Lauderdale et al. examined the association between perceived fair/poor health and sleep duration by several methods [1]. The authors concluded that U-shaped relationship between sleep duration and prevalence of fair/poor health was observed only with measuring sleep with survey sleep hours and survey calculated sleep time. In contrast, there was no association between long sleep duration and increased prevalence of fair/poor health by actigraphy. However, I am not fully convinced by their arguments.

First, the authors well understand the limitation of actigraphy, named Actiwatch Spectrum, for sleep evaluation and the need of validation study of actigraphy against sleep polysomnography. There is a difference between brain activity and physical movement during sleep, and the discrepancy of sleep parameters between polysomnography and actigraphy is obvious for insomniacs [2]. When calculating total sleep time by actigraphy, the authors selected default sleep/awake sensitivity setting (40 counts per minutes), and also carried out a sensitivity analysis with a lower threshold of 20. Although the correlation coefficient between sleep duration with different threshold setting was greater than 0.99, actual sleep duration differs 18 minute in an average. Kushida et al. reported the best sleep/awake threshold of Actiwatch for detecting wakefulness as "high-sensitivity" setting (20 counts per minutes) [3]. Peterson et al. adopted default sleep/awake sensitivity setting of Actiwatch, and described the overestimation of total sleep time and underestimation of wake-after sleep onset [4]. These reports present that level of sleep/awake threshold is important for estimating sleep duration by actigraphy.

Second, the numbers of subjects in each category of sleep duration seems useful information in their study. The authors described the mean value of sleep duration by survey sleep hours and by actigraph total sleep time were 7.5 hours and 7.2 hours, but the prevalence of fair/poor health in subjects with sleep duration >9 hours by survey sleep hours was two- fold higher than that by actigraph total sleep time. The correlation coefficient between sleep duration by survey sleep hours and by actigraph total sleep time was 0.29, and I suspect that some subjects with long sleep duration by actigraphy do not actually keep enough sleep duration.

Anyway, sleep polysomnography study is required to confirm the lack of U-shaped association between sleep duration and prevalence of fair/poor health.

References

1 Lauderdale DS, Chen JH, Kurina LM, et al. Sleep duration and health among older adults: associations vary by how sleep is measured. J Epidemiol Community Health 2015 Nov 3. doi: 10.1136/jech-2015-206109

2 Natale V, Leger D, Martoni M, et al. The role of actigraphy in the assessment of primary insomnia: a retrospective study. Sleep Med 2014;15:111-5.

3 Kushida CA, Chang A, Gadkary C, et al. Comparison of actigraphic, polysomnographic, and subjective assessment of sleep parameters in sleep- disordered patients. Sleep Med 2001;2:389-96.

4 Peterson BT, Chiao P, Pickering E, et al. Comparison of actigraphy and polysomnography to assess effects of zolpidem in a clinical research unit. Sleep Med 2012;13:419-24.

Conflict of Interest:

None declared