Association of rest-activity and light exposure rhythms with sleep quality in insomnia patients

The relevance of altered rest-activity rhythm (RAR) and light exposure rhythm (LER) in insomnia patients under natural conditions remains unclear. The aim of this study was to compare the parametric and nonparametric circadian variables of RAR and those of LER under natural conditions between insomnia patients and normal controls (NC) in a community-dwelling setting. The relationship of the nonparametric variables with sleep quality was also explored in both groups. Participants above 18 years old were recruited from three Public Health Centers in a rural area of Korea. Actigraphy (Actiwatch 2; Philips Respironics, Murrysville PA, USA) recording was conducted for 7 days. Subjects were eligible for our study if they had an insomnia disorder (ID) for at least 1 month. Actigraphy data of 78 normal control (NC) subjects (Age, 55.95 +/- 13.22 years) and 104 patients with insomnia disorder (ID) (Age, 62.14 +/- 12.34 years) were included for the analysis. Acrophases and amplitudes of RAR and LER were estimated using cosinor analysis. Interdaily stability (IS), intradaily variability (IV), and relative amplitude (RA) of these rhythms were determined using nonparametric methods. Parametric cosinor and nonparametric variables of RAR and LER were compared between the NC and ID groups. Generalized linear models (GLMs) were applied to evaluate the main effects of group and each nonparametric variable as well as a group by each variable interaction on the sleep onset latency (SOL), sleep efficiency (SE), and wake after sleep onset (WASO) reflecting sleep quality. Among sleep parameters, the ID group showed significantly lower SE and greater WASO than the NC group. There were no significant differences in the acrophase and amplitude of RAR and LER between the two groups. There were no significant differences in IV, IS, and RA of RAR and LER between the two groups either. GLMs for RAR revealed a significant interaction between the group and IS on the SOL (beta = -46.39, p < 0.01), indicating a negative relationship of the IS with SOL in ID unlike its positive relationship in NC. There were no significant main effects of IV on the SOL, SE, and WASO, but significant main effects of RA on the SE and WASO (beta = 63.65 and beta = -221.43, respectively, p < 0.01). GLMs for LER revealed no significant main effects of IS, IV or RA on the SOL, SE, and WASO, but significant interactions between group and RA on the SE and WASO (beta = 56.17 and beta = -171.93, respectively, p < 0.05), indicating a stronger positive relationship of the RA with SE in ID compared to NC, and a negative relationship of the RA with WASO in ID, unlike its positive relationship in NC. Although our study did not reveal group differences in circadian variables of RAR and LER, it suggested that the regularity of RAR could be positively associated with sleep initiation, while the robustness of LER could be positively associated with sleep maintenance in insomnia patients.