despite an arousing, aversive stimulus (recorded traffic noise in these studies).
Zolpidem CR was evaluated in a 6-month, double-blind, placebo-controlled study by Erman et al. (6). Results demonstrated that zolpidem CR was differentiated from placebo with regard to its capacity to improve sleep initiation, sleep maintenance, and satisfaction with sleep quality over this 24-week study, without development of tolerance.
Sodium oxybate is another name for γ-hydroxybutyrate, also known as GHB. Although this compound is widely know as a drug of abuse, it has been approved by the FDA for the treatment of cataplexy and residual daytime sleepiness in patients with narcolepsy. It is marketed under the trade name Xyrem and is available under a tightly controlled distribution system.
Sodium oxybate is a potent sedative agent, but sleep produced by sodium oxybate lasts for only a few hours. Its use is being investigated in conditions such as fibromyalgia and chronic fatigue syndrome, in which disrupted sleep has been proposed as a possible etiologic or exacerbating factor. At the present time, any use for conditions other than narcolepsy is off label. Although it is possible that access to this drug may be eased in the future if and when other indications are approved, sodium oxybate is a significant "street" drug of abuse and has a high potential for lethality in overdosage. Its use should be restricted to patients with narcolepsy by clinicians familiar with the potential risks and benefits of this agent.
Melatonin receptor agonist
Ramelteon is a melatonin receptor agonist and is currently the only agent of this type approved for the treatment of insomnia. Its therapeutic effects are believed to be generated by melatonin agonist activity at MT1 and MT2 receptor sites in the suprachiasmatic nucleus. It appears to have minimal activity at the GABAA receptor complex, the site of action of benzodiazepine and BZRA hypnotic agents.
Ramelteon has been studied in outpatient and inpatient environments. In a crossover study involving 107 adults with chronic primary insomnia, doses of 4, 8, 16, and 32 mg all yielded reductions in sleep latency and increases in total sleep time, without evidence of next-day residual effects. A 5-week outpatient study was performed involving 829 older adults (aged 65 years or greater) with chronic insomnia. Using the patient-reported sleep day, sleep latency was significantly reduced compared with placebo at weeks 1, 3, and 5, without evidence of rebound insomnia or withdrawal effects after treatment discontinuation (7).
Ramelteon has a half-life of 1 to 2.6 hours and is only indicated for sleep initiation. It appears to have minimal abuse potential or evidence of psychomotor or cognitive toxicity, even at doses of up to 160 mg, 20 times the therapeutic dose (8). Ramelteon is not restricted to short-term use.
The recommended dose of this compound is 8 mg for adult and older adult (age >65) populations as well. This compound has been demonstrated to be effective for sleep initiation, with little benefit for improvement in sleep maintenance. The effects of the agent appear to become greater with longer administration (i.e., 5 weeks versus 3 weeks versus 1 week), without any suggestion of the development of tolerance. It is unlikely to provoke side effects, with symptoms such as fatigue, dizziness, and somnolence seen in the 4% to 5% range (as contrasted with the 2% to 3% range for placebo).
Off-label use of antidepressants for the treatment of insomnia is extremely common in the United States. Prominent among these agents is trazodone, a heterocyclic antidepressant, although studies examining its use in the treatment of insomnia are quite limited. One such study, a large parallel-group study of patients with well-defined primary insomnia compared trazodone 50 mg, zolpidem 10 mg, and placebo during 2 weeks of administration (9). Compared with placebo, trazodone and zolpidem shortened subjective sleep latency and increased total sleep duration, with greater reduction in sleep latency seen with zolpidem than with trazodone. The sleep-promoting effects of trazodone compared with those of placebo were seen only during the first week. In week 2, zolpidem was superior to placebo for sleep latency, but not for total sleep time. The findings of this study suggested that trazodone 50 mg may have some short-term hypnotic benefits in primary insomnia, but it appeared to be less potent than zolpidem.
Trazodone use has increased dramatically in recent years, with increases in use for the treatment of insomnia of almost 150% between 1987 and 1996. The use of trazodone and other antidepressants (including tricyclic antidepressants such as amitriptyline and doxepin and newer agents such as mirtazapine) to promote sleep is based on the sedative effects of these agents rather than on demonstrated sleep-promoting properties. Trazodone use is associated with the risk of side effects such as cardiac rhythm disturbances, orthostatic hypotension, arrhythmias, and priapism. Side effects associated with tricyclic compounds include hypotension with syncope, ventricular arrhythmias, and cardiac conduction disturbances (10).
Several factors have probably driven the use of these agents. They are not scheduled, and, thus, they are not perceived to be "abusable" compounds. They have a long history of relatively safe use, although death (from tricyclic drug overdose) and priapism (from trazodone exposure) certainly belie the notion that these are entirely safe compounds. Cost and ready access are major determinants of the use of these agents. There are typically no restrictions to their use in insurance formularies, and because they are (for the most part) available as generic preparations, their cost is a fraction of that for brandname prescription hypnotic agents.
The recent NIH State of the Science Conference on insomnia reviewed data supporting the use of antidepressant agents and found it lacking. The following statement about use of antidepressants for the treatment of insomnia was made: "All antidepressants have potentially significant adverse effects, raising concerns about the risk—benefit ratio. There is a need to establish dose-response relationships for all of these agents and communicate them to prescribers" (1).
Antipsychotics have long been used to promote sleep. In the 1960s, agents such as thioridazine or chlorpromazine were likely to be prescribed for patients with histories of substance abuse. These agents were perceived to have an acceptable risk profile and were perceived to be superior to available agents (chloral hydrate and flurazepam) with regard to safety and abuse potential. Unfortunately, some patients who were treated with these agents developed tardive dyskinesia.
Currently, atypical antipsychotics such as quetiapine and olanzapine are used to promote sleep for patients with and without major psychiatric problems. Why are these compounds used? For some patients, it may be argued that the sedative and anxiolytic effects of these medications are helpful in reducing daytime anxiety and agitation or that the antipsychotic effect of the medications help to stabilize mood or reduce "acting out" or other unwanted behaviors. However, the common motivations for choosing these agents appeared to be access and perceived safety. These agents are readily available within most formularies and are not controlled substances. It is typically relatively easy to have a prescription for one of these compounds approved for treatment of insomnia.
However, there are number of issues associated with these agents. They are expensive. The doses for the treatment of insomnia are uncertain, as there are no published studies demonstrating the effectiveness of these agents in patients with primary insomnia. The risk of side effects with antipsychotics, including neuroleptic malignant syndrome, tardive dyskinesia, and metabolic syndrome, is high. The capacity of these agents to lead to hyperglycemia is well recognized, and the capacity of both newer and older antipsychotic agents to increase the risk of development of diabetes has recently been demonstrated.
The NIH Conference also reviewed the safety and efficacy of these agents and stated the following observation with regard to their use: "Studies demonstrating the usefulness of these medications for either short- or long-term management of insomnia are lacking. Furthermore, all of these agents have significant risks. Thus, their use in the treatment of chronic insomnia cannot be recommended" (1).
Sedating antihistamines (such as hydroxyzine and diphenhydramine) and muscle relaxants (cyclobenzaprine) have long been used for treatment of insomnia. Just as for the antidepressants and the antipsychotics, the percentage of their total prescription use that is directed to treatment of insomnia is unknown. Diphenhydramine is the most widely used active ingredient in over-the-counter sleep aids.
Although these agents are readily available as over-the-counter preparations, antihistamines are associated with a number of risks that should limit their use, especially in senior patient populations. They have been demonstrated to impair alertness when given in the daytime and are likely to promote residual sedation when taken to promote sleep, especially for patients who spend inadequate time in bed after taking these medications or who are sensitive to their effects. The risk of anticholinergic toxicity as a consequence of the use of the agents is also a source of concern, particularly for elderly patients and for patients taking other compounds with anticholinergic effects (such as tricyclic antidepressants).
A number of agents are in the process of evaluation as treatments for insomnia. One, indiplon, may be released to the marketplace in the near future. It is a BZRA compound (chemically a pyrazolopyrimidine compound), closely related to zaleplon. Similar to zaleplon, it has a short half-life of about 1.5 hours and, as such, would be expected to be of interest primarily for the treatment of sleep initiation problems.
The company developing indiplon, Neurocrine Biosciences, elected to utilize a modified release formulation as a core element of its plans for this compound. The expectation was that a modified release formulation of a short half-life compound of this sort would allow for sustained effect, generating improved sleep maintenance with a low risk of residual sedation.
For reasons not yet publicly announced, the FDA did not approve the 15-mg modified release formulation of indiplon, approving only the immediate release forms at doses of 5 and 10 mg. Neurocrine is continuing to meet with the FDA to determine what would be needed to have the 15-mg modified release formulation approved for treatment of insomnia.
A low-dose formulation of doxepin for the treatment of insomnia is also being researched. The company performing this research, Somaxon Pharmaceuticals, has obtained the rights to a use patent for treatment of insomnia with very low doses of doxepin. Research conducted by the company has shown therapeutic effects in patients with insomnia at doses of 1, 3, and 6 mg of doxepin. Research is also being conducted on other antidepressant compounds, such as mirtazapine, which are known to have sedative properties and have been used off label for insomnia treatment.
Another agent being tested for treatment of insomnia is gaboxadol, a novel hypnotic that acts on GABAA receptors by a nonbenzodiazepine mechanism. This agent reportedly acts at extrasynaptic sites, differentiating it from BZRA compounds. On the basis of animal data, it does not appear to have significant abuse potential, and animal and human data suggest that it promotes the production of slow-wave sleep.
A number of companies are working on agents hoped to improve sleep depth and quality by promoting the production of slow-wave (delta) sleep. These agents have a mechanism of action based on 5-hydroxytryptamine receptor antagonism, a property of a number of other agents (such as ritanserin) that have been explored for their utility as hypnotic agents. Although these agents may promote slow-wave sleep, as did ritanserin, to be successful as hypnotic agents they must be able to generate a perception of improved sleep depth or quality, something that was not accomplished with ritanserin.
Another area of excitement in the treatment of insomnia is research involving the neurotransmitter orexin, also known as hypocretin. This neuropeptide became of interest in sleep research with the discovery that its dysregulation causes narcolepsy in several animal species. Humans with narcolepsy were discovered to have essentially absent orexin/hypocretin levels, and it is now recognized that orexin/hypocretin neurons excite various brain nuclei with important roles in sleep and wakefulness. Research is being conducted to determine whether stimulation of this pathway may provide stimulant effects and whether orexin/hypocretin could promote sleep.
Although the focus of this article is on pharmacological therapy, some attention must be paid to behavioral therapies as well. Various behavioral therapies have been demonstrated to be effective in the treatment of chronic insomnia (11) and an effort should be made to include behavioral therapies in the treatment of all patients with insomnia. At a minimum, a sleep history should be reviewed with every patient to try to find elements of sleep hygiene that may be responsive to treatment.
Cognitive behavioral therapy, an effective modality in the treatment of depression, has also been demonstrated to be effective in multiple research studies as a treatment for insomnia. Typically, cognitive behavioral therapy for insomnia works to address maladaptive behaviors and thought patterns that patients with insomnia develop over long periods of time. These disturbances become an entrenched part of these patients’ thought and behavior patterns, tend to develop no matter how the insomnia originally developed, and include behavioral and cognitive elements that interfere with the patient’ s ability to relax and fall off to sleep. Examples of mal-adaptive behaviors include irregular sleep-wake patterns, frequent daytime naps, clock watching, and going to bed at too early an hour. Among maladaptive thought patterns are trying too hard to fall asleep (which provokes greater arousal), pathological anxiety generated by the immediate inability to fall asleep, and excessive worry about sleep loss and its possible next day consequences.
Cognitive therapy uses various techniques to alter dysfunctional sleep cognitions, beliefs, attitudes, and expectations. Treatment targets include unrealistic expectations (i.e., "If I do not get 8 hours of sleep every night I will not be able to function"), faulty causal attributions ("My insomnia is something I was born with and is beyond my control"), and pathological anxiety about consequences of insomnia ("My insomnia is causing my poor health"). Among treatment strategies used are instructions to maintain realistic expectations about the amount of sleep one can attain, to never try to sleep, to try to minimize the importance of sleep, and to not worry excessively during or after a night of poor sleep.
Education in sleep hygiene is always beneficial to patients, even for those who feel they fully understand how to improve their sleep. Core elements include attempting to maintain regular bedtime and awakening hours, avoiding caffeine late in the day, minimizing clock watching in the bedroom, and ensuring that excessive noise or light in the bedroom does not interfere with or prevent sleep. Clinicians should discuss computer use with their patients; using a computer when one cannot sleep involves motor activity, cognitive activation, and light exposure, all of which will interfere with the ability to fall asleep.
Naps are another area for patient education. Patients may deny that they nap, often reporting that they just "rest" on the couch or in bed during the afternoon for 30 or 60 minutes. Others may acknowledge sleeping for 90 or 120 minutes over the course of the day but have difficulty understanding that this reduces sleep drive at night, making entry into sleep and sustained sleep more difficult.
Milton Erman M.D.; Clinical Professor Department Psychiatry, UCSD School of Medicine; President, Pacific Sleep Medicine Services.
Consultant: Sanofi, Mallinckrodt, Cephalon, Takeda, Neurocrine. Grants/Research Support: Sanofi, Cephalon, Takeda, Aventis, Pfizer, Pharmacia, ResMed, Merck, Schwatz, Organon, GlaxoSmith Kline, Lilly. Advisory Board: Sanofi, Cephalon, Takeda, Neurocrine. Speakers Bureau: Sanofi, Takeda, Stock Shareholder: Cephalon, Forest, Neurocrine, Pfizer, Sepracor, Merck, Sanofi, Somaxon.