Non-invasive assessment of fibrosis in alcoholic liver disease

Mélissa Salavrakos(1), Hubert Piessevaux(1), Mina Komuta(2), Nicolas Lanthier(3), Peter Stärkel(3) Published in the journal : July 2018 Category : Mémoires de Recherche Clinique

Summary :


Transient elastography (TE), or Fibroscan®, has been validated for the diagnosis and staging of liver fibrosis in chronic hepatitis C, but there is no clear consensus on the optimal TE cut-off values in alcoholic liver disease (ALD). The objectives of the study were: (a) to evaluate the cut-off values reported in the literature and determine new ones for the diagnosis of fibrosis in alcoholic patients; (b) to investigate the impact of 2 weeks of abstinence on TE results; (c) to evaluate the diagnostic accuracy of TE for detecting clinically significant portal hypertension; (d) to evaluate the accuracy of non-invasive blood tests (AST-Platelet Ratio Index [APRI], Forns index, Fibrosis-4 [FIB-4]); (e) to study the potential histological (steatosis, alcoholic hepatitis, perisinusoidal fibrosis) and biochemical (transaminases and cholestasis) confounding factors leading to misclassification by TE.



Patients admitted for alcohol withdrawal underwent a first Fibroscan ®; patients whose TE value was suggestive of significant fibrosis (≥F2) were proposed a liver biopsy. A second Fibroscan® was proposed 2 weeks later to a subgroup of patients who had remained abstinent.



A total of 118 patients were included in the study, 57 of whom underwent a second Fibroscan®. Fibroscan® was well correlated with the histological score (ρ=0.680, p<0.01) and showed a very good negative predictive value of 92% for ruling out severe fibrosis (≥F3) and 93% for cirrhosis (F4). Our optimal cut-offs were found at ≥11.7kPa for F2, ≥15.2 for F3, and ≥21.2 for F4. After 2 weeks of abstinence, a mean decrease in TE values of 2.7kPa (+/-0.9) was observed, along with a significantly better agreement between Fibroscan® and histology. Fibroscan® also correlated well with the hepatic venous pressure gradient (HVPG) (ρ=0.753, p<0.01); a TE value of 30.6 kPa for predicting an HVPG >10mmHg yielded a 94% specificity. Fibroscan® performed better than all non-invasive blood tests. We did not find any impact of the confounding factors on misclassification.



Fibroscan® is currently the most accurate non-invasive method for diagnosing fibrosis in ALD patients. We can confidently assert that significant fibrosis can be ruled out for TE values <11.7kPa and esophageal varices for TE values<30.6kPa. We did not find any factor that could explain the Fibroscan®’s trend to overestimate histological fibrosis, but a period of abstinence reduces this effect. Future studies should focus on this point.