Bidirectional Association Between Psoriasis and Obstructive Sleep Apnea: A Systematic Review and Meta-Analysis

The link between psoriasis and obstructive sleep apnea (OSA) has not been confirmed. We aimed to investigate the relationship between psoriasis and OSA. We conducted a systematic review and meta-analysis of case-control, cross-sectional, and cohort studies on the association between psoriasis and OSA. We searched MEDLINE and Embase for relevant studies on May 11, 2019. The Newcastle-Ottawa Scale was used to evaluate the risk of bias of included studies. We performed random-effects model meta-analysis to calculate pooled odds ratio (ORs) with 95% confidence intervals (CIs) for case-control and cross-sectional studies as well as pooled incidence rate ratio (IRR) with 95% CIs for cohort studies in association between psoriasis and OSA. A total of 4 case-control or cross-sectional studies and 3 cohort studies with a total of 5,840,495 subjects were included. We identified a significantly increased odds for OSA in psoriasis patients (pooled OR 2.60; 95% CI 1.07–6.32), and significantly increased risk for psoriasis in OSA patients (pooled IRR 2.52; 95% CI 1.89–3.36). In conclusion, our study identified a bidirectional association between psoriasis and OSA. Sleep quality should be inquired in patients with psoriasis. Respirologist consultation or polysomnography may be indicated for those presenting with night snoring, recurrent awaking, and excessive daytime sleepiness.

gating the association between psoriasis and OSA, including case-control, cross-sectional, and cohort studies; (2) research on human subjects; (3) the case group consisted of patients with psoriasis and the control group consisted of people without psoriasis in studies that examined the odds or risk of OSA in psoriasis patients; the study group was composed of patients with OSA and the control group was composed of people without OSA in studies that evaluated the odds or risk of psoriasis in OSA patients. We screened titles and abstracts for the initial study selection and obtained the full text of potentially eligible studies to confirm if they met our inclusion criteria. Two authors (T.G. and Y.F.) independently selected studies and disagreement was resolved by consulting the other author (C.C.). Data extraction. We extracted the following data from the included studies: first author, year of publication, country of study, number and gender of study subjects, definition of the case group, definition of outcome, and risk estimates including hazard ratio (HR), incidence rate ratio (IRR), and odds ratio (OR) with 95% confidence interval (CI) on the association between psoriasis and OSA.
Risk of bias evaluation. The risk of bias of included studies was assessed by using the Newcastle-Ottawa Scale 29 . For case-control and cross-sectional studies, we examined the following eight domains: adequacy of case definition, representativeness of cases, selection of controls, definitions of controls, comparability of cases and controls, ascertainment of exposure, same method of ascertainment for cases and controls, and non-response rate. For cohort studies, we evaluated the following eight domains: representativeness of exposed cohort, selection of non-exposed cohort, ascertainment of exposure, absence of outcome of interest at start of study, comparability of cohorts, assessment of outcome, follow-up duration, and adequacy of follow up of cohorts.

Statistical analysis. The Review Manage version 5.3 (Copenhagen: The Nordic Cochrane Centre, The
Cochrane Collaboration, 2014) was used for conducting all analyses. We calculated the OR with 95% CI in included case-control and cross-sectional studies and IRR in included cohort studies. If OR was not reported in an included study, we calculated the crude OR based on published data for example the number of events in the case and control groups. The most fully adjusted OR, HR, or IRR were adopted if reported. We performed meta-analyses to evaluate the bidirectional association between psoriasis and OSA. For included case-control and cross-sectional studies, we calculated pooled OR to examine the association between the two diseases. For cohort studies, we treated HR as IRR and calculated the pooled IRR 30 . The statistical heterogeneity was assessed by the I 2 statistic across the included studies. An I 2 of >50% represents substantial heterogeneity 31 . We chose random-effects model for meta-analyses because clinical heterogeneity was anticipated. Therefore the DerSimonian and Laird method that takes between-study variability into account was used to obtain the pooled OR and IRR estimates 32 . Also, we conducted a sensitivity analysis after excluding studies that were rated with a high risk of bias.

Results
Characteristic of included studies. The PRISMA flow chart of study selection is shown in Fig. 1. Our systematic literature search yielded 98 records after removing duplicates. We identified one additional relevant study in a review article. After screening the titles and abstracts, 54 records were excluded. After assessing the full text for eligibility, 38 articles were excluded due to no relevant data, review articles, no comparison group or irrelevant comparison group. Ultimately seven studies with a total of 5,840,495 study subjects were included. One cohort study, two case control studies, and one cross-sectional study investigated the association of psoriasis with OSA. Three cohort studies and one case-control study investigated the association of OSA with psoriasis. There was one cohort study that reported the bidirectional association between psoriasis and OSA. The main characteristics of included studies are summarized respectively in Table 1.

Risk of bias of included studies.
The risk of bias of included studies was summarized in Figs. 2 and 3. Of four included case-control and cross-sectional studies, no items were rated as high risk of bias (see Fig. 2). The Tsai 2011 study was rated with an unclear risk of bias in the 'adequacy of case definition' domain because only the International Classification of Disease (ICD) diagnostic codes were used to identify patients 33 . Two studies were rated with an unclear risk of bias in the 'representativeness of cases' and 'selection of controls' domains because patients were selected from hospital not community 34,35 . We rated two studies as unclear risk of bias in the 'comparability of cases and controls' domain because only age and gender, but not body mass index, were controlled 33,35 . Moreover, three studies were rated as unclear risk in the 'ascertainment of exposure' domain because only medical records or ICD codes were employed 24,33,34 . In three included cohort studies, we rated the Cohen 2015 study at high risk of bias in the 'representativeness of exposed cohort' domain (see Fig. 3) 23 . The reason was that the study subjects were all female nurses.
Conversely, three cohort studies 23,25,27 and one case-control study 34 found a consistent increase in psoriasis among patients with OSA. The meta-analysis of three included cohort studies with 5,544,674 study subjects showed a significant association of OSA with psoriasis (pooled IRR, 2.52; 95% CI, 1.89-3.36; Fig. 5) 23,25,27 . There was no statistical heterogeneity within these studies (I 2 = 0%). The Cohen 2015 study was rated with a high risk of bias in the representativeness of the exposed cohort because all the study subjects were nurses 23

Discussion
To the best of our knowledge, the present study is the first meta-analysis investigating the bidirectional association between psoriasis and OSA. The evidence from included case-control and cross-sectional studies found a 2.6-fold greater odds for prevalent OSA in relation to psoriasis 24,33,35 and a 13-fold increased odds for prevalent psoriasis in relation to OSA 34 . Similarly, the evidence from an included cohort study demonstrated consistently increased risk for incident OSA among patients with mild and severe psoriasis as well as psoriatic arthritis 25 . Conversely, OSA patients were 2.52-fold more likely to develop incident psoriasis when compared with non-OSA controls 23,25,27 . Our meta-analysis possesses high generalizability because of including studies from various ethnicities for example the United States, France, Denmark, Greece, Israel, and Taiwan.
Obesity is a trigger for inflammation and has been linked to chronic inflammatory diseases 36 . Patients with psoriasis have been found to have increased body mass index (BMI) than healthy controls 37 . Besides, OSA was related to obesity and thus weight loss is the recommended fist-line therapy 38,39 . Obesity may promote upper www.nature.com/scientificreports www.nature.com/scientificreports/ airway inflammation, reduce muscle contractibility, and induce airway collapse during sleep 40 . BMI has been correlated with psoriasis severity 37 and OSA 26 . Among most of included studies in our meta-analysis 23   www.nature.com/scientificreports www.nature.com/scientificreports/ the confounding from obesity has been considered and BMI was adjusted in the statistical analysis. The bidirectional association between psoriasis and OSA remained significant after adjustment for obesity or BMI. Only one included case-control study of Tsai 2011 did not contain any information about BMI or body weight 33 . Another included cross-sectional study of Sacmaci 2019 did not perform adjustment for obesity, but there was no significant difference in BMI between psoriasis patients and non-psoriasis controls (28.5 kg/m 2 vs. 26.9 kg/m 2 ; P = 0.051) 35 . Therefore, confounding by obesity could not fully explain the association between psoriasis and OSA.
The bidirectional association between psoriasis and OSA suggest the shared common systemic inflammatory pathogenic pathways 19 . OSA is considered a systemic inflammatory disorder. High levels of systemic inflammatory molecules are part of mechanism leading to OSA 41,42 . The activities of IL-17, TNF, IL-6, IL-7 and C-reactive protein are significantly increased in patients with OSA compare to obese patients [43][44][45] . These pro-inflammatory cytokines decreased after continuous positive airway pressure (CPAP) treatment 44 . Studies have shown that the circadian rhythm of TNF release was significantly disturbed in patients with OSA 46,47 . The inflammatory process may predispose them to the development of psoriasis 27 . Elevated circulating levels of IL-17, TNF, IL-6, and IL-22 were also associated with psoriasis 48 . One previous study found that treatment with etanercept significantly improved symptoms of OSA in moderate to severe psoriasis 49 . The increase of IL-17 is related to atherosclerotic vascular disease, which is a risk factor for OSA 50 . Moreover, psoriasis causes pruritus and may lead to sleep disturbance, which contributes to systemic inflammation and consequently induces OSA 51-53 . Autonomic activation is considered an important factor between OSA and psoriasis [53][54][55] . Some scholars assumed that psoriasis-related itching and pain could disrupt sleep and thus increase autonomic activation, which possibly induces OSA 18,54 . On the other hand, Gabryelska et al. proposed low grade inflammation in OSA patients with elevated levels of IL-1 and TNF could stimulate hypothalamus and increase hypothalamic-pituitary-adrenal activity with resultant surging of autonomic activity 56 . The inflammation and autonomic activation involved in OSA may be a risk factor for psoriasis 55 .
Intermittent hypoxia during sleep in OSA patients induces oxidative stress. Some markers including nuclear factor-κB (NF-κB) and hypoxia-inducible factor-1α (HIF-1α) could detect the oxygen stress and activate the inflammatory pathway 57 . HIF-1α, a transcriptional regulator of cell oxygen metabolism, is a specific marker for diagnosing OSA 58 . The levels of HIF-1α correlate with the severity of OSA and can be effectively reduced by CPAP treatment 57 . Besides, vascular risk response to hypoxia was detected in OSA patients. A significant upregulation of NF-κB, HIF-1α, endothelial nitric oxide synthase, vascular cell adhesion molecule 1, and vascular endothelial growth factor (VEGF) were expressed in the skin biopsy specimens of OSA patients 59 . Increased circulation of VEGF promotes angiogenesis which is also an important process in initiating psoriasis 59,60 . VEGF is a marker of tissue response to hypoxia and activated by HIF-1α 59 . There is evidence indicating activation of excess HIF-1α and VEGF in psoriatic skin when compare to normal skin 61,62 . In addition, HIF-1α is involved in T-cell regulation and survival which are crucial in psoriasis formation 62 . The effects of hypoxia with elevated levels of HIF-1α and VEGF may explain the association between OSA and psoriasis.
There are a few limitations of the present study. First, OSA was identified by the use of ICD codes in four included studies 24,25,27,33 and misclassification bias might have been present. Second, high statistical heterogeneity was detected as to the association of psoriasis with OSA (I 2 = 89%; see Fig. 4), but all the include studies consistently reported positive association. . Forest plot for case-control and cross-sectional studies on the association of psoriasis with obstructive sleep apnea. The meta-analysis illustrated a significant association of obstructive sleep apnea with psoriasis (pooled odds ratio 2.60; 95% confidence interval 1.07-6.32).

Figure 5.
Forest plot for cohort studies on the association of obstructive sleep apnea with psoriasis. The metaanalysis illustrated a significant association of obstructive sleep apnea with psoriasis (pooled incidence rate ratio 2.52; 95% confidence interval 1.89-3.36). (2020) 10:5931 | https://doi.org/10.1038/s41598-020-62834-x www.nature.com/scientificreports www.nature.com/scientificreports/ In conclusion, the evidence to date supports a bidirectional association between psoriasis and OSA. All patients with psoriasis should be informed about the risk of OSA, and vice versa. OSA is a comorbidity of psoriasis that should not be ignored, and sleep quality should be inquired in patients with psoriasis. Respirologist consultation or polysomnography may be indicated for those presenting with night snoring, daytime sleepiness, and insomnia.