Analytical workflow for comprehensive blood metabolomics analysis by GC-MS. Application to children with ventilator associated pneumonia

Metabolomics is a widely used approach for analyzing a vast array of low molecular weight compounds such as amino acids, organic acids, vitamins, biogenic amines and carbohydrates in biological samples, with the aim of investigating biomarkers in personalized medicine studies. Advancements in gas chromatography- mass spectrometry (GC–MS) instrumentation, along with the availability of commercial and public spectral libraries, have highlighted the relevance of GC–MS analysis as a valuable tool for metabolomics applications. Stability assessment in derivatisation and GC–MS analysis is a crucial yet often overlooked aspect of metabolomics studies. In this study, an untargeted GC-MS method workflow for large scale metabolomics studies is presented after assessment and optimization of whole blood sample's stability. The method consists of a common two-step derivatisation procedure including methoximation using methoxyamine hydrochloride, followed by silylation with N-methyl-N-(trimethylsilyl) trifluoroacetamide (MSTFA). To ensure the stability of the studied metabolites, extensive stability experiments were performed. The stability of the derivatives was evaluated over 24 h in the autosampler at room temperature, as well as after storage for 24 and 48 h at -20 °C for both derivatized and dried extracts. While derivatised samples remained stable for 24–48 h in the freezer, dried extracts exhibited variability after 48 h. Findings support the storage of derivatised samples over dried extracts, ensuring greater stability. To increase condidence in metabolite identification data from the analysis of 120 standard compounds were utilized. The developed method was applied to analyze blood samples from 32 children with ventilator-associated pneumonia (VAP), collected at four different time points during ICU hospitalization. This analysis led to the identification of 43 metabolites. The results of multivariate and univariate statistical analyses demonstrated several statistically significant metabolites, including aspartic acid, alanine, and pyroglutamic acid, which showed a strong correlation with the disease's manifestation and may potentially serve as biomarkers in the diagnosis of ventilator-associated pneumonia VAP at the stage of clinical suspicion.