Prevalence of sleep disturbance in patients with ankylosing spondylitis: a systematic review and meta-analysis

Ankylosing Spondylitis (AS) patients face several challenges due to the nature of the disease and its physical and psychological complications. Sleep disorders are among the most important concerns. Sleep disorders can aggravate the signs and symptoms of the disease and ultimately reduce the quality of patients’ lives. This study uses a systematic review and meta-analysis to pool the reported prevalence of sleep disorders among AS patients. To find related studies, the WoS, PubMed, ScienceDirect, Scopus, Embase, and Google Scholar databases were systematically searched without a lower time limit. Heterogeneity among the identified studies was checked using the I2 index, and the Begg and Mazumdar correlation test examined the existence of published bias. Comprehensive Meta-Analysis (v.2) software was adopted to analyze the data. In the review of 18 studies with a sample size of 5,840, the overall pooled prevalence of sleep disorders among AS patients based on the random effects method was found to be 53% (95% CI: 44.9–61). The highest and lowest prevalence was in Egypt at 90% and Australia at 19.2%, respectively. Our meta-regression results show that with the increase in ‘sample size’ and ‘year of publication’, the overall prevalence of sleep disorders in patients with AS decreases (p < 0.05). The results of the present study indicate a high and significant prevalence of sleep disorders among AS patients. Thus, health policymakers and healthcare providers must focus on timely diagnosis and effective educational and therapeutic interventions for the prevention and proper treatment of sleep disorders in this population of patients.


Background
Ankylosing spondylitis (AS) is a chronic progressive spinal inflammatory arthritis and belongs to a group of arthritis called spondyloarthritis. Clinical manifestations usually emerge in the third decade of life. AS characteristically affects the axial skeleton, sacroiliac joints, entheses, and extra-skeletal sites such as the bowel, skin, and eye, and rarely the lungs and heart can be affected [1][2][3][4].
Inflammation in AS can cause bone erosion, new bone formation, and excessive bone remodeling and slowly fills the interarticular spaces and eventually fuses the joints in the spine, which triggers severe pain and reduces spinal mobility and stiffness [1][2][3][4][5][6]. This bone destruction and misplaced growth lead to major limitations in spine mobility, physical activity, and impaired functional ability. [5,6]. In addition, these patients are at a greater risk of suffering from depression and anxiety [7,8]. Such burdens negatively affect the quality of life of the patients [9].
Sleep quality is one of the important factors affecting quality of life [10]. Sleep disorders are one of the public health problems, so that most people have experienced sleep disorders in their life [11][12][13]. The findings from previous research indicate that patients with AS suffer from more sleep disorders than healthy individuals [14][15][16]. Patients with AS are affected by various sleep disorders, including sleep onset and continuation disorders, restless leg syndrome, and obstructive sleep apnea (OSA) [10,14,[16][17][18]. Nevertheless, previous research findings present different sleep disorder prevalences among AS patients. For instance, in the studies conducted in Egypt, China, and Australia, the prevalence of sleep disorders was reported as 90%, 31%, and 19.2%, respectively [10,16,19].
Sleep disorders in AS patients can be influenced by the experience of axial pain and inflammatory back pain after midnight [20,21], severeness of the disease [10], and fatigue and functional limitations experienced by the patients [6,20]. In addition, under the influence of symptoms and complications of the disease, these patients experience anxiety and depression, which can affect sleep quality [16,20,[22][23][24].
Sufficient and high-quality sleep regulates several internal bodily processes, such as metabolism, tissue restoration and recovery, immune-inflammatory response, and synaptic homeostasis, and improves cognitive function, memory, and mood [25,26]. Accordingly, sleep disorders are related to increasing the risk of impaired cognitive functions, signs of depression, sleepiness during the day, and thus increasing the probability of accidents, falls and related deaths, fatigue, and reducing the physical capacity of a person in performing daily activities [12,26]. Similarly, and even more sternly, sleep disorders affect the quality of life of AS patients [6]. The findings of existing research highlight the role of sleep disorder in the occurrence of depression and anxiety [16,20,22] and fatigue [20,27] and that there is a significant positive relationship between sleep disorder and AS [6,28].
Considering the negative effects of sleep disorders on an AS patient's life, there is a need for policymakers and health professionals to be trained on the impacts of the disease. It is also imperative to allocate the necessary resources for pertinent therapeutic interventions. Considering dissimilar reported results of existing studies on the relationship between AS and sleep disorders, this study aims to obtain the pooled prevalence of sleep disorders among patients with AS.

Methods
This study is a systematic review and meta-analysis, which was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [29].
Additionally, lists of references within the identified articles were manually searched to ensure the comprehensiveness of the search and to find further relevant studies. No restrictions were placed on the year of publication of the articles. Study information was transferred into the EndNote (x8) reference management software.

Inclusion and exclusion criteria
Studies that fulfilled the inclusion criteria were retained in the study selection process. The inclusion criteria entailed cross-sectional, cohort, and case-control studies that reported the prevalence of sleep disorders among AS patients and had sufficiently reported sufficient related information. Exclusion criteria were case reports, case series, intervention studies, review studies, studies not published in English, and those whose full text was unavailable.
The International Classification of Sleep Disorders-Third Edition (ICSD3) is a reliable and authoritative source for diagnosing sleep disorders. According to this text, sleep disorders are divided into seven main categories that are insomnia disorders, sleep-related breathing disorders, central disorders of hypersomnolence (hypersomnia), circadian rhythm sleep-wake disorders, sleep-related movement disorders, parasomnias, and other types of sleep disorders [30][31][32][33]. This study did not focus on the prevalence of insomnia disorders, sleeprelated breathing disorders such as obstructive sleep apnea, hypersomnia, circadian rhythm sleep-wake disorders, sleep-related movement disorders such as restless leg syndrome and parasomnia. Therefore, our analysis did not include several studies that only reported the prevalence of obstructive sleep apnea and restless leg syndrome.

Study selection
Following the initial search, duplicated studies in different repositories were removed, and only a copy was retained. Subsequently, the titles and abstracts of articles were reviewed per the inclusion and exclusion criteria, and irrelevant articles were removed. In the next step, the full texts of the remaining studies were evaluated based on the inclusion and exclusion criteria, and at this step, further irrelevant studies were omitted. To avoid bias, two researchers conducted all the steps of reviewing sources and data extraction independently, and the reasons for omitting articles were noted. In cases where there was a difference of opinion between two researchers, the review of the article was finalized by a third person to reach a consensus.

Quality evaluation
The Newcastle-Ottawa Scale is a quality assessment tool for observational studies that the Cochrane Collaboration recommends. The NOS assigns up to a maximum of nine points for the least risk of bias in three domains: selection of study groups, comparability of groups, and ascertainment of exposure and outcomes for case-control and cohort studies. Eventually, articles were classified as high quality (scoring ≥ 5 points) or low quality (scoring < 5 points).

Data extraction
Data extraction was conducted by two researchers using a separate pre-prepared checklist with the following fields: First author's name, year of publication, place of study, type of study, AS diagnostic criteria, tools used to investigate sleep disorder, age of patients, sample size, and prevalence of sleep disorder.

Statistical analysis
The heterogeneity among the studies was assessed using the I 2 test. The Begg and Mazumdar correlation test at a significance level of 0.1 was used to examine publication bias, and the corresponding funnel plot was drawn. Data analysis was conducted within the Comprehensive Meta-Analysis (v.2) software.

Results
After systematically searching the research repositories and databases, 1,177 studies were identified and transferred into the EndNote reference management software. After removing 366 duplicate studies, the titles and abstracts of the remaining 811 articles were examined based on inclusion and exclusion criteria. Following this examination, 174 studies were retained for secondary evaluation. At this stage, the full texts of the 174 studies were examined following the inclusion and exclusion criteria, and a further 150 articles were excluded. After evaluating the quality of the remaining 24 studies, six further articles were omitted as they were deemed low quality. Finally, 18 studies were included in the meta-analysis (Fig. 1). The information from 18 studies is outlined in Table 1 [6,10,14,16,19,20,22,27,28,[34][35][36][37][38][39][40][41][42].
In the review of 18 studies with a sample size of 5,840, the I 2 test illustrated high heterogeneity (I 2 : 96.05), and thus random effects method was adopted to conduct the analysis. Our meta-analysis shows that the pooled prevalence of sleep disorders among AS patients is 53% (95% CI: 44.9-61) (Fig. 2). The included studies were published between 1999 and 2022. The highest reported prevalence, with a rate of 90%, belonged to the study of Abdulaziez and Asaad, which was conducted in Egypt [19], and the lowest prevalence, with a rate of 19.2%, was reported in the work of Tymms et al. in Australia [10].
The Begg and Mazumdar correlation test was conducted at the significance level of 0.1, which indicated the absence of publication bias among the studies (p: 0.288) (Fig. 3).
We also conducted a meta-regression analysis to examine the effect of other factors that may be affecting the heterogeneity of the studies. Our meta-regression analysis showed that with the increase in 'sample size' , the prevalence of sleep disorders in AS patients decreased (p < 0.05) (Fig. 4). Moreover, with the increase in 'year of publication' , the prevalence of sleep disorders among AS patients decreased (p < 0.05) (Fig. 5).

Discussion
Ankylosing Spondylitis is a chronic inflammatory disease that has a negative effect on the quality of life [21]. The findings of previous studies indicate the existence of disorders in the quality, duration, and efficiency of sleep, difficulties in falling asleep, the presence of restless leg syndrome, and obstructive sleep apnea among AS patients [37,38,[46][47][48][49].
This systematic review and meta-analysis were conducted to pool the prevalence of sleep disorders in patients with AS. As a result, 18 studies with a sample size of 5,840 people were evaluated, and their reported results were analyzed. Accordingly, the pooled prevalence of sleep disorders among AS patients is 53%.
Moreover, considering our meta-regression analyses, it was found that with the increase in 'sample size' and 'year of publication' , the prevalence of sleep disorders significantly decreased.
In a systematic review by Leverment [20,36,38,40]. The lowest prevalence (19.2%) among the retained studies was related to the work of Tymms et al. in Australia [10]. The differences in reported results of existing literature can be due to the variations in the tools adopted for assessing sleep disorders, alternative approaches to defining the cut-off point of the tools used, and the working methods within the studies.
As highlighted earlier, the findings show that the prevalence of sleep disorders decreases with the increase of 'sample size' . Among the included studies, the highest prevalence belonged to the two studies of Abdulaziez and Asaad and Yüksel et al., with sample sizes of 20 people, 90% and 80%, respectively [19,42].
In the study of Aydin et al., the scores of 5 out of the seven subscales of PSQI, i.e., subjective sleep quality, sleep duration, habitual sleep efficiency, sleep disturbances, and daytime dysfunctions in patients, were significantly higher than in healthy individuals. However, the scores of the other two subscales, sleep latency and the use of sleeping pills, were insignificant between the two groups. In fact, in the study, the main concern was related to the continuation of sleep and not the sleep onset; this may be because the inflammatory back pain intensifies in the second half of the sleep duration [46]. Inflammatory back pain and axial pain are among the In a report by Li et al., AS patients had higher scores than the control group in the subscales of subjective sleep quality, sleep latency, sleep efficiency, sleep disorders, and medication use. However, the difference between the two groups was insignificant regarding sleep duration and disruption in daily functioning [53]. Compared to that study, the current research has selected more databases for search and covers a longer time range (until 2022), and accordingly, the sensitivity of the present study is higher. Also, the analysis of factors causing heterogeneity based on meta-regression is considered so that an analysis can be performed on the high heterogeneity of the study. Abdulaziez and Asaas conducted a study using polysomnography (PSG) to objectively evaluate sleep quality in patients with AS. They reported that compared to the healthy individuals, patients had a lighter sleep with an increase in Non-rapid eye movement (NREM) stage I and II, which means a reduction in deep sleep. In addition, slow-wave sleep was reduced among the patients, indicating a reduction in deep sleep [38,[43][44][45].
The pro-inflammatory cytokine, tumor necrosis factor-alpha (TNF-α), is one of the important cytokines in the inflammation process in patients. Several TNF-α inhibitors have been developed to reduce spinal pain and inflammation in this disease [54][55][56]. The findings of the study of Karatas et al. demonstrate that in patients with severe AS activity who underwent anti-TNF therapy, compared to patients who were in remission and were  not treated with anti-TNF-α drugs three months after the treatment, the PSQI scores decreased significantly, which means that their sleep quality improved meaningfully compared to the healthy group. However, PSG-related parameters such as NREM stage I and II after three months of treatment did not show a difference between the two groups [37]. In the study of In et al., After evaluating the sleep quality of the patients using PSG, the total sleep time and its efficiency in the anti-TNF group was significantly higher than the group receiving NSAIDS, stage I was considerably shorter, and the rapid eye movement (NRM) stage was markedly longer than the NSAIDs group [57]. Therefore, according to the literature, the type of treatment can affect the quality of sleep and the prevalence of sleep disorders.
Disease activity can be another factor in reports on the severity of sleep disorders. In this regard, the study of Tymms et al. showed that patients with a more severe AS experience the symptoms of insomnia seven times more [10]. The findings of other studies also indicate the negative impact of more severe disease activity on the sleep quality of AS patients [6,14,18,21,37]. More disease activity causes increased structural damage and disruption of spine mobility, which can cause difficulty in changing position while sleeping and thus disturb the patients' sleep [18,40]. Also, the findings indicate a significant positive correlation between sleep disorder scores and metrics (measurements related to the pelvis and spine) measured by Bath Ankylosing Spondylitis Metrology Index (BASMI) [6,19,28]. However, reported  [38].
Night pain is one of the prominent features of AS disease, which can affect the quality of sleep among patients [16]. The findings of existing literature indicate a significant relationship between pain experience and sleep quality disruption [18,20,22,37,38]. In a study by Nie et al., nighttime back pain was one of the key factors resulting in sleep disturbance. Experiencing pain before and during the night causes problems in falling asleep and reduces the duration and quality of sleep [20].
AS also significantly impacts a patient's psychological state [23]. In the study by Zhang et al., the prevalence of depression among AS patients was reported about 35% [23]. Furthermore, the findings of Jiang et al. and Nie et al. showed that 31.6% and 48% of AS patients experience anxiety, respectively [16,20]. Additionally, the prevalence of anxiety disorders in patients suffering from insomnia and patients without insomnia was reported as 76.1% and 33.3%, respectively [22].
Regarding the relationship between educational level and sleep quality, the results of the study by Li et al. showed that there is a significant negative relationship between the overall score of sleep quality and years of education, which means that the higher a patient's education, the less likely she/he is to experience a sleep disorder [38]. Also, in the study of Jiang et al., higher education patients experienced less anxiety and depression [16]. Therefore, according to pertinent literature, higher education seems to enable a patient to gain sufficient knowledge and awareness about the disease and its associated considerations. This also can help the patient to be able to control the disease and its symptoms. Nonetheless, in the findings of the study of Nie et al. no relationship between sleep disorder and the level of education of a patient was found [20].
The findings of the study of Günaydin et al. showed that the experience of fatigue in AS patients is influenced by sleep disorder [27]. Moreover, as sleep disturbance can increase the experience of pain, it seems that sleep disturbance lowers the threshold of tolerance and strengthens the pain signals, thus resulting in the person focusing more on the feeling of pain [38]. In the study of Purabdollah et al., it was reported that a significant relationship was found between sleep disorders and pain [58][59][60][61][62]. This study states that the severity of pain and sleep problems can predict inflammatory markers that can be clues to the severity of the disease. Therefore, relieving pain and improving sleep can reduce the severity of the disease [63].
This study found a relatively high pooled prevalence of sleep disorders among AS patients. Also, it is known that sleep disorders affect various aspects of patients' lives. Therefore, regular follow-ups and therapeutic interventions, including non-pharmacological treatments, as well as effective and quality education offered by healthcare providers, are vital to improving the patients with AS physical and mental health. Moreover, pertinent interventions and policies will be instrumental in treating and preventing sleep disorders among patients.

Limitations
One of the limitations of the study is that we included a number of researchers that had not used specific sleep disorder questionnaires to report the prevalence of sleep disorders. Additionally, two used PSQI cut-off points were used for measuring sleep disorders within the included studies. Moreover, the included studies used different sleep disorder questionnaires to report the prevalence. The report of sleep disorder was based on self-report questionnaires that provide the possibility of subjective evaluation of sleep quality. The mentioned cases may have affected the accurate reporting of the prevalence of sleep disorders in AS patients. Another limitation of the current research was that many studies could not be included due to either lack of access to their full text or their omissions after the quality evaluation stage.

Conclusion
The findings of the present study indicate the high prevalence of sleep disorders among AS patients, which means that policymakers and service providers in the field of healthcare, including nurses, should pay more attention to this population of patients and devise necessary interventions for timely diagnosis and treatment. Effective and quality educational and therapeutic interventions for patients and their families require regular follow-up. Additionally, more support for AS patients would be necessary to prevent and appropriately treat sleep disorders and associated complications.