We thank Drs. Yilmaz and Kurt for their comment on our article by Joka et al.1 In that study we evaluated different CK-18 cell death biomarkers for their diagnostic value to differentiate between various fibrosis stages in sera of patients with chronic liver disease (n = 121). We applied the internationally accepted Ishak score for staging of fibrosis.2 Compared to the Metavir score, the Ishak score allows for a more differentiated histopathological assessment, which enables its application not only for chronic viral hepatitis but also for other disease entities. It was not our intention to selectively evaluate the biomarkers in hepatitis C virus (HCV) patients but rather in a broad range of chronic inflammatory liver diseases. As the definition of disease stages, however, does not differ substantially between scores, we would not expect any relevant differences by applying alternative fibrosis scoring systems. Dr. Yilmaz and colleagues found higher levels of M30 in nonalcoholic fatty liver disease (NAFLD) patients with minimal fibrosis compared to those with Wilson's disease and severe fibrosis. In our NAFLD cohort, patients showed no or only minimal fibrosis, which did not influence the diagnostic value of the M30 or M65 assay for prediction of nonalcoholic steatohepatitis (NASH). However, we found that, in contrast to the M65 assays, detection of NASH by the M30 enzyme-linked immunosorbent assay (ELISA) depends on alanine aminotransferase (ALT) levels. Thus, inflammatory disease activity might contribute to higher M30 levels in NAFLD patients despite lower fibrosis stages. In this context, it is also worth mentioning that differences of M30 antigen stability between sera and plasma might contribute to discrepant observations, since concerns about M30 antigen stability in plasma samples during storing have been reported.3 Using serum samples, we did not observe stability issues of the M30 antigen during storing. We certainly agree that NAFLD patients represent a relatively small subgroup in our analysis. However, despite the limited number of NAFL (nonalcoholic fatty liver; n = 10) and NASH (n = 12) patients included, we could demonstrate that, unlike the M30 ELISA, the M65 assays are able to significantly (P < 0.01) distinguish between healthy individuals and patients with NAFL or NASH. This observation is in line with a previous study of NAFLD patients that failed to distinguish between patients with simple steatosis and healthy controls by the M30 assay but demonstrated a higher predictive value of the M65 compared to the M30 assay to identify NASH patients.4 Furthermore, we disagree that the study was substantially underpowered. Boxplot analysis demonstrated a wide overlap for the M30 assay with regard to NAFL and NASH or healthy controls. To be acceptable as a marker for clinical decision making, the discriminatory power of an assay must yield significant results if subgroups of more than 10 patients are studied. In our opinion, one major factor contributing to the lower diagnostic accuracy of the M30 ELISA might be the lower values obtained by the M30 compared to the M65 assays, especially in patients with mild liver diseases. Heike Bantel M.D.*, Matthias J. Bahr M.D. , Jerome Schlue M.D. , Klaus Schulze-Osthoff Ph.D.§, * Department of Gastroenterology, Hepatology and Endocrinology, Hannover Medical School, Hannover, Germany, Sana Kliniken Luebeck, Luebeck, Germany, Institute of Pathology, Hannover Medical School, Hannover, Germany, § Interfaculty Institute for Biochemistry, University of Tuebingen, Tuebingen, Germany.
Read full abstract
Jul 1, 2018
+ 3
Open Access