Liver fibrosis and cirrhosis result from an imbalance of fibrogenesis and fibrolysis. Serum levels of matrix components may be useful for assessing hepatic matrix turnover or to predict fibrosis stage,1 which is a problem in children in whom fibrosis markers are influenced by body growth.1, 2
We therefore determined serum levels of tenascin, laminin-2, hyaluronan, collagen IV and VI, procollagen III N propeptide (PIIINP), matrix metalloproteinase-2 (MMP-2), tissue inhibitor of matrix metalloproteinase-1 (TIMP-1), and the MMP-9/TIMP-1 complex after an overnight fast in 47 children (mean age 8 years; range 4–16) with chronic HBe-Ag positive hepatitis B prior to interferon-α therapy. Measurements were performed on the Bayer Immuno 1 Analyzer (Bayer AG, Leverkusen, Germany) using fluorescein and alkaline phosphatase–labeled monoclonal antibodies to the protein antigens.3, 4 Liver biopsies were obtained at the time of serum sampling, and the protocol was approved by the ethics committee of the Medical University of Białystok. Fibrosis and inflammation were assessed by a single-blinded pathologist (M. S. Ł). Staging was according to Ishak and colleagues (early fibrosis: stage 1–3 vs. advanced fibrosis: stage 4–6),5 METAVIR,6 and Batts and Ludwig (stage 1–2 vs. 3–4).7 Receiver operating characteristics (ROC, ACCUROC, Canada) and area under curve analysis were measured to calculate the power of the markers to detect advanced fibrosis.
Seven,5 nine,6 or eight7 children had advanced fibrosis. Using the Batts and Ludwig score, serum hyaluronan above 27 ng/mL had a sensitivity of 100% and a specificity of 50%, and serum laminin-2 above 34.9 ng/mL had a sensitivity of 62.5% and a specificity of 92.1% to predict advanced fibrosis. The combination of both markers was superior. With laminin-2 above 34.9 ng/mL and/or hyaluronan above 27 ng/mL, the area under curve was 0.8429 (P = .003), the sensitivity was 87.5%, and the specificity was 73%. Similar results for single and combined markers were obtained for the Ishak and METAVIR scores (Fig. 1). Using the combined markers and the Batts and Ludwig score, 28 and 6 out of the 45/47 children (for two children only one of the two parameters was determined) could be allocated with 100% confidence either to the group with mild or severe fibrosis, respectively, potentially avoiding biopsy in 34 (75.6%) of children. All other serum markers and their combinations yielded a lower predictive power. None of the markers was a good predictor of histologic inflammation.
Liver biopsy remains the gold standard for the grading and staging of liver disease. However, this risky procedure is dispensible for children who have only minor fibrosis. Notably, among a broad panel of serum fibrosis markers, only hyaluronan and laminin-2 accurately predicted the presence of advanced fibrosis. Other serum fibrosis markers were not deemed useful in children because they are influenced by body growth (PIIINP and procollagen IV) or bone metabolism (procollagen I).1, 2 The predictive power of hyaluronan and laminin-2 may be due to their short biologic half life (hyaluronan)8 or their selective expression in fibrosing sinusoidal basement membranes (laminin-2),9 which precludes a major contribution by (growing) bones or ubiquitous interstitial tissues. The usefulness of these parameters for children with other chronic liver diseases (e.g., chronic hepatitis C, biliary atresia) has yet to be demonstrated.