Abstract
Objective To investigate sleep quality in juvenile fibromyalgia syndrome (JFS) and its effect on the global burden of the disease.
Methods Consecutive patients with JFS who performed full-night polysomnography (PSG) were included in this cross-sectional study. JFS-related symptoms, neuropsychiatric features, and sleep quality were assessed using self-report measures. PSG sleep parameters, including N3 distribution index, were obtained from patients and age-matched healthy controls.
Results We included 25 patients (20 females, median age 15.7 yrs). Nonrestorative sleep was reported by 22 of 25 (88%) patients. Patients with JFS showed significantly longer sleep period time (P = 0.004) and increased wake after sleep onset (P = 0.03) compared to healthy peers. The N3 distribution index was significantly lower in patients than in the control group (P = 0.02). Subjective poor sleep quality was related to Widespread Pain Index (WPI; rs −0.65), Symptom Severity Scale (rs −0.64), depressive symptoms (rs −0.58), fatigue (rs −0.44), and symptom severity upon awakening (rs −0.65). The N3 distribution index was correlated to depressive symptoms (rs 0.41) and irritability (rs 0.40). On multiple regression analysis, WPI was predicted by subjective sleep quality (β −0.32, P = 0.04), whereas depressive symptoms were predicted by subjective sleep measures (β −0.32, P = 0.04) and PSG parameters (N3 min: β −0.07, P = 0.03).
Conclusion Sleep complaints are a key hallmark of JFS and have significant effect on relevant clinical domains of the disease, such as pain and depression.
Juvenile fibromyalgia syndrome (JFS) is a disabling condition characterized by widespread diffuse musculoskeletal pain, fatigue, and mood disturbances that cause functional impairment and have a marked effect on the patients’ quality of life.1 JFS is more common in girls than in boys, and has a mean age at symptoms onset of 11.4 years to 13.7 years.2 The estimated prevalence of JFS ranges from 1.2% to 6.2%,3,4 making this condition an emerging healthcare problem in female adolescents.
Sleep disturbances are a well-known component of the clinical spectrum of fibromyalgia (FM), and are reported in 75% to 90% of adult patients5-7 and in 67% to 96% of patients with JFS.1,2,8,9 Self-reported subjective sleep problems in FM and JFS include difficulty falling asleep, sleep fragmentation with multiple arousals, nonrestorative sleep, and excessive daytime somnolence.5-8 These subjective sleep complaints are substantiated by objective sleep-wake patterns on polysomnography (PSG) examination. Indeed, PSG and electroencephalography (EEG) spectral analysis have revealed greater sleep onset latency (SOL), an increased arousal index, decreased total sleep time (TST) and sleep efficiency (SE), reduced percentage of slow-wave sleep (SWS), and increased alpha wave activity during SWS (the alpha-delta sleep) in patients with FM compared to control subjects (CSs).10-15 These PSG findings suggest a disorder of homeostatic and circadian mechanisms during sleep in patients with FM. A wide range of evidence indicates a significant relationship between sleep disturbances and FM-related symptoms, such as unrefreshing sleep, sleepiness, fatigue, lower pain thresholds, mood swings, and cognitive impairment.16-18
In contrast to adults, JFS has been investigated less extensively in regard to sleep characteristics. Sleep assessment in children with chronic pain has primarily relied on subjective measures of sleep quality8,9,19,20 and, to date, only 3 studies that we know of have used PSG to investigate the pathophysiological bases of sleep disturbances in JFS.21-23 Similarly to FM, abnormal sleep patterns in JFS included shortened TST, prolonged sleep latency, decreased SE, longer awake periods during sleep, and alpha intrusions into SWS.21,22
However, there is conflicting evidence regarding the effect of sleep disturbances on JFS-related symptoms. Roizenblatt et al have reported an association between alpha-delta sleep with subjective sleep complaints and musculoskeletal pain in 34 patients with JFS.21 These results have been questioned by Olsen et al who found that alpha-delta sleep was unrelated to pain, disability, or subjective sleep difficulties in an observational study involving 10 patients with JFS.23 Overall, however, the available data are insufficient to establish a causal relationship between sleep quality and JFS-related symptoms.
The purposes of the present study are to investigate sleep quality and describe the PSG sleep patterns of adolescents with JFS. We also aimed to determine the effect of sleep disturbances on the global burden of the disease by exploring their relationships with JFS-related symptoms.
METHODS
Patients with JFS meeting the 2010 American College of Rheumatology Adult Fibromyalgia Criteria24,25 were recruited from the Pediatric Rheumatology Department at the Gaslini Children’s Hospital of Genoa, Italy. All consecutive patients who performed a full-night PSG between 2019-2020 were included in this cross-sectional study. Patients who had received sleep medications within 1 week prior the study or had comorbid illnesses that could affect sleep were excluded.
The Widespread Pain Index (WPI), the Symptom Severity Scale (SSS), as well as the total number of tender points25 at the time of inclusion in the present study were recorded. The WPI includes a list of 19 painful areas and reflects the bodily distribution of pain. The SSS comprises 2 parts: part SSS2a evaluates the severity of fatigue, waking unrefreshed, and cognitive symptoms using a 4-point Likert scale (total score range: 0-9). Part SSS2b includes a checklist of 41 somatic symptoms (score range: 0-3, based on the number of symptoms reported by the patients: 0 symptoms [score of 0], 1-10 symptoms [score of 1], 11-24 symptoms [score of 2], and ≥ 25 symptoms [score of 3]). The SSS is obtained from the sum of the SSS2a and SSS2b scores and ranges from 0 to 12.25 Tender point evaluation was performed at 18 standardized points by digital pressure.
Participants were asked to rate their average level of pain in the previous 2 weeks using the standard 100-mm pain numerical rating scale (NRS) that ranges from 0 mm (no pain) to 100 mm (pain as bad as it can be).26,27 Self-report measures (100-mm NRS) were also used to rate the severity of symptoms in the following domains: fatigue, headache, symptom severity upon awakening, irritability, and global assessment of disease severity.
Depressive symptoms were assessed using the Children’s Depression Inventory (CDI), a self-report scale with a total score ranging from 0 to 54.28 A CDI score ≥ 19 is considered the cutoff to identify patients with clinically significant depressive symptomatology. Anxiety symptoms were evaluated using the Multidimensional Anxiety Scale for Children (MASC), a self-report questionnaire which consists of 39 items. The total score is pathological if ≥ 60.29 Perceived sleep quality was assessed using the Sleep Condition Indicator (SCI), an 8-item, patient-reported questionnaire based upon Diagnostic and Statistical Manual of Mental Disorders (DSM-5) criteria.30 Scores of these 8 components are summed to obtain a global score ranging from 0 to 32; a score ≤ 16 identifies an insomnia disorder.30 Daytime sleepiness was assessed using the Epworth Sleepiness Scale for Children and Adolescents, a self-administered 8-item questionnaire. The total score is marked out of a maximum of 24, and a score > 10 is considered abnormal and indicative of excessive daytime sleepiness.31
Patients underwent a full-night PSG recording at their home. The recording montage included: F4-M1, C4-M1, O2-M1, F3-M2, C3-M2, and O1-M2 electroencephalograms, right and left electrooculograms for eye movement detection and rapid eye movement (REM) sleep identification, a bipolar mental electromyogram (EMG) for muscle tone measurement, nasal airflow, chest, abdominal, and leg movements (piezo-electric bands), arterial oxygen saturation, and a bipolar electrocardiogram (ECG). Patients were set up in the morning in the clinic, and the device recording time was aligned with the patients’ usual sleep schedule. After this procedure, patients spent the night at home in order to avoid the first-night effect.
All recordings were computed and scored using RemLogic software, in accordance with the scoring rules proposed by the American Academy of Sleep Medicine.32 PSG variables included sleep period time (SPT), TST, SOL, wake after sleep onset (WASO), SE, sleep staging with total minutes of non-REM (N1, N2, and N3) and REM sleep stages, and awakening index.. N1 corresponds to light transitional sleep and N2 to fully developed sleep that occurs throughout the sleep period. N3 (also referred to as delta sleep or SWS) corresponds to deep sleep; it is thought to be the most restful form of sleep and it is more represented in the first part of the sleep period. Time spent in stages N1, N2, N3, and REM was analyzed and expressed as a percentage of TST and SPT. In addition, wake time was defined as a relative percentage in the SPT that was computed as the elapsed time from sleep onset through to the last epoch of sleep. TST was computed as the total duration of epochs scored as asleep. SOL was computed as duration of time from light off to falling asleep. WASO, defined as the duration of nighttime awakenings, was computed as the total duration of epochs scored as wake after initial sleep onset and prior to final awakening. SE was computed as the proportion of time scored as asleep relative to the total time in bed. Arousals were defined as an abrupt shift in EEG frequency lasting from 3 seconds to 14 seconds and accompanied by ≥ 1 of the following changes in recording signals: increased chin or EMG activity, increased heart rate, ECG distortion, and distortion of the respiratory signals. Awakenings were computed as a shift from a sleep stage epoch to awake epoch. Awakening index was the number of awakenings per hour of sleep.
We also investigated the distribution of N3 in the course of the night by calculating the N3 distribution index, defined as ([number of epochs of N3 in the first half of SPT – number of epochs of N3 in the second half of SPT]/total number of N3 epochs).33 The N3 distribution index ranges from −1 to 1, with higher values being correlated with predominance of N3 sleep in the first half of the night, as expected in physiological sleep. All the aforementioned PSG variables were also extracted from a population of 27 CSs matched for age and sex of the patients. CSs underwent a full-night PSG recording at their home. All CSs had a negative history of musculoskeletal pain and underlying medical conditions including cardiac, neurological, or respiratory diseases that are known to have an effect on sleep architecture. None of them reported current sleep disturbances.
This study was approved by the institutional review board for human subjects research (CER Liguria: 333/2022 – DB id 12460). All participants and legal guardians provided the appropriate informed consent/assent for the use of recordings for research and publications purposes.
Statistical methods. Descriptive statistics were first performed; categorical data were reported in terms of absolute frequencies or percentages, whereas continuous quantitative data were reported in terms of median (IQR). Comparison of quantitative data was made by means of the nonparametric Mann-Whitney U test. Correlation between quantitative data was evaluated by means of the nonparametric Spearman correlation coefficient (rs); correlations were considered high if > 0.70, moderate if 0.40 to 0.69, and low if < 0.40.34
We hypothesized that in a multivariate model, increased pain, SSS, and anxiety/depressive symptoms could be predicted by poorer sleep quality. Multiple regression models for different outcomes (ie, WPI, SSS, fatigue, CDI, MASC, etc.) and selected independent variables (ie, subjective and objective sleep indicators) were performed. Independent variables that were clinically relevant or were linearly correlated to the outcomes were included in multivariable models. β coefficients and standard errors of the models were calculated and reported for each model and the classical t tests were used to evaluate the β coefficients. R2 was used as an indicator of goodness-of-fit.
All statistical tests were 2-sided, and P values < 0.05 were considered statistically significant. The software “Statistica” (release 9.0, StatSoft Corporation) was used for descriptive and bivariate analyses.
RESULTS
Twenty-five consecutive patients with JFS, whose demographic and clinical features are reported in Table 1, were included in the study. Twenty patients were not receiving any medications before and during PSG. Of the remaining patients, 2 discontinued their medications (1 using amitriptyline and 1 using amitriptyline + clonazepam) 7 days before the study and 3 elected to continue their medications (1 using sertraline, 1 using gabapentin, and 1 using duloxetine).
Description of the study patients.
We also included 27 CSs (19 females and 8 males; median age 15.3 [IQR 14.9-16.0] yrs).
Eighteen patients with JFS (72%) complained of subjective poor sleep quality according to the SCI. Sleeping problems included difficulties initiating sleep or frequent awakenings, as reported in Table 2. Nonrestorative sleep (ie, awakening tired with the sensation of not having slept during the night) was reported by 22 patients (88%). Although excessive daytime sleepiness was reported in only 4 patients (16%), 9 patients (36%) reported regular napping in the afternoon.
Results of the SCI in the 25 patients with JFS. The last column represents the percentage of patients whose answers are within the pathological range.
Table 3 shows the comparison of full-night PSG between patients with JFS and age-matched CSs. Patients with JFS showed a significantly longer SPT (P = 0.004) and increased WASO (P = 0.03) compared to their healthy peers. No differences in SE, number of arousals, or sleep latency were found in patients with JFS compared to CSs. Although the time spent in N3 sleep stage did not differ between patients and CSs, the N3 distribution index was significantly lower in patients with JFS (0.5) than in the control group (0.6; P = 0.02), indicating a higher number of epochs of N3 during the second part of the night.
Results of the full-night polysomnographic recording: comparison between patients with JFS and controls.
As reported in Table 4, subjective poor sleep quality and daytime sleepiness were related to bodily distribution of pain (rs −0.65 and 0.50, respectively), increased SSS (rs −0.64 and 0.61, respectively), depressive symptoms (rs −0.58 and 0.54, respectively), fatigue (rs −0.44 and 0.51, respectively) and symptom severity upon awakening (rs −0.65 and 0.49, respectively). Reduced N3 sleep stage was related to SSS (N3 min: rs −0.41; N3 % SPT: rs −0.46) and CDI scale (N3 min: rs −0.40). The N3 distribution index correlated to depressive symptoms (rs 0.41) and irritability (rs 0.40).
Correlations between subjective and objective sleep quality measures and the JFS-related symptoms.
On multiple regression analyses (Table 5), the number of painful body regions (WPI) was predicted by subjective poor sleep quality (β −0.32, P = 0.04), whereas depressive symptoms were predicted by both subjective poor sleep quality (β −0.32, P = 0.04) and objective PSG measures (N3 min: β −0.07, P = 0.03; N3 distribution index: β 11.54, P = 0.002).
Multiple regression analyses.
DISCUSSION
This study confirms that disturbed sleep is a key hallmark of JFS and provides important insights into the effect of sleep complaints on relevant clinical domains of the disease, such as pain and depression. In line with previous studies,1,2,8,9 the majority of our patients complained of difficulties in falling asleep, frequent awakenings, nonrestorative sleep, and excessive daytime somnolence. More than 70% of patients had an abnormal SCI score and satisfied the criteria for insomnia disorder according to DSM-5.
Of note, these results do not appear to be fully reflected by the objective measures of sleep obtained from PSG recording. In contrast to previous studies,21,22 our patients did not differ in SE, number of arousals, SOL, and sleep staging from CSs, except for WASO. Although the last variable was significantly increased in our patients as compared to healthy controls, the TST appeared to be preserved, resulting in an overall increase in the total SPT.
The discrepancy between subjective and objective sleep variables suggests the presence of a certain amount of sleep misperception in patients with JFS. This should always be kept in mind when evaluating these patients, especially when considering the pharmacological treatment of sleep complaints in this population.
Although JFS might present with a set of symptoms and PSG findings distinct from adult forms, we cannot exclude that the preservation of the TST might disappear as disease progresses, leading to the same PSG changes described in adult patients. In support of this hypothesis, Roizenblatt et al found that PSG alterations are less prominent in children with JFS than in their affected mothers, which suggested that sleep disturbances may worsen over time.21 In line with these observations, the short disease duration of our patients might be explained in part by the lack of major alterations in sleep macrostructure as revealed by PSG examinations. As TST and total N3 sleep appear to be quantitatively similar to CSs, morning symptoms reported by patients cannot be explained by sleep deprivation. Although sleep macrostructure was preserved, sleep EEG in our patients showed a significantly different distribution of the N3 sleep stage throughout the night. In fact, although the time spent in N3 was similar between patients and CSs, the N3 distribution index was significantly lower in patients with JFS, indicating a higher representation of N3 sleep during the second part of the night.
The N3 stage, also referred to as delta sleep or SWS, is the deepest and most restorative sleep stage. Physiologically, episodes of N3 are longer at the beginning of the night and shorten gradually as the night progresses.32 The altered distribution of N3 sleep in our patients could be related to the increased length of wakefulness periods during the night. An alternative hypothesis is that the higher amount of N3 epochs during the second part of the night might be the expression of an impairment in the process of physiological release of the homeostatic drive to sleep. Regardless of the mechanism, this shift of N3 sleep could contribute to explaining the severity of morning symptoms reported by our patients. Indeed, an increased amount of N3 sleep during the second part of the night might enhance the physiological phenomenon of sleep inertia, thereby exacerbating transitory hypovigilance, confusion, and impaired cognitive and sensory–motor performance that physiologically follows awakening.35
In support of the major role of N3 in the restorative process, we found a significant inverse correlation between N3 sleep duration and the SSS, which evaluates the severity of fatigue, waking feeling unrefreshed, and cognitive symptoms. The relationship of FM symptoms with N3 sleep deprivation is well-established in adult patients.6,7,10,11,21 Our results confirm the pivotal role of N3 sleep for energy restoration, providing a reasonable explanation for nonrestorative sleep in JFS.
We found that self-reported sleep complaints were significantly and independently correlated to the WPI, which reflects the distribution of pain throughout the body. Similarly, Roizenblatt et al found that alpha-delta sleep was significantly associated with the number of tender points and inversely related to pain thresholds in patients with JFS.21 Moreover, there is preliminary evidence that multidisciplinary pain treatment programs, including intensive physical exercise, produce subjective improvement in sleep quality in patients with JFS.23 Collectively, these results confirm the interplay between sleep and pain, and provide support to nonpharmacological treatment strategies aimed at enhancing sleep quality in JFS. This approach could also help to reduce the risk and severity of comorbid psychiatric disorders, such as anxiety and depression, which represent a significant problem in JFS.1,8,36 In fact, anxiety and depression disorders have been correlated with school absenteeism and functional disability in JFS.37,38 An alarming finding is that depression and anxiety have been associated with a high risk of suicide in a cross-sectional study involving 31 patients with JFS.39 Therefore, identification of predictors of anxiety and depressive symptoms in JFS is of utmost importance to improve treatment outcomes.
Our results suggest a relationship between sleep disturbances and depressive symptoms in JFS. In fact, both subjective and objective sleep measurements were significantly and independently correlated with the CDI score. They also confirm the results of previous studies in FM, which provided evidence of a bidirectional link between sleep disorders, depression, and anxiety.40,41 Such a link could be mediated by common pathophysiological mechanisms, including alterations in serotoninergic activity and melatonin secretion,42,43 although further studies are needed to establish the degree of this possible bidirectional relationship. Because sleep complaints and depressive symptoms are known to have a significant effect on daily activities and pain thresholds,37,38,41 studies examining the interrelationship between these variables over time could help to develop more effective therapeutic strategies for JFS.
There are some limitations of our study that should be addressed in future research. First, the statistical power is limited by the small sample size. Second, spectral analysis was not performed and, thus, data on sleep microstructure are lacking. Such data could improve our understanding of the mechanisms underlying poor sleep quality in JFS, in light of the previous descriptions of alpha-delta sleep phenomenon in FM.10,11 This altered sleep pattern is well described in adult patients with FM, but not in the juvenile form of the disease. Third, a potential influence of medications on the study results cannot be excluded despite discontinuation of therapies before patient inclusion into the study. Finally, we only made single timepoint PSG recordings. Longitudinal interventional studies are needed to elucidate whether interventions addressing sleep disturbances may improve pain, mood disorders, and emotional dysfunction in adolescents with JFS. The strengths of our analysis include the use of objective measurements of sleep quality and the age- and sex-matched control group used for PSG examinations. This requirement is essential because of the differences in sleep patterns related to age and pubertal status.44
In conclusion, our study is among the first to combine self-reported measurements and PSG examinations in the evaluation of sleep quality in JFS, and to demonstrate the effect of sleep disruption on the global burden of the disease. Although we found a discrepancy between subjective sleep measures and PSG findings, our results revealed consistently that more sleep alterations were associated with more severe symptoms across a variety of disease-related domains, especially those concerning pain and depression. Elucidation of the influence of sleep complaints on disease severity could provide insights for the development of more comprehensive therapeutic approaches. Longitudinal studies are needed to demonstrate whether treatment of sleep disturbances may help to decrease the severity of JFS-related symptoms and improve disease outcomes.
Footnotes
This study was supported by the Italian Ministry of Health, 5 × 1000 project 2017 “SOLAR: Sleep disorders in children: an innovative clinical research perspective.”
The authors declare no conflicts of interest relevant to this article.
- Accepted for publication December 12, 2022.
- Copyright © 2023 by the Journal of Rheumatology