Abstract
Objectives: To highlight the occurrence of familial cases and addresses, whether familial Kleine-Levine syndrome (KLS) presents the same spectrum of disease, as that seen in sporadic KLS.
Methods: Between September and December 2014, reports of familial cases of KLS were identified by searching the Library of Congress, PubMed, and Web of Science databases restricted to the English language, with no restriction on date of publication. All cases were reviewed to identify familial cases consistent with current diagnostic criteria for sporadic KLS.
Results: Six reviews and 11 case reports describing cases of familial KLS were identified. In 17 of the 29 familial cases identified, sufficient clinical details were described to be confident that these cases were familial and consistent with the description of KLS in the International Classification of Sleep Disorders 3rd edition (ICSD-3), and recent detailed reviews of sporadic KLS.
Conclusion: A significant number of familial cases of KLS have been described that are consistent with the ICSD-3 description of KLS, and indistinguishable from sporadic KLS. This suggests that study of familial KLS using modern genetic techniques may be useful in elucidating the pathogenesis of this rare condition.
Kleine-Levin syndrome (KLS) is rare, but is still the most common form of recurrent hypersomnia with a prevalence of 1-2 per million.1 The International Classification of Sleep Disorders 3rd edition (ICSD-3)1 describes KLS as being characterized by “relapsing-remitting episodes of severe hypersomnolence in association with cognitive, psychiatric, and behavioral disturbances”. The ICSD-3 diagnostic criteria is shown in Table 1. Kleine first described 2 cases of recurrent hypersomnia, hyperphagia, and cognitive disturbance in adolescent boys in 1925,2 Lewis in 1926,3 and then Levin in 19294 and 19365 described further similar cases. Critchley and Hoffman6 described 2 cases in naval personnel in 1942, and first suggested the eponymous name for the disease: Kleine-Levin syndrome. Critchley later described 9 further cases,7 describing the onset during adolescence and the trend to spontaneous remission, excluding female cases in this series. Most importantly, as in Kleine’s original series, female cases are clearly described, and were no longer excluded from the diagnosis. Further clarity with regard to diagnosis has been provided by a case series described by Arnulf et al,8 and the reviews of Miglis and Guilleminault9 and Billiard et al.10 Despite this progress with definition and diagnosis, the etiology of KLS remains unclear. A number of possibilities have been considered including hypothalamic damage, autoimmunity,11 and infectious triggers.12 Familial cases with a preceding history of trauma or brain injury have been described.13 However, such a history and/or hypothalamic-pituitary endocrine dysfunction is not consistently found.10 Kleine-Levin syndrome has been associated with human leukocyte antigen (HLA) DR2,14 and HLA-DQB1*0211,15 but more extensive studies have failed to confirm HLA disease’ associations.16 Sporadic and familial cases describing an infectious prodrome have been previously described.17,18 However, no single causative agent has been identified, and it is possible that the initial KLS attack mimics an infectious illness, rather than being associated with infection as recognized by Katz and Ropper.17 The description of familial KLS suggests a genetic component, either conferring susceptibility, or directly involved in pathogenesis. Investigation of such families may identify genetic loci associated with KLS. This study will review familial cases with careful consideration of diagnostic criteria, to clarify if familial KLS present the same clinical features as sporadic KLS. This is essential to validate extrapolation of findings in familial KLS to the pathophysiology of KLS in general.
The Third International Classification of Sleep Disorders Diagnostic (ICSD-3) criteria for Kleine-Levin syndrome.
Methods
Literature review
The Library of Congress, PubMed, and Web of Science databases were searched from September to December 2014, restricted to the English language, using the terms ‘Kleine Levin’, ‘Kleine-Levin’, ‘Kleine-Levin syndrome’, ‘periodic hypersomnia’, ‘Familial Kleine-Levin’, ‘Familial Kleine-Levin syndrome’, ‘Twins Kleine-Levin’, and ‘Twins Kleine-Levin syndrome’. As much of the KLS literature is more than 10 years old, no restriction by date of publication was imposed. This identified 326 articles. Publications referenced in these articles were also reviewed; all publications describing cases with affected family members, or making reference to such cases were selected. Seven reviews (including one book chapter) and 11 case reports documenting familial cases were found.
The diagnosis of familial KLS was reviewed according to the diagnostic criteria described by ICSD-3.1 Case reports or series with insufficient data to confirm the diagnosis were excluded. In their case control study, Arnulf et al16 were fastidious in their review of the clinical characteristics. They included 2 of the early reported familial cases,13,17 and we therefore included these cases here. In keeping with the ICSD-3 classification and approach taken by Arnulf et al16 and Billiard et al,10 we have not excluded female cases, and did not consider hyperphagia an absolute pre-requisite for the diagnosis. From the 11 reports describing KLS with affected family members, we have excluded 2 reports. One case cited in the literature as being familial has been excluded as in the original report the affected uncle was described only as having “committed suicide due to asocial behavior”.19 The second report is available in abstract form only, describing 9 family members with KLS over 3 generations.20 Although the case histories strongly suggest a diagnosis of KLS, the clinical features of each individual case are not described in detail. This series has not therefore been considered further here.
A final report where the affected sib had menstrual related hypersomnia (MRH)21 has been included and is worthy of comment. It is of note that ICSD-3 reclassifies MRH as menstrual related KLS. The purpose of this review is to highlight the study of familial cases in searching for loci contributing to the pathogenesis of KLS, whilst emphasizing important aspects of such analyses. Studies in autism have emphasized that closely related, but clearly distinct manifestations of a syndrome may have distinct genetic associations. It is therefore important to recognize that, although now clearly described as part of the KLS spectrum, menstrual related KLS may have unique factors contributing to its pathogenesis. Study of such cases using modern genetic techniques may be very informative, and DNA should be collected from such cases. However, in analysis of such cases the possibility of unique genetic contributions should be considered. From the remaining 8 reports describing familial cases we were able to confidently identify 17 cases. Four reports describe sib pairs,17,18,22,23 including 2 pairs of monozygotic (MZ) twins.18,23 In a further report, 6 affected members of a Saudi family were described,15 5 of whom had classical KLS with one family member with otherwise classical disease episodes, but that were restricted to the puerperium. It is noted that Billiard et al10 consider episodes of hypersomnia restricted to the puerperium to be more consistent with MRH than KLS. In the 3 remaining reports,13,24,25 single cases were described and a family history noted. Although the affected family members were not described in detail, sufficient clinical information was recorded to be confident of the diagnosis of KLS in these additional family members. These reports identified 5 additional family members in total. Thus, we have been able to consider 17 reported cases (including one with disease episodes restricted to the puerperium) with 5 additional family members noted also.
Data collection
A recent review10 commented on the variability in reporting of clinical features of KLS. This is true also of familial cases of KLS with no consistent dataset being reported. The following data have been collected where available: gender; age at onset; family history; medical and psychiatric history (including birth history); HLA typing; number and frequency of episodes; duration of KLS episodes; disease duration, presence and nature of potential precipitating events at KLS onset; events precipitating disease episodes; presence of hypersomnia; eating behavior disorder; sexual behavior disorder; cognitive impairment; hallucinations, delusional states; derealization; mood disorders and irritability; and mood and cognitive impairment between episodes. Finally, the age at remission, and whether this was spontaneous, or drug-related have been recorded.
Results
The publication of the ICSD-31 and recent studies16 and reviews of KLS9,10,26,27 have set out clear diagnostic criteria. Billiard et al10 documented 9 cases of familial KLS in their review of 339 cases of recurrent hypersomnia, 239 of which had classical KLS (3.7%). In their series, Arnulf et al16 commented on 5/104 (4.8%) being familial cases, although does not detail individual cases. This series includes 2 cases reported by Billiard et al.10 They commented also on the cases of Janicki et al13 and Katz and Ropper.17 Dauvilliers et al11 documented one of 25 cases as having an affected family member. As clinical details specifically of this case are not provided, it has not been considered further here. The case described by Thacore et al19 has been cited as a case of familial KLS. However, the affected relative (uncle) is described only as having ‘attempted suicide after asocial acts’, and this is not therefore considered here as a familial case. Rocamora et al21 described a brother that has KLS, and a sister considered to have MRH. Although clustering of related sleep disorders within a family is of interest, this report has not been considered as a confirmed report of familial KLS, as discussed above. We therefore considered 17 cases of KLS described in the literature as being familial KLS (Table 2). These cases conform to the diagnostic criteria for ‘sporadic’ KLS. However, the diagnostic criteria as set out by ICSD-3 are broad (for example, absence of hyperphagia does not preclude the diagnosis), and the time frame specified in ICSD-2 has been challenged,10 a similar criterion being retained in ICSD-3. Specific consideration will therefore be given to whether the familial cases identified conform also with the description of sporadic KLS set out in the case controlled study of Arnulf et al,16 and reviews of Arnulf et al,27 Billiard et al,10 and Miglis and Guilleminault.9 The clinical features for these cases have been summarized in Table 3.
Summary of cases of Kleine-Levin syndrome (KLS) described in the literature as being familial KLS.
Clinical features reported for cases of familial Kleine-Levin syndrome. In reports where multiple cases are presented. Data for each case are shown separately.
In one of the first reports of familial KLS, Suwa and Toru24 described a 29-year-old Japanese male whose paternal grandmother, father, and older sister had episodes of hypersomnia consistent with KLS. In the grandmother and father, physical and mental fatigue was considered to be triggers, with disease episodes in the sister being preceded by fever. The propositus is described as having episodes of hypersomnia typical for KLS, often preceded by mental and physical exhaustion. A craving for sweet foods preceded disease onset, and many individual disease episodes, and episodes were associated with depression and altered body perception. Intravenous administration of glucose induced ‘light sleep’ confirmed by EEG. Janicki et al13 described an 18-year-old female patient who experienced a 3-week period of hypersomnia following a minor head injury and consumption of vodka. Mood change and a craving for ‘junk food’ are described during this episode. A subsequent relapsing remitting course, consistent with a diagnosis of KLS is described with a further episode possibly precipitated by alcohol. The family history included one male cousin who developed KLS at the age of 10 years. One paternal aunt had narcolepsy.
In 2002 Katz and Ropper17 described a brother and sister with KLS, with a 5-month separation in onset, these being the first familial cases where both affected family members presented for clinical assessment by the authors. The first case is of a 15-year-old boy with a preceding flu like illness without fever. Mechanical eating of salty foods is described during attacks when he is noted to be irritable and withdrawn. During episodes, sensations were described as feeling bizarre, unpleasant, or wrong. Eight episodes are described in 10 months with duration of 7-12 days. Over the next 5 years, disease episodes became less frequent but lasted up to 81 days, with 16 episodes of hypersomnolence over 5 years in total. Episodes were preceded on occasion by stress or slight alcohol intake, but these were not universally implicated. The sister was described as having a disease onset age of 13 years, also with a preceding flu like illness. The mental state described was for her brother, with a persistent sense of unreality and disconnection. Episode duration was initially typical for KLS, but similar to her brother, increased in duration in up to 72 days with 18 spells over 5 years. Both sibs were normal between attacks. No physical abnormalities were described including normal brain imaging. In the brother, during a disease episode, lumbar puncture, positron emission CT, and cerebrospinal fluid hypocretin were normal. Both sibs expressed DR2 and DQ1, and had no response to medication. The authors17 discuss the possibility of a psychiatric diagnosis with the second sib imitating the symptoms of the first sib, although they believed this to be unlikely. Although the length of the later episodes for both sibs was atypical, this does not preclude a diagnosis of KLS. Most importantly, with regard to the ‘infectious’ prodrome, these authors17 comment that these symptoms may be part of KLS, rather than suggesting an infectious trigger.
Bonkalo22 also describes a brother and sister pair. The sister suffered 8 sleep attacks between the ages of 16 and 21 years lasting 3-11 days. She complained of nausea and occasional vomiting, and her food intake was described as ‘poor’ to ‘good’. Her brother had trouble waking from adolescence with a more protracted episode of daytime hypersomnia at the age of 21 years, and an episode of sleeping for 3 days later the same year. He is reported also as being ‘always hungry’ although the temporal relationship to the hypersomnia is not known. Although the brother’s history is not absolutely typical of KLS, Billiard et al26 have reviewed this case, accepting the author’s suggestion that this was a ‘forme fruste’ of KLS. Poppe et al25 describe a boy presenting shortly after onset at the age of 15 years, whose maternal uncle was reported to have had KLS in adolescence. Typical symptoms are described with megaphagia and an infectious prodrome, disease episodes lasting 8-9 days. There was a possible reduction in frequency with lithium treatment, however, compliance was poor.
Two pairs of MZ twins with KLS have been described. Peraita-Adrados et al23 described male MZ twins who were 33 years of age at presentation. The first-born twin reported a first episode of hypersomnolence at age 16 years, and a last episode age 29 years. These were typical KLS disease episodes, except for decreased eating. They lasted a mean of 15 days, progressively shortening, with a cycle length of 40 days. The second-born twin experienced a first disease episode at age 17 years after the break up of his first love affair. The last disease episode was at age 31 years. This twin described compulsive eating and sexual disinhibition. Disease episodes for the second twin had a mean duration of 2 weeks with a cycle length of 40 days lengthening to 180 days. Neither twin benefited from lithium. The twins expressed DQB1*0601 as opposed to DQB1*0201, the latter being the DQB allele suggested to be over-represented in KLS.11 Ueno et al18 described also male MZ twins with KLS. The first twin presented at age 15 years with the first disease episode at 13 years of age, one month after treatment for influenza with zanamivir. The episodes lasted 7-10 days occurring monthly. This twin described reduced eating with attacks often preceded by flu-like symptoms. Twenty-three episodes had been recorded to the time of the report, with a reduction in frequency with lithium. The second twin described a similar pattern of disease with onset at age 14 years, the first episode again being preceded by influenza treated with zanamivir. These twins expressed DQB1*0302/0601. The largest family, in which the authors15 studied all affected members, is a multiplex Saudi family with 6 of 12 individuals affected. Diagnoses were carried out according to ICSD-2 criteria, and are consistent with the ICSD-3 criteria. A mean age of onset of 16.7 years, and mean episode duration of 10.7 days were recorded. One female family member experienced disease episodes only during the puerperium. This family was consanguineous with homozygosity for HLA DQB1*02 in 4/6 affected members, and only 2/6 unaffected. However, all but one family member (affected and unaffected) expressed at least one copy of the suggested susceptibility haplotype. Laboratory findings for these cases have not been consistently reported. Data available have been summarized in Table 4.
Results of polysomnography, cranial imaging (CI), electroencephalography studies (EEG), and laboratory investigations reported for cases of familial Kleine-Levin syndrome as found in different studies.
Discussion
Billiard et al26 stated that published data support a ‘strongly genetic basis’ for KLS. Elucidation of the nature of this genetic basis may be undertaken by study of sporadic cases. However, a study of familial cases may be a particularly powerful approach. As emphasized by Moghadam and Plazzi28 when reviewing the genetic basis of autosomal dominant nocturnal frontal lobe epilepsy (ADNFLE), in order to extrapolate findings in familial cases to sporadic KLS, it is essential to confirm as far as possible, that sporadic and familial cases represent the same condition. The purpose of this article is to review published cases of familial KLS to address whether familial cases have been described that present a typical picture of KLS. Similarly, careful clinical diagnosis and differentiation of subgroups is an essential pre-requisite for genetic studies.
Arnulf et al8 reviewed 186 published KLS cases, and subsequently carried out a case control study of 108 cases.16 They found 75.9% of cases to be adolescent boys with an overall mean age of onset of 15.7±6 years (range: 6-59 years) with 81.7% of cases having disease onset in the second decade.16 Billiard et al10 documented a male:female ratio of 4.08 in classical KLS, with disease onset before the age of 10 years in 3.7% of cases, and 6.7% after the age of 30 years. In keeping with this, 12 of the 17 cases discussed here were males with 15/17 having onset in the second decade (Table 2). The occurrence of 5 female cases presenting with classical features emphasizes the importance of not excluding female patients from this diagnosis. Billiard et al10 described episode lengths of 1-180 days in men, and 1-60 days in women with significant intra-individual variability. These authors10 question the validity of the ICSD-2 criterion “recurrence within one year” as they describe cases with otherwise classical KLS with cycle lengths of up to 1095 days in men, and 1460 days in women. Although modified to recurrence within 18 months in ICSD-3, these cases would still lie outside these criteria, despite being typical in all other respects. A decrease in frequency was described in 25.9% of cases with KLS, as seen in some familial cases. The familial cases reviewed in this present study fit comfortably within the spectrum defined in sporadic cases.
Arnulf et al16 describe 89% of patients as remembering a preceding event, 72% reporting an infectious prodrome, alcohol use in 23%, sleep deprivation in 22%, unusual stress in 20%, or physical exertion in 19%, and head trauma in 9%. Billiard et al10 describe a significant event at onset in 66.5% of cases with a history suggesting an infectious illness being most frequent. Again the familial cases described here conform to this description.
Billiard et al10 describe hyperphagia in 66%, or increased food intake in 56% of cases, but decreased appetite in 34%. In keeping with this the ICSD-3 classification1 describes altered alimentary behavior, rather than hyperphagia. Thus, the description of cases with reduced appetite by Peraita-Adrados et al,23 Ueno et al,18 and BaHammam et al15 does not preclude consideration of these cases in this review of familial KLS. Arnulf et al16 describe automatic eating in 37% of cases as described for some cases considered here.
The cases reported here accurately reflect the spectrum of cognitive impairment, altered perception, and psychological change reported in the published literature on sporadic KLS. Arnulf et al16 describe a continuously altered dreamlike perception, as described by one of the familial cases reviewed here, as having a 100% sensitivity for KLS. They describe also the occurrence of a brief ‘overshoot’ of insomnia, as described in some familial cases. Restoration of normal function between episodes is considered a prerequisite for the diagnosis, and was reported in all cases considered here. Altered sexual drive was considered part of KLS in the early literature. Only 3 of the 17 patients described here were documented as reporting alteration in sexual behavior. However, this is not a prerequisite for the diagnosis based on ICSD-3 criteria,1 and Arnulf et al16 found altered sexual behavior in only 59% of cases, when enquiring specifically regarding this criterion. Billiard et al10 described the 4 classical behavioral ‘components’ of KLS (hypersomnia, compulsive eating, sexual disinhibition, and odd behavior) to have all occurred during at least one episode in only 14.2% of patients with the first 3 in 30.5% of cases described. A number of authors15,16 have studied HLA associations in KLS. In their case control study of 104 patients, Arnulf et al16 found no evidence of preferential transmission of the DRB1*0301-DQB1*0201 and DRB1*0701-DQB1*0202 alleles described as being associated with KLS in a previous smaller study. As discussed above, a multiplex Saudi family has been described, in which 4/6 affected individuals were homozygous for DQB1*02 with only 2/6 unaffected individuals being homozygous.15 However, homozygosity is not required for HLA alleles to confer susceptibility to autoimmunity, this finding potentially suggesting a linked non-HLA gene to be important.
The main limitations of this study are the variable reporting of clinical features and small number of cases identified. However, inclusion of only studies where sufficient information was presented to be confident of the diagnosis by comparison with international criteria and detailed reviews of sporadic KLS ensures that problems associated with variable recording of clinical features are avoided. Although the number of cases documented are small, the familial cases of KLS described here demonstrate the same pattern of disease as sporadic cases as suggested by Billiard et al.10 As long as diagnostic criteria set out here are strictly adhered to, study of familial cases, may therefore shed light also on the pathogenesis of sporadic KLS.
There is an 800-4000 fold increased risk of developing KLS in first degree relatives of affected individuals.27 Shared environmental effects cannot be excluded including common exposure to an infectious trigger. However, even when sibs have both described a preceding infectious illness, the significant difference in time of onset argues against an infectious trigger in these cases. Furthermore, extensive search for infectious triggers in sporadic KLS has not identified a causative agent. The suggestion of a genetic component for KLS is supported by the increased incidence of KLS in the Jewish population,16,26 possibly due to Jewish specific polymorphisms, or enriched polymorphisms within this population. The description of familial cases reviewed here, including a consanguineous Saudi family with 6 affected individuals, makes enrichment for a polymorphism found widely more likely.
Taken together, these findings argue that genetic analysis of cohorts of individuals diagnosed according to criteria discussed in this review may help to further our understanding of the pathogenesis of KLS. Family studies are appropriate and may facilitate such analyses.
The purpose of this review is to identify cases that could confidently form the basis for genetic analysis of familial KLS. This does not suggest that cases not included here do not represent KLS. More in that it is essential that initial genetic studies focus on cases where diagnostic confidence is highest. Given that many of the cases described in the literature are not available for such studies, it is hoped that this review will set a standard to be considered when further familial cases are identified and considered for genetic analysis.
Footnotes
Disclosure. Author has no conflict of interests, and the work was not supported or funded by any drug company.
- Received July 7, 2015.
- Accepted October 7, 2015.
- Copyright: © Saudi Medical Journal
This is an open-access article distributed under the terms of the Creative Commons Attribution-Noncommercial-Share Alike 3.0 Unported, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.