Background
Epidemiological studies reported controversial results regarding the relationship between cholelithiasis, cholecystectomy and colorectal cancer (CRC). In the presence of reverse causality and confounding factors, findings of our previous retrospective study that it was gallbladder disease rather than cholecystectomy that was a risk factor for colorectal cancer was not sufficiently convincing. Therefore, we used Mendelian randomization (MR) to further explore the relationship between cholelithiasis or cholecystectomy and CRC.
Methods
We performed a univariate MR (UVMR) and multivariate MR (MVMR) analysis of 1,054,773 samples and 37,970,958 SNPs from three European genome-wide association studies (GWAS) to explore the causality of cholelithiasis, cholecystectomy and CRC. The forward analysis, with cholelithiasis or cholecystectomy as exposure and CRC as outcome, included UVMR and MVMR analysis. In UVMR, 21 and 30 independent SNPs strongly (P < 5*10− 8) associated with cholelithiasis on CRC and cholecystectomy on CRC were extracted as valid instrumental variables (IVs); in MVMR, 14 and 26 valid IVs were extracted, respectively. The inverse analysis, with CRC as exposure and cholelithiasis or cholecystectomy as outcome, only included MVMR. 9 and 7 IVs strongly associated with CRC on cholelithiasis and CRC on cholecystectomy were extracted, respectively. MR results were estimated using multiplicative random effects-inverse variance weighted (MRE-IVW), simple mode (SM), weighted median (WME), weighted mode (WMO) and MR-Egger regression methods, respectively. Sensitivity analysis was performed using heterogeneity test, pleiotropy test and leave-one-out test. MR results were presented using scatter plots, forest plots and funnel plots.
Results
In the forward analysis, neither UVMR nor MVMR analysis estimated by MRE-IVW revealed a causal relationship between cholelithiasis on CRC (OR = 1.0002, 95% CI = 0.999–1.001, P = 0.729 and OR = 1.0003, 95% CI = 0.998–1.003, P = 0.799, respectively) or between cholecystectomy on CRC (OR = 0.9917, 95% CI = 0.963–1.022, P = 0.582 and OR = 0.9897, 95% CI = 0.936–1.046, P = 0.715, respectively). In the inverse analysis, MRE-IVW estimates also revealed little evidence for a causal relationship between CRC on cholelithiasis (OR = 0.0504, 95% CI = 0.001–2.871, P = 0.147) or between CRC on cholecystectomy (OR = 0.9894, 95% CI = 0.792–1.236, P = 0.925). Results from other MR estimation methods were consistent with MRE-IVW. Sensitivity analysis confirmed the stability and reliability of results.
Conclusions
Our two-sample univariate and multivariate MR analysis revealed neither cholelithiasis nor cholecystectomy was causally associated with colorectal cancer.