Systematic review with meta-analysis : Diagnostic accuracy of transient elastography for staging of fibrosis in people with alcoholic liver disease

Background The progression of hepatic fibrosis into cirrhosis is a main prognostic factor for survival in people with alcoholic liver disease. The range of cut-off values characterising the stage of hepatic fibrosis seems to be dependent on the aetiology of the liver disease. Aims To determine the diagnostic accuracy of transient elastography (the index test) for diagnosis of fibrosis in alcoholic liver disease when compared with liver biopsy (the reference standard), using the METAVIR scoring system. To establish the optimal cut-off values for the hepatic fibrosis stages. Methods We followed Cochrane Methodology for diagnostic test accuracy reviews. We identified 14 studies. Among the study participants with alcoholic liver disease, 834 provided numerical data for analysis (August 2014). Only half of the studies were monoaetiology studies. We used the bivariate model and estimated the summary sensitivities and summary specificities. Hence, we calculated the summary likelihood ratios (LRs) to rule in or rule out hepatic fibrosis. We investigated pre-defined sources of heterogeneity. Results Severe fibrosis (F3 or worse): summary (95% CI) sensitivity 0.92(0.89-0.96) and specificity 0.70(0.61-0.79); LR+ 3.1(2.1-4.1), LR- 0.11(95% CI 0.06-0.16). Cirrhosis (F4): summary (95% CI) sensitivity of 0.95(0.87-0.98) and specificity 0.71(0.56-0.82); LR+ 3.3(2.1-5.0); LR- 0.07(0.03-0.19). Conclusions Transient elastography may be used as a diagnostic method to exclude cirrhosis or severe fibrosis when the test is negative. Cut-off values of around 12.5 kPa for cirrhosis may be used in clinical practice, but caution is needed, as the values reported in the review are not yet prospectively validated.