Inclusion criteria:

Exclusion criteria:

Population:

Exposure:

Outcomes:

A systematic search was performed across electronic databases, including Medline via PubMed, CINAHL via EBSCO, EMBASE, Scopus, ScienceDirect, and Web of Science. The structure of the search strategy was prepared in accordance with the research question’s requirements and the specific formats required by the databases. Additionally, gray literature was explored using Google Scholar, ProQuest, and Sodhganga.

“Sleep disorders,” “sleep deprivation,” “sleep disturbance,” “sleep variation,” “sleep quality,” “sleep quantity,” “sleep wake disorders,” “circadian rhythm,” “sleep disorders, circadian rhythm,” “osteoarthritis,” “knee osteoarthritis,” and “hip osteoarthritis” have been applied as search terms. The comprehensive search strategy for the above databases is available in Supplementary Table 1.

Once the data and references were collected, they were converted into multiple formats, such as CSV and RIS. The studies were then imported into the Rayyan QCRI software to identify and remove duplicates. The screening process was performed in two stages: first, an initial title and abstract screening, followed by a full-text review.

During the title-abstract screening, three independent reviewers evaluated the studies based on predefined inclusion and exclusion criteria. Any conflicts were resolved through discussion with a third reviewer. Full-text articles that satisfy the inclusion criteria were then thoroughly assessed by two impartial reviewers. Disagreements arose regarding the eligibility of certain studies, prompting a third consulting reviewer to intervene and facilitate a resolution.

A record of excluded articles from the full-text screening was maintained, along with documented reasons for exclusion based on eligibility criteria. Relevant studies were categorized according to study design, osteoarthritis, and sleep components, and then systematically organized into a tabulated excel sheet.

A pre-prepared pilot standard coding form containing the study type, study setting, study period, sample size, publication year, authors, osteoarthritis assessment criteria/tool, and sleep-related judgment criteria, was utilized for significant data extraction from included studies for evidence synthesis concerning the impact of KHOA on sleep including the overall objective. The literature screening process and results are provided through the PRISMA flowchart (Figure 1).

The data extraction sheet included study characteristics, participant characteristics, KHOA assessment tools, sleep assessment methods, outcomes related to sleep disturbances and sleep quality, main findings and conclusions.

Quality appraisal was carried out independently by two reviewers. The risk of bias was assessed in accordance with the Joanna Briggs Institute (JBI) guidelines. (19)

A narrative synthesis was conducted in accordance with the Synthesis Without Meta-Analysis (SWiM) guidelines (20) to examine sleep disturbances related to KHOA. Data have been analyzed using RStudio (version 2024 12.0 + 467) software. Both non-continuous and continuous variables were reported with 95% confidence intervals (CIs), and forest plots were generated. Heterogeneity across studies was evaluated using I² statistics and the Q-test. When the P-value was ≥ 0.10 and I² < 50%, indicating homogeneity, a fixed-effects model was applied. Conversely, when the P-value was < 0.10 and I² ≥ 50%, suggesting significant heterogeneity, a random-effects model was utilized. A P-value less than 0.05 is considered statistically significant. A sensitivity analysis was conducted to determine the extent of variation in the pooled effect size across various studies by systematically removing individual studies. Subgroup analysis could not be performed due to limited data for comparisons and the presence of extreme heterogeneity.

After applying the JBI critical appraisal tool, studies were classified as low quality if they achieved a score below 50%, moderate quality if they achieved a score between 50% and 70%, and high quality if they scored above 70%.

The study characteristics demonstrated that all articles fell within the medium to high quality range. The included studies consisted of one case-control study, two cohort studies, and 14 cross-sectional studies. The studies covered cases from Brazil, Spain, China, the United Kingdom, Japan, Australia, Nigeria, and Canada (each with one study), South Korea (three studies), the United States, and Turkey (each with three studies).

Knee and/or hip osteoarthritis diagnosis and severity were determined using various criteria, including the National Institute for Health and Care Excellence (NICE) guidelines (21), American College of Rheumatology classification (22–26), International Classification of Diseases- Tenth revision (ICD-10) codes (27), and the Western Ontario and McMaster Osteoarthritis Index (WOMAC) questionnaire (16, 22–24, 28–30), which assesses pain, stiffness, and physical function. Additional measures included the Numerical Pain Rating Scale (NPRS, 0–10) for pain intensity (24, 25, 30, 31), Kellgren/Lawrence (K/L) grading for severity (15, 22, 23, 25, 29, 31–34), Knee Injury and Osteoarthritis Outcome Score (KOOS) (32), Lequesne Index for knee OA severity, Visual Analog Scale (VAS) (22, 26) for knee OA pain intensity, Hip Outcome Score (HOS) (16), modified Harris Hip Score (mHHS) (16), Steinbrocker functional classification, Hip Injury and Osteoarthritis Outcome Score (HOOS) questionnaire (25), and radiographic assessments.

The Pittsburgh Sleep Quality Index (PSQI) is a widely accepted and frequently used questionnaire for assessing sleep quality, completed subjectively by patients. A higher PSQI score indicates poorer sleep quality. Sleep disturbance assessment relied on the PSQI questionnaire in nine studies (16, 21–26, 30, 32), the Epworth Sleepiness Scale (ESS) in one study (21), the National Sleep Foundation’s sleep duration criteria in one study (31), the Medical Outcomes Study sleep questionnaire in one study (15), the ICD-10 disease code in one study (27), other self-assessment questionnaires in two studies (33, 34), and the Center for Epidemiologic Studies Depression Scale (CES-D) in two studies (28, 29). One study did not specify the sleep assessment tool used. The PSQI provides a global sleep score in addition to six sub-scores assessing sleep disturbance, sleep latency, sleep duration, subjective sleep quality, habitual sleep efficiency, use of sleep medications, and daytime somnolence, with individual domain scores from 0 to 3. Higher scores indicate poorer performance in that category.

Osteoarthritis diagnosis and sleep assessment tools were not specified clearly in one study by Feehan et al. (35).

Seven studies reported sleep disturbances or poor sleep quality among KHOA patients using various measures. The occurrence of sleep disturbances is higher in individuals with OA. All studies found a significant association between sleep quality and KHOA, indicating that greater KHOA severity is linked to poorer sleep quality. Four studies explicitly stated in their results that more than 55% of KHOA patients had a PSQI score of ≥ 5. Six studies identified additional influencing factors, including anxiety, depression, high pain scores, greater KHOA severity, and comorbidities. Regarding sleep disorders, one study reported that 20% of patients had obstructive sleep apnea (OSA), while another study found that 71% experienced a recent sleep problem, 56% reported having insomnia, and 56% reported inadequate sleep.

The assessment was conducted using the critical appraisal tools from the JBI (19), as required (Supplementary Table 2). The quality assessment of the included articles revealed that cross-sectional studies received scores ranging from 5 to 8 out of 8, indicating moderate to high quality. Cohort studies received scores ranging from 8 to 11 out of 11, reflecting high-quality research. Similarly, case-control studies attained a perfect score of 10 out of 10, confirming their high quality. Table 1 represents the characteristics of the included studies of this systematic review.

Meta-analysis was conducted using seven studies that employed a common sleep assessment tool, with PSQI score as the outcome measure for KHOA patients, reported in terms of mean and standard deviation (Figures 2–4). The pooled mean PSQI score among individuals with KHOA was 8.53, which is above the standard cutoff of 5, indicating poor sleep quality. In other words, the average PSQI score among KHOA patients was 8.53, which exceeds the established threshold of 5 for poor sleep quality. So, we can be 95% confident that the true effect size lies within the interval 7.18–9.87, indicating a statistically significant overall effect. The relatively narrow confidence interval suggests a precise estimate with a low likelihood of results occurring by chance. Heterogeneity analysis revealed substantial variability across studies, with an I² of 94.96%, indicating extremely high heterogeneity. A p-value < 0.0001 and H² = 39.31 further confirm considerable heterogeneity. The estimated level of variance (τ² = 3.0317) across studies justified the application of the random-effects model in the meta-analysis. The forest plot showed that Afsar et al. (22) had the lowest mean score (5.00), while Alburquerque-García et al. (24) had the highest (11.10). The funnel plot indicated an outlier in the lower-right region Alburquerque-García et al. (24), suggesting potential true heterogeneity due to variations in study populations. A leave-one-out sensitivity analysis was performed to evaluate the robustness of a meta-analysis systematically and to identify whether a single study had a disproportionate influence on the overall results. Ggplot showed that removing Afşar et al. (22), Alburquerque-García et al. (24) significantly impacted the effect size, increasing it to approximately 9.0 or decreasing it to around 8.2, respectively, highlighting heterogeneity in the overall estimate. This high heterogeneity (I² = 94.96%) makes interpretation of the pooled estimate more challenging. Several factors may explain this variability. First, differences in study populations could have contributed; some studies included younger patients, while others had older or more severe cases, which can affect sleep patterns. Second, cultural and regional differences in sleep habits and healthcare access may have influenced the reporting of PSQI. Third, although all studies used PSQI, the application of the tool may have varied (for example, in cut-off thresholds, versions used, or interviewer versus self-administered assessments). Fourth, differences in study design and sample size may also contribute to variability. Finally, co-morbidities such as depression, obesity, or medication use were inconsistently reported across studies and might explain variations in sleep quality.

Even though subgroup analysis was not feasible due to the small number of studies, a narrative interpretation suggests that both clinical and methodological diversity are likely sources of heterogeneity. This indicates that while the pooled effect supports an overall trend of poor sleep in KHOA patients, caution is warranted in generalizing the exact magnitude of the effect. These findings underscore the need for future research with more standardized study designs and better control for confounding variables, which will be critical to refining our understanding of sleep problems in KHOA and guiding targeted clinical interventions.

This is the inaugural systematic review focused on mapping the existing literature and assessing sleep appraisal in KHOA, a critical and significant area of concern, by analyzing data from 17 studies. Our review confirms a significant positive correlation between sleep disturbances/quality and osteoarthritis concerning KHOA. The first limitation is that only studies focusing on knee and/or hip OA are included to assess their impact on sleep quality, while other types of osteoarthritis are excluded. Additionally, the search is restricted to studies published within the last 20 years, potentially omitting relevant findings from earlier years. Furthermore, limiting the analysis to English-language studies may exclude valuable insights available in other languages. Further research should explore sleep-specific interventions for individuals with KHOA and their potential effect on disease activity and overall quality of life. Many of the studies used self-reported questionnaires, such as the PSQI, ESS, and CES-D, with very few employing objective measures like polysomnography or other formal sleep studies. This is particularly important because self-reported data are inherently subject to bias and may not accurately reflect actual sleep patterns or disturbances.