BackgroundEvidence supports the observational correlations between human blood metabolites and sepsis. However, whether these associations represent a causal relationship is unknown. In this study, we applied two‐sample Mendelian randomization (MR) analyses to examine causality between genetically proxied 486 blood metabolites and sepsis risk.MethodsWe used summary data from genome‐wide association studies (GWAS) on 486 metabolites involving 7824 individuals as exposure and a sepsis GWAS including 11,643 cases and 474,841 controls as the outcome. The inverse‐variance weighted (IVW) was the primary method to estimate the causal relationship between exposure and outcome, with MR‐Egger and weighted median serving as supplements. Sensitivity analyses were implemented with Cochrane's Q test, MR‐Egger intercept, MR‐PRESSO and leave‐one‐out analysis. In addition, we performed replication MR, meta‐analysis, Steiger test, linkage disequilibrium score (LDSC) regression and multivariable MR (MVMR) to thoroughly verify the causation.ResultsWe identified that genetically determined high levels of 1‐oleoylglycerophosphoethanolamine (odds ratio (OR) = .52, 95% confidence interval (CI): .31–.87, p = .0122), alpha‐glutamyltyrosine (OR = .75, 95% CI: .60–.93, p = .0102), heptanoate (7:0) (OR = .51, 95% CI: .33–.81, p = .0041) and saccharin (OR = .84, 95% CI: .74–.94, p = .0036) were causally associated with a lower risk of sepsis. MVMR analysis demonstrated the independent causal effect of these metabolites on sepsis.ConclusionsThese findings indicated that four blood metabolites have a protective impact on sepsis, thus providing novel perspectives into the metabolite‐mediated development mechanism of sepsis by combining genomics and metabolomics.