Exploring sleep challenges in CDKL5 Deficiency Disorder

Abstract

Background

Sleep disturbance is a common and significant issue for individuals with CDKL5 Deficiency Disorder (CDD) and their families. The study aimed to investigate experiences of sleep disturbance in CDD and associated factors.

Methods

Data were sourced from the baseline and follow-up questionnaires completed by caregivers of 258 individuals in the International CDKL5 Disorder Database. Outcome variables were scaled scores for selected domains of the Sleep Disturbance Scale for Children: Disorders of Maintaining Sleep (DIMS), Disorders of Excessive Somnolence (DOES), Sleep Breathing Disorders (SBD), and Sleep-Wake Transition Disorders (SWTD). For DIMS and SBD, t scores were calculated and dichotomised into high (≥ 70) and low (< 70) groups. Through univariable and multivariable Poisson and logistic regression analyses, the relationships between the sleep disturbance domains and covariates, including sex, age group, genetic variant, motor skills, seizure frequency and patterns, medication use and side effects, constipation, and emotional behaviours, were examined. Caregivers also provided qualitative data on their children’s sleep abnormalities. Content analysis assessed caregivers’ responses to the prompt, “Is there anything about your child’s sleeping that you consider unusual or problematic?”.

Results

Individuals  ≥18 years old had an adjusted mean DIMS scaled score 6.49 (95% CI -12.44,-0.55) points lower than those aged 3–5 years. Daily seizures were associated with lower DIMS (-6.65, 95% CI -11.52,-1.78) and higher SBD (7.09, 95% CI 2.39,11.79) scores. Those whose parents reported constipation had more than twice the odds of a high DIMS t score (2.39, 95% CI 1.24,4.62). The mean DOES-mod score was 6.92 (95% CI 0.40,13.44) points higher in those able to sit and walk than in those unable to sit and walk, and was higher in adolescents and adults. Occurrence of seizures during night-time sleep was associated with higher SWTD-mod scores (4.09, 95% CI -0.15,8.32). Caregivers reported insomnia (sleep initiation and maintenance), “night parties” (possibly reflecting disrupted circadian rhythms), and daytime somnolence affecting functioning as major concerns, with seizure timing and frequency, gastrointestinal symptoms, and medication use identified as contributing factors.

Conclusions

Functional abilities and comorbidities influenced sleep. Reducing the impacts of constipation and side-effects of anti-seizure medications on sleep could improve sleep and enhance child and family wellbeing.

Supplementary Information

The online version contains supplementary material available at 10.1186/s11689-025-09666-w.

Introduction

CDKL5 Deficiency Disorder (CDD) is a rare developmental and epileptic encephalopathy caused by a pathogenic variant in the cyclin-dependent kinase-like 5 (CDKL5) gene on the X chromosome [1]. The initial presentation is generally early-onset infantile seizures that typically occur within the first 2 weeks to 6 months of life. These early-onset infantile seizures may disappear, such that a “honeymoon” period is observed whereby no seizures occur for weeks to months before their re-occurrence [2]. Many individuals experience seizures that are refractory to treatment [2] (anti-seizure medications (ASMs), ketogenic diet [3] and, used less frequently, vagal nerve stimulation) [4]. In addition to this epileptic encephalopathic presentation, a variety of other symptomatology is seen in CDD [1]. This includes severe developmental impairment; gastrointestinal issues including feeding difficulties, reflux and constipation; cortical visual impairment; musculoskeletal problems (including scoliosis); respiratory infections; and behaviour and sleep disturbances [1]. Much research identifying these associated comorbidities has been implemented using the International CDKL5 Disorder Database (ICDD) [1], which was established in 2012 and now includes over 400 cases.

Sleep disturbances have been reported in over 80% of individuals with CDD at some point in their life [5]. These sleep disturbances could have a detrimental impact on the individual both by lowering the seizure threshold and by contributing to behavioural difficulties. The association between sleep disturbances and the quality of life of the individual with CDD has been described in two publications. In the first, we noted that severe sleep difficulties (calculated as the average score of the 22 items included from the Sleep Disturbance Scale for Children [SDSC], developed by Bruni and colleagues [6], and subsequently stratified into quartiles) were associated with lower scores in the physical health and negative emotions domains of a quality of life instrument, QI Disability [7]. In the second, two individual domains of the SDSC, Disorders of Initiating and Maintaining Sleep (DIMS) and Disorders of Excessive Sleepiness (DOES) were examined [8]. Similar findings concerning physical health and negative emotions were observed for the DOES, but for the DIMS, the only association was with negative emotions. Not only do sleep disturbances have a negative impact on the individual with CDD, but it has also been shown that with progressively worsening sleep, the mental health of the caregiver and family unit declines, especially in terms of maternal emotional well-being [9].

Over 200 pathogenic variants have been identified to date to cause CDD [10], and each could potentially have different effects on sleep. For ease of analysis, these variants have been grouped into four categories based on location on the CDKL5 protein: (A) No functional protein; (B) Missense or in-frame mutations within the catalytic domain; (C) Truncations between amino acid 172 and amino acid 781; and (D) Truncations after amino acid 781 [10, 11]. The earlier study using 167 genetically verified cases from the ICDD found that those with no functional CDKL5 protein were slightly more likely to experience any sleep difficulty than those in the other variant groups [12]. Further investigation in a larger dataset is warranted.

Individuals with CDD have been reported to experience a median of two seizures per day, with the majority having 1 to 5 per day [2]. Epilepsy in CDD is particularly refractory to medication, with a small minority being seizure-free and almost half (45%) taking three or more anti-seizure medications [2]. Other non-pharmaceutical treatments include ketogenic diet [3] and vagal nerve stimulation [4]. Epilepsy in general has also been shown to be associated with a twelve times higher rate of sleep difficulties [13]. Difficulty with sleep may lead to a reduced seizure threshold and can further perpetuate the frequency and severity of seizures in individuals with epilepsy [14, 15]. Given this, there may be a relationship between seizure frequency and sleep disturbances in CDD.

Not a great deal is known about the use of polypharmacy for management of seizures and its effects on sleep in people with intellectual disabilities [16] and even less for individuals with CDD. Many of the pharmaceuticals used for seizure control in CDD may be associated with adverse sleep effects, particularly daytime drowsiness. Specifically, phenobarbital, levetiracetam, and sodium valproate have been linked to increased daytime somnolence [17]. Therefore, there may also be relationships between the types of anti-seizure medications used and sleep disorders in children with developmental disorders such as CDD [16].

The most frequently used sleep medication in CDD (and also in Rett syndrome, a closely related developmental encephalopathy [18] as well as in other neurodevelopmental disorders [19]) was melatonin, prescribed for 29/129 in a recent ICDD study [8]. In another mainly European study of early seizure semiology identified from EEG patterns and parental interviews, melatonin had been used by 10/30 patients with no improvement apart from a small benefit in one using slow-release melatonin [20]. Other medications used include clonidine and trazodone [8]. Despite the use of these medications, sleep difficulties have generally persisted [8, 20]. Understanding the relationship between poor sleep and medication use in CDD will be important.

Additional to severe developmental impairment children and adults with CDD experience many comorbidities that may impact sleep. In a study involving children with intellectual disability, specifically Rett syndrome, Down syndrome, cerebral palsy and autism, those who had more functional impairment were more likely to experience sleep breathing disorder (SBD) or excessive sleepiness but not insomnia [16]. Comorbidities associated with recurrent pain or frequent coughing also had adverse associations with each of these sleep problems [16]. At present, we have no information on whether there is any relationship between the degree of functional impairment or the presence of non-seizure related comorbidities and sleep disturbances in CDD.

Although sleep disturbance is a common and significant issue for individuals with CDD and their families, to date it has received relatively little attention in the medical literature. Moreover, many factors may negatively impact sleep, including seizures, pharmaceutical management, comorbidities, and disrupted sleep cycles or circadian rhythm. The purpose of this mixed methods study was: (1) to describe and quantify the degree of disturbances in initiating and maintaining sleep, daytime sleepiness, sleep-breathing, and sleep-wake transition in CDKL5 Deficiency Disorder using relevant domains of the Sleep Disturbance Scale for Children (6); (2) to identify whether these outcomes have any relationships with sex, agegroup, genotype, gross motor function and comorbidities, specifically seizures and their treatment, constipation and behaviour; and (3) to analyse caregiver text responses to identify factors and experiences affecting sleep. This was completed using the latest data provided to the ICDD by mid-2023 on 258 individuals with CDD aged 3 years and over.

Methods

Data source and study population

Data for this study were obtained from the ICDD, which contained responses from a baseline questionnaire administered between 2012 and 2023 and a follow-up questionnaire administered between 2018 and 2019. This retrospective cohort study included individuals with a pathogenic or likely pathogenic CDKL5 variant aged 3 years and over at ascertainment.

Outcome variables

The Sleep Disturbance Scale for Children (SDSC) is a validated parent-report paediatric sleep questionnaire that comprises 26 items that load into six factors, “Disorders of Maintaining Sleep” (DIMS); “Disorders of Excessive Somnolence” (DOES); “Sleep Breathing Disorders” (SBD); “Sleep Wake Transition Disorders” (SWTD); “Disorders of Arousal and Nightmares” (DA); and “Sleep Hyperhidrosis” (SHY)] [6]. For this study, we only consider the DIMS, DOES, SBD and SWTD factors as we considered these four potentially relevant to our research question. Parents or caregivers (“caregivers”) were asked to recall their child’s sleep during the past 6 months and respond to all selected items on a 5-point Likert scale. All items for the DIMS and SBD factors had been included in the family questionnaire. On the other hand, two of the five items in the DOES (“Unable to move when waking up in the morning” and “Awakes in the morning feeling tired”) and one of the six items in the SWTD factors (“Experiences vivid dream-like scenes while falling asleep”) were excluded from our questionnaire because these particular items could be inappropriate for a disorder as severe as CDD. Thus these two domains were renamed DOES-mod and SWTD-mod respectively. The first item was omitted from the DOES because these children have severe gross motor impairment and their movement is already limited. Cross-sectional data has indicated that less then two thirds could sit without support and fewer than a quarter could take steps independently [1, 21]. The second item was omitted because the majority have no expressive language and therefore they would not be able to indicate to their parent that they were feeling tired. We previously reported that less than a quarter were able to communicate using spoken language, signs, or abstract symbols [1, 21]. This was also the reason for excluding the item about “dream-like scenes” in the SWTD domain and the items about recalling events including “nightmares” in the DA domain. Although we did ask about sleep-walking, one of the items in the DA domain, we did not consider the responses valid given such a small proportion was able to walk. Caregivers were also asked if there was anything about their child’s sleeping that they considered unusual or problematic. If yes, they were asked to elaborate in an open question. There was no word limit for the response, and the length of responses varied from one sentence to multiple paragraphs.

Covariates

Data for sex, age at questionnaire completion, variant groups and country of residence were extracted from the ICDD. Age was categorised into four groups: 3–5 years, 6–10 years, 11–17 years, and 18 years or above. Variants were categorised as has been done previously [10, 22] into the following groups: (A) No functional protein, comprising any mutation that prevents function in the catalytic domain; (B) Missense or in-frame mutations within the catalytic domain,; (C) Truncations occurring between amino acid (aa) 172 and aa781; and (D) Late truncations occurring after aa781. Country of residence was categorised based on broad geographic region (USA and Canada, Western, Northern and Southern Europe, Australia and New Zealand, and others).

Gross motor function was assessed through the ability to sit on the floor for 10 s and to walk 10 steps through the following categories: (1) No assistance; (2) Mild or moderate assistance; and (3) Maximum assistance or unable to. These functional abilities were then condensed into a ternary variable for better characterisation: (1) Able to sit and walk; (2) Able to sit but unable to walk (unable was defined as mild, moderate, severe or unable to); and (3) Unable to sit and walk. Seizure frequency was categorised as (1) Seizure free; (2) Monthly or less/Weekly; (3) Daily (1–4 per day); and (4) Daily (≥5 per day). Caregivers also reported whether their child was having seizures more frequently when asleep during the night and/or during the day. A binary variable was developed for sleep-related side effects of anti-seizure medications that flagged the reported occurrence of any symptom such as sleepiness, fatigue, drowsiness, oversedation and insomnia. Furthermore, caregivers had been asked whether their child had constipation at the time of questionnaire completion. Behavioural issues were characterised using the Rett Syndrome Behaviour Questionnaire (RSBQ) recently developed factor “Emotional and Disruptive Behaviours” [23], which consists of 12 items with a maximum score of 24. However, four of these items (“Spells of inconsolable crying for no apparent reason during the night”, “Spells of screaming for no apparent reason during the night”, “Spells of laughter for no apparent reason during the night”,and “Spells of laughter for no apparent reason during the day”) were thought not to be relevant to CDD and, thus had not been asked in the questionnaire. The exclusion gave a reduced summed score of 16. The scale was subsequently dichotomised into a low and high behavioural problem group using a cut-off of 8 points.

Data management

Relevant variables from the baseline and follow-up questionnaires were ascertained from the ICDD and initially checked using descriptive statistics and exploratory data analysis. Information from the last observation point was used, so if both baseline and follow-up data were available, the latter was chosen. For the outcome variables (i.e., the SDSC factors), missing data existed, and a simple imputation exercise was conducted. The median score of the factor was first calculated and then applied to the items with missing data within the factor. This was only relevant to records where there was a response for at least half of the items in the factor. For example, the median DIMS score was applied to missing items provided the caregivers had responded to at least four of the seven items.

Statistical analysis

Quantitative analysis (sleep disturbance scale for children)

Descriptive statistics were used to summarise the characteristics of the study population. The imputed outcome variables were transformed using two approaches: (1) Scaled score – scaled to a range of 20 to 100, as done previously [8] (2) Binary variable – t score was first calculated based on normative data from Bruni [6]. This only applied to the DIMS and SBD factors, as some items in the DOES and SWTD factors were excluded, as previously mentioned. The DIMS and SBD t score was then dichotomised into two groups as low (< 70) or high (≥70) sleep disturbance. Poisson and logistic regression were used to model the association of the covariates with the skewed scale score and binary outcome, respectively. The corresponding effect sizes (Poisson: mean differences; Logistic: odds ratios) and their 95% confidence intervals were estimated. Univariable and multivariable (adjusted for all covariates) analyses were carried out. Complete case analysis was used because the percentage missing was below five per cent for all covariates except “Patterns of occurrence of seizures during sleep”. All analyses were carried out using Stata version 18.0 (StataCorp. 2023. College Station, TX).

Qualitative analysis (open question)

Text data responses to the open question were imported into NVivo software [24], which was used for content analysis. ES developed the initial coding scheme based on in-depth reviews of half of the qualitative data using an inductive approach which involved classifying the data after reading it multiple times. This approach allowed the broad categories to emerge from the caregiver responses rather than having them predefined. This coding scheme with examples was discussed with HL and JD. Following that discussion, ES reviewed and coded the second half of the qualitative data. HL co-coded 45% of the qualitative data and any adjustments to the coding scheme were discussed with HL and JD.

Results

The characteristics of the study population (n = 258) are shown in Table 1. Females comprised 83% of the individuals, and just under two-thirds (62.8%) were between 3 and 10 years old. Almost half (49.2%) were residents in North America and just under a quarter (23.6%) in Europe. Almost 60% were able to sit independently, and almost a quarter (24.8%) could walk independently. Over half were experiencing one or more seizures per day. In just over a quarter (27.5%), parents reported that seizures occurred more frequently during night-time sleep and in 10.1% during daytime sleep. A sleep study had been conducted among 21.3% (55/258) of individuals and about 11.2% (29/258) of them were reportedly diagnosed with some degree of sleep apnoea. Despite this, only a very small proportion, 1.2% (3/258), of caregivers mentioned that their children were prescribed CPAP technology. All but 10% were taking ASMs, and two-thirds (65.9%) took between 2 and 4 ASMs per day. Over a quarter of caregivers (27.9%) had independently reported sleep-related side effects concerning the ASMs that their children were taking. Over a fifth (17.8%) had taken melatonin to promote sleep with a median dose of 3 mg (IQR 3–6 mg). Constipation was reported for over half (59.3%) of individuals. Nearly a third (32.9%) had (≥8) high scores for the emotional and disruptive behaviours domain.

Table 1.

Characteristics of 258 individuals aged 3 years or above with CDKL5 deficiency disorder (CDD)

	n (%) or median (interquartile range; range)
Sex
Female	214 (82.9)
Male	44 (17.1)
Age at completion of questionnaire, in years	8.8 (5.6,13.6; 3.0–42.4.0.4)
3–5	78 (30.2)
6–10	84 (32.6)
11–17	64 (24.8)
≥18	32 (12.4)
Variant group
No functional protein	62 (24.0)
Missense/in-frame variants within catalytic domain	78 (30.2)
Truncating variants between aa172 and aa781	70 (27.1)
Truncating variants after aa781	30 (11.6)
Other variants	18 (7.0)
Country of Residence
USA and Canada	127 (49.2)
Western, Northern and Southern Europea	61 (23.6)
Australia and New Zealand	27 (10.5)
Othersb	43 (16.7)
Floor sitting/Walking 10 steps
Able to sit and walk	64 (24.8)
Able to sit but unable to walk	90 (34.9)
Unable to sit and walk	101 (39.1)
Missing	3 (1.2)
Seizure frequency
Seizure free	34 (13.2)
Monthly or less/Weekly	70 (27.1)
Daily (1–4)	78 (30.2)
Daily (≥5)	65 (25.2)
Missing	11 (4.3)
Patterns of occurrence of seizures during sleep
During daytime sleep
No	202 (78.3)
Yes	26 (10.1)
Missing	30 (11.6)
During night
No	159 (61.6)
Yes	71 (27.5)
Missing	28 (10.9)
Count of antiseizure medications currently used
None	26 (10.1)
1	47 (18.2)
2–4	170 (65.9)
≥5	10 (3.9)
Missing	5 (1.9)
Sleep related side-effects of antiseizure medications
No	186 (72.1)
Yes	72 (27.9)
Constipation – Current
No	96 (37.2)
Yes	153 (59.3)
Missing	9 (3.5)
Emotional and Disruptive Behaviours
Low (< 8)	170 (65.9)
High (≥8)	85 (32.9)
Missing	3 (1.2)
Sleep study performed
Yes	55 (21.3)
Diagnosis made of sleep apnoea
Yes	29 (11.2)
Use of melatonin to promote sleep
Yes	46 (17.8)

NA, not applicable; n, number of individuals

^a includes Belgium, France, Germany, Luxembourg, Switzerland, Netherlands, Denmark, Finland, Norway, Sweden, Ireland, United Kingdom, Greece, Italy, Portugal, Slovenia, and Spain

^b includes Argentina, Belarus, Bermuda, Brazil, Bulgaria, Chile, China, Georgia, Hungary, Israel, India, Japan, Kazakhstan, Malaysia, Mexico, Peru, Philippines, Poland, Puerto Rico, Romania, Russia, Saudi Arabia, Turkey, Ukraine, United Arab Emirates

Sleep disturbance scale for children (quantitative)

Univariate analysis

Disorders of maintaining sleep (DIMS)

The quantitative results are presented in Tables 2 and 3. The median scaled score was 45.7 (interquartile range [IQR] 37.1–57.1), and, compared to those aged 3–5 years, the mean value was 5.18 points lower (95% confidence interval [CI] −10.18,−0.17) in those aged 6–10 years and 8.00 points lower (95% CI −14.46,−1.55) in those aged 18 years. Those whose mutation was classified as another variant were almost four times (odds ratio [OR] 3.96, 95% CI 0.88,15.86) as likely to have a high t score. At the same time, those for whom sleep-related side-effects of ASMs had been reported were almost twice (OR 1.85, 95% CI 1.04,3.32) as likely to have a high t score. For those with 1 or more seizures daily the mean DIMS scaled score was 46.3 (SD 13.5, IQR 37.1–54.2, n = 143) Those with 1–4 seizures daily had a mean score of 7.31 points lower (95% CI −12.03,−2.58) than those who were seizure-free or had less than daily seizures. In contrast, those currently experiencing constipation had a mean score 6.32 points higher (95% CI 2.35,10.29) and were over twice (OR 2.60, 95% CI 1.45,4.62) as likely to have a high t score than those without constipation. Similarly, compared to the low emotional and disruptive behaviour group, those whose emotional and disruptive behaviour scores were in the high emotional and disruptive behaviour group had a mean score 7.24 points higher (95% CI 2.75,11.73) and nearly three times as likely (OR 2.91, 95% CI 1.60,5.30) to have a high t score.

Table 2.

Descriptive statistics of selected factors of the sleep disturbance scale for children (SDSC) in 258 individuals with CDD

	n (%) or median (interquartile range; range), n
DIMS scaled score	45.7 (37.1,57.1;20.0–97.1.0.1),258
DIMS t score	69.6 (60.0,82.5;40.7–127.5.7.5),258
Minimal difficulties (< = 50)	16 (6.2)
Mild difficulties (50–70)	123 (47.7)
Moderate difficulties (70–90)	76 (29.5)
Severe difficulties (≥90)	43 (16.7)
DOES-mod scaled score	40.0 (33.3,60.0;20.0–93.3.0.3),257
SBD scaled score	26.7 (20.0,40.0;20.0–100.0.0.0),257
SBD t score	51.6 (44.7,65.4;44.7–127.4.7.4),256
Minimal difficulties (< = 50)	96 (37.2)
Mild difficulties (50–70)	114 (44.2)
Moderate difficulties (70–90)	33 (12.8)
Severe difficulties (≥90)	13 (5.0)
Missing	2 (0.8)
SWTD-mod scaled score	44.0 (32.0,52.0;20.0–100.0.0.0),258

Note: The raw subscale scores underwent a linear transformation to convert them to a standardised scale (i.e., scaled score) ranging from 20 to 100, with a higher score indicating more frequent sleep disturbance (i.e., worse sleep) within that specific domain. The t scores were calculated using normative data from Bruni [6] for the DIMS and SBD factors only. Two items from the DOES (“Unable to move when waking up in the morning” and “Awakes in the morning feeling tired”) and one item from the SWTD (“Experiences vivid dream-like scenes while falling asleep”) were excluded as potentially inappropriate for children with CDD. These domains were therefore designated as DOES-mod and SWTD-mod

DIMS Difficulty initiating and maintaining sleep, DOES Disorders of excessive somnolence, SBD Sleep Breathing Disorders, SWTD Sleep-Wake Transition Disorders, n number of individuals

Table 3.

Unadjusted association between SDSC factors and explanatory variables in 212* individuals with CDD aged 3 years or above

	SDSC factors
	DIMS scaled score	DIMS t score High v. Low#	DOES-mod scaled score	SBD scaled score	SBD t score High v. Low#	SWTD-mod scaled score
Explanatory variables	Mean difference (95% CI), P value or OR (95% CI), P value
Sex
Female	Ref	Ref	Ref	Ref	Ref	Ref
Male	−1.20 (−6.78,4.37),0.67	0.73 (0.35,1.54),0.41	1.96 (−4.50,8.42),0.55	2.60 (−2.55,7.76),0.32	1.15 (0.44,3.03),0.78	−0.03 (−5.66,5.61),0.99
Age at completion of questionnaire, in years
3–5	Ref	Ref	Ref	Ref	Ref	Ref
6–10	−5.18 (−10.18,−0.17),0.04	0.60 (0.30,1.20),0.15	0.21 (−5.36,5.78),0.94	−3.20 (−7.84,1.44),0.18	0.51 (0.18,1.51),0.23	−5.19 (−10.15,−0.24),0.04
11–17	3.38 (−2.39,9.15),0.25	1.55 (0.74,3.25),0.25	5.97 (0.20,11.73),0.04	0.77 (−4.79,6.33),0.79	1.58 (0.62,4.01),0.34	−2.12 (−7.40,3.16),0.43
≥18	−8.00 (−14.46,−1.55),0.02	0.42 (0.16,1.10),0.08	2.59 (−5.77,10.94),0.54	1.66 (−5.91,9.23),0.67	1.62 (0.52,5.04),0.40	−9.95 (−15.54,−4.37),<0.01
Variant group
No functional protein	Ref	Ref	Ref	Ref	Ref	Ref
Missense/in-frame variants within catalytic domain	−2.82 (−8.71,3.06),0.35	0.81 (0.39,1.69),0.58	−0.96 (−7.47,5.54),0.77	2.43 (−2.66,7.51),0.35	1.51 (0.55,4.17),0.42	1.37 (−4.32,7.07),0.64
Truncating variants between aa172 and aa781	−1.90 (−7.44,3.64),0.50	1.01 (0.48,2.12),0.98	−2.73 (−9.21,3.76),0.41	1.59 (−3.32,6.50),0.53	1.13 (0.39,3.29),0.82	3.63 (−1.63,8.88),0.18
Truncating variants after aa781	−3.11 (−11.43,5.20),0.46	0.69 (0.26,1.88),0.47	−0.70 (−9.53,8.13),0.88	0.40 (−5.90,6.70),0.90	0.96 (0.23,4.12),0.96	−2.31 (−8.18,3.57),0.44
Other variants	3.93 (−4.68,12.55),0.37	3.96 (0.99,15.86),0.05	2.53 (−7.01,12.07),0.60	8.83 (−0.24,17.89),0.06	1.75 (0.39,7.89),0.46	−0.88 (−8.43,6.68),0.82
Floor sitting/Walking 10 steps
Able to sit and walk	Ref	Ref	Ref	Ref	Ref	Ref
Able to sit but unable to walk	−0.22 (−6.15,5.72),0.94	1.10 (0.52,2.30),0.80	2.11 (−3.60,7.82),0.47	2.94 (−0.33,6.22),0.08	6.57 (0.81,53.12),0.08	−0.50 (−5.62,4.63),0.85
Unable to sit and walk	0.44 (−4.95,5.84),0.87	1.25 (0.61,2.56),0.54	4.52 (−1.11,10.15),0.12	8.96 (4.81,13.12),<0.01	15.10 (1.97,115.92),0.01	−1.96 (−6.96,3.04),0.44
Seizure frequency
Seizure free/Monthly or less/Weekly	Ref	Ref	Ref	Ref	Ref	Ref
Daily (1–4)	−7.31 (−12.03,−2.58),<0.01	0.59 (0.31,1.13),0.11	0.90 (−4.46,6.27),0.74	2.14 (−1.54,5.83),0.25	1.58 (0.59,4.26),0.36	−5.50 (−9.84,−1.15),0.01
Daily (≥5)	−1.18 (−6.30,3.95),0.65	1.32 (0.67,2.59),0.42	4.41 (−1.05,9.87),0.11	9.14 (3.96,14.32),<0.01	3.62 (1.43,9.16),0.01	−2.57 (−7.40,2.25),0.30
Patterns of occurrence of seizures during sleep
During daytime sleep
No	Ref	Ref	Ref	Ref	Ref	Ref
Yes	1.11 (−5.02,7.24),0.72	1.14 (0.49,2.66),0.77	−2.24 (−9.19,4.71),0.53	0.65 (−5.32,6.61),0.83	1.04 (0.33,3.26),0.95	1.24 (−4.42,6.90),0.67
During night
No	Ref	Ref	Ref	Ref	Ref	Ref
Yes	3.93 (−0.51,8.36),0.08	1.43 (0.80,2.54),0.23	−0.04 (−4.62,4.54),0.99	−0.21 (−4.14,3.73),0.92	0.71 (0.31,1.61),0.41	6.08 (1.77,10.39),0.01
Count of antiseizure medications currently used
None or 1	Ref	Ref	Ref	Ref	Ref	Ref
2–4	0.38 (−4.27,5.04),0.87	1.11 (0.59,2.06),0.75	2.09 (−3.07,7.24),0.43	0.48 (−3.51,4.46),0.81	1.03 (0.43,2.46),0.95	0.84 (−3.46,5.14),0.70
≥5	0.36 (−8.31,9.04),0.93	1.50 (0.36,6.19),0.58	7.38 (−6.43,21.19),0.29	18.46 (2.17,34.76),0.03	4.70 (1.03,21.35),0.05	0.27 (−10.94,11.49),0.96
Sleep related side-effects of antiseizure medications
No	Ref	Ref	Ref	Ref	Ref	Ref
Yes	2.15 (−2.13,6.42),0.33	1.85 (1.04,3.32),0.04	5.69 (0.86,10.52),0.02	−0.06 (−4.22,4.09),0.98	0.71 (0.31,1.61),0.41	0.93 (−3.07,4.92),0.65
Constipation – Current
No	Ref	Ref	Ref	Ref	Ref	Ref
Yes	6.32 (2.35,10.29),<0.01	2.60 (1.46,4.62),<0.01	0.51 (−4.25,5.27),0.83	0.03 (−4.15,4.21),0.99	0.86 (0.41,1.82),0.70	4.68 (0.77,8.59),0.02
Emotional and Disruptive Behaviours
Low (< 8)
High (≥8)	7.24 (2.75,11.73),<0.01	2.91 (1.60,5.30),<0.01	7.70 (2.86,12.54),<0.01	−1.21 (−5.02,2.60),0.53	1.16 (0.54,2.50),0.71	8.06 (3.87,12.25),<0.01

Note: bold text = P value <0.05. The raw subscale scores underwent a linear transformation to convert them to a standardised scale (i.e., scaled score) ranging from 20 to 100, with a higher score indicating more frequent sleep disturbance (i.e., worse sleep) within that specific domain. The t scores were calculated using normative data from Bruni [6] for the DIMS and SBD factors only. Two items from the DOES (“Unable to move when waking up in the morning” and “Awakes in the morning feeling tired”) and one item from the SWTD (“Experiences vivid dream-like scenes while falling asleep”) were excluded as potentially inappropriate for children with CDD. These domains were therefore designated as DOES-mod and SWTD-mod

n number of individuals, CI confidence interval, OR odds ratio, Ref reference category

* complete case analysis

^# High was defined as t score ≥70, indicating high sleep disturbance

Disorders of excessive somnolence (DOES-mod)

The median scaled score was 40 (IQR 33.3–60.0). The mean score was 5.97 (95% CI 0.20,11.73) points higher in those aged 11–17 years than in those aged 3–5 years and was 4.52 points higher in those unable to sit and walk. Compared with those with no or one seizure daily, those with 5 or more seizures daily had a mean score of 4.41 (95% CI −1.05,9.87) points higher. The children of those caregivers who had independently expressed that their child’s ASM was associated with sleep-related side effects had a mean score of 5.69 (95% CI 0.86,10.52) points higher than those whose caregivers did not. Similarly, those whose emotional and disruptive behaviour scores were in the high group had a mean score of 7.7 (95% CI 2.86,12.54) points higher than those in the low group.

Sleep breathing disorders (SBD)

The median scaled score was 26.7 (IQR 20.0–40.0), and the mean score was 8.96 (95% CI 4.81,13.12) points higher in those unable to sit and walk compared with those who could and this group was more than 15 times (OR 15.10, 95% CI 1.97,15.92) as likely to have a high t score than those who were able to sit and walk independently. Compared with those who had none or one seizure daily, those who had ≥ 5 seizures daily had a mean score 9.14 (95% CI 3.96,14.32) points higher and were nearly four times (OR 3.62, 95% CI 1.43,9.16) as likely to have a high t score. Those currently taking 5 or more ASMs had a mean score 18.46 (95% CI 2.17,34.76) points higher and were nearly five times (OR 4.70, 95% CI 1.03–21.35) as likely to have a high t score than those on none or one ASM.

Sleep wake transition disorders (SWTD-mod)

The median scaled score was 44.0 (IQR 32.0–52.0), and compared with those aged 3–5 years, the mean score was 5.19 (95% CI −10.15,−0.24) points lower in those aged 6–10 years and 9.95 points lower (95% CI −15.54,−4.37) in those aged 18 years. Compared with those with none or one seizure daily, those with one to four seizures daily had a mean score of 5.50 points lower (95% CI −9.84,−1.15). On the other hand, those who frequently experienced seizures during sleep at night had a mean score of 6.08 (95% CI 1.17,10.39) points higher than those who did not. Moreover, those with constipation had a mean score of 4.68 (95% CI 1.17,10.39) points higher than those without while those whose emotional and disruptive behaviour scores were in the high group had a mean score of 8.06 (95% CI 3.87,12.25) points higher than those in the low group.

Multivariate analysis

The multivariate analyses, as shown in Table 4, produced some changes but the effect sizes for the DIMS were minimally altered with the mean scaled score 6.49 points lower (95% CI −12.44,−0.55) in those aged 18 years compared with those aged 3–5 years. Those with 1–4 seizures daily still had a DIMS mean score of over 6 (6.65) points lower (95% CI −11.52,−1.78) than those who were seizure-free or had less than daily seizures. Similarly, those currently experiencing constipation were over twice (OR 2.39, 95% CI 1.24,4.62) as likely to have a high DIMS t score than those without constipation. Moreover, those whose emotional and disruptive behaviour scores were in the high group were still three times as likely (OR 2.96, 95% CI 1.42,6.16) to have a high DIMS t score.

Table 4.

Adjusted^{^} association between SDSC factors and explanatory variables in 212* individuals with CDD

	SDSC factors
	DIMS scaled score	DIMS t score High v. Low#	DOES-mod scaled score	SBD scaled score	SBD t score High v. Low#	SWTD-mod scaled score
Explanatory variables	Mean difference (95% CI),P-value	OR (95% CI),P-value	Mean difference (95% CI),P-value	Mean difference (95% CI),P-value	OR (95% CI),P-value	Mean difference (95% CI),P-value
Sex
Female	Ref	Ref	Ref	Ref	Ref	Ref
Male	−2.83 (−8.09,2.43),0.29	0.43 (0.17,1.10),0.08	1.56 (−4.16,7.29),0.59	−0.15 (−5.28,4.99),0.95	0.89 (0.29,2.74),0.83	1.28 (−4.24,6.80),0.65
Age at completion of questionnaire, in years
3–5	Ref	Ref	Ref	Ref	Ref	Ref
6–10	−4.14 (−8.94,0.66),0.09	0.66 (0.30,1.46),0.30	1.75 (−3.40,6.90),0.51	−2.00 (−6.03,2.03),0.33	0.71 (0.21,2.39),0.58	−3.74 (−8.59,1.11),0.13
11–17	3.42 (−1.86,8.70),0.20	1.60 (0.69,3.73),0.28	7.31 (1.62,13.01),0.01	3.12 (−2.07,8.30),0.24	3.13 (1.06,9.28),0.04	−2.71 (−7.64,2.21),0.28
≥18	−6.49 (−12.44,−0.55),0.03	0.54 (0.18,1.67),0.29	4.23 (−3.00,11.46),0.25	3.27 (−3.19,9.73),0.32	2.30 (0.60,8.80),0.22	−8.43 (−13.90,−2.96),<0.01
Variant group
No functional protein	Ref	Ref	Ref	Ref	Ref	Ref
Missense/in-frame variants within catalytic domain	−4.16 (−9.81,1.49),0.15	0.64 (0.28,1.48),0.30	−0.36 (−6.17,5.44),0.90	2.21 (−2.10,6.52),0.32	1.82 (0.57,5.84),0.32	0.89 (−4.17,5.94),0.73
Truncating variants between aa172 and aa781	−2.65 (−8.11,2.80),0.34	0.86 (0.36,2.05),0.74	−2.53 (−8.02,2.95),0.37	1.55 (−2.58,5.68),0.46	1.39 (0.40,4.81),0.61	3.82 (−1.13,8.76),0.13
Truncating variants after aa781	−1.83 (−9.55,5.89),0.64	0.68 (0.22,2.13),0.51	0.14 (−7.68,7.96),0.97	1.37 (−4.69,7.43),0.66	1.60 (0.31,8.35),0.57	−0.65 (−6.06,4.76),0.81
Other variants	6.83 (−1.68,15.33),0.12	7.16 (1.56,32.79),0.01	2.64 (−6.45,11.73),0.57	8.25 (0.81,15.68),0.03	2.42 (0.42,13.96),0.32	−0.54 (−8.15,7.06),0.89
Floor sitting/Walking 10 steps
Able to sit and walk	Ref	Ref	Ref	Ref	Ref	Ref
Able to sit but unable to walk	−1.08 (−6.14,3.98),0.68	0.82 (0.33,2.08),0.68	1.71 (−4.02,7.43),0.56	3.56 (0.02,7.09),0.05	8.63 (0.97,77.14),0.05	−1.61 (−6.33,3.10),0.50
Unable to sit and walk	4.17 (−0.92,9.27),0.11	1.88 (0.73,4.84),0.19	6.92 (0.40,13.44),0.04	7.48 (3.42,11.54),<0.01	23.40 (2.54,215.76),0.01	0.21 (−4.48,4.90),0.93
Seizure frequency
Seizure free/Monthly or less/Weekly	Ref	Ref	Ref	Ref	Ref	Ref
Daily (1–4)	−6.65 (−11.52,−1.78),0.01	0.67 (0.30,1.48),0.32	1.15 (−4.58,6.89),0.69	0.33 (−3.12,3.78),0.85	1.16 (0.35,3.86),0.81	−3.16 (−7.47,1.15),0.15
Daily (≥5)	−2.61 (−7.58,2.36),0.30	1.22 (0.53,2.79),0.64	2.11 (−3.50,7.71),0.46	7.09 (2.39,11.79),<0.01	2.64 (0.85,8.18),0.09	−2.55 (−7.08,1.98),0.27
Patterns of occurrence of seizures during sleep at night
No	Ref	Ref	Ref	Ref	Ref	Ref
Yes	2.68 (−1.51,6.86),0.21	1.47 (0.74,2.93),0.27	−0.07 (−4.85,4.71),0.98	2.04 (−1.70,5.79),0.28	0.94 (0.36,2.48),0.90	4.09 (−0.15,8.32),0.06
Count of antiseizure medications currently used
0–1	Ref	Ref	Ref	Ref	Ref	Ref
2–4	2.07 (−2.20,6.34),0.34	1.26 (0.61,2.59),0.54	0.75 (−3.99,5.50),0.76	−0.78 (−4.48,2.93),0.68	0.95 (0.34,2.69),0.93	2.14 (−1.82,6.10),0.29
≥5	1.89 (−5.83,9.61),0.63	1.27 (0.22,7.21),0.79	5.37 (−6.37,17.11),0.37	12.75 (0.47,25.02),0.04	3.49 (0.53,22.86),0.19	3.34 (−6.18,12.86),0.49
Sleep related side-effects of antiseizure medications
No	Ref	Ref	Ref	Ref	Ref	Ref
Yes	1.44 (−2.51,5.39),0.47	2.00 (1.01,3.94),0.05	5.34 (0.81,9.86),0.02	−1.10 (−4.94,2.73),0.57	0.50 (0.20,1.29),0.15	0.73 (−3.19,4.66),0.71
Constipation – Current
No	Ref	Ref	Ref	Ref	Ref	Ref
Yes	4.17 (0.35,8.00),0.03	2.39 (1.24,4.62),0.01	−1.90 (−6.54,2.73),0.42	−0.80 (−4.73,3.13),0.69	0.67 (0.27,1.63),0.37	3.17 (−0.61,6.95),0.10
Emotional and Disruptive Behaviours
Low (< 8)	Ref	Ref	Ref	Ref	Ref	Ref
High (≥8)	5.10 (0.87,9.33),0.02	2.96 (1.42,6.16),<0.01	10.84 (5.68,16.01),<0.01	0.42 (−2.99,3.83),0.81	2.38 (0.84,6.74),0.10	6.43 (2.03,10.83),<0.01

Note: bold text = P -value <0.05. The raw subscale scores underwent a linear transformation to convert them to a standardised scale (i.e., scaled score) ranging from 20 to 100, with a higher score indicating more frequent sleep disturbance (i.e., worse sleep) within that specific domain. The t scores were calculated using normative data from Bruni [6] for the DIMS and SBD factors only. Two items from the DOES (“Unable to move when waking up in the morning” and “Awakes in the morning feeling tired”) and one item from the SWTD (“Experiences vivid dream-like scenes while falling asleep”) were excluded as potentially inappropriate for children with CDD. These domains were therefore designated as DOES-mod and SWTD-mod

n number of individuals, CI confidence interval, OR odds ratio, Ref reference category

^{^} adjusted for all explanatory variables

^# High was defined as t score ≥70, indicating high sleep disturbance

* complete case analysis

For the DOES-mod, the effect of functional ability (inability to sit and walk) was greater in the multivariate analysis with a mean score 6.92 (95% CI 0.40,13.44) points higher in those unable to sit and walk compared with those able to walk, while those aged 11–17 years had a mean score 7.31 (95% CI 1.62,13.01) points higher than those aged 3–5 years. The magnitude of the association with emotional and disruptive behaviours increased so that those whose emotional and disruptive behaviour scores were in the high group had a mean DOES-mod score of 10.84 (95% CI 5.68,16.01) points higher than those in the low group.

For the SBD, the relationships with functional abilities, seizure frequency and number of ASMs generally persisted. In terms of functional ability, those who could not sit nor walk were more than 23 times (OR 23.40, 95% CI 2.54,215.76) as likely to have a high t score than those who were able to sit and walk independently.

For the SWTD-mod, the effects of age persisted. Compared with those aged 3–5 years, the mean SWTD-mod score was 8.43 points lower (95% CI −13.90,−2.96) in those aged 18 years. The effect of seizure frequency became less obvious a

nd the effect of having night-time seizures a little less with a mean score of 6.08 (95% CI 1.17,10.39) points higher in those who frequently experienced seizures during sleep at night than those who did not. The effect of constipation was also attenuated, although the relationship with emotional and disruptive behaviour was retained with those whose emotional and disruptive behaviour scores were in the high group having a mean score of 6.43 (95% CI 2.03,10,83) points higher than those in the low group.

Open question

Caregivers were asked to respond to the prompt, “Is there anything about your child’s sleeping that you consider unusual or problematic?”, and if so to elaborate further about this. Their responses spanned from single small sentences to paragraphs of information about their child’s sleep habits. Five broad categories (summarised in Table 5) were identified as described below.

Table 5.

Categories and sub-categories identified for sleep challenges in individuals with CDD

Categories	Sub-categories
Insomnia	• Sleep initiation insomnia • Slight maintenance insomnia • “Night Parties” and disruption to circadian rhythm
Daytime somnolence	• Impact of daytime somnolence
Effect of seizures on sleep	• Seizures causing insomnia • Post-ictal somnolence • Fatigue causing seizures
Effect of comorbidities on sleep	• Spontaneous breathing disorders • Gastrointestinal discomfort
Medication	• Melatonin  - Beneficial association with sleep  - No beneficial association with sleep • Other medications

Insomnia

A considerable number of caregivers chose to elaborate on their experience of insomnia with their child. Caregivers describe issues with sleep initiation, “Some nights she does not sleep at all and stays awake into early morning hours.”, as well as trouble maintaining sleep, “she doesn’t go into a deep sleep. Any small noise will wake her. She is a terrible sleeper”. Many of the caregivers also elaborated on the difficulty experienced in both sleep initiation and sleep maintenance, “she is a very light sleeper the first 2 hours of sleep and wakes easily. Sometimes she cannot go back to sleep for several hours (3 to 6 hours)”.

Another predominant sub-theme within the experience of insomnia was the caregivers’ recounts of “night parties”. Caregivers recounted days and nights of wakefulness, “I can’t remember the last time he slept through the entire night. Some days he has what we call CDKL5 Nights or diaper parties”. Caregivers often described periods of hyperactivity associated with these nights, “.at least 1 to 2 times a week she will stay awake, happy and very, very energetic for a full 24 to 36 hours straight (sometimes more)” with caregivers often referring to this phase as the “CDKL5 nights” or that their child is having “a party in [their] crib”.

Daytime somnolence

Caregivers also chose to elaborate in detail on their experience with daytime somnolence. Some caregivers reported difficulty rousing their child, “we wake her in the mornings, she stays asleep during breakfast while we feed her and she eats beautifully. She will remain asleep until she’s at school.” Others reported that their child had many naps during the day, “her night sleep is very disturbed, so she makes up with naps in the day time.” Similar to the link between insomnia and a disrupted circadian rhythm, daytime somnolence is often experienced as a result of a disrupted circadian rhythm. Caregivers describe a period of extreme wakefulness followed by excessive daytime somnolence, “She has had numerous occasions where she would not go to sleep at all. She would be awake (happy and playing) for 24 to 36 hours straight. She would finally have a seizure and them want to sleep for the next 24 to 36 hours straight.”

Effect of seizures on sleep

Responses from caregivers described how seizures had a considerable impact on sleep. Occurrence of seizures was found to contribute to insomnia (sleep initiation and sleep maintenance), post-ictal fatigue and daytime somnolence. One caregiver reported that their child experienced “seizure activity [that] can cause her to not sleep for a full 48 hours. Then her sleep cycle is off and this is a problem”, illustrating that the seizure activity disrupted the initiation of sleep. Another caregiver reported issues associated with sleep initiation and maintenance insomnia for their child, “she needs longer to fall to sleep… In the night, she wakes up because of epileptic attacks, between 5–10 times.” Another caregiver outlined that, “myoclonic jerks” prevent their child from falling asleep.

Post-ictal somnolence is experienced broadly in many epileptic encephalopathies, and individuals with CDD appeared to be no exception. Caregivers reported an association between seizure occurrence and subsequent daytime somnolence, “if she has a seizure upon waking up, she may go back to sleep for as much as 2–3 hours” and that if the individual was going through a period of time with increased seizure frequency, then, “he will sleep all day.”

Some caregivers identified that seizures were often necessary for the initiation of sleep, outlining that it would take a, “large seizure to put [their child] to sleep” or that they would have a seizure and then sleep for “24–36 hours straight” following a period of hyperactive insomnia.

As previously mentioned, “CDKL5 nights” were experienced frequently by caregivers and individuals with CDD. Caregivers linked periods of prolonged hyperactive wakefulness with subsequent increased seizure activity. One caregiver has summarised this experience, “he is staying awake all night long for up to 3 days, which leads to having seizures and being in bad form.”

Effect of comorbidities on sleep

One caregiver wrote that they, “thought it was a seizure waking him up, but it was the apnoea… they startle him awake”. Some caregivers just wrote “sleep apnoea” or “periods when they stop breathing” as their response. Caregiver responses also indicated that gastrointestinal disturbances such as gastrointestinal reflux and constipation were impacting sleep. One caregiver reported that, “her reflux will wake her up and trigger seizures”. Another caregiver recounted her child’s experience where, “[the] build-up of gas while sleeping…[leads to] a seizure, and it goes on until she can release the gas”.

Medications

Some caregivers expressed reliance on melatonin for their child’s sleep, “.she will not go to sleep without melatonin. She might stay awake the whole night without it”. Others found melatonin to be ineffectual or variable in its response, “melatonin works most of the time but also will not put her to sleep”, “Tried Melatonin to promote sleep, seemed to help initially now have doubled the dosage (4 mg) and not working at all and made him very cranky the following day”.

There were a variety of medications that caregivers reported to have a beneficial effect on their child’s sleep. These included medications within the class of benzodiazepines, anti-epileptics (levetiracetam, phenobarbital), alpha-2-receptor agonists, anti-depressants (mirtazapine), sedatives (chloral hydrate), with some caregivers just stating “medications” that have observed benefit in their child’s sleep. Medications that caregivers reported to have a negative impact on sleep included anti-epileptics (lamotrigine), steroids, and with most stating “sleeping medication” or “sedative” and its associated negative effect on their child’s sleep.

Discussion

This study sought to obtain a quantitative picture of the sleep disturbances experienced by individuals with CDD using domains of the Sleep Disturbance Scale developed by Bruni [6]. This was complemented by insights from a content analysis of caregiver responses to an open question where caregivers elaborated on anything unusual or problematic they had noticed about their child’s sleep. The major themes identified by caregivers in the open question included the experience of insomnia, daytime somnolence, the timing and frequency of seizures affecting sleep habits and patterns, comorbidities associated with CDD, and medication use.

Insomnia was originally identified as a comorbidity among individuals with CDD, affecting over 80% of individuals to some extent [5]. In the current study, the median DIMS t score was 69.6 and approximately 45% of those aged 3 years and over would have moderate or severe difficulties with insomnia. This compares with 53% in a previous smaller study (n = 129) investigating the relationship between sleep disturbances and quality of life in CDD and unlike the current one, the previous study was restricted only to follow-up data such that the cohort would be very slightly older [8]. In the present study the scaled score was much lower in those aged 18 years or above than in those aged 3–5 years, suggesting that this problem may improve in adulthood although it was worse in teenagers than in those younger. This observation has to be tempered by the fact that the SDSC has not been validated for individuals over age 18 years and recent evidence indicates that five of the seven insomnia items give a the best model fit for CDD [25].

Most caregivers reported in the open text either sleep initiation, maintenance, or both forms of insomnia experienced either currently or in the past. Many outlined that their child was often unsettled during periods of insomnia. Poor sleep among individuals with CDD has been shown to be associated with poorer caregiver emotional health [9]. In this study many caregivers made mention of how their lives were often controlled by the uncertainty of the sleep cycle. In particular, those who wrote about their child’s “night parties” elaborated on how this had a negative effect on their personal and family unit well-being, mostly due to the sleep deficit they themselves experienced. Whilst sleep difficulties are a common occurrence amongst children with intellectual disability, a “night party” disruption to the sleep cycle of this extent has rarely been identified in other disorders [26].

Unlike the DIMS we did not find that the scores for DOES-mod (Disorder of Excessive Sleepiness) improved in adulthood. However, we did find that the scores were higher in those who could not walk independently, especially in the adjusted analysis, consistent with analyses in individuals with other intellectual disabilities [16]. We also found that these scores were higher in those whose parents had reported side effects of ASMs that were sleep-related suggesting that the use of some ASMs has a role to play in the high levels of daytime sleepiness in this disorder. It is also possible that individuals who are more severely affected by the disorder in terms of physical functioning are more likely to experience daytime sleepiness. Some caregivers also reported that these children often slept during meal times and social interactions. e.g., during school and family time, and that they were therefore more dependent on their carer due to this excessive somnolence. These findings would be consistent with the associations we previously found between sleep disturbance and quality of life both in the physical health (7) and negative emotions domains [7, 8].

This current study identified that many caregivers reported their children having seizures either prior to bed, during sleep, or upon waking. Caregivers wrote about how during periods of increased seizure activity, their child experienced more restlessness and irritability. Many also linked the frequency and timing of seizures to their child’s sleep difficulties, a relationship previously described in the general literature where epilepsy was associated with a twelve times higher rate of sleep difficulties than in the general population [13]. A more recent CDD-specific study has investigated this relationship, suggesting that early in life in CDD spasms may start in slow-wave or quiet sleep with features mimicking sleep terrors [20]. The quantitative data also identified that those with five or more seizures per day had higher scores mainly in the SBD but not in the DIMS, as did those who were taking five or more antiseizure medications (ASMs).

The SWTD domain comprises items that describe symptoms that occur during the transition between wakefulness and sleep or from one sleep stage to another. Items are designed to measure hypnic jerks, rhythmic movement disorder, vivid dreamlike hallucinations, nocturnal hyperkinesia, sleeptalking and bruxism [6]. We excluded the item describing dreamlike hallucinations because this cannot be evaluated in children with CDD. However, it could be that some items capture aspects of epilepsy which influences sleep, rather than a SWTD. Consistently, we found that individuals with frequent seizures during sleep at night had SWTD-mod scores approximately four points higher than those for whom the caregiver did not report epilepsy during sleep. This raises the important question of what the SWTD domain questions measure in CDD. Parents of children with Dravet syndrome, another developmental epileptic encephalopathy, have criticized available sleep questionnaires for not querying epilepsy during sleep [27]. Further, the case series of infants with CDD observed clusters of spasms during quiet sleep that mimicked a sleep terror and were often responsive to an ASM [20]. A sleep video-EEG or polysomnographic study would be required to investigate epilepsy during sleep for appropriate diagnoses and management.

Another theme identified amongst the caregivers was the significance of comorbidities and sleep, specifically spontaneous breathing disorders and gastrointestinal issues. In one small Dutch case series (n = 4) a sleep breathing disorder was identified in one patient [28]. However, to our knowledge there has been no further specific research in this area. Sleep breathing disorders can often be managed successfully using CPAP or by adenoidectomy and therefore it is important for them to be identified. We found that the likelihood of a sleep breathing disorder was increased in those who were unable to sit up independently, in those who were having multiple (>5) seizures per day and/or in those who were on multiple (>5) ASMs per day. This would suggest that those affected by a sleep breathing disorder, although possibly small in number accounting for ~ 11% of the total population under study, are among those with the greatest severity.

Gastrointestinal disturbance leading to sleep difficulties was another common theme identified by caregivers. Caregivers in this study in open text frequently attributed the sleep disturbance to gastroesophageal reflux and constipation. We also found quantitatively that the presence of constipation was associated with higher DIMS scores and to a lesser extent SWTD-mod scores. Effective management of these comorbidities may contribute to improved sleep quality and well-being.

Although there is some evidence of an association between sleep disturbance and quality of life in CDD [7, 8], literature on the behavioural phenotype of CDD is lacking and, although we can hypothesise that there is an association with sleep disturbances the quantitative evidence is absent. A precedent for investigating this relationship is our research on Rett syndrome where we found a relationship between both insomnia and excessive sleepiness and high anxiety scores [29]. On this account we felt it was worth also considering whether behavioural disturbance might be a factor in this study. In the absence of a validated instrument for measuring behaviour in CDD, we decided in this case to use available items from the newly developed RSBQ domain for assessing ‘emotional and disruptive behaviour’ in Rett syndrome [23]. As might be expected we found that there were clear relationships between high”behaviour” scores and higher DIMS and DOES-mod scores. What we do not know is the direction of this relationship.

Consistent with previous literature [8, 20], reference was commonly made to melatonin by caregivers in their open text. Reported effects on their child’s sleeping habits were varied, with some caregivers reporting that it has been the mainstay of consistent sleep cycles and others claiming that it has no effect. Although there has been a recent review of melatonin use in children with autism and neurogenetic disorders its findings may not be totally applicable to the CDD population where behavioutal management techniques could be challenging [30].

It is also unclear whether the daytime somnolence is a sequelae of insomnia (e.g.during the previous night), medication-induced or another unknown factor associated with CDD. When asked to describe benefits and side-effects of particular ASMs we found that the children of those who had specifically commented on the presence of sleep-related side effects had much higher DOES-mod scores. This would suggest that ASM use does contribute somewhat to daytime sleepiness and does warrant further investigation in a condition where sleep disturbance is already so prevalent.

Strengths and limitations

The sample size of this study from a quantitative perspective is a significant benefit. Combined with the textual analysis of the open question, it has allowed us to gather broad information about the experience and factors affecting sleep among individuals with CDD. However, the content analysis was limited in that it was based on responses to one open question that could have lacked depth in caregiver responses as the ability to probe for further elaboration by the caregivers was not possible. Moreover, caregivers who responded to the open question were more likely to be experiencing issues regarding their child’s sleeping habits, resulting in potential bias. Further qualitative studies involving in-depth interviews would allow exploration of the themes identified here and their impact on the child and the family. Regarding the quantitative analysis, this was based on four domains: DIMS; DOES-mod; SBD and SWTD-mod. For reasons similar to the exclusion of items from the DOES and SWTD domains two of the three questions relating to the DA domain were not administered to families and therefore, as has been done in some other fairly recent literature [8, 16, 18, 25, 31], we did not report a total sleep disturbance score. We acknowledge concerns about the validity of measures which have been modified from their original format. However, our modifications were weighed up against the inclusion of items clearly not applicable to children with as severe a disorder as CDD and the impact of such inclusion on the caregiver respondents and at the least, items in scales need to be relevant and clear to the target population [32]. As mentioned in this helpful primer relating to development of sleep questionnaires for children [33] it is important to take into account the target population and clearly the SDSC was not specifically designed for this population of children with profound functional impairment-hence our minor modifications. It was also designed for the purpose of screening for different types of sleep disturbances [33] such that it would seem reasonable as we did to evaluate individual domains. Our questionnaire study of 258 individuals who live in different regions of the world forms part of a natural history study and it was not practical or feasible to undertake a clinical review to confirm SDSC findings as could be done in clinical practice or assess test-retest reliability. We acknowledge that sleep medicine for severe neurodevelopmental disorders is an emerging field and our observations will inform efforts to ensure that sleep measurement scales with good content validity are available for this group.

Caregivers’ ability to accurately recall and report information may vary, potentially leading to inaccuracies. For information about sleep apnoea we were dependent on information provided in relation to the question about a sleep study being performed and there may have been other treatments provided in addition to what we reported about CPAP technology. We did not specifically ask a question about treatments provided or used for sleep apnoea and this could be considered a limitation. Nevertheless, caregiver-reported data still remains a valuable source, especially in contexts where self-reporting is impractical. These analyses used available SDSC items rather than basing item selection on factor analyses using CDD data which has recently identified different scoring methods [25]. Where there were missing data items i.e. in the DOES-mod and SWTD-mod domains we were not able to calculate a t score which allows a comparison with the Italian normative data. However in the absence of a t score we were able to evaluate relationships with the various covariates using the scaled score we created. For the domains where we were able to measure the t score in addition to the scaled scores the results showed a congruence validating the use of the scaled scores. Despite our original rationale for opting not to include all items, we suggest that future studies should consider using the items suitable for CDD as evaluated using confirmatory factor analysis [25].

Missing data, particularly if non-random, can also lead to bias. While we have employed a simple imputation method to mitigate the effects of missing outcome data, more sophisticated approaches for managing missing data may become available for future analysis. These advanced techniques could address the limitations of variability reduction of simple imputation and the inherent drawbacks of complete case analysis, ensuring a more robust and reliable analysis.

Conclusion

This study described sleep difficulties and associated factors among individuals with CDD. The features identified included insomnia (initiation and maintenance of sleep), disrupted circadian rhythms leading to “night parties” and daytime drowsiness affecting daytime functioning. Contributing factors included seizure timing and frequency, comorbid conditions such as gastrointestinal symptoms and medication use. This study will help to inform the clinical management of sleep difficulties among individuals with CDD and guide further study in this important area of CDD research.

Authors’ contributions

ES, HL, JD, MJ and KW contributed to the study’s conception and design. ES undertook the textual analysis. MJ and KW performed data management and statistical analysis and prepared tables. ES, HL, MJ, JS and KW wrote the first draft of the manuscript, and all authors commented on subsequent versions. All authors read and approved the final manuscript.

Funding

Funding for the International CDKL5 Disorder Database has been provided by the International Foundation for CDKL5 Research. Contributions from the Orphan Disease Centre at the University of Pennsylvania (CDKL5-19-D-101-3) assisted with analytical support. JD is supported by a Fellowship from the Stan Perron Charitable Foundation.

Data availability

The dataset analysed during the current study is not publicly available but may be available from the corresponding author on reasonable request following an application to and with approval from the local ethic committee.

Declarations

Ethics approval and consent to participate

The Human Research Ethics Committee of the University of Western Australia provided approval for this study in accordance with the Declaration of Helsinki.

Consent for publication

Written or electronic informed consent was obtained from the caregivers of individuals whose data was used in the study.

Competing interests

HL provided consultancy for Marinus Pharmaceuticals, Acadia Pharmaceuticals, Novartis Gene Therapies and Orion Corporation; and clinical trials with Anavex Life Science and Newron Pharmaceuticals. All consultancies are unrelated to this work, and all remuneration has been made to her department.JD provided consultancy for Marinus Pharmaceuticals, Neurogene, Ultragenyx Pharmaceutical, Acadia Pharmaceuticals, Novartis Gene Therapies, Orion Corporation and Taysha Gene Therapies; and clinicaltrials with Anavex Life Science and Newron Pharmaceuticals. All consultancies are unrelated to this work, andall remuneration has been made to her department.