This study presents comprehensive data on polysomnographic characteristics of adults from the general population with and without physician-diagnosed AD. All measured parameters were similar between individuals with and without AD except for the number of wakefulness and movement episodes and the percentage of REM sleep.

When comparing sleep parameters as measured in our study with normative values obtained from a meta-analysis of PSG parameters including more than 5000 healthy adults [14], we found lower values in study participants with AD than in the reference population for total sleep duration (363.4 vs. 394.6 min) and sleep efficiency (79.7 vs. 85.7%), but a higher value for sleep latency (17.1 vs. 15.4 min). They also showed higher scores for WASO (67.4 vs. 48.2 min), indicating a higher extent of sleep fragmentation. Regarding the stages of sleep, participants with AD showed higher scores for N1 (30.0 vs. 7.9%), but lower scores for N2 (38.9 vs. 51.4%) and REM sleep (12.0 vs. 19.0%) compared to the reference population. Scores for N3 stage were comparable (19.1 vs. 20.4%). In participants with AD, but also in the total study population, higher AHI scores in comparison with the reference population were found (8.9 vs. 2.9 and 10.7 vs. 2.9, respectively).

Few previous studies investigating sleep characteristics in individuals with AD have applied objective measures such as PSG or actigraphy. Actigraphy is a non-invasive method for the assessment of sleep-wake cycles by involving a small wrist-worn device, and measures have been found to be highly correlated with PSG findings in individuals with AD [3]. With respect to the interpretation of the parameters measured in the present study, comparative data are extremely limited. A recently published meta-analysis based on case-control and cohort studies using PSG or actigraphy in adults and children with AD identified seven studies with 173 patients with AD and 112 controls [13]. Overall analyses showed that patients with AD had decreased total sleep time and sleep efficiency, and prolonged WASO and REM latency compared to controls. Among the studies included in this review, two had performed actigraphy in adults with moderate to severe AD. In contrast to our findings, the first study including 14 patients and 14 controls found that patients with AD slept less, awoke more often, and spent more time awake during these waking episodes, resulting in lower sleep efficiency compared to controls [15]. Similarly, the second study based on data from 15 patients and 19 controls showed that patients with AD had lower total sleep time and lower sleep efficiency, and higher WASO compared to controls [16].

Our analyses revealed that participants with AD had a higher number of wakefulness and movement episodes than participants without AD. Likewise, a study investigating nocturnal movements using accelerometers among adults with pruritic dermatoses including AD found that patients had twice as many movements as controls [17]. REM sleep, the “dreaming” state, most often occurs in the early morning hours of a normal sleep cycle and is considered essential for memory consolidation [18]. As an indicator of sleep impairment, we found that participants with AD had a lower percentage of REM sleep compared to those without.

The association of AD with obstructive sleep apnea (OSA) is not well investigated yet. A recent cross-sectional investigation among 2648 U.S. adults reported an association of OSA symptoms with self-reported AD [19]. According to our descriptive data, mean AHI scores among participants with and without AD correspond to mild OSA. This finding probably reflects a peculiarity of our study population since individuals with sleep problems were potentially more likely to participate in the PSG examination.

There is strong evidence to suggest that disease severity and specific AD symptoms such as pruritus and scratching are significant contributors to sleep disturbances in adult patients with AD [4]. For example, a study applying PSG and actigraphy in 20 adults with AD revealed that increasing disease severity was related to more scratching and reduced sleep quality [20]. Another study among 35 adult patients with AD used PSG and an infrared video camera system to investigate nocturnal scratching over 112 nights [21]. That study found longer durations of scratching and being awake after a scratching in patients than in controls. Moreover, numerous studies based on self-reports demonstrated an association of disease severity and scratching with sleep disturbances in adults with AD [22,23,24,25,26,27].

In view of aspects of feasibility, self-reported measures are important tools for the assessment of the quality and patterns of sleep, and the PSQI is a well-established and widely used instrument for this purpose. In contrast to previous studies applying the PSQI in patients with AD [15, 23], no significant differences regarding subjective sleep parameters between individuals with and without AD were found in the present study population. Notably, patients with AD are supposed to have higher disease severity than individuals with AD from the general population, which is associated to poor sleep quality [13, 27]. Therefore, the comparability of data from patient populations with data from the general population in this regard is limited.

When interpreting the existing evidence, several significant conceptual and methodological concerns should be discussed. Limiting the validity of the results and the comparability of the findings, most studies were designed as cross-sectional investigations with small sample sizes, and show strong heterogeneity regarding the included study populations, the methods used for evaluation of AD status and disease severity, and the availability of data on AD symptoms with impact on sleep quality. Notably, the majority of studies have analyzed sleep as secondary outcome, and self-reports, which might be subject to recall bias, were commonly applied. However, although PSG provides objective data, it is time- and cost-consuming, not widely available, and can be uncomfortable and burdensome for the patient [5], which limits its application in large trials. Besides these methodological issues, the pathophysiology of sleep disturbances in AD is extremely complex and difficult to map and measure in studies. In this regard, several contributing factors have been discussed including disease flares with itch and scratching, sleep habits such as co-sleeping and behavioral insomnia, environmental factors, melatonin dysregulation, nocturnal pruritus due to the circadian rhythm of the skin, and cytokine dysregulation [3].

Taken together, the results of the present study do not provide evidence for sleep disturbances as measured with PSG in adults with AD from the general population. As an indicator of sleep difficulties, we found a higher number of movement and wakefulness episodes and a lower proportion of REM sleep among individuals with AD compared to those without. Along with the existing literature, our study provides important directions for future research. To further elucidate the relationship between disease severity, AD symptoms, and sleep characteristics, larger-scaled longitudinal studies allowing for the assessment of the temporal course of disease severity and of symptoms with impact on sleep such as nocturnal pruritus and scratching are indispensable. Concerning the assessment of sleep characteristics, the application of actigraphy in combination with self-reported data on both sleep quality and symptom intensity might be indicated in view of feasibility and costs.

The present study is based on data from a population-based study applying high quality standards of the performed examinations including standard operating procedures, certification procedures, and quality reports, which enhances the reliability of the results. AD status of study participants was evaluated by physicians in a standardized skin examination, which is considered the gold standard in epidemiological surveys [28]. PSG provides objective data on sleep characteristics, and the PSQI is a well-established and widely used questionnaire to measure the subjective quality and patterns of sleep in adults [29]. Also, important limitations should be noted. First, our study is based on data from a single PSG recording, and the so-called “first night effect” (FNE) might have impacted our results. It has been argued that first night data may reflect a period of adaptation that is unrepresentative of usual sleep patterns [30], and previous studies showed that the FNE is characterized by decreased sleep duration and decreased sleep quality [31, 32]. Therefore, we cannot rule out that the FNE might have biased our results. Second, data on AD severity and symptoms such as nocturnal pruritus and scratching, which are considered critical determinants of sleep quality in individuals with AD, were not available. Third, since AD is characterized by often seasonal fluctuations and temporal remission for a certain time in a year, one-time skin examination as applied in the present study might exclude mild or transient AD [28], leading to potential underestimation of AD cases. Fourth, the overall sample size was small. Only 26.9% of the SHIP-TREND-0 participants took part in both the dermatological examination and the PSG, which must be evaluated in the context of the broad examination program in SHIP lasting several hours. Also, the potential time delay between the dermatological examination and the PSG measure might have impacted our findings since participation in both investigations was not in close temporal connection in every case.