目的·利用孟德尔随机化法（Mendelian randomization，MR）研究1型糖尿病与结直肠癌间潜在的因果关系。方法·采用两样本双向MR分析评估1型糖尿病与结直肠癌的因果关系。研究数据均来自IEU Open GWAS Project数据库。1型糖尿病的数据集包括患者9 266例和对照人群15 574例，包含12 783 129个单核苷酸多态性（single nucleotide polymorphism，SNP）的关联分析；结直肠癌的数据集包括患者5 657例和对照人群372 016例，包含29 999 696个SNP的关联分析。筛选工具变量SNP，以逆方差加权（inverse variance weighted，IVW）法结果作为效应的主要指标，同时将MR-Egger回归、加权中位数法、基于众数的简单估计、基于众数的加权估计4种方法结果作为参考。采用留一法检验敏感性，采用IVW法和MR-Egger法进行Cochran's Q检验判断异质性，MR-pleiotropy函数检验多效性，采用Steiger检验进行方向性研究。采用共定位分析评估1型糖尿病和结直肠癌之间的效应是否由相同的SNP引起，采用交叉性状连锁不平衡得分回归（linkage disequilibrium score regression，LDSC）分析2种疾病之间的遗传相关性。所有检验采用R语言软件（4.3.1版本）执行分析。结果·经筛选后，共采用工具变量（SNP）33个。异质性检验结果发现，SNP之间存在一定的异质性（IVW法和MR-Egger法结果均P<0.05），因此效应评估采用随机效应模型的结果。正向MR分析结果显示，IVW法、MR-Egger法、加权中位数法、基于众数的加权估计均发现1型糖尿病对结直肠癌存在显著的因果效应（均P<0.05）；敏感性分析显示，结果稳定。多效性检验未检测到多效性（P>0.05）。Steiger检验发现，1型糖尿病对结直肠癌的效应未受反向作用干扰。反向MR分析未发现结直肠癌对1型糖尿病存在因果效应（均P>0.05）。共定位分析结果显示，H₄假设概率为45.7%，2种疾病间的因果关系不是由两者基因序列中相同的SNP引起的。LDSC分析显示2种疾病不存在遗传相关性。结论·1型糖尿病可能促进结直肠癌发生，但结直肠癌对1型糖尿病不存在影响。

Objective ·To investigate the potential causal relationship between type 1 diabetes and colorectal cancer by using Mendelian randomization (MR). Methods ·Two-sample bidirectional MR was used to investigate the causal relationship between type 1 diabetes and colorectal cancer. All research data were collected from the IEU Open GWAS Project database. The dataset of type 1 diabetes included 9 266 cases and 15 574 controls, with correlation analysis in 12 783 129 single nucleotide polymorphisms (SNPs); the dataset of colorectal cancer included 5 657 cases and 372 016 controls, with correlation analysis in 29 999 696 SNPs. The instrumental variables SNPs were screened. The results derived from the inverse-variance weighted (IVW) method were used as the main indicator of effect. The results derived from other four methods, namely MR-Egger regression, weighted median, simple mode, and weighted mode, were used as reference. Sensitivity was analyzed with the leave-one-out method. Heterogeneity was analyzed with Cochran's Q test by using both IVW and MR-Egger methods. Pleiotropy was analyzed with MR-pleiotropy function, and Steiger test was used for directional research. The colocation analysis was used to find out whether the causal relationship between type 1 diabetes and colorectal cancer was caused by the same SNP. The genetic correlation between 2 diseases was analyzed by using the linkage disequilibrium score regression (LDSC). All tests were analyzed by using R language software (version 4.3.1). Results ·After being screened, a total of 33 instrumental variables (SNPs) were used. The heterogeneity test results showed that there was heterogeneity among the SNPs (IVW and MR-Egger: P<0.05), so the effect evaluation was based on the results of the random effect model. MR analysis showed that type 1 diabetes had a significant causal effect on colorectal cancer (P<0.05) by using IVW, MR-Egger, weighted median and weighted mode. Sensitivity analysis showed that the results were stable. Pleiotropy was not detected in pleiotropy test (P>0.05). Steiger test showed that the effect of type 1 diabetes on colorectal cancer was not interfered with by the reverse effect. Reverse MR analysis showed no causal effect of colorectal cancer on type 1 diabetes (P>0.05). The results of colocalization analysis showed that the probability of H₄ hypothesis was 45.7%, and the causal relationship between the 2 diseases was not caused by the same SNP in the gene sequences. LDSC analysis demonstrated that there was no genetic correlation between the two diseases. Conclusion ·Type 1 diabetes may promote colorectal cancer, but colorectal cancer has no effect on type 1 diabetes.