The AS/HFA group consisted of 32 children (M = 10.8 yrs, range 8.5–12.8), 28 boys and 4 girls. These 32 children were selected out from a total of 122 children with a clinical diagnosis of AS, registered at three PDD-habilitation centres in Stockholm. The initial aim of the present study, which was undertaken 2001–2002, was to assess whether AS in school-age children is associated with disturbed sleep. Therefore, the inclusion criteria were that the child was born in the period 1989–1992 and had a clinical diagnosis of AS. Exclusion criteria were intellectual disability, comorbid physical disability, seizure disorder or long-term medication, since all of these factors are known to have an impact on sleep [16–18]. In the first stage of the selection procedure, after the first review of the medical records of the eligible 122 children, 34 children were excluded: 5 due to epilepsy; 5 due to essential language delay (which is inconsistent with an ICD-10 diagnosis of AS); 4 due to physical disabilities (severe allergic complaints in 2 children, ataxia in one child, and ataxia and a visual impairment in another child); and 20 due to pharmacological treatment (psychostimulants in 10 children, antidepressants in 8, and neuroleptics in 2 children). Notably, the initial review of the medical records did not indicate intellectual disability in any of these children.
In the second stage of the selection procedure, the remaining 88 families were given written invitation to participate in the study. Fifty-one families expressed willingness to take part in the investigation. However, another 19 out of these 51 children were excluded in this stage of the selection procedure: 15 due to current use of psychotropic medication (psychostimulants in 9 children, neuroleptics in 3, and antidepressants in 3 children); and 4 due to suspicion of mental retardation.
To further clarify matters, the initial clinical AS diagnoses of the 32 children were based on comprehensive multidisciplinary assessments, performed on average 40 months prior to the present study by independent clinicians at child psychiatric and paediatric clinics. These assessments had included neuropsychiatric examination, speech and communication testing, and neuropsychological testing. The neuropsychological assessments were performed by use of the Wechsler Intelligence Scale for Children [WISC-R], the Leiter International Performance Scale, the Wechsler Preschool and Primary Scale of Intelligence [WPSSI], or Griffiths' Development Scale. Results of neuropsychological testing had shown that these 32 children were of normal intelligence. In the majority of cases (n = 21), Intelligence Quotient (IQ) was assessed using the WISC-R.
Finally, before entering our study, the 32 participating children were also subjected to a diagnostic reassessment which was made by the first author in order to ensure that these children fulfilled ICD-10 research criteria for AS . The reassessment which was based on structured interviews with parents and children and an additional review of medical records did not lead to any further exclusion of participants. Unexpectedly, however, 13 children (11 boys and 2 girls) with a clinical diagnosis of AS displayed a history of essential language delay and fulfilled ICD-10 research criteria for autistic disorder. These children were rediagnosed as having HFA. Nineteen children (17 boys and 2 girls) fulfilled ICD-10 research criteria for AS. The sampling procedure and the diagnostic reassessment of the PDD sample has been presented elsewhere [15, 20].
The control group, 32 typically developing children (M = 10.9 yrs, range 8.5 -13.4), 28 boys and 4 girls, were matched pairwise with the children in the AS/HFA group with respect to age, gender and residency. The controls were recruited via mainstream schools. In a first step of recruitment procedure, school nurses selected children of suitable age and gender who attended regular classes in mainstream schools and were without mental, developmental, or physical disabilities and long-term medication according to school medical records. In a second stage of recruitment of controls, a school nurse or the first author telephoned the parents of selected children and asked whether the families were willing to participate in the study. An introductory letter was then sent to families who agreed to participate. The controls were not IQ-tested.
A sleep questionnaire, sleep diary and actigraphs and the behavioural screening forms were distributed to all families in conjunction with home visits, and parents conveyed the relevant instruments to their children's teachers. Teachers mailed the completed forms to the first author, and all other instruments were returned to the first author via a home visit, a parental visit to the clinic, or by mail.
Sleep-wake behaviour during the previous six months
A parental paediatric sleep questionnaire previously used in population-based studies of Swedish children [21–23] was utilised for a detailed survey of retrospective sleep-wake behaviour pertaining to the previous six months. Twenty one items (items Q1 – Q21) (see Additional file 1) were categorised according to a 5-point rating scale ("never," "rarely," "once or twice per week," "3 or 4 times per week," and "at least 5 times per week"). The parental paediatric sleep questionnaire used for this study also comprised additional items: a global question regarding whether the child had a current sleeping problem Q22, and questions about the following consequences of that sleeping problem: distress Q23, and impaired daytime functioning Q24 in the child, and a burden for the family Q25.
Definition of insomnia
We used the DSM-IV-adapted criteria for paediatric insomnia, proposed by Glaze et al :
the complaint is significant difficulty (defined by frequency, severity, and/or chronicity) initiating or maintaining sleep. The difficulty is viewed as problematic by the child and/or a caregiver;
the sleep disturbance causes clinically significant impairment in school performance, behaviour, mood, learning, or development, for the child as reported by the child and/or a caregiver;
the sleep disturbance does not occur exclusively in the context of an intrinsic dyssomnia such as narcolepsy, restless legs syndrome, or sleep-related breathing disorders; a circadian rhythm disorder; or a parasomnia; and
the sleep disturbance is not attributable to either the direct physiologic effect of a drug or the abuse or misuse of a prescribed medication.
Recent sleep patterns
As described in a previous report , we used a one-week sleep diary and actigraphy to describe the child's recent sleep patterns. The actigraphs, used in the current study (Actiwatch, Cambridge Neurotechnology, Ltd, Cambridge, UK), were worn on the child's nondominant arm. All movements exceeding the 0.05 g threshold were sampled at the medium sensitivity level with the epoch length 30 seconds, and stored as activity counts per epoch in the Actiwatch's 16-K memory. Data from actigraph memory were thereafter downloaded to the Actiwatch Sleepwatch software . The following seven sleep variables, one from the sleep diary and six from the actigraphy, were selected for the present report: 1) Bed time (the time when the child went to bed according to the parent-report in the sleep diary); 2) Sleep start (the first minute after bedtime that was identified as sleep by the Actiwatch Sleepwatch algorithm, and was followed by at least 10 consecutive minutes of recorded immobility); 3) Sleep latency (the time from bedtime to sleep start); 4) Actual sleep time (the calculated difference between sleep end and sleep start in minutes minus actual time spent awake during the sleep period); 5) Sleep end (the last epoch of immobility before the start of at least 10 minutes of consecutive activity); 6) Actual time awake (the amount of time spent awake as determined by the algorithm); and 7) Sleep efficiency (the percentage of time spent asleep while in bed). All these seven sleep variables were averaged for each child into school day and weekend mean values, using the occurrence of school attendance the next day as the definition of a school day (school day: Sunday, Monday, Tuesday, Wednesday, Thursday; weekend: Friday, Saturday). Thirty children in the AS/HFA group and 32 controls were monitored for seven days, and two children from the AS/HFA group were monitored for six days. Children within each matched pair were monitored by the same actigraphy device. Four different devices were used and the numbers of participants in each group were 9, 7, 8, and 8 pairs of children, respectively.
The High-Functioning Autism Spectrum Screening Questionnaire (ASSQ), a 27-item checklist was used to evaluate the extent of autism-related symptoms in children with AS/HFA and to screen for autism-related symptoms in the control group . Eleven items covering social interaction, 6 items covering communication problems, and 5 items covering aspects of restricted and repetitive behaviour were included. The remaining 5 items embrace motor clumsiness and other associated symptoms, including motor and vocal tics. Parent and teacher ASSQ versions have shown satisfactory test/retest reliability, inter-rater reliability, and validity .
The Strengths and Difficulties Questionnaire (SDQ), a 25-item checklist, was used in order to measure aspects of social competence and psychopathology of the child [26, 27]. The SDQ probes behaviours and psychological attributes reflecting the child's difficulties such as hyperactivity/inattention, emotional symptoms, conduct and peer problems, as well as strengths such as social competence (prosocial behaviour). The Swedish version of parent SDQ has shown satisfactory reliability and validity . Both parent and teacher SDQ ratings were used in the current study.
The analyses were divided into two main parts: 1) comparisons between children in the AS/HFA group (n = 32) and the pairwise matched control group (n = 32); and 2) comparisons between children with (n = 10) and those without insomnia (n = 22) in the AS/HFA group. The frequency of occurrence of parent-reported sleep-wake behaviours (items Q1 – Q21), parent-reported sleeping problems (Q22), and consequent distress and impaired daytime function for the child and burden for the parents (Q23 – Q25) was compared between children in the AS/HFA group and children in the control group. The Wilcoxon Signed Ranks Test was used for these pairwise comparisons. Further, in the AS/HFA group, the frequency of occurrence of coexisting sleep-wake behaviours during the past 6 months was compared between children with and those without insomnia, using ordinal regression. Separate analyses were performed, using the occurrence of each sleep-wake behaviour as dependent and the presence or absence of insomnia as independent variables, while controlling for the age of child. Recent sleep patterns were compared between the children with and those without insomnia in the AS/HFA group using logistic regression. In these analyses, the presence or absence of insomnia and each of the seven sleep variables (school day and weekend means) were entered as dependent and independent variables respectively, while controlling for the age of child and for the actigraphy device (n = 4) . The nominal scale describing the actigraphs was transformed into four index variables. Logistic regression was also conducted to determine the relationship between insomnia and behavioural characteristics. In these analytical operations, the presence or absence of insomnia and the ASSQ-, and SDQ-scores were entered as dependent and independent variables respectively, while controlling for age of the child.
In addition, continuous sociodemographic data (parental age, number of children in the family, parental employment status) were compared between the AS/HFA and control groups by using t tests for paired samples. The Wilcoxon Signed Ranks Test was used for comparisons of categorical sociodemographic data (family status, school situation). Significance level p < 0.05 was regarded as statistically significant, using SPSS .
The study was approved by the Ethical Committee at the Karolinska Hospital, Stockholm, Sweden.