Comparison of 7 versus 14 days wrist actigraphy monitoring in a sleep disorders clinic population

Abstract

Wrist actigraphy is a valid measure to assess sleep and circadian rhythm abnormalities. It is listed in the diagnostic criteria for sleep disorders where single night polysomnography is insufficient (ICSD-2). However, an optimal recording time remains unclear. We hypothesised that seven days would provide sufficient data for analysis, similar to recordings for 14 days. We analysed three consecutive years of actigraphy data obtained within a tertiary sleep referral centre. Data were recorded continuously for two weeks using an AW4 actiwatch (Cambridge NeuroTechnology, Cambridge, UK; Mini Mitter Co, Sunriver, OR). Parameters, including sleep efficiency (SE), sleep latency (SL), sleep fragmentation index (SFI), total sleep time (TST) and wake after sleep onset (WASO) were analysed using GraphPad Prism (Version 5.02, GraphPad Software Inc, San Diego, CA) and classified into week one, week two and an overall average for the duration of 14 days. In addition, two experienced consultants working in the sleep laboratory compared the results of week one versus week two independently, visually analysing the data for circadian rhythmicity and fragmentation of the pattern, allowing calculation of the intraclass correlation coefficient (ICC), κ. The actigraphies of 239 patients (51.9% male; age 42 (16) years) were analysed. There was no difference in SE, SL, SFI or WASO between week one, week two and 14 days average recording. A small difference was found between TST week one (399.9 minutes, 95% CI 389.9–409.9 minutes) and TST week two (388.7 minutes, 95% CI 378.3–399.1 minutes), but not between TST for 14 days average recording (394.3 minutes, 95% CI 384.7–403.9 minutes) and either week. Independent scorers achieved a good agreement in the rhythmicity of the sleep pattern (ICC κ 0.734, p < 0.001) and a low agreement for the fragmentation of the pattern (ICC κ 0.380, p < 0.001). One week of wrist actigraphy recording provides similar data to two week actigraphies, despite subtle differences between the weeks. One week wrist actigraphy could be recommended as standard compared to longer recordings to maximise efficiency of the clinical service. Further studies are required to validate our results in specific clinical subgroups.

• Actiwatch
• circadian rhythm
• insomnia
• polysomnography
• sleep apnoea
• sleep efficiency

Acknowledgments

We acknowledge the help of Dr Abdul Dauri, Senior Lecturer in Medical Statistics, Kings College London.