Measurement, Classification and Evaluation of Sleep Disturbance in Psoriasis: A Systematic Review

Background Psoriasis is a long-term immune-mediated inflammatory disorder mainly, but not only, affecting skin, and is associated with significant medical and psychological morbidity. Evidence suggests that sleep is disrupted in psoriasis, however high quality empirical evidence is lacking. Given the importance of sleep for health, characterisation of sleep disruption in psoriasis is an important goal. We therefore conducted a systematic review of the sleep-psoriasis literature. Methods Searches were conducted in Pubmed, SCOPUS and Web of Science from inception to May 2016. Studies were compared against inclusion/exclusion criteria and underwent a quality evaluation. Given the heterogeneity of studies, we conducted a narrative synthesis of the findings. Results Searches revealed 32 studies which met our predetermined inclusion/exclusion criteria. Whilst 93.7% of studies reported sleep disruption in this population, ranging from 0.05% to 85.4%, many had important methodological shortcomings. Over half of all quantitative studies (54.8%; 17/31) relied on non-validated measures, contributing to heterogeneity in study findings. In those that employed valid measures, assessing sleep was often not the primary objective. We frequently found the absence of adequate sample size calculations and poor statistical reporting. Conclusion This review showed that in psoriasis, reported sleep rates of sleep disturbance varied substantially. Most studies lacked a hypothesis driven research question and/or failed to use validated measures of sleep. We were unable to draw firm conclusions about the precise prevalence and nature of sleep disturbance within the psoriasis population. We offer suggestions to help advance understanding of sleep disturbance in psoriasis.


Introduction
Psoriasis is a long-term, immune mediated inflammatory skin disorder, characterised by scaly plaques on knees, elbows and scalp, but any skin surface may be affected [1]. Population-based cohort studies indicate that psoriasis is associated with other systemic inflammatory conditions, including: psoriatic arthritis (PsA) [2]; inflammatory bowel disease [3]; cardiovascular disease (CVD) [4]; and diabetes [5].
Contemporary evidence suggests that sleep is important for daytime functioning and health, [6][7][8] subserving optimal physiological [9] and psychological functioning [10][11][12][13]. Moreover, disturbed sleep may causally drive disease processes. For example, persistent sleep disturbance is a risk factor for the future development of diabetes [14], CVD [15,16], hypertension [17,18], and depression [19][20][21][22]. Several studies [23][24][25][26][27][28][29][30][31][32][33][34][35][36][37][38], including four reviews, have examined sleep domains in psoriasis populations. One review evaluated factors linked to sleep disturbance in psoriasis, concluding that mood, obstructive sleep apnoea (OSA), itch and pain as possible sources of sleep disturbance [29]. Another review provided an overview of the sleep and dermatology literature with a focus on the role of itch, suggesting skin temperature, circadian rhythm and psychological factors, such as depression, influenced itch, which in turn disrupted sleep [30]. Similarly, a further review assessed the associations between sleep disorders and the broad category of skin disorders [31]. More recently, Gupta et al. conducted a systematic review of psoriasis and sleep disorders and found an increased prevalence of OSA and restless leg but inconclusive evidence of increased insomnia risk [32]. Whilst these reviews provide some insight into the nature and predictors of sleep in psoriasis the majority are narrative and none critically examine the measures used to assess sleep nor do they assess the quality of current research. Additionally the most recent review [32] only included papers where a sleep disorder diagnosis was made, thus excluding studies where sleep was assessed without a diagnosis, potentially missing research where sleep was assessed but no diagnosis was made. Accordingly, a more rigorous and nuanced appraisal of the sleep-psoriasis literature is required. We present a systematic review focusing on (1) the evidence for sleep disturbance in psoriasis and (2) the quality of measures and methodology employed to assess sleep in patients with psoriasis. We conclude by outlining a research agenda aimed at advancing understanding of sleep disturbance in the context of psoriasis.
MeSH terms (Table 1). Studies were assessed against inclusion/exclusion criteria determined by an expert group of reviewers (Table 2). Reference lists of eligible studies were forward and backward searched for completeness. Reviews were included to identify additional studies that were not identified via database searches.

Data extraction
Data extracted from studies included: authors, title, year, study aim(s), journal, study design, type of psoriasis, sample size, measure of sleep used and findings relating to sleep (Tables 3-5).

Quality Evaluation
We adapted a standard quality scoring tool [34] for quantitative and qualitative study designs to include information about the quality of the sleep measure used, details of potential confounding variables and the sleep findings reported by each paper. Members of the review team (ALH, SB, CB, AC) tested the quality tool on a sample of four papers. The first author then reviewed all papers, discussing any problems with the fourth and sixth authors with amendments discussed and agreed upon during subsequent meetings until there was full agreement on the domains included in the tool. Subsequently ten randomly selected papers were coded by the third author (SB) independently with these scores compared to those of the first author (ALH). Disagreement in scores was found for one domain which was resolved following discussion with the rest of the team resulting in 100% agreement.
Fifteen domains were used for quantitative papers and seven for qualitative papers (supplementary data). Each quality evaluation domain had specific criteria that had to be fulfilled in order to obtain a score from 0 to 2. A standardised rating scale was used to score each item thus: a score of 2 indicates that criteria had been fulfilled, a score of 1 indicates partial fulfillment, and a score of 0 indicates no fulfillment of the criteria. Quantitative studies, depending on their design could score a maximum score of 40, whereas qualitative studies could score a maximum of 22 points.     Each paper's total quality score was then standardised as a percentage, obtained by dividing the study's score by the total points available. Quality appraisal thresholds determined by the team were then applied to each study based upon its percentage score (good: !80%; fair: !61-79%; and poor: 60%).

Results
We identified 337 original papers. Abstracts were screened for eligibility and 42 full papers were obtained and scrutinised (Fig 1). In total, 32 papers were included in the final review. Thirty-one studies employed quantitative designs, and one study used a qualitative framework. We structure our review findings under the following category headings: study quality; sleep methodology and prevalence of sleep disruption; and factors associated with sleep disturbance.

Study Quality
Fewer than half of all studies (15/32; 46.9%) were rated as good quality (i.e. scoring ! 80%), over fair (15/32;46.9%) (scoring ! 61-79%), and the remainder (2/32; 6.25%) as poor (scoring 60%) ( Tables 3-5). Overall quality scores were fair (mean % = 76.9), with a number of methodological limitations identified across studies: 27 studies failed to perform sample size calculations, 17 had inappropriately small or large sample sizes, 7 studies did not provide demographic data, 2 studies failed to report any data on central tendency or variability, and 5 studies reported this data for only one time point and 14 studies failed to control for confounding variables (S1 Table). Of the 14 studies using valid measures of sleep, the majority (9) scored in fair (scoring ! 61-79%), and poor (2) categories (scoring 60%) and only 3 were rated as good (scoring !80%). Five studies used polysomnography (PSG), and focused on OSA [26,27,35,36] or itch [37], with one of these also using the Functional Outcomes of Sleep Questionnaire [26]. Four used the Pittsburgh Sleep Quality Index (PSQI) [23,24,38,39], one of which was conducted on a PsA sample [39] and one study used a condensed version of the PSQI [40]. One study used the Insomnia Severity Index (ISI) [23] while a further four studies used other validated self-report measures specifically, the General Sleep Disturbance Scale [41] and the Medical Outcomes Study Sleep Scale (MOS-SS) [28,42,43]. One study was qualitative, evaluating the impact of pruritus using interviews with physicians alongside patient focus groups [44]. The remaining 17 (54.8%) quantitative studies used non-validated sleep measures. Six studies used broad measures of health and functioning, which incorporated sleep items, such as measures health-related quality of life (HRQoL) (General Health Questionnaire, modified Dermatology Life Quality Index, Children's Dermatology Life Quality Index, General Health Questionnaire Hindi) or other general health-related questionnaires (Brief Pain Inventory, McGill Pain Questionnaire) [45][46][47][48][49][50]. Four studies used non-validated questionnaires [25,[51][52][53] three of which contained a single item pertaining to sleep disturbance [25,52,53] (e.g "in a typical month, how many days did your disease interfere with sleeping?") [25]. Four studies solely used medical records [54,55,56,57], and another used a clinical diagnostic interview with a sleep disorder diagnosis [58]. In two validation studies HRQoL was assessed using a willingness-to-pay paradigm, where participants were asked to indicate how much money they would allocate to relieve themselves of a particular symptom [59,60]. Interestingly, in one of the studies using a PsA sample, sleep was allocated the highest median amount of money ($10,000) [60], whereas in the other study, with a psoriasis sample, sleep was the lowest ranked issue ($625) [59].
Of all studies reviewed, 11 (34.4%) had an a priori aim to assess sleep, however, only three (9.4%) had the primary objective of assessing sleep quality and quantifying sleep disturbance in psoriasis [23,24,37]. The other studies were not primarily focused on general sleep in psoriasis. Two studies investigated associations between PsA and sleep [25,39] while four studies assessed the link between OSA and psoriasis [26,27,35,36]. The remaining two studies measured the influence of itch on psoriasis [37,51], one of which used a mixed dermatological sample [37] with just 1 of 15 participants (6.67%) having psoriasis. The former [51] used a questionnaire and the latter [37] used PSG to assess sleep. One study aimed to assess the links between sleep disorders in psoriasis and cardiovascular risk [56].
Thirty (93.7%) of 32 studies, including one qualitative study, observed sleep disturbance. However, in studies where sleep was measured prospectively, values varied from 0.05% to 87.5% (Table 6). Two papers reported no sleep problems relative to healthy controls [24,59]. In one of these studies, whilst sleep disturbance was reported by the majority (55%) of participants, relative to other domains of life, sleep disturbance was the least bothersome [59].
In contrast, another study reported sleep as being the second-most disrupted domain as a consequence of psoriasis, across young, middle aged and older-aged respondents; indicated by 20%, 22% and 22% respectively [53]. Sleep quality was worse in psoriasis patients (8.8±4.4 vs. 6.3±4.4) [23] and PsA patients (9.70±3.9 vs. 4.05±1.85) [39] relative to healthy controls in two out of three studies using PSQI with the third paper showing no difference between psoriasis patients and controls (5.56 vs. 5.13) [46]. The fourth paper using the PSQI showed that on two components of the measure scores were significantly worse in psoriasis subjects than healthy controls (subjective sleep quality; 1.48±0.17 vs. 1.02±0.13 and habitual sleep efficiency; 0.79 ±0.19 vs. 0.32±0.14) [38]. Nevertheless, they did not report data for one component nor did they provide global PSQI scores for both groups [38]. Furthermore, one of these studies showed significantly increased ISI scores in psoriasis patients (11.0±7.0) relative to healthy controls (6.3 ±6.0) [23], with psoriasis patients' mean score at the threshold for clinical insomnia [61].
Four studies found high rates of OSA in psoriasis patients ranging from 36% to as high as 56.3% [27,35,36]. Moreover in a large epidemiological study, psoriasis patients had a significantly increased prevalence of sleep disorder diagnosis than non-psoriasis counterparts [57] and in another, psoriasis patients with a sleep disorder had elevated cardiovascular risk relative to those without a sleep disorder [56].

Factors associated with sleep disturbance
Sixteen of the 32 reviewed studies reported factors associated with sleep disturbance. Itch, depression, PsA and pain were the most commonly reported factors associated with sleep disturbance however others were reported across studies. Itch was reported to occur more frequently at night [50] and in six studies it was associated with disrupted sleep [25,37,42,44,50,52]. Itch predicted lower scores on the MOS-SS, indicating worse sleep [42]; increased scratching was associated with more awakenings during stages 1 and 2 of sleep [37]; and itch was reported by subjects as causing difficulty initiating and maintaining sleep as well as performing daily activities [44]. Additionally, in one study 65% of participants reported that sleep ameliorated itch [52]. The association between itch and sleep was documented further in one study finding the association between depressive symptoms and itch severity was partially mediated by sleep quality [40]. In another study daytime sleepiness was associated with depression [28]. Contrasting findings were indicated by one study in which itch levels were reported as equivalent across those with and without sleep disruption, whereas depression was significantly higher in those who experienced sleep disruption relative to those who did not [51]. This suggests that depression may have played a greater role in sleep disruption than itch. This is supported by another study showing that when controlling for depression variables, psoriasis patients are no more likely to experience poor sleep or have greater insomnia symptoms than controls [23], and another study showing that while PSQI scores were higher in psoriasis subjects, there were no differences in psychiatric symptoms relative to controls [38].
Pain appears to contribute to poorer sleep, and in one study over 85% of those reporting increased pain specified sleep as the most disrupted function [47]. Another study showed that sleep disturbance mediated the relationship between pain and reduced HRQoL [41]. Four studies showed associations between PsA and sleep [25,28,39,60]. PsA and generalized pain was associated with reduced PSQI scores relative to healthy controls [39], increased want for improved sleep [59] and predicted sleep disturbance [25]. Additionally sleep disturbance was associated with anxiety, enthesitis, levels of C-reactive protein and erythrocyte sedimentation rate [39].
Sleep disturbance was consistently associated with reduced QOL [28,47,54,58]. Moreover, in four treatment studies using biologics, improvements were observed in sleep quality, disturbance, daytime sleepiness and adequacy [26,41,43,46] suggesting that improvements in psoriasis disease processes may lead to concomitant improvements in sleep. Furthermore in a longitudinal cohort study sleep showed the greatest improvement from baseline to follow-up across varied psoriasis therapies (topicals and systemic) relative to other items on the Children's Dermatology Life Quality Index [48].

Key findings from the review
This systematic review examined the evidence for sleep disturbance in psoriasis and the measures and methodology used to assess sleep. The majority of studies showed sleep problems were associated with itch, low mood and pain. Itch appears to be associated with increased sleep fragmentation, disturbance and reduced quality of sleep. These findings are consistent with research in other pruritic conditions [62][63][64]. Low mood and pain were also implicated in poor sleep as previously shown in both healthy individuals and other illness populations [65,66]. It is likely that mood, pain and itch interact disturb sleep; although the nature of this interaction requires examination within the context of purposively-designed studies.
Prevalence rates varied substantially and while studies were of fair-to-good quality, many demonstrated a number of methodological and statistical shortcomings. In particular, the lack of validated measures of sleep, coupled with limited theoretical work-up may have accounted for the variation in rates of sleep disturbance across studies. Nevertheless, prevalence clustered around the 50-60% mark; higher than current estimations of sleep pathology in the general population [67]. Additionally, rates of OSA appear to be higher in psoriasis (36-56.3%) relative to the general population (3-7%) [68]. Although the reasons for this are unclear; OSA is multifaceted, involving both inflammatory pathways and lifestyle factors including high BMI and physical inactivity [69]. OSA can have a significant impact on health and contribute to cardiovascular disease risk [70], which is already elevated in psoriasis [4]. Further research should identify mechanisms underpinning the link between OSA and psoriasis.
Lack of detailed assessments and incomplete data reporting have impeded further conclusions and provide a somewhat superficial picture of the true problem. For instance, studies using the PSQI often failed to report full data for the measure (sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbance, use of sleeping medication and daytime dysfunction). Providing these data would facilitate a more thorough and itemised understanding of sleep quality.
Moreover, despite the 'gold standard' PSG being used in five studies, it provides limited information about the range of factors that may impact upon sleep [71,72]. Comprehensive assessment of sleep involves the integration of objective and subjective measures, profiling of sleep-wake patterning over time, and the assessment of factors that may contribute to night-tonight variation in sleep quantity and quality [71,72]. We found no studies in this review that conformed to these standards and neither actigraphy nor sleep diaries have been applied in psoriasis populations. These data would support a more nuanced understanding of sleep disruption in psoriasis and help to design appropriate interventions [71,72].

Limitations
This review only included studies published in English and those that have undergone peer review. Therefore, it is possible that some relevant findings may have been missed. Further, due to the subjective nature of the quality scoring, there is greater opportunity for personal bias to influence scores. However we attempted to minimize this risk by the authors scoring papers independently and there was close to 100% agreement in paper scores following independent scoring.

Concluding remarks: Recommendations for future research
There is a need to systematically and consistently examine sleep in psoriasis populations, employing comprehensive and validated measures of sleep in specifically designed studies. Accurate prevalence rates of sleep disturbance must first be established; from here prospective studies can explore the relationship between sleep and psoriasis, focusing on potential precipitating and perpetuating factors that may drive and maintain poor sleep, (candidate targets: itch, mood, pain and OSA). Work should also explore sleep-wake variability in this population and assess any concomitant consequences on daytime functioning.
Obtaining a comprehensive understanding of sleep in psoriasis is of importance due to the implications of sleep disruption for health. Sleep disturbance may result in disruptions to immune and sympathetic nervous system functioning possibly leading to the maintenance and/or exacerbation of psoriasis, conferring risk for adverse psychological and medical outcomes. This has been suggested in a recent study which found that the presence of a sleep disorder alongside psoriasis confers increased risk for CVD morbidity [56].
The potential role of inflammation in the relationship between poor sleep and psoriasis has also been suggested. Four studies included in this review [26,41,43,46] showed improvements but not a mitigation of sleep problems following the administration of biologic medication. This suggests that reductions in systemic inflammation may confer sleep improvement possibly through a reduction in itch and associated skin discomfort. Further research is required to examine mediating pathways and treatment mechanisms in greater detail.
Supporting Information S1