Are Neurodegenerative Diseases Associated with an Increased Risk of Inflammatory Bowel Disease? A Two-Sample Mendelian Randomization Study.
Guanghui Cui,Shaojie Li,Hui Ye,Yao Yang,Qiuyue Huang,Yingming Chu,Zongming Shi,Xuezhi Zhang
DOI: https://doi.org/10.3389/fimmu.2022.956005
IF: 7.3
2022-01-01
Frontiers in Immunology
Abstract:Background:Several studies have shown that neurodegenerative diseases (e.g., Parkinson's disease [PD] and Alzheimer's disease [AD]) are associated with inflammatory bowel disease (IBD), but the causality and direction of their associations remain unclear. Mendelian randomization (MR) studies have explored the causal effects of IBD on PD and AD. However, only a few studies examined this reverse association. Thus, this study aimed to explore whether there are causal associations of genetically predicted PD and AD with IBD, using a two-sample MR study. Methods:Summary statistics for IBD, ulcerative colitis (UC), and Crohn's disease (CD) were derived from a genome-wide association study (GWAS) meta-analysis, which included the International IBD Genetics Consortium and the UK IBD Genetics Consortium (n=59,957). Genetic variants associated with the largest meta-analysis of GWAS of PD (n=1,474,097) and AD (n=455,258) were used as instrumental variables. We used multiple methods, including inverse variance weighted (IVW), weighted median (WM), MR-Egger regression, weighted mode, and Robust Adjusted Profile Score (RAPS) methods, to estimate the effects of genetically predicted PD and AD on IBD. To confirm the validity of the analysis, we also evaluated the pleiotropic effects, heterogeneity, and leave-one-out sensitivity analysis that drive causal associations. Results:The results of the IVW method, WM, and RAPS showed that genetically predicted PD was significantly associated with an increased risk of UC (odds ratio [OR]IVW=1.068, OR WM=1.107, OR RAPS=1.069, all P<0.05). Additionally, we found that there were significant associations of genetically predicted PD with CD (OR IVW=1.064, OR RAPS=1.065, all P<0.05) and IBD (OR IVW=1.062, OR RAPS=1.063, all P<0.05) using the IVW method and RAPS. However, there was no significant causal evidence of genetically predicted AD in IBD, UC, or CD among all MR methods. In all MR analyses, there were no horizontal pleiotropy (all P>0.05), or statistical heterogeneity. The sensitivity analysis results of the leave-one-out sensitivity analysis showed that the causal effect estimations of genetically predicted PD and AD on IBD were robust. Conclusions:Our MR study corroborated a causal association between genetically predicted PD and IBD but did not support a causal effect of genetically predicted AD on IBD. More animal experiments or population-based observational studies are required to clarify the underlying mechanisms of PD and IBD.