Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
NARCOLEPSY AND IDIOPATHIC HYPERSOMNIA Idiopathic Hypersomnia: A Study of 77 Cases Kirstie N. Anderson, MB BS, D Phil1; Samantha Pilsworth, Bsc1; Linda D. Sharples, PhD1,2; Ian E. Smith, MA, MD1; John M. Shneerson, MA, DM1 Respiratory Support and Sleep Centre, Papworth Hospital Papworth Everard, Cambridge, United Kingdom; 2MRC Biostatistics Unit, Cambridge, United Kingdom 1 Study Objectives: To review the clinical and polysomnographic characteristics of idiopathic hypersomnia as well as the long-term response to treatment. Setting: The Respiratory Support and Sleep Centre at Papworth Hospital, Cambridge, UK. Patients and Design: A large database of more than 6000 patients with sleep disorders was reviewed. A retrospective study of the clinical and polysomnographic characteristics of 77 patients with idiopathic hypersomnia was performed. Comparison with a similar group of patients with narcolepsy was performed. The response to drug treatment was assessed in 61 patients over a mean follow-up of 3.8 years. Measurements and Results: Idiopathic hypersomnia was 60% as prevalent as narcolepsy. Comparison with a similar group of patients with narcolepsy showed that those with idiopathic hypersomnia were more likely to have prolonged unrefreshing daytime naps, a positive family history, increased slow-wave sleep, and a longer sleep latency on the Multiple Sleep Latency Test. The results of the Multiple Sleep Latency Test were not helpful in predicting disease severity or treatment response. The clini- cal features were heterogeneous and of variable severity. The majority of patients with idiopathic hypersomnia had symptoms that remained stable over many years, but 11% had spontaneous remission, which was never seen in narcolepsy. Two thirds of patients with idiopathic hypersomnolence had a sustained improvement in daytime somnolence with medication, although a third needed high doses or combinations of drugs. Conclusions: Idiopathic hypersomnolence has characteristic clinical and polysomnographic features but the prolonged latency on the Multiple Sleep Latency Test raises doubt about the validity of this test within the current diagnostic criteria. The disease often responds well to treatment and a substantial minority of patients appear to spontaneously improve. Keywords: Idiopathic hypersomnia, narcolepsy, polysomnography, treatment Citation: Anderson KN; Pilsworth S; Sharples LD; Smith IE; Shneerson JM. Idiopathic hypersomnia: a study of 77 cases. SLEEP 2007;30(10):12741281. THERE ARE MANY CAUSES OF EXCESSIVE DAYTIME SLEEPINESS (EDS), INCLUDING DRUGS, TOXINS, METABOLIC DERANGEMENTS, RESPIRATORY CAUSES SUCH as obstructive sleep apnea, and structural lesions such as stroke or head injury.1 In 1880, Gelineau2 first used the term narcolepsy to describe a primary sleep disorder causing irresistible sleep attacks and “astasia” (falling), with Adie3 later using the term cataplexy to describe the loss of muscle tone with emotion. Later, the typical features of cataplexy, hypnagogic hallucinations, and sleep paralysis suggested that patients with narcolepsy had disturbed rapid eye movement (REM) sleep. This was confirmed with the discovery of the sudden onset of REM sleep in these patients.4 As the disease became better defined, it became apparent that there were a number of other patients who were excessively sleepy, often with prolonged nocturnal sleep, who did not fit the diagnostic criteria for narcolepsy. Idiopathic hypersomnia was identified relatively recently by Bedric Roth. He described a large cohort of patients distinct from Disclosure Statement This was not an industry supported study. Dr. Shneerson has participated in research activities supported by Boehringer Ingelheim, UCB Pharma, GlaxoSmithKline, Lundbeck and Pfizer. Drs. Anderson, Sharples, Smith and Ms. Pilsworth have reported no financial conflicts of interest. Submitted for publication February, 2007 Accepted for publication May, 2007 Address correspondence to: Dr. Kirstie N. Anderson Respiratory Support and Sleep Centre, Papworth Hospital Papworth Everard, Cambridge, United Kingdom CB8 3RE; Tel: 44 0 1480 830541; Fax: 44 (0)1480 364568; E-mail: [email protected] SLEEP, Vol. 30, No. 10, 2007 1274 those with narcolepsy with EDS. These patients had marked sleepiness, often with prolonged nocturnal sleep but without clinical or electrophysiologic features of REM sleep disturbance.5, 6 Roth described patients with EDS alone (monosymptomatic) but also patients with prolonged night sleep who exhibit sleep drunkenness on waking (polysymptomatic). Sleep drunkenness refers to prolonged difficulty waking with automatic behavior, confusion, and repeated returns to sleep. A family history was common, and the condition was reported as chronic with a poor response to treatment. The disease was defined as a distinct disease entity in 1979 in the International Classification of Sleep Disorders (ICSD). Further attempts have been made to characterize idiopathic hypersomnia on clinical and electrophysiologic grounds, and there have been several series of patients published since Roth’s original report.7-12 However, there has been a lack of consensus in the literature about the definition of idiopathic hypersomnia and the presence or absence of diagnostic subcategories. The diagnosis of idiopathic hypersomnia relies upon the exclusion of other causes of EDS, but the absence of its own pathognomonic clinical and laboratory features can make a positive diagnosis more difficult. This may account for the wide variation in the reported prevalence of the condition. The most recent ICSD diagnostic criteria13 for both idiopathic hypersomnia and narcolepsy are summarized in table 1. Idiopathic hypersomnia is divided into 2 groups, depending on whether night sleep is prolonged (> 10 hours) or not. It is important to note that the case series published since the original description by Roth have either not used Multiple Sleep Latency Test (MSLT) scores or used variable mean MSLT scores as part of their patient entry criteria. Very little has been published regarding treatment, particularly with newer drugs such as modafinil. It is common practice Idiopathic hypersomnia—Anderson et al nographic data were then reviewed for all those with possible idiopathic hypersomnia or narcolepsy to see if they satisfied the inclusion criteria for the study. To retain a diagnosis of idiopathic hypersomnia, the following criteria had to be met: (1) symptoms of EDS lasting for longer than 1 year and an Epworth Sleepiness Scale (ESS) score 14 of 10 or higher with daytime napping on most days, (2) no improvement after a trial of increased nighttime sleep, (3) more than 2 sleep-onset REM periods (SOREMPS) on Multiple Sleep Latency Test (MSLT) when MSLT was performed, (4) no features to suggest cataplexy, and (5) an apnea and hypopnea index (AHI) of less than 10. Although the diagnosis of obstructive or central apnea is defined by an AHI of 5 or more, this would be considered mild disease, unlikely to require treatment with continuous positive airway pressure or to cause significant daytime somnolence. Apnea was defined as airflow cessation for more than 10 seconds and a hypopnoea was defined as longer 10 seconds of decreased airflow or respiratory effort with a decrease in oxyhemoglobin saturation of at least 4% or followed by an arousal. Additional inclusion criteria included (6) fewer than 20 periodic limb movements per hour of sleep and (7) no other apparent cause for sleepiness. Patients were excluded if they had incomplete polysomnography or had a secondary brain lesion or injury as a cause of hypersomnolence, if they were on sedative medication, or if they had any coexistent medical or psychiatric conditions that might have contributed to EDS. Of note, all patients completed the Beck Depression Inventory on the night of their in-patient stay for polysomnography, and patients with a Beck Depression Inventory score of greater than 9, who were felt to have significant depression, were excluded. These criteria fulfil the most recent diagnostic criteria used by the ICSD (2005), with the exception of MSLT scores; we did not define an upper limit of sleep latency on the MSLT in our patients. Seventy-seven patients within the database fulfilled the inclusion criteria and made up the study cohort. A comparative group of patients with narcolepsy was obtained from the patient database. They had all had the same detailed clinical assessments and all met the following criteria: (1) EDS for greater than 1 year, (2) definite cataplexy, (3) 2 or more SOREMP on MSLT, (4) AHI of less than 10, and (5) periodic limb movement index of less than 20. Sixty-four patients from the 126 existing patients within the database satisfied these criteria and were included as a group for comparison within the study. The majority of patients with narcolepsy who were excluded had associated obstructive sleep apnea, periodic limb movements of sleep, or coexistent medical conditions or had not had complete polysomnography at Papworth. Detailed clinical notes were available for all patients and were reviewed. All patients in the study group had polysomnography. The ESS score was recorded at every clinic visit for each patient. Wherever possible, patients were seen 2 to 3 weeks after the introduction of any treatment to document any benefit and side effects. Improvement after treatment was defined as a drop in the ESS score by at least 4 points, as this was the criterion used in the randomized controlled trials examining the use of modafinil in patients with narcolepsy.15 A prospective clinical assessment was also performed when possible and 1 author (KA) attempted to contact all patients with idiopathic hypersomnia; 33 were not available. The remaining 44 were interviewed in person in the clinic or by telephone with a detailed assessment covering medi- Table 1—Criteria for Idiopathic Hypersomnia and Narcolepsya Idiopathic hypersomnolence without long sleep time A Complaint of EDS for > 3 months B Normal nocturnal sleep < 6 h but < 10 h C Nocturnal polysomnography has excluded other causes of EDS D Polysomnography findings: major sleep period > 6 h and < 10 h with short sleep latency E MSLT findings: mean sleep latency < 8 minutes and < 2 SOREMP F Hypersomnia is not better explained by another sleep disorder, medical or neurologic disorder, mental disorder, medication use or substance abuse Idiopathic hypersomnolence with long sleep time A Complaint of EDS for > 3 months B Prolonged sleep time > 10 h with laborious wakening in the morning or from naps C Nocturnal polysomnography has excluded other causes of EDS D Polysomnographic findings: a major sleep period > 10 h in length with a short sleep latency E MSLT findings if performed: mean sleep latency < 8 minutes and < 2 SOREMP F Hypersomnia is not better explained by another sleep disorder, medical or neurologic disorder, mental disorder, medication use or substance abuse. Narcolepsy A Complaint of EDS for > 3 months B Cataplexy (sudden and transient episodes of loss of muscle tone triggered by emotions)b C Either polysomnographic findings: nocturnal sleep of at least 6 h followed by mean sleep latency on MSLT < 8 minutes with > 2 SOREMP or hypocretin levels in cerebrospinal fluid < 110 pmol D Hypersomnia is not better explained by another sleep disorder, medical or neurologic disorder, mental disorder, medication use or substance abuse Taken from the current International Classification of Sleep Disorders criteria.13 EDS refers to excessive daytime sleepiness; MSLT multiple sleep latency test. b In the absence of cataplexy, the patient should have definite sleeponset REM periods (SOREMP) a to treat these patients with stimulants but, unlike in patients with narcolepsy, planned naps are unhelpful, as they are both long and unrefreshing. Therefore, a number of questions remain unanswered, including appropriate diagnostic criteria and, in particular, the helpfulness of the MSLT, disease prevalence, the natural history of the condition, and response to treatment. We have reviewed a large database of patients with sleep disorders and identified 77 patients with idiopathic hypersomnia. Their clinical and polysomnographic characteristics as well as their long-term response to treatment are described. PATIENTS AND METHODS A database of more than 6000 patients who attended the Respiratory Support and Sleep Centre at Papworth Hospital from 1994 to 2006 was reviewed to identify all patients with a diagnosis of idiopathic hypersomnia or narcolepsy. The clinical and polysomSLEEP, Vol. 30, No. 10, 2007 1275 Idiopathic hypersomnia—Anderson et al RESULTS cal history, family history (defined as EDS with onset younger than age 40 years with no other known primary sleep disorder), nap duration, dream quality, length of nighttime sleep, sleep drunkenness (defined as prolonged awakening difficulties for longer than 30 minutes with confusional behavior), drug history and recent completion of an ESS. Polysomnography was performed using a standard procedure, including video recording, a sleep electroencephalogram (leads C4-A1 and C3-A2), bilateral eye movements (contralateral electrodes placed 2 cm out from the mid outer canthus and 2 cm up on 1 side and 2 cm down on the other, with these electrodes referenced to the opposite A electrode position), submental electromyography, and bilateral anterior tibialis electromyogram to record any periodic leg movements of sleep. Respiratory effort was detected with chest and abdominal bands measuring inductance, airflow was detected with nasal cannulae measuring pressure, and oxygen saturation of arterial blood was also measured in order to exclude sleep apnea syndrome. Airflow limitation and changes in respiratory effort were used to detect increased upperairway resistance. All respiratory events were scored according to standard criteria.16 Sleep stages were scored according to standard criteria.17 Sleep latency, sleep efficiency, and stage 1, stage 2, stage 3, stage 4, and REM sleep (as a percentage of total sleep time) were recorded independently. MSLTs were recorded on the day after the polysomnogram and scored according to standard guidelines.18 SOREMPs were defined as sleep onset followed by periods of REM sleep within 15 minutes. Typing of Class 11 HLA haplotypes was available in 39 of the 77 patients with idiopathic hypersomnia and 61 of the 63 patients with narcolepsy. Clinical Characteristics of Patients with Idiopathic Hypersomnia Hypersomnia began at a mean ± SD age of 16.6 ± 9.4 years (range 0-46 years) with 49 of 77 patients developing symptoms under the age of 18 and, in 7 cases, worsening over several years. Some patients and their families were sure that the symptoms had been present from the first year of life, often in comparison with other siblings. The mean age of diagnosis was 30 years. There was no precipitant in the majority of cases, although 3 subjects reported a transient viral illness at the time of symptom onset, and 1 subject reported onset of symptoms over a day. The mean ESS score at initial presentation prior to treatment was 16.3 ± 3.3 (range 11-24), with the majority of patients having had difficulty with work and social activities as a consequence of their daytime somnolence. The mean length of nighttime sleep reported was 9.2 ± 1.8 hours, and this was not related to disease severity as measured by the ESS. There were 23 patients with nighttime sleep of 10 hours or longer, and they had a mean ESS of 16 ± 2.7; 54 patients had nighttime sleep of less than 10 hours, with a mean ESS score of 16.8 ± 3.5 (P = 0.96). There was no significant difference in MSLT between those with a long sleep time (mean MSLT of 8.9 ± 3.5) and those with a sleep time shorter than 10 hours (mean MSLT of 7.9 ± 2.6; P = 0.91). Seventy-six patients had daytime naps on most days (5 or more days a week). These lasted for longer than 60 minutes in 87% and were described as typically unrefreshing by 78% of patients. Vivid dreams occurred in 25%, but hypnagogic hallucinations or sleep paralysis were rarely seen (5% and 4%, respectively). Sleep drunkenness was found in 52% of subjects but was not related to disease severity, as measured by ESS; those with sleep drunkenness had a mean ESS score of 16.7 ± 2.9, compared with those without having a mean ESS score of 15.8 ± 3.6 (P = 0.90). Patients with normal sleep time often reported sleep drunkenness (25/54), and not all patients with long sleep time reported sleep drunkenness (16/23). There was not a significant association between prolonged nighttime sleep and sleep drunkenness (P = 0.083). Twenty-six of the 77 patients (34%) had a family history of similar problems, and, in 8 cases, more than 1 family member was affected. One patient had a family member with a diagnosis of narcolepsy, but the details of the history and investigations were not available to us. The median (interquartile range) body mass index was 25 kg/m2 (4). These clinical characteristics are summarized in Table 2. There was no relevant past medical history in 31 patients. Other medical conditions reported included hypothyroidism (n = 5, 6%), migraine (n = 6, 8%) and minor depression (n = 11, 14%). The hypothyroidism was treated in all cases but without improvement in the symptoms of hypersomnolence. Of the patients who had haplotype analysis the percentage of carrier frequency was 18% (7/39) for the HLA DQB1*0602 antigen. This was in contrast with the group with narcolepsy, in which the percentage of carrier frequency was 98% (61/62) (P < 0.001). There was a carrier frequency of 10% (4/39) for the Cw2 antigen in the patients with idiopathic hypersomnia. The follow-up in the clinic was for a mean duration of 3.4 ± 2.2 years. The patients who received drug treatment were followed up for a mean of 3.8 ± 2.1 years. Statistical Analysis Continuous measurements are summarized as the mean ± SD or the median (interquartile range) as appropriate. Categorical measurements are summarized as the number (percentage of the diagnosis group). The Kolmogorov-Smirnoff test was used to determine whether measurements were normally distributed. Of the clinical and polysomnography data, only REM sleep latency and body mass index were nonnormally distributed, and, for these, the statistical significance of the median difference between groups was established using the Mann Whitney test. For all other continuous variables, the significance of mean differences between groups was assessed using Student t test. Statistical significance refers to P < 0.05. For further analysis of differences between the idiopathic hypersomnia and narcolepsy groups, all variables were categorized as high or low according to prespecified normal ranges or commonly used thresholds. These categorical variables were tabulated against diagnosis group and 2-sided Fisher exact tests were used to establish whether associations between predictive factors and diagnosis group were significant. From these tables, the sensitivity and specificity for idiopathic hypersomnia were calculated. Sensitivity for idiopathic hypersomnia was estimated as the number of patients with idiopathic hypersomnia with the factor divided by the number of patients with idiopathic hypersomnia; specificity was the number of nonpatients with idiopathic hypersomnia who did not have the factor divided by the number of patients without idiopathic hypersomnia. SLEEP, Vol. 30, No. 10, 2007 1276 Idiopathic hypersomnia—Anderson et al Table 2—Clinical and Polysomnographic Characteristics of Patients with Idiopathic Hypersomnia in Comparison with Patients with Narcolepsy Parameter Subjects, no. Sex, male:female, no. Age of symptom onset, y Age at diagnosis, y Family historyb Sleep drunk Vivid dreams Hours of night sleep reported BMI, kg/m2 DQB1*0602 positive Initial ESS score, off treatment MSLT SOREMPs Sleep latency, min Sleep efficiency, % Slow-wave sleep, % Spontaneous improvementc Number treated Idiopathic hypersomnia 77 38:39 Narcolepsy 63 37:26 0.395 20.4 ± 9.8) 44.0 ± 17.6) 11 (18) 2 (3) 51 (81) 0.024 <0.001 0.035 <0.001 <0.001 9.2 ± 1.8 25 (4) 7 (18) 8.0 ± 0.6 27 (6) 60 (98) <0.001 0.008 <0.001 16.3 ± 3.3 8.3 ± 3.1 8 (3) 11.5 ± 8.2 94.3 ± 4.2 22.9 ± 8.7 18.6 ± 3.3 4.1 ± 2.6 173 (68) 10.2 ± 11.1 81.5 ± 13.0 17.6 ± 7.1 <0.001 <0.001 <0.001 0.415 <0.001 <0.001 0/63 (0) 62/63 (99) Factor Nap duration ≥ 60 min Sleep drunk, yes Vivid dreams, no Night sleep ≥ 9 hours REM latency ≥ 50 minutes MSLT ≥ 8 min Sleep efficiency ≥ 90% P Valuea 16.6 ± 9.4 34.3 ± 12.0) 26 (34) 42 (55) 19 (25) 11 (14) 61 (79) Table 3—Factors Helping to Distinguish Idiopathic Hypersomnia From Narcolepsya Specificity 55/63 (87) 61/63 (97) 51/63 (81) 57/62 (92) 33/63 (48) 57/61 (93) 41/63 (65) P Valueb <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 Data are presented as number of subjects with the factor/total in group (percentage). b P value was determined based on Fisher exact test from cross-tabulation of predictive factor and diagnosis group a sleep efficiency of less than 81% (see Table 2). Latency to REM sleep and percentages of light sleep and REM sleep were normal, compared with normal ranges.19 Four patients had repeat polysomnography off and then on stimulant medication. Those on stimulant medication did not show changes in sleep architecture, sleep efficiency, or sleep latency after treatment, although 3 had subjective improvement in their daytime somnolence, and all 4 had a decrease in their ESS scores. The mean ± SD sleep latency on MSLT of patients with idiopathic hypersomnia was 8.3 ± 3.1 minutes, which was significantly higher than that of the narcolepsy group (4.1 ± 2.6; P < 0.001). The mean MSLT data were normally distributed and ranged from 3.3 minutes to 13 minutes. Forty patients had a mean sleep latency on MSLT of shorter than 8 minutes and 32 had a mean sleep latency on MSLT of 8 minutes or longer. These 2 groups (MSLT ≥ 8 and MSLT < 8) were compared with each other to see if there was any difference in the other clinical or polysomnographic parameters. No statistically significant differences between these 2 MSLT groups were found in the duration of light sleep, slow-wave sleep, REM sleep, sleep latency, sleep efficiency, REM sleep latency, body mass index, or untreated ESS. Three patients repeated their MSLT after starting medication, with 2 patients having an increased MSLT on medication and a decreased ESS score, but 1 patient who had a shorter sleep latency on MSLT despite a subjective improvement in her symptoms of daytime somnolence. <0.001 <0.001 Data are presented as mean ± SD or number (percentage), except body mass index (BMI), which is presented as median (interquartile range). All data were based on total number of subjects in each group (ie, 77 and 63, respectively), except Multiple Sleep Latency Test (MSLT), which included data from 72 and 61 subjects, respectively, and sleep-onset REM periods (SOREMP), which included 308 and 254 as the denominators, respectively. a P values were derived from Fisher exact test, Mann-Whitney U test, or Student t test, as appropriate b Family history refers to a family member with onset of excessive daytime sleepiness (EDS) at age < 40 years but without symptoms of sleep apnea or restless legs syndrome. c Spontaneous improvement refers to normal Epworth Sleepiness Scale (ESS) score without symptoms of EDS for > 6 months. Polysomnography and MSLT Differences Between Idiopathic Hypersomnolence and Narcolepsy Polysomnography was performed in all patients with idiopathic hypersomnia. Seventy-three patients with idiopathic hypersomnia were studied without taking any wake-promoting or stimulant medication, but 4 were studied while on antidepressant medication. Five patients did not have an MSLT. Because an AHI of 10 was used as an exclusion criteria, the AHI data were reviewed, and 4 of 77 patients had an AHI between 5 and 10 (1 patient had an AHI of 6 and the other 3 had an AHI of 7). None of these patients had arousals associated with their hypopneas, and all 4 had their symptoms of EDS from childhood. We felt that it was appropriate to include these subjects and that sleep apnea was not a cause of their EDS. Polysomnographic analysis showed a short mean sleep latency of 11.5 minutes ± 8.2, increased mean slow wave sleep of 22.9% ± 8.7, and a high mean sleep efficiency of 94.3%. Only 10 patients had a sleep efficiency of less than 89%, and none had a SLEEP, Vol. 30, No. 10, 2007 Sensitivity 67/77 (87) 42/77 (55) 58/77 (75) 45/77 (58) 72/77 (94) 35/72 (49) 64/77 (84) The clinical and polysomnographic features of the patients with idiopathic hypersomnia were compared with the features of the patients with narcolepsy to see if there were any statistically significant differences. Comparisons are summarized in Table 2. A number of predictive factors made a diagnosis of idiopathic hypersomnia more likely, and these are summarized in Table 3; the sensitivity and specificity of the key predictive factors are also shown. Patients with idiopathic hypersomnia were significantly more likely to be sleep drunk (P < 0.001) and to have a positive family history (P = 0.035), nighttime sleep of 9 or more hours (P < 0.001), and a daytime nap duration of 60 or more minutes (P < 0.001). They were also more likely to have a body mass index of less than 30 (P = 0.015) and, without treatment, to have an ESS score of less than 20. When we reviewed the polysomnography data, we found that patients with idiopathic hypersomnia were significantly more 1277 Idiopathic hypersomnia—Anderson et al Table 4—Clinical and Polysomnographic Characteristics of Patients Who Spontaneously Improved Parameter Chronic idiopathic hypersomnia Subjects, no Sex, male:female Age of symptom onset, y Duration of symptoms, y Family history Sleep drunk Vivid dreams Hours of night sleep BMI Initial ESS score off treatment MSLT SOREMPs, no. Sleep latency, % Sleep efficiency, % Slow-wave sleep, % Number treated 66 33:33 Spontaneous improvement 11 5:6 P Valueb 1.000 17.2 ± 9.5 13.4 ± 8.6 0.216 15.9 ± 10.8 20/66 (30) 35/66 (53) 14/66 (21) 9.2 ± 1.9 25 (4) 20.6 ± 19.6 6/11 (55) 7/11 (64) 5/11 (45) 9.3 ± 1.4 26.4 (8) 0.458 0.168 0.745 0.127 0.918 0.709 16.3 ± 3.4 8.5 ± 3.1 7/264 12.1 ± 8.5 94.4 ± 4.1 22.4 ± 8.7 55/66 (80) 16.6 ± 2.5 7.2 ± 2.5 1/44 7.8 ± 4.4 94.1 ± 5.2 26.0 ± 8.3 8/11 (73) 0.803 0.205 1.000 0.103 0.831 0.204 0.689 Figure 1—Treatment effects on patients with idiopathic hypersomnia. The figure shows the number of patients treated (blue bars) and the number who improved on treatment with a drop in the Epworth Sleepiness Scale score of > 4 points (red bars). Four different treatments are shown: modafinil (MOD), dexamphetamine (DEX), modafinil and dexamphetamine (MOD/DEX), and modafinil and caffeine (MOD/CAF). Data are presented mean ± SD or as number of subjects with the factor/total in group (percentage), except for body mass index (BMI), which is presented as median (interquartile range) b P value derived from Fisher exact test, Mann-Whitney U test or Student t test, as appropriate a Of the patients who used another drug or combinations of drugs, 57% had a significant drop in ESS their score, with 6 of these patients having an ESS score less than 11. The mean reduction in ESS score for those who switched from modafinil to dexamphetamine was 7 ± 6.8. Twenty-one patients reported side effects on modafinil. The most common were transient nausea (n = 7) and headache (n = 5), often alleviated by dose reduction and more cautious dose escalation. Four patients reported mood disturbance, and 2 patients lost weight. One patient developed orofacial dyskinesia. Two patients treated with dexamphetamine reported side effects of paranoid ideation. The reduction in ESS scores achieved in the narcolepsy group after medication was 6.0 ± 4.8, but it was not possible to make any further direct comparisons because a far wider range of medication was used in this group, and many patients were treated for both narcolepsy and symptomatic cataplexy. Eleven patients within the idiopathic hypersomnia cohort spontaneously improved and became asymptomatic with normal ESS scores off all medication. The follow-up time of this group, after resolution of symptoms, was longer than 1 year in all cases. The clinical and polysomnographic characteristics of this group were compared with the patients with chronic symptoms. The group that spontaneously improved had shorter sleep latencies on MSLT, shorter sleep latency on polysomnography, and a younger age of symptom onset than the patients with persistent symptoms, but the differences were not statistically significant. The initial ESS score was not different between the groups, and those whose symptoms spontaneously improved did so both on and off medication (3 had never used any medication). There was no history of illness or a precipitating cause for the hypersomnia in those whose symptoms improved. These data are shown in Table 3. Two women improved during pregnancy and then remained well postpartum despite cessation of medication during the pregnancy. Following this result, the other women within the cohort were questioned about symptoms during pregnancy. One likely to have more than 20% of slow-wave sleep (P = 0.026), a sleep efficiency of greater than 90% (P < 0.001), and a sleep latency on MSLT of longer than 8 minutes (P < 0.001). Of the factors in Table 3, the single most useful indentifier in the clinical history to reliably distinguish idiopathic hypersomnia from narcolepsy was daytime napping of greater than an hour, which had both a sensitivity and a specificity of 87%. Treatment and Outcome The various treatments that were used and the number of subjects who responded to treatment (improvement in ESS score > 4 points) are shown in Figure 1. Sixty-one of 77 patients were treated at Papworth, with the remaining patients either lost to follow-up or not wishing treatment. Of 54 patients starting modafinil (mean dose 400 mg, range 100-1000 mg), 39 remained on this alone, 8 switched to dexamphetamine because of side effects, and 7 reported lack of efficacy, with 3 adding caffeine (50-100 mg) and 3 adding dexamphetamine. Of the 7 remaining treated patients, 2 patients used dexamphetamine then modafinil, 2 used alternative stimulants (pemoline and methylphenidate), and 3 patients were treated with sleep hygiene. Modafinil alone helped 62%, with the respondents showing a drop of more than 4 points on their ESS scores and reporting decreased daytime napping. Nineteen of these patients had a normal ESS score (<11) on medication. Two patients had transient benefit from modafinil with decreased efficacy after a few weeks, and they were counted as nonresponders. The mean reduction in ESS scores for those on modafinil alone was 6.0 ± 5.4. SLEEP, Vol. 30, No. 10, 2007 1278 Idiopathic hypersomnia—Anderson et al other patient had noted incomplete resolution of symptoms during pregnancy but recurrence postpartum. In contrast, none of the patients with narcolepsy who have attended the Sleep Disorders Clinic in Papworth have spontaneously improved or reported improvement in pregnancy. onset. Billiard and colleagues evaluated 23 patients and again emphasized Roth’s initial distinction between a polysymptomatic form with sleep drunkenness and a monosymptomatic form.11 Laffont and colleagues21 further subdivided 128 nonnarcoleptic patients with EDS into 8 categories, emphasizing variability of symptoms and age of onset. In common with other studies, we found long unrefreshing naps, positive family history, sleep drunkenness, and an insidious onset as a teenager in many cases. The limitations of the familyhistory data must be taken into account because family members could have other causes of EDS, but it is interesting to note that many patients reported other family members with longstanding EDS from childhood or from early teens, with a similar pattern of prolonged unrefreshing daytime naps. Distinguishing narcolepsy without cataplexy from idiopathic hypersomnia with normal sleep time can be difficult. With the exception of cataplexy, the single most useful factor in the clinical history that distinguished idiopathic hypersomnia from narcolepsy was nap duration of longer than 60 minutes, which had 87% sensitivity and specificity in our cohort. No other combination of variables gave a higher predictive accuracy. Further attempts to find a multivariate model led to instability due to small group numbers. Clinically, the patients with idiopathic hypersomnia virtually always described not only long, but unrefreshing, naps, in marked contrast with the brief refreshing naps of the narcolepsy group. Some previous authors, including Roth, have divided idiopathic hypersomnia into polysymptomatic patients who have long nighttime sleep and exhibit sleep drunkenness in the morning and monosymptomatic patients with EDS alone. This has led to the current diagnostic criteria that divide idiopathic hypersomnia into those with prolonged nighttime sleep who often exhibit sleep drunkenness and those with nighttime sleep between 6 and 10 hours. In our cohort, 24 patients had nighttime sleep duration of 10 hours or longer, and 53 had nighttime sleep duration of less than 10 hours, but we did not find an association between the presence or absence of sleep drunkenness and prolonged nighttime sleep. Patients with both normal and prolonged sleep exhibited symptoms of sleep drunkenness. There was also no significant correlation between severity of EDS as measured by the ESS and the presence or absence of prolonged nighttime sleep and sleep drunkenness. There was also no significant difference between the MSLT values for the 2 groups. Our group is therefore clinically heterogeneous, with symptoms of variable severity, and we did not find the distinction between those with normal and those with prolonged nighttime sleep to be an important distinction when assessing symptoms. The polysomnography findings were typically of normal amounts of light sleep and REM sleep but slightly prolonged slowwave sleep and high sleep efficiency. This correlates well with the clinical description of deep, often prolonged, nighttime sleep given by many patients. It also contrasts markedly with the nighttime sleep of the patients with narcolepsy who complain of disturbed nighttime sleep and usually have much lower sleep efficiency. Roth’s original studies also described deep and unrefreshing sleep,5 but, in contrast, some authors have reported disrupted nighttime sleep in 15% to 45% of patients.8, 22 Variability in inclusion criteria may well have affected this, and the exclusion of patients with significant depression or anxiety, factors that can disrupt night sleep, may have reduced this in our cohort in which only 10 patients had sleep efficiency below 90% and none below 81%. DISCUSSION This review of the database of patients within the sleep center at Papworth Hospital provides an opportunity to describe the clinical and polysomnographic characteristics of a cohort of patients with idiopathic hypersomnia. It also allows a study of the effects of treatment, in particular, the newer wakefulness-promoting drugs, in a group that is under long-term follow-up care. This is the largest published group of such patients in the medical literature. The Diagnosis and Characterization of Idiopathic Hypersomnia There are a relatively small number of case series in the literature to guide the current ICSD diagnostic criteria. Roth first described idiopathic hypersomnia as distinct from narcolepsy with or without cataplexy.5 He collected information from 642 patients over a 30year period who had hypersomnia with and without a secondary cause. He classified 174 as having functional or idiopathic hypersomnia. The characteristic features of his patient group included deep, unrefreshing nocturnal sleep with long unrefreshing daytime naps, and he described 2 types, with monosymptomatic patients having EDS alone but polysymptomatic patients having EDS, sleep drunkenness, and prolonged nighttime sleep. He noted a family history in a third of patients and reported that symptoms were lifelong and often responded poorly to treatment. Not all patients had polysomnography, so conditions such as sleep apnoea and periodic limb movements of sleep were not excluded. A number of retrospective case series have studied the clinical and polysomnographic features of patients with idiopathic hypersomnia, compared with those with narcolepsy. Van den Hoed and colleagues studied 100 consecutive patients with EDS not caused by sleep apnea.20 Forty-six were diagnosed with narcolepsy, whereas 17 were diagnosed with idiopathic hypersomnia. The mean sleep latency on MSLT was longer on average in the idiopathic hypersomnia group, with 14 patients having a mean latency of longer than 11 minutes. A cooperative study reviewing polysomnography from several different centers9 found idiopathic hypersomnia to be 35% as frequent as narcolepsy, with more nocturnal arousals and myoclonic jerks found in the narcoleptic group. Baker and colleagues noted higher blood pressure and more disrupted nighttime sleep in patients with narcolepsy, as compared with patients with idiopathic hypersomnia. They defined idiopathic hypersomnia in 37 patients with sleep latency on MSLT of less than 5 minutes, no SOREMP, definite EDS, and no cataplexy or hypnagogic hallucinations.10 Bassetti and Aldrich evaluated 42 patients with idiopathic hypersomnia,8 with detailed follow-up in 28 cases. They described 3 groups: those with “classic” idiopathic hypersomnia (8 cases), “narcoleptic-like” hypersomnia (9 cases), and “mixed” idiopathic hypersomnia (11 cases). Their patients had inclusion criteria similar to criteria used in previous studies with the exception of a sleep latency on MSLT of less than 10 minutes. They noted occasional patients who spontaneously improved but only in association with secondary causes of hypersomnolence such as viral illness at the time of symptom SLEEP, Vol. 30, No. 10, 2007 1279 Idiopathic hypersomnia—Anderson et al In our idiopathic hypersomnia cohortm the mean sleep latency on MSLT was significantly longer than that of our patients with narcolepsy, which has been noted in previous studies.8, 20 Many patients who had a typical clinical story for idiopathic hypersomnia had mean sleep latencies on MSLT of at least 8 minutes, which is greater than that allowed within the current ICSD diagnostic criteria. The 2 subgroups of patients (mean MSLT < 8 minutes and mean MSLT ≥ 8 minutes) were compared to see if there were any differences between them. There were no statistically significant differences in any of the other clinical or polysomnographic features. The MSLT is well established as a measure of propensity to daytime sleep but has poor correlation with subjective measures of daytime somnolence, such as the ESS.23 There is also overlap between the normal and pathologic values. At least 15% of patients with narcolepsy have an MSLT sleep latency of 8 minutes or longer,24 and up to 30% of the general population has a mean sleep latency of less than 8 minutes.25 In our patient population, the mean sleep latency was not always short, and a normal MSLT did not predict less severe symptoms or a particular response to treatment. An increased prevalence of the Cw2 antigen has been reported in patients with idiopathic hypersomnia,26 although this finding has not been replicated by other groups.27 We also did not find an increase in this haplotype in those subjects tested and did not find an increase in the HLA DQB1 *0602 haplotype beyond that expected for the background population (estimated at 25% in previous population studies in the UK). To our knowledge, no other group has attempted to replicate the data on the Cw2 antigen. atically looked at the long-term use of medication and whether or not any effect was sustained. To examine effects of treatment without confounding factors, the exclusion criteria for this study were strict, and possible causes of hypersomnia—such as medication, relevant medical history, and in particular significant psychiatric history and brain injury—were excluded. Modafinil is a nonamphetamine wakefulness-promoting medication of uncertain action that was first licensed for use in 1997 following trials in narcolepsy.15 In this case series, 24 of 39 patients had sustained benefit from modafinil over a mean follow-up period of 3.8 years. The mean dose was 400 mg, with few patients achieving symptom control at the lower doses and 18 patients requiring greater than the licensed recommended dose (400 mg). In our experience, there have been no adverse effects relating to the higher doses of the drug. Modafinil did have some effects on behavior, with a number of patients feeling aggressive and agitated and 8 of 54 patients first treated switched to dexamphetamine. Dexamphetamine was also effective in this group, who reported few side effects at an average dose of 30 mg per day. There were a number of patients who felt that they had functional benefit and decreased daytime napping with stimulant medication but did not have a significant drop in ESS scores. They still chose to remain on medication. The ESS is the most widely used questionnaire to assess EDS, but it may be that other measures of alertness may have shown benefit. The retrospective nature of this study did not allow other measures to be applied to this patient group. In total, 61% of patients (32/54) treated with some form of stimulant medication had a significant improvement in their symptoms. How Common is Idiopathic Hypersomnia? Natural History of Idiopathic Hypersomnia There have been no epidemiologic studies, and variable criteria have been used to diagnose idiopathic hypersomnia. The prevalence of the disease remains unclear. Narcolepsy is said to affect up to 0.05% to 0.1% of American and European populations.28 Over time, the ratio of idiopathic hypersomnia to those patients with narcolepsy appears to have decreased, with the most recent studies showing the condition to be 10% to 15% as common as narcolepsy when studying referrals to sleep centers.8, 27 The referral base for the Papworth Sleep Centre is both regional and national, and it is not possible to make any estimate of prevalence based on our data. However, it is possible to compare the numbers of patients with narcolepsy with the number of those with idiopathic hypersomnia. Within our cohort, idiopathic hypersomnia occurred approximately 60% as often as narcolepsy (77/126 patients). Restricting the mean sleep latency on MSLT to less than 8 minutes would still make idiopathic hypersomnia occur 40% as frequently as narcolepsy in our series. This suggests that idiopathic hypersomnia may have been underdiagnosed in other reports of the condition. There are a number of possible reasons for this, including differing prevalence across different populations, differing diagnostic criteria in different case series, and an increase in referrals to sleep units following the introduction of newer safer treatments of hypersomnia. Those patients whose symptoms spontaneously improved were interesting for a number of reasons. Idiopathic hypersomnia was initially described as chronic and unremitting5; however, more recent reports have suggested that some patients’ symptoms spontaneously improve.8 This is the first case series to look at improvers in detail. The 11 patients whose symptoms improved were younger, with shorter sleep latencies on polysomnography and MSLT, but none of the differences reached statistical significance. This may be because insufficient numbers were studied. Patients improved both on and off medication, and the group included patients who had never been treated, making a long-term effect of drug treatment on the condition unlikely. None of the improvers had any clear precipitant for their hypersomnia, as reported in the Bassetti and Aldrich group,8 and many patients had symptoms for several years without variation prior to improvement. Two reported postpartum improvement, despite the obvious sleep disturbance this entails, raising the possibility of a hormone effect. Whether there is a definite association remains to be studied in more detail. In contrast, none of the patients with narcolepsy spontaneously improved once their symptoms were established, and remission has not been reported in the literature. Remission has been reported over time in some other conditions causing EDS, such as Kleine-Levin syndrome, but the reasons for this also remain unclear. Idiopathic hypersomnia deserves further study to try and identify factors that might lead to a remission. These data suggest a better prognosis than previously thought for idiopathic hypersomnia and might alter the information given to patients at diagnosis. Treatment of Idiopathic Hypersomnia The only case series in the literature studied 18 patients with idiopathic hypersomnia, of whom 15 took modafinil for 2 months after ESS scores were recorded. There have been no placebo-controlled studies of any stimulant and no studies that have systemSLEEP, Vol. 30, No. 10, 2007 1280 Idiopathic hypersomnia—Anderson et al CONCLUSIONS 18. Carskadon MA, Dement WC, Mitler MM et al. Guidelines for the multiple sleep latency test (MSLT): a standard measure of sleepiness. Sleep. 1986;9:519-524 19. Ohayon MM, Carskadon MA, Guilleminault C, Vitiello MV. Metaanalysis of quantitative sleep parameters from childhood to old age in healthy individuals: developing normative sleep values across the human lifespan. Sleep. 2004;27:1255-1273 20. van den Hoed J, Kraemer H, Guilleminault C et al. Disorders of excessive daytime somnolence: polygraphic and clinical data for 100 patients. Sleep. 1981;4:23-37 21. Laffont F, Mallet A, Mayer G et al. Study of a patient population investigated for excessive daytime sleepiness (EDS). Neurophysiol Clin. 2002;32:343-351 22. Lesperance P, Lapierre O, Gosselin A, Montplaisir J. Effect of exogenous melatonin on excessive daytime somnolence associated with disturbed sleep: an open clinical trial. Sleep Res. 1997;26:111-115 23. Chervin RD, Aldrich MS, Pickett R, Guilleminault C. Comparison of the results of the Epworth Sleepiness Scale and the Multiple Sleep Latency Test. J Psychosom Res. 1997;42:145-155 24. Folkerts M, Rosenthal L, Roehrs T et al. The reliability of the diagnostic features in patients with narcolepsy. Biol Psychiatry. 1996;40:208-214 25. Mignot E, Lin L, Finn L et al. Correlates of sleep-onset REM periods during the Multiple Sleep Latency Test in community adults. Brain. 2006;129:1609-1623 26. Poirier G, Montplaisir J, Decary F et al. HLA antigens in narcolepsy and idiopathic central nervous system hypersomnolence. Sleep. 1986;9:153-158 27. Billiard M, Dauvilliers Y. Idiopathic Hypersomnia. Sleep Med Rev. 2001;5:349-358 28. Hublin C, Kaprio J, Partinen M et al. The prevalence of narcolepsy: an epidemiological study of the Finnish Twin Cohort. Ann Neurol. 1994;35:709-716 We have described a large series of patients with idiopathic hypersomnia and have found a more heterogeneous group than has previously been reported. Idiopathic hypersomnia remains a diagnosis of exclusion, with polysomnography needed to exclude other causes of hypersomnia. Unlike narcolepsy, it is characterized by long unrefreshing naps and high sleep efficiency. The more prolonged mean sleep latency on MSLT raises doubts about the place of the MSLT within the diagnostic criteria. The response to modafinil and dexamphetamine was good in 61% of our patients, with a substantial reduction in ESS, although a number of patients required high doses of medication or drug combinations. A substantial minority spontaneously improved even after many years of symptoms, so that the prognosis may be better than previously thought. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. Shneerson JM. Excessive daytime somnolence. Sleep Medicine: A Guide to Sleep and its Disorders. 2nd ed. London: Blackwell Publishing, 2005:125-160 Gelineau JBE. De la narcolepsy. Gaz Hop Par. 1880;53:626-628, 635-637 Adie WJ. Idiopathic narcolepsy: a disease sui generis; with remarks on the mechanism of sleep. Brain. 1926;49:257-306 Vogel G. Studies in psychophysiology of dreams. III. The dream of narcolepsy. Arch Gen Psychiatry. 1960;3:421-428 Roth B. Narcolepsy and hypersomnia: review and classification of 642 personally observed cases. Schweiz Arch Neurol Neurochir Psychiatr. 1976;119:31-41 Roth B. Functional hypersomnia. In: Guilleminault C, Dement W, Passouant P, eds. Narcolepsy. New York: Spectrum Publication, 1976:333-349 Aldrich M. The clinical spectrum of narcolepsy and idiopathic hypersomnia. Neurology. 1996;46:393-401 Bassetti C, Aldrich MS. Idiopathic hypersomnia. A series of 42 patients. Brain. 1997;120 ( Pt 8):1423-1435 Coleman RM, Roffwarg HP, Kennedy SJ et al. Sleep-wake disorders based on a polysomnographic diagnosis. A national cooperative study. Jama. 1982;247:997-1003 Baker TL, Guilleminault C, Nino-Murcia G, Dement WC. Comparative polysomnographic study of narcolepsy and idiopathic central nervous system hypersomnia. Sleep. 1986;9:232-242 Billiard M, Merle C, Carlander B et al. Idiopathic hypersomnia. Psychiatry Clin Neurosci. 1998;52:125-129 Komada Y, Inoue Y, Mukai J et al. Difference in the characteristics of subjective and objective sleepiness between narcolepsy and essential hypersomnia. Psychiatry Clin Neurosci. 2005;59:194-199 The International Classification of Sleep Disorders, 2nd ed: Diagnostic and coding manual. 2nd ed. Westchester, Illinois: American Academy of Sleep Medicine, 2005 Johns MW. Sleepiness in different situations measured by the Epworth Sleepiness Scale. Sleep. 1994;17:703-710 Broughton RJ, Fleming JA, George CF et al. Randomized, doubleblind, placebo-controlled crossover trial of modafinil in the treatment of excessive daytime sleepiness in narcolepsy. Neurology. 1997;49:444-451 Sleep-Related Breathing Disorders in Adults: Recommendations for syndrome definition and measurement techniques in clinical research,The report of the American Academy od Sleep Medicine Task Force. Sleep. 1999;22;667-689 Rechstaffen A, Kales A. A manual of standardised terminology, techniques and scoring system for sleep stages in human subjects. 204 vol. Washington: Government Printing Office, 1968 SLEEP, Vol. 30, No. 10, 2007 1281 Idiopathic hypersomnia—Anderson et al