Combined and interactive associations of body mass index and physical activity patterns with osteoarthritis risk: a prospective cohort study in the UK Biobank

Abstract Background Obesity and insufficient physical activity are modifiable factors associated with osteoarthritis (OA), but their combined and interactive associations with OA risk remain unclear. We examined the independent, combined, and interactive associations of body mass index (BMI) and physical activity (PA) with OA risk and characterized their dose-response relationships. Methods This prospective cohort study included 422,584 participants from the UK Biobank. PA patterns were classified as inactive, weekend warrior, or regularly active, and BMI was categorized as normal and below weight, overweight, or obesity. Cox proportional hazards models were used to assess associations of PA and BMI with OA risk. Additive and multiplicative interactions were evaluated using Aalen additive hazards models and Cox interaction models, respectively. Restricted cubic spline models were used to characterize dose-response relationships. Results During follow-up, 62,203 participants developed OA. Compared with participants with normal and below weight, those with overweight and obesity had higher OA risks (HR = 1.34, 95% CI: 1.32–1.37; HR = 2.01, 95% CI: 1.97–2.06, respectively). Inactivity was associated with a modestly higher OA risk than regular activity (HR = 1.04, 95% CI: 1.02–1.06), whereas the weekend warrior pattern was not significantly associated with OA risk. Obesity combined with inactivity showed the highest OA risk and a positive additive interaction, corresponding to 92.16 additional OA cases per 100,000 person-years (95% CI: 29.74–154.58), while no multiplicative interaction was observed. The PA ranges associated with lower OA risk were 150.0–626.1 min/week among participants with normal and below weight, 150.0–374.7 min/week among those with overweight, and 150.0–448.9 min/week among those with obesity. Conclusions Higher BMI and inactivity were associated with increased OA risk, with obesity and inactivity showing a positive additive interaction. PA did not fully offset the higher OA risk associated with elevated BMI, although the PA ranges associated with lower OA risk differed across BMI categories.