When to Order a Sleep Study and How to Read the Report: Part II

Authors: Donna Arand, PhD, Kettering Medical Center, Wallace Kettering Neuroscience Institute; George Burton, MD, Kettering Medical Center, Wallace Kettering Neuroscience Institute, Wright State University; and Michael Bonnet, PhD, Kettering Medical Center, Wallace Kettering Neuroscience Institute, Dayton Department of Veterans Affairs Medical Center, and Wright State University.

Editor’s Note—This is the final in a 2-part series on when to order a sleep study and how to read the report. The sleep study report contains essential information for diagnosing common sleep-related physiological disorders. The report format is not standardized, so a variety of formats are used by different facilities. However, generally the same information should be found in any sleep study report. Typically the information provided includes data on sleep parameters, respiration, leg movements, ECG, and descriptions of any abnormalities noted in the various parameters that are recorded. Understanding the information provided in the sleep study report and knowing how to use it is crucial for accurate diagnosis and effective treatment. The combined information from the sleep history and the sleep report allows diagnosis of all sleep disorders.

Sleep Architecture

Sleep architecture refers to the temporal pattern of sleep and wake across the sleep period. The generation of sleep follows a consistent progression of sleep stages and repetitive cycles that are represented by the sleep architecture. Sleep stages are based on the EEG and consist of 5 stages (Stages 1, 2, 3, 4, and rapid eye movement [REM] sleep). Stages 1-4 are collectively called non-REM sleep (NREM). Adults usually cycle through the NREM stages in consecutive order and then into REM about every 90 minutes throughout sleep. REM periods get longer across the night, while stage 3 and 4 get shorter and occur primarily in the first half of the night. Stages of sleep are usually sustained in normal sleep architecture without frequent brief awakenings that characterize disturbed sleep.

The sleep report may contain a graphic representation of the stages of sleep across the night called a hypnogram. This is a visual representation of the sleep architecture. A neat stepwise graph reflects normal uninterrupted sleep while a ragged, comb-like appearance depicts disturbed sleep with frequent stage changes and awakenings (see Figure 1). This fragmented sleep is unlikely to be restorative. The general appearance of the hypnogram gives a quick impression of the amount of sleep as well as the degree of sleep disruption present. The numerical data represented by the hypnogram is usually provided in the sleep report (see Table 1).

Total Sleep Time

The total sleep time (TST) in minutes or hours must be reported (see Table 1). Typically patients are recorded for 8 hours and they will sleep between 6-8 hours. TST divided by the time in bed (TIB) is called the sleep efficiency, and it is expressed as a percent. A normal adult sleeper would be expected to have a sleep efficiency > 90% at home, but in the laboratory > 80% may be considered good since sleep time in the laboratory is usually shorter and more disrupted than sleep at home. Sleep efficiencies greater than 95% are suspicious in adults. This may represent the effects of sleep deprivation due to insufficient sleep, idiopathic hypersomnia or possibly narcolepsy. An MSLT is useful in the differential diagnosis of these problems.

Reports should document an adequate amount of sleep time on the back and side, to evaluate a positional component in the diagnosis of sleep apnea. Often sleep apnea is worse in the supine position as the tongue lapses back to occlude the airway. In some patients, apnea may occur only in the supine position so an effective treatment option is simply a behavioral technique to avoid the supine sleep position (eg, use of a tennis ball in a back pocket at the mid-scapular level on a sleep shirt). If a patient always avoids a certain sleep position, then recording from that position is unnecessary.

Recent revisions by the Centers for Medicare and Medicaid Services (CMS) require a minimum of 2 hours of sleep time as a basis for computing the apnea and hypopnea index for sleep apnea.1 (See Table 2.) Shorter sleep times can inflate the index—especially with sleep times of less than 1 hour. This CMS requirement primarily affects split night studies. To meet CMS requirements, the technician performing a split night study must accurately estimate whether 2 hours of sleep, not recording time, have occurred and judge whether there will still be adequate sleep time to titrate the patient.

Sleep Latency

Sleep latency is the time from lights out to the first 30-second epoch of scored sleep (sleep onset). Sleep latency must be interpreted in the context of the patient’s sleep history and medication use. Sleep latencies longer than 30 minutes warrant further investigation. A common cause of a long latency is situational insomnia due to difficulty sleeping in an unfamiliar laboratory setting. This is a likely explanation if the patient’s sleep history or if sleep logs do not indicate a problem with sleep onset insomnia. If the patient’s history does indicate chronic difficulty falling asleep, then physiological causes such as pain or restless legs may be identified from the sleep history. In the absence of an identifiable physiological cause for a long sleep latency as reported in the PSG and indicated in the sleep history, a diagnosis of psychophysiological insomnia is supported.

Latencies to other stages of sleep may be reported, and these are measured from sleep onset to the first epoch of that stage of sleep. Only the latency to stage REM is diagnostically useful. Normal REM latency is 90 minutes or longer in adults. Shorter REM latencies of 30-50 minutes are classic findings in untreated depressed patients.2 Very short REM latencies (< 15 min) are common in narcolepsy patients and support the diagnosis.3 However, a short REM latency may also be induced in otherwise normal individuals who are partially sleep deprived, depressed, or withdrawing from stimulant medication or alcohol. In these cases, the shortened REM latency is also accompanied by an increase in the percent of REM sleep, a phenomenon called REM-rebound. Whenever the REM percent is very high, REM-rebound should be suspected. A REM-rebound phenomenon must be excluded as an explanation for a shortened REM latency before attributing diagnostic meaning to it.

Sleep Stages

Sleep stages are determined for each 30 seconds of EEG activity based on standard scoring criteria.4 The minutes and percent of stages 1-4 and REM are usually reported (see Table 3). The minutes of each stage are divided by the TST to give the percentage of each stage. The percentage of stages is more useful for comparison than sleep time, since it corrects for variations in TST. Normal sleep stage percentages vary with age. The percent of each sleep stage is useful in evaluating the degree of sleep disturbance and can help in the differential diagnosis of some sleep disorders.

Stage 1 is a transitional stage between wake and sleep. In normal sleepers it would comprise a few minutes of TST. The higher the percentage of stage 1, the more disrupted the patient’s sleep. The cause of increased stage 1 is often due to repetitive sleep apnea episodes or leg movements that cause brief arousals or awakenings that are followed by a retransition to sleep.

Stage 2 sleep normally follows stage 1 sleep. Stage 2 occupies 50% or more of TST in adults. It is a "filler" stage that will decrease when other stages rebound or increase in their absence. The percent of stage 2 is not useful for diagnostic purposes or as an indication of sleep disruption.

Stages 3 and 4 are characterized by increasing delta waves in the EEG and are collectively called delta sleep or slow wave sleep (SWS). Delta sleep decreases dramatically with age and is often absent or minimal in normal sleepers older than age 40. In patients still capable of generating delta sleep, the presence of frequent arousals or awakenings, as occur in sleep apnea, may not allow enough time between awakenings for delta sleep to occur since it follows stage 2 sleep. With initiation of effective PAP treatment, arousals are eliminated and there may be a rebound of delta sleep during the titration night. In patients without frequent sleep disruptions, the amount of delta sleep is useful for diagnostic purposes since some parasomnias such as sleepwalking and night terrors only occur during stage 3 or 4. If there is little or no stage 3 or 4 recorded, parasomnias can be ruled out as the cause of abnormal behaviors during sleep.

The presence of REM sleep is very useful for diagnostic purposes. REM sleep is physiologically different from NREM. REM sleep is characterized by muscular atonia (eg, paralysis of volitional muscles), irregular breathing, and hypoventilation that may result in lower oxygen values compared to wake. Due to these changes, the type of disordered breathing events may be different in REM (eg, central rather than obstructive apnea), the events may be longer, and the oxygen desaturation may be greater. In some patients, disordered breathing events may only appear in REM sleep. Baseline or diagnostic studies without REM sleep may miss the presence of sleep apnea or underestimate its severity. In addition, disorders such as petit mal epilepsy are activated by REM sleep, and REM behavior disorder only occurs during REM sleep. If however, a significant disorder can be documented without REM sleep, the recording of REM is not crucial. The percentage of REM sleep is typically around 20-25% in normal adults. However, dramatic increases (called REM-rebound), can be seen following withdrawal of REM-suppressant medications, recovery from sleep deprivation and during the initiation of PAP. REM sleep is often decreased when sleep is frequently disrupted.


Data concerning sleep-related breathing disorders are usually of most concern. The 2 types of breathing events scored are apneas and hypopneas. Apnea refers to a complete cessation of airflow while hypopnea refers to a decrease in airflow. These are further defined as obstructive, central, or mixed events. In obstructive apnea events, there is continued respiratory effort without airflow. In central apnea events, there is an absence of respiratory effort and airflow. A mixed apnea event is a combination of central and obstructive apnea. Hypopneas may be similarly classified depending on the relationship between the decreased airflow and the amount of respiratory effort. There can be variations in the definitions of these terms between sleep programs, so definitions of the terms must be provided to accurately compare reports from different facilities.

Recently, CMS has adopted standardized definitions that are used to determine qualification for CPAP coverage for Medicare patients.1 (See Table 2.) According to CMS recommendations, apnea is defined as a cessation of airflow for at least 10 seconds. Hypopnea is defined as an abnormal respiratory event lasting at least 10 seconds with at least a 30% reduction in thorocoabdominal movement or airflow as compared to baseline, and with at least a 4% oxygen desaturation. However, airflow and chest movement are usually measured qualitatively in the PSG; therefore, the percent of decrease in flow or effort depends on the judgment of the scorer and interpreting physician (scoring bias).

The total number and type of apneas and hypopneas scored should be included in the report. When the number of apneas and hypopneas is divided by total sleep time, the result is the Apnea Hypopnea Index (AHI). An index can be computed in the same way for each type of breathing event (see Table 3). An Apnea Index (AI) includes only apneas per hour of sleep while the Hypopnea Index (HI) is based on total hypopneas divided by total sleep time. The Respiratory Disturbance Index (RDI) is a term used in some reports, and it is generally considered the same as the AHI. However, the more general term "respiratory disturbance" may include other breathing irregularities in addition to apneas and hypopneas depending on the definition used by the laboratory. The indices are used as a measure of the severity of sleep apnea, with higher indices reflecting increasing severity. Terms such as mild, moderate, or severe are not as useful as the actual index. In fact, no standards for mild, moderate, or severe exit. Since the effect of sleep-disordered breathing depends on factors other than the AHI (such as degree of sleep disturbance, amount of oxygen desaturation, and duration of sleep-disordered breathing events), it is simplistic to infer the severity of sleep-disordered breathing based on the AHI alone.

The use of an index controls for variations in number of events due to differences in TST and provides a more reliable measure for comparison. However, the index is extremely inflated when computed for sleep times less than 1 hour. CMS requires a minimum of 2 hours of sleep before accepting an AHI for determining treatment coverage. Generally an AHI < 5 is considered within normal limits.

Apnea and hypopnea indices can be computed for sleep position (side or back) as well as sleep state (REM vs NREM) by counting events in each condition and dividing by the total sleep time for that condition. An AHI by position is useful for treatment purposes. If the index is > 5 on the back but < 5 on the side, techniques to help avoid back sleep can be extremely effective. By contrast, a REM-related AHI > 5 with a NREM index < 5 is not useful for treatment purposes as treatment cannot be confined to the REM state. A high REM index underscores the effect of differential control of breathing during REM and the associated atonia that results in loss of rib cage musculature to aid respiration. REM sleep occurring in the supine position usually results in the most severe apnea episodes, with the longest duration and the greatest oxygen desaturation.

Oxygen Saturation

Apnea and hypopneas can be associated with severe drops in oxygen saturation. The minimum oxygen saturation occurring during the night should be noted in the report. (See Table 3.) This value may be reported for the REM and NREM states independently. However, for treatment purposes, only the lowest value is needed. If oxygen during sleep falls below 88% at any time and is not associated with apneas or hypopneas, nocturnal oxygen administration may be warranted. Significant oxygen desaturation associated with apneas or hypopneas should be eliminated with effective PAP treatment in individuals without respiratory compromise. Treatment of any underlying pulmonary disease must be achieved before CPAP or oxygen therapy is initiated.

The CMS definition of hypopnea requires an associated drop of 4% or more in oxygen saturation. However, not all hypopneas are associated with oxygen desaturation. Some hypopneas produce arousals without oxygen desaturation. PAP treatment of hypopneas associated with arousals using CPAP is effective in eliminating disturbed sleep and daytime sleepiness. A definition of hypopnea that requires a degree of oxygen desaturation results in a lower reported AHI.


Arousals refer to brief awakenings (> 3 seconds) during sleep. They typically occur at the termination of apnea events as the patient awakens to breathe. Arousals may also be triggered by events such as PLMS or they may occur spontaneously. Sleepiness increases as the number of arousals increase.5-7 The sleepiness occurs even without changes in TST or stages of sleep.8 The sleep fragmentation is considered the cause of excessive sleepiness regardless of the trigger for the arousal. The number of arousals per hour of sleep is computed and reported as the arousal index. (See Table 1.) Indices greater than 20 are generally considered elevated.

Cardiac Parameters

Cardiac information including rate, rhythm, and general morphology is collected during the sleep study. The cardiac rhythm during sleep is similar to wake although the heart rate is slower. Relative bradycardia and tachycardia (that may not be abnormal) typically occurs in association with apneas, with heart rate decreasing during the apnea and then increasing during the arousal that follows. Due to the slower heart rate during sleep and apnea events, escape beats are more likely to occur. Some patients may have cardiac arrhythmias or conduction disturbances only during sleep. These will be described in the sleep study report and may require further cardiac evaluation.

EEG Activity

Sleep can activate epileptiform discharges or seizures. However, epileptiform activity may be difficult to detect in the sleep study EEG recordings since typically only 2 EEG placements are used. If epilepsy is suspected, a full EEG montage should be performed and specified in the PSG order. Any EEG abnormalities noted in the study will be included in the sleep study report. Significant findings may require further neurologic evaluation, including a seizure-montage EEG, interpreted by a neurologist.


Diagnostic impressions or provisional diagnoses are included at the end of the report. These are used in the context of all other clinical information, to make final diagnoses. Often patients will have more than 1 sleep diagnosis when sleep study findings and sleep history information are both carefully evaluated.

The list of diagnoses on the report may also include the corresponding ICSD and International Classification of Disease (ICD) codes. The ICSD lists more sleep disorders than the ICD, so some sleep disorders diagnoses will not have a corresponding ICD code. Codes for common sleep disorders, such as sleep apnea, are similar.

Treatment Recommendations

Treatment recommendations should be included with the report. In sleep centers accredited by the American Academy of Sleep Medicine (AASM), all sleep study reports must include treatment recommendations. These may be found at the end of the report following diagnoses or in a separate summary letter to the referring physician.

Specialized Sleep Tests

Titration Studies
Titration reports for PAP or bilevel PAP include the same information as the baseline studies in addition to the AHI for the various pressures used from 4 to 20 cm/H2O. (See Table 4.) Generally, the AHI decreases with increasing pressure. However, if pressure is raised too high for a patient, an increase in the AHI may occur. Generally, the pressure producing the lowest AHI is recommended for treatment, but this is not always the case. The pressure with the lowest AHI may not include supine sleep or REM sleep, the conditions when breathing problems are usually worse. If the baseline study indicated that the breathing problem was REM-related or only occurred in the supine position, then titration indices that include those periods are needed to determine the most effective pressure. PAP may not always eliminate all apneas and hypopneas, but the recommended pressure should result in an AHI less than 5. In some cases, the effective pressure may eliminate apneas and hypopneas but oxygen saturation may remain low. In these cases, supplemental nocturnal oxygen can be bled into the PAP mask and this may be also titrated during the sleep study. Information concerning supplemental oxygen will be included in the sleep study report.

Multiple Sleep Latency Test
Performance of MSLT is generally indicated when narcolepsy is suspected. It is not part of the routine evaluation of sleep apnea. The MSLT is used as an objective measure of daytime sleepiness and to aid in the diagnosis of narcolepsy. The MSLT is performed following a polysomnogram to assure that an adequate amount of normal sleep was obtained in the sleep period prior to the MSLT. In the MSLT, the mean sleep latency is determined from a series of naps and is used as a quantitative measure of sleepiness (see Table 5.)

The MSLT consists of 4-5 naps given 2 hours apart, beginning 1.5-3 hours after awakening. The MSLT requires the patient to lie down in a quiet, dark room and try to fall asleep. The patient is given 20 minutes to fall asleep or the nap is terminated. If the patient does fall asleep, the nap is terminated 15 minutes after sleep onset. The mean sleep latency is computed across all naps using 20 minutes if no sleep occurs in a trial. A mean sleep latency < 5 minutes is considered abnormal while a latency > 10 minutes is considered normal. Latencies between 5 and 10 minutes represent borderline pathological sleepiness. The cause of the sleepiness needs to be determined for latencies < 10 minutes. If no physiological cause can be found for latencies < 5 minutes then a diagnosis of idiopathic hypersomnia is made and stimulant medication is recommended.

The presence of REM sleep or latency to REM in each nap is also reported. If REM sleep occurs in 2 or more naps (REM onsets), a diagnosis of narcolepsy is supported. A final diagnosis of narcolepsy is made if the sleep history is also consistent with narcolepsy and other causes of REM onsets are ruled out. REM onset sleep is characteristic of narcolepsy, but the presence of REM in the MSLT is not always indicative of narcolepsy. REM onset sleep can also be caused by withdrawal of medications, sleep deprivation, depression, and disturbed nocturnal sleep, such as that caused by sleep apnea.2,9 A preceding night study is used to document adequate sleep time without sleep-related physiological abnormalities to eliminate these as possible causes for the REM onsets. Medication use is documented and drug screens are often performed on the day of the MSLT to help assure an interpretable MSLT study. If other conditions are found that may produce REM onsets, the diagnosis of narcolepsy is uncertain. Consequently, strict guidelines are imposed on the patients and the procedure for the MSLT to assure validity for diagnostic purposes.

In some cases, the MSLT is used for objective documentation of the degree of daytime sleepiness when an underlying sleep disorder is known to exist. This commonly occurs when patients are involved in hazardous work activities that put them or others at risk due to job malperformance from sleepiness. Objective testing is needed since subjective estimates of sleepiness may be overestimated or underestimated and because the subjective estimates are not always related to objective measurements.10

Common Sleep Disorders

Narcolepsy is a usually inherited sleep disorder with symptoms typically beginning between 15 and 30 years of age.11 Age of onset is useful in differentiating narcolepsy from sleep apnea. Narcolepsy is estimated to affect .02-.15% of the US population.12,13 It is characterized by sudden attacks of irresistible sleepiness, cataplexy, sleep paralysis, and hypnogogic hallucinations. These symptoms may occur in varying combinations and severity. Some symptoms such as sleep paralysis and hypnogogic hallucinations may occur in individuals without narcolepsy, but cataplexy is believed to be specific. Very short naps of 10-20 minutes are usually extremely refreshing in narcoleptic patients and often consist of REM sleep. REM onset sleep periods occur because of impaired sleep-wake regulation rather than an excessive need for REM sleep.14 This particular characteristic of narcolepsy is useful for diagnosis since the MSLT will often demonstrate REM onset sleep in 2 or more nap trials in narcoleptics. This finding, along with an appropriate clinical history, is the basis for diagnosing narcolepsy.

Patients with narcolepsy have a decrease in hypocretin (orexin) production as a result of decreased functioning in hypocretin-producing cells of the hypothalmus.15,16 Activity of these cells are also associated with arousal and hunger.

Use of new drugs that selectively target the hypothalamus, such as a modafanil, have been found to be effective in narcolepsy without the side effects of amphetamine and without significant changes in sleep architecture. Modafanil affects specific cells in the hypothalamus rather than producing general activation throughout the brain.17 Other common treatments include amphetamines, methylphenidate, and pemoline.18 Methylphenidate and dextroamphetamine, that have been shown to produce the most improvement in patients’ alertness compared to modafanil, pemoline, protriptyline, ritanserin, and gamma-hydroxybutyrate.19 Selegeline and modafanil have fewer adverse effects and less abuse potential than the amphetamine-containing compounds.20-23

Severe cataplexy may require additional treatment. Tricyclic antidepressants, such as imipramine, as well as serotonin reuptake inhibitors, such as fluoxetine and citalopram, have been found to be effective.24 Roboxetine has been found to be effective for both excessive sleepiness and cataplexy.25

Although narcolepsy has a genetic basis and is HLA-DR2 linked, about one third of the general population has the "narcolepsy gene," which makes genetic testing useless for screening.

Insomnia without obvious physiological causes is called "psychophysiological insomnia." Psychophysiological insomnia is due to increased sympathetic arousal that is present throughout the 24-hour day.26 The dysphoria of being hyperaroused throughout the day and night results in the insomnia symptom complex of "fatigue, tiredness, sadness, and general malaise" as well as accounting for the paradoxical inability to sleep despite being very tired. Treatment that focuses on decreasing activation only at bedtime may be ineffective in promoting sleep in a strong hyperaroused system or eliminating the subjective complaints associated with insomnia. In patients without strong hyperaroused systems, treatment with hypnotics such as benzodiazepines, zolpidem tartrate, or zeleplon, generally provide fast and effective short-term treatment but not long-term effectiveness. The use of cognitive behavioral therapy (CBT) including sleep hygiene, stimulus control, cognitive restructuring, and sleep restriction have longer response times than using sedating hypnotics but the effects are longer lasting.27-29 Combined treatment of hypnotics and CBT is more efficacious in the short term, but not as good as CBT alone in the long term.

Shift work
An estimated 3 million Americans are involved in shift work.30 When sleep time is altered so that sleep occurs during another part of the 24 hours cycle, there can be a misalignment of body rhythms. Alterations of the sleep wake cycle affect sleep and performance. Sleep time during the daylight hours is shorter by 2-4 hours with more awakenings compared to night sleep, even in permanent night shift workers.

Hypnotics and behavioral interventions can improve the amount and quality of daytime sleep. Typically, hypnotics with short or intermediate half lives such as temazepam, sonata, or ambien are used when sleep periods are initially shifted. This reduces the possibility of residual drug effects called "drug hangover" during the subsequent wake period. Nonpharmacological treatments are preferred long term. A 2-3 hour prophylactic nap prior to a nocturnal work shift significantly reduces sleepiness. Prophylactic naps in combination with caffeine at intervals during the work shift can maintain performance through the wake period.

Restless Legs Syndrome
The restless legs syndrome (RLS) is a common cause of sleep disturbances and affects an estimated 10-15% of the population.31 The unpleasant sensory symptoms are typically described as a crawling or tingling sensation. It is temporarily relieved by moving the affected limbs. Symptoms usually occur at rest and commonly when an individual lays down to sleep. RLS contributes to long sleep latencies and can prolong awakenings during the night. This is reflected in a low sleep efficiency. RLS is diagnosed only through a sleep history. The underlying pathophysiology is not known, although it is believed due to an alteration in neuronal dopaminergic or opioidergiec pathways since treatment with opioidergic or dopaminergic agonists is often beneficial.32-34 Mirapex and neurontin have become commonly used medications to treat RLS. Iron deficiency is associated with RLS, and iron replacement can improve symptoms of those patients with a low serum ferritin, so it is useful to screen for iron deficiency in RLS patients.35

Sleep Apnea
Sleep apnea refers to pauses in breathing during sleep. It is characterized by loud snoring, punctuated with pauses during the cessation of airflow. The prevalence of sleep-disordered breathing increases with age ranging from 5% to 25% in middle age to about 24% in healthy older adults.36,37 In addition to age, the prevalence of sleep apnea and hypopnea is greater in individuals with hypertension, obesity, and other medical conditions including cardiac arrhythmias.

Obstructive apnea or hypopnea is due to a complete or partial collapse of the upper airway while central apnea reflects a loss of CNS stimulation causing an absence of diaphragmatic effort. The exact cause of the collapse or disruption in neural control is not known.

Individuals with sleep apnea often present with complaints of daytime sleepiness and loud snoring. A bed partner is often able to give an accurate account of witnessed apneas. On awakening, the patient often has a dry mouth and may experience morning headaches. Examination of the oral pharynx often reveals a crowded oral airway due to any number of factors including large tonsils and/or adenoids, thick or large tongue relative to the size of the oral cavity, narrowed opening between palatal pillars, long uvula, or soft palate.

A PSG is needed for diagnosis in patients presenting with symptoms of sleep apnea. The AHI indicates the frequency of the sleep-related respiratory events and is a measure of the severity of the sleep apnea along with the minimum oxygen saturation associated with events. Treatment is typically initiated with indices more than 5, but cardiovascular sequellae have been demonstrated with even lower levels of sleep-disordered breathing.38

The most common treatment recommended is PAP, including CPAP and BiPAP. PAP treatment requires that the patient undergo a titration study to determine the effective pressure. Other treatment options include surgery such as UPPP, LAUP, somnoplasty, hyoid advancement or the use of dental appliances to pull the lower jaw forward during sleep, thereby moving the tongue forward and increasing the clearance between the base of the tongue and the oral pharynx. If the apnea occurs only in the supine position, behavioral measures to avoid the supine position are very effective. The use of a tennis ball sewn into a sleep shirt is a common recommendation but has never been subjected to rigid scrutiny and should not be recommended for people with significant sleep-disordered breathing or symptoms. In overweight patients, weight loss is recommended to reduce or eliminate sleep-related breathing problems as well as aid other treatments undertaken. Repeat sleep studies should be performed to evaluate the effectiveness of surgical interventions or dental appliances to treat sleep apnea. Surgical approaches have a success rate of less than 50% and were actually less effective than oral appliances in a recent head-to-head trial.39

Sleep Disorders Centers

Referral to a sleep disorder center is appropriate when sleep testing is needed and usually is valuable in the diagnosis and treatment of patients with challenging clinical symptoms. At sleep programs accredited by The American Academy of Sleep Medicine (AASM), an individual boarded in sleep medicine is required to review the raw data of every sleep study and provide diagnosis, interpretation, and treatment recommendations to the referring physician. Mechanisms for treatment and long-term follow-up are also available at accredited programs if needed.


1. Centers for Medicare & Medicaid Services. Coverage Issues Manual Section 60-17. Available at: www.cms.hhs.gov/coverage/8b3-bbb.asp.

2. Kupfer DJ. REM latency: A psychobiologic marker for primary depressive disease. Biol Psychiatry. 1976;11:159-174.

3. Sangal RB, et al. Maintenance of wakefulness test and multiple sleep latency test. Chest. 1992;101:898-902.

4. Rechtscaffen A, Kales T. A manual of standardized terminology, techniques, and scoring system for sleep stages in human subjects. Number 204 in National Institutes of Health Publications. Washington DC: US Government Printing Office; 1968.

5. Bonnet MH. Performance and sleepiness as a function of frequency and placement of sleep disruption. Psychophysiology. 1986; 23:263-271.

6. Martin SE, et al. The effect of clustered versus regular sleep fragmentation on daytime function. J Sleep Res. 1999;9:305-311.

7. Stepanski E, et al. Experimental sleep fragmentation in normal subjects. Int J Neurosci. 1987;33:207-214.

8. Stepanski E, et al. Sleep fragmentation and daytime sleepiness. Sleep. 1984;7:18-26.

9. Chervin RD, Aldrich MS. Sleep onset REM periods during multiple sleep latency tests in patients evaluated for sleep apnea. Am J Resp Crit Care Med. 2000;161:426-431.

10. Benbadis SR, et al. Association between the Epworth Sleepiness Scale and the multiple sleep latency test in a clinical population. Ann Intern Med. 1999;130:289-292.

11. Aldrich MS. Narcolepsy. N Engl J Med. 1990;323(6)389-394.

12. National Institutes of Health. National Heart, Lung and Blood Institute. Narcolepsy. Bethesda, Md: National Institutes of Heath, National Heart, Lung and Blood Institute. NIH Publication No 96-3649, 1996.

13. Mignot E. Genetic and familial aspects of narcolepsy. Neurology. 1998;50:S16-S22.

14. Thorpy M. current concepts in the eriology, diagnosis and treatment of narcolepsy. Sleep Med. 2001;2:5-17.

15. Peyton C, et al. A mutation in a case of early onset narcolepsy and a generalized absence of hypocretin peptide in human narcoleptic brains. Nat Med. 2000;6:991-997.

16. Thannickal TC, et al. Reduced number of hypocretin neurons in human narcolepsy. Neuron. 2000;27:469-474.

17. Chemelli RM, et al. Narcolepsy tin orexin knockout mice: molecular genetics of sleep regulation. Cell. 1999;98(4):437-451.

18. Mitlerr MM, et al. Narcolepsy and its treatment with stimulants. Sleep. 1994;17:352-371.

19. Mitler M, Hajduckovic R. Relative efficacy of drugs for the treatment of sleepiness in narcolepsy. Sleep. 1991;14:218-220.

20. Mayer G, et al. Selegelide hydrochloride treatment in narcolepsy: a double-blind placebo controlled study. Clin Neuropharmacol. 1995;18(4):306-319.

21. Ferraro l, et al. Modafinil: An antinarcoleptic drug with a different neurochemical profile to d-amphetamine and dompanine uptake blockers. Biol Psychiatry. 1997;42(12):1181-1183.

22. Gold LH, Balser R. Evaluation of the cocaine-like discriminative stimulus effects and reinforcing effects of modafinil. Psychopharmacology. 1996;126:286-292.

23. Broughton RJ, et al. Randomized double-blind placebo-controlled crossover trial of modafnil in the treatment of excessive sleepiness in narcolepsy. Neurology. 1997;49:444-451.

24. Thirumalai SS, Shubin RA. The use of citalopram in resistant cataplexy. Sleep Med. 2000;1:313-316.

25. Larrosa O. Stimulant and anticataplectic effects of roboxetine in patients with narcolepsy: A pilot study. Sleep. 2001;3:282-285.

26. Bonnet MH, Arand DL. Hyperarousal and insomnia. Sleep Med Rev. 1997;1:97-108.

27. Hauri P. Can we mix behavioral therapy with hypnotics when treating insomniacs? Sleep. 1997;20:1111-1118.

28. Morin C, et al. Behaviroal and pharmacological therapies for late-life-insomnia. A randomized controlled trial. JAMA. 1999;281: 991-999.

29. Rosen RC, et al. Psychophysioligcal insomnia: Combined effects of pharmacotherapy and relaxation-based treatments. Sleep Med. 2000;1:279-288.

30. US Congress, Office of Technology Assessment. Biological Rhythms: Implication for the Worker. (OTA-BA-463) Washington, DC, US Government Printing Office. September 1991.

31. Lavigne GJ, Montplaisir JY. Restless legs syndrome and sleep bruxism: Prevalence and association among Canadians. Sleep. 1994;17:739-743.

32. Winkelmann J, et al. Opioid and dopamine antagonist drug challenges in untreated restless legs syndrome patients. Sleep Med. 2001;2:57-61.

33. Wetter TC, et al. A randomized controlled study of pergolide in patients with restless legs syndrome. Neurology. 1999;52: 944-950.

34. Hening WA, et al. Dyskinesias while awake and periodic movements in sleep in restless legs syndrome: Treatment with opioids. Neurology. 1986;36:1363-1366.

35. Sun ER, et al. Iron and the restless legs syndrome. Sleep. 1998; 21(4):371-377.

36. Ancoli-Israel S, et al. Sleep disordered breathing in community dwelling elderly. Sleep. 1991;1:3-17.

37. Young T, et al. The occurrence of sleep disordered breathing among middle-age adults. N Engl J Med.1993;328:1230-1235.

38. Shahar E, et al. Sleep-disordered breathing and cardiovascular disease. Cross-sectional results of the Sleep Heart Health Study. Am J Respir Crit Care Med. 2001;163:19-25.

39. Engstrom ML, et al. 4-year follow-up of treatment with dental appliance or uvulopalatopharyngoplasty in patients with obstructive sleep apnea: a randomized study Walker. Chest. 2002;121: 739-746.