A bidirectional two-sample Mendelian randomization study on oral disease and risk of inflammatory bowel disease
10.3760/cma.j.cn311367-20231125-00177
- VernacularTitle:口腔疾病与炎症性肠病风险的双向两样本孟德尔随机化研究
- Author:
Yaxin XU
1
;
Yanan GAO
;
Jun XU
;
Ting YAO
;
Yamei CHEN
Author Information
1. 同济大学医学院，上海 200092
- Publication Type:Journal Article
- Keywords:
Mendelian randomization analysis;
Mouth diseases;
Inflammatory bowel diseases;
Causality
- From:
Chinese Journal of Digestion
2024;44(8):532-539
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To explore the bidirectional causal relationship between genetically predicted oral diseases and inflammatory bowel disease (IBD) by Mendelian randomization (MR) and meta analysis.Methods:Genetic datas of 4 oral diseases, including periodontitis, lichen planus, oral ulcer, and dental caries were extracted from the Integrative Epidemiology Unit (IEU) open genome-wide association study project (IEU open GWAS), and genetic datas related to IBD, ulcerative colitis (UC), and Crohn′s disease (CD) were extracted from 3 independent databases, including the meta-analysis of de Lange et al, the FinnGen database, and the International Inflammatory Bowel Disease Genetics Consortium (IIBDGC). Single nucleotide polymorphism (SNP) identified as instrumental variable (IV) for " exposure-outcome" relationship was screened by setting the association strength and eliminating linkage disequilibrium, and confounder SNP and palindromic SNP were removed at each database level. Weak IV bias was excluded according to the F-value. Inverse variance weighted (IVW) method was used as the main analysis method, and Cochran′s Q test was used to analyze the heterogeneity of a single database, and random effect or fixed effect model was selected for MR analysis according to the presence or absence of heterogeneity. The MR-Egger test and Mendelian randomization-pleiotropy residual sum and outlier(MR-PRESSO) test were used to remove outliers and evaluate horizontal pleipotency across multiple databases, and then meta-analysis was performed to summarize the effect value of multiple databases. The effect of individuol SNP on the outcome was determined by a sensitivity test using the leave-one-out method. Results:The results of the MR analysis indicated that genetically predicted lichen planus had a significant effect on IBD ( OR=1.069, 95% confidence interval (95% CI) 1.043 to 1.097, Z=5.25, P<0.001), CD ( OR=1.077, 95% CI 1.018 to 1.139, Z=2.56, P=0.010) and UC ( OR=1.075, 95% CI 1.040 to 1.111, Z=4.31, P<0.001). The reverse MR analysis revealed that IBD was associated with an increased risk of periodontitis ( OR=1.051, 95% CI 1.020 to 1.083, Z=3.25, P=0.001), lichen planus ( OR=1.166, 95% CI 1.011 to 1.344, Z=2.11, P=0.035), and oral ulcer ( OR=1.003, 95% CI 1.001 to 1.004, Z=4.28, P<0.001). There was a causal association between UC and periodontitis ( OR=1.043, 95% CI 1.009 to 1.077, Z=2.51, P=0.012). CD was significantly correlated with an increased risk of lichen planus ( OR=1.088, 95% CI 1.038 to 1.141, Z=3.48, P<0.001). Conclusions:Lichen planus may be a risk factor of the progression of IBD, and IBD may be associated with increased risk of multiple oral diseases. There may be a potential bidirectional causal relationship between oral diseases and IBD.
