The predictive value of FIB-4 versus FibroTest, APRI, FibroIndex and Forns index to noninvasively estimate fibrosis in hepatitis C and nonhepatitis C liver diseases


  • Potential conflict of interest: Nothing to report.

The Predictive Value of FIB-4 Versus FibroTest, APRI, FibroIndex and Forns Index to Noninvasively Estimate Fibrosis in Hepatitis C and Nonhepatitis C Liver Diseases

To the Editor:

Noninvasive indirect biochemical markers of liver fibrosis have gained popularity, but hepatologists are now facing the difficult dilemma of choosing among the 20 fibrosis markers that are now available.1 Stanislas Pol's group2 claim that FIB-4, initially developed in a human immunodeficiency virus (HIV)-hepatitis C virus (HCV) coinfected population, is a simple, inexpensive, and readily available marker which is reliable for predicting fibrosis in HCV monoinfected patients. It combines platelets, aspartate aminotransferase (AST), alanine aminotransferase (ALT), and age as measurement factors.

However, they did not present comparative area under the receiver operating characteristic curves (AUROCs) with other inexpensive, simple, and readily available scores such as APRI (aspartate aminotransferase-to-platelet ratio index), Forns index, FibroIndex, and FibroTest, which have been widely validated in the HCV3–5 and non-HCV population.6–8 They do not provide data for other liver diseases besides HCV and, finally, they do not give data related to the value of FIB-4 for the prediction of significant (Metavir F2-4 or Brunt S2-4) fibrosis, another important issue for the clinician because this is the cut-off above which progression of fibrosis occurs and medical treatment is needed.

We sought to compare, independently from the promoters, the diagnostic accuracy using AUROCs of the 5 biochemical scores in our HCV population (n = 152) and our global series (n = 290) of patients with chronic liver disease including also hepatitis B virus (n = 16), alcoholic (n = 54), nonalcoholic (n = 38), and other (n = 23) diseases. The prevalence of significant fibrosis (F2-4 or S2-4), advanced fibrosis (F3-4 or S3-4), and cirrhosis (F4 or S4), evaluated histologically using the METAVIR (in the HCV, HBV, and other populations) or the BRUNT classification (in alcoholic and nonalcoholic liver disease) was 77%, 23%, and 12% and 77%, 34%, and 21%, respectively, in the HCV series and the global series. This compares to 36%, 17%, and 7% in Pol's HCV series.

Comparing the performance of the 5 tests in our HCV and global series for the diagnosis of significant fibrosis, advanced fibrosis, and cirrhosis (Table 1), FibroTest and FIB-4 have equal excellent diagnostic power, FibroIndex being the least accurate.

Table 1. HCV and Global Series: Comparison of AUROCs Between the 5 Different Biochemical Scores for the Prediction of Significant Fibrosis (F/S2-4), Advanced Fibrosis (F/S3-4), and Cirrhosis (F/S4).
TestsFibrosis stage
HCVAll EtiologiesHCVAll EtiologiesHCVAll Etiologies
  • *

    Statistically different from Fibrotest and from FIB-4.

  • **

    Statistically different from FIB-4.


Table 1 and Fig. 1 show a trend for FibroTest to have better AUROCs, even if they are not statistically significant.

Figure 1.

AUROCs of FibroTest, FIB-4, and FibroIndex in predicting cirrhosis (F/S4) in our HCV population.

In the global series, the Forns index is less accurate than FibroTest and FIB-4 for the diagnosis of cirrhosis and APRI also has less discriminative power than FibroTest and FIB-4 for the diagnosis of significant fibrosis, advanced fibrosis, and cirrhosis.

Thus, the results of our study reinforce the last recommendations9 of the HAS (Haute Autorité de la Santé) in France, which equally ranked the following methods for the diagnosis of liver cirrhosis: liver biopsy, Fibroscan, and FibroTest.

Our data are in accordance with those of Pol's group2 and confirm, in a HCV series where the prevalence of advanced fibrosis is similar, that FIB-4 is an alternative to FibroTest albeit with less expense for 3 clinically pertinent issues: the noninvasive diagnosis of significant fibrosis, advanced fibrosis, and cirrhosis. We also show that FIB-4, which includes simple and common biochemical parameters, is equally valuable and accurate across the 4 most frequent etiologies of chronic liver disease.

Michael Adler*, Béatrice Gulbis†, Christophe Moreno*, Sylvie Evrard*, Gontran Verset*, Philippe Golstein*, Brigitte Frotscher†, Nathalie Nagy‡, Philippe Thiry†, * Departments of Hepato-Gastroenterology, Hôpital Erasme, Brussels, Belgium, † Clinical Biochemistry, Hôpital Erasme, Brussels, Belgium, ‡ Pathology, Hôpital Erasme, Brussels, Belgium.