Causal relationship of genetically predicted gut microbiota with thyroid cancer: a bidirectional two-sample mendelian randomization study

Abstract

Background: Previous investigations have demonstrated a correlation between the composition of gut microbiota and the development of thyroid cancer (TC). Nonetheless, there was no consensus on the causal effect of gut microbiota composition on TC risk. Therefore, the present study aimed to perform a bidirectional two-sample Mendelian randomization (MR) analysis to explore potential causal associations between gut microbiota and TC risk. Methods: Utilizing data from the MiBioGen consortium’s genome-wide association studies (GWAS) meta-analysis involving a sample size of 18,340, we identified instrumental variables for 211 gut microbiota taxa. The summary statistics for TC was from relevant large-scale GWAS conducted by the FinnGen consortium. In the first stage, the Inverse-variance weighted (IVW) method was used as the primary estimate method, and the stability of estimations was tested by a battery of sensitivity analyses. In the second stage, a reverse MR analysis was applied to determine whether reverse causality existed. Results: According to the IVW method, we identified 9 genetically predicted gut microbiota that were causally correlated with TC risk. Among them, we observed a positive causal effect of Family Christensenellaceae (OR = 1.664, 95% CI: 1.103–2.511, P = 0.015), Family Victivallaceae (OR = 1.268, 95% CI: 1.009–1.594, P = 0.042), Genus Methanobrevibacter (OR = 1.505, 95% CI: 1.049–2.159, P = 0.027), Genus Ruminococcus2 (OR = 1.846, 95% CI: 1.261–2.704, P = 0.002), Genus Subdoligranulum (OR = 1.907, 95% CI: 1.165–3.121, P = 0.010), Phylum Verrucomicrobia (OR = 1.309, 95% CI: 1.027–1.668, P = 0.029) on TC risk, while Class Betaproteobacteria (OR = 0.522, 95% CI: 0.310–0.879, P = 0.015), Family Family XI (OR = 0.753, 95% CI: 0.577–0.983, P = 0.037), Genus Sutterella (OR = 0.596, 95% CI: 0.381–0.933, P = 0.024) might be correlated with a decreased risk of TC. Subsequently, various sensitivity analyses indicated no heterogeneity, directional pleiotropy or outliers. In addition, reverse analysis demonstrated a negative causal effect of TC risk on the abundance of the gut microbiota (Genus Ruminococcus2, OR = 0.947, 95% CI: 0.907–0.989, P = 0.014). Conclusion: Genetic evidence suggested that bidirectional causal associations of specific bacteria taxa and the risk of TC, highlighting the association of the “gut-thyroid” axis. Further exploration of the potential microbiota-related mechanisms might have profound implications for public health in terms of the early prevention and treatment of TC.