Untargeted Plasma Metabolomics Reveals Extensive Metabolic Alterations Among Treatment-Naive Human Immunodeficiency Virus/Hepatitis C Virus Co-Infected Patients with Liver Disease Progression

Both human immunodeficiency virus (HIV) and hepatitis C virus (HCV) may induce metabolic disorders and cause liver complications. Therefore, we aim to analyze the metabolite differences among treatment-naive HIV/HCV co-infected patients with versus without liver disease progression (LDP) and HIV mono-infected patients. A cross-sectional study was conducted in 65 HIV/HCV co-infected patients (22 with LDP and 43 without) and 65 HIV mono-infected patients in Dehong prefecture of Yunnan province, China. Plasma metabolomics were measured by gas chromatography-mass spectrometry (MS) and liquid chromatography-MS. Discrimination analysis, pathway enrichment analysis, generalized linear model with binomial distribution, and area under the receiver-operating characteristic curve (AUC) were conducted to identify bilateral differences in metabolites and pathways in different comparison groups. A total of 10,831 with 673 named and 10,158 unnamed metabolites were detected. Compared with HIV/HCV co-infected patients without LDP, phenylalanine, tyrosine, and tryptophan biosynthesis pathway with the increased level of tyrosine were significantly altered among HIV/HCV co-infected patients with LDP. Compared with HIV mono-infected patients, the decreased level of glutamine and increased levels of glutamic acid, arachidonic acid, and its derivatives were identified among HIV/HCV co-infected patients. Metabolite panels adjusted for baseline information had a higher accuracy than baseline model (without metabolite information) in distinguishing HIV/HCV co-infected patients with versus without LDP (AUC 0.951 vs. 0.849, p = .027) and HIV/HCV co-infected patients from HIV mono-infected patients (AUC 0.889 vs. 0.766, p < .001). A novel set of metabolites were found to discriminate HIV/HCV co-infected patients with versus without LDP, and from HIV mono-infected patients, which may have mechanistic and interventional implications.