Article Text
Abstract
Objective The gold standard assessment for sleep quality is polysomnography (PSG). However, actigraphy has gained popularity as an ambulatory monitor. We aimed to assess the value of actigraphy in measuring sleep fragmentation in children.
Methods 130 children aged 2–18 years referred for assessment for sleep disordered breathing (SDB) were recruited. The arousal index (AI) scored from PSG was compared to the actigraphic fragmentation index (FI) and number of wake bouts/h.
Results The ability of actigraphic measures to correctly classify a child as having an AI>10 events/h rated as fair for the FI and poor for wake bouts/h (area under the receiver operator characteristic curve, 0.73 and 0.67, respectively).
Conclusion Actigraphy provides only a fair indication of the level of arousal from sleep in children. While the limitations of actigraphy prevent it from being a diagnostic tool for SDB, it still has a role in evaluating sleep/wake schedules in children.
Statistics from Altmetric.com
Evaluation of sleep disturbances in children often involves parental reports. However, parents may not be able to accurately assess sleep disturbance throughout the night, highlighting the need for objective ambulatory measurement of sleep fragmentation in children. Actigraphy is the recording of movement via an accelerometer to identify periods of sleep versus wake, which has gained increasing popularity as an ambulatory monitor. Our group has recently validated actigraphy for determining sleep and wake in children with sleep disordered breathing (SDB).1
Actigraphy analysis can also provide information on sleep quality, such as the amount of sleep fragmentation. Examples of actigraphic measures of sleep fragmentation include the total number of periods of wakefulness during sleep (wake bouts) and the sleep fragmentation index (FI).2 The aim of this study was to evaluate how well the actigraphic measures of sleep fragmentation (wake bouts/h and FI) reflect the arousal index measured from polysomnography (PSG) in a clinical paediatric population.
Methods
Subjects
A total of 130 children aged 2–18 years referred for investigation of SDB were studied. The Monash Medical Centre Human Ethics Committee granted ethical approval for this project. Written informed consent was obtained from parents.
Physiological measurements
Electroencephalography (C4/A1, O2/A1), electrooculograms and submental electromyogram (EMG) were recorded. Continuous electrocardiogram was recorded to provide instantaneous heart rate (HR). Leg EMG, oronasal airflow (nasal cannula and oronasal thermistor; Compumedics), thoracic and abdominal movements (Resp-ez Piezo-electric sensor, EPM Systems), oxygen saturation (Biox 3700e Pulse Oximeter, Ohmeda) and both end tidal and transcutaneous carbon dioxide (PetCO2: Capnocheck Plus, BCI Inc; TCO2: TCM3, Radiometer) were also recorded. Recordings were made on a Compumedics Series-S Sleep System. An actigraph (Actiwatch AW64, Mini Mitter Company) was placed on the non-dominant wrist. The time on the actigraph was synchronised to the time on the polygraphic trace.
Data analysis
Polysomnography
Sleep staging was performed according to standard criteria. Arousals were scored according to the criteria of the American Sleep Disorders Association, as well as autonomic/behavioural criteria (≥2 of an increase in submental EMG, increase in HR or gross body movement3), to yield a combined cortical and subcortical arousal index (AI). Respiratory events ≥2 respiratory cycles in duration were scored according to current paediatric guidelines, and an obstructive apnoea hypopnoea index (OAHI) was calculated. Diagnostic criteria for classification of SDB severity followed our current clinical practice: primary snoring (PS; OAHI<1 event/h); mild obstructive sleep apnoea (OSA; OAHI between 1 and 5 events/h) or moderate/severe OSA (OAHI>5 events/h).
Actigraphy
Actigraphy was analysed using the lights out/lights on and sleep onset/sleep offset data obtained from the scored PSGs. Data from the actigraph were coded into sleep and wake in 30-s epochs using commercially available software (Actiware®-Sleep V.3.3, Mini Mitter Company). Data were analysed at the automatic activity threshold setting.1 The FI was calculated as the percentage of epochs moving within the sleep period, plus the percentage of immobility phases of 1 min out of all immobility phases.2 Thus, the FI is the addition of two percentages and is reported in arbitrary units. The total number of wake bouts/h was calculated as the actual number of episodes of wakefulness divided by the sleep period.
Statistics
Subject demographics, polysomnographic and actigraphic characteristics across the severities of SDB were compared using one-way ANOVA on ranks with Dunn's post hoc testing or one-way ANOVA with Bonferroni post hoc comparisons where appropriate. Sex distribution across the severities of SDB was compared using χ2 analysis. Data are presented as mean±SEM or median and interquartile range (IQR). Statistical significance is set at p<0.05.
The FI and wake bouts/h were analysed singly for their ability to distinguish children with AI>10 events/h and OAHI>1 event/h. For each actigraphic parameter, a receiver operator characteristic (ROC) curve was generated. Values for the area under the ROC curve (AUC) were used as indicators of performance. A guide for classifying the accuracy of a diagnostic test is 0.90–1=excellent, 0.80–0.90=good, 0.70–0.80=fair, 0.60–0.70=poor, <0.60=fail. Various cut-off values for each parameter were tested with regard to sensitivity and specificity; those reported are at a cut-off where most subjects were correctly classified (ie, the highest agreement rate). The positive predictive value and negative predictive value were also calculated.
Results
Subject demographics and polysomnographic characteristics across the severities of SDB are shown in table 1. The AI and OAHI were significantly increased in the mild OSA and Mod/Sev OSA groups compared to PS (p<0.05 for all). There was a significant difference in the FI across the three severities of SDB (p<0.01). Post hoc analysis revealed that the FI was significantly increased in the Mod/Sev OSA group compared with the PS group (p<0.05). There was no difference in wake bouts/h across the three severities of SDB.
The performance of FI and wake bouts/h in correctly classifying a child as having an AI>10 events/h or an OAHI>1 event/h varied considerably (table 2). The AUC ranged from 0.59 to 0.73. The best performance was FI (at a cut-off value of 17.8) detecting AI. Thus, the actigraphy measures unfortunately rate overall as “poor” with the best measure (FI for AI) rating as “fair”.
Discussion
We found that the actigraphic FI, often regarded to reflect restlessness and sleep fragmentation in adults, was increased in children with moderate/severe OSA compared with primary snorers but rated as a poor diagnostic test for OSA and only a fair diagnostic test for a high arousal index.
Similarly, wake bouts/h performed as a poor measure of sleep disruption in children. A limitation of this measure is that epochs with short arousals may still be defined as “sleep” by the actigraphic sleep algorithm and therefore not counted. The FI, on the other hand, does not rely on the sleep algorithm but rather on the movement count, and so epochs with a small amount of movement that are still determined to be “sleep” via the scoring algorithm are still taken into account.
The FI, which measures both movement and short periods of immobility,2 has been used by many researchers as an index of restlessness during sleep, for example, in many pharmacological studies.4 Now for the first time, the FI has been directly compared with polysomnographic data in children and has only been shown to be a fair index of sleep fragmentation as measured by the arousal index. However, while the FI is not a powerful measure of traditional sleep fragmentation, it may be providing a more global measure of restlessness in children.
Our data show that although the FI increased with increasing severity of SDB, the crossover between groups makes it a poor tool to discriminate OSA severity in children. This study supports the view that actigraphy cannot accurately determine the presence or absence of breathing abnormalities.5
In conclusion, actigraphy provides only a fair indication of the level of sleep fragmentation in children, as compared with PSG-derived arousal index. PSG remains the best test to determine sleep fragmentation in children, but the benefits of actigraphy as an ambulatory monitor do not rule it out as an adjunct test. While the limitations of actigraphy prevent it from being a diagnostic tool for SDB, it still has a role in evaluating sleep/wake schedules in children.
Acknowledgments
The authors would like to thank all children and their parents who participated in this study. The authors would also like to thank Ms Nicole Verginis and all the staff of the Melbourne Children's Sleep Unit for invaluable technical assistance.
Footnotes
-
Funding Dr O'Driscoll was part-funded by the Kaarene Fitzgerald Research Fellowship from SIDS & Kids Victoria.
-
Competing interests None.
Ethics approval This study was conducted with the approval of the Monash Medical Centre Human Ethics Committee.
Provenance and peer review Not commissioned; externally peer reviewed.
Patient consent Obtained..