Causal association between social support and Osteoarthritis: a two- sample Mendelian randomization study

Abstract Objective: We aimed to study whether social support (SS), including the frequency of friend / family visits (FFV), leisure / social activities (LSA), and ability to talk (AC) are causally associated with Osteoarthritis (OA). Methods: Data of FFV, LSA, AC and OA in the European population was obtained from the UK Biobank public genome-wide association study (GWAS) summary statistics included in the Mrbase.Three Mendelian randomization (MR) methods, including Inverse variance weighted method(IVW), Weighted median method(WM) and MR-Egger regression were used to evaluate the potential causal effect of FFV, LSA, AC on OA ，and the potential causal effect of OA on FFV, LSA, AC. Heterogeneity was tested using the Cochran's Q test and the I2 statistic. A leave-one-out analysis and test for directional level pleiotropy by MR-Egger regression test served as a sensitivity analysis. Results: When assessing the causal effect of SS on OA, the IVW method supports causal relationship between FFV and OA risk (beta = -0.02883, SE=0.01293, P=0.02574), and causal relationship between LSA and OA risk (beta =-0.2255, SE=0.067, P =0.0007641) ; The WM method supports a causal relationship between FFV and OA risk (beta = -0.02968, SE=0.01272, P =0.01965), No significant causal relationship between LSA and OA risk was supported by the WM method (beta = -0.1748, SE=0.09266, P=0.05929); The MR-Egger regression did not support a significant causal relationship between FFV, LSA and OA risk. However, the evaluation did not find a causal relationship between AC and OA, nor find a causal relationship between OA and FFV, LSA, AC.The Cochran's Q test and I2 statistics of FFV showed some heterogeneity (MR Egger P=0.003158, I2=0.53452; IWW P = 0.00282, I2=0.53075), and the Cochran's Q test and I2 statistics of LSA showed low heterogeneity (MR Egger P=0.2916, I2=0.100719; IWWP = 0.4652, I2=0.306336). MR-Egger regression analyses of FFV, LSA on OA showed horizontal pleiotropy unlikely to affect the assessment results (FFV intercept =0.0011, P =0.369; LSA intercept =0.0037, P = 0.65). Conclusion: The results of MR analysis support that FFV and LSA may be causally associated with an decreased risk of OA.FFV and LSA may reduce or delay the development of OA.