Internight sleep variability: its clinical significance and responsiveness to treatment in primary and comorbid...

Although sleep diary and actigraphy data are usually collected daily for 1 or 2 weeks, traditional analytical approaches aggregate these data into mean values. Internight variability of sleep often accompanies insomnia. However, few studies have explored the relevance of this 'construct' in the context of diagnosis, clinical impact, treatment effects and/or whether having 'variable sleep' carries any prognostic significance. We explored these questions by conducting secondary analyses of data from a randomized clinical trial. The sample included primary (PI: n = 40) and comorbid insomnia (CMI: n = 41) sufferers receiving four biweekly sessions of cognitive-behavioural therapy (CBT) or sleep hygiene education. Using the within-subject standard deviations of diary- and actigraphy-derived measures collected for 2-week periods [sleep onset latency (SOL), wake after sleep onset (WASO), total sleep time (TST) and sleep efficiency (SE)], we found that CMI sufferers displayed more variable self-reported SOLs and SEs than PI sufferers. However, higher variability in diary and actigraphy-derived measures was related to poorer sleep quality only within the PI group, as measured by the Pittsburgh Sleep Quality Index (PSQI). Within both groups, the variability of diary-derived measures was reduced after CBT, but the variability of actigraphy-derived measures remained unchanged. Interestingly, the variability of actigraphy measures at baseline was correlated with PSQI scores at 6-month follow-up. Higher SOL variability was associated with worse treatment outcomes within the PI group, whereas higher WASO variability was correlated with better treatment outcomes within the CMI group. Sleep variability differences across insomnia diagnoses, along with their distinctive correlates, suggest that mechanisms underlying the sleep disruption/complaint and treatment response in both patient groups are distinct. Further studies are warranted to support variability as a useful metric in insomnia studies.