Background
Higher adiposity and related metabolic alterations have been implicated in endometrial cancer (EC) aetiology, yet the precise mechanisms underlying these relationships remain incompletely understood. 
 
Objectives
We aimed to identify circulating metabolites associated with EC risk and to evaluate the extent to which the metabolomic signatures of EC risk could help identify biological processes relating to adiposity traits and EC. 
 
Methods
We assessed associations between pre-diagnostic circulating metabolite levels and EC risk, using data from Prostate, Lung, Colorectal and Ovarian screening trial (PLCO), Nurse’s Health study (NHS) I, NHS II for discovery analyses and data from the European Prospective Investigation into Cancer and Nutrition (EPIC) for replication. Mendelian randomisation (MR) analyses were conducted to investigate the potential causal roles for identified risk-associated metabolites in EC. Secondary analyses, including MR and profile comparison analyses, were performed to evaluate the extent to which the metabolomic signature of EC risk could be explained by body mass index (BMI) and five adiposity distribution traits. 
 
Results
Prospective analyses identified 44 metabolites associated with EC, including 18 amino acids showing positive associations. Formiminoglutamate showed the strongest association: Hazards ratio (HR) for 1 SD increase1.63; 95% confidence interval (CI): 1.38, 1.93; p=1.39x10-8. Eleven lipid metabolites, particularly several sphingolipids, were positively associated with EC, while 2 glycerophospholipids, 1-(1-enyl-palmitoyl-2-oleoyl-GPC and C34-3-PC-plasmalogen, showed inverse associations. In addition to the 44 metabolites associated with EC, 124 metabolites from related pathways exhibited directionally consistent estimates and were associated with EC (p<0.05) in EPIC. Of the risk-associated metabolites, 15 were found to be potentially causally associated with EC in MR analyses, though these likely reflect pathway involvement rather than effects of specific metabolites due to the non-specific nature of the instruments. Crude assessment of possible adiposity involvement through adjustment for BMI partly attenuated the HR estimates for all metabolites. MR and profile comparison analyses suggested that some metabolites associated with EC are likely to be influenced by BMI. For example, higher BMI was associated with higher levels of mannose (beta per SD increase in BMI was 0.39; p=1.16x10-19), which was positively associated with EC (HR: 1.49; 95%CI: 1.28, 1.75; p=6.96x10-7). We also found evidence that abdominal subcutaneous adipose tissue (ASAT) was associated with some metabolites associated with EC risk, many of which are robustly associated with BMI. 
 
Conclusions/Implications
This study provides evidence of shared biology between adiposity traits and EC by highlighting a range of metabolites associated with the risk of developing EC and the extent to which adiposity traits influence changes in these metabolite levels. The strategic use of prospective and MR studies enabled the strengths of these study designs (direction and cause) to be exploited for the assessment of EC aetiology. Work leading on from this is now investigating whether obesity management, such as bariatric surgery, modify adiposity-driven metabolic pathways linked to EC risk through metabolomic analyses of blood and endometrial tissue samples collected pre- and post-bariatric surgery.