Abstract
BACKGROUND: Previous studies have shown conflicting results regarding the impact of circulating antioxidants on the risk of inflammatory bowel disease (IBD). In this study, our intent was to investigate the causal relationship between circulating antioxidants and IBD using Mendelian randomization (MR). METHODS: Instrumental variables for absolute circulating antioxidants (ascorbate, retinol, lycopene, and β-carotene) and circulating antioxidant metabolites (α-tocopherol, γ-tocopherol, ascorbate, and retinol) were screened from published studies. We obtained outcome data from two genome-wide association study (GWAS) databases, including the international inflammatory bowel disease genetics consortium (IIBDGC, 14,927 controls and 5,956 cases for Crohn's disease (CD), 20,464 controls and 6,968 cases for ulcerative colitis (UC), and 21,770 controls and 12,882 cases for IBD) and the FinnGen study (375,445 controls and 1,665 cases for CD, 371,530 controls and 5,034 cases for UC, and 369,652 controls and 7,625 cases for IBD). MR analysis was performed in each of the two databases and those results were pooled using meta-analysis to assess the overall effect of exposure on each phenotype. In order to confirm the strength of the findings, we additionally conducted a replication analysis using the UK Biobank. RESULTS: In the meta-analysis of the IIBDGC and FinnGen, we found that each unit increase in absolute circulating level of retinol was associated with a 72% reduction in the risk of UC (OR: 0.28, 95% CI: 0.10 to 0.78, P=0.015). The UC GWAS data from the UK Biobank also confirmed this causal relationship (OR: 0.99, 95% CI: 0.97 to 1.00, P=0.016). In addition, there was suggestive evidence that absolute retinol level was negatively associated with IBD (OR: 0.41, 95% CI: 0.18 to 0.92, P=0.031). No other causal relationship was found. CONCLUSION: Our results provide strong evidence that the absolute circulating level of retinol is associated with a reduction in the risk of UC. Further MR studies with more instrumental variables on circulating antioxidants, especially absolute circulating antioxidants, are needed to confirm our results.