ObjectiveTo systematically evaluate the causal relationship between blood metabolites and pancreatic cancer (PC) risk using Mendelian randomization (MR). MethodsWe conducted a two-sample MR analysis using genetic instruments for 8 299 blood metabolites derived from a European genome-wide association study (GWAS) and PC outcome data from the GWAS Catalog. The primary analysis employed inverse-variance weighted (IVW) regression, with sensitivity analyses including MR-Egger, weighted median, weighted mode, and simple mode methods. Heterogeneity was assessed using Cochran’s Q test, pleiotropy was evaluated via MR-Egger intercept tests, and outliers were identified using MR-PRESSO. Robustness was confirmed through leave-one-out analyses. For metabolites showing significant associations (P<0.05), we performed independent replication using the same European PC GWAS cohort, followed by meta-analysis of all results. Reverse causation was excluded using Steiger directionality tests and bidirectional MR, while genetic confounding was assessed via linkage disequilibrium score regression (LDSC). ResultsAfter multi-stage screening, 26 blood metabolites were identified as significantly associated with pancreatic cancer risk (P<0.05), comprising 18 known metabolites (including lipids, amino acids, xenobiotics, coenzymes and vitamins, nucleotides, and peptides), 3 metabolite ratios, and 5 unannotated metabolites. Notably, 1-palmitoleoylglycerol 16:1 exhibited the lowest P value [OR=0.78, 95%CI (0.69, 0.89), P<0.001]. While repeated validation confirmed significance for 6 of these 26 metabolites, meta-analysis demonstrated that all 26 metabolites remained significantly associated with pancreatic cancer risk (P<0.05). Linkage disequilibrium score regression analysis revealed that, except for myristoyl glycerol (Rg=1.534, se=0.571, P=0.007), which indicated potential genetic confounding of MR estimates, the MR associations between the remaining blood metabolites and pancreatic cancer risk were not significantly influenced by potential confounders (P>0.05). ConclusionThis study provides causal evidence within the European population that some blood metabolites are associated with PC risk, identifying 1-palmitoleoylglycerol 16:1 as a novel protective biomarker and highlighting targeting lipid metabolic pathways as a promising therapeutic strategy for PC.
