BACKGROUND: The potential effects of insulin therapy on osteoarthritis (OA) risk are poorly understood. This study aimed to explore the causal relationship between insulin therapy and OA. METHODS: Mendelian randomization (MR) analysis was performed to examine the association between genetically proxied inhibition of insulin targets and the risk of overall, hip (HOA) and knee OA (KOA). We then performed univariable MR using summary statistics regarding insulin target genes derived from the DrugBank database. Data related to blood glucose reduction levels were used as a proxy for insulin levels. Two phenotypes, type 2 diabetes, and glycosylated hemoglobin levels, were selected as positive controls to confirm the direction and validity of the proxies. The OA datasets were derived from the UK Biobank cohort. Multivariable MR was adjusted for body mass index, sedentary behavior, cigarette smoking, frequency of alcohol intake, age, and genetic sex. RESULTS: Genetically proxied insulin therapy was associated with an increased risk of overall OA [odds ratio (OR):1.2595; 95% confidence interval (CI):1.0810-1.4675] and HOA (OR:1.4218; 95%CI:1.1240-1.7985), which remained consistent across multiple MR methods. After adjusting for confounders, we found evidence supporting a significant causal link with a higher risk of overall OA and HOA. A further two-step MR analysis revealed no significant mediation effects from the six mediators in the associations. CONCLUSION: There was a causal association between genetically proxied insulin therapy and a higher risk of OA, especially HOA.