Gut-Liver Axis Metabolites and Sepsis: Insights From Mendelian Randomization.

Background: Sepsis is a life-threatening syndrome characterized by a dysregulated host response to infection. Alterations in gut-liver axis metabolites, particularly bile acids, are commonly observed in sepsis. However, the associations between bile acids and sepsis risk or outcomes remain unclear. This study aimed to investigate the potential associations between genetically predicted levels of gut-liver axis metabolites-primarily bile acids-and sepsis risk and prognosis using bidirectional two-sample Mendelian randomization (MR), multivariable MR, and two-step mediation MR analyses.

Methods: Genetic instruments for circulating bile acids were obtained from genome-wide association studies (GWAS) curated in the OpenGWAS database. Summary-level data for sepsis and 28-day mortality were derived from the UK Biobank. We conducted bidirectional two-sample MR to assess the associations between nine bile acids and both sepsis incidence and short-term prognosis. In addition, two-step mediation MR was performed to evaluate whether the associations between specific bile acids and sepsis risk might be mediated through intermediate traits, such as liver function markers. The statistical significance of mediation effects was further tested using both the Sobel test and bootstrap resampling methods.

Results: Univariable MR analyses suggested that higher genetically predicted levels of taurodeoxycholate acid (TDCA) were associated with a lower risk of sepsis (OR = 0.797, 95%CI: 0.668-0.952, p = 0.012). In contrast, glycocholate acid (GCA) (OR = 1.964, 95%CI: 1.220-3.164, p = 0.005) and taurochenodeoxycholate acid (TCDCA) (OR = 1.998, 95%CI: 1.085-3.678, p = 0.026) were positively associated with an increased 28-day mortality risk among sepsis patients. Results from the two-step mediation MR analysis indicated that alanine aminotransferase (ALT) may act as a mediator in the association between ursodeoxycholate acid (UDCA) and sepsis risk. The statistical significance of this mediation effect was further supported by both the Sobel test and bootstrap resampling analysis, suggesting that UDCA may be associated with a reduced risk of sepsis, at least in part, through its influence on circulating ALT levels.

Conclusions: This MR study provides genetic evidence consistent with potential relationships between specific bile acids and sepsis risk and prognosis. Taurodeoxycholate acid (TDCA) may be associated with a reduced risk of sepsis, whereas glycocholate acid (GCA) and taurochenodeoxycholate acid (TCDCA) might relate to worse outcomes. Moreover, among these liver enzymes, ALT exhibited the most significant mediation effect, suggesting that it may play a crucial role in the process by which UDCA influences the occurrence of sepsis. These findings suggest a possible role of bile acids in the pathophysiology of sepsis and may inform future mechanistic studies or therapeutic considerations.