BackgroundWe aimed to investigate the effects of blood lipids and lipid-lowering agents on osteoarthritis (OA) risk.Materials and methodsWe performed Mendelian randomization (MR) analyses to estimate the causal effect of blood low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), and triglyceride (TG) levels on knee and hip OA. Single nucleotide polymorphisms (SNPs) were selected from large genome-wide association studies (GWASs) of individuals of European ancestry as genetic instruments for blood lipid levels. The associations of selected genetic instruments with knee and hip OA were estimated in a recent GWAS of the UK Biobank and arcOGEN datasets. Univariate and multivariate MR analyses were performed to detect and adjust for potential pleiotropy. Furthermore, genetic instruments in HMGCR, NPC1L1, and PCSK9 regions were used to mimic LDL-C-lowering effects of statin, ezetimibe, and evolocumab, respectively.ResultsGenetically determined LDL-C increments led to reduced risks of both knee OA (OR = 0.91 per 1-SD increment, 95% CI: 0.86–0.95, P = 6.3 × 10−5) and hip OA (OR = 0.92, 95% CI: 0.85–0.99, P = 0.027). Multivariate MR analysis proved that the effect was independent of HDL-C, TG, and body mass index. TG increment was associated with reduced risks of hip OA in the univariate MR analysis; however, this was not supported by the multivariate MR analysis. Genetically proxied LDL-C-lowering effects of statins are related to increased risks of knee OA but not hip OA.ConclusionsThe findings suggested that LDL-C increments have independent protective effects on both knee and hip OA. LDL-C-lowering effects of statins may increase the risk of knee OA.