Author information
1Centre de Recherche sur l'Inflammation (CRI), UMR 1149 Inserm, Université Paris Diderot, Clichy, France; Service d'Hépatologie, AP-HP Hôpital Beaujon, Clichy, France. Electronic address: patrick.marcellin@bjn.aphp.fr.
2Centre de Recherche sur l'Inflammation (CRI), UMR 1149 Inserm, Université Paris Diderot, Clichy, France; Service d'Hépatologie, AP-HP Hôpital Beaujon, Clichy, France.
3Roche Diagnostics Ltd., Forrenstrasse, 6343 Rotkreuz, Switzerland.
4IST GmbH, Mannheim, Germany.
5Institute of Translational Hepatology, Beijing 302 Hospital, Beijing, China; Humanity and Health Gastroenterology and Liver Clinic, Hong Kong SAR, China.
6Liver Research Unit, Chang Gung Memorial Hospital, Chang Gung University College of Medicine, Taipei, Taiwan.
Abstract
BACKGROUND & AIMS:
Non-invasive techniques are needed to assess hepatic fibrosis in patients with chronic hepatitis B. We developed a scoring system to determine degree of fibrosis in patients with genotype B or C hepatitis B virus (HBV) infection and positive for the e antigen (HBeAg).
METHODS:
We performed a retrospective study to identify baseline variables associated with severity of fibrosis (METAVIR scores, F0‒F4) in a large Phase 3 clinical trial of predominantly Asian patients (n=710), using multivariate logistic regression analyses. Significant variables were used to construct predictive models using optimal cut-off values. The final model was validated in similar patients from a large Phase 4 clinical trial (n=465).
RESULTS:
We developed 2 prediction scoring systems (PS). PS1 analyzed data on HBV genotype (B vs C), patient age (< vs ≥30 ys), level of hepatitis B surface antigen (HBsAg) (≤ vs >17,500 IU/mL), and level of alanine aminotransferase (≤ vs >3-fold the upper limit of normal). PS2 analyzed data on only age and level of HBsAg. PS1 identified patients with F0-F1 vs F2‒ F4 fibrosis with >87% specificity and a positive predictive value (PPV) >75; it identified patients with F0-2 vs F2‒F4 fibrosis with approximately 95% specificity and a PPV of ∼97%. PS2 identified patients with F0-F1 fibrosis with less accuracy than PS1, but patients with F0-F2 fibrosis with an almost identical level of sensitivity and PPV.
CONCLUSIONS:
We developed a simple scoring system to determine the severity of fibrosis in patients with genotype B or C HBV infection who are HBeAg positive. Our system differentiated patients with no or mild fibrosis (F0-F1) from those with marked or severe (F2-F4) fibrosis with a high PPV. The high level of specificity for the identification of non-severe fibrosis (F0-2) limits the risk of overlooking patients with severe fibrosis (F3-4).