BACKGROUND/AIMS: Chronic hepatitis C (CHC) is characterised by hepatic fibrosis, used as a proxy measure of prognosis. Liver biopsy is a flawed reference standard and serum markers of fibrosis offer an attractive alternative. METHODS: A systematic review was conducted to assess the performance of panels of serum markers of hepatic fibrosis in CHC, incorporating analyses placing markers in a clinical context. RESULTS: 14 studies were included with 10 different panels. Median AUC in validation populations was 0.77 and training populations 0.81. Likelihood ratios (LR) ranged from -LR 0.1 to 0.9, + LR 1.2 to 33.1, diagnostic odds ratios (DOR) were 9.0 (median) with a range of 5 to 27- mostly below values of robust tests. Tests perform with either high sensitivity with low specificity or vice versa. Cut-off levels that gave clinically relevant predictive values for the presence/absence of significant fibrosis were applicable to 35% of the population. CONCLUSIONS: Serum markers can rule-in or rule-out fibrosis in up to 35% of patients, but cannot differentiate stages of fibrosis reliably. Improvement of index and reference test in needed including evaluation of clinical outcomes as reference. Improved test reporting is needed to derive LR and DOR as performance indicators.
BACKGROUND/AIMS: Chronic hepatitis C (CHC) is characterised by hepatic fibrosis, used as a proxy measure of prognosis. Liver biopsy is a flawed reference standard and serum markers of fibrosis offer an attractive alternative. METHODS: A systematic review was conducted to assess the performance of panels of serum markers of hepatic fibrosis in CHC, incorporating analyses placing markers in a clinical context. RESULTS: 14 studies were included with 10 different panels. Median AUC in validation populations was 0.77 and training populations 0.81. Likelihood ratios (LR) ranged from -LR 0.1 to 0.9, + LR 1.2 to 33.1, diagnostic odds ratios (DOR) were 9.0 (median) with a range of 5 to 27- mostly below values of robust tests. Tests perform with either high sensitivity with low specificity or vice versa. Cut-off levels that gave clinically relevant predictive values for the presence/absence of significant fibrosis were applicable to 35% of the population. CONCLUSIONS: Serum markers can rule-in or rule-out fibrosis in up to 35% of patients, but cannot differentiate stages of fibrosis reliably. Improvement of index and reference test in needed including evaluation of clinical outcomes as reference. Improved test reporting is needed to derive LR and DOR as performance indicators.
Authors: Dalia Omran; Ayman Yosry; Samar K Darweesh; Mohammed M Nabeel; Mohammed El-Beshlawey; Sameh Saif; Azza Fared; Mohamed Hassany; Rania A Zayed Journal: Clin Exp Med Date: 2017-05-31 Impact factor: 3.984