Abstract: BACKGROUND: In recent years, epidemiological studies have shown that sleep patterns are risk factors for osteoarthritis, but the causal relationship between sleep characteristics and osteoarthritis remains unknown.
OBJECTIVE: To investigate the causal relationship between seven sleep phenotypes and osteoarthritis, thereby providing a theoretical foundation for clinical prevention and intervention of osteoarthritis. 
METHODS: Seven sleep-related features, namely sleep duration, wake-up time, daytime napping, morning/evening preference, snoring, insomnia, and hypersomnia, were selected from published genome-wide association studies. Instrumental variables for these sleep-related features were extracted. Instrumental variables for knee osteoarthritis and hip osteoarthritis were obtained from publicly available genome-wide association studies. Causal relationships between sleep characteristics and outcome risks were evaluated using two-sample and multivariable Mendelian randomization analyses. The inverse variance weighted method was employed as the primary Mendelian randomization approach. Various methods, including weighted median, weighted mode, Mendelian randomization-Egger regression, Mendelian randomization pleiotropy-residual sum and outlier, were utilized to detect and correct for the presence of pleiotropy.
RESULTS AND CONCLUSION: The results of the inverse variance-weighted method in the two-sample Mendelian randomization study revealed a detrimental causal association between the duration of sleep and the incidence risk of knee osteoarthritis [odds ratio (OR)=0.621, 95% confidence interval (CI): 0.470-0.822, P=0.001]. Concurrently, insomnia displayed a positive causal connection with hip osteoarthritis risk (OR=2.016, 95% CI: 1.249-3.254, P=0.005). Sensitivity analysis affirmed the robustness of these causal relationships, and Mendelian randomization-Egger intercept analysis found no evidence of potential horizontal pleiotropy (knee osteoarthritis: P=0.468, hip osteoarthritis: P=0.551). Moreover, the results from the multivariable Mendelian randomization analysis showed that the causal association between insomnia and hip osteoarthritis lacked statistical significance (P=0.715). In contrast, sleep duration exhibited a direct negative causal relationship with the incidence risk of knee osteoarthritis (OR=0.526, 95% CI: 0.336-0.824, P=0.005). Reverse Mendelian randomization analysis indicated that knee osteoarthritis did not influence sleep duration (P=0.757). These findings indicate a negative correlation between sleep duration and incidence risk of knee osteoarthritis, suggesting that correcting insufficient sleep might mitigate the incidence risk of knee osteoarthritis.

Key words: sleeplessness, sleep duration, knee osteoarthritis, Mendelian randomization study, causal association