Actigraphy versus Polysomnography to Measure Sleep in Youth Treated for Craniopharyngioma

ABSTRACT

Objective/Background

Youth with craniopharyngioma are at increased risk for excessive daytime sleepiness and narcolepsy. Polysomnography (PSG) is the gold standard for diagnosing sleep disorders, but is time-intensive, costly, and does not offer an in vivo measure of typical sleep routine. We determined the sensitivity, specificity, and accuracy of actigraphy compared with PSG in measuring nocturnal sleep in pediatric craniopharyngioma.

Participants

Fifty youth with craniopharyngioma (age 3–20 years) were assessed by overnight PSG and concurrent actigraphy after surgical resection and before proton therapy.

Methods

PSG and actigraphy data were synchronized utilizing an epoch-by-epoch comparison method. Sensitivity, specificity, and accuracy were calculated using measures of true wake, true sleep, false wake, and false sleep. Bland-Altman plots were conducted to further assess level of agreement.

Results

Actigraphy was 93% sensitive (true sleep [TS]) and 87% accurate (ability to detect TS and true wake) in measuring sleep versus wakefulness and was a reliable measure of sleep efficiency (SE) and sleep latency (SL). Specificity (true wake) was poor (55%) and total sleep time (TST) was underestimated by an average of 15.1 min. Wake after sleep onset (WASO) was overestimated by an average of 14.7 min.

Conclusions

Actigraphy was highly sensitive and accurate and was a reliable measure of SE and SL. Although there were differences in TST and WASO measurements by actigraphy and PSG, our findings provide the basis for future studies on the use of actigraphy to monitor treatment response to wakefulness-promoting medications in youth with craniopharyngioma who demonstrate excessive daytime sleepiness.

Acknowledgments

The study was supported by the St. Jude Cancer Center and ALSAC. We acknowledge Breya Walker, MS, for data collection and Lisa Meltzer, PhD for her assistance with epoch-by-epoch coding. We also thank the patients and families who agreed to participate in this study.

Disclosure statement

The authors have no relevant financial or nonfinancial relationships to disclose.