Causal relationship between Physical activity, leisure sedentary behaviors and hip/keen osteoarthritis risk: a Mendelian randomization study

Abstract Objective: We used Mendelian randomization to examine the causal effects of physical activity (PA) and leisure sedentary behaviour (LSB) on the risk of knee and hip osteoarthritis (KOA/HOA). Methods: We used large-scale datasets of PA (N=377,000) and LSB (N=422,218) as exposures, and KOA (N=403,124), HOA (N=398,873), and KOA and/or HOA (N=417,596) as outcomes. We applied inverse-variance weighted, MR-Egger and weighted median methods to estimate causal effects. The validity of the instrumental variables was assessed using Cochran’s Q test, MR-Egger intercept test and MR-PRESSO test. We performed leave-one-out analysis and funnel plot to examine the influence of individual variants and potential publication bias. We also conducted risk factor analyses to explore possible mediators. Results: The genetic prediction of PA increased the risk of KOA (OR: 1.29; P: 0.02). Television viewing raised the risk of KOA (OR: 1.34; P: 5.30e-5) and HOA (OR: 1.66; P: 1.77e-15). However, computer use lowered the risk of KOA (OR: 0.80; P: 0.01) and KOA and/or HOA (OR: 0.83; P: 0.01). No effect of PA, fraction of accelerations >425 milli-gravities and driving on KOA/HOA was observed. Risk factor analysis indicated that multiple risk factors may mediate the above causal relationships.