Abstract
ObjectivesTo examine the dose-response relationship of accelerometer-derived physical activity and sedentary behaviour with the risk of hip or knee osteoarthritis (OA).MethodsWe used data from an accelerometer study of the UK Biobank. From 2013 to 2015, 103,660 individuals worn a wrist-worn accelerometer for seven days. Associations between moderate-to-vigorous physical activity (MVPA), total physical activity (TPA), light physical activity (LPA), and sedentary behaviour were investigated using Cox proportional hazard regression models. We examined the dose-response associations using restricted cubic spline models.ResultsOur analyses included 81,436 individuals, with a mean age of 61.58 [SD, 7.85] years and 56.6% of them being female. During a median follow-up of 8.0 years, 2,086 hip OA and 2,729 knee OA occurred. TPA and LPA showed positive linear relationships with incident OA, and sedentary behaviour indicated a negative linear correlation, while MVPA exhibited a nonlinear association. Compared to MVPA of < 75 min/week, the hazard ratios for incident hip OA were 0.81 (95% CI, 0.70-0.93) for 75-149.9 min/week, 0.78 (95% CI, 0.69-0.89) for 150-299.9 min/week, and 0.94 (95% CI, 0.82-1.06) for ≥ 300 min/week. However, no significant association were observed between MVPA and knee OA.ConclusionA higher level of TPA and LPA is linked to a higher risk of OA, but a moderate level of MVPA is associated with a lower risk of hip OA. Compared to MVPA < 75 min/week, adhering to MVPA of 150-300 min/week reduces hip OA risk by 22% in middle-aged and older adults.</p>