Original article
Biomarkers of liver injury for hepatitis clinical trials: a meta-analysis of longitudinal studies
Thierry Poynard, Yen Ngo, Mona Munteanu, Dominique Thabut, Julien Massard, Joseph Moussalli, Anne Varaud, Yves Benhamou, Vlad Ratziu
Corresponding author name: Thierry Poynard
Corresponding author e-mail: tpoynard@teaser.fr
Citation: Antiviral Therapy 2010; 15:617-631
doi: 10.3851/IMP1570
Date published online: 25 June 2010
Abstract
Background: Liver biopsy and virological end points are standard references for assessing the effect of viral hepatitis treatments. We aimed to review evidence-based published data of biomarkers that have been validated as non-invasive alternatives to biopsy as end points for HBV and HCV infection trials.
Methods: Studies were included if there were at least two repeated estimates of fibrosis per patient using biomarkers with at least two studies and a control group. Meta-analysis of the percentage of fibrosis progression per year (pFPy) was performed.
Results: Two biomarkers were included, FibroTest® and liver stiffness measurement (LSM; FibroScan®). A total of 1,413 patients with chronic hepatitis C (11 populations) and 772 with chronic hepatitis B (6 populations) were analysed. In a comparison of HCV patients with controls, the FibroTest® pFPy was -18% (95% confidence interval [CI] -23–-14; P<0.001) in treated patients and the LSM pFPy was -15% (95% CI -28–-1; P=0.01), both with differences according to virological responses. In HBV patients, there was a significant decrease of the pFPy only in patients with baseline advanced fibrosis (mean difference -5% [95% CI -10–-0.1]; P=0.02). In patients with advanced fibrosis, stratified by virological response, there were similar differences between pFPy estimated either using FibroTest® or biopsy, both in HCV and HBV infections. Repeated LSM in HBV patients had an early variability related to necroinflammatory activity.
Conclusions: In patients with chronic hepatitis C and B, the treatment effect on fibrosis progression rate was similarly estimated using FibroTest® or biopsy. The same concordance was observed for FibroScan® but with a possible overestimation of the fibrosis regression during the first weeks of treatment.