OBJECTIVES: The accuracies of biomarkers checked in the hepatic vein (HV) and peripheral vein (PV) were compared in the prediction of advanced fibrosis (AF) of liver. METHODS: Patients with chronic viral hepatitis (n=101) who underwent hepatic venous pressure gradient, liver biopsy, and paired HV-PV samples (6 biomarkers: hyaluronic acid [HA], haptoglobin, matrix metalloproteinase-2 [MMP2], tissue inhibitor of metalloproteinases-1 [TIMP1], procollagen III N-terminal peptide [PIIINP], and apolipoprotein-A1 [Apo-A1]) were enrolled. RESULTS: Differences were displayed between the HV and PV in the predictive logit-models for predicting AF (-3.13+0.017×MMP2-0.019×haptoglobin and -0.270+0.007×HA-0.018×haptoglobin, respectively). In the area under the receiver operating characteristic curves, PIIINP (0.74/0.68, p=0.03), MMP2 (0.72/0.63, p=0.04), HA (0.79/0.76, p=0.94), Apo-A1 (0.56/0.48, p=0.73), and predictive logit-model (0.81/0.78, p=0.68) showed higher diagnostic value in the HV sample. CONCLUSIONS: While most biomarkers were correlated better with hepatic fibrosis in HV than in PV, individually and in predictive logit-models, they were inadequate to determine the degree of advanced fibrosis.
OBJECTIVES: The accuracies of biomarkers checked in the hepatic vein (HV) and peripheral vein (PV) were compared in the prediction of advanced fibrosis (AF) of liver. METHODS:Patients with chronic viral hepatitis (n=101) who underwent hepatic venous pressure gradient, liver biopsy, and paired HV-PV samples (6 biomarkers: hyaluronic acid [HA], haptoglobin, matrix metalloproteinase-2 [MMP2], tissue inhibitor of metalloproteinases-1 [TIMP1], procollagen III N-terminal peptide [PIIINP], and apolipoprotein-A1 [Apo-A1]) were enrolled. RESULTS: Differences were displayed between the HV and PV in the predictive logit-models for predicting AF (-3.13+0.017×MMP2-0.019×haptoglobin and -0.270+0.007×HA-0.018×haptoglobin, respectively). In the area under the receiver operating characteristic curves, PIIINP (0.74/0.68, p=0.03), MMP2 (0.72/0.63, p=0.04), HA (0.79/0.76, p=0.94), Apo-A1 (0.56/0.48, p=0.73), and predictive logit-model (0.81/0.78, p=0.68) showed higher diagnostic value in the HV sample. CONCLUSIONS: While most biomarkers were correlated better with hepatic fibrosis in HV than in PV, individually and in predictive logit-models, they were inadequate to determine the degree of advanced fibrosis.