No abstract

Recent evidence suggests that patients with traumatic brain injuries (TBIs) have a distinct circulating metabolic profile. However, it is unclear if this metabolomic profile corresponds to changes in brain morphology as observed by magnetic resonance imaging (MRI). The aim of this study was to explore how circulating serum metabolites, following TBI, relate to structural MRI (sMRI) findings. Serum samples were collected upon admission to the emergency department from patients suffering from acute TBI and metabolites were measured using mass spectrometry-based metabolomics. Most of these patients sustained a mild TBI. In the same patients, sMRIs were taken and volumetric data were extracted (138 metrics). From a pool of 203 eligible screened patients, 96 met the inclusion criteria for this study. Metabolites were summarized as eight clusters and sMRI data were reduced to 15 independent components (ICs). Partial correlation analysis showed that four metabolite clusters had significant associations with specific ICs, reflecting both the grey and white matter brain injury. Multiple machine learning approaches were then applied in order to investigate if circulating metabolites could distinguish between positive and negative sMRI findings. A logistic regression model was developed, comprised of two metabolic predictors (erythronic acid and myo-inositol), which, together with neurofilament light polypeptide (NF-L), discriminated positive and negative sMRI findings with an area under the curve of the receiver-operating characteristic of 0.85 (specificity = 0.89, sensitivity = 0.65). The results of this study show that metabolomic analysis of blood samples upon admission, either alone or in combination with protein biomarkers, can provide valuable information about the impact of TBI on brain structural changes.

Acute traumatic brain injury (TBI) is associated with substantial metabolic abnormalities, both centrally and in the periphery. We have previously reported extensive changes in the circulating metabolome resulting from TBI, including changes proportional to disease severity and associated with patient outcomes. The observed metabolome changes in TBI likely reflect several pathophysiological mechanisms supporting the concept that TBI is a systemic disease after the primary injury. However, one of the main metabolic changes we have observed following a TBI are changes in lipids, including the structural lipids that are known to be present in the myelin in the brain. Here, we conducted a study to investigate the relationship between traumatic microstructural changes in white matter seen on magnetic resonance imaging (MRI) and quantitative lipidomic changes in the blood in a subset of patients with TBI recruited to the MRI sub-study of the Collaborative European NeuroTrauma Effectiveness Research in TBI (CENTER-TBI) study. In total, there were 103 patients who had both a magnetic resonance imaging (MRI) scan and serum samples available for analysis. From serum, 201 known lipids were quantified. Diffusion tensor fitting generated fractional anisotropy (FA) and mean diffusivity (MD) maps for the MRI scans, in addition to volumetric data. Association matrices and partial correlation networks were built to elucidate the connections between the lipid groups and the maps. We found that there are distinct directions of associations between the neuroimage data (FA and MD sets) and the concentrations of circulating lipids after injury. The FA and MD values were in inverse relationship with the severity of TBI (higher MD values, lower FA). We also observed that the lipid associations to FA and MD show different metabolic signatures. Lysophosphatidylcholines (LPC) associate mostly with FA while sphingomyelins (SM) associate with MD. Only phosphatidylcholines( PC) have strong associations with both as well as with the volumetric data. Finally, we found that the lipid changes are not associated with the number of regions with abnormalities. In conclusion, we have identified groups of lipids which assocate with specific MRI imaging metrics following TBI. There appears to be consistent patterns of lipid changes associating with the specific microstructure changes in the CNS white matter. There is also a pattern of lipids with regional specficity, suggesting that blood-based lipidomics may provide an insight into the underlying disease mechanisms in TBI.