Download Idiopathic Hypersomnia: A Study of 77 Cases

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