Children and adolescents with chronic fatigue syndrome (CFS) frequently report sleep disturbances. However, little is known about the nature and severity of sleep disturbance and factors associated with sleep problems in pediatric CFS. The purpose of this review was to synthesize and critically appraise existing literature relating to sleep disturbances in pediatric CFS.
Embase, CINAHL, PsychINFO, PubMed. and Medline databases were searched to retrieve all studies that included an assessment of sleep in pediatric CFS. Two reviewers independently assessed eligibility, extracted data, and systematically assessed reporting quality.
Six studies were included and these were mostly case-controlled designs. Findings varied across studies; however, most studies found that children and adolescents with CFS had significantly more sleep disturbances when compared to healthy controls. Significant methodological variations and limitations were apparent.
This review suggests that children and adolescents with CFS experience sleep disturbances. However, results need to be interpreted cautiously given the limited evidence available and its overall low quality. More research is required to elucidate the nature and extent of sleep disturbance in pediatric CFS and should focus on (1) identifying the specific types, causes, and severity of sleep disturbances; (2) the specific consequences of sleep disturbances; and (3) the most effective interventions for sleep problems in this population.
Snodgrass K, Harvey A, Scheinberg A, Knight S. Sleep disturbances in pediatric chronic fatigue syndrome: a review of current research. J Clin Sleep Med 2015;11(7):757–764.
Chronic fatigue syndrome (CFS) is a profoundly debilitating illness characterized by a period of prolonged, excessive fatigue together with a range of other core symptoms.1 Originally thought to occur only in adults, it is becoming increasingly recognized within the pediatric population with estimates of incidence ranging from 0.001% to 2%.2–5 With typical age of onset in mid to late adolescence, the effect of CFS often occurs in the critical stages of higher education where missed schooling due to illness can have a major effect on future academic and social outcomes. Sleep problems are commonly reported by children and adolescents with CFS.6
It is well accepted that adequate amount and quality of sleep is crucial for healthy child development and the functional consequences of poor sleep for children and adolescents are well known.7 Although studies show that most children and adolescents with CFS complain of sleep disturbances,6 the type and severity of sleep disturbance that these children and adolescents experience has not been well characterized. There has been limited investigation of factors that might contribute to sleep disturbances (physiological, psychological, and social) in this group, as well as little or no evidence supporting the efficacy of sleep interventions in children and adolescents with CFS. Therefore, there is little guidance for clinicians in how to best manage sleep disturbances in this population, which may potentially exacerbate disease burden in this already vulnerable group.
In contrast to the limited research evaluating sleep in children and adolescents with CFS, there is more research focusing on sleep in adults with CFS. Adult studies have shown differences in subjective measures of sleep where patients with CFS consistently report poorer sleep quality compared to controls, as well as complaining of unrefreshing sleep, increased daytime sleepiness, and napping.8,9 In contrast, studies using objective sleep measurements, such as polysomnography (PSG), have reported discrepant findings, and not all studies have identified sleep difficulties in their samples. There is emerging evidence that physiological mechanisms, such as heart rate variability and altered cortisol levels, may contribute to the sleep disturbances in adults with CFS.8
Although children and adolescents with CFS may experience similar sleep disturbances to those found in adults, there are a range of unique considerations for sleep in the pediatric CFS population. For example, in addition to the well-known hormonal, psychosocial, and physical changes of adolescence, this developmental phase is characterized by striking maturational changes in sleep and its neurobiological regulation.10 The transition to adolescence is associated with decrease in total sleep time, later bedtimes, increase in sleep onset latency, reductions in slow wave sleep and rapid eye movement (REM) sleep and widening variability in weekday/weekend sleep behavior.10 Differentiating sleep disturbances in adolescents with CFS from the typical changes expected during this developmental period is therefore complex and application of developmental considerations in research design is critical.
Understanding the current literature on sleep disturbances in children and adolescents with CFS is crucial in guiding future research directions and allowing for targeted and effective treatment of sleep disturbances in the future. The aim of this review is to synthesize and critically appraise the current literature assessing sleep in children and adolescents with CFS. This will allow us identify areas requiring further research.
This review was performed in line with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement.11 Searches were performed using Embase (1987–2014), Medline (1987–2014), CINAHL (1987– 2014), PsycINFO (1987–2014), and PubMed (2013–2014) databases. The searches were restricted to English-language publications and human studies. The Medical Subject Headings (MeSH) or key words included (1) fatigue syndrome, chronic; (2) adolescent or child; AND (3) sleep or actigraphy or polysomnography or circadian rhythm. Inclusion criteria were kept deliberately broad due to the anticipated limited literature available. The inclusion criteria were any studies specifically investigating sleep in children and adolescents (age younger than 19 y) in whom CFS had been diagnosed. Any diagnostic criteria for CFS were accepted due to variable use of case definitions apparent in the CFS literature. Excluded were non-English publications and animal studies. As the official diagnosis of CFS was first published in 1987, the database search was limited to dates from 1987 to present (March 2014). This search returned a total of 67 articles of which 18 were duplicates, leaving 49 articles to be considered for inclusion (Figure 1). Reference lists were searched; however, no further references were found. The titles and abstracts of the remaining 49 articles were screened independently by two reviewers (KS and SK), of which 39 were excluded, leaving 10 articles for full text review. Any disagreements regarding inclusion were resolved in discussion with a third reviewer (AH) until consensus was achieved. Six articles were selected for final inclusion in this review.
Flow diagram of literature search.
CFS, chronic fatigue syndrome.
Flow diagram of literature search.CFS, chronic fatigue syndrome.
Data Extraction, Quality and Risk of Bias Assessment, and Data Synthesis
Due to limited papers and wide variation in methods used, a meta-analysis was not possible. The STROBE statement12 and minimum data elements recommended for CFS research by Jason et al.13 were used as guidelines to critically analyses the reporting of included studies. The STROBE statement is a systematically developed checklist of recommendations on what should be included in an accurate and complete report of an observational study in general. Jason et al.13 details the minimum data elements that should be included in research papers specifically reporting on samples of patients with CFS.
Descriptions of Studies
Design characteristics of the included studies are reported in Table 1. Of the six studies, five were case-control studies and one was a case series.14 Two studies were conducted in a clinical setting15,16 and the setting was not reported in four studies. The clinical sample sizes of studies were generally small, with some variation among the studies (range: 4–57). All studies included healthy control samples recruited as convenience samples, with the exception of the case series. Overall, recruitment methods, study settings, and participation rates were not adequately described.
Study design characteristics.
Assessment of Sleep
The studies also varied in the measurements used for sleep assessment (Table 1). Two studies assessed sleep using PSG (one in a home setting and one in a laboratory setting)14,15; two used actigraphy16,17 (one in conjunction with a sleep/wake diary)17; one used a researcher-designed sleep questionnaire18; and one used a researcher-designed sleep-wake diary.19 All studies were performed in the home environment except for the case series by Ambrogetti and Olson.14 Of the studies utilizing objective or continuous daily subjective sleep measures, the time frame of data collection varied greatly, with the two studies using PSG only recording data for 1 day. The time frame for data collection varied within the samples of both studies using actigraphy. Ohinata et al.17 recorded for between 1–2 w, while Kawabata et al.16 recorded actigraphy for a minimum of 3 days and a maximum of 4 w. Only Ohinata et al.17 recorded subjective sleep quality in combination with objective methods and used researcher-designed rather than validated questionnaires.
Two studies18,19 investigated biological factors that may underlie sleep disturbances in their samples. Along with subjective sleep questionnaires, Tomoda et al.19 investigated circadian variation in core body temperature (CBT) and cortisol in saliva, whereas Knook et al.18 investigated melatonin levels in adolescents with CFS compared to controls. With the exception of these biological markers, other factors that may be associated with sleep quality (e.g., age, sex, illness duration, illness severity, comorbid psychopathology, sleep hygiene) were not investigated in any of the included studies.
Demographic and clinical characteristics of study samples are shown in Table 2. Overall, there were significant limitations in the reporting of patient characteristics and the use of diagnostic criteria for CFS varied across the studies. Three17–19 of the six papers used the Centers for Disease Control and Prevention (CDC 1994)20 diagnostic criteria. Stores et al.15 used criteria published by Sharpe et al.1 modifying the illness duration to at least 2 mo from at least 6 mo. The other two papers14,16 did not clearly describe the criteria they used to diagnosis CFS. Most studies, with the exception of Tomoda et al.,19 did not specify the method of diagnosis (e.g., diagnosis confirmed by pediatricians, medical examination/medical investigations conducted), nor methods used to exclude other medical or psychiatric conditions. One study specified that they included patients who complained of sleep problems only,17 whereas the remaining studies did not specify whether a complaint of sleep disturbance was a required inclusion criteria or whether all patients with a diagnosis of CFS were included.
Demographic and clinical characteristics of study sample.
Demographic and clinical characteristics of study sample.
Ages of included participants ranged from 9–21 y. All studies reported sex distribution of their samples, with the percentage of female participants ranging from 41% to 80%. The duration of illness was only reported in two studies,15,19 with durations ranging from 5 mo to 9 y. Only two studies18,19 reported that they excluded patients on medication or asked patients to withdraw from medication prior to study participation. One further study described the medication status of their sample and conducted analyses both including and excluding those on medication.15
As shown in Table 2, descriptions of sample characteristics (e.g., socioeconomic status, illness factors) were in general lacking in detail or not included. Other potential confounding factors such as comorbidities and other symptomatology such as pain were seldom described.
Four of the six studies15–18 compared sleep in children with CFS to healthy controls and these studies consistently found increased sleep disturbances in patients with CFS. Knook et al.,18 using a researcher-designed questionnaire, found that patients with CFS were more likely to subjectively report nocturnal wake-ups and restless sleep than controls. However, time of sleep onset and total sleep duration were similar in both groups. Using actigraphy, Ohinata et al.17 reported that patients with CFS in their study had significantly longer total daily sleep time and significantly more disrupted sleep (as indicated by the percentage of sleep less than 180 min of total sleep time) compared to controls. They also identified two groups of patients with CFS based on sleep/wake diaries: irregular sleep and delayed sleep phase type. They suggested that the circadian rhythm may be disrupted in some children with CFS, particularly those with irregular sleep, with these children presenting as less active during the day and more active at night compared to healthy peers. Comparing nighttime sleep in patients with CFS to controls was not a main focus of the study by Kawabata et al.16; however, they noted differences in the temporal distributions of activity and rest between patients with CFS and healthy controls using actigraphy.
Using PSG, Stores et al.15 also reported significant differences in the sleep physiology of patients with CFS compared to controls. Although Stores et al. found no significant differences in the actual sleep time or sleep latency of participants with CFS compared to controls, the participants with CFS had more disrupted sleep (disrupted by both brief and long awakenings) and therefore significantly reduced sleep efficiency. The median number of awakenings less than 2 min in the CFS group was 8.8 (range 2.4–17.1) compared to 2.0 (range 0.3–5.0) in the healthy controls. For awakenings longer than 2 min, the median number was 19.0 (range 2.5–47.6) in the CFS group compared to 0.5 (range 0 – 1.0) in the healthy controls. Percentage of nonrapid eye movement (NREM) sleep stage 1 was not significantly different across groups; however, significantly less NREM stage 2 sleep was observed in the patient group (median = 21.1%) compared to healthy controls (median = 36.6%). Small but statistically significant reductions in REM sleep were also observed (median, CFS = 18.0%, healthy control = 22.0%), but the REM latency was not significantly different across groups. No significant differences in slow wave sleep (NREM stages 3 and 4) were found between groups. Periodic leg movements during sleep were not investigated in this study.
Two studies18,19 investigated potential biological markers for the sleep disturbances in patients with CFS. Knook et al.18 found significantly higher levels of melatonin in saliva in 13 patients with CFS compared to healthy controls. In a larger sample of patients with CFS, Tomoda et al.19 found that the amplitude of CBT variation was significantly smaller and the appearance time of lowest CBT was significantly delayed in the group of patients with CFS compared to controls. They also found differences in the circadian rhythm of cortisol in patients with CFS compared to controls.
Sleep disturbance is a frequent complaint in children and adolescents in whom CFS is diagnosed. This study aimed to systematically review existing literature on sleep disturbances in pediatric CFS. We identified only six published studies investigating sleep in pediatric CFS. Both subjective and objective methods have been utilized in the measurement of sleep in this area, although no validated sleep questionnaires have been employed. Findings of the six studies identified in this review all suggest that increased sleep disturbances occur in children and adolescents with CFS compared to healthy peers. However, the small number of studies, together with their significant methodological inadequacies, limits the capacity to synthesize, interpret, and generalize findings.
Interestingly, regardless of measurement method (questionnaire, actigraphy, PSG), a significantly increased number of night-time awakenings were apparent in the CFS groups compared to control groups. Findings were inconsistent as to whether the total sleep time differs between patients with CFS and healthy controls. For example, using actigraphy, Ohinata et al.17 found longer total sleep time in their CFS sample compared to healthy controls, whereas Stores et al.15 and Knook et al.18 found a similar total sleep time across groups using PSG and questionnaires, respectively. Although longer sleep onset latencies have been described in adults with CFS,8 there is insufficient evidence to determine whether children and adolescents also experience longer sleep onset latencies than healthy peers due to limited studies and variable findings. Stores et al.15 and Knook et al.18 found no significant differences in sleep onset latencies in children with CFS compared to controls and sleep onset latencies were not directly assessed in the other studies included in this review.
The only study investigating sleep architecture in children with CFS identified in the current review found differences in sleep stage architecture between children with CFS and healthy controls.15 Statistically significant but relatively small reductions were found in NREM stage 2 sleep and REM sleep, but NREM stage 1, slow wave sleep, and REM latency were not significantly different from same-aged peers. Studies assessing sleep architecture in adults with CFS have revealed generally variable findings.8 Consistent with current results, reduced REM sleep has been reported in some studies in adults with CFS. However, this finding is not universal and some studies have reported a higher percentage of REM sleep relative to controls.8 In contrast to current findings, most studies have found that adults with CFS experience reduced SWS relative to controls.8 Given that only one case-control study has investigated the sleep architecture in pediatric CFS, further research is required to clearly understand the nature and extent of alterations in sleep stage dynamics in children and adolescents with CFS. The possibility that pediatric CFS may not be associated with a unitary sleep profile is alluded to by Ohinata and colleagues17 and proposed in the adult CFS literature.21 It is worth considering that sleep in pediatric CFS may in fact be heterogeneous and a number of sleep phenotypes may explain the varying symptomatology. Therefore, future research that further characterizes the possible heterogeneity of sleep in CFS will be useful.
CFS is a debilitating and complex condition with numerous symptoms and comorbidities. Although current research is beginning to elucidate the nature of sleep disturbances in adults with CFS, this review highlights the paucity of research focusing on sleep in pediatric CFS. Given the unique sleep behaviors related to physiological and developmental differences of children and adolescents, greater focus on this developmental stage in CFS sleep research is necessary. Further, existing literature is hampered by many methodological concerns, which include variable use of case definitions, limited reporting of important factors such as duration of illness and comorbidities, and unimodal approaches to measuring sleep. Reporting of medical and/or psychiatric status and medication use varied considerably among the papers (and were often not reported at all), making it difficult to evaluate the role of these variables in sleep related outcomes and complicating comparisons between studies. It will be crucial for future studies to collect and report sufficient information about their study samples to allow generalizability of findings to the larger pediatric CFS population. Studies have not adequately investigated factors that may affect sleep quality and quantity in pediatric CFS. It has been proposed that multiple factors interact to contribute to the development and maintenance of poor sleep in both healthy and chronically ill populations.22 Specifically in CFS, disease-related factors, such as fatigue and pain symptoms, may interact with developmental, physiological, psychological/behavioral, and sociocultural factors to affect sleep quality and quantity (Figure 2). Through these complex bidirectional interactions with sleep, participation and quality of life may be affected. Applying such a model to future research in CFS will help us to understand the main factors that affect sleep in pediatric CFS and how they interact, and hence guide the development of targeted interventions for this vulnerable group.
Proposed conceptual model: In chronic fatigue syndrome, physiological, disease-related, developmental, psychological/behavioral, and socio-cultural factors interact to influence sleep quality and quantity.
These factors and the complex interactions between them in turn influence participation and quality of life.
Proposed conceptual model: In chronic fatigue syndrome, physiological, disease-related, developmental, psychological/behavioral, and socio-cultural factors interact to influence sleep quality and quantity.These factors and the complex interactions between them in turn influence participation and quality of life.
In terms of physiological and disease-related factors, existing studies suggest that circadian rhythm variations in CBT, melatonin, or cortisol may underlie sleep disturbances in pediatric CFS; however, due to methodological limitations these findings are inconclusive and replication in larger, well-designed studies is required to confirm these findings. Although some support for cortisol and melatonin variations can be found in the adult CFS literature,23–26 the circadian rhythm of CBT in adults with CFS has been found to be indistinguishable from healthy controls.27,28
The proposed complex relationships between sleep and pain in CFS also warrant investigation, given the frequency at which children and adolescents with CFS experience pain symptomatology. Pain has shown to be an important contributor to sleep disturbances in children and adolescents with chronic pain.29 A recent systematic review of sleep disturbances in pediatric chronic pain29 proposed a complex, bidirectional relationship between sleep and pain that is influenced by physiological and mood variables. The interaction between pain and sleep is yet to be investigated in the pediatric CFS population; however, the coexistence of fibromyalgia was not found to be associated with increased sleep disturbance in adults with CFS.30 The possible unidirectional or bidirectional relationships between sleep and comorbid mood disorders in CFS have not been investigated. With emerging evidence that sleep problems may represent a precursor to the onset of depression and anxiety among adolescents,31,32 exploring these relationships in patients with CFS will be an important direction for future research given the heightened risk of mood disorders in this population.33,34 Further, the sleep hygiene practices of children and adolescents with CFS, and how they may contribute to disturb sleep, have not yet been investigated. Given the potential for effective behavioral intervention for poor sleep hygiene,35 characterizing the sleep hygiene practices of children and adolescents with CFS represents an important future direction.
To date, studies investigating sleep in pediatric CFS have been cross-sectional in design. Future studies utilizing longitudinal designs will allow us to understand how changes over time in sleep in pediatric CFS compare to normal maturational sleep changes. A longitudinal design would also provide opportunities to explore of the directionality and complexity of relationships between sleep and developmental, physiological, psychological/behavioral, sociocultural, and disease-specific factors. Further, the functional consequences of sleep disturbances in pediatric CFS have not been investigated, with the effect of sleep disturbances on participation, school function, and quality of life, warranting investigation.
Methodological limitations inherent to a systematic review of this nature should be acknowledged. Although PRISMA guidelines were carefully adhered to, there is a possibility that some studies may not have been detected in the search phase. Further, limiting inclusion to English language and peer-reviewed publications may introduce a potential publication bias. Limited reporting on disease and sociodemographic information by the included studies cautions the generalizability of these data to the wider pediatric CFS population.
Further research is required to inform clinical guidelines into the identification and management of sleep disturbances in pediatric CFS. Establishment of evidence-based guidelines regarding the treatment of sleep disturbance in pediatric CFS awaits further high quality research into the nature of sleep disturbances, as well as well-designed randomized controlled trials to determine the efficacy of treatments in this particular patient group.
Current review findings suggest that sleep disturbances occur in children and adolescents with CFS. However, the limited evidence available in conjunction with its low methodological quality necessitates caution in the interpretation of results of the current review. Given the potential complex relationship between CFS and sleep, this review highlights the pressing need for more high-quality and well-designed research to focus on this area in order to inform clinical guidelines. Combining both validated subjective and objective measures in future studies will provide a more comprehensive understanding of sleep in this patient group. Research needs to focus on identifying the specific types, severity, and causes and contributory factors, of sleep disturbances, as well as the most appropriate treatment strategies for sleep problems in children and adolescents with CFS. Improving sleep in children and adolescents with CFS may consequently improve participation, functional outcomes and quality of life in this vulnerable patient group.
This study was funded by the Mason Foundation through the Australian and New Zealand Banking Group Limited Trustees and the Murdoch Childrens Research Institute. Research at the Murdoch Childrens Research Institute is supported by the Victorian Government's Operational Infrastructure Support Program. The authors have indicated no financial conflicts of interest.