AbstractPancreatic ductal adenocarcinoma (PDAC) is an uncommon but highly fatal malignancy. Identifying causal metabolite biomarkers offers an opportunity to facilitate effective risk assessment strategies for PDAC. In this study, we performed a two‐sample Mendelian randomization (MR) study to characterize the potential causal effects of metabolites in plasma on PDAC risk. Genetic instruments were determined for a total of 506 metabolites from one set of comprehensive genome‐wide association studies (GWAS) involving 913 individuals of European ancestry from the INTERVAL/EPIC‐Norfolk cohorts. Another set of genetic instruments was developed for 483 metabolites from an independent GWAS conducted with 8299 individuals of European ancestry from the Canadian Longitudinal Study on Aging (CLSA) cohort. We analyzed GWAS data of the Pancreatic Cancer Cohort Consortium (PanScan) and the Pancreatic Cancer Case‐Control Consortium (PanC4), comprising 8275 PDAC cases and 6723 controls of European ancestry. The association of metabolites with PDAC risk was assessed using the inverse‐variance weighted (IVW) method, and complemented with sensitivity analyses of MR‐Egger and MR‐PRESSO tests. Potential side effects of targeting the identified metabolites for PDAC intervention were further evaluated by a phenome‐wide MR (Phe‐MR) analysis. Forty‐four unique metabolites were identified to be significantly associated with PDAC risk, of which four top‐ranking metabolites (X: 12798, X: 11787, X: 11308 and X: 19141) showed replication evidence when using instruments developed from both two cohorts. Our results highlight novel blood metabolites related to PDAC risk, which may help prioritize metabolic features for PDAC mechanistic research and further evaluation of their potential role in PDAC risk assessment.