Chapter 85 Narcolepsy
Diagnosis and Management
Narcolepsy is a syndrome with a prevalence close to 0.04%. It is a chronic neurologic sleep disorder resulting from the abnormal regulation of the sleep–wake cycle, resulting in abnormal sleep tendencies, including excessive daytime sleepiness, disturbed nocturnal sleep, and manifestations related to rapid eye movement (REM) sleep such as cataplexy (an abrupt drop in muscle tone), sleep paralysis, and hypnopompic or hypnagogic hallucinations. Its peak age of onset is the second decade. Multiple sleep latency testing following overnight polysomnography shows short sleep latencies (mean sleep latency ≤8 minutes) and two or more sleep-onset REM periods (SOREMPs). The treatment of narcolepsy has improved in recent years, with the newest compounds providing fewer side effects than the classic stimulants for excessive daytime somnolence and tricyclic antidepressants for cataplexy. Modafinil and armodafinil improve daytime somnolence, and sodium oxybate improves both cataplexy and daytime somnolence. Narcolepsy is a complex syndrome that involves dysfunction in the timing of changes in wake and sleep stages. Medications with an alerting effect can help but do not completely control the multifaceted problems associated with this syndrome.
Gélineau1 first coined the term narcolepsy in 1880 to designate a pathologic condition characterized by irresistible episodes of sleep periods of short duration recurring at close intervals. Although Westphal2 and Fisher3 previously published reports of patients with sleepiness and episodic muscle weakness, Gélineau was the first to characterize narcolepsy as a distinct syndrome. He wrote that falls, or “astasias” sometimes accompanied attacks. Henneberg later referred to these attacks as cataplexy.4 In the 1930s, Daniels5 emphasized the association of daytime sleepiness, cataplexy, sleep paralysis, and hypnagogic hallucinations. Referring to these symptoms as “the clinical tetrad,” Yoss and Daly6 and Vogel7 reported nocturnal sleep-onset rapid eye movement (REM) periods (SOREMPs) in narcoleptic patients, a finding confirmed in the following years.8–11
The current definition of narcolepsy strongly emphasizes the dysfunction of REM sleep, but it ignores the notion that genetic factors are often involved in the development of narcolepsy. Moreover, although most cases of narcolepsy are idiopathic, secondary causes of narcolepsy have been described.12 The discovery of the hypocretin (orexin) mutation in narcoleptic Doberman pinschers led to a better understanding of the syndrome and changed, to some degree, our approach to diagnosis and treatment of disorders of excessive daytime sleepiness, particularly narcolepsy. Because hypocretins are involved in many hypothalamic functions, narcolepsy may be considered more than just a disorder of sleep. In addition, it may be considered a syndrome of instability of wake and sleep stages rather than merely a disorder of dysfunctional REM sleep: Patients have the capacity to achieve wakefulness, non-REM (NREM) sleep, and REM sleep, but they are unable to maintain the state. They appear to lack the modulator responsible for maintaining the active sleep state long enough for the normal physiologic switches to change the state. Thus, patients with narcolepsy dissociate into the various states of consciousness at inappropriate times. This dissociation is often incomplete, leading to a mixture of normal and abnormal states of consciousness, such as cataplexy, which represents a combination of the waking state and the paralysis of REM sleep.13
The International Classification of Sleep Disorders includes two entities: narcolepsy with cataplexy and narcolepsy without cataplexy.14 The classic tetrad for narcolepsy includes excessive daytime sleepiness, cataplexy, sleep paralysis, and hypnagogic hallucinations (Video 85-1). Automatic behavior and disrupted nighttime sleep also commonly occur. Not all of these symptoms are present in all patients. Many of the symptoms of narcolepsy can occur in any person who is severely sleep deprived; only cataplexy is unique to narcolepsy.
Unwanted episodes of sleep recur several times a day not only under favorable circumstances, such as during monotonous sedentary activity or after a heavy meal, but also in situations in which the patient is fully involved in a task. The duration of these sleep attacks can vary from a brief episode of a few seconds to several minutes, if the patient is in an uncomfortable position, to longer than an hour if the patient is reclining. Narcoleptic patients characteristically wake up feeling refreshed, and there is a refractory period of one to several hours before the next episode occurs. These relatively short but refreshing naps can help differentiate patients with narcolepsy from patients with idiopathic hypersomnia, who take long and unrefreshing naps. Despite having sleep attacks during the day, narcoleptic patients generally do not spend a greater amount of time sleeping than people without narcolepsy. Apart from sleep attacks, patients might feel abnormally drowsy, resulting in poor performance at work, memory lapses, and even gestural, ambulatory, or speech automatisms.
Cataplexy occurs in 60% to 70% of narcoleptic patients.15,16 Cataplexy is an abrupt and reversible decrease or loss of muscle tone, most commonly elicited by emotional responses such as laughter, anger, or surprise. It can involve certain muscles or the entire voluntary musculature. Most typically, the jaw sags, the head falls forward, the arms drop to the side, and the knees unlock or buckle (Video 85-2). Awareness is preserved throughout the attack.
The severity and extent of cataplectic attacks can vary from a state of absolute powerlessness, which seems to involve the entire voluntary musculature, to a limited involvement of certain muscle groups, or to no more than a fleeting sensation of weakness extending more or less throughout the body. Although the extraocular muscles supposedly are not involved, weakness can occur, and the patient might complain of blurred vision. Complete paralysis of extraocular muscles has never been reported, although the palpebral muscle may be affected.
Speech may be impaired and respiration can become irregular during an attack. Long breathing pauses have never been recorded, but short pauses similar to those seen during nocturnal REM sleep in healthy subjects can occur. Speech may be broken or slurred because of intermittent weakness affecting the arytenoid muscles. If the weakness involves only the jaw or speech, the subject might have wide masticatory movement or an unusual attack of stuttering speech.
Rarely in a cataplectic attack, there is a complete loss of muscle tone that can lead to total body collapse, and serious injuries such as skull or other bone fractures can occur. However in most cases, cataplectic attacks are not that extreme and might even go unnoticed by persons nearby. An attack might consist of only a slight buckling of the knees. Patients might perceive this abrupt and short-lasting weakness and simply stop or stand against a wall. If the attack involves the upper limbs, the patient might complain of “clumsiness,” reporting activity such as dropping cups or plates or spilling liquids when surprised, laughing, and so forth. As seen during nocturnal REM sleep, the abrupt muscle inhibition is interrupted by sudden bursts of muscle tone returning, which at times even seem enhanced.
The short cataplectic attacks are the most common manifestation of cataplexy. Because they do not resemble the classic full-blown attack of cataplexy, they are often missed even by skilled physicians without the aid of electromyographic recording. By the same token, the skilled physician must be cautious not to overdiagnose as cataplexy normal phenomena such as the “rubber knees” that precede the anxiety of public speaking or “rolling on the ground laughing.” The duration of each cataplectic attack—whether partial or complete—is variable, lasting from a few seconds to 30 minutes, but in most cases 30 seconds to 2 minutes. Attacks can be elicited by emotion or stress. Laughter and anger seem to be the most common triggers, but these attacks can also be induced by a feeling of elation while listening to music, reading a book, or watching a movie. Cataplexy can be induced merely by remembering a happy or funny situation, and it occurs rarely without clear precipitating acts or emotions, particularly if the patient is sleepy. It often occurs when the patient is telling a joke and even when the patient simply anticipates saying something humorous.
This is a terrifying experience that occurs in the narcoleptic patient upon falling asleep or awakening. Patients find themselves paralyzed, suddenly unable to move the limbs, speak, or even breathe deeply. This state is often accompanied by hallucinations. During an episode of sleep paralysis, the patient is unable to move the extremities, speak, or open the eyes, although he or she is fully aware of the condition and able to recall it completely later. In many episodes of sleep paralysis, but especially the first occurrence, the patient may be subjected to extreme anxiety associated with the fear of dying. This anxiety is often greatly intensified by the terrifying hallucinations that can accompany the sleep paralysis. Patients often interpret the experience as being “scared stiff.” With more experience with the phenomenon, however, the patient usually learns that episodes are brief and benign, rarely lasting longer than a few minutes and always ending spontaneously. Sleep paralysis can occur as an independent and isolated phenomenon, and 3% to 5% of the general population may be affected by it.
Sleep onset, either during daytime naps or at night, may be unpleasant due to vivid auditory or visual hallucinations, known as hypnagogic hallucinations (Video 85-3). Similar hallucinations can occur upon awakening; these hypnopompic hallucinations may be more characteristic of narcolepsy than the hypnagogic ones. The visual hallucinations usually consist of simple forms (colored circles, parts of objects, and so forth) that are either constant or changing in size. The image of an animal or a person can occur abruptly and more often in color. Auditory hallucinations are also common, although other senses are seldom involved. The auditory hallucinations can range from a collection of sounds to an elaborate melody. The patient may also be frightened by threatening sentences or harsh invective. Often hypnopompic hallucinations are perceived as so vividly realistic that the patient acts on them upon awakening. For example, patients with narcolepsy have been known to call the police due to intruders being in the home, only to discover after authorities have searched the house that it was all a hallucination. The exact boundary between hypnagogic and hypnopompic hallucinations and dreams is not a clear one. In some cases of unrecognized narcolepsy with daytime hypnagogic or hypnopompic hallucinations, the patient may be mistakenly determined to have a delusional psychosis.17
Another common and interesting type of hallucination reported at sleep onset involves elementary cenesthopathic feelings (i.e., experiencing picking, rubbing, or light touching), changes in location of body parts (e.g., arm or leg), or feelings of levitation or extracorporeal experiences (e.g., moving the body in space or floating above the bed) that may be quite elaborate. For example, the patient might report, “I am above my bed and I can also see my body below” or “I am a few feet up and people are jumping over my body.” The association of sleep paralysis has led researchers to postulate gamma loop involvement in some of these hallucinations. The abrupt motor inhibition that involves the spinal cord motor neurons can lead to significant decrease in the feedback of information that is normally used by the central nervous system (CNS) to gauge the position of the body and the relation of the limb segments to each other.
Narcolepsy patients commonly experience disturbed nocturnal sleep. Night sleep is often interrupted by repeated awakenings and sometimes accompanied by terrifying dreams. Ironically, patients complain of difficulty falling asleep and staying asleep at night, although they might fall asleep repeatedly during the daytime. Rarely, insomnia with secondary daytime tiredness is the initial complaint. The sleep disruption may be enhanced by the presence of periodic limb movements, which appear to be common in these patients. These movements can be associated with typical episodes of REM sleep behavior disorder.
The first symptom often develops near the age of puberty; the peak age at which reported symptoms occur is 15 to 25 years, but narcolepsy and other symptoms have been noted as early as 2 years, and at 6 months of age in the case with hypocretin gene mutation.18 A second, smaller peak of onset has been noted between 35 and 45 years and near menopause in women. In those who developed narcolepsy at age 60 years or later, cataplexy was the most common initial symptom.19 There is often a marked delay of more than 10 years before diagnosis is made, especially if cataplexy is initially absent (Video 85-4).20 A survey of 157 narcoleptic patients found that the majority experienced symptoms before 30 years of age; only 21% of the patients had their first symptom after 30 years of age.19
Excessive daytime sleepiness and irresistible sleep attacks usually occur as the first symptoms, either independently or associated with one or more symptoms. They are enhanced by high environmental temperature, indoor activity, and idleness. Symptoms might abate with time but never phase out completely. Attacks of cataplexy usually appear in conjunction with abnormal episodes of sleep but can occur as long as 20 years later. They occasionally occur before the abnormal sleep episodes, in which case they are a major source of difficulty in diagnosis. They can vary in frequency from a few episodes during the subject’s entire lifetime to one or several episodes per day. Cataplectic attacks have an overall tendency to decrease in frequency with aging.
Hypnagogic and hypnopompic hallucinations and sleep paralysis do not affect all subjects, are often transitory, and occur commonly in the general population.21 Disturbed nocturnal sleep seldom occurs in the first stages and usually worsens with age.22 Narcolepsy leads to a variety of complications, such as accidents related to driving or operating machines; difficulties at work leading to disability, forced retirement, or job dismissal; impotence; and depression.23
The Stanford Sleepiness Scale,24 a seven-point scale, was developed to quantify the subjective sleepiness of patients throughout the day, but it is often difficult for patients to accurately rate themselves every 15 to 20 minutes. The Epworth Sleepiness Scale (ESS) is also used as an index of subjective sleepiness (see Chapter 143). The ESS is not validated in children. The Pediatric Daytime Sleepiness Scale, a validated scale, is more appropriate for use in children and teenagers.
Several tests have been designed to evaluate sleepiness objectively. The multiple sleep latency test (MSLT; see Chapter 143) was designed to measure physiologic sleep tendencies in the absence of alerting factors.25 This test consists of five scheduled naps, usually at 10 AM, noon, and 2, 4, and 6 PM, during which the subject is polygraphically monitored in a comfortable, soundproof, dark bedroom, while wearing street clothes. The MSLT records the latency for each nap (time between lights out and sleep onset), the mean sleep latency, and the presence or absence of REM sleep during any of the naps.26 On the basis of polygraphic recording, REM sleep that occurs within 15 minutes of sleep onset is considered a SOREMP.27 After each 20-minute monitoring period, patients stay awake until the next scheduled nap.
In the normal population, MSLT scores vary with age, and puberty is a critical landmark. Prepubertal children between the ages of 6 and 11 years appear to be hyperalert. In postpubertal subjects, mean MSLT scores under 8 minutes are generally considered to be in the pathologic range; those over 10 minutes are considered to be normal. When the range is between 8 and 10 minutes, age factors interact, so the test must be interpreted with greater care; mean scores of 8 to 10 minutes represent a gray area.28 An MSLT performed alone has drawbacks: it measures sleepiness regardless of its cause, which may simply be sleep deprivation. The MSLT also ignores repetitive microsleeps that can lead, in borderline cases, to daytime impairment not scored by conventional analysis. To be clinically relevant, the test must be conducted under specific conditions. Subjects must have abstained from medication for a sufficient period (usually 15 days) so that drug interaction is avoided. On the basis of sleep diaries, their sleep–wake schedules are stabilized. On the night preceding the MSLT, the subjects undergo a standard nocturnal polysomnogram. Throughout the total nocturnal sleep period, any sleep-related biologic abnormalities responsible for sleep fragmentation and sleep deprivation are recorded. The nocturnal polysomnogram can indicate the underlying cause of sleepiness; the MSLT indicates the severity of the problem. Once the nocturnal sleep recording has eliminated specific diseases, the MSLT confirms the diagnosis of narcolepsy with the presence of two or more SOREMPs and a mean sleep latency of 8 minutes or less.
Browman and colleagues29 proposed adding to the MSLT a test for maintaining wakefulness (MWT). The MWT tests the patient’s ability to remain awake in a comfortable sitting position in a dark room for five 20-minute trials given at 10 AM, noon, and 2, 4, and 6 PM. The MWT may be helpful in specific pharmacologic trials, in evaluating response to a treatment for alertness, and in evaluating the risk of falling asleep associated with specific jobs or activities, but it has proved to be unsatisfactory as a diagnostic procedure.29 Traditionally, the test is composed of five nap opportunities at 2-hour intervals, similar to the MSLT. Each opportunity for sleep lasts 20 minutes. More recently, the test consists of four 40-minute sessions administered at 2-hour intervals, with a cutoff point around 30 minutes before falling asleep. MWT is typically useful in distinguishing between a pathologic process and normality.
Another procedure that has been used to document sleepiness is a continuous 24- or 36-hour polysomnogram that provides information about the number, duration, times, and types of daytime sleep episodes, as well as documenting nighttime sleep disruptions. In addition, this long polysomnographic recording can identify the dissociated REM sleep–inhibitory process characterizing cataplexy by showing the absence of chin and muscle twitches, which are typical of REM sleep, in the awake patient. These recordings may be performed using ambulatory equipment.
By consensus, the diagnosis of narcolepsy requires a clinical history of excessive daytime sleepiness and a positive MSLT, with a mean sleep latency of no more than 8 minutes and two or more SOREMPs.14 However, there is controversy concerning the criteria needed to confirm narcolepsy in patients with excessive daytime sleepiness. There is a progressive decrease in the number of SOREMPs and an increase in the mean sleep latency on the MSLT as a function of age, suggesting that the current criteria used for diagnosis may be too stringent in older patients.30
Clinicians and researchers in Japan have indicated that their U.S. counterparts have given too much credit to polysomnographic criteria and the presence of two or more SOREMPs on MSLT.31 General population investigations have shown that normal subjects can have two or more SOREMPs, and there seems to be a greater chance of having narcolepsy with the presence of two or more SOREMPs, but the likelihood of having narcolepsy in relation to an increasing number of SOREMPs is unknown.32
In Japan, a positive history of cataplexy associated with excessive daytime sleepiness is systematically required for the diagnosis of narcolepsy.31 Undoubtedly, the presence of cataplexy is pathognomonic of narcolepsy, but it may be difficult to rely on this criterion alone, particularly when cataplexy is partial (i.e., limited to the head and neck or neck and upper arms). Anic-Labat and colleagues33 developed a self-administered questionnaire that validated cataplexy in 1000 subjects.
Moscovitch and coworkers34 also cautioned against the diagnosis of narcolepsy if excessive daytime sleepiness and two or more SOREMPs at MSLT are the only positive findings. These authors found that only 84% of the persons with complaints of excessive daytime sleepiness and documentation of cataplexy had two or more SOREMPs on MSLT, a replication of findings by Van den Hoed and colleagues28 10 years earlier. They found that all subjects with excessive daytime sleepiness and cataplexy had two or more SOREMPs at one MSLT if the MSLT was repeated daily for 4 days. Moscovitch and coworkers34 recommended that the term narcolepsy be used only when excessive daytime sleepiness and at least a positive history of cataplexy are associated with two or more SOREMPs. If there is no history of cataplexy, a more descriptive term such as excessive daytime sleepiness with several SOREMPs could be used, even if other symptoms such as sleep paralysis or hypnagogic hallucinations exist.
Narcolepsy occurs sporadically in 99% of patients, but genetic factors probably play a role in which environmental factors to act on. Genetic testing has been used to aid in the clinical diagnosis of narcolepsy. The HLA DQB1*0602 is the most specific genetic marker for narcolepsy across all ethnic groups, and it is found in 95% of narcoleptic patients with cataplexy.35,36 Mignot’s group35 showed that 40% of subjects with narcolepsy and no cataplexy were positive for the human leukocyte antigen (HLA) DQB1*0602, and the HLA is positive in 18% to 35% of the general population (without narcolepsy or cataplexy). Thus, genetic testing alone is not sufficient for the diagnosis of narcolepsy. The systematic requirement of a positive history of cataplexy would eliminate subjects in the developing phase of the syndrome, which can take several years. However, the use of strict criteria will allow better epidemiologic studies to be conducted. Subjects would be classified as narcoleptic only after the presence of excessive daytime sleepiness and cataplexy had been reported and abnormal daytime alertness had been confirmed by polysomnographic recording with MSLT.
The discovery of cerebrospinal fluid (CSF) hypocretin (orexin) has led to a new diagnostic approach. In 1998, two different groups independently discovered that a hypothalamic peptide neurotransmitter (named hypocretin by one group and orexin by the other)37,38 played a major role in maintenance of wakefulness. Hypocretin is produced exclusively by a group of neurons located in the lateral hypothalamus and has projections widely throughout the CNS.39
In 1999 Stanford researchers discovered an autosomal recessive mutation in the hypocretin receptor 2 gene (HCRTR2) responsible for narcolepsy in canines.40 This discovery identified hypocretins as the major sleep-modulating neurotransmitter.40 However, narcolepsy in humans is not a result of gene mutations; rather, narcoleptic humans were found to have very low levels of CSF hypocretin. Hypocretin neurons and functions are selectively damaged in patients with narcolepsy, although the cause is still unknown.39 By performing a lumbar puncture, a very low or nonexistent level of hypocretin can confirm the diagnosis of narcolepsy with cataplexy.40–43 CSF hypocretin-1 levels less than 110 ng/L (measured using the Stanford University technique), have a high positive predictive value (94%) for narcolepsy with cataplexy.39,44
Currently, CSF hypocretin measurement is the most accurate diagnostic technique available; however, narcolepsy is recognized in a low percentage (less than 9%) of subjects despite normal levels of CSF hypocretin. In rare cases, low CSF hypocretin levels may be due to Guillain-Barré syndrome, brain tumors, encephalitis, vascular diseases, and brain trauma.39
At present, there is no cure for narcolepsy. The goal of all therapeutic approaches is to control narcoleptic symptoms and allow the patient to have a full personal and professional life.44a Other sleep disorders may co-exist (Video 85-5). Treatment goals specifically include control of excessive daytime sleepiness, cataplexy attacks, hypnagogic or hypnopompic hallucinations, sleep paralysis, improvement in nocturnal sleep, and psychosocial difficulties. Drug prescriptions must take into account possible side effects because narcolepsy is a lifelong illness and patients have to receive medication for years. Tolerance or addiction can occur with some compounds. In addition, hypertension, abnormal liver function, and psychosis are the most commonly reported complications associated with the long-term use of stimulant medications. The treatment of narcolepsy must balance maintaining an active life with avoiding side effects and developing tolerance to medications.45
Part of the difficulty in treating patients who have narcolepsy involves the patient’s frustration over the delay in diagnosis. In 500 narcoleptic patients surveyed in United States in the 1980s, the mean time to diagnosis from symptom onset was 15 years. The consequence of this, particularly in a young patient, is the development of reactive depressive symptoms. One of the most important initial treatments is a referral to patient support groups organized by sleep disorders centers, such as the National Sleep Foundation or the Narcolepsy Network. Other support groups exist in most western European countries and in North America. The American Academy of Sleep Medicine is an integral resource.
Career counseling is also important because patients and their employers must be educated regarding jobs that patients with narcolepsy should avoid, including shift-work, on-call schedules, driving and the transportation industry, or any job necessitating continuous attention for long hours without breaks, particularly under monotonous conditions. Some of these difficulties can be overcome if the employer recognizes the importance of short 15 to 20 minute naps every 4 hours during the daytime. In addition to scheduled naps, other important behavioral treatment targets include following a regular sleep–wake schedule, avoiding frequent time zone changes, and practicing good sleep hygiene.