As noted above, both behavioral treatments and pharmacological treatments can be applied to the treatment of almost all sleep disorders. In this section, we describe how the general characteristics of these treatments, described above, are applied to specific sleep disorders.

Primary Sleep Disorders: Dyssomnias

Primary Insomnia
Primary insomnia is a condition involving difficulty initiating or maintaining sleep, or nonrestorative sleep, for at least 1 month. The sleep disturbance (or associated daytime fatigue) must cause significant distress or impairment in social or occupational functioning. In primary insomnia, the sleep disturbance is not due to an underlying mental disorder (e.g., major depression), medical illness (e.g., congestive heart failure), substance abuse disorder, or another sleep disorder (e.g., sleep apnea).

In treatment outcome research, the definition of insomnia is often more operational and typically refers to a latency to sleep onset and/or time awake after sleep onset greater than 30 minutes per night, with a corresponding sleep efficiency of less than 85%, on 3 or more nights per week for at least 6 months.

Several subtypes of primary insomnia have been described. These include psychophysiological insomnia, a conditioned sleep problem that results from learned sleep-preventing associations and somatized tension; idiopathic insomnia, a lifelong inability to obtain adequate sleep, which may represent a defect in neuronal sleep/wake mechanisms; and sleep state misperception, a gross discrepancy between subjective and objective sleep measurements. To date, distinctions among these subtypes have not been strongly supported by empirical research, and many patients with chronic insomnia have some evidence for each subtype (Edinger et al. 1996; Reynolds et al. 1991).

Behavioral/psychotherapeutic treatment Insomnia has received more research attention than any other sleep disorder. The findings of more than 50 treatment studies (involving more than 2,000 patients) of nonpharmacological interventions for insomnia have been summarized in two separate meta-analyses (Morin et al. 1994a; Murtagh and Greenwood 1995) and in a recent practice parameter paper by a task force of the American Academy of Sleep Medicine (Morin et al. 1999b). The data indicate that behavioral treatment (lasting an average of 4-6 weeks) produces reliable and durable changes in several sleep parameters of patients with chronic insomnia. For example, sleep-onset latency is reduced from an average of 60-65 minutes at baseline to about 35 minutes at posttreatment. The duration of awakenings is similarly decreased, from an average of 70 minutes at baseline to about 38 minutes following treatment. Total sleep time is increased by a modest 30 minutes, from 6 hours at baseline to 6 hours after treatment, but perceived sleep quality is significantly enhanced with treatment. Overall, between 70% and 80% of patients benefit from treatment, about one-third become good sleepers, and a small proportion do not benefit from treatment. Sleep improvements achieved with behavioral interventions are well sustained over time, with follow-up evaluations averaging 6 months after treatment completion.

These results represent fairly conservative estimates of treatment efficacy because they are based on average effect sizes computed across all treatment modalities and all insomnia subtypes (primary and secondary). Although treatment effects are usually larger for primary than for secondary insomnia, recent evidence indicates that behavioral treatment modalities can also be of significant benefit for insomnia associated with medical and psychiatric conditions (this evidence is reviewed in later sections). Also, comparison of treatment procedures suggests that stimulus control and sleep restriction are the most effective single treatment modalities, whereas more variability exists in outcome with relaxation methods. Cognitive therapy is particularly useful in alleviating emotional distress surrounding the perceived effect of chronic insomnia and in teaching coping skills to manage residual sleep disturbances after treatment. Multifaceted interventions, which incorporate behavioral, educational, and cognitive components, are often indicated to target maladaptive sleep habits and dysfunctional beliefs and attitudes that frequently contribute to sleep difficulties (Espie et al. 2001; G. D. Jacobs et al. 1996; Morin 1993; Morin et al. 1999b).

An important limiting factor with nonpharmacological interventions is the time and training required for their implementation; as a result, these methods are underused in clinical practice (National Institutes of Health 1996). Recent studies have shown that behavioral interventions can be successfully implemented in primary care medicine either by family physicians (Baillargeon et al. 1998) or by nurse practitioners (Espie et al. 2001). Other studies have also documented significant benefits with behavioral self-help treatment for primary insomnia (Mimeault and Morin 1999; Riedel et al. 1995).

Pharmacological treatment Numerous studies have also shown the safety and efficacy of benzodiazepine receptor agonists and related medications in the treatment of symptoms that meet the diagnostic criteria for primary insomnia. A recent meta-analysis of benzodiazepine treatment effects found moderately large effect sizes for subjective outcome measures, including sleep quality, wakefulness during sleep, and sleep duration (Nowell et al. 1997). Other hypnotic medications, including barbiturate and older nonbarbiturate, nonbenzodiazepine drugs, should not be used because of their systemic toxicity and potential for dependence and serious withdrawal effects and the availability of safer agents. Antidepressant medications, such as sedating TCAs and trazodone, are also sometimes used. A few controlled trials support the efficacy of this approach (Hajak et al. 1996; Hohagen et al. 1994; Montgomery et al. 1983). Controversies surrounding the pharmacological treatment of primary insomnia focus not on efficacy but on the optimal duration of treatment, the prevalence of dependence and other side effects, and the role of medications in overall treatment planning.

Almost all treatment studies of benzodiazepines or related compounds in patients with primary insomnia (or patients who could be diagnosed with this disorder) have been short term, lasting 7-28 days (Nowell et al. 1997). Polysomnographic sleep studies, however, have documented the continued efficacy of medications such as triazolam, flurazepam, and zolpidem for as long as 5 weeks (Mitler et al. 1984; Vogel and Morris 1992), and studies that used subjective reports have shown continued effects for up to 1 year (Maarek et al. 1992). It is equally clear that abrupt discontinuation of benzodiazepine medications from regular nightly use can lead to “rebound” insomnia (A. Kales et al. 1983). This effect is more severe in patients with worse sleep efficiency at baseline (Merlotti et al. 1991) and after higher doses of hypnotics (Roehrs et al. 1986b). In general, withdrawal insomnia is more severe with short-acting than with long-acting medications and can be reduced by tapering the dose, even of short-acting agents (Greenblatt et al. 1987). However, the actual prevalence of withdrawal symptoms in a call-in sample of patients was less than 10% for both short- and long-acting benzodiazepines, a figure equal to that obtained with over-the-counter preparations (Balter and Uhlenhuth 1992). A recent placebo-controlled study showed that rebound insomnia following 4 weeks of treatment was significantly more frequent among patients who received placebo than among those who received benzodiazepine receptor agonists and was more frequent among treatment nonresponders than responders (Hajak et al. 1998).

Although insomnia is a very prevalent problem, most (about 85%) individuals who complain of it have not used medications for sleep in the previous year (Mellinger et al. 1985). The prevalence of tolerance with increasing medication doses is also difficult to estimate accurately, but Balter and Uhlenhuth (1992) found that the number of patients who increased benzodiazepine doses was similar to the number who increased antidepressant doses - less than 10%. Population data also indicated that the large majority of patients - about 75% - use hypnotics for less than 1 month and that fewer than 15% use them for more than 12 months (Balter and Uhlenhuth 1992; Mellinger et al. 1985). This contrasts with the pattern seen in other countries, where up to 75% of use is for longer than 1 year (Ohayon and Caulet 1996).

Debate has also focused on whether the use of medications impairs the efficacy of behavioral and psychotherapeutic treatments, as mentioned above. Case series and controlled trial data indicate better long-term results from behavioral measures alone (Morin et al. 1999a). Despite this experimental evidence, however, clinical experience supports a role for pharmacotherapy in patients with primary insomnia. The rapid relief of insomnia symptoms produced by medications may decrease anxiety related to sleep, interrupt the “vicious cycle” of preoccupation with sleeplessness, and allow the patient to focus more effectively on behavioral treatments. Ideally, the intermittent use of benzodiazepines or related medications in patients with chronic primary insomnia can be accomplished by limiting the number of doses to a maximum of three to four per week. Some clinicians recommend that benzodiazepines or related medications be taken every night for approximately 2-4 weeks and then gradually reduced in frequency. These clinicians argue that such dosage schedules more effectively interrupt the vicious cycle.