Background:  Liver biopsy is still recommended to determine liver fibrosis in patients with hepatitis C virus (HCV) infection, but is an at-risk invasive procedure. Several non-invasive tests providing a scoring system after combination of routine laboratory parameters have been proposed to predict liver fibrosis in HCV monoinfected patients. There is little information on the validity of these tests in HIV/HCV-co-infected patients.

The aim of our study was to determine the diagnostic performance of FIB4, APRI, and FORNS tests for the prediction of liver fibrosis in HIV/HCV-co-infected patients.

Methods:  Overall, 200 HIV/HCV-co-infected patients of the ANRS CO3 Aquitaine Cohort, who underwent liver biopsy between 1999 and 2005 and had complete data to validate all the considered tests, were investigated. Liver fibrosis was assessed on liver biopsy using METAVIR scoring system. Kappa’s coefficient was used to measure the concordance between liver fibrosis (gold standard) and non-invasive tests results. Diagnostic performance was estimated for each test by calculating sensitivity, specificity, positive and negative predictive values. The area under the ROC curve and the percentage of patients correctly identified were estimated.

Results:  The concordance between the liver biopsy and FIB4 was moderate (cut-off 1.45, κ = 0.40, CI 0.28 to 0.53), weak but statistically significant (p <0.05) for all the other tests at their respective cut-off values. APRI index showed at cut-off ≤1, nevirapine (NPV) of 93% to predict absence of cirrhosis (F4) and at cut-off >2, positive predictive value of 53% to predict cirrhosis. Similarly, FIB4 index had at cut-off ≤1.45, NPV of 83% to predict absence of severe fibrosis (F3-F4) and at cut-off >3.25, PPV of 71% to predict severe fibrosis. Applied at these cut-offs, APRI and FIB4 classified correctly 65.5 and 56.5 of patients, respectively. When APRI was used at cut-off ≤0.5 and >1.5, and FORNS at cut-off ≤4.2 and >6.9, to predict absence or presence of significant fibrosis (F2-F4), 35.5 and 19% only of patients were correctly diagnosed, respectively.

Conclusions:  In this cohort study, we confirmed that APRI (cut-off ≤1, >2) and FIB4 index, could be applied routinely to predict cirrhosis and severe fibrosis in HIV/HCV-co-infected patients, whereas APRI (cut-off ≤0.5, >1.5) and FORNS index had low diagnostic performances to predict significant fibrosis in HIV/HCV-co-infected patients.