Predictors of Liver Fibrosis in HCV+ Hemophilic Men.

Hepatitis C is a major cause of chronic liver disease in hemophilia. Liver biopsy is typically not performed in this population because of potential bleeding risk, and, thus, the proportion with liver fibrosis is not known. A number of laboratory tests, including AST, ALT, ALT/AST ratio, PT, platelet count and AST/platelet ratio (APRI) have been evaluated as surrogate tests for liver fibrosis in other HCV risk groups, but not in individuals with hemophilia. If laboratory tests could be validated as surrogate markers of fibrosis in the hemophilic population, it might be possible to predict the extent of their HCV liver disease. We, therefore, evaluated the above laboratory tests as potential surrogate markers of fibrosis in HCV (+) hemophilic men undergoing liver biopsy and enrolled in an ongoing prospective, cross-sectional multi-center NHLBI-funded study of hepatitis C in hemophilia (HHH Study). Liver biopsies were evaluated by independent, central review for fibrosis using the Ishak, Metavir, and Knodell classification systems. Statistical analyses were performed by Spearman rank correlation, logistic regression, and area under receiver operating characteristic curves (AUROC). Only significant predictors in univariate models were used in the final multivariate model. Of 301 enrolled subjects, 75 (24.9%) had undergone liver biopsy, of whom 53 (17.6%) with slides/blocks/tissue available for pathologic review, were analyzed. Of the 53, 24 (45.3%) had biopsy evidence of fibrosis by Ishak score ≥ 3 (F3–F6). Those with fibrosis (F3–F6) did not differ in proportion HIV (+) or in proportion treated with interferon/ribavirin, from those without fibrosis (F0–F2). As compared with the HIV(–) group (n= 26), the HIV(+) group (n=27) was older, 40 vs. 33 yr, and had higher AST, 64 vs. 40 IU/ml, higher APRI scores, 0.48 vs. 0.21, and lower platelet counts, 180,000/μl vs. 235,000/μl, all p < 0.05. The HIV (+) men also tended to use alcohol more frequently, ≥ twice monthly vs. ≤ once monthly, and tended to use analgesics more frequently, at least weekly vs. monthly or less, both p = 0.03; but did not differ in hemoglobin level, 14.7 gm% vs. 15.4 gm%, or in the proportion treated with interferon/ribavirin, 24% vs. 28%, both p > 0.1. Although all three fibrosis scoring systems, Ishak, Metavir, Knodell showed good correlation with all lab markers studied, the Ishak system showed the best correlation. In the logistic regression model, the best predictor of Ishak fibrosis (F3–F6), was the AST/platelet ratio (APRI), p = 0.008. The APRI score was a better predictor of fibrosis (F3 –F6) in HIV (−) than among HIV (+) men, area under the receiver operating characteristics curve (AUROC), 88.9% vs. 66.1%, respectively. In the HIV (+) group, only the Knodell system showed good correlation with lab markers. In conclusion, APRI score is a potentially useful surrogate marker for fibrosis on liver biopsy in HIV (−) hemophilic men with hepatitis C, but less so in HIV (+) hemophilic men with hepatitis C. The reason for differences in the APRI score as a predictor of fibrosis between HIV(−) and HIV(+) is not known, but may suggest that AST and/or platelet count may be affected by HIV infection, rendering these parameters less specific and/or predictive for HCV liver fibrosis. It will be important to test this hypothesis in a large number of patients.