Identificaiton of Non‐Alcoholic Steatohepatitis (NASH) Using Plasma Metabolome in Humans

Abstract

The plasma metabolomic profile of subjects with non‐alcoholic fatty liver disease (NAFLD), hepatic steatosis and steatohepatitis (NASH), was examined to identify specific disease‐related patterns and to identify physiological and biochemical perturbations. Plasma samples were obtained after an overnight fast from histologically confirmed subjects with NASH (N=24) or steatosis (N=11), and compared with healthy age and sex‐matched controls (n=25). Subjects with steatosis and NASH were obese and were insulin resistant, as determined by HOMA. Using GC‐MS and LC‐MS, 345 metabolites were detected in the plasma samples, of which 137 compounds were different amongst the groups (p<0.05, q<0.1). Significant differences between NASH and controls were observed for branched chain amino acids, serine, threonine, glutamine, glutamate and aromatic amino acids. Other changes were related to glutathione, purine and fatty acid metabolism. Plasma concentrations of hexanoylcarnitine, cholate and glycocholate were higher in NAFLD when compared with controls. Random forest analysis and recursive partitioning could separate subjects with NASH from controls with an error rate of ∼13%. A recursive partitioning model using body mass index (BMI) and an unnamed compound had an excellent predictive value for subjects with NASH. We conclude that the plasma metabolome can be used to identify subjects with NASH, generate metabolic pathway related hypotheses, and used to follow response to therapeutic interventions without the need for repeat liver biopsies.