孟德尔随机化研究：谷氨酰胺如何降低LOMG风险？2024稳健性分析详解

It is important to note that the robustness of different exposure factors in the aforementioned results varies. The negative association between glutamine and LOMG is significant in both IVW and weighted median methods, and the direction is consistent with MR-Egger, suggesting that this result is relatively robust. Although citrulline shows a significant protective effect on LOMG, its instrumental variable is based on only three independent SNPs, and thus the results should be interpreted with caution. The negative association between testosterone and MG did not reach statistical significance in the weighted median method, and due to insufficient instrumental variables after outcome data filtering, neither testosterone nor lactate could undergo reverse Mendelian randomization validation. Therefore, this article will focus on the sensitivity analysis and Bayesian colocalization analysis results for glutamine and LOMG, while the sensitivity analysis results for other exposure factors are detailed in Supplementary Table S3.