Conclusion: Implications for Medical Practice

  • Wallace B. Mendelson


We are beginning to have some understanding of the phenomenology of sleep, but its function remains elusive. It may be studied in terms of the EEG sleep measures, but their relationship to the subjective qualities of sleep—e.g., loss of consciousness, a sense of restoration—are not entirely clear. Something about this experience of sleeping has gone amiss in the perhaps one-third of Americans who complain of sleeplessness or lack of refreshment. Studies of large numbers of these insomniacs show that although as a group they sleep less and take longer to go to sleep than noncomplaining individuals, the EEG measures of the two groups overlap substantially. This seems to indicate that factors other than a decrease in the total amount of EEG sleep are involved in the genesis of the uncomfortable experience of insomnia. Among these may be a dissociation of the usual relationship of the sensation of arousal to EEG-defined waking, disorders of the circadian rhythm of sleep, and disruptions of the nonREM/REM cycle by multiple brief awakenings. It is also becoming clear that insomniacs are a heterogeneous group, many of whom have specific disturbances such as sleep apnea, nocturnal myoclonus, and depressive illness.

The response of many persons—perhaps 5–15% of the population—to insomnia is to take prescription hypnotics at least occasionally, and 1–3% take them very frequently. An additional 6% or so, perhaps a different group from those who take prescription hypnotics, purchase over-the-counter sleep aids. Given this large rate of consumption, it behooves us to learn as much as possible about the characteristics of these agents, their benefits and possible hazards.

The pharmacology of the hypnotics is very diverse. They include a number of different chemical classes which vary in many qualities including the ease of absorption, biological half-life, route of excretion, presence of psychoactive metabolites, and other factors. One consequence of this diversity is that each drug has its own particular set of advantages and disadvantages. Thus no single hypnotic is generally “better,” and in various clinical situations different hypnotics may be more advantageous (or at least less harmful).

The meaningfulness of evaluations of efficacy of hypnotics is limited by our very incomplete understanding of the nature of insomnia. It is not clear what the appropriate measures should be, and each of the commonly used approaches—the sleep EEG, patient reports, and nursing observations—have very clear limitations. In short-term use, virtually all prescription hypnotics have been reported to be more beneficial than placebo in some combination of these measures. Studies of longest duration in insomniacs have suggested that flurazepam increases total sleep for at least 28 days, but even this is shorter than the duration of the average hypnotic prescription. Studies in normal volunteers have also reported increased total sleep with chloral hydrate for 28 days and with nitrazepam for over 10 weeks. Whether this translates into effectiveness for insomniacs is not known. There are virtually no available studies of efficacy during intermittent usage, which may well be the most common way these agents are taken in practice. This problem is compounded by the almost complete lack of longitudinal data on the natural course of patients with various types of insomnia.

It is not at all clear which patients should be given hypnotics. Most investigators agree that hypnotics have limited usefulness when insomnia is accompanied by an associated disorder such as nocturnal myoclonus or phase lag syndrome. It has been argued that among the remaining patients with “primary” insomnia, only those with objective EEG evidence of disturbed sleep should be given hypnotics; others believe that the subjective experience of insomnia sufficiently justifies the use of hypnotics. The former argument appeals to scientific rigor, but at present limits the prescribing of hypnotics to a minority of patients. Another difficulty is that our ability to define objective disturbances of sleep is changing constantly. As but one example, the use of multivariate analysis may rapidly increase the percentage of patients who can be objectively identified as insomniacs. The second problem is that many drugs, including tranquilizers, antidepressants, and analgesics, are often given on the basis of the patient’s reports of his subjective feelings and behavior; it is not clear that an electronic measure of disturbed physiology is needed before a medication is given. These are, however, nothing but speculations, best resolved by the outcome of careful follow-up studies, which are waiting to be performed.

Although systematic hypnotic efficacy studies are generally incomplete and suggest limited usefulness of hypnotics, these agents are of course very widely prescribed and many practitioners claim to see benefits which exceed those documented by investigators. Similarly, several polls of large populations of users report that a large proportion- 80% in one study-believe that their prescription hypnotics help them substantially. The sources of this discrepancy are not clear. It is possible that the patients are experiencing a placebo effect, or that sleeping pills serve some beneficial function in the physician’s relationship to a patient seeking help, or that hypnotics produce some other effect (such as euphoria or tranquilization) which is unrelated to sleep but which places them in high demand. Alternatively, it may be that both patients and many physicians are recognizing some clinical benefit which is not picked up in systematic studies—in short, investigators might not be measuring the right things. There is also a possibility that although benefit to groups of patients are very modest in systematic studies, some individual patients are helped greatly.

In contrast to our very tentative knowledge about the efficacy of hypnotics, evidence is accumulating rapidly about the many hazards of widespread hypnotic use. The importance of this problem was emphasized by a prospective population study now several years old which suggested that persons with no history of medical illness but who used hypnotics often were 1.5 times more likely to die in the next 6 years than those who never used such drugs. The negative aspects of hypnotic use which have received the most attention are toxicity in acute overdose and drug addiction. Although these are very real problems, it is essential not to lose sight of others which are often less dramatic but still very important. These include interactions with other drugs (including ethanol), daytime residual effects, and enhanced toxicity in the elderly.

The interaction of hypnotics with ethanol must be carefully assessed, as they are frequently taken together in both routine use and overdose. There is a growing literature on the relation of ethanol-hypnotic interactions on driving skills. Although the interpretation of these studies is difficult, the very least that can be said is that many drivers who are arrested or are involved in accidents have been taking both ethanol and a hypnotic or anxiolytic, and it behooves us to understand the effects of such combinations. Studies of ethanol-hypnotic interactions indicate that performance may be affected variably, depending on which measure is used, dose, duration of administration, and other factors. The potential for adverse interactions also varies for different drugs in the same chemical class. Diazepam, for instance, seems more clearly to interact with ethanol than do some other benzodiazepines such as chlordiazepoxide or oxazepam. There is relatively little work on the combination of ethanol with flurazepam, although this drug accounts for over half of all hypnotic prescriptions.

Since the half-lives of many hypnotics are substantially longer than 8 hr, one consequence of taking hypnotics is that a patient may have substantial amounts of active drug in his body during the daytime. Hypnotics may alter the waking EEG up to 18 hr after administration; although the clinical significance is uncertain, this at least demonstrates that these agents continue to have physiological effects well into the next day after bedtime use. Tests of psychomotor function show a variety of deficits at least 16–19 hr after a single dose, and over the entire day during chronic administration. Although the relation of such tests to practical skills such as driving is not entirely clear, these results certainly raise the possibility that persons who take hypnotics may have decreased performance in practical daytime chores. Perhaps most worrisome is the finding that persons with drug-induced performance deficits may be subjectively unaware of their impairment.

The elderly may be particularly at risk for deleterious residual effects or toxic reactions to hypnotics. This may stem from increased nervous system sensitivity, as well as altered absorption, distribution, and excretion of drugs. Decreased clearance and/or increased volume of distribution of many hypnotics may lead to prolonged plasma half-lives of these agents in the elderly. Any clinical difficulties associated with this would seem likely to be magnified with the traditional benzodiazepines such as flurazepam, which even in young adults have both psychoactive metabolites and substantially longer half-lives than most other hypnotics. Empirically, the incidence of toxic reactions to flurazepam does increase with age, and in one study was a problem in one-fourth of nursing home patients receiving it. Disinhibition reactions in the elderly resulting in restlessness and aggression have long been associated with the barbiturates but may occur in all classes of hypnotics. All of these problems are compounded by the disproportionately high rate of use of hypnotics by the elderly. In contrast, hypnotics are tested most extensively in young adults. Thus we tend to test hypnotics in one generation and then administer them to another.

All of the hazards of hypnotic use would seem to be enhanced by the tendency for many persons—perhaps 15–20% of those started on hypnotics—to continue to take them for prolonged periods. A large number of these chronic users—over 20% in one study—were started on these agents while hospitalized for medical, surgical, or psychiatric disorders. There is also some evidence that prolonged usage may develop in patients who seem medically and psychologically healthy, but are given hypnotics during a period of stress. Tentative data suggest that patients started on those hypnotics which are not generally considered to have substantial abuse potential, such as nitrazepam, may be as likely to desire long-term use as those started on amobarbital. The factors which contribute to such long-term use are not clear, and need to be studied. As but one example, it is uncertain to what degree disturbed sleep from withdrawal of clinical doses contributes to a desire to continue medication, nor is it documented how frequently such sleep disturbance occurs.

Alternative approaches to the use of prescription hypnotics include OTC hypnotics and nonpharmacologic treatments for insomnia. The OTC hypnotics are used perhaps even more widely than their prescription counterparts. The few efficacy studies which are available show mixed results, and studies of what until recently have been the two major ingredients have indicated a carcinogenic risk from one (methapyrilene) and a tendency for toxic anticholinergic effects from the other (scopolamine aminoxide). Developments in the large and lucrative market for OTC hypnotics should be carefully followed.

The most common nonpharmacologic approaches to insomnia are the traditional forms of psychotherapy and behavior modification techniques. Although the experience of many psychotherapists leads them to believe in the efficacy of their approach, there are no available systematic studies demonstrating substantial benefits to insomniacs. Many of the behavioral approaches to treating insomnia are based on the questionable hypothesis that excessive autonomic stimulation is important in keeping patients from sleeping. Systematic studies of techniques to enhance relaxation such as EMG biofeedback, autogenic training, and progressive relaxation suggest that they are of little benefit to groups of insomniacs. Two techniques which may have the most promise are stimulus control and behavioral self-management, although they first need to pass the test of controlled trials. One of the difficulties in this area has been that most studies were done on groups of insomniacs who have not been subjected to careful diagnosis of associated conditions (e.g., sleep apnea) unlikely to be helped by these techniques. Thus it may be that some nonpharmacologic methods are beneficial to selected patients, but that these benefits are obscured in studies of heterogeneous groups of insomniacs.

We are faced, then, with large numbers of Americans seeking help for insomnia, while the therapeutic armamentarium, although varied and widely used, has little scientific support. Our meager understanding of insomnia and of the possible benefits of available hypnotics cannot justify a conclusive set of recommendations for the use of hypnotics at this time. It might be well, however, to end by briefly summarizing a practical approach to the insomniac patient and some tentative conclusions on the use of hypnotics.


Sleep Apnea Depressive Illness Nursing Home Patient Autogenic Training Acute Overdose 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Plenum Publishing Corporation 1980

Authors and Affiliations

  • Wallace B. Mendelson
    • 1
  1. 1.National Institute of Mental HealthBethesdaUSA

Personalised recommendations