Plasma Circulating Vitamin C Levels and Risk of Endometrial Cancer: A Bi-Directional Mendelian Randomization Analysis.

BACKGROUND

Observational studies indicated that circulating vitamin C (VitC) levels may be correlated with the risk of endometrial cancer (EC). However, the causal effects and direction between them were still unclear.

METHODS

In this study, 11 single nucleotide polymorphisms (SNPs) robustly correlated with plasma VitC levels were extracted from the latest genome-wide association study (GWAS), containing 52,018 individuals. Genetic data of EC were obtained from the Endometrial Cancer Association Consortium (ECAC) (12,906 cases and 108,979 controls). An inverse-variance weighted method was utilized as the primary analysis of Mendelian randomization (MR), supplemented by the weighted median, MR Pleiotropy Residual Sum and Outlier test (MR-PRESSO), and MR-Egger methods. Additional sensitivity analyses excluding 3 SNPs with secondary phenotypes were conducted to rule out the possible pleiotropic effects. Potential impacts of several risk factors of EC, such as obesity, body mass index (BMI), hypertension, and diabetes on VitC levels, were assessed. We additionally evaluated the effects of VitC on LDL cholesterol levels, HDL cholesterol levels, and triglycerides levels to probe into the possible mediators in the VitC-EC pathway.

RESULTS

Genetically predicted higher plasma VitC levels (per 1 SD increase, approximately 20 μmol/L) were causally associated with an increased risk of EC overall [odds ratio () 1.374, 95% 1.128-1.674, = 0.0016], supported by complementary sensitivity analyses. In the subgroup analyses, genetically predicted higher levels of VitC were associated with a tendency of increased risks of both endometrioid ( 1.324, 95% 0.959-1.829, = 0.0881) and non-endometrioid histology ( 1.392, 95% 0.873-2.220, = 0.1647) while without statistical significance. The association remained significant after the exclusion of the three pleiotropic SNPs ( 1.394, 95% 1.090-1.784, = 0.0082). The confounders and mediators were unlikely to affect the VitC-EC relationship. The causal effect of EC on VitC levels was not supported ( 1.001, 95% 0.998-1.004, = 0.4468).

CONCLUSIONS

This bi-directional MR study demonstrated a causal risk role of higher circulating VitC at physiological levels on an increased risk of EC, which was independent of confounders and mediators. Further studies are warranted to elucidate the possible mechanisms.