Abstract Background Osteoarthritis (OA) is a common chronic degenerative joint disease globally. While sleep-related factors are linked to its pathogenesis, the specific association and nonlinear characteristics between sleep midpoint (the midpoint of the sleep cycle) and OA remain unclear. Methods Data for this study were obtained from the National Health and Nutrition Examination Survey (NHANES 2015–2020). Sleep midpoint was calculated from self-reported weekday bedtime and wake-up time. Restricted cubic spline (RCS) models analyzed the relationship between sleep midpoint and OA, adjusted for demographic, lifestyle, and clinical covariates. Subgroup analyses and propensity score matching verified result robustness. Results A total of 7640 participants were included in this study, of whom 977 had OA. RCS revealed a “J-shaped” association with an inflection point at 2:30 AM. No significant association was found for sleep midpoint < 2:30 AM (OR = 0.95, 95% CI = 0.88–1.02). For sleep midpoint ≥ 2:30 AM, each 30-min delay was associated with an 8% increased OA prevalence (OR = 1.08, 95%CI = 1.00–1.18). Subgroup and matched analyses confirmed robustness. Conclusions Our study results indicate that in a representative sample of US adults, a delayed sleep midpoint (≥ 2:30 AM) is independently associated with OA, suggesting that a delayed sleep midpoint beyond 2:30 AM is linked to a higher prevalence of OA. This highlights the potential role of sleep midpoint in OA prevention and management. Key Points • Sleep midpoint shows a nonlinear association with osteoarthritis prevalence, with an inflection around 02:30 AM. • After approximately 02:30 AM, each 30-min delay in sleep midpoint is associated with ~ 8% higher adjusted odds of osteoarthritis, independent of demographics, lifestyle factors, comorbidities, and sleep duration. • Results are consistent across subgroups and remain robust in propensity score–matched and sensitivity analyses using survey weighted, nationally representative US data. • Sleep timing is modifiable, indicating a practical target for risk reduction that warrants prospective and interventional evaluation.
Yang et al. (Sat,) studied this question.