Narcolepsy genetic variants associated with sleep efficiency in a community dwelling older cohort

Abstract

Narcolepsy type I (NT1) is a life-long debilitating autoimmune neurological condition characterised by excessive daytime sleepiness (EDS); the only symptom universal to all patients. Issues regarding sleep efficiency is also prevalent in individuals with NT1, however it remains relatively understudied due to the difficulty in measuring the effect. Genetic traits have shown to predispose an individual to NT1 and while HLA-DQB1 * 06:02 remains the most impactful genetic risk factor additional genes that contribute to immune cell processing have also been identified. In this retrospective study we impute 13 non-MHC narcolepsy associated single nucleotide polymorphisms (SNPs) from 1,558 non-pathological elderly volunteers who have been followed for up to a 24-year period to determine the association with sleep efficiency. Utilising a healthy cohort allows us to independently assess the potential contribution of each SNP on the impact of the sleep cycle disruption. We observed significant associations between SNPs and various elements of the sleep process; however, the main findings were the associations with disturbed night sleep (DNS). We observed an association with rs10915020 and rs1551570 with an increased number of wake episodes during the night, conversely rs2859998 and rs2834168 showed a protective effect—reducing the frequency of nighttime disturbances. While the association with NT1 and DNS has long been established, this is the first investigation that attributes elements of DNS to the genetic profile of the patient. This suggests that the issues with sleep efficiency reported by patients may be due to genetic predispositions and supports the variation seen in the co-morbidities associated with the condition.

Introduction

Narcolepsy type 1 (NT1) is a debilitating neurological disorder with a prevalence of 0.02% in European and US populations (Khatami et al. 2016). Symptoms of NT1 often present in early-adolescence and include excessive daytime sleepiness (EDS), hypnagogic/hypnopompic hallucinations and enhanced rapid eye movement (REM) sleep pressures. Symptoms present due to a lack of hypocretin (also termed orexin) in the spinal fluid resultant of the destruction of hypocretin/orexin producing neurons, however the pathophysiological mechanism remains unclear. Evidence suggests that NT1 is autoimmune in origin, with strong genetic and environmental risk factors associated with onset (Association with H1 N1 Infection and Vaccination. 2016). A subset of individuals with NT1 also experiences and uncontrollable loss of muscle tone, termed cataplexy, triggered by intense emotions such as laughter, excitement or fear (Dauvilliers et al. 2014). Patients experiencing a cataplectic episode may fall due to the sudden loss of strength at the knees, their head may drop while laughing or they may experience slurred speech due to a weakening of the facial muscles (Scammell 2015). Patients with narcolepsy symptoms but with normal levels of hypocretin are classified as Narcolepsy Type 2 (NT2). NT2 is less prevalent than NT1 and while it has been suggested that NT2 could be an early or mild form of NT1 however there is a lack of clear marks to define the condition (Ferrazzini, et al. 2024). Diagnosis of narcolepsy is complex due to difficulties in symptom recognition and overlap with other hypersomnolence disorders, in particular idiopathic hypersomnia which is less prevalent than narcolepsy but exhibits a number of the same symptoms (Lopez, et al. 2017) (Thorpy and Krieger 2014).

EDS is the only symptom universal to all narcolepsy patients (Sateia 2014) but additional symptoms may also be present, including sleep paralysis—the temporary inability to move while falling asleep or waking (Chavda, et al. 2022). Episodes of paralysis are self-limiting, lasting only a short time (1–10 min) and often disappear when interrupted by another person touching the patient (Chavda, et al. 2022). Behavioural and cognitive disturbances may also be experienced, but these are often self-reported incidents which are rarely confirmed by neurological evaluation. Such symptoms are often attributed to manifestations of depression resulting from the original diagnosis, or the adverse effects of prescribed medication (Naumann et al. 2006). Obesity is another common comorbidity of narcolepsy, and research has shown that paediatric patients with narcolepsy have up to a 20% higher body mass index (BMI) than non-narcolepsy patients (Poli et al. 2013; Chabas et al. 2007). Low metabolism has been proposed for the increased incidence of weight gain as individuals with narcolepsy burn less calories, further exacerbated by the increase in daytime sleepiness making sufferers comparatively less active than healthy individuals. Obesity combined with other common comorbidities of the condition including diabetes and depression also increases the risk of cardiovascular disease (Jennum et al. 2021; Black et al. 2017).

Due to variability in definition and lack of validated measure, a symptom which is often overlooked for NT1 patients is disrupted/disturbed nighttime sleep (DNS) (Maski et al. 2022). This is despite 30%—95% of NT1 patients and paediatric care givers reporting frequent awakenings during the night (Pizza et al. 2014; Roth et al. 2013). DNS is characterised by sleep fragmentation resultant of stage shifts and is described by patients as “difficulty staying asleep” and is measured by polysomnography (PSG) and actigraphy tests (Roth et al. 2013). DNS is exacerbated by motor control dysfunctions including restless leg syndrome (Plazzi et al. 2010) and mental health issues including depression and anxiety (Barateau et al. 2020). In narcolepsy, the contributing role of DNS is yet to be fully determined as multiple symptoms in these patients could disrupt sleep. However, DNS is significant in 65% of NT1 patients and has a severe impact on nocturnal sleep quality, and consequently EDS (Bassetti et al. 2021). In 2011, Roth et al. objectively measured sleep problems and the number of arousals after sleep onset in NT1 patients, concluding that frequent nocturnal awakenings and generally poor quality sleep are a feature of the disorder (Roth et al. 2013) although REM sleep disturbances, sleep paralysis and hallucinations also contribute (Sansa et al. 2010). DNS in NT1 patients is consequently considered to be a secondary symptom as a consequence of EDS or the medication administered to combat EDS.

The calculation to determine sleep efficiency is Total Sleep Time (TST) divided by Time in Bed (TBT) multiplied by 100. However, despite this objective calculation, the prevalence of sleep disturbance throughout the population in general is poorly defined. Genetic polymorphisms have been associated with nocturnal sleep disturbances in the general population, including DNS, trouble falling asleep, nocturnal arousal, low amounts of REM sleep and an increased need for daytime sleep (Barclay and Gregory 2013; Byrne et al. 2013; Shi et al. 2017). In addition, there are strong bidirectional links between sleep and the immune system, whereby the activation of the immune system has shown to promote an increase in both intensity and duration of sleep, but equally can also cause sleep disruption (Besedovsky et al. 2019). Sleep disruption can form a feedback loop with a lack of undisrupted nocturnal sleep potentially leading to chronic, systemic low grade inflammation which is linked with neurodegenerative conditions (Besedovsky et al. 2019).

Human leukocyte antigen (HLA) -DQB1* 06:02 encoded with the major histocompatibility complex (MHC) remains the strongest genetic marker that predisposes an individual to NT1, and is present in over 95% of confirmed cases (Rogers et al. 1997; Mignot et al. 2001). Despite the strong association with the presence of a single copy of HLA-DQB1* 06:02, and with risk onset increased up to five times with two copies HLA-DQB1* 06:02 (Mignot et al. 2001) only 0.001% of individuals who express this allele will develop narcolepsy (Kornum et al. 2017). The low risk associated with HLA-DQB1* 06:02 is due to the comparatively high frequency of this HLA allele in several global populations (Gonzalez-Galarza et al. 2020), implying that HLA-DQB*06:02 alone cannot be a single causative agent for disease onset. As a result, variants outside the HLA genes have been a focus for recent genome wide association studies (GWAS). The most promising SNPs outside the MHC are encoded within genes that interact with the HLA molecule or are involved in immunoregulatory function (Ollila et al. 2023; Hallmayer et al. 2009; Kornum et al. 2011).

DNS remains an understudied symptom of narcolepsy due to the difficulty in measuring the effect. Clarification of any genetic predispositions that contribute to DNS in narcolepsy patients could lead to a better understanding of the underlying pathophysiological mechanisms and an improved diagnosis. The presence of HLA-DQB1*06:02 in NT1 patients with cataplexy is also associated with an increase in wake episodes following sleep onset and a reduction in the N2 stage of the sleep cycle (Hong et al. 2000). Within the sleep cycle, N2 is the dominant sleep stage contributing to 50% of the total sleep time and a low N2 would indicate sleep fragmentation (Shrivastava et al. 2014).

Aim

In this retrospective study we aim to determine the relationship between non-MHC derived SNP variants associated with narcolepsy onset and their impact on sleep efficiency using longitudinal sleep data from The University of Manchester Longitudinal Study of Cognition in Normal Healthy Old Age (UMLCHA) cohort. By choosing a healthy population for this study, it is anticipated that the contribution of genes other than HLA-DQB1*06:02 to the symptoms observed in narcolepsy patients could be assessed independently for their potential impact on sleep cycle disruption.

Methods

Study cohort

The UMLCHA cohort consisted of 6,063 healthy adult volunteers (4,238 female, 1,825 male, median age 65 years) from Greater Manchester and Newcastle upon Tyne, recruited between 1982 and 1994 (Rabbitt et al. 1993). Sleep data was collected using a Personal Details Questionnaire (PDQ) which also collected several demographics including age, marital status, current health status, alcohol consumption, smoking status and usage of sleep medication completed at up to seven waves over a 24-year period (1985–2010). The first PDQ was completed at the initial recruitment stage until 1993. The second questionnaire was completed between 1984–1996, third questionnaire completed between 2001–2003 and the fourth and fifth were completed by participants in 2001 and 2010 respectively. As the recruitment into the study was distributed throughout the 24-year period, there was a degree of overlap as each participant completed the questionnaire at different dates and recruitment into the study continued over a twelve-year period. The number of participants who attended to each wave were, 6063, 3176, 624, 751, and 575, respectively. Mean ages at each wave were, 65.19 ± 7.45, 67.87 ± 6.82, 76.23 ± 5.70, 80.56 ± 5.34, and 82.88 ± 5.20 years (mean ± SD), respectively (Table 2). PDQs were completed either in person or by post. Written informed consent was obtained from all respondents at the onset of data collection. All PDQs were collected under the approval of the University of Manchester research ethics committee and this study includes secondary analysis of the anonymised dataset.

Sleep assessment

Questions within the PDQ relating to sleep (SQ) included: SQ1—“Generally at what time do you get up in the morning?”, SQ2—“Generally what time do you go to bed at night?”, SQ3—“On average how many hours sleep do you get every night?”, SQ4—“How many times during the night do you wake up?”, SQ5—“Do you have difficulty getting to sleep?”, SQ6 – “Do you take sleeping tablets?” and SQ7 – “Do you sleep through the night without waking?”. Answers to SQ2 spanned a 2-day period and were therefore converted into a 24:00 h measure (range − 12 to 12). The midpoint of sleep (CAL1) was calculated using the formula [SQ2 + (duration between SQ2 and SQ1)/2]. Sleep efficiency (CAL2) was determined as a percentage, calculated using the formula (SQ3/duration between SQ2—SQ1) * 100. Resting napping hours (CAL3) was determined per month and was calculated by dividing the resting hours by 31. Comparison of sleep parameters were generated using Pearson’s correlation performed using Stata (Stata Statistical Software. College Station, TX, StataCorp LP). Longitudinal changes in sleep timing were investigated using linear-mixed models. Growth curve analysis modelled sleep trajectories of the population adjusting for inter-individual and within individual variations and missing data.

Genetic variant selection

Thirteen non-MHC narcolepsy associated SNPs were identified using GWAS studies published before January 2022. All SNPs were selected using the EBI GWAS Catalog (https://www.ebi.ac.uk/gwas/) and were only included if they had an odds ratio (OR) greater than or equal to 1.00 and a minor allele frequency (MAF) greater than 0.05. See Table 1. for the list of candidate SNPs selected.

Table 1 SNPs selected or the study

Genotyping SNPs

The DNA Archive for Ageing and Cognition was established following invitation to all participating volunteers between 1999 and 2004. Ethical approval was obtained from University of Manchester and of the original 6063, 1563 volunteers agreed to supply a peripheral blood sample for DNA isolation. Following informed consent, venesected whole blood was collected for DNA extraction. DNA was obtained using a standard phenol/chloroform technique from whole blood on 1,563 volunteers. GWAS data was generated using the Illumina Human610-Quad BeadChip and was available for 1,558 volunteers, five samples were excluded as they failed to meet either the missing call rate threshold (≥ 2%) or were outside Hardy–Weinberg Equilibrium (p-value ≤ 0.001). The data was subjected to standard GWAS quality control and was imputed using the 1000 genome reference panel (Genomes Project and C., 2015).

Statistical analysis

Linear regression analysis was used to determine the association between the chosen SNPs and the continuous measures of sleep (baseline and longitudinal) – SQ1, SQ2, SQ3, SQ4, CAL1, CAL2 and CAL3. Linear regression results are reported in terms of beta and p-values. For longitudinal data the average decline in the general population for each sleep characteristic is expressed as 0. A SNP with a beta value of > 0 suggests that the SNP is associated with a slower than average rate of decline in that sleep function, conversely if the beta value is < 0 it indicates that the SNP is associated with a faster rate of decline. Logistical regression analysis was performed for all binary outcomes; SQ5, DQ6 and SQ7 for both baseline and longitudinal data. In addition to P values, odds ratios (OR) were determined for each SNP.

Linear and logistic regression analysis of all SNPs was performed using Plink Software (version 1.9, Shaun Purcell, https://www.cog-genomics.org/plink/) with age and sex included as covariates (Purcell et al. 2007). P values of ≤ 0.05 were considered significant.

Results

Demographics

The cohort was predominantly female (69.91%), with a mean age at initial time of recruitment of 65.19 years and 82.88 years at the last time of data collection. The incidence of smoking was at its highest at the initial time of recruitment (17.78%) which then decreased to 3.9% at the time of last data collection. The trend of alcohol consumption increased throughout the years of data acquisition with a steady increase in volunteers consuming alcohol more than once per day. Self-assessment of health remained constant throughout the years of data collection, whereby much of the cohort defined themselves as of good health. There was a decline in the use of sleep medication throughout the years of data acquisition, calculated as 12.09% at wave 1 and steadily decreasing to 5.91% at the final time of data collection – see Table 2.

Table 2 Volunteer demographics

Sleep demographics

The mean “Get Up” and “Go to Bed” times were consistent throughout the five waves of data collection, varying by only seven and twenty-seven minutes over the 22 years respectively. The mean number of times waking during the night was at its peak at wave 5 (average of twice per night), indicating that the volunteers are reporting an increase in disturbed night’s sleep as their age increases. This coincides with the percentage number of volunteers sleeping through the night (SQ7) which decreases with age. However, this contrasts with the number of volunteers who reported difficulty sleeping (35.63% in wave 1 compared to 26.45% in wave 5) which is consistent with a reduction on reliance on sleeping tablet – see Table 3.

Table 3 Sleep demographics

Regression analysis

Baseline linear analysis

Both the linear and logistical regression analysis showed nominal significant associations for the SNPs of interest and the impact on sleep characteristics. There is an association between rs4290173 and the time the volunteers would go to bed (P = 0.028), indicating that the presence of the minor allele for this SNP is associated with individuals going to bed earlier. rs4290173 was also nominally associated with a reduction in the mid-point of sleep at the time of initial data collection – P = 0.021. The “Go to Bed” time is incorporated in the calculation to determine the midpoint of sleep so this association is potentially expected. The logistic regression showed an additional association with rs1154155 and ability to sleep through the night (P = 0.043, OR = 1.242). This correlated with the linear regression analysis for this SNP and Wake Episodes which showed a non-significant trend (P = 0.056) – see Table 4.

Table 4 Linear regression P values for all phenotypes at the time of initial data collection

Longitudinal linear analysis

Rs10995245 was associated with a protective effect against waking up (P = 0.012) suggesting that the presence of this minor allele marginally correlates with individuals remaining asleep for longer. rs306336 correlated significantly with a reduction in the number of resting hours per month (P = 0.035). The sleep characteristic that showed the most significance with the targeted SNP panel was the number of wake episodes. rs10915020 and rs1551570 were statistically associated with an increase in the number of wake episodes during the nocturnal sleep – P = 0.024 and 0.018 respectively. Conversely, rs2859988 (P = 0.042) and rs2834168 (P = 0.005) were associated with a reduction in the number of wake episodes suggesting that the minor alleles of these SNPs are protective against waking multiple times. We found no significant association with the narcolepsy associated SNPs and the number of hours sleep (SQ3), sleep efficiency (CAL2), difficulty sleeping (SQ5) and the use of sleep tablets (SQ6) in the baseline and longitudinal data sets—see Table 5.

Table 5 Linear regression P values for all phenotypes for the longitudinal results showing the rate of decline over a 12–18 year period

Logistic analysis

We found no significant association with the narcolepsy associated SNPs and the binary outcomes of difficulty sleeping (SQ5) and the use of sleep tablets (SQ6) in both the baseline and longitudinal data sets. There was however a nominal association with rs1154155 (P = 0.043) and the occurrence of individuals sleeping through the night (SQ7) at the initial time of data collection, but this was not replicated in the longitudinal dataset – see Tables 6 and 7.

Table 6 Logistic regression analysis—Cross-sectional results at the time of initial data collection

Table 7 Logistic regression analysis—Longitudinal results showing the rate of decline over a 12–18 year period

Discussion

EDS is experienced by all narcolepsy patients, and in NT1 patients the sleepiness is accompanied by hypocretin deficiency and cataplexy. NT1 is symptomatically heterogenous with a high degree of variability in expression and severity. Despite being present in 30–95% of sufferers, DNS remains relatively understudied (Barateau, et al. 2022). The wide range in DNS prevalence is mainly due to the variability in definition, assessment methods and the exclusion of comorbid sleep disorders (Maski et al. 2022). DNS can be defined as sleep instability which includes frequent brief awakenings, increased wake time after sleep, poor sleep quality and altered sleep stage transitions, determined using a polysomnography (PSG) and actigraphy. Although the gold standard, there are limitations with the PSG including the first night effect (poor sleep due to a change in environment), interference by medication, false readings and instrument malfunctions. Actigraphy has the benefit of being able to be performed over a series of days/weeks but often provides less information, is prone to artefacts and is unable to distinguish between wake and sleep (Sadeh 2011).

In this study, we analysed self-reported sleep characteristics in a large sample of non-narcolepsy adult patients longitudinally over a 24-year period to determine whether there was an association with SNPs associated with narcolepsy onset. While some of the results generated nominal P values, the characteristic that showed the greatest promise was the frequency of arousal, whether enhanced or suppressed in the longitudinal data.

The minor allelic variants of both rs2859998 T and rs2834168 A were significantly associated with a reduction in the number of nocturnal awakenings, whereas rs10915020G and rs1551570G showed an association with an increase in the number of nocturnal arousals. rs2859998 is an intronic variant encoded on chromosome 8 within the UBX domain protein 2B (UBXN2B) gene complex. UBXN2B is expressed by monocytes and neutrophils, playing a role in the cell cycle and regulating spindle orientation during mitosis (Kress et al. 2013). The SNP rs2834168 is located on chromosome 21 within the genes encoding the interferon alpha receptor subunit 2 and interleukin 10 receptor subunit beta (IFNAR2-IL10RB), expressed on dendritic cells and non-classical monocytes. The main function of IFNAR2-IL10RB is to regulate the production of upstream and downstream gene products. Both observations suggest that both these variants conferring a protective effect against excessive nocturnal awakenings are involved in cell maintenance and regulation. The SNP rs10915020, located on chromosome 1 within the genes that encode the Small Integral Membrane Protein 12 (SMIM12), encodes proteins predicted to be an integral component of the cell membrane of eosinophils. Located on chromosome 19, rs1551570 lies within the Peter Pan Homolog – Purinergic Receptor P2Y11 (PPAN-P2RY11) genes which is a member of a large family of more than 20 purinergic receptors ubiquitously expressed in the spleen and lymph nodes. The purinergic signalling pathway plays a fundamental role in immune regulation, proliferation, apoptosis and chemotaxis in lymphocytes and monocytes (Bours et al. 2006). Irregularities in purinergic signalling have been shown to impact sleep efficiency and have been extensively studied in Alzheimer disease and depression (Ribeiro et al. 2023).

Other nominal findings included the minor allelic variant of rs4290173 A which was associated with volunteers going to bed earlier and in the initial data cohort, a reduction in their midsleep component. rs4290173 is located 4.3kb upstream of APOBEC1 complementation factor (A1 CF) which functions as an RNA binding subunit, docking APOBEC1 and deaminating the upstream cytidine (Snyder et al. 2017). A mutation of A1 CF leads to the lack of RNA C – U editing for APOB, producing APOB- 48 which results in the increased adsorption of lipids from the intestinal lumen. This process has been shown to increase the weight of an individual which is also correlated to poor sleep efficiency (Sun et al. 2015). rs10995245 A was associated with later wake up time. rs10995245 is an exonic SNP located within the genes of zinc finger protein 365 (ZNF365) on chromosome 10, the main function of ZNF365 is to repair DNA damage and to maintain the stability of the genome. We have previously reported the protective effects of rs10995245 on cognitive decline which is also experienced by a number of individuals with narcolepsy (Jervis and P.A., Lowe M, Didikoglu A, Verma A, Poluton K, 2023). Despite the minor allele of rs10995245 A being associated with disease onset, it would be reasonable to suggest that based on these and previous findings the presence of this minor allele is protective against some of the other reported symptoms experienced by narcolepsy patients.

Sleep is homeostatically regulated, implying that an extended period without sleep would be followed by a longer period of sleep. In addition to this homeostatic cycle, the circadian system modulates the timing of sleep and synchronises the body to a 24-h sleep–wake cycle which is impacted by behavioural traits and body functions (Besedovsky et al. 2019). Both systems are controlled by the central (CNS) and autonomic nervous systems (ANS). The CNS/ANS and immune system are closely linked and regulate each other, i.e. when systems under the control of the CNS/ANS are activated, it often leads to the induction of an inflammatory response (Bierhaus et al. 2003). Equally, activation of the immune system following stimulation by foreign peptide prompts activation of the CNS/ANS. Therefore, it is reasonable to suggest that polymorphisms in genes that encode elements of the immune system that bring about modifications to the “normal” processes would impact on the CNS/ANS equilibrium. Interaction between the neurological and immunological systems has been extensively studied in multiple sclerosis (MS). In MS, GWAS analysis has identified over 200 genes which are classed as genetic risk factors for the condition and are primarily expressed in a wide range of immune cell types (Parnell and Booth 2017). The variants (combined with environmental stimuli) lead to a loss of peripheral tolerance causing inflammation and myelin breakdown. Consistent with NT1, MS sufferers also experience issues with sleep regulation which impact on quality of life.

The association with narcolepsy and DNS has long been established. However, this is the first investigation that attributes elements of DNS to the genetic profile of the patient. Our findings are consistent with others, who have shown that narcolepsy patients do not have difficulty falling asleep but struggle to maintain prolonged periods of sleep and suffer with frequent (not prolonged) wakefulness (Roth et al. 2013). This is often regarded to be a consequence of the primary symptom, EDS, implying that daytime sleepiness episodes impact the nocturnal sleeping efficiency. The other suggested rational for DNS is the pharmacological therapy used to combat EDS including stimulants administered prior to nocturnal sleep. While there is strong evidence that both contribute to DNS, is it also reasonable to suggest that the susceptibility to DNS is further enhanced for patients who carry certain genetic risk traits. While genetic association studies can only suggest that DNS could result from genotypic variants, introducing genetic screening for these polymorphisms at the time of diagnosis can only be of benefit when determining the appropriate course of treatment. A better understanding on the causes of DNS and the wider impact of a patient’s genetic profile is key to the improvement of individualised therapy in the routine care of narcolepsy patients.

The present study represents the first attempt to show significant associations with SNPs previously correlated with narcolepsy onset and the co-morbidities experienced by patients. Nevertheless, there are some limitations that should be acknowledged. Firstly, disrupted night’s sleep can be attributed to several other conditions including obstructive sleep apnoea, restless leg syndrome and others (Bogan 2006; Adekolu and Zinchuk 2022), particularly in an aging population. No data was collected regarding sleep apnoea or restless leg syndrome so this should be considered a weakness of this study. Additional confounders to sleep in an elderly population include chronic pain (e.g. arthritis), neurodegenerative disease (e.g. Alzheimer’s disease), respiratory disorders (e.g. COPD), depression/anxiety, loneliness, medications (e.g., beta-blockers or antidepressants), physical inactivity, and daytime napping – we didn’t have any data on these so they couldn’t be controlled for as covariates. In addition, we did not have data on mortality rates, so this could not be controlled for as a covariate. Secondly, there are inherent inaccuracies often seen in subjective sleep quality assessments. Evaluations based on subjective measures have been shown on occasion to be unreliable, especially in measuring sleep fragmentation (Oakes et al. 2022). Finally, in this exploratory research study, the aim was to identify associations that could guide future studies. Due to the small effect size of the SNPs investigated the inclusion of stringent statistical correction for multiple analyses (i.e. Bonferroni correction) was not considered. Such correction methods increase the chance of type II error incidence, potentially resulting in the loss of meaningful results. However, it is acknowledged that omission of statistical correction may have increase the false discovery rate and introduced the element of chance. Further work in this field should look to address these limitations and may require a more detailed exploration of the relationship of SNPs defined with narcolepsy onset and the co-morbidities of the condition.

Conclusion

The aim of this preliminary study was to determine whether variants previously associated with narcolepsy onset also influence sleep in healthy individuals. Whilst triggers may be different, we may infer from any associations that the same pathways are involved. We identified several SNPs that may contribute to DNS in this patient cohort. Despite being a frequently reported symptom, the burden of DNS in narcolepsy patients has not been fully established, primarily due to the difficulty in assessment and evaluation. Even with the benefit of PSG and actigraphy, there is a lack of consistency in the definition of DNS between patients and clinicians. A genetic screening assay could be used to support this element of the diagnosis and would offer an additional level of diagnostic stratification. Genetic risk scores for onset have been developed in the field of narcolepsy and an expansion of this to include co-morbidities would be of great interest (Ouyang et al. 2020). In the general population, it is well established that DNS is associated with an increased risk of comorbidities seen in narcolepsy patients, including cardiovascular disease, depression and fibromyalgia (Jennum et al. 2021; Li et al. 2021; Disdier et al. 1994). While not all narcolepsy sufferers will experience these conditions, the support of a genetic screening assay may help guide clinicians when diagnosing the primary condition. Identification of individuals with narcolepsy who are at a higher risk of certain co-morbidities will allow for therapeutic intervention to be appropriately tailored.