Causal relationships between hundreds of plasma metabolites and PTSD: a bidirectional mendelian randomization study.

Abstract

Background: Recent studies have indicated a connection between plasma metabolites and Post-traumatic stress disorder (PTSD). Nevertheless, the precise causal relationship remains unclear.

Methods: We performed bidirectional Mendelian Randomization (MR) using two metabolite and two PTSD GWAS datasets to examine causal relationships between PTSD and 1009 plasma metabolites. Forward MR tested metabolite causally effects on PTSD, while reverse MR assessed PTSD causally effects on metabolites. Primary analysis employed the IVW method, supported by four supplementary methods. Four IVW results per direction were meta-analyzed to identify high-credibility metabolites. Venn diagrams intersected results from the four IVW analyses, and this intersection was further compared with meta-analysis findings to generate a second Venn diagram. Sensitivity analyses addressed horizontal pleiotropy for robust results.

Results: After sensitivity analyses, a robust set of 775 metabolites in the forward MR analysis and a set of 566 ones in the reverse process were identified. The meta-analysis of IVW method results (four results between two metabolites GWAS and two PTSD GWAS) revealed that 58 metabolites were significantly associated with the risk of PTSD (P < 0.05) in the forward MR analysis, and 19 metabolites might exhibit significant changes in PTSD (P < 0.05) in the reverse progress. Further Venn diagram intersection analysis among those four IVW results unveiled 4 metabolites with promoting or inhibiting effects on PTSD (P < 0.05) and 1 metabolites with notably increased plasma levels in PTSD (P < 0.05). The subsequent Venn diagram intersection analysis of the meta-analysis outcomes and the initial Venn diagram results identified 3 metabolites. In the forward analysis, 5-hydroxy-2-methylpyridine sulfate (OR = 1.05, P = 0.004) and levulinoylcarnitine (OR = 1.08, P = 0.005) from the Xenobiotics pathway were significantly associated with an increased risk of PTSD. Additionally, cysteinylglycine from the Amino Acid pathway significantly reduced the risk of PTSD (OR = 0.918, 95%CI: 0.868-0.971, P = 0.003). In the reverse analysis, no significant changes in plasma metabolites at the genetic level were found to causally influence the development of PTSD.

Conclusions: Our findings provide potential biomarkers for predicting and preventing PTSD, as well as possible therapeutic targets for that. However, further research is needed to confirm the clear mechanism.