NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

AHCPR Health Technology Assessments. Rockville (MD): Agency for Health Care Policy and Research (US); 1990-1999.

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

Cover of AHCPR Health Technology Assessments

AHCPR Health Technology Assessments.

Show details

Polysomnography and Sleep Disorder Centers

Health Technology Assessment Reports 1991, Number 4

, M.D.

Created: .


The Office of Health Technology Assessment (OHTA) evaluates the risks, benefits, and clinical effectiveness of new or unestablished medical technologies. In most instances, assessments address technologies that are being reviewed for purposes of coverage by Federally funded health programs.

OHTA's assessment process includes a comprehensive review of the medical literature and emphasizes broad and open participation from within and outside the Federal Government. A range of expert advice is obtained by widely publicizing the plans for conducting the assessment through publication of an announcement in the Federal Register and solicitation of input from Federal agencies, medical specialty societies, insurers, and manufacturers. The involvement of these experts helps assure inclusion of the experienced and varying viewpoints needed to round out the data derived from individual scientific studies in the medical literature.

After OHTA receives information from experts and the scientific literature, the results are analyzed and synthesized into an assessment report. Each report represents a detailed analysis of the risks, clinical effectiveness, and uses of new or unestablished medical technologies. If an assessment has been prepared to form the basis for a coverage decision by a Federally financed health care program, it serves as the Public Health Service's recommendation to that program and is disseminated widely.

OHTA is one component of the Agency for Health Care Policy and Research (AHCPR), Public Health Service, Department of Health and Human Services.

  • Thomas V. Holohan, M.D.
  • Director
  • Office of Health Technology Assessment
  • J. Jarrett Clinton, M.D.
  • Administrator, AHCPR
  • Questions regarding this assessment should be directed to:
  • Office of Health Technology Assessment
  • Executive Office Center, Suite 400
  • 2101 East Jefferson Street
  • Rockville, MD 20857
  • (301)227-8337


Sleep disorders are common in the general population. On the basis of a survey conducted in 1979 and extrapolated to the general population, about 50 million adults in the United States have difficulty sleeping at night and are sleepy during the day. Approximately 10 million Americans discuss their sleep disturbances with a physician each year, and about 5 million receive sleeping pill prescriptions.(1). Other groups, such as shift workers and long-distance drivers, face the risk of poor productivity and accidents due to sleepiness and loss of alertness. To highlight the importance of sleep disorders, a National Commission on Sleep Disorders Research was established in 1988 by an act of Congress. The purpose of this commission is to study sleep disorders and develop a long-range plan for the use and organization of national resources to deal effectively with sleep disorder research and medicine.(2)

In addition, several professional associations have been formed, including the Association of Professional Sleep Societies, the Sleep Research Society, the American Sleep Disorders Association (ASDA), the Association of Polysomnographic Technologists, the Academy of Sleep Disorders Medicine, and the American Board of Sleep Disorders Medicine. One bimonthly journal (Sleep) and one annual abstract volume (Sleep Research) are devoted to sleep disorders, and hundreds of related articles and abstracts are published annually in the general medical literature. The National Institutes of Health (NIH) has had two consensus development conferences on sleep: a 1983 conference entitled "Drugs and Insomnia"(3). and a 1990 conference entitled "The Treatment of Sleep Disorders in Older People."(4).

A number of different types of sleep disturbances will be discussed in detailed below. At the present time, some of the more complex and serious sleep disorders are evaluated in sleep disorder centers (SDCs). The purpose of this assessment is to assess the usefulness of SDCs in diagnosing and treating certain disorders of sleep.


Normal Sleep

Sleep, like many other physiologic processes when normal, appears natural and effortless. However, underlying normal physiologic processes such as sleep, kidney function, hematopoiesis, and heart function are an enormous array of interacting complex feedback mechanisms and controls. Only when an abnormality appears is one made aware of such substructures, and often the analysis of abnormalities of these systems gives a clue to the normal submechanisms that allow the process to function normally.

Although sleep and its disturbances have been the subject of many ancient literary allusions and comments, the modern scientific study of sleep dates back only 40 or 50 years. A very large body of literature on this subject now exists. Sleep research was made possible in part by the development of sophisticated monitoring devices capable of recording a variety of electrical and physiological parameters from the sleeping subject. The application of such recording devices to the study of sleep disturbances has evolved in an organized manner only during the past 30 years.

"Sleep" is a general word that defines and describes a certain state; as will be seen below, sleep is not a single entity but rather several different physiologic processes grouped together under the term "sleep".

According to a simple behavioral definition, sleep is a reversible behavioral state of perceptual disengagement from and unresponsiveness to the environment. It is also true that sleep is a very complex amalgam of physiological and behavioral processes. Sleep is usually (but not necessarily) accompanied by postural recumbency, quiescence, closed eyes, and all the other indicators one commonly associates with sleeping. In the unusual circumstance, other behaviors can occur during sleep. These behaviors may include sleepwalking, sleeptalking, toothgrinding, and other physical activities. Anomalies involving sleep processes also include intrusions of the sleep processes --sleep itself, dream imagery, or muscle weakness --into wakefulness.(5)

Sleep consists of two distinct states: rapid eye movement sleep (also known as REM sleep, D-sleep, paradoxical sleep, or dreaming sleep) and non-REM or NREM sleep (also known as S-sleep, orthodox sleep, or slow-wave sleep).

Rapid eye movement sleep is characterized on a psychological level by dreaming and on a physiologic level by cortical activation (a mixed-frequency, low-voltage electroencephalographic (EEG) brain wave pattern), bursts of extraocular and middle ear muscle activity, variability of heart and respiratory rates, actively induced atonia of major antigravity and locomotor muscles, increased cerebral blood flow, and, in many instances, increased activity of some individual neurons. In short, REM sleep is a very active brain in a paralyzed body. In the normal adult, 20-25 of the total night's sleep, or about 90-120 minutes per night, is spent in REM. Rapid eye movement sleep occurs in approximately three to five regularly spaced periods, which begin about 70-100 minutes after sleep onset and occur at intervals of about 90 minutes from the onset of one period to the next.

Non-REM sleep is usually subdivided into four stages on the basis of relatively distinguishable EEG brain wave patterns. Stage 1, a brief transitional stage between wakefulness and sleep, is about 5 of the total night's sleep and has a low-voltage, mixed-frequency EEG pattern. Stage 2, defined on the basis of sleep spindles and K complexes, usually constitutes 40-60 of total sleep in the young adult. Stages 3 and 4 are often referred to as delta sleep because they are characterized by moderate to large numbers of slow delta waves. Most stage 3 and stage 4 sleep occurs during the first 1-3 hours of the night in normal young adults. Stage 3 and 4 sleep are the deepest levels of sleep; that is, during these stages the person is most difficult to arouse.

Normal (nocturnal) sleep invariably begins with NREM sleep. As a person falls asleep, he or she enters stage 1, then stage 2, and finally stages 3 and 4. After sleeping for about 1-1/2 hours, the individual enters the first period of REM sleep, which is usually brief (5-15 minutes). The NREM/REM cycle then begins again and is repeated throughout the night. As noted, most stage 3 and stage 4 sleep occurs during the first one or two cycles. Until about the age of 45, growth hormone secretion occurs during the first or second NREM phase, usually in association with delta sleep.

Sleep patterns change with aging. Newborns, for example, sleep about 55-70 of each 24-hour period, half of which may be REM sleep. Sleep stages at this age are not well defined, nor are circadian rhythms. By contrast, in middle and old age, the deepest levels of sleep (stages 3 and 4) markedly diminish as sleep becomes fragmented with periods of arousal and wakefulness.(6)

The particular arrangement of those various stages and conditions of sleep in a person is called the "sleep architecture," and characteristic changes in sleep architecture can be observed in various sleep disorders. The exact function and purpose of each of these different sleep stages are unknown, as is the reason for sleep in general. Restorative theories suggest that sleep is necessary to restore a deficit in body and cerebral tissue that builds up after a period of wakefulness.

Alternatively a sleep-inducing substance may accumulate during wakefulness. Other theories suggest that sleep fulfills an adaptive response of animal and human behavior to the day/night cycle. For example, sleep might be a behavioral mechanism that promotes survival by ensuring temporal integration of a species in its ecological niche. The diurnal placement of sleep might prevent foraging at times when it would be a waste of energy or might incur predators. The reduced responsiveness in sleep blocks potentially harmful reactions to external stimuli that could occur if the animal were merely resting. "Keeping us out of harm's way" makes nonresponding necessary for survival. These theories are certainly not mutually exclusive. Thus sleep serves a number of social, adaptive, and physiologic functions that may be necessary for the survival of the organism.(7)

The two main factors that are thought to control cycles of sleep and wakefulness are the circadian rhythm and certain homeostatic mechanisms.(8) Animals and people living without any external time cues continue to function rhythmically. People living in time isolation, for instance, adopt a roughly 25-hour day. A number of physiologic functions are synchronized to this day length, and seemingly to each other, including sleep and wakefulness, core body temperature, and adrenocorticotropic hormone and cortisol secretion. Such rhythms are called circadian (occurring in approximately 1 day) rhythms. The maintenance of this rhythmicity in the absence of environmental cues suggests the presence of some internal clock within the organism that can keep time on its own.

The circadian rhythm, however, can be influenced by properly timed exposure to light. In addition to the circadian forces, there are thought to be important homeostatic mechanisms controlled by the continual buildup of sleep-inducing chemical factors. An integration of both these systems is necessary for normal sleep.(8) A more detailed exposition of the considerable literature regarding the biochemical, neurophysiologic, and electrical concomitants of sleep is beyond the scope of this assessment. As noted above, knowledge of the field of sleep is in its infancy and much more detailed information will be forthcoming.

Sleep Disorders Classification

There are several organizations that have produced classifications of sleep disorders. The Association of Sleep Disorders Centers and the Association for the Psychophysiological Study of Sleep provided a classification that was published in the journal Sleep in 1979.(9) Because a number of new disorders have been described since 1979 and much additional information about sleep disorders has been obtained, a revised classification was produced by the ASDA in 1990.(10) This manual gives a comprehensive description of the 75 known human sleep disorders. A very condensed classification outline from this manual is as follows:

  1. Dyssomnias
    1. Intrinsic sleep disorders
    2. Extrinsic sleep disorders
    3. Circadian rhythm sleep disorders
  2. Parasomnias
    1. Arousal disorders
    2. Sleep-wake transition disorders
    3. Parasomnias usually associated with REM sleep
    4. Other parasomnias
  3. Medical/psychiatric sleep disorders
    1. Associated with mental disorders
    2. Associated with neurologic disorders
    3. Associated with other medical disorders
  4. Proposed sleep disorders

Other organizations have produced classifications of sleep disorders that differ from the above. For example, the International Classification of Diseases (9th revision; ICD-9) uses the following classifications:

  • 780.5
  • Sleep disturbances (excluding that of nonorganic origin)
  • 780.50
  • Sleep disturbance, unspecified
  • 780.51
  • Insomnia with sleep apnea
  • 780.52
  • Other insomnia
  • 780.53
  • Hypersomnia with sleep apnea
  • 780.54
  • Other hypersomnia
  • 780.55
  • Disruptions of 24-hour sleep-wake cycle Inversion of sleep rhythm Irregular sleep-wake rhythm Non-24-hour sleep-wake rhythm
  • 780.56
  • Dysfunctions associated with sleep stages or arousal from sleep

Sleep disorders of nonorganic origin include the following:

  • 307.45
  • Phase-shift disruption of 24-hour sleep-wake cycle
  • Jet-lag syndrome
  • Rapid time-zone change
  • Shifting sleep-work schedule
  • 307.46
  • Somnambulism or night terrors
  • 307.47
  • Other dysfunctions of sleep or arousal from sleep Nightmares: not otherwise specified REM-type sleep Sleep drunkenness
  • 307.48
  • Repetitive intrusions of sleep (with: atypical polysomnographic features, environmental disturbances; repeated REM-sleep interruptions)

In addition, the Diagnostic and Statistical Manual of Mental Disorders (3rd edition, revised; DSM-III-R) lists the following types of sleep dyssomnias (insomnia disorders of various types); sleep-wake schedule disorders of several types; and parasomnias (dream anxiety disorders, sleep terror disorders, sleepwalking disorders).

Thus there are at present several sleep disorder classifications. Efforts to coordinate and unify these various classifications are in progress (vide infra).

For the purpose of this assessment, only four of the more serious sleep disorders will be discussed in some detail. These are 1) sleep apnea syndromes/sleep-disordered breathing, 2) narcolepsy, 3) parasomnias, and 4) persistent insomnia. Note that the tests used to identify the abnormal characteristics of the disorders mentioned in this section will be discussed in detail in a later section.

Sleep Apnea Syndromes/Sleep-Disordered Breathing


Sleep apnea syndromes and sleep-disordered breathing are classified as dyssomnias. Sleep apnea is defined as cessation of airflow at the nose and mouth for 10 seconds or more during sleep. There are several types of sleep apnea: 1) obstructive apnea occurs when respiratory effort is present with upper airway obstruction preventing airflow; 2) central apnea is the lack of respiratory effort and also airflow; and 3) mixed apnea is a combination of obstructive and central apnea. Hypopnea is a reduction of airflow of 50 or more accompanied by decreased blood oxygen saturation of 4 or more. Some patients with sleep apnea will have several types of apnea. In the opinion of some observers, it is abnormal to have more than five episodes of apnea or hypopnea per hour of sleep; blood oxygen saturation levels below 80 are often observed in such patients. The most prevalent types of apnea in a given patient are used for classification. Hypopneas can be either obstructive or central in nature; both hypopneas and apneas are labeled as disordered breathing events. In comparison with central apneas, obstructive apneas tend to last longer and the oxygen desaturation may be more pronounced during REM sleep than during NREM sleep.

Sleep apnea is particularly likely to develop in overweight, middle-aged men with short, thick necks. Obstructive apnea is also relatively common in elderly men, occurring in up to 30 of this population.(14) Snoring and excessive daytime somnolence are the cardinal features of sleep apnea syndromes. Typically, the snoring is loud and intermittent, with, in the case of obstructive apnea, notable periods of silence despite persistent respiratory efforts.

A major symptom of sleep apnea is severe daytime sleepiness, which correlates best with the degree of disruption of sleep caused by repetitive episodes of apnea. The tendency to fall asleep inappropriately during the day has serious social and economic consequences and is a life-threatening complication for those who operate motor vehicles or industrial equipment.

The patient may experience other troublesome although less common symptoms, such as morning headache, loss of libido, memory deficit, depression, and intellectual deterioration. Systemic hypertension, pulmonary hypertension, and cardiac rhythm disturbances --such as cyclic bradycardia and tachycardia or ventricular premature contractions -- reflect responses to repetitive episodes of apnea and oxygen desaturation. Hypoxemia, hypercapnia, and cor pulmonale may complicate the late stages of the obstructive sleep apnea syndrome, especially when obesity and chronic lung disease are present. The obese patient with daytime sleepiness, sleep apnea, and cor pulmonale has been referred to as "pickwickian," in reference to the character in Dickens' Pickwick Papers.(12). Death can result from severe episodes of apnea.

Sleep apnea is thus a potentially serious medical problem. After proper diagnosis, the patient can often be successfully treated with one of a variety of procedures and/or drugs.

Narcolepsy, which is classified as a dyssomnia, (11,15-17 is a syndrome of excessive sleepiness and intrusion of fragments of REM sleep into the wakeful state. It is a distinct neurological disorder withcharacteristic clinical symptoms and sleep-laboratory findings. Narcolepsyassociated with cataplexy (see below). is often called "classic narcolepsy."(16). Narcolepsy is strongly linked to certain class II HLA antigens, indicating the presence of a "narcolepsy-susceptibility" gene on chromosome 6. Pharmacological and pathological studies of human and canine narcolepsy have demonstrated abnormalities of monoaminergic and cholinergic functioning in the brain. The prevalence of narcolepsy in the United States is estimated variously to be between 1 in 1,000 and 1 in 10,000; it is twice as common as multiple sclerosis. Men and women are affected equally. Narcolepsy is neither psychogenic nor a form of epilepsy, and for the most part it occurs independently of other diseases. Although narcolepsy can begin before the age of 10 or as late as the age of 50, gradual onset between 15 and 35 years of age is usual. Changes in the sleep schedule, psychological stress, head trauma, and infections sometimes appear to precipitate symptoms, but these events may be coincidental.(16)

The two most important symptoms of narcolepsy are irresistible sleepiness and cataplexy. As with normal sleepiness, narcoleptic sleepiness is most apparent in boring, sedentary situations and is partially alleviated by motor activity and mental stimulation. It differs from normal sleepiness, however, in that it cannot be fully relieved by any amount of sleep. The severity of sleepiness in different patients varies considerably. Excessive urge to sleep is more characteristic of the disorder than increased total sleep time, and periods of daytime sleep are usually brief.

Cataplexy, muscular weakness usually brought on by excitement or emotion, is virtually unique to narcolepsy. Laughter is the most common precipitant, but other forms of emotion and athletic activities often induce cataplexy. Severe attacks cause complete paralysis except for the respiratory muscles, whereas the more common partial episodes may cause patients to drop objects or to sit down or stop walking. Momentary attacks are common, and most last less than a minute. Sleep paralysis (defined as brief episodes of inability to move during the onset of sleep or on awakening) and hypnagogic hallucinations (hallucinatory experiences that accompany the onset of sleep or awakening) occur in about 60 of persons with narcolepsy. Along with sleepiness and cataplexy, they constitute the classic clinical tetrad of this disorder. During sleep paralysis, which lasts up to 10 minutes, the patient may feel as if he or she is struggling to move; hallucinations may accompany the paralysis or occur independently and have a dreamlike quality, although some awareness of the surroundings is preserved. These hallucinations are usually visual, but they can have auditory or tactile components.(16)

Narcolepsy is a serious medical problem. The effects of narcolepsy on the patient's life are substantial; 60-80 of patients with narcolepsy report that they have fallen asleep while driving, at work, or both. Persons with narcolepsy often have a preference for shift work or night work, perhaps because on-the-job drowsiness is more acceptable under such circumstances. Treatment at present is limited primarily to combinations of various drugs as well as changes in sleeping patterns, such as regular employment of naps.


Parasomnias are a group of conditions that represent undesirable or unpleasant occurrences during sleep. Behavior during these times can often lead to damage to the surroundings and injury to the patient or to others.

Among these conditions is sleepwalking. Sleepwalking consists of a series of complex behaviors that are initiated during NREM slow-wave sleep and result in walking during sleep.(18)

Episodes can range from simple sitting up in bed to walking, and even to apparent frantic attempts to escape. The patient may be difficult to awaken, but when awakened often is mentally confused. There usually is amnesia for the sleepwalking event which occurs during the first third of the night or during other times of increased slow-wave sleep. The motor activity may terminate spontaneously or the sleepwalker may return to bed, lie down, and continue to sleep without reaching alertness at any point. Sleep talking can also be observed during these events.(19)

During sleepwalking, inappropriate behavior (especially in children) such as urinating in a closet can occur. Sleepwalking can result in falls and injuries. Physical harm can result from the attempt to "escape" or simply from walking into dangerous situations; walking out of a door into the street or through a window is not uncommon. Rarely, homicide or suicide during an apparent sleepwalking episode has been reported. The person attempting to awaken the patient can be violently attacked. Other parasomniac activity, such as sleep terrors, can also occur in patients who are sleepwalkers.(19)

Sleepwalking episodes can occur as soon as a child is able to walk, but such episodes reach a peak prevalence between ages 4 and 8 and usually disappear spontaneously after adolescence. They can occur several times a week or only when precipitating factors are present. Sleepwalking can also be observed in adults.

Another condition, sleep terrors, can present a slightly different constellation of behaviors. Sleep terrors are characterized by a sudden arousal from slow-wave sleep with a piercing scream or cry, accompanied by autonomic and behavioral manifestations of intense fear. There is a severe autonomic discharge, with tachycardia, tachypnea, flushing of the skin, diaphoresis, mydriasis, decreased skin resistance, and increased muscle tone. The patient usually sits up in bed, is unresponsive to external stimuli, and, if awakened, is confused and disoriented. Amnesia for the episode occurs, although sometimes there are reports of fragments or very brief vivid dream images or hallucinations. The episode may be accompanied by incoherent vocalizations or micturition.(20)

Attempts to escape from bed or to fight can result in harm to the patient or others. Psychiatric evaluations of adults indicate that psychopathology may be associated with sleep terrors. Children with sleep terrors do not have a higher incidence of psychopathology than the general population.

Typically, sleep terrors are observed in children between the ages of 4 and 12, and, as in sleepwalking, they tend to resolve spontaneously during adolescence. This condition is more unusual in adults.

In rare cases, loss of REM sleep atony results in the syndrome of REM sleep behavior disorder.(21) Patients who have this disorder seem to act out their dreams literally. Unintentional injury to self or a bed partner may result. The reporter precipitants of this disorder include neurologic lesions, adjustment disorders, post-traumatic stress disorder, and discontinuation of drugs that suppress REM sleep. All of the above may lead to undesirable outcomes.(22)

Persistent Insomnia

Persistent insomnia (lasting more than 6 months) is classified as a dyssomnia. Insomnia is a perception by patients that their sleep is inadequate or abnormal. This symptom includes difficulty initiating sleep, frequent awakenings from sleep, a short sleep time, early awakening, and "nonrestorative" sleep. Epidemiologic studies have estimated the number of people in the United States who complain of insomnia to be in the millions.(23) Insomnia is not a single entity but rather a symptom, and therefore, an accurate differential diagnosis is essential to effective management of patients.(18,24,25). The symptom of insomnia can be associated with a variety of medical disorders, psychiatric disorders, or specific sleep disorders.

It should be noted that individuals may vary considerably in the amount of sleep they need to function normally, as well as their preference as to when to sleep, e.g., "larks" vs "owls." Hauri and Esther(26). have outlined conditions associated with insomnia:

  1. Stress, psychophysiologic, situational, social
  2. Psychiatric disorders
  3. Use of drugs and alcohol
  4. Respiratory impairments during sleep
  5. Periodic movements of sleep (PMS) and restless legs syndrome
  6. Other medical (e.g., chronic pain), toxic, or environmental conditions

In addition, some individuals think they have insomnia but after objective testing are found to have normal sleep. An individual patient may have a combination of several causes of insomnia, and the insomnia can be mild, moderate, or severe. The latter describes a complaint of not being rested after what the patient perceives to be an insufficient amount of sleep. Severe insomnia is associated with feelings of restlessness, irritability, anxiety, daytime fatigue, tiredness, depression, and sleepiness; the symptoms of severe sleepiness can produce marked impairments of social and occupational functions. Insomnia is more common in elderly patients.(27) Thus, insomnia, especially persistent insomnia, associated with severe daytime sleepiness, is a serious and complex medical problem the cause of which requires an accurate diagnosis for successful treatment.

According to proponents of polysomnography (PSG), evaluation of such patients' need for diagnostic testing by PSG requires that their insomnia meet certain criteria: 1) the insomnia has been present for a minimum of 6 months, at least 4 nights a week; 2) it has not responded to behavioral/sleep hygiene 3) the insomnia has not responded to a therapeutic trial with sleep-promoting medication or other psychotropic agents such as antidepressants or such medication is contraindicated; and most important, 4) a medical or psychiatric etiology has been excluded (vide infra).

Other Sleep Disorders

There are a number of other sleep disorders that lead to lack of adequate sleep and symptoms of excessive daytime sleepiness. (Some of these were also included under conditions associated with insomnia, see p.5.) These include periodic limb movements during sleep (PMS), (28). which characterized by periodic episodes of repetitive and highly stereotyped limb movements that occur during sleep. The movements occur typically every 20-40 seconds, most often in the legs, and consist of extension of the big toe in combination with partial flexion of the ankle, knee, and sometimes hip; these movements can also occur in the upper limbs. The movements are often associated with a partial arousal or awakening; however, the patient is usually unaware of the limb movements or the frequent sleep disruption.

Another syndrome that can be quite bothersome usually occurs just before sleep and is called the restless legs syndrome.(28) It is disorder characterized by disagreeable leg sensations, usually prior to sleep onset, that cause an almost irresistible urge to move the legs. The most characteristic feature is the partial or complete relief of the sensation with leg motion and return of the symptoms upon cessation of leg movements. The sensations and associated leg movements usually interfere with sleep onset. A variety of words may be used to describe the sensations, usually including "ache," "discomfort," "creeping," "crawling," "pulling," "tingling," or "itching. "

The sleep-wake cycle also has an important influence on epilepsy.(29-31) This has led to the classification of epilepsies as awakening (diurnal), sleep (nocturnal), and diffuse (random). Sleep is an important activator of potentially epileptogenic EEG activity and seizures. Any type of seizure can occur during sleep, but there are certain types that are more likely, such as generalized tonic-clonic seizures (grand mal), partial seizures with motor symptomatology (partial or focal motor), and partial seizures with complex symptomatology (complex partial).

Epileptic episodes may cause awakenings from sleep, but they rarely cause a primary complaint of insomnia. Frequent episodes during sleep may disrupt sleep to the extent that excessive sleepiness may result and be exacerbated by sedating anticonvulsants. Sleep-related disorders associated with motor activity, such as sleepwalking or sleep terrors, occasionally have an epileptic etiology. Rarely, epilepsy may also cause respiratory difficulty during sleep, which should be distinguished from other causes of sleep-related breathing difficulties.

Sleep disturbances accompany many mental illnesses and may influence their development and course. Sleep abnormalities differ among psychiatric diagnoses, though particular clusters of specific sleep problems may characterize a variety of mental illnesses. A higher percentage of individuals with either insomnia or hypersomnia suffer from a mental illness compared with normal individuals. Sleep disturbances of particular kinds may provide markers of biological processes contributing to the development of mental disorders; conversely, in certain disorders, the sleep disturbances themselves may contribute to the psychiatric manifestations. Specifically, the general neurophysiologic mechanisms involved in the control of REM and NREM sleep have important implications for disorders associated with sleep-onset REM periods or short REM latency, including depression, obsessive-compulsive disorder, schizophrenia, and alcoholism. The treatment of sleep disturbances may aid in or improve the management of mental illness.

Diagnosis and Evaluation of Sleep Disorders

First, an in-depth sleep history is obtained from the patient. This should include a review of the patient's activities over an entire 24-hour period. In addition, a formatted questionnaire may be used. A detailed description should also be recorded of abnormal events that may or may not be disturbing to the patient during sleep such as night terrors, respiratory disturbances, seizures, cardiovascular abnormalities, gastrointestinal disturbances, enuresis, and so on. Information taken from other family members may be essential, particularly for revealing events during sleep of which the patient may not be aware.

In addition to understanding the primary complaint and the pattern of sleep and wakefulness, the clinician must have a clear knowledge of the patient's general physical health and the use of medication, alcohol, or drugs.

Psychological testing with the Minnesota Multiphasic Personality Inventory (MMPI) and other psychometric instruments that measure depression and anxiety may be indicated. A complete physical examination and, when indicated, a thorough psychiatric evaluation should also be performed. Finally, a detailed 2-week sleep log that documents time of going to bed and estimates the number of hours of sleep and the time of awakening should be obtained.(32) Advocates of PSG believe that if there is enough evidence of significant sleep disturbance, referral to an SDC may be considered.(32)

The SDC is a medical facility that allows the simultaneous monitoring and recording of several different physiological processes that occur as a patient is going to sleep, during sleep, and on awakening. The basic concept, according to the proponents of SDCs, is that since the patient is suffering from a complaint about sleep, the SDC allows for a detailed objective measurement and examination of the various processes that occur just prior to and during sleep and awakening.(33). The disorders of sleep mentioned above, according to the proponents of SDCs, often have characteristic findings that can be detected by the tests performed in an SDC. The data obtained from SDC examinations are then correlated with the history and physical findings to help provide an accurate diagnosis of the patients' problems, which is essential for appropriate therapy.

The typical physical equipment for an SDC includes: quiet sleeping rooms equipped with two-way intercom and low-light or infrared video monitor and recording instruments; polygraphs with 10-to 20-channel capabilities for recording inputs from several different devices; a work area where laboratory data are processed and stored; an office from which the patients' evaluations are coordinated; and bathroom and shower facilities for patient convenience.

The staff required includes several technicians, for both daytime nd nighttime recording, and a clinical coordinator. A physician medical director with experience in sleep medicine and physiology is responsible for the whole unit. Such physicians usually have training in psychiatry, neurology, or pulmonary medicine.

Physiologic functions monitored during sleep in SDCs include measurement of electrical activity of the brain using the multichannel EEG; measurement of eye movement using electro-oculogram; and measurement of muscle tone and muscle movement using the electromyogram (EMG). All of these are highly useful measurements of the stages and depth of sleep. Respiration at the nose and mouth are recorded with thermistors that measure the temperature of the air stream; strain gauges measure movements of the chest or abdomen. The electrical activity of the heart and heart rate are measured by the electrocardiogram (ECG). For special purposes esophageal pressure and blood pH can also be measured. The degree of oxygenation of the blood is measured by an oximeter attached to the ear or finger. Pupillary diameter, which changes during different stages of sleep, can also be measured with specialized equipment. If necessary, rectal temperatures can be recorded with a probe placed in the rectum. When some or all of these tests are done simultaneously, the procedure is called polysomnography. The interpretation of all of the above tests, which are recorded simultaneously by pens on a moving paper strip, gives an accurate picture of the onset and relationships between stages of sleep and depth of sleep, eye movements during sleep, frequency of apneic attacks, brain electrical activity, muscle tone, heart activity, blood oxygen levels, abnormal movements, etc.

Another test often done in the SDC is the multiple sleep latency test (MSLT) (34) This test quantifies in a standardized and objective manner the degree of daytime sleepiness. This test is optimally performed just after an overnight PSG in an SDC, because disturbances in the quality or quantity of sleep the night before will influence the results obtained on the MSLT and may lead to a false-positive result. The MSLTs are performed at 2-hour intervals beginning at 9:30 or 10:00 a.m., and measure the patient's tendency to fall asleep when in a relaxed state in bed. The patient is allowed 20 minutes to fall asleep on each of five nap opportunities. The mean latency to sleep onset is helpful in differentiating patients with normal alertness, who have a mean sleep latency of more than 5 minutes, from those with excessive daytime sleepiness, whose mean latency is less than 5 minutes. The types of sleep that occur are also diagnostically useful: It is quite common in narcolepsy to observe two or more sleep-onset REM periods in the five naps in a patient who otherwise had a normal night of sleep preceding the test. The MSLT is purported to be helpful in making a diagnosis of several other types of sleep disorders as well as in assessing the results of treatment.

Depending on the clinical circumstances, all-night PSG and MSLTs are alleged to be necessary for the evaluation of excessive sleepiness, particularly when sleep apnea and/or narcolepsy is suspected, and for the evaluation of severe behavioral disturbances during sleep (e.g., aggressive or assaultive behaviors), particularly when the history suggests REM sleep behavior disorder. Two nights are believed to be generally sufficient for most evaluations; however, three consecutive nights of PSG may be employed especially as part of the evaluation of treatment-resistant chronic insomnia. Proponents recommend that testing should include screening for sleep-disordered breathing, periodic leg movements, and cardiac arrhythmias. In the case of excessive sleepiness, they believe a minimum of one night and one day (for MSLTs) is necessary. Many advocates recommend two consecutive nights of PSG to assess internight variability in severity of sleep-disordered breathing. They believe that a minimum of two consecutive nights is required for the proper physiologic assessment of insomnia in order to allow for first-night effects and to assess internight variability. Typically, two nights for the assessment of sleep-related behavioral disturbances have been recommended, including all-night EEG with the 10/20 montage. This is helpful in the differential diagnosis of epilepsy vs parasomnia.

The purpose of this assessment is to evaluate the usefulness and effectiveness of SDC in the diagnosis and treatment of the several sleep disorders discussed in the previous sections, to determine which patients should be evaluated in SDCs, and finally to determine whether a physician needs to be present during overnight studies in an SDC.

Review of Available Information

This section will be divided two parts: first, a brief overview of the PSG and MSLT findings obtained in SDCs that may be observed in certain sleep disorders; second, a summary of the relevant literature that analyzes the role and usefulness of SDCs for the diagnosis and treatment of sleep disorders compared with such diagnoses and treatments prescribed without the use of SDCs.

Overview of PSG and MSLT Findings

The PSG examination may, according to proponents, reveal characteristic measurements and findings in certain sleep disorders. For example, in sleep apnea the frequency and duration of the apneic periods can be measured as well as the relationship between the cessation of breathing and airflow at the nose and mouth.(13) In addition, the degree of hypoxia during the apneic periods can be measured, and cardiac arrhythmias, if they occur, can also be detected with the ECG.

The degree of sleep disruption produced by the apnea can be documented by the EEG patterns. Stage 3 and 4 deep sleep can be markedly reduced whereas the amount of lighter sleep, stage 1 and 2, is increased. Furthermore, REM sleep can occur earlier than usual. In the case of sleep apnea, the PSG can also provide direct information regarding therapeutic intervention; for example, the amount of air pressure and flow used in intermittent positive pressure breathing can be adjusted and the effect of this therapy can be measured.

In the case of narcolepsy, a common feature is the very rapid onset of REM sleep --within 10 minutes of the onset of sleep.(16) Particular findings using pupillometry have also been described. The MSLT gives the most specific diagnostic findings. Here, the patient will go to sleep within 5 minutes, and REM sleep will be observed in the very early phase of such sleep episodes.

Various parasomnias can also demonstrate abnormalities on PSG. Sleepwalking(19,35). usually begins in stage 3 and 4 sleep, most commonly at the end of the first or second episode of slow-wave sleep, and EEG patterns will not show evidence of epileptiform activity during these attacks.

Sleep terrors(36,37). begin in stage 3 or 4 sleep, usually in the first third of the major sleep episode. However, episodes can occur in slow-wave sleep at any time. Partial arousals from slow-wave sleep without full terror are more frequent in patients with sleep terrors. Tachycardia usually occurs during both clinical episodes and partial arousals. The EEG portion of the PSG can distinguish sleep terrors from sleep-related epilepsy.

Another disorder that can clinically resemble the above is the REM sleep behavior disorder. The PSG feature of this disorder is the loss of muscle atony that is usually found during REM sleep.(21) This is identifiable on the EMG portion of the PSG. Because of normal muscle function, the patient can actually act out the dreams that occur during REM sleep. Thus, the cardinal feature of this disorder is that complex and vigorous behavior of the patient occurs during REM sleep. The PSG, according to proponents, is therefore highly useful in distinguishing sleepwalking and sleep terrors from REM sleep disorders; this differential diagnosis is claimed to be a useful guide for therapy.(21)

Epileptic attacks occurring at night can also occasionally produce symptoms that resemble the above states. The EEG portion of the PSG may be needed to make the differential diagnosis in such patients, since the sleep-wake cycle can also influence epilepsy and certain types of epilepsy can be seen more commonly during sleep. Careful EEG recordings with more channels during PSG are useful in revealing the exact type of epilepsy that is occurring.(29-31)

Another disorder that has specific PSG findings is the periodic movement disorder.(28). Here, periodic limb movements can appear immediately with the onset of NREM stage 1 sleep, are frequent during stage 2 sleep, and decrease in frequency in stage 3 and 4 sleep. The periodic limb movements are usually absent during REM sleep. The EMG shows repetitive muscle contractions lasting 0.5-5 seconds approximately every 20-40 seconds. Flurries of movements can last from a few minutes to several hours.

Studies in SDCs have also been used for the diagnosis of the multiple causes of persistent insomnia. According to proponents, studies in an SDC can rule out insomnia, since a small percentage of patients may complain of insomnia, but upon testing in an SDC the PSG reveals that the patient slept longer than he or she thought and that the quality and depth of sleep were also normal.

Insomnia can also be the primary complaint in disorders such as sleep apnea(13). and PMS, (28). which may have specific PSG findings. Medical disorders such as chronic obstructive pulmonary disease (COPD) can also lead to insomnia, and the PSG demonstrates a decreased sleep efficiency, decreased total sleep time, increased number of stage shifts and arousals, and decreased delta and REM sleep. Typically, hypoxemia occurs and is worsened during REM sleep.(12) However, neither the diagnosis nor treatment of COPD depends on PSG

Insomnia is also commonly seen in mental illness such as depression and anxiety disorders.(18) Here the sleep tracings show abnormalities in sleep continuity, sleep architecture, sleep efficiency, and, of particular note, several measures of REM sleep. In depression, sleep latency is typically prolonged in younger persons; older individuals may fall asleep normally but then experience a continuity disturbance of frequent awakenings that may be related to dreams. Sleep architecture changes consist of reduced delta sleep and increased REM sleep. The delta sleep abnormality often includes a relative shift of delta waves from the first NREM period to later in the night. This shortened first NREM period results in an earlier onset of the first REM period, i.e., a short REM latency, which constitutes the most characteristic feature of the PSG findings. Nevertheless, as with medical conditions that do not have sleep-dependent physiologic changes and that have adequate diurnal features for accurate diagnosis, diagnosis and selection of appropriate treatment for depression may be helped by, but are not dependent on, PSG.

Another form of insomnia is psychophysiologic insomnia, (23). a disorder of somatized tension and learned sleep-preventing associations that results in a complaint of insomnia and associated poor functioning during wakefulness. On PSG the usual features indicating objective insomnia are found, such as increased sleep latency, increased wakefulness after sleep onset, and decreased sleep efficiency. There is increased stage 1 sleep and, possibly, a decrease of delta sleep. There may be increased muscle tension, and increased EEG alpha waves may be observed. As with other disorders mentioned above, the role of PSG in primary diagnosis or treatment selection is not established.

Summary of Relevant Literature

A number of studies have used PSG to investigate sleep disorders in SDCs. Some features of these representative papers are summarized in the Table. Several points regarding the studies are important. No prospective studies of the clinical utility of PSG have been published and the available data are all retrospective. Scientific study of sleep disorders is a relatively new field, with organization of individuals in this field not occurring until the late 1970s. In part because of the lack of centralized direction and funding of trials, clinical studies of patients in SDCs were not coordinated but were instead performed by a multitude of individual groups. In this respect, the use of SDCs does not differ from many currently used medical technologies.

Table. Summary of clinical studies.


Table. Summary of clinical studies.

In a 1982 paper by Coleman et al, (38). a large cooperative study of 5,000 patients at 11 centers was reported; almost all these patients had PSG examinations. However, results on only 3,900 of the 5,000 were reported. The largest single diagnostic category in the study was patients with disorders of excessive daytime sleepiness, accounting for 51 of the 3,900 diagnoses made. The sleep apnea syndrome was the most prevalent diagnosis (22) among the 3,900 patients and accounted for 43 of hypersomnia diagnoses. Most had obstructive sleep apnea. Narcolepsy was both the second most prevalent diagnosis and second most frequent hypersomnia diagnosis. Narcolepsy and sleep apnea together accounted for 68 of the hypersomnia diagnoses and 35 of the cases evaluated. In the diagnoses of narcolepsy, the clinical history of daytime sleepiness, cataplexy, and other auxiliary symptoms was combined with PSG, although the relative contributions of each were not directly assessed.

With regard to insomnia, the etiologic diagnoses included psychiatric disorders (35), psychophysiologic insomnia (15), drug and alcohol dependence (12), myoclonus and restless legs syndrome (12), and no objective insomnia (9). The authors(38). concluded that a history of severe daytime sleepiness persistent for longer than a month requires thorough evaluation since sleep apnea syndrome and narcolepsy are the most likely causes.

In a paper in which persistent insomnia was studied, Zorick et al(25). evaluated 84 patients by medical, psychiatric, and PSG examinations. On the basis of these studies, patients were placed into 10 diagnostic categories. The authors noted that only half the patients had objectively defined difficulty inducing or maintaining sleep when compared with 20 normal control subjects. However, 68 of the 84 patients showed PSG evidence of some sleep disorder. Diagnostic categories within the insomnia groups could be distinguished from one another by PSG. In only 3 of the 10 diagnostic categories did patients show evidence of psychological distress on the MMPI.

Zorick et al(25). stated that sleep evaluation not only results in a diagnostic classification of the insomnia that has treatment implications but may also prevent the use of ineffective or potentially dangerous treatments. They believed that psychiatric, medical, psychological, and PSG evaluation of patients with the complaint of persistent insomnia is necessary for thorough assessment. However, the ultimate implications of the study are unclear, given that of 84 referred chronic insomnia patients, only half had objective evidence on PSG of abnormalities in inducing or maintaining sleep when compared with a smaller normal control group.

In 1982 Kryger et al(39). evaluated sleep breathing disorders in 48 referred patients in a general hospital They concluded that evaluation of breathing during sleep was most useful in those having excessive daytime sleepiness or unexplained polycythemia. The sleep apnea syndrome was documented in 19 of 24 patients with excessive daytime sleepiness, and 10 of 15 patients with unexplained polycythemia were found to have severe hypoxemia during sleep. The authors concluded that sleep breathing problems are not rare and that in their experience, evaluation of breathing during sleep provides objective quantitative documentation of a disorder that is only suspected clinically.

A paper by Hauri et al, (40). published in 1982, evaluated the effectiveness of SDCs in a group of patients referred by their private physicians for consultation and testing by means of followup questionnaires. Both patients and their private physicians were sent a questionnaire 9 months after SDC evaluation. Approximately 30 of patients reported that their sleep problems had improved. Forty-five percent of all physicians thought that the evaluation had been "clearly helpful," the term having been defined in the questionnaire as "indicating new and useful treatments that had not previously been considered." Forty-two percent of all physicians thought that the evaluation had been "somewhat helpful" (defined as confirming previously held diagnostic impressions and/or treatments). The authors did not address confounding issues of referral bias (e.g., referring physicians were more likely to believe in the utility of PSG) nor of recall bias on the part of the patients.

In a retrospective chart review study of chronic insomnia by Jacobs et al, (41). 123 patients were evaluated with or without PSG studies. Patients were evaluated by means of a sleep-wake log, a sleep habits questionnaire, psychiatric and clinical interviews, and a minimum of two consecutive nights of PSG. Sixty-three percent of the sample evidenced Research Diagnostic Criteria-diagnosed psychopathology, and 11 had "medical conditions" believed to be responsible for insomnia. The specific medical conditions were not enumerated, except for examples such as "endocrinopathy." The authors stated that in 49 of the patients, laboratory results added to, refuted, and/or failed to support the clinical impression and that this resulted in substantial modification of the initial diagnostic formulation and increased specificity of diagnosis. The authors admitted that the conclusions that can be drawn from their study must be qualified by the absence of a control group and that the crucial issue (i.e., whether laboratory studies contribute substantially to improved quality of sleep) can be addressed only in a prospective study. However, they believed that their data suggest laboratory measurement should be combined with clinical assessment in the evaluation of insomnia.

In contrast to the above studies, in 1987 Kales and Kales(42). reviewed their experiences(42). in sleep disorder evaluations and concluded without reference to specific quantitative data that patients with sleep disorders can be thoroughly evaluated in the office setting. "A careful sleep history, drug history, general medical assessment, and psychiatric evaluation along with an appraisal of the interplay between the patient's condition and his environment can provide all of the elements needed for diagnosis and treatment formulation." They stated that their experience gave them the impression that in general, sleep laboratory diagnostic studies are of limited usefulness and are indicated primarily when sleep apnea is suspected or when the sleep attacks of narcolepsy are present in the absence of auxiliary symptoms.

A study byt Edinger et al(44). published in 1989 looked into the question of whether PSG should be used in the routine evaluation of chronic insomnia. One hundred patients with insomnia were evaluated on an outpatient basis by a comprehensive medical, psychiatric, and behavioral series of tests and by PSG. An experienced treatment team then evaluated all the medical records except for results of the PSG examination. Fourteen of 25 patients with a PMS disorder were diagnosed clinically, and the additional 11 by PSG. Only one of three patients with apnea was diagnosed without the added information from PSG, and in seven patients a clinical impression of sleep apnea was ruled out by PSG findings. One of six cases of subjective insomnia was predicted by clinical assessment only. However, 44 of this group of patients had psychiatric disorders, and the authors pointed out that psychiatric, psychophysiologic, and substance abuse etiologies could usually be established on the bases of the history, physical examination, and psychological testing. The authors concluded that PSG often yields new findings in insomniac patients not suspected of having sleep apnea, PMS, restless legs syndrome, or subjective insomnia. They also suggested that in patients older than 40, PSG examinations are more likely to reveal abnormal findings.

The questions of whether PSG findings have prognostic significance were evaluated in a study by He et al(45). published in 1988. Three hundred eighty-five patients with obstructive sleep apnea were studied by PSG in an SDC. An apnea index was used to quantify the severity of the disease. In patients younger than 50 years of age, there was a significant positive correlation between a higher apnea index and mortality. The measurement of the apnea index was claimed to be a guide for effective therapy of obstructive sleep apnea such as uvulopalatopharyngoplasty and continuous positive airway pressure.

The usefulness of the MSLT and very early sleep onset REM periods (SOREMPs) was evaluated in a retrospective study of 144 patients with suspected narcolepsy by Amira et al.(46) The MSLT and SOREMPs tests were administered to these patients, who were described as having a "clinical diagnosis" of narcolepsy on the basis of the following criteria: 1) excessive daytime sleepiness plus hypnagogic hallucinations, or sleep paralysis, or cataplexy; 2) "clinical characteristics" such as natural history; 3) "other diagnostic studies" such as EEG or PSG. Eighty of the 144 patients in the study had no episodes of SOREMPs. In this subgroup, the authors determined an "ultimate clinical diagnosis" and concluded that only one patient had narcolepsy, and six remained undiagnosed. Unfortunately, the criteria for the "ultimate" diagnoses were not elaborated. Fifty-two of the 144 patients had more than two episodes of SOREMPs. The authors noted that 51 had "clinically apparent narcolepsy," but they did not further define this group. In all, narcolepsy was diagnosed in a total of 61 patients (42). The authors calculated the sensitivity and specificity of both MSLT and SOREMPs in diagnosing narcolepsy; however, in the absence of specified diagnostic criteria other than the SOREMPs or MSLT, it is not clear precisely what those calculations represent. It is significant that in the group of 61 patients in whom narcolepsy was diagnosed, 10 (16) failed to exhibit two or more episodes of SOREMPs, and 26 (43) had mean sleep latencies greater than 5 minutes. Amira et al concluded that laboratory standards for those tests may be "too exclusionary," and they questioned whether "center-specific norms" might be necessary.

The role of PMS in patients with insomnia (12 patients) or excessive daytime sleepiness (11 patients) was examined in a paper by Saskin et al.(47) All patients had a complete medical examination, detailed sleep history, and psychiatric examination, and they all had PSG performed in an SDC. The patients with insomnia had a greater number of "bursts" (runs of frequent muscle movement) per hour of sleep than did patients with excessive daytime sleepiness. In addition, those patients with more severe PMS showed a greater fragmentation of sleep. Patients with excessive daytime sleepiness had a better sleep efficiency. This paper did not examine the success of treatment strategies based on these data.

Rosenthal et al(48). performed another study of the relationship between PMS and sleep complaints in two groups of patients, 51 who complained of insomnia and 29 who complained of excessive daytime sleepiness. The authors reported that sleepy patients differed from patients with insomnia in that they fell asleep faster, slept longer, and exhibited more frequent arousals (shifts to stage 1 sleep and number of awakenings). Patients with insomnia had more PMS, whereas sleepy patients had more PMS bursts per series. Rosenthal et al concluded that their data suggest that the pattern of sleep fragmentation associated with PMS is consistent with the person's complaints. Infrequent but long awakenings are associated with insomnia complaints, and frequent but brief awakenings or arousals are associated with excessive daytime sleepiness complaints. People who show PMS without arousals or awakenings would be expected to have no sleep-wake complaints. In terms of therapy, these data would suggest that these patients can be managed symptomatically. Drugs (e.g., sedative-hypnotics) that blunt that arousal response might be effective even if they have no direct effect on the frequency of PMS. The authors, however, did not study the effects of treatment in these patients.

A paper by Schenck et al(37 examined PSG findings in 100 consecutive adult patients with sleep-related injuries, such as lacerations and fractures of extremities and vertebrae. Most of the patients were referred to the SDC by private physicians. All patients had a sleep history; medical, psychological, and psychiatric history; PSG; and MSLTs. In this group of patients, there was typically a long delay from onset to diagnosis and treatment, with a range of 2 months to more than 40 years. Ninety-one of the 100 patients were segregated by distinct, positive PSG criteria into five diagnostic categories: 1) night terrors/sleepwalking (complex and/or vigorous behaviors appearing during abrupt arousals from NREM sleep); 2) REM sleep behavior disorder (loss of EMG atonia with excessive limb jerking and complex vigorous behaviors during REM sleep); 3) dissociative disorders (sequences of altered behaviors appearing during EEG wakefulness); 4) nocturnal seizures; and 5) sleep apnea/periodic movements. For the remaining nine patients, the normal (or negative-yield) PSG study results were concluded to be useful in supporting the presumptive clinical diagnoses by excluding various alternative diagnoses. In this study, information was used for therapy; patients with night terrors/sleepwalking and those with REM sleep behavior disorders were successfully treated with appropriate

The NIH sponsored a consensus development conference in 1990 entitled "The Treatment of Sleep Disorders of Older People."(4). Since the topic directly addresses the issue at hand in this assessment, and since the consensus conference process is generally acknowledged to be a comprehensive and an objective review of a controversial topic of diagnosis or treatment, it is appropriate that the conclusions of the panel be reviewed in some detail. The 14 panel members were presented data by 25 of the leading researchers and clinicians in the field over a period of 1-1/2 days. The panel then prepared a "Consensus Statement" representing the conclusions arrived at on due consideration of the information presented.(4) The panel noted that disturbances of sleep are common, affecting more than one-half of people 65 or older who live at home and about two-thirds of those living in long-term care facilities. However, they concluded that "There is little agreement among workers in the field about what is clinically normal and what is clinically abnormal except in extreme cases...." The natural history of certain sleep disorders is not well described (e.g., whether they may spontaneously remit), and it is difficult to determine which diagnostic criteria are important in establishing clinical abnormalities. While severe forms of sleep disturbance such as sleep apnea are generally accepted and criteria are agreed on, mild and moderate clinical entities are not. For example, while PMS is common in the older patient, there is "insufficient evidence at this time to indicate whether or not the disease state or its treatment affects morbidity in the older patient." The panel called for further research in areas of the classification, diagnostic criteria, basic mechanisms, and efficacy of treatment of sleep disorders and concluded that "Current and new therapies and technologies must be evaluated by randomized controlled clinical trials."

The Consensus Statement contains specific reference to the utility of PSG:

Polysomnography is indicated when a sleep related breathing disorder is suspected, and may be useful for certain behavior or movement disorders during sleep. Polysomnography followed by a multiple sleep latency test is useful for establishing the diagnosis of narcolepsy and for quantitating daytime sleepiness. At present, there are insufficient data to assess the value of polysomnography in the routine evaluation of insomnia, depression, or dementia.(4)

An editorial in The Lancet commented favorably on the conclusions of this conference, particularly in terms of increasing physician education as regards sleep disorders. However, the editorial expressed concern over the practicality of an approach to sleep disorders that involves early referral for PSG in lieu of the employment of other "practical, effective responses."(49).

Part of the problem in sleep disorder medicine is that many of the disorders described by the classifications cited above do not have a single "gold standard" absolute diagnostic criteria. Rather the diagnosis of a particular type of sleep disorder has to depend on a conglomeration of many diverse clinical and laboratory findings. In an attempt to remedy this difficulty in the case of insomnia, the National Institute of Mental Health (NIHM) of the Alcohol, Drug Abuse, and Mental Health Administration (ADAMHA) is currently funding a multicenter prospective trial of the reliability of insomnia classifications. Few data exist on the reliability, coverage, and overlap of existing classification systems of insomnia [e.g., the DSM-IV criteria, the ICD-10 criteria, and the ASDA's International Classification of Sleep Disorders (vide supra)]. The goal of this trial is to develop a clinical set of diagnostic criteria for insomnia that will be compatible with more specific definitions and to determine what, if any, information on the results of PSG should be included in the DSM-IV. The study will evaluate differences in interrater reliability among the three classification systems, whether certain diagnoses are associated with recommendations for PSG, what impact PSG data have on these diagnoses, and whether specific diagnoses reliably lead to specific treatment recommendations. Five sites will accrue a total of 250 patients, and the trial is projected to be completed within 24 months.

Responses from the Public Health Service

As part of the assessment process other agencies of the Public Health Service (PHS) were consulted.

The National Institutes of Health

One institute responded that PSG is indicated when a sleep-related breathing disorder is suspected and may be useful for certain behavior or movement disorders during sleep. Polysomnography followed by an MSLT is useful for establishing the diagnosis of narcolepsy and for quantitating daytime sleepiness. The complaint of hypersomnolence may be indicative of a potentially lethal condition (obstructive sleep apnea), a more benign neurological condition (narcolepsy), or a mental disorder such as depression.

Diagnosis of sleep apnea. To evaluate and quantitate sleep apnea objectively and thoroughly, a nocturnal PSG is essential. As subjective response of apnea to treatment can be seen in the absence of objective improvement, close followup is necessary.

Diagnosis of narcolepsy. The differential diagnosis of excessive daytime sleepiness, in addition to narcolepsy, includes sleep apnea, insufficient sleep, the effects of medication, and several other less common disorders such as PMS. Patients with excessive sleepiness and definite cataplexy are often given a diagnosis of narcolepsy on the basis of the history alone; laboratory studies are helpful, however, because they corroborate a diagnosis that can lead to disability and other socioeconomic difficulties, they identify persons who feign the symptoms of narcolepsy to obtain controlled stimulant medications, and they may disclose the presence of other disorders characterized by excessive sleepiness, such as sleep apnea. The MSLT is clinically the most useful technique for objectively quantifying daytime sleepiness.

Diagnosis of parasomnias. Electrophysiologic studies of sleep in patients with sleepwalking are useful for diagnostic purposes, although often the diagnosis can be made from the typical history and age of onset. However, in some situations, differentiation from other disorders that can produce motor behavior during sleep may be necessary.

Diagnosis of REM behavior disorder. The history of prominent, violent, or potentially injurious motor behavior occurring after 90 minutes from sleep onset is highly suggestive of REM behavior disorder, particularly if there is vivid dream mentation reported appropriate to the observed action. Although the diagnosis may be suspected clinically, extensive polygraphic study is mandatory because of the seriousness and legal aspects of injuries related to this behavior.

Diagnosis of insomnia. Use of objective quantifiable laboratory testing is an important adjunct in informed decisionmaking regarding the need for further investigation and-or making specific treatment recommendations. This is in concurrence with the NIH consensus conference on drugs and insomnia in 1983, which recommended that the physician attempt to establish the specific (and possible etiologic) correlates of the patient's complaint of disturbed, inadequate, or excessive sleep, with the goal of formulating specific intervention.

General conclusions and recommendations regarding SDCs. There is no question that this monitoring will substantially improve diagnostic accuracy about whether or not the patient is sleeping and what is disrupting his or her sleep. The next question is, does improved diagnostic accuracy lead to better treatment? While the most effective therapies for some sleep disorders may be controversial and need more study, many other sleep disorders are now clearly treatable once they have been accurately diagnosed. Electrophysiologic monitoring of the sleeping patient is as safe a procedure as such monitoring of the waking patient, which is routinely done in clinical neurophysiology and EEG laboratories around the country without the actual physical presence of a physician during the testing.

Another NIH institute responded that there are limited objective data to substantiate the effectiveness of therapies based on diagnoses obtained by PSG compared with therapies used in patients not having PSG-based diagnoses.

The conclusion that PSG can assist in the diagnosis of sleep-disordered breathing and of certain cases of narcolepsy is appropriate. However, as noted in the report, there are some cases where a diagnosis of sleep apnea is rather straightforward and may not require an extensive PSG workup. Certainly, a presumptive diagnosis can be made based on history of hypersomnolence, family or bed partner observations, physical examinations, and to some extent daytime alertness.

This Institute agreed that there is no compelling case to be made for the use of PSG in the evaluation of insomnia. It is not totally clear how a PSG would be used to delineate the cardiovascular complications associated with sleep apnea.

Sleep disorders centers provide an excellent focal point for patients requiring medical attention for a wide variety of sleep-related problems. However, it has not been firmly established that specialized centers of expertise such as SDCs are the most medically appropriate or cost-effective means of diagnosing, treating, and managing sleep disorders. Other alternative approaches need to be examined compared with SDCs. Unfortunately, there is no standard for diagnosing sleep disorders; the PSG is about the best measure and has been used for the longest period of time. However, the PSG should be compared with alternative low-technology approaches such as questionnaires, home monitoring, patient care plans, and other screening approaches for making initial diagnoses.

The Alcohol, Drug Abuse, and Mental Health Administration

The SDC permits an extensive objective measurement and examination of sleep processes. Polysomnography is widely used and accepted in the diagnosis of narcolepsy and sleep apnea. There is much less consensus on its use in the diagnosis of insomnia. Most studies evaluating the efficacy of diagnostic testing in SDCs are retrospective. A DSM-IV field trial funded by the NIMH involving six SDCs at academic medical centers is now under way. The purpose of this study is to evaluate the reliability of the diagnosis of insomnia by sleep disorder experts and psychiatrists by using DSM-IV and ICD-10 criteria. This prospective, controlled study will provide important information about the impact of PSG on diagnostic reliability and precision.

The following responses to the questions raised by the Office of Health Technology Assessment (OHTA) are based on an assessment of past research (retrospective studies) conducted on this technology that examines the impact of PSG on diagnosis and therapeutic services. Symptoms or complaints of excessive sleepiness constitute the most compelling indication for diagnostic PSG. The major diagnostic categories are sleep-disordered breathing (suggested by the combination of excessive sleepiness and loud snoring) and narcolepsy-cataplexy. Objective diagnostic confirmation provides the necessary basis for selection of therapy and for assessing the effects of intervention. In the case of narcolepsy, MSLT is highly specific. It is advisable for physicians to have the diagnosis objectively confirmed before prescribing stimulant medication. Diagnostic PSG is indicated as a first-line laboratory test under these clinical circumstances.

Polysomnography evaluation of sleep-related behavioral disturbances (particularly when they are violent)is frequently indicated to diagnose whether the behavior represents an epileptic or nonepileptic parasomnia. This determination has important implications for choice of therapy. If the clinical events are occurring at a frequency of four to seven times per week, there is a reasonably high probability of detection during diagnostic PSG.

There is growing agreement among sleep disorder specialists that PSG evaluation of persons older than 45 years of age with insomnia complaints is indicated as part of a comprehensive evaluation, particularly if the insomnia is refractory to behavioral and psychiatric intervention and/or there is marked disability due to symptoms. In the case of chronic insomnia, diagnostic PSG is not indicated as early in the workup as it is for excessive sleepiness. Rather, there should first be evidence of failure to respond to behavioral or psychiatric intervention. Polysomnography is indicated after the diagnosis of a mental disorder has been ruled out when there is a severe complaint of insomnia on four or more of seven nights per week, for longer than 6 months, or in a patient who is not responsive or who relapses after a 4-week trial of sedative-hypnotics or antidepressants for treatment of the sleep disturbance. A physician need not be present during the night to monitor the sleep evaluation. If all-night medical supervision is necessary, then the patient should be hospitalized as an inpatient. The patient populations most likely to benefit from assessment in a sleep disorders clinic include 1) patients with excessive sleepiness; 2) patients with persistent insomnia (lasting longer than 1 month) who are unresponsive to behavioral and psychiatric interventions; and 3) patients with sleep-related behavioral disturbances, particularly when epilepsy is a diagnostic consideration.

Response from Groups and Individuals

In response to the Federal Register notice of this assessment and solicitation of information and opinion from individuals and groups having experience with SDCs, the OHTA has received a number of letters and comments regarding SDCs.

American Sleep Disorder Association

This organization believes that the presence of a physician is no more (and no less) required than for most other medical laboratory tests. Polysomnography is indicated for accurate diagnosis of sleep disorders much as ECGs are indicated for accurate diagnosis of cardiac disorders, both of which would require tests only if the clinical diagnosis would not otherwise be accurate.

It is appropriate to refer a patient for PSG when there i no other adequate and less complex way to acquire the information necessary to understand and treat problems related to sleep and wakefulness. Specific medical conditions in which PSG is indicated include sleep apnea, narcolepsy, violent behavior and movement disorders, and persistent insomnia.

In the absence of controlled clinical trials assessing the value of PSG in the diagnosis of insomnia, ASDA recently surveyed 12 psychiatrists and psychologists recognized as experts in the diagnosis and management of insomnia. These individuals reached a consensus that patients meeting all of the following criteria would be appropriate for PSG assessment of insomnia:

  1. The insomnia complaint has been present for a minimum of 6 months (at least four nights per week).
  2. The insomnia has not responded to behavior/sleep hygiene intervention.
  3. The insomnia has not responded to a therapeutic trial with sleep-promoting medication, or sleep-promoting medication is contraindicated.
  4. A medical or psychiatric etiology has been excluded.

American Academy of Neurology

Sleep-disordered breathing. Specific findings on PSG studies in cases of sleep-disordered breathing (obstructive apnea, central apnea, mixed apnea, sleep-associated hypoventilation, and Cheyne-Stokes respiration) define the condition that cannot be objectively evaluated in any other manner. The significance of obstructive sleep apnea is emphasized by its estimated incidence of 1-10 in the general population. Polysomnography is also beneficial in the selection and assessment of therapeutic options for sleep apnea. The use of the MSLT in the evaluation of the degree of excessive daytime sleepiness in patients with sleep-disordered breathing is useful, as these patients often underestimate the degree of, or deny, hypersomnia.

Narcolepsy. It should be remembered that at presentation, 30 of patients with narcolepsy will not exhibit cataplexy --and only about half of this group will eventually develop cataplexy, but not for up to 20 years. Only 20-25 of narcoleptics experience the full range of symptoms. Definitive diagnosis is established by PSG study that reveals no other primary cause of hypersomnia, the abnormal occurrence of REM sleep, and an MSLT, which documents severe daytime sleepiness. A single nap test for the determination of narcolepsy is inadequate, as REM sleep on an isolated nap is not specific for narcolepsy. The diagnosis of idiopathic central nervous system (CNS) hypersomnia is made following an unmarkable PSG with severe sleepiness by MSLT --without the appearance of REM sleep. Objective confirmation of narcolepsy or idiopathic CNS hypersomnia is highly desirable.

Sleep-related violent or injurious behaviors. In many cases, it is not possible to accurately diagnose conditions leading to abnormal behavior or movement disorders during sleep by history alone. The differential diagnosis encompasses a wide variety of parasomnias. A complete evaluation with meticulous PSG assessment and, when appropriate, psychiatric and neurologic consultation is highly desirable.

Persistent insomnia. Insomnia is the most common sleep-related complaint. Most authors have recommended that sleep studies be used in a restricted sense for the diagnosis and management of chronic insomnia. In general, the diagnosis of acute insomnia does not require PSG. On the other hand, management of chronic insomnia may, in certain instances, require clinical PSG. (Here, the American Academy of Neurology uses the criteria for PSG assessment of chronic insomnia also advocated by the ASDA and reproduced above.)

Safety of sleep studies. There are no serious safety issues posed by this technology and its common clinical use. The presence of a physician during the PSG or MSLT is not necessary.

American Narcolepsy Association

Polysomnographic examinations are judged to be inherently safe. They are noninvasive and do not require the attendance of a physician for their implementation. The diagnostic procedures of nocturnal PSG followed by the MSLT are absolutely essential for the proper differential diagnosis of the excessively somnolent patient who does not have a significant sleep-related respiratory abnormality. The MSLT is a four-or five-nap procedure, not a three-nap procedure. The diagnosis of narcolepsy requires two abnormal transitions to REM sleep on the MSLT after evidence is obtained that there is no other significant nocturnal sleep pathology.

California Thoracic Society

Polysomnography should be performed in patients having the following conditions: COPD; restrictive ventilatory impairment secondary to chest wall and neuromuscular disturbances; disturbances of respiratory control; excessive daytime sleepiness or sleep maintenance insomnia; nocturnal cyclic brady-or tachyarrhythmias, nocturnal abnormalities of atrioventricular conduction, and ventricular ectopy during sleep that appears increased relative to wakefulness. When diagnostic tests are performed in an SDC, it is not necessary for a physician to be present during the performance of the study.

American Thoracic Society

Indications for cardiopulmonary sleep studies are listed below.


  1. Polysomnography is indicated in patients with COPD whose awake Pa(O2) > 55 mm Hg but whose illness is complicated by pulmonary hypertension, right ventricular failure, or polycythemia.
  2. Overnight ear oximetry can be very useful in prescribing an appropriate O(2) flow rate during sleep in patients in whom continuous O(2) therapy is indicated (i.e., patient whose awake Pa(O2) < 55 mm Hg).
  3. Overnight ear oximetry or PSG is not indicated in patients with COPD whose awake Pa(O2) > 55 mm Hg and who are free of complications (pulmonary hypertension, right ventricular failure, or polycythemia).

Restrictive ventilatory disorders


Polysomnography is indicated in patients with restrictive vetilatory impairment secondary to chest wall and neuromuscular disturbance whose illness is complicated by chronic hypoventilation, polycythemia, pulmonary hypertension, disturbed sleep, morning headaches, or daytime somnolence and fatigue.


Overnight sleep studies are not indicated in patients with restrictive chest wall, neuromuscular, or interstitial lung diseases who are not chronically hypoventilating and who are free of polycythemia, pulmonary hypertension, disturbed sleep, morning headaches, or daytime somnolence and fatigue.

Disorders of respiratory control


Polysomnography is indicated in patients with disturbances of respiratory control whose awake Pa(CO2) > 45 mm Hg or whose illness is complicated by pulmonary hypertension, polycythemia, disturbed sleep, morning headaches, or daytime somnolence and fatigue.

Risk factors for sleep apnea


Although snoring and obesity are risk factors for obstructive sleep apnea, the mere presence of snoring or obesity, without other symptoms, is not an indication for a cardiopulmonary sleep study.

Symptoms arising from sleep apnea


Polysomnography is indicated in patients with excessive daytime sleepiness or sleep maintenance insomnia.

Cardiovascular manifestations of sleep apnea


Polysomnography is indicated in patients with nocturnal cyclic brady-or tachyarrhythmias, nocturnal abnormalities of atrioventricular conduction, and ventricular ectopy during sleep that appears increased relative to wakefulness.


In the absence of other features of sleep apnea, overnight sleep studies are not indicated in patients with systemic hypertension or in patients with nocturnal nonspecific cardiac arrhythmias.

National Association of Medical Directors of Respiratory Care

It is safe and reasonable for Medicare to loosen its requirement for a physician to be present during sleep studies.

The following are representative responses from individuals and SDCs.

Valley Neurology Associates, Ltd (Mesa, AZ)

It is extremely important that Medicare coverage be available for this testing. It is not necessary to have a clinical PSG study performed in the presence of a physician.

Cedar-Sinai Medical Center (Los Angeles)

Although it would be nice to have a controlled study comparing outcomes of patients who have and have not been examined during sleep, with many sleep disorders, such a study would seem inappropriate. You would not know if half the patients had the disease in question.

National Capital Sleep Center (Bethesda, MD)

Physician presence is not required during PSG. Overnight PSG and MSLT are useful for making many diagnoses including sleep apnea syndrome, narcolepsy, and nocturnal myoclonus. In addition, PSG should be used in patients with persistent insomnia. The EEG portion of the PSG can also point to a possible underlying depression.

Scott and White Sleep Disorder Center (Temple, TX)

The Center does not believe physician presence is required during PSG. The medical director of this SDC believes that PSG studies should be performed for "cause rather than because." That is, specific guidelines should be used for each specific suspected condition for which patients are studied in an SDC.

University of Colorado Health Science Center Regional Sleep Disorder Center (Denver)

An all-night sleep recording (PSG) is mandatory in cases of sleep apnea syndromes. With respect to the other large group of patients with complaints of excessive sleepiness, and in whom apnea or narcolepsy is a possibility, a PSG, and very often an MSLT, performed the day after the PSG, is necessary to establish the diagnosis and to quantify the severity of the disorder. In the case of parasomnias, the use of PSG must be dictated by careful clinical examination by competent physicians. Most parasomnias do not require PSG for their diagnosis or treatment. The primary exceptions are those cases of nocturnal seizure disorders that present clinically as atypical parasomnias. The REM sleep behavior disorder generally requires a PSG with video monitoring for accurate diagnosis. With respect to a physician's presence during the PSG recording, this is completely unnecessary and would constitute a most extraordinary waste of professional time.

VA Medical Center (Martinez, CA) and University of California at Davis

Controlled sleep laboratory studies of insomniacs and normal subjects do not reveal diagnostically useful differences between the two groups and very few consistent differences of any kind. Nocturnal myoclonus is a parasomnia in which periodic movements of the legs occur during sleep. It is asserted by the ASDA that this phenomenon is a major cause of insomnia and perhaps the most important indication for sleep laboratory examination. However, the evidence that nocturnal myoclonus occurs with equal incidence in insomniacs and controls renders this assertion questionable. Recently, other kinds of disturbed behavior during sleep have been described. Some of these behavioral parasomnias could eventually prove clinically important, for example, the REM-related behavior found in some demented patients, which may be related to nocturnal delirium and wandering. However, the clinical significance and a role for sleep laboratory assessment must be established for this or any other parasomnia before public funding is justified. Neither does the analysis of the brain wave (EEG) patterns of sleep offer important assistance in evaluating insomnia. Careful and controlled studies of sleep EEGs of insomniacs and controls show only a few minor group differences. Narcolepsy is also an important clinical problem, but it is less common than sleep apnea. However, the value of sleep laboratory examinations for the diagnosis of narcolepsy resembles that for insomnia with one exception. There are some sleep abnormalities (early REM onset) that occur frequently in those narcoleptics who also have cataplexy. However, when cataplexy is present, no diagnostic tests are required because this grossly observable behavior is itself diagnostic of narcolepsy.

Department of Psychiatry, University of California, San Diego

It is essential that patients suspected of suffering from narcolepsy be evaluated in SDCs. Likewise, all-night PSG and MSLT are indicated for patients with sleep apnea syndrome. The role of the sleep laboratory examination in the evaluation of chronic insomnia remains somewhat controversial, largely because there have yet to be a sufficient number of well-conducted studies of the effectiveness of PSG evaluation in patients with uncomplicated chronic insomnia. On the other hand, preliminary data strongly suggest that the SDC has much to offer these patients. At the present time, most authorities would accept the recommendation that patients with chronic insomnia be referred to an SDC if there is strong evidence for sleep apnea, narcolepsy, sleep-related movement disorders, or intractable insomnia. This institution strongly recommends that the presence of a physician not be required whenever a diagnostic test is performed in an SDC.

Sleep Research and Treatment Center, Pennsylvania State University College of Medicine (Hershey)

The sleep laboratory is useful for the diagnosis of sleep apnea and certain cases of narcolepsy. Moreover, the diagnostic information obtained in the sleep laboratory for sleep apnea has clear-cut therapeutic utility in that it determines to a large extent the therapeutic management of the disorder. In contrast, sleep laboratory studies of the parasomnias have only limited diagnostic indications, and for insomnia they are not diagnostically indicated or helpful. Accordingly, information obtained in cases of insomnia has no therapeutic value. A physician thoroughly trained in the diagnosis and treatment of sleep disorders should be in direct supervision of the sleep laboratory at all times. However, this does not mean that he or she needs to be physically present in the laboratory. Rather, the physician should, at a minimum, be available by phone whenever tests are performed.


The medical discipline studying sleep disorders gradually arose more than 40 years ago through the efforts of a few pioneers in the field. The discovery of REM sleep in 1953(50,51). was an important clue to the complexity of sleep, and this observation helped to initiate the modern era of sleep research. Several conferences on sleep disorders took place in the 1960s and 1970s and classifications of sleep disorders were developed as described earlier. The enormous body of human and animal research that has emerged during the past 40 years has revealed a great deal of new information; nevertheless, at the present time much has yet to be learned about the basic mechanisms of sleep.

The major organization in the United States concerned with sleep disorders is the ASDA. Among their activities are continuing medical education for physicians and an accrediting program for SDCs; as of July 1990, 150 SDCs in 45 States have been accredited by the ASDA. It is estimated that there are between 900 and 1,000 nonaccredited SDCs (personal communications, ASDA, October 1990).

The majority of clinical reports summarized in this assessment indicate that overnight PSG examinations and MSLTs in SDCs provide useful diagnostic information and that characteristic PSG findings could establish or rule out diagnoses in a proportion of patients studied. In addition, careful physiologic monitoring of patients with certain forms of sleep disorders appears to be a logical component of the development of the scientific basis for understanding the pathophysiology of sleep. In some reports, the authors concluded that PSG may be of prognostic value and might also be used to evaluate the progress of patients being treated for sleep disorders.(52) However, as noted earlier, there are few prospective studies and no prospective controlled trials of the clinical utility of PSG, nor are there prospective studies that compare the effectiveness of therapies based on diagnoses obtained by PSG compared with therapies used in patients not having a PSG-based diagnosis. Furthermore, other alternative low-technology approaches to these problems have not been carefully evaluated. This lack of controlled trials demonstrating that studies in SDCs lead to more accurate diagnoses and better treatment was addressed by the National Institute of Neurological Disorders and Stroke (NINDS), ADAMHA, and the National Institute of Aging (NIA); these agencies informed OHTA that they are planning to encourage the development of such trials by the community of sleep disorder specialists. One such trial is the study currently funded by ADAMHA, discussed above.

Most of the responses from individuals and groups to the Federal Register notice and mailings were highly supportive of the use of SDCs. However, some correspondents indicated the belief that SDCs are not necessary for the diagnosis and treatment of some sleep disorders. The majority of expert opinion provided to OHTA, including that from the PHS, supported the use of PSG in the evaluation of sleep-related breathing disorders, narcolepsy, epilepsy, parasomnias, and insomnia. It should be noted that in deriving the conclusions of an assessment, the quality of evidence is of paramount importance. That is, greatest weight is given to data from clinical trials that are designed to minimize the potential for error and in which the results are unlikely to be due to measurement artifacts, deficiencies in study design, bias, or misinterpretation. Expert opinion unverified by objective data is not a sufficient basis for an assessment conclusion that the technology in question is or is not safe, efficacious, clinically effective, or appropriate.

On the basis of the limited evidence available, it appears most reasonable to conclude that PSG can aid in the diagnosis of sleep-related breathing disorders, such as sleep apnea, although it is also clear that in some cases a diagnosis may be arrived at with a high degree of accuracy by information based on the history, family or bed partner observations, and physical and routine laboratory examinations. Furthermore, not every component of a full PSG examination may be necessary to provide useful diagnostic information regarding sleep apnea. For example, a recent study(53). using PSG of 200 patients with suspected sleep apnea/hypopnea syndrome (SAHS) was performed. As part of this study, oxygen saturations were recorded separately and were interpreted by one of the authors without knowledge of the patient's identity or results of the other findings of PSG. The results of oximetry alone correctly identified 66 of SAHS patients. The authors concluded that many of the neurophysiologic measurements of sleep performed during a full PSG may not be necessary to make a diagnosis of SAHS. The development of newer techniques involving home monitoring appears promising in this regard. Clinical studies are currently in progress, and data should be available in the near future. Further, PSG may be warranted in the diagnostic workup of certain otherwise unexplained clinical syndromes that are known to be a potential consequence of sleep-related breathing disorders, such as cardiac arrhythmias, polycythemia, idiopathic heart failure, etc. In addition, the available evidence indicates that PSG may be useful in the diagnosis of certain cases of suspected narcolepsy wherein the history and physical and neurologic examinations are insufficient to make a diagnosis with reasonable certainty. However, PSG and/or MSLT in all suspected cases may not be warranted. Likewise, there are circumstances wherein a reasonable clinical suspicion of epilepsy presenting as a sleep disturbance or altered sleep behavior may dictate the prudent use of noctural EEG. The PHS has addressed this issue in some detail in the assessment on EEG video monitoring.(54)

In patients with suspected parasomnias, the nature of these conditions may be established by careful clinical evaluation in the majority of cases. In those instances of repeated violent or injurious sleep-related behavior, and where epileptic seizures have been ruled out, PSG may be beneficial; examples of such benefit include the ability to rule out psychiatric etiologies and/or determine the probability of spontaneous remission of the disorder.

However, the evidence showing PSG to be clinically useful in the evaluation of all types of insomnia is not compelling. Much of the published data indicate that a significant proportion of insomniac patients studied with PSG could have had presumptive diagnoses based on history, physical examination, or simpler diagnostic workups (vide supra). Moreover, there is virtually no published information available that addresses outcome measures, i.e., the relationship between PSG-based diagnoses and the effectiveness of therapy for insomnia so chosen. The paucity of current data regarding insomnia has resulted in the ADAMHA-funded cooperative, prospective trial to evaluate the reliability and clinical utility of diagnostic schema for insomnia and the contribution of PSG to those schema. The American Academy of Neurology and the ASDA provided identical suggestions with regard to the use of PSG in chronic insomnia. While these recommendations may be logical, no objective data were provided in support of, and no published study has employed, those criteria for PSG in evaluating insomnia. The ASDA stated that their recommendations resulted from their survey of the opinions of 12 experts. Furthermore, after evaluating all the evidence presented at the 1990 Consensus Development Conference "The Treatment of Sleep Disorders in Older People,"(4). the panel concluded that there were insufficient data to assess the value of PSG in the routine evaluation of insomnia. "Chronic insomnia" fulfilling the ASDA's and American Academy of Neurology's criteria was not addressed.

Patients with sleep disorders that can be attributed to other obvious medical conditions, such as alcoholism, Parkinsonism, dementia and Alzheimer's disease, depression, asthma, and various types of cardiovascular disease, do not usually need PSG studies.

There are nearly 150 SDCs that have been approved by the ASDA, and 1,000 without such approval (personal communication, ASDA, October 1990). Thousands of patients have been studied and evaluated in these centers, yet for many of the conditions described in the ASDA classification of sleep disorders, there are few objective, reliable data that would allow for a clear elaboration of the clinical utility of PSG. This lack of reliable data was also addressed by the NIH Consensus Development Conference.(4) In addition to the need for prospective controlled studies, which has been recognized by several agencies of the PHS (vide supra), it appears that the establishment of a centralized formal system of patient registry and patient followup would go far toward providing reliable evidence of the proper clinical role, including the effects of treatment, for PSG, particularly in parasomnias and insomnias. The registration of consecutive cases, use of consistent patient selection criteria, uniform diagnostic and classification schemes, and results of patient outcome data would provide information on which to assess objectively the clinical effectiveness of PSG. Such registries are already being used successfully in fields such as transplantation, interventional cardiology, oncology, positron emission tomography scanning, and other technologies.

Polysomnography is an entirely noninvasive procedure and would have the same degree of risk as, for example, an ECG examination. If anything, the patient sleeping in an SDC is at less of a risk than when sleeping at home because he or she is being observed and monitored continuously. The question of whether a physician needs to be present during PSG was not directly addressed in the papers cited earlier in this assessment, but the responses obtained from a notice in the Federal Register and solicitation of information from professional organizations and individuals unanimously supported the opinion that a physician does not have to be directly present during a PSG.


Sleep disorders are common in our society. It is estimated that there are 50 million people in the United States who suffer to varying degrees from sleep problems. A great deal has been learned about sleep during the past 40 years. Much of this knowledge has been obtained by the use of PSG, which consists of the simultaneous recording of several physiologic parameters from a patient just prior to and during sleep. Much of the technology utilized in PSG are based on individual tests developed many years ago. Current published data permit the conclusion that PSG is useful for the diagnostic evaluation of patients with sleep-related breathing disorders, may be helpful in the evaluation of suspected cases of narcolepsy wherein other findings are in inconclusive or contradictory, and may be helpful in cases of parasomnias and/or suspected epilepsy wherein the distinction between seizure activity and other forms of sleep disturbance is uncertain. Current data do not permit a firm conclusion as to the clinical effectiveness of PSG in other symptoms of sleep disturbance such as insomnia. Current, ongoing clinical trials are expected to provide information addressing this point, and several agencies (NINDS, ADAMHA, and NIA) have expressed their intent to encourage the organization of prospective trials to determine the ultimate clinical utility of SDC and PSG techniques. A physician need not be present during PSG in an SDC.


Richardson JW Symposium on sleep disorder: Mayo sleep disorders update Mayo Clin Proc 1990. 65:857–860. [PubMed: 2195246]
[no authors listed] Public Law 92-463, as amended. (5 US Code Appendix 2) 1988.
[no authors listed] Drugs and Insomnia, National Institutes of Health Consensus Development Conference, Bethesda, MD: Public Health Service 1983; 1984. 4(10).
[no authors listed] The Treatment of Sleep Disorders in of Older People: Consensus Statement National Institutes of Health Consensus Development Conference, Bethesda, MD: Public Health Service, 1990; . 8(3)
Carkasdon MA, Dement EC Normal sleep: An overview in Kryger MH, Roth T, Dement WC (Eds): Principles and Practices of Sleep Medicine., Philadelphia: WB Saunders Co, 1989p 3.
[no authors listed] Project Sleep, appendix A in Kupfer DJ, Crook VT (Eds): Physician's Guide to the Recognition of Sleep Disorders in the Elderly, New Canaan: Mark Powley Associates, Inc. 1989p 3.
Zepelin H Mammalian sleep in Kryger MH, Roth T, Dement WC (Eds): Principles and Practices of Sleep Medicine., Philadelphia: WB Saunders Co, 1989p 30.
Mistlberger R, Rusak B Mechanism and models of the circadian time keeping system in Kryger MH, Roth T, Dement WC (Eds): Principles and Practices of Sleep Medicine., Philadelphia: WB Saunders Co, 1989p 141.
[no authors listed] Diagnostic classification of sleep and arousal disorders Sleep 1979. 2:1–137. [PubMed: 531417]
The International Classification of Sleep Disorders American Sleep Disorders Association, Lawrence, KS: Allen Press, 1990.
Kales A [Vela-Bueno A, Kales JD.] Sleep disorders: Sleep apnea and neurology narcolepsy Ann Intern Med 1987. 106: 434–443. [PubMed: 3544995]
Kaplan J, Staats BA Obstructive sleep apnea syndrome Mayo Clin Proc 1990. 65: 1087–1094. [PubMed: 2201837]
Guillminault C Sleep and breathing in Guilleminault C (Ed): Sleeping and Walking Disorders: Indications and Techniques., Menlo Park, CA: Addison-Wesley Publishing Co, 1982.
Ancoli-Israel , Kripke DF, Mason W. Characteristics of obstructive and central sleep apnea in the elderly: An interim report Biol Psychiatry 1987; . 22:741–750. [PubMed: 3593813]
Richardson JW, Fredrickson PA, Lin S Narcolepsy update Mayo Clin Proc 1990. 65:991–998. [PubMed: 2198399]
Aldrich MS Narcolepsy N Engl J Med 1990. 323:389–394. [PubMed: 2196464]
Guilleminault C Narcolepsy in Kryger MH, Roth T, Dement WC (Eds): Principles and Practices of Sleep Medicine., Philadelphia: WB Saunders Co, 1985p 338.
FredericksonFredrickson PA, Richardson JW, Esther MS, et al Sleep disorders in psychiatric practice Mayo Clin Proc 1990. 65:861–868. [PubMed: 2195247]
[no authors listed] The International Classification of Sleep Disorders American Sleep Disorders Association, 1990p 145.
[no authors listed] The International Classification of Sleep Disorders American Sleep Disorders Association, 1990p 148.
Schenck CH, BundieBundlie JR, Patterson AL, et al Rapid eye movement sleep behavior disorder JAMA 1987. 257: 1786–1789. [PubMed: 3820495]
Raschka LB Sleep and violence Can J Psychiatry 1984. 29:132–134. [PubMed: 6722704]
Zorick F. Overview of insomnia in Kryger MH, Roth T, Dement WC (Eds): Principles and Practices of Sleep Medicine., Philadelphia: WB Saunders Co, 1985p 431.
Gillin CJ, Byerley WF The diagnosis and management of insomnia N Engl J Med 1990. 322:239–248. [PubMed: 2242104]
Zorick FJ, Roth TA, Hartze KM, et al Evaluation and diagnosis of persistent insomnia Am J Psychiatry 1981. 138:769–773. [PubMed: 7246806]
Hauri PJ, Esther M Insomnia Mayo Clin Proc 1990. 65:869–882. [PubMed: 2195248]
Prinz PN, Vitgiello Vitiello MV, Raskind MA, et al Geriatrics: Sleep disorders and aging N Engl J Med 1990. 329:520–525. [PubMed: 2198467]
Montplaisir J, Godbout R Restless legs syndrome and periodic movements during sleep in Kryger MH, Roth T, Dement WC (Eds): Principles and Practices of Sleep Medicine., Philadelphia: WB Saunders Co, 1989p 402.
[no authors listed] The International Classification of Sleep Disorders, American Sleep Disorders Association. 1990. p. p 247.
Peled R, Lavie P Paroxysmal awakening awakenings from sleep associated with excessive daytime somnolence: A form of nocturnal epilepsy Neurology 1986. 36:95–98. [PubMed: 3941791]
Shouse MN Epilepsy and seizures during sleep in Kryger MH, Roth T, Dement WC (Eds): Principles and Practices of Sleep Medicine., Philadelphia: WB Saunders Co, 1989p 364.
Thorpy MJ Diagnosis, evaluation, and classification of sleep disorders in Williams RL, Kartaca I, Moore CA (Eds): Sleep Disorders, Diagnosis, and Treatment, New York: John Wiley and Sons, 1988, p 3.
Guilleminault, C (Ed) Sleeping and Waking Disorders: Indications and Techniques, Menlo Park, CA: Addison-Wesley Publishing Company, 1982.
Carskadon MA. Guidelines for the multiple sleep latency test (MSLT): A standard measure of sleepiness Sleep 1986. 9:519–524. [PubMed: 3809866]
Kavey NB, Whyte S, Resor, SR, et al Somnambulism in adults Neurology 1990. 40:749–752. [PubMed: 2330099]
[no authors listed] The International Classification of Sleep DisordersAmerican Sleep Disorders Association, 1990p 147.
Schenck CH, Millner Milner DM, Hurwitz TD. A polysomnographic and clinical report on sleep related injury in 100 adult patients Am J Psychiatry 1989; . 146:1166–1173. [PubMed: 2764174]
Coleman RM, Roffwarg HP, Kennedy SJ, et al Sleep wake disorders based on polysomnographic diagnosis JAMA 1982. 247: 997–1003. [PubMed: 7057593]
Kryger MH, Mezon BJ, Acres JC Diagnosis of sleep breathing disorders in a general hospital Arch Intern Med 1982. 142:956–958. [PubMed: 7082118]
Hauri PJ, Hayes B, Sateria M. Effectiveness of a sleep disorder disorders center: A 9-month followup follow-up Am J Psychiatry 1982. 139:663–666. [PubMed: 7072857]
Jacobs EA, Reynolds CF, Kupfer DS The role of polysomonography polysomnography in the differential diagnosis of chronic insomnia Am J Psychiatry 1988. 145:346–349. [PubMed: 3344849]
Kales A, Kales JD Evaluation and diagnosis in of sleep disorder disorders patients Semin Neurol 1987. 7: 243–248. [PubMed: 3332459]
Tan TL, Kales JD, Kales A, et al Inpatient multimensional multidimensional management of treatment resistant insomnia Psychosomatics 1987. 28:266–272. [PubMed: 3423177]
Edinger JD, Hoelscher TJ, Webb MD, et al Polysomnographic assessment of DIMS: Empirical evaluation of its diagnostic value Sleep 1989. 12:315–322. [PubMed: 2762687]
He J, Kryger MH, Zorick FJ. Mortality and apnea index in obstructive sleep apnea Chest 1988. 94:9–14. [PubMed: 3289839]
Amira SA, Johnson TS, Logowitz NB Diagnosis of narcolepsy using the multiple sleep latency test: Analysis of current laboratory criteria Sleep 1985. 8:325–331. [PubMed: 3880172]
Saskin P, Moldofsky H, Lue FA Periodic movement movements in sleep and sleep-wake complaints Sleep 1985. 8:319–324. [PubMed: 3880171]
Rosenthal L, Roehrs T, Sicklesteel J, et al Periodic movement during sleep, sleep fragmentation and sleep[-]wake complaints Sleep 1984. 7:326–330. [PubMed: 6515247]
[no authors listed] Sleeping giant (editorial) Lancet 1991; . 337:1067.
Aserinsky E, Kleitman N Regularly occurring periods of eye mobility, and concomitant phenomena, during sleep Science 1953. 118:273–274. [PubMed: 13089671]
Dement W Some Must Watch While Some Must Sleep, Stanford, CA: Stanford Alumni Association, 1972p 24.
Orr WC Utilization of polysomnography in the assessment of sleep disorders Med Clin North Am 1985. 69:1153–1167. [PubMed: 3934480]
Douglas NJ, Thomas S, Jan MA Clinical value of polysomnography LancettLancet 1992. 339;347–350. [PubMed: 1346422]
Erlichman M Electroencephalographic (EEG) Video Monitoring. Health Technology Assessment Report 1990. No. 4., Rockville, MD: Agency for Health Care Policy and Research.

AHCPR Pub No. 92-0027


  • PubReader
  • Print View
  • Cite this Page

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...