Exploring potential causal associations between autoimmune diseases and colorectal cancer using bidirectional Mendelian randomization

Abstract

Several observational studies have revealed an association between autoimmune diseases (AIDs) and colorectal cancer (CRC), although their causal association remained controversial. Therefore, our study used a two-sample Mendelian randomization (MR) analysis to verify the causal association between AIDs and CRC. We employed three common MR approaches, including inverse variance weighted (IVW), weighted median, and MR-Egger methods, to assess the causal association between type 1 diabetes (T1D), systemic lupus erythematosus, rheumatoid arthritis, psoriasis, multiple sclerosis, juvenile idiopathic arthritis, celiac disease, and primary sclerosing cholangitis (PSC) and CRC. The reverse MR analysis was performed to assess the possibility of reverse causation. To evaluate the validity of the analysis, we also performed sensitivity analysis, such as the heterogeneity test, the horizontal pleiotropy test, and the leave-one-out sensitivity analysis, and validated the results in the validation cohort. Our results showed that genetically predicted T1D was nominally associated with a lower risk of CRC (IVW OR = 0.965, 95% CI = 0.939-0.992, P = 0.012). However, genetic susceptibility to psoriasis nominally increased the risk of CRC (IVW OR = 1.026, 95% CI = 1.002-1.050, P = 0.037). Genetically predicted PSC had a significant causal effect on the increasing risk of CRC (IVW OR = 1.038, 95% CI = 1.016-1.060, P = 5.85 × 10^-4). Furthermore, the MR analysis between PSC and the CRC validation cohort indicated consistent results. We found no causal association between genetically predicted other five AIDs and CRC (P > 0.05). The results of reverse MR analysis showed that genetically predicted CRC had no causal effect on T1D, psoriasis, and PSC (P > 0.05). The sensitivity analysis demonstrated that the results of the MR analysis were reliable. Our findings help to understand the causal association between AIDs and CRC, which deserves further investigation.

Figure 1

Flow diagram illustrating the core assumptions and study design of two-sample bidirectional MR analysis. Assumption 1: IVs are closely related to exposure. Assumption 2: IVs are not associated with confounders. Assumption 3: IVs influence the outcome only through the exposure. T1D type 1 diabetes, SLE systemic lupus erythematosus, RA rheumatoid arthritis, MS multiple sclerosis, JIA juvenile idiopathic arthritis, CD celiac disease, PSC primary sclerosing cholangitis, CRC colorectal cancer, AIDs autoimmune diseases.

Figure 2

Forest plot for results of the Mendelian randomization analysis. T1D type 1 diabetes, SLE systemic lupus erythematosus, RA rheumatoid arthritis, MS multiple sclerosis, JIA juvenile idiopathic arthritis, CD celiac disease, PSC primary sclerosing cholangitis.

Figure 3

Scatter plots of the causal effect of T1D, SLE, RA, psoriasis, MS, JIA, CD, and PSC on CRC in the MR analysis. (A) T1D on CRC. The F-statistics of IVs were from 30.38 to 196.60. (B) SLE on CRC. The F-statistics of IVs were from 29.96 to 374.85. (C) RA on CRC. The F-statistics of IVs were from 29.70 to 321.46. (D) psoriasis on CRC. The F-statistics of IVs were from 30.60 to 728.17. (E) MS on CRC. The F-statistics of IVs were from 26.86 to 525.02. (F) JIA on CRC. The F-statistics of IVs were from 33.50 to 121.17. (G) CD on CRC. The F-statistics of IVs were from 29.83 to 920.79. (H) PSC on CRC. The F-statistics of IVs were from 27.32 to 574.41. T1D type 1 diabetes, SLE systemic lupus erythematosus, RA rheumatoid arthritis, MS multiple sclerosis, JIA juvenile idiopathic arthritis, CD celiac disease, PSC primary sclerosing cholangitis, CRC colorectal cancer.