Abstract
BACKGROUND This study was designed to study the serum metabolites of patients with liver failure. MATERIAL AND METHODS The study included 50 patients with liver failure, 30 patients with chronic hepatitis B treated with an artificial liver, 11 patients with an artificial liver, and 32 healthy controls. Clinical data were recorded, and blood samples were analyzed by gas chromatography-mass spectrometry (GC-MS). The random forest algorithm was used to construct a multidimensional scale map to preliminarily reflect the differences between samples. The data were then analyzed to obtain the correlation of different variables among samples, from which the differential metabolites were screened. RESULTS Thirty-five metabolites were identified by GC-MS. There were significant differences in serum metabolites levels before and after treatment in the liver failure group and in the chronic hepatitis group, healthy control group, and artificial liver group. Different metabolites were screened according to the importance of different variables among samples. Significant differences were found between the liver failure group, the chronic hepatitis group, and the healthy control group. In addition, there were significant differences in the liver group before and after treatment with an artificial liver, including differences in boric acid, 2-(methoxyamino)-propionic acid, glycine, l-methionine, aminopropionic acid, glyceryl monostearate, cholesterol, and other substances. CONCLUSIONS A variety of differences in metabolites were found in each group, some of which revealed possible metabolic pathways leading to differences between groups. Blood metabolomics analysis has great potential in real-time dynamic monitoring of liver failure and evaluation of artificial liver therapy.