目的·采用两样本孟德尔随机化法（Mendelian randomization，MR）探究2型糖尿病（type 2 diabetes mellitus，T2DM）和认知功能障碍之间的因果关系。方法·从大规模全基因组关联研究（genome-wide association study，GWAS）汇总数据集中汇总与T2DM相关的工具变量，采用逆方差加权作为主要分析技术，以MR-Egger回归、加权中位数法和简单中位数法为辅，并联合应用meta分析对不同结局进行合并，分析T2DM与痴呆症、阿尔茨海默病（Alzheimer's disease，AD）和帕金森痴呆之间因果关系的可能性。采用MR-PRESSO整体测试和MR-Egger分析检查水平多效性。结果·遗传预测的T2DM与痴呆症（OR=1.11，95%CI 1.02~1.20，P=1.96×10^-2）和AD（OR=1.16，95%CI 1.04~1.30，P=8.41×10^-3）之间存在因果关系。meta分析支持T2DM与认知障碍的关联（OR=1.12，95%CI 1.05~1.20，P=4.22×10^-4）。一系列敏感性分析提示不存在异质性和水平多效性。反向MR分析结果显示各种类型痴呆对T2DM没有明显的反向因果关系。结论·T2DM与多种痴呆症的患病风险呈正相关，提示T2DM可能是认知障碍的一个重要风险因素。

Objective ·To investigate the causal relationship between type 2 diabetes mellitus (T2DM) and cognitive dysfunction using two-sample Mendelian randomisation (MR). Methods ·Instrumental variables associated with T2DM were pooled from a large-scale genome-wide association study (GWAS) dataset. Inverse variance weighting was used as the primary analytical technique, supplemented by MR-Egger regression, weighted median and simple median analyses. Meta-analysis was jointly applied to combine different endpoints and to analyse the possibility of a causal relationship between T2DM and dementia, Alzheimer's disease, and Parkinson's dementia. Horizontal pleiotropy was examined by MR-PRESSO global test and MR-Egger analysis. Results ·There was a causal relationship between genetically predicted T2DM and dementia (OR=1.11, 95%CI 1.02~1.20, P=1.96×10^-2) and AD (OR=1.16, 95%CI 1.04~1.30, P=8.41×10^-3). Meta-analysis also supported the association between T2DM and cognitive impairment (OR=1.12, 95%CI 1.05~1.20, P=4.22×10^-4). A series of sensitivity analyses suggested the absence of heterogeneity and horizontal pleiotropy. Reverse MR analysis showed no significant causal relationship of various types of dementia on T2DM. Conclusion ·T2DM is positively associated with the risk of developing various types of dementia, suggesting that T2DM may be an important risk factor for cognitive impairment.