Adiposity, metabolites and endometrial cancer risk: Mendelian randomization and Observational analyses

IntroductionIncreased and excess adiposity is associated with increased risk of endometrial cancer (EC) and both of these are associated with circulating metabolite profiles. However, how metabolites relate to the adiposity-EC relationship remains unclear. MethodsWe have brought together evidence from Mendelian randomization (MR) and observational analyses to evaluate the effect of i) adiposity traits on endometrial cancer, ii) adiposity traits on circulating metabolites and iii) adiposity-associated metabolites on EC. We have also evaluated the potential role of metabolites in the adiposity-EC relationship using multivariable MR. Observational analyses were conducted using individual level data from UK Biobank (N = 1,005 cases and 215,339 controls). MR analyses were performed using female-specific summary statistics from genome-wide association studies of body mass index (BMI; N up to 434,794), waist-to-hip ratio (WHR; N up to 381,152), 249 metabolites and ratios from targeted nuclear magnetic resonance metabolomics (N up to 140,768) and EC risk (12,906 cases and 108,979 controls). ResultsIn observational analyses, higher BMI and WHR were associated with elevated odds of overall EC (odds ratio (OR) per standard deviation (SD) increase in BMI = 1.37; 95% confidence interval (CI) = 1.19, 1.57; OR per SD increase in WHR= 1.15; 95% CI = 1.01, 1.32). In MR analysis, higher BMI was associated with elevated odds of overall EC risk (OR per SD increase in BMI = 1.80; 95% CI = 1.56, 2.07), endometrioid cancer (OR = 1.71; 95% CI = 1.45, 2.02) and non-endometrioid cancer (OR = 2.20; 95% CI = 1.55, 3.12). There was weaker evidence for a causal relationship with WHR. BMI was associated with 165 metabolites and ratios after Bonferroni-correction in MR analyses, several of which were associated with EC and 25 of which were directionally consistent with an intermediate role in the effect of BMI on EC risk from two-step MR and observational analyses. In MVMR analyses, there was evidence suggesting that the effect of BMI on non-endometrioid EC was mediated by several lipid metabolites; for example, the univariable MR OR for non-endometrioid EC per 1 SD increase in BMI was 2.51 (95%CI = 1.47, 4.29), whereas on adjusting for free cholesterol to total lipids ratio in medium LDL, the MVMR OR for non-endometrioid EC per 1 SD increase in BMI was 1.18 (95%CI = 0.53, 2.66). Further bioinformatic analyses highlighted a mixture of other potential shared pathways (including height, adiposity traits and blood cell traits) that could influence the risk of EC. ConclusionEvidence here suggests that higher BMI causes a higher risk of overall and all histological subtypes of EC and variation in numerous circulating metabolites. Several of these metabolites showed relationships consistent with an intermediate role between BMI and non-endometrioid EC, however, further bioinformatic analyses highlighted other potential shared mechanisms that could influence the risk of EC.