Abstract
Introduction
A high prevalence of sleep disturbances in children with ADHD has been often reported in the literature. Studies based on subjective (parental) reports showed an increased prevalence of bedtime struggles, difficulty falling asleep, night awakenings, movements during sleep, and tiredness upon waking, when compared with normal controls (Corkum, Tannock, & Moldofsky, 1998; Gruber, 2009; Hoeppner, Trommer, Armstrong, Rosenberg, & Picchietti, 1996; Kaplan, McNicol, Conte, & Moghadam, 1987; J. Owens et al., 2009; Van der Heijden, Smits, Van Someren, Ridderinkhof, & Gunning, 2007). On the contrary, studies using objective measures (e.g., polysomnography, actigraphy) yielded little evidence of differences in sleep structure between ADHD patients and normal controls, with similar sleep duration and only few differences in polysomnographic and actigraphic parameters (Corkum, Tannock, Moldofsky, Hogg-Johnson, & Humphries, 2001; Cortese, Faraone, Konofal, & Lecendreux, 2009; Hvolby, Jorgensen, & Bilenberg, 2008; J. Owens et al., 2009; Sadeh, Pergamin, & Bar-Haim, 2006).
The major difference between objective sleep measures and parental reports is related to sleep onset insomnia and night awakenings (Corkum et al., 2001); on the contrary, the report of increased motor activity at night has been confirmed by different studies (Chervin et al., 2002; Konofal, Lecendreux, Bouvard, & Mouren-Simeoni, 2001; Moreau, Rouleau, & Morin, 2014) and acknowledged by two meta-analyses of polysomnographic studies (Cortese et al., 2009; Sadeh et al., 2006). These two meta-analyses showed that in children with ADHD a high number of periodic limb movements during sleep is often reported as well as a high number of stage shifts per hour, slightly increased apnea–hypopnea index and low sleep efficiency (percentage of the night that the child was asleep from original sleep onset to waking up) and total sleep time.
A very recent systematic review analyzed 24 studies comprising 2,179 unmedicated ADHD school-age children versus controls, using several actigraphic measures as an outcome: sleep duration, activity mean, sleep onset latency, sleep efficiency, and wake after sleep onset. The meta-analysis indicated that children with ADHD have a moderately increased mean activity compared with typically developing (TD) children and this increase was more evident during daytime structured experimental sessions. Sleep latency shows a significant and moderate increase and sleep efficiency parameter was lower in ADHD compared with TD children. There was no evidence that children with ADHD compared with TD have different sleep duration or wake after sleep onset (De Crescenzo et al., 2016).
Another accurate evaluation of the different actigraphic studies in school-age children with ADHD versus TD children showed conflicting results: Some studies failed to detect major differences in sleep measures (Bergwerff, Luman, & Oosterlaan, 2016; Corkum et al., 2001; Poirier & Corkum, 2018; J. Owens et al., 2009; Wiggs, Montgomery, & Stores, 2005) while other studies reported greater night-to-night variability (Gruber & Sadeh, 2004; Gruber, Sadeh, & Raviv, 2000; Moreau et al., 2014) and an increase of mean activity (Dagan et al., 1997; Moreau et al., 2014) as well as difficulty initiating sleep (Hvolby et al., 2008; Moreau et al., 2014) and shorter total sleep time (Moreau et al., 2014; J. Owens et al., 2009).
The high variability of sleep patterns of children with ADHD (Gruber et al., 2000; Palm, Persson, Bjerre, Elmqvist, & Blennow, 1992) has been advocated as the major factor for the inconsistency between objective and subjective measures. Gruber et al. (2000) found an increased night-to-night variability in the sleep–wake pattern of children with ADHD compared with TD children. This finding has been corroborated by another recent study that found that children with ADHD have longer sleep onset latency and an increased intra-individual night-to-night variability compared with healthy children and with children with other psychiatric diagnoses (Hvolby et al., 2008).
ADHD is increasingly diagnosed at preschool age (Egger & Angold, 2006; Hardy et al., 2007; Healey, Miller, Castelli, Marks, & Halperin, 2008; Marakovitz & Campbell, 1998; Posner et al., 2007); furthermore, Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM-5; American Psychiatric Association, 2013) considers the possibility to diagnose ADHD starting from 4 years of age. The prevalence of preschoolers with ADHD varies in different countries: In the United States, about 6% of preschoolers received a diagnosis of ADHD in a community sample of 1,073 children (Lavigne Lebailly, Hopkins, Gouze, & Binns, 2009) while in Europe a recent study on 2,475 preschoolers in Norway estimated a prevalence of 1.9% (Wichstrøm et al., 2012). Several studies used a procedure that required the integration of the emotional and behavioral questionnaires, such as the Child Behavior Checklist (CBCL) or the Conners Comprehensive Behavior Rating Scales (Conners CBRS) as a first screening and a semi-structured psychiatric interview to parents, such as the Pre-School-Age Psychiatric Assessment (PAPA; Egger & Angold, 1999; Egger et al., 2006). In preschoolers, ADHD shares many features with the disorder diagnosed in school-age children and in adolescents (Sonuga-Barke, Thompson, Stevenson, & Viney, 1997; Sterba, Egger, & Angold, 2007) although most affected preschoolers display predominant hyperactivity/impulsivity or both subcomponents. The similarity of ADHD symptoms profiles in preschool and school-age children and the presence of sleep disorders in early age, suggests that 4 to 6 years should be considered an important time-window to identify not only ADHD but also the specific sleep features of this disorder.
There are several examples of children with sleep disorders and ADHD symptoms in whom an improvement in attention and overactivity with treatment of the sleep disorder was reported (Cortese et al., 2013; Kirov & Brand, 2014; Konofal et al., 2007) but also other studies showed that dopaminergic therapy in children with restless legs syndrome (RLS) improved RLS but not ADHD symptoms (England et al., 2011). In light of the present studies, the nature of sleep problems in ADHD appears complex and multifactorial, and therefore, it is not clear how specific sleep problems/alterations in ADHD may affect various daily neuro-behavioral functions.
As all of the actigraphic studies have been carried out in school-age children with ADHD and almost none in preschool children (Goodlin-Jones, Waters, & Anders, 2009; Miyahara, Healey, & Halperin, 2014) and as the early identification of ADHD participants with sleep disturbances and increased motor activity might guide the choice of pharmacological and non-pharmacological interventions, the aim of the present study was to evaluate the characteristics of sleep of ADHD preschoolers by means of multiple measures of sleep (sleep questionnaire, sleep diary, and actigraphy).
Method
Participants
A group of 25 preschoolers with a clinical diagnosis of ADHD, according to the Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; American Psychiatric Association, 1994) criteria aged between 3 and 5.11 years was consecutively recruited from a clinical population referred to a Child and Adolescent Neuropsychiatry Outpatient Center located in Rome (Italy) in the last 2 years. Children with neurological sensory-motor deficits, cognitive impairment (IQ <70), or autistic disorder were excluded. All children with ADHD were medication naive.
Inclusion criteria for the ADHD group were (a) significant ADHD symptoms, as indicated by pathological scores at the Attention Problems, Aggressive Behaviors, and Externalizing scales of the CBCL and (b) clinical diagnosis of ADHD according to DSM-IV criteria based on the PAPA interview (Egger & Angold, 1999; Egger et al., 2006) and behavioral observation.
A total of 21 age-matched, TD preschool children whose parents accepted to participate in the study were recruited in a day care center located in Rome (Italy). Inclusion criteria for the TD group were (a) the absence of any developmental or behavioral disorder (established by parent report) and (b) normal score at the Leiter International Performance Scale–Revised (Leiter-R; Leiter, 1979) and no psychiatric diagnoses at the PAPA interview (Egger & Angold, 1999; Egger et al., 2006).
Procedure
For the preschool children with ADHD, the assessment procedure needed different visits, on separate days, in two steps. In the first step, a child psychiatrist collected the history from the primary caregiver with particular attention to the emotional-behavioral problems and all parents were asked to fill out the CBCL 1.5-5 (Achenbach & Rescorla, 2000). Thirty-three children, for whom parents reported ADHD symptoms and/or borderline/clinical T-scores in the externalizing CBCL scale, underwent a second assessment step. All parents accepted to proceed with the assessment.
In the second step, the children participated to an unstructured behavioral observation session and neurological examination. To exclude participants with intellectual disabilities, all children underwent cognitive assessment by the Leiter International Performance Scale–Revised (Leiter-R; Leiter, 1979). The parents (mother or father or both) were administered the PAPA interview (Egger & Angold, 1999; Egger et al., 2006) by a child psychiatrist expert in developmental and psychiatric disorders in preschool age. Based on the behavioral observation and PAPA interview, the child psychiatrist made the diagnosis of ADHD according to the DSM-IV criteria in 25 children. The remaining eight children did not fulfill the criteria and were excluded.
Similarly, the parents or primary caregivers of 21 TD children who accepted to participate to the study were asked to fill out the CBCL 1.5-5 as a first screening. None of the children showed clinical T-scores at the CBCL externalizing scale. Furthermore, parents were asked to participate in the PAPA interview to exclude the presence of clinical ADHD criteria symptoms. All parents accepted to proceed with the assessment.
After the interviews, actigraphs were provided to the parents or caregivers with general information on how the actigraph works and then it was placed on the child’s non-dominant hand wrist. At the end of the recording period, the parents returned the actigraph to the Center. Parents filled out and returned the sleep questionnaire and sleep diaries concurrently with the actigraphs.
In children with intercurrent diseases, the actigraphic study was replicated when in healthy condition.
Socioeconomic status was derived by calculating the Hollingshead four-factor index based on both parents’ levels of occupation and education (Hollingshead, 1975) even if only one parent was employed. If both parents were employed, the Hollingshead scores were added and divided by 2 to receive a combined score that was then divided into 3 homogeneous categories (low, middle, and high).
This study was performed at the Department of Social and Developmental Psychology, Sapienza University, Rome, Italy, and at the Center of Child and Adolescent Neuropsychiatry, Rome, Italy, in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments.
The institutional review board approved the study. Informed consent was obtained from parents before enrollment in the study.
Psychodiagnostic Assessment
CBCL 1
5-5. The CBCL 1.5-5 (Achenbach & Rescorla, 2000) was filled out by parents of children with ADHD and TD children. The CBCL 1.5-5 for preschoolers is composed of 99 items inherent to seven dimensions, namely, Emotionally Reactive, Anxious/Depressed, Somatic Complaints, Withdrawn, Sleep Problem, Attention Problems, and Aggressive Behavior designed to rate emotional and behavioral problems in preschool age. Items are scored by parents as 0, 1, or 2, corresponding to not true (as far as you know), somewhat or sometimes true, and very true or often true, respectively. In addition, the computerized scoring system of the questionnaire reports 3 composite scores (internalizing, externalizing, and total problems). The reliability and validity of the CBCL in predicting stable patterns of behavior throughout childhood are well established for both ADHD and for other diseases (Mattison & Spitznagel, 1999). The questionnaire has been translated into Italian and validated for its use in Italy (Achenbach & Rescorla, 2000)
PAPA interview
The PAPA interview (Egger & Angold, 1999; Egger et al., 2006) was administered by interviewers experienced with pre-school-age children. The PAPA is a semi-structured, clinical interview with the parent to diagnose psychiatric disorders in preschool children aged 2 to 5.11 years and provides a comprehensive and reliable assessment of DSM-IV diagnoses. It uses an interviewer-based approach and includes a detailed glossary that provides guidelines for rating the symptom severity, frequency, duration, and impairment. The PAPA is the only comprehensive psychiatric interview to our knowledge with demonstrated test–retest reliability and validity for assessing psychiatric symptoms and disorders in toddlers and preschool children. The original study reported a diagnostic reliability similar to that obtained with interviews of older children and adults with K’s ranging from .36 to .79. Test–retest intra-class correlation (ICC) for DSM-IV syndrome scale scores ranged from .56 to .89. There were no significant differences in reliability (either 0 or ICCs) with respect to age, gender, or race (Egger et al., 2006).
Sleep Assessment
Sleep questionnaire
The items of the sleep questionnaire were derived from the Sleep Disturbance Scale for Children (SDSC; Bruni et al., 1996) used to define the presence of any sleep disorders. The questionnaire was filled in by a parent or a caregiver and consisted of 24 items coded 0, 1, and 2 based on the frequency of the disturbed behavior (0 = never, 1 = sometimes or 1-2 times per week, 2 = more than 2 times per week). The areas of sleep behavior/disturbance in the questionnaire were divided into four sections: (a) bedtime, (b) behavior during sleep, (c) morning awakening, and (d) daytime somnolence. For the purpose of this study, the item frequencies were evaluated based on a dichotomic scale based on the absence (score 0) or the presence (score 1 or 2) of the behavior. The internal consistency was high in controls (.79) and in sleep disorder participants (.71); the test–retest reliability was adequate (r = .71; Bruni et al., 1996).
Sleep diary
In the absence of validated sleep diaries for children in the literature, an ad hoc sleep diary was developed from previously validated pediatric symptom diary scales, and it was successfully used in a previous actigraphic study (Bruni, Russo, Violani, & Guidetti, 2004). Parents were trained to keep a daily log of their child’s sleep and indicated the days of the week and the time when the children went to bed and when they woke up. The parents completed this sleep diary during the actigraphic recording period to provide subjective assessment of sleep–wake patterns and collected information on sleep onset time, wake-up time, bedtime difficulties, and nocturnal awakenings, with identical questions each day. Moreover, parents noted when the actigraph was not worn (e.g., when swimming or bathing), if the child was napping, and if any change in sleep routine occurred. The information derived from the daily sleep log was also used to check possible discrepancies or missing data from the actigraph. Parents were asked to complete the diary just before their children went to sleep and as soon as possible after they got up in the morning.
This child sleep diary was previously administered to a pilot group of parents to evaluate the ease of use and compliance with its use.
Actigraphy
Actigraphy is a simple method to objectively evaluate sleep patterns, consisting of a wristwatch-sized activity sensor worn on the non-dominant wrist, to discriminate between sleep–wake states through documentation of body movements; it is used in the child’s natural environment, and it allows for multiple-day data collection (Sadeh, Sharkey, & Carskadon, 1994).
Watch-Motionloggers from Ambulatory Monitoring, Inc. (Ardsley, NY) were used in the current study, to measure sleep parameters and nocturnal motor activity. These actigraphs employ a piezoelectric sensor, have a fixed sensitivity at 2 to 3 Hz and detect accelerations greater than 0.01 g force. The mechanism is housed in a metal, waterproof case and has a 32K memory. The actigraphs were programmed to employ a zero-crossing mode using an auto actigraph interface.
Data were extracted using the ACT software and analyzed by means of the ACTIONW2 program, according to the validated sleep estimation algorithm developed by Sadeh, Lavie, Scher, Tirosh, and Epstein (1991).
Parents were provided with general information on how the actigraph works, which was then placed on the wrist of the child’s non-dominant hand by the researcher. Parents were instructed to allow the child to wear the watch continuously during the recording period except during baths or water play. During actigraphy data collection, parents were asked to keep their child’s regular sleep routine and record sleep–wake times. Moreover, parents were instructed to mark “lights out” and “get out of bed” events or when the actigraph was not worn by pressing the event marker of the actigraph.
The following actigraphic data variables were computed:
Total Sleep Duration: the period of time in minutes from the marked “lights out” to the marked “get out of bed”;
Sleep Latency: minutes from the time the parents’ note of “lights out” to the first period of sleep lasting more than 20 min (sleep onset);
Sleep Minutes: the period of sleep time in minutes from the sleep onset to morning awakening time minus the wake after sleep onset;
Wake After Sleep Onset: the times in minutes of awakenings during the Sleep Minutes;
Mean Wake Episode: mean duration of Wake Episodes (minutes);
Sleep Efficiency: the ratio between the Sleep Minutes and the Total Sleep Duration × 100;
Activity Mean: the mean number of activity counts per 1 min epoch during Total Sleep Duration;
Activity Index (index of motor activity): the percentage of epochs with >0 Activity Mean;
Sleep Fragmentation Index: the ratio between the number of awakenings and the Sleep Minutes × 100.
Similar to previous studies, intra-individual variability was estimated for each of the above parameters by calculating the within-subjects standard deviation of each variable (Gruber et al., 2000, Gruber & Sadeh, 2004; Moreau et al., 2014).
Data Analysis
The statistical analysis was performed using the software SPSS for Windows Version 20.0 (Armonk, NY: IBM Corp.). Descriptive statistics were used to explore the data and to describe the behavioral and sociodemographic characteristics of the groups. The age difference between TD children and ADHD was evaluated with the one-way ANOVA. Frequency data (and percentages) were produced for parent-reported sleep disturbance questionnaire items and differences were evaluated using the Yates-corrected chi-square test.
For each child, actigraphic data were averaged over 4 to 5 weekday nights, as recommended by the literature (Sadeh & Acebo, 2002), and comparisons between TD and ADHD preschoolers were carried out by means of the ANOVA.
Furthermore, the intra-individual variability for each parameter was estimated by calculating the within-subject standard deviation of each variable. Also, for these parameters, the two groups were compared by means of the ANOVA.
Results
Group Demographics
Demographic information on the study groups is reported in Table 1. The ADHD group included 21 males (84%) with a mean age of 5.4 ± 0.3 years (range = 48-69 months).
Demographic and Clinical Characteristics of the Study Groups.
Note. TD = typically developing.
Significant differences are indicated in bold.
2 × 2 chi-square test with Yates correction when appropriate.
Six children (24%) were diagnosed with the inattentive subtype, nine (36%) with the hyperactive-impulsive subtype, and 10 (40%) with the combined subtype. Regarding psychiatric comorbidity, oppositional defiant disorder was found in 10 cases (40%), generalized anxiety disorder in three cases (12%), and depression in two cases (8%).
Family history for ADHD symptoms in parents was found in four cases and for depressive symptoms in five cases; only two siblings presented with ADHD symptoms.
The group of TD preschool children included 15 males (71%) with a mean age of 5.1 ± 1.8 years (range = 48-69 months). Family history of TD children was unremarkable.
There was no difference between children with ADHD and TD children for age (p = .87) and sex (Yates-corrected chi-square = 0.45; p = .50). Both groups were of Caucasian ethnic background. None of the children was taking naps during the day.
Almost no significant differences have been found on demographic variables analyzed. Regarding the socioeconomic status, no significant difference emerged between the two groups. Interestingly, a higher positive family history for sleep disorders in parents was found in the ADHD group (21%) than in the TD group (0).
CBCL Analysis
Table 2 reports the differences between the two groups in the CBCL dimensions and composite scores. As expected, children with ADHD showed significant differences in Withdrawal (58.83 vs. 51.15; p < .0001), Attention Problems (69.88 vs. 51.54; p < .0001), Aggressive Behavior (59.46 vs. 51.08; p < .0001), and Internalizing (55.21 vs. 46.77; p < .05), Externalizing (63.38 vs. 42.62; p < .0001), and Total (59.29 vs. 44.15; p < .0001) composite scores.
Comparisons of the CBCL Mean Scores Between ADHD and TD Children.
Note. CBCL = Child Behavior Checklist; TD = typically developing.
Significant differences are indicated in bold.
Sleep Assessment
The frequency distribution of the items of the sleep questionnaire was evaluated by means of the Yates-corrected chi-square test (Table 3). A significant difference emerged only for the frequency of waking up with a “bad mood.” In particular, percentage of children with ADHD who wake up in bad mood (52.4%) is significantly higher than that of children in the TD group (20.0%). The sleep diary was essentially used to support the actigraphic recording and to record any specific activities that would affect sleep. No differences were found between the two groups in the sleep diary for sleep onset time and wake-up time.
Frequency (%) of Parent-Reported Sleeping Problems in ADHD and TD Children.
Note. TD = typically developing.
Significant differences are indicated in bold.
Actigraphy
The 46 participants included in the study wore the actigraph for an average of 5.24 days (SD = 0.84; range = 4-7); 40 participants (87.0%) had actigraphic data for 5 to 7 days while six participants (13.0%) had actigraphic data for 4 days. No data were lost because of a failure of the actigraphic equipment. No significant differences were found between ADHD and TD children for the number of recorded days.
Table 4 shows the comparison of the actigraphic parameters obtained from ADHD and TD children. Children with ADHD showed a higher level of nocturnal activity with significant differences in the actigraphic parameters Activity mean (9.34 vs. 7.17; p < .05) and Activity index (31.57 vs. 25.74; p < .05). Furthermore, children with ADHD showed an increased night-to-night variability (i.e., SD of the parameters) for sleep minutes (56.44 vs. 32.79; p < .01) and mean wake episodes (1.34 vs. 0.98; p < .05) as well as for the parameters related to the nocturnal activity, namely, the Activity mean (2.64 vs. 1.71; p < .05) and the Activity Index (5.15 vs. 3.77; p < .05).
Actigraphic Sleep Measures.
Note. TD = typically developing.
Significant differences are indicated in bold.
We further analyzed the correlation between the CBCL scales and the actigraphic parameters in children with ADHD (Table 5). The Sleep Problems scale was positively correlated with the Activity index (r = .36) and wake after sleep onset (r = .31). The Attention Problems scale (conceptually related to the ADHD phenotype) showed a positive correlation with mean wake episode (r = .46) and Activity mean (r = .31).
Correlation Analysis Between CBCL Scale Items and Actigraphic Parameters in Children With ADHD.
Note. Significant correlations are indicated in bold. CBCL = Child Behavior Checklist.
Discussion
To our knowledge, this is the first pilot study evaluating objectively the sleep features of preschool children with a diagnosis of ADHD and showing significant changes in actigraphic sleep parameters, versus TD children involving in particular an increase of the mean activity and of the activity index while only few differences were found in the parental reported sleep problems, evaluated by questionnaire.
Questionnaire data showed only one significant difference with preschool ADHD children reporting more frequently to wake up in bad mood, compared with TD children.
Differently from our findings, several previous studies have reported a high prevalence of parent-reported sleep disturbances in school-age children with ADHD (Cortese et al., 2013; Corkum et al., 1998; Gruber & Sadeh, 2004) with approximately one third of medication-free children who experienced chronic sleep onset insomnia (Van der Heijden et al., 2007), as well as bedtime resistance, night awakening, restless sleep, and difficult morning awakening (Corkum et al., 1998; Corkum, Moldofsky, Hogg-Johnson, Humphries, & Tannock, 1999; Gruber et al., 2006, Gruber, 2009; J. A. Owens, Maxim, Nobile, McGuinn, & Msall, 2000; Stein, 1999).
Unfortunately, there are no studies assessing sleep disorders in preschool children with ADHD that can be compared with our data. The only previous study assessing sleep problems in preschool children with ADHD symptoms (without a clinical full diagnosis of ADHD) showed that participants whose parents reported a “generic” sleep problem were more likely to exhibit the clinical ADHD Problems profile at the CBCL (Goodlin-Jones et al., 2009).
However, our finding is partially corroborated by a recent longitudinal study (Scott et al., 2013) aimed at evaluating the association between ADHD diagnosis, sleep duration, and night awakenings in infancy and childhood. This study showed that the proportion of participants who woke up three or more times during the night was higher in children with ADHD at every time point but was not significantly different from controls during infancy or in preschool children. Probably, sleep disruption develops or worsens at a later age, in relation to other factors occurring during development, such as the later use of medication or the subsequent occurrence of stable and chronic psychiatric comorbidity (Konofal et al., 2010).
Actigraphic findings in our study showed an increase of the mean activity and of the activity Index in preschoolers with ADHD but showed no differences in the traditional sleep parameters, such as sleep onset time, sleep duration, night awakenings and sleep percentage, in agreement with almost all previous actigraphic studies in school-age children with ADHD (Corkum et al., 2001; Gruber et al., 2000; Hvolby et al., 2008; Waldon, Vriend, Davidson, & Corkum, 2018; Wiebe et al., 2013; Wiggs et al., 2005).
The only actigraphic study on 2- to 5-year-old children with clinical and non-clinical ADHD profiles derived from the CBCL questionnaire showed no differences in actigraphic parameters between the two groups (Goodlin-Jones et al., 2009). This discrepancy with our results might be explained by the different sample analyzed (children with neurodevelopmental disabilities and TD children) and by the different methodological procedure.
Our findings of an increase of the mean activity and of the activity index are in contrast with other actigraphic studies in school-age children with ADHD that reported no differences in the mean activity (Corkum et al., 2001; Wiebe et al., 2013; Wiggs et al., 2005); however, the most recent actigraphic study by Moreau et al. (2014) is in agreement with our finding showing a significant increase in mean activity in school children with ADHD. They also found a shorter sleep duration, a decrease in sleep onset latency, and sleep efficiency. Accordingly, we found a shorter sleep duration that approached statistical significance but no differences for sleep onset latency and sleep efficiency.
We hypothesize that increased motor activity also in preschoolers with ADHD may represent an important parameter useful to evaluate the continuity from young to older children with ADHD, when the early irregularity of sleep patterns might play a role in the endurance, persistence, and worsening of ADHD symptoms. It has been demonstrated that irregular sleeping habits occurring “often” between 2 and 4 years of age were associated with later attention problems at the age of 5 years and with a large effect for persistent attention problems from 5 to 14 years (O’Callaghan et al., 2010).
The finding of increased activity during sleep is not always reported in older children with ADHD; this could be also explained by the higher prevalence of hyperactivity versus inattention that characterizes younger children with ADHD (Lahey, Pelham, Loney, Lee, & Willcutt, 2005).
In fact, current evidence points to the emergence of overactive and impulsive symptoms at around 3 to 4 years of age, suggesting that temperamental overactivity during early childhood may be the most important predictor of the developmental emergence of ADHD. The cognitive symptoms of ADHD emerge between the ages of 6 and 9 years, with impairments in attention and impulse control (Price et al., 2005). Following this trend, it is possible that the overactivity is more evident in preschool age and then attenuates in school age with the emergence of cognitive symptoms. These changes in daytime behavior might reflect changes also in sleep behavior and might justify the differences found in our study.
We also found higher intra-individual night-to-night variability for sleep duration and for mean activity, activity index, and mean wake episodes, confirming the increased instability of the sleep–wake system in ADHD also in the preschool age. Accordingly, actigraphic studies have suggested that children with ADHD tend to have unstable sleep patterns or inconsistency in their sleep and wake times with an increased intra-individual night-to-night variability (i.e., higher SDs across a 5-day period; Gruber & Sadeh, 2004; Gruber et al., 2000; Moreau et al., 2014). However, other studies did not find this night-to-night variability of actigraphic parameters (Corkum et al., 2001; Goodlin-Jones et al., 2009; Wiggs et al., 2005).
The higher intra-individual variability of sleep parameters in children with ADHD might explain the discrepancy between subjective and objective measures of sleep. Parents often report the more problematic nights but the evaluation of multiple nights with actigraphy possibly picks up problematic nights intertwined with better nights, and the average across multiple nights might blur the data. Other possible explanations for the discrepancy are the use of medication and/or the presence of psychiatric comorbidity (Konofal et al., 2010).
A more recent research demonstrated that a similar degree of night-to-night variability in sleep duration can be found in both ADHD and TD children (Poirier & Corkum, 2018).
Our correlation analysis between CBCL questionnaire and actigraphic measures in preschool children with ADHD showed that the CBCL scale of Sleep problems correlated with wake after sleep onset and activity index and the CBCL scale of Attention Problem with mean wake episode and activity mean.
Even if evaluated with a different methodology, a correlation between inattention and actigraphic sleep parameters (sleep minutes and sleep efficiency) has been found in children with ADHD (Waldon et al., 2018). A recent study showed an interaction effect between ADHD symptoms and internalizing and externalizing behavior on total sleep time, time in bed, and average sleep bout duration, indicating a complex interplay between psychiatric symptoms and sleep (Bergwerff et al., 2016).
This relationship might be linked to the neurobiological mechanisms underpinning ADHD psychopathology and those of sleep regulation, which may share common dopamine and noradrenaline deficits; these neuromodulators are neurochemical determinants of ADHD symptoms (Pliszka, 2005) and play a crucial role for the non-rapid eye movement (REM)-REM sleep alternation in the course of the night (Monti & Monti, 2007). Furthermore, serotonin deficit has been proposed as an important neurochemical mechanism of ADHD psychopathology and is involved in non-REM/REM sleep switching (Kirov & Brand, 2014). Moreover, it has been demonstrated that periodic limb movements in sleep (PLMS) and RLS (highly comorbid with ADHD) may be connected with dopamine deficit in the nigrostiatal brain region and the administration of L-DOPA or metilphenydate and other stimulants increases the synaptic levels of dopamine and noradrenaline, improving ADHD symptoms (Kirov & Brand, 2014).
Therefore, we can advocate that the imbalance between these neurotransmitters in preschool children with ADHD might be responsible not only for the night-to-night variability but also for the increased night-time motor activity.
Sleep problems have also been related to ADHD subtypes: Children with combined subtypes of ADHD have more sleep problems compared with children with the inattentive or hyperactive/impulsive subtypes of ADHD (Chiang et al., 2010; Corkum et al., 1999; Mayes et al., 2009) and their sleep patterns may be characterized by forced non-REM/REM sleep alternation and increased movement time and sleep stage shifts (Kirov et al., 2004). The findings of our study might be therefore influenced by the relative increased presence of the combined subtype and hyperactive/impulsive versus inattentive subtypes explaining both increased mean activity and night-to-night variability.
Limitations
The present study has some limitations that must be duly acknowledged. Although actigraphy is considered a reliable objective method for sleep evaluation, it does not provide information on sleep architecture or on the presence of eventual specific sleep disorders (i.e., sleep disordered breathing and periodic leg movement disorder). However, based on the sleep questionnaire results, it is unlikely that our participants suffered from specific sleep disorders.
We did not evaluate specifically the influence of psychiatric comorbidity because of the small number of participants and therefore we cannot exclude the influence of psychiatric comorbidity on sleep parameters.
A further limitation is that we included children with all three subtypes of ADHD and it is possible that there might be different results for different presentations of ADHD, although in light of the current research, this is unlikely (Speth, Benoit, & Corkum, 2014).
In this study, as in other investigations (Hvolby et al., 2008), we were not able to differentiate between sleep patterns on weekdays and weekends, and between seasons, which might have an influence on the results. However, actigraphic sleep recording took place on random days equally in both groups.
Conclusion
This study provides objectively sleep features in preschool children with ADHD, using actigraphy, and could offer implications for treatment and understanding of different developmental trajectories of children with ADHD.
Different studies have shown an improvement of ADHD symptoms following treatment of underlying sleep disorders (mainly sleep apnea and sleep-related movement disorders; Cortese et al., 2013; England et al., 2011; Kirov & Brand, 2014). A better understanding of the relationship between sleep-related movement disorders and ADHD might allow a more specific clinical management of patients who present with both of these disorders (Walters, Silvestri, Zucconi, Chandrashekariah, & Konofal, 2008).
Because this is the first study on young children, it is impossible to compare the data with other studies in preschool age, but our results of increased motor activity during sleep are corroborated by different clinical studies and meta-analyses performed on school children with ADHD.
If this study is replicated with specific ADHD subgroups, the results might define better if our data indicate the presence of sleeplessness related to behavioral traits such as hyperactivity or a specific developmental characteristic of the disorder in preschool age, predictive of other sleep difficulties in older children with ADHD.
Further studies are needed to identify if some specific sleep features in preschool children with ADHD are related to certain comorbidities and to longitudinally explore the evolution of sleep problems and comorbid disorders.
Footnotes
Author Contributions
Each author made a substantive intellectual contribution to the study. Maria Grazia Melegari: conceptualization and study design, data collection and interpretation, revision of the manuscript, and approved the final manuscript as submitted. Elena Vittori: conceptualization and study design, data collection and interpretation, revision of the manuscript, and approved the final manuscript as submitted. Luca Mallia: data analysis and interpretation, revision of the manuscript, and approved the final manuscript as submitted. Alessandra Devoto: data analysis and interpretation, revision of the manuscript, and approved the final manuscript as submitted. Fabio Lucidi: data analysis and interpretation, revision of the manuscript, and approved the final manuscript as submitted. Raffaele Ferri: revision of the manuscript and approved the final manuscript as submitted. Oliviero Bruni: conceptualization and study design, data analysis and interpretation, preparation and revision of the manuscript, and approved the final manuscript as submitted.
Authors’ Note
This was not an industry-supported study. This work was performed at the Department of Social and Developmental Psychology, Sapienza University, Rome, Italy, in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
