Abstract
Keywords
Recent Australian estimates place the prevalence of attention deficit hyperactivity disorder (ADHD) in children aged six to 17 years as high as 11 per cent, but the prevalence of attention problems alone at six per cent [1]. However, conservative figures out of the US and Great Britain place the rates at between three and five per cent of prepubertal children [2]. In either circumstance, ADHD remains the most common disorder presenting to children's mental health services, and therefore attracts a great deal of media, government and research interest.
A range of factors have been proposed as playing a causal role in the pathogenesis of ADHD, including genetic, environmental and psycho-social factors [3]. However, despite vast research into this area, there remains no proven unified theory as to the cause of the disorder. It has been established that stimulant medication is the most common form of pharmacotherapy, and appears to be the most effective single treatment currently available for the treatment of ADHD [4]. Stimulant treatment is said to improve abnormal behaviours, self-esteem, and both family and social interactions [5]. It is important to remember that a favourable response to stimulant medication is not proof of a diagnosis of ADHD [3].
The use of stimulant medication in children continues to supply the medical profession with controversy. The major concerns seem to centre around fears of abuse, the effect on the developing brain, legal issues (particularly in the US), and side-effect profiles [5]. The majority of children with ADHD who are treated with stimulants (methylphenidate or dexamphetamine) will experience some adverse side-effects, including insomnia, decreased appetite, stomachache, and headache [6]. Side-effects of insomnia provoke some interesting discussion as symptoms of inattention and impulsiveness (i.e. the core symptoms of ADHD) are also the characteristics of sleep deprivation [9–11] while severe sleep problems in infancy appear to be associated with subsequent development of ADHD [12]. The National Health and Medical Research Council (NH&MRC) recommend the use of DSM-IV criteria for the diagnosis of ADHD in Australia [7]. These criteria define ADHD as a chronic condition involving the core symptoms and behavioural characteristics of hyperactivity, inattention and impulsivity [8].
Sleep disturbances and difficulties have been associated with ADHD for many years, and were included as part of the diagnostic criteria for attention deficit disorder in DSM-III [13]. In 1998 Corkum et al. [2] reviewed empirical research of ADHD-related dyssomnia that was published since 1970. This review concluded however, that most of the research in this area used small sample sizes, a variety of diagnostic criteria, and relied heavily on parental reporting and other subjective measures of sleep disturbance. More recently, for instance, Stein [14] re-investigated parental reports of sleep problems of more than 200 stimulant-treated and untreated children diagnosed with ADHD. It was found that moderate to severe sleep problems reportedly occurred at least once a week in about 20% of children with ADHD versus 13% of the psychiatric and 6% of the paediatric control subjects. Stimulant treatment was found to be associated with a higher prevalence of nightly severe sleep problems than did untreated children with ADHD (i.e. 29% vs.10%). The author concluded that, despite the high prevalence of sleep related problems in ADHD, the significance of the association between delayed sleep onset and ADHD with regard to aetiology and management of ADHD is still poorly understood.
Although the notion of dyssomnia in ADHD is widely accepted, Corkum and colleagues [2] concluded that objective verification was inconclusive [e.g. 15]. Corkum et al. also reviewed some preliminary evidence supporting the hypothesis of stimulant medication leading to changes in sleep latency. However, the main outcome of their review was that the exact nature of sleep problems in children with ADHD still remains to be determined. As a result of this, and the vast quantities of such early research, it became necessary in this review to limit the examination to the objective studies published since 1999.
Materials and methods
An online literature search of the National Library of Medicine (PubMed) and the Cochrane Library was performed in June 2004. The search of PubMed using the search terms ‘Attention Deficit Hyperactivity Disorder’ (8743 hits), and ‘sleep’ (71 837 hits) in a combined search revealed 255 articles. Limiting these articles to ‘human’, ‘English language’ and those written 1999–2004 revealed 16 articles. On review of these 16 articles, 10 relevant objective studies were identified [11],[16–24]. However, the search of the Cochrane Library revealed no relevant articles for this time period.
Results
The 10 reviewed studies used objective measures of sleep disturbances, either alone or in combination. Objective measures of sleep involved recording information of sleep patterns. This information is captured in real time and includes actigraphy, video analysis, and polysomnography.
Actigraphy and video recordings are commonly used to objectively examine movement disorder in sleep, but they do not measure EEG sleep stages or architecture [10]. Actigraphy uses computerized devices to quantify movement during sleep, and thereby distinguish between sleep and wake stages. Video analysis involves video-taping the child whilst sleeping, which is then scored and analysed. Polysomnography involves the examination of sleep architecture and the different stages of EEG sleep [10]. This is achieved by recording a number of physiological measures, and is normally performed in a laboratory setting. Of the studies under review, six studies used polysomnography alone [11], [16], [17],[22–24], two used actigraphy alone [19], [21], one study used a combination of polysomnography and actigraphy [18] while one study used video recording analysis in combination with polysomnography [20].
Sleep architecture can be divided into rapid eye movement (REM) and non-rapid eye movement (NREM) sleep, which describe specific anatomical, electrophysiological and behavioural characteristics. The REM stage of sleep is typically characterized by distinct low voltage cerebral activity, well-defined episodic rapid-eye movements and absent muscle tone. REM and NREM sleep stages make up a cycle of sleep, which normally lasts about 90minutes. There are between four and five cycles a night, with the duration and vigour of the REM stage increasing with each subsequent cycle.
REM sleep percentage refers to the ratio of duration of REM sleep to the total sleep duration. Interestingly, this percentage remains constant throughout life (approx. 23%), whilst the actual time spent sleeping may differ. REM sleep latency refers to the time taken to reach REM sleep [10]. Children tend to have a high quality of sleep, with short sleep latencies, a long duration of consolidated sleep, few spontaneous arousals, and sleep efficiencies, which are often greater than 95% [9].
Using overnight polysomnography, the effect of stimulants on sleep characteristics in children with ADHD was investigated by O'Brien et al. [16] in a sleep clinic and community sample. The study also performed subjective measures (i.e. sleep habits questionnaire). The subjects were selected from two sources; a chart review of all children with ADHD who were referred to a specific sleep clinic over a 2 year period, and a community survey of 5–7 year old children during the same time period. The study divided these subjects into three groups: stimulant-medicated children diagnosed with ADHD (n = 53), nonmedicated children with ADHD (n = 34), and a control group (n = 53). The main results of this study identified parents of ADHD children (both medicated and non-medicated) as reporting more sleep disturbances than controls. The results of the objective measurements demonstrated decreased REM sleep percentage in children with ADHD (see also [11]) without establishing major differences in subjective or objective measures between medicated and non-medicated ADHD children.
More recently Kirov and coworkers [22] largely confirmed this pattern of changes in sleep architecture in 17 unmedicated boys diagnosed with ADHD versus 17 closely age and intelligence matched controls, while Lecendreux et al. [23] did not record significant changes in REM sleep when comparing 30 children with ADHD aged between five and 10 years with 22 age and gender matched controls.
Polysomnography studies have also been employed to quantify periodic limb movements in sleep in ADHD. For instance, Picchietti et al. [24] compared data of 14 stimulant-naïve children diagnosed with ADHD with 10 control children. The authors reported increased rates of periodic limb movements during sleep in the children with ADHD that was also associated with arousals. Interestingly, the incidence of restless leg syndrome was also significantly increased in the biological parents in the ADHD sample when compared to the parents of the control group, thus suggesting a possible link of the genetic predisposition for ADHD and the movement disorder (see also [12]).
O'Brien and colleagues [17] provided a more detailed analysis of sleep patterns of their former sample [16] by combining polysomnography and actigraphy analysis. Significant differences between the groups were confirmed for REM sleep latency and percentage, and periodic limb movement with associated arousal. REM sleep latency and percentage of REM sleep was increased in children with ADHD compared with controls, for both the community and clinic samples. Periodic limb movement was increased in the clinic sample of children with ADHD but not significantly different between the ADHD community and control sample. These findings supports parental reporting of increased restless legs syndrome and periodic leg movements during sleep in children with high scores on the DSM-IV-derived hyperactivity index [25].
A retrospective review of 97 children diagnosed with ADHD, who were referred to a paediatric sleep medicine centre, was performed by Crabtree et al. [18]. A comparison was made between the subjective symptoms reported and results of an objective sleep assessment. Polysomnography and actigraphy were performed on all 97 children. There was no control group comparison made in this study. Thirty-six per cent of the children demonstrated periodic limb movement disorder and 16 children demonstrated substantial sleep variability. In a small proportion of cases only, however, were parental subjective sleep complaints confirmed by objective measurements.
Corkum et al. [19] assessed a variety of sleep parameters in children with ADHD and compared these to an age and gender matched control group. Each sample group consisted of 25 children (20 boys and 5 girls) ranging in age from seven to 11 years. The data was obtained through actigraphy over seven consecutive nights (in the home setting), whilst parents recorded a sleep diary. This trial showed through questionnaires that parents of children with ADHD report more sleep difficulties than parents of children without ADHD. However, objective verification was not obtained with either actigraphy or the use of a sleep diary, apart from the observation that ADHD children showed increased sleep duration.
Konofal et al. [20] compared sleep patterns of 30 children with ADHD with an age and gender matched group of 19 controls. The assessment was performed with the use of polysomnography and video analysis of observable nocturnal movements. The results suggested that children with ADHD have significantly more nocturnal movements than control group children. The movements tended to be of the upper and lower limbs, and the duration of activity was longer in the ADHD group. This trial showed no demonstrable difference in REM sleep percentage or latency between ADHD and control children.
Gruber et al. [21] performed a comparison between the sleep-wake cycles of children with ADHD and a control group. The comparison was made using data collected from subjective (sleep diary) and objective measures (five night actigraphy) on 38 school-aged boys diagnosed with ADHD and 64 school-aged boys in the control group. The samples were recruited from the same school district in Tel Aviv, Israel, and were specified as having homogenous psychosocial and cultural backgrounds. The findings of this study support the hypothesis that instability of the sleep-wake system is characteristic of children with ADHD. For instance, ADHD children in this study demonstrated an increased instability of sleep onset, duration and REM sleep when compared with the control group. A discrimant analysis of the data suggests that classification of children with ADHD in this trial can be significantly predicted on the basis of their sleep measures.
Discussion
Despite polysomnography, actigraphy, and video analysis studying different aspects of sleep disturbances, they have proved themselves as valid and reliable measures of overall sleep function. For the purposes of this review, actigraphy and polysomnography are considered as providing equally valid measurements of sleep disturbance, due to the high rate of agreement (85–90%) between them [10]. Actigraphy has also been established as a valid and reliable measure of distinguishing between sleep disturbed children and controls [9] despite its inability to record information regarding the architecture of sleep.
Although polysomnography is highly regarded in terms of its usefulness in sleep analysis due to its ability to measure sleep architecture, it is important to consider the ramifications of recording information in a laboratory setting. Laboratory studies often impose an unnatural sleep-wake cycle [9] and place children in an unfamiliar environment. In an effort to negate these effects, all the trials, which used polysomnography, examined the children in a darkened room, with an ambient temperature of 24 degrees, and one parent present. None of the trials used medication to induce sleep. Only one study allowed time for the children to adapt to the new sleep environment (three nights prior to polysomnography), to minimize the first-night effect [20].
Studies, which used actigraphy as the sole objective measure of sleep disturbances, did so with the aim of reproducing a natural sleep-wake cycle. These trials recorded information over a longer time period (i.e. five [21] to seven [19] days) in the child's normal sleep environment.
In the 10 studies reviewed here, only two studies [20], [23] showed no differences between sleep variables in ADHD children and controls. However both studies did illustrate increased levels of nocturnal activity in the ADHD sample. Two studies produced results suggestive of increased inter- and intrasubject variability in the sleep-wake patterns of children with ADHD [18], [21]. Of the 10 studies reviewed, six demonstrated prolongation of REM sleep latency and decreased percentage of REM sleep in the ADHD group compared with controls [11],[16–19],[22]. This is in comparison with children without ADHD who tend to have short REM sleep latencies and a very high sleep efficiency percentage.
The small sample sizes were an area identified as a possible methodological limitation of earlier studies [2]. The precise numbers used in each study are mentioned earlier, however, the studies reviewed here, mostly used larger sample sizes than the earlier research. Nine of the trials had a control group, which was age and gender matched, with one prospective study not using a control group at all [18].
Diagnostic criteria used to identify children with ADHD was identified as an area of weakness in the earlier studies performed on sleep disturbances [2]. All 10 studies reviewed here, included children with a previous diagnosis of ADHD (by a physician or psychologist) in combination with parental reporting in their ADHD group. Five of the studies specified meeting DSM-IV criteria for ADHD as part of their inclusion criteria [19–23], with three of the other studies using the Conners Parent Rating Scale [26] to confirm the presence of hyperactive behaviours [11], [16], [17]. The Conners Parents Rating Scale is not a diagnostic tool for ADHD, however, the ADHD index is well validated, with a score of two standard deviations above the mean providing a recognized measure of children at high risk of a diagnosis of ADHD [27]. All of the trials excluded children with other significant medical problems, and all but one of the studies [16] required the children to be unmedicated or medication-naïve (e.g. [24]).
The study which analysed medication use and its effects on sleep characteristics in children with ADHD, required the subjects in the medicated group to be administered stimulants only, as per their normal medication regime [16]. It was found that stimulant medication was not associated with differences in objective sleep measures despite a tendency towards prolonged REM sleep latency in the medicated group, although this result was not significant. However, there were several limitations in this study. These include the fact that information regarding actual medication schedules (morning and/or night administration), whether the medication was actually taken, whether the nonmedicated children were stimulant-naïve (or the time elapsed since discontinuation), and rigorous diagnostic criteria for the ADHD sample, were not addressed.
Conclusions
This literature search produced 10 relevant articles from over 70 000 articles in the topic area, which suggests the high specificity of the search. The sensitivity is more difficult to evaluate, however, it has been assumed that studies, which have not been included in the extensive Medline/Life Sciences database, are more likely to be of lower quality [28]. The conclusions reached by these studies supported much of the early research on sleep disturbances in ADHD. The lack of correlation between subjective retrospective parental reporting and objective sleep measures was apparent in all six studies. Parental reporting in interviews and questionnaires of children with ADHD showed alarmingly higher reports of frequent nocturnal wakening, increased bedtime resistance, difficulties initiating sleep, and nightmares, than non-ADHD children [14], [16]. However, these subjective reports were not substantiated by polysomnography or actigraphy results in any of the reviewed reports. Higher parental reporting of sleep disturbances in children with ADHD without objective verification, may suggest an extrinsic cause (e.g. behavioural or environmental) for the disturbances.
The studies also produced similar conclusions to the earlier trials with regard to sleep latency, REM sleep percentages, and nocturnal activity in their ADHD samples. The studies reviewed here demonstrate that children with ADHD have increased sleep latency, decreased REM sleep percentages and increased nocturnal activity [11],[16–19],[22]. These results may suggest an intrinsic cause for the sleep disturbances in ADHD, with the implication of dysregulation of sleep architecture as the causal mechanism.
Although parental reporting of sleep disturbances in ADHD is out of proportion to the disturbances found in objective studies, there does appear to be a link between ADHD and sleep. Review of the polysomnography studies suggests a significant association between children with ADHD and architectural problems of sleep [11],[16–19],[22]; c.f. 23. Whether these results implicate a causal role of sleep disturbance in the pathogenesis of ADHD or whether sleep architecture is adversely affected by behavioural problems remains uncertain.
Initial evidence seems to suggest that treatment of ADHD with stimulant medication appears to impose little effect on objective sleep characteristics [16]. There was however, some suggestion of a mild increase of REM sleep latency in children with ADHD who are stimulant-medicated. Further research is required in this area; particularly with regard to the effects of different medication schedules (e.g. higher dose levels and newer long-acting stimulants) on sleep architecture.
Footnotes
Acknowledgements
Ulrich Schall received infrastructure support from NSW Health through the Brain and Mental Health Research Program of the Hunter Medical Research Institute (HMRI) and the Neuroscience Institute of Schizophrenia and Allied Disorders (NISAD).