BACKGROUND: Many recent studies have shown a close relationship between inflammatory cytokines and osteoporosis and bone mineral density (BMD). However, the causal relationship between inflammatory cytokines and BMD has not been fully revealed. OBJECTIVE: To explore the potential causal relationship between inflammatory cytokines and BMD using a two-sample Mendelian randomization analysis. METHODS: The single nucleotide polymorphisms associated with 41 circulating inflammatory cytokines were selected from the open database of genome-wide association studies (GWAS) as instrumental variables. The GWAS data about BMD were from the Genetic Factors for Osteoporosis Consortium, involving a total of 32 735 individuals of European ancestry. Inverse variance weighting was used as the primary analysis to evaluate the causal effect. Weighted median, MR Egger regression, simple mode, and weighted mode methods were used to supplement the explanation. We used the MR-Egger intercept and MR-PRESSO method to conduct a pleiotropy test, the Cochran’s Q test was used to determine whether there was heterogeneity in the results, and the leave-one-out method was used to evaluate the stability of the results. In addition, to more accurately assess the causality, the Bonferroni-corrected test was used to identify inflammatory cytokines that have a strong causal relationship with BMD. RESULTS AND CONCLUSION: (1) According to the results of the inverse variance weighting method, we found a positive causal relationship between interleukin-8 and lumbar spine BMD [β=0.075, 95% confidence interval (CI): 0.033-0.117, P=0.000 5), while a negative causal relationship between interleukin-17 and lumbar spine BMD (β=-0.083, 95% CI: -0.152 to -0.014, P=0.018). There might be a negative causal relationship between tumor necrosis factor b and femoral neck BMD (β=-0.053, 95% CI: -0.088 to -0.018, P=0.003), while a positive causal relationship between basic fibroblast growth factor and femoral neck BMD (β=0.085, 95% CI: 0.016-0.154, P=0.015). There might be a negative causal relationship between macrophage inflammatory protein-1a and total body BMD (β=-0.056, 95% CI: -0.105 to -0.007, P=0.025). There was a negative causal relationship between interleukin-5 (β=-0.019, 95% CI: -0.031 to -0.006, P=0.004), stromal cell-derived factor-1a (β=-0.022, 95% CI: -0.038 to -0.005, P=0.010), hepatocyte growth factor (β=-0.021, 95% CI: -0.041 to -0.002, P=0.030), interleukin-4 (β=-0.016, 95% CI: -0.032 to -0.001, P=0.034) and heel BMD, while a positive causal relationship between nerve growth factor (β=0.019, 95% CI: 0.002-0.036, P=0.033), granulocyte colony-stimulating factor (β=0.011, 95% CI: 0.000-0.022, P=0.050), and heel BMD. Meanwhile, after the Bonferroni-corrected test, there was a strong positive causal effect between interleukin-8 and lumbar spine BMD (P=0.000 5). And consistent directional effects for all analyses were observed in MR Egger, weighted median, simple mode, and weighted mode methods. (2) Sensitivity analyses revealed no heterogeneity, pleiotropy, or outliers for the causal effect of circulating inflammatory cytokines on BMD. [ABSTRACT FROM AUTHOR]