关键词: 2型糖尿病, 脂质, 孟德尔随机化, 因果关系

Abstract: To systematically investigate the causal relationship between lipids and type 2 diabetes mellitus (T2DM), this study used a Mendelian randomization (MR) approach and summary data from large-scale genome-wide association studies (GWAS) to infer causal associations between 179 lipids and T2DM. Single nucleotide polymorphisms significantly associated with lipid levels were selected as instrumental variables. Analyses including inverse variance weighting and MR-Egger regression were performed, together with sensitivity analyses to verify the robustness of the findings. The results indicated that 12 lipids, such as phosphatidylcholine (O-18:0_20:4) (OR=1.075, 95% CI: 1.011–1.143, P=0.022), were positively associated with T2DM risk, suggesting their potential role in promoting disease development. In addition, 19 lipids, including phosphatidylcholine (15:0_18:2) (OR=0.960, 95% CI: 0.933–0.988, P=0.005), exhibited protective effects and may reduce the risk of T2DM. Sensitivity analyses revealed no significant horizontal pleiotropy or heterogeneity, indicating the robustness and high reliability of the findings. From a genetic perspective, this study confirms the causal roles of specific lipid molecules in the development of T2DM, providing new evidence for elucidating its underlying pathological mechanisms and identifying potential therapeutic targets.

Key words: lipids, type 2 diabetes mellitus, Mendelian randomization; causal relationship