Metabolic Profiling: Shedding New Light On The Pathophysiological Process And Prognosis Of Hbv-Related Acute-On-Chronic Liver Failure

Liujuan Ji,Xia Liu,Yu Liu,Hui Zhu,Yuyi Zhang,Zhengguo Zhang,Hongying Guo,Wei Yuan,Zhiping Qian,Naixia Zhang,Jiefei Wang

2017-01-01

Abstract:Severe global metabolic deterioration is a key feature of hepatitis B virus-related acute-on-chronic liver failure (HBV-ACLF). To investigate the global metabolic changes of HBV-ACLF patients and explore the prognostic potential, extracted metabolic profiling information will shed new light on the pathophysiological process and prognosis of this disease. An NMR-based metabolomics study and its coupled survival prediction ability analysis were applied to the serum samples of healthy controls (HC, n=23), chronic hepatitis B patients (CHB, n=27), surviving HBV-ACLF patients (ACLF-S, n=24) and non-surviving HBV-ACLF patients (ACLF-NS, n=30). The score plots from the metabolomics analysis reveal clear metabolic separations of the HBV-ACLF patients from the CHB individuals and of the surviving HBV-ACLF individuals from the non-surviving ones. Ultimately, 14 discriminative metabolites, including lipids, LDL and twelve amino acids, presented consecutive changes with the progression of HBV-ACLF. These metabolites potentially play important roles in the progression of HBV-ACLF. The averaged sensitivity and specificity values of the survival prediction models based on metabolomics data are 76.7% and 83.3%, respectively. The combined use of metabolomics data and clinical parameters raises the sensitivity and specificity values of the predicted score plot up to 91.7% and 90%, respectively. The phenotypic metabolomics analyses of HC, CHB, ACLF, ACLF-S and ACLF-NS individuals provides an encouraging basis for deeper studies of the HBV-ACLF pathophysiological process and demonstrates a proof of concept that global metabolic profiling data is a reliable source to improve the accuracy of predicting the survival outcome of HBV-ACLF patients.