Narcolepsy with cataplexy: Does age at diagnosis change the clinical picture?

Abstract Objective To compare symptoms and sleep characteristics in patients diagnosed with narcolepsy‐cataplexy (NC) before and after the age of 18 years. Methods De novo patients with NC diagnosis completed a standardized questionnaire and interview, followed by a sleep study. The clinical and sleep measures were compared between patients diagnosed before (46 children, median age: 12 year old) and after (46 adults, median age: 28.5 year old) 18 years of age. Results The frequency of obesity (54% vs 17%), night eating (29% vs 7%), parasomnia (89% vs 43%), sleep talking (80% vs 34%), and sleep drunkenness (69% vs 24%) were higher in children than in adults, the frequency of sleep paralysis was lower (20% vs 55%) but the frequency of cataplexy and the severity of sleepiness were not different. Children scored higher than adults at the attention‐deficit/hyperactivity disorder (ADHD) scale. Depressive feelings affected not differently children (24%) and adults (32%). However, adults had lower quality of life than children. There was no difference between groups for insomnia and fatigue scores. Quality of life was essentially impacted by depressive feelings in both children and adults. Obstructive apnea‐hypopnea index (OAHI) was lower in children with higher mean and minimal oxygen saturation than in adults. No between‐group differences were found at the multiple sleep latency test. The body mass index (z‐score) was correlated with OAHI (r = .32). Conclusion At time of NC diagnosis, children have more frequent obesity, night eating, parasomnia, sleep talking, drunkenness, and ADHD symptoms than adults, even if sleepiness and cataplexy do not differ. These differences should be considered to ensure a prompt diagnosis.


| INTRODUC TI ON
Narcolepsy with cataplexy (NC) is a rare neurological disorder characterized by excessive daytime sleepiness (EDS) with irresistible sleep attacks and cataplexy (sudden loss of muscle tone triggered by emotions), associated sometimes with other abnormal rapid eye movement (REM) manifestations such as hypnagogic hallucinations, sleep paralysis, REM sleep behavior disorder (RBD), and disturbed nocturnal sleep. 1 The frequent comorbidities of NC include obesity, restless legs syndrome (RLS), periodic limb movement syndrome (PLMS), depressive mood, and symptoms of attention-deficit/hyperactivity disorder (ADHD). The prevalence of narcolepsy ranges from 0.02% to 0.05% in European and North American populations. 2,3 The NC prevalence increased after the H1N1 infection and vaccine. 4,5 Narcolepsy with cataplexy is caused by a loss of hypocretin-1 neurons located in the lateral hypothalamus, 6 probably via an autoimmune mechanism. 7,8 Indeed, low cerebrospinal fluid (CSF) hypocretin-1 levels (less than or equal to 110 pg/mL) are found in more than 90% of patients with NC (also called narcolepsy type 1 or hypocretin deficient). In opposite, in narcolepsy type 2, there is no decrease in hypocretin levels. Additionally, the human leukocyte antigen (HLA) DQB1*0602 genotype is closely associated with NC. 9 The question of age at onset and age at diagnosis is important in the context of NC. Despite more than half of patients with NC have a disease onset prior to the age of 18 years, 10 there is a delay of more than 10 years between the disease onset and its diagnosis. 11 Indeed, many people with an NC onset during childhood or adolescence are diagnosed only when they are adults, after they completed and possibly failed their academic studies, gained an almost irreversible obesity and experienced low self-esteem. There may be several reasons for this delay, including the rarity of NC, reduced awareness about this disorder among clinicians, but also changes in the clinical picture of NC depending on age at diagnosis. Indeed, Aran et al reported that children with postpubertal onset of narcolepsy had a higher prevalence of sleep paralysis and hypnagogic hallucinations, and shorter mean sleep latency (MSL) than children with peripubertal and prepubertal onset. 12 In large groups, there was a decrease in the number of sleep onset rapid eye movement periods (SOREMPs) and an increase of MSL with age gain, with no difference across age in terms of clinical complaints. 10,13 Comorbidities may also vary as a result of age in NC. Obesity in NC affects more than 50% of children 14,15 and only 30% of adults. 16,17 Aran et al 12  group compared to patients younger than 20 years, including obstructive sleep apnea (OSA) (53% vs 0%), RBD (40% vs 32%), PLMs (34% vs 16%), and RLS (21% vs 0%). The authors noticed that cataplexy was more frequent in the adults (80%) than in children and adolescents (52%). 18 Because these results were obtained in mixed groups (among these patients, 83% had NC and 17% had narcolepsy without cataplexy, which is a different form of narcolepsy), and could be secondary to treatments (76% were treated with stimulants or anti-cataplexy medications at time of study) which could promote RBD, RLS, PLMs, and weight changes, there is a need for studying de novo, untreated patients in order to obtain a genuine picture.
There is a need for studying de novo, untreated patients in order to obtain a genuine picture. Therefore, we aimed at systematically characterizing the symptoms, signs, HLA positivity and sleep measures in nighttime polysomnography and multiple sleep latency tests (MSLT) at time of diagnosis in de novo (untreated) patients, and at comparing these measures according to the age at diagnosis (≥18 years or <18 years). In theory, it would be more interesting to evaluate patients at disease onset before and after the age of 18 years but patients often did not clearly remember the onset time of the first symptoms or did not have a complete investigation at this time. The aim of this study was to identify differences which could drive the attention of clinicians towards an earlier recognition of NC in children at the moment of the diagnosis.

| Patients
In this retrospective study, all children (age < 18 years old) and adults (age ≥ 18 years old) presenting in the 4 national reference centers for narcolepsy with newly diagnosed NC between 2008 and 2011 were proposed to take part in the research program NARCOBANK (PHRC AOM07-138, principal investigator Isabelle Arnulf). Most patients multiple sleep latency test. The body mass index (z-score) was correlated with OAHI (r = .32). At time of NC diagnosis, children have more frequent obesity, night eating, parasomnia, sleep talking, drunkenness, and ADHD symptoms than adults, even if sleepiness and cataplexy do not differ. These differences should be considered to ensure a prompt diagnosis.

K E Y W O R D S
cataplexy, children, Narcolepsy, obesity, sleep (99.6%) agreed to take part in the program, which included a systematic interview of patients (and parents) with the physician in charge (ML, PF, YD, IA) and a collection of the results of sleep study and HLA genotyping. Both the adult patients or pediatric patients and their parents signed written consent forms. This study was approved by the ethics committee (Comité de Protection des Personnes Ile de France -6).
The cohort included 46 de novo patients diagnosed with NC before 18 years and 46 patients diagnosed with NC after 18 years. All the patients had clear-cut cataplexy evaluated by sleep specialists.
No patient was treated at time of diagnostic procedure.

| Questionnaires
The severity of daytime sleepiness was evaluated in adults using the score at the Epworth sleepiness scale (ESS, range 0-24) 19 and in children using the score at the adapted Epworth sleepiness score (AESS), in which the item "falling asleep while in a car stopped in the traffic" was replaced to "falling asleep at school". 20 Scores greater than 10 were considered abnormal.
The severity of cataplexy was evaluated using the Cataplexy Severity Rating Score (1 = moderate weakness, for example, head drop or jaw opening; 2 = can maintain posture with external support; 3 = loses posture and falls to the ground). 21 However, a patient can have different types of cataplexy (type 1, 2, 3). The frequency of cataplexy attacks was also evaluated using a Likert scale (0: less than one episode per year; 1: more than 1 attack per year; 2: more than 1 attack per month, 3: more than 1 episode per week; 4: more than 1 episode per day). 22 For the frequency of partial cataplexy, there was no distinction between type 1 or type 2.
To assess depressive symptoms, the Beck Depression Inventory (BDI) was used in adults 23 and the Children's Depression Inventory (CDI) in children, 24 with scores greater than or equal to 16 considered as abnormal for both scales.
The symptoms associated with attention-deficit/hyperactivity disorder were scored by adults with NC 25 and by parents of children with NC using the Conners Parents Rating Scale-Revised (Conners RS-R). 26 Moderate to severe symptoms were defined with a cutoff above 65, severe symptoms with a cutoff above 75.
Fatigue was scored from 0 to 14 using the Chalder's fatigue scale 27 in adults and children, with abnormal cutoff score as greater than 10.
Insomnia was evaluated using the Insomnia Severity Index (ISI) 28 in adults and children. The total score is considered as pathological when it is higher than 10.
The severity of the disease was estimated using Narcolepsy Severity Scale (NSS), 29 in children the item concerning the driving was not filled.
Health-related quality of life (QoL) was assessed in adult patients with the SF36 health survey, 30 including the physical, psychological, general well-being as well as physical activity, pain, limitation, perceived heath, psychological health, vitality, and relations domains.
QoL was evaluated in children and adolescents using a questionnaire named Vécu et Santé Perçue adapted for adolescents (11-18 years) (VSP-A) 31 and for children (< 11 years) (VSP-E). 32 VSP is a self-questionnaire exploring the following dimensions: psychological and physical well-being, body image, vitality, friends, parents, teachers, medical staff, leisure, school performance, and a global QoL index (range: 0-100). Lower scores correspond to a poorer quality of life.

| Diagnosis procedure
The sleep and wake monitoring procedure included:

| Criteria for narcolepsy with cataplexy
As CSF hypocretin levels were not done routinely for narcoleptic diagnosis in our sleep laboratories, we used the criteria for NC

| Physical measures
Height and weight were measured, and body mass index (BMI) (weight/height 2 ) was calculated. The BMI z-score, which represents a measure of weight adjusted for height, sex, and age, relative to a smoothed reference distribution was computed. 35 In adults, overweight was defined by a BMI greater than 25 kg/m 2 and lower than 30, and obesity by a BMI greater than 30. In children, the World Health Organization (WHO) adolescence BMI-forage curves at 19 years closely coincide with adult overweight (BMI 25) at + 1 SD and adult obesity (BMI 30) at + 2 SD. As a result, these SD classifications are extended down to 5 years. 36 We used this definition in our study. Overweight and obesity were defined respectively when BMI was above + 1 SD and above + 2 SD for sex and age. Information concerning secondary sexual characteristics at symptoms onset were not known, and hence, the presence of precocious puberty could not be established. Early menarche was defined as menarche between age 9 and 11 years, using the 5th percentile of the French Health Behavior in school aged children distribution. 37

| General clinical characteristics
At time of NC diagnosis, the study group included 46 children (61% male) with a median age of 12 years old and 46 adults (61% male) with a median age of 28.5 years old, with no sex-ratio difference.
The BMI, BMI z-score, and waist size were higher in adults than in children. Overweight was more prevalent in adults than in children; in contrast, obesity was more frequent in children than in adults. The age at menarche is similar between children and adults. The delay from disease onset to diagnosis seems to be equivalent between children and adults. Nine children and 4 adults had the H1N1 seasonal flu vaccination prior to disease onset. HLA DQB1*0602 positivity was more commonly found in children than in adults. There was no difference between groups for the CSF hypocretin values.
Clinical characteristics of adults and children with NC are presented in Table 1.

| Narcolepsy symptoms
In terms of timing, children had earlier narcolepsy symptoms (sleepiness, cataplexy, hypnagogic hallucinations, and sleep paralysis) than adults (Tables 1 and 2). The frequency and severity of excessive daytime sleepiness was similar between groups. However, sleep paralysis was more frequent in adults than in children. Hypnagogic hallucinations also tended to be more frequent in adults, but not significantly. In contrast, there were more frequent parasomnia, sleep talking, night eating, and sleep drunkenness (defined as being tired when waking up in the morning) in children than in adults. There was no difference between groups for the presence of dyssomnia (defined by poor quality of sleep with frequent awakenings), familial parasomnia, and severity of the disease. The total disease severity score was 26 (range from 2 to 54) in children vs 30.5 (range from 7 to 48) in adults. Narcolepsy related characteristics are presented in Table 2.
The prevalence, frequency, and severity (partial vs total) of cataplexy were similar between groups (Table S1).
Although depressive feelings, insomnia, and fatigue symptoms did not differ between groups, the scores at the ADHD scale of Conners RS-R were on average higher in children than in adults, with a higher number of children with a pathological score. Adults reported a lower quality of life than children. Results for mood, fatigue, insomnia, attention disorders, and quality of life are presented in Table 3.

| Sleep measures
On nighttime polysomnography, the total sleep time was longer in children than in adults (Table 4) The mean and minimal transcutaneous oxygen saturation were lower in adults than in children. Higher BMI z-score were correlated with higher OAHI (r = .32, P = .002).
As for MSLT, the MSL tended to be lower in children than in adults. The number of SOREMPs was similar across groups; however, the number of SOREM/MSLT tests was higher in children than Two children had less than 2 SOREM but all with sleep latency less than 8 minutes.

| Determinants of quality of life
In adult patients, the bivariate association between the Q oL dimensions and the continuous covariates is provided in Supplemental  In children (Table S3 and Figure 2), depression was the factor that most affected the quality of life (44%), followed by insomnia (30%) and ADHD symptoms (17%). The factors mostly impacted psychological health and physical activity. However, there were few data for children.
In the multivariate regression model in children, when the CDI score was entered into the linear model adjusted for gender and age, no other continuous independent variable could significantly increase the likelihood of the model. The retained model with gender, age, and CDI score as independent variables "explained" 49% of the variability of the QoL score (N = 27).
In adults, the retained model with gender, age and Beck and fatigue scores as independent variables "explained" 56% of the variability of the QoL score (N = 36). In adults, QoL was mostly impacted by depression but also by fatigue and ESS scores. Similar results were found between Beck and Epworth scores, and these independent variables explained 52% of the variability of the QoL score (N = 38). There was a positive correlation between fatigue and Epworth scores in adults (r = .40, P = .01).

TA B L E 3
Mood, fatigue, insomnia, hyperactivity and quality of life in adults and children with narcolepsy-cataplexy obesity prevalence observed in childhood narcolepsy. Experimental studies suggest that the hypocretin system could play a role in the regulation of metabolism. 40 However, in the few patients reported in our study, no differences were found in CSF hypocretin-1 levels between children and adults. We recently reported an impairment in histamine neurotransmission, another brain wake-promoting system, in children with NT1, 41 which was not found in adults. 42 Histaminergic neurons could be also involved as studies in animal models have suggested that lack of either hypocretin or histamine could lead to a mild obesity. 43 Since this was a retrospective study, it would be interesting to conduct a prospective study in these children to follow their outcome during adult life.
In conclusion, compared to NC adults, children with NC presented with a higher prevalence of obesity, night eating, parasomnia, sleep talking, and ADHD. Conversely, adult patients showed more sleep paralysis, OSA, and a lower quality of life. Such differences in clinical characteristics and complications according to age need to be considered in order to ensure a prompt diagnosis and better management to prevent such complications.

ACK N OWLED G M ENTS
This work was supported in by a China Grant "China Scholarship Landel for providing language help.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.