We read with interest the article by Sterling et al.1 The authors derived from the AIDS Pegasys Ribavirin International Coinfection Trial (APRICOT)2 an index (FIB-4) to predict liver fibrosis in patients with HIV and hepatitis C virus (HCV) coinfection. The variables entered in the FIB-4 are as simple as age, aspartate aminotransferase (AST), alanine aminotransferase (ALT), and platelet (PLT) count. Using the following formula, age [yr] × AST [U/L])/((PLT [109/L]) × (ALT [U/L])1/2), the authors show that the FIB-4 index can accurately differentiate nil-moderate fibrosis (Ishak 0-3) from bridging fibrosis and cirrhosis (Ishak 4-6) in HIV/HCV coinfected patients with a cutoff < 1.45 or > 3.25, respectively.
We wondered whether the FIB-4 index could also discriminate nil-moderate fibrosis (Ishak 0-3 or Metavir 0-2) from bridging fibrosis and cirrhosis (Ishak 4-6 or Metavir 3-4) in HCV-mono-infected patients. We explored the accuracy of the FIB-4 index in a series of 919 liver biopsies performed in 848 patients at our center. The FIB-4 index enabled the correct identification of patients with severe fibrosis (F3-F4 in the Metavir classification) with an area under the receiver operating characteristic (AUROC) curve of 0.85 [95% confidence interval (CI) 0.82-0.89]. A FIB-4 index < 1.45 had a negative predictive value of 95.2% to exclude extensive fibrosis (F3-F4) with a sensitivity of 80.4%. A FIB-4 index > 3.25 had a positive predictive value of 81.2% and a specificity of 98.3%. The FIB-4 efficiently identified cirrhosis with an AUROC of 0.91 (95% CI 0.86-0.93). Using these cutoffs, 69.2% of the 919 biopsies with FIB-4 values outside 1.45-3.25 were correctly classified.
The FIB-4 index was concordant (kappa 0.537, P < .01) with the FibroTest (Biopredictive SAS, Paris, France), the most evaluated commercially-available noninvasive marker of liver fibrosis.3 A FIB-4 index < 1.45 was concordant in 92.1% of the cases with the FibroTest (fibrosis score < 0.56) to exclude severe fibrosis (Ishak 4-6 or Metavir F3-F4) in a series of 824 FibroTests performed in 595 patients. The post-hoc review of the files of the patients with a FIB-4/FibroTest discordance concluded, in two thirds of the cases, in a FibroTest failure (low haptoglobin, Gilbert disease, data capture failure, or unexplained). With an index > 3.25, the FIB-4 was concordant with the FibroTest in 70.4% of the cases (fibrosis score > 0.7). The discordances in this subgroup were explained in three quarters of the cases as a FIB-4 failure (very young age, unexplained thrombocytosis).
In conclusion, a FIB-4 index beyond 1.45-3.25 is accurate to differentiate nil-moderate (F0-F2) from severe (F3-F4) liver fibrosis in HCV mono-infected patients. In these ranges, the FIB-4 index could replace advantageously expensive and/or invasive methods to assess liver fibrosis, especially in endemic countries were these techniques are barely available.