Background:  Among 77 consecutive liver biopsies in HIV/HCV co-infected patients referred to Johns Hopkins (JHU) HCV clinic, 58% had cirrhosis. Since these and other existing estimates of liver disease are subject to referral bias, our objective was to estimate the true magnitude of liver disease among HIV/HCV co-infected patients by randomly selecting patients undergoing HIV care.

Methods:  We randomly selected 137 HIV/HCV members of the JHU HIV clinical cohort, of whom 115 were eligible for liver biopsy (n = 112) or had evidence of end-stage liver disease (n = 3). Biopsies were evaluated by a single pathologist and scored according to the Ishak modified histological activity index scoring system from F0 (no fibrosis) to F6 (cirrhosis).  Multivariate logistic regression analysis was used to identify non-histological and histological correlates of fibrosis (>F3).  Factors considered in analysis included demographics, alcohol, antiretroviral therapy (ART), liver enzymes over time, necroinflammatory activity and liver fat.

Results:  The mean age of the 115 patients was 38 (+7) years, 61% were male and 91% African-American. At the time of the biopsy, 59 (51%) individuals were on ART, including 52 on HAART; 43% had no fibrosis (F0), 30% had some evidence of significant fibrosis (>F3), and 17% had cirrhosis (F5 or F6).  The only independent, non-histological predictor of any significant fibrosis was cumulative proportion of aspartate aminotransferase values >5X upper limit of normal (odds ratio [OR] for >10%, 8.7; 95%CI: 1.2 to 10.4). Hepatic necroinflammatory activity (>5) was the only significant histological predictor of liver fibrosis (OR, 5.6; 95%CI:  2.3 to 13.9) so we further examined correlates of activity.  Independent predictors of greater activity included cumulative HAART exposure (OR per year of exposure, 0.8; 95%CI:  0.7 to 0.98), male gender (OR, 3.4; 95%CI: 1.6 to 7.2), and cumulative proportion of CD4 measurements <200 cells/mm³ (OR for <25%, 0.4; 95%CI:  0.2 to 0.9).

Conclusions:  The prevalence of significant liver fibrosis is substantially lower in this randomly selected sample of HIV/HCV co-infected patients compared to previous estimates.  Although very little of the variability in fibrosis was explained by previously observed correlates, these data underscore the importance of prospective study of the interrelationship between use of HAART, liver inflammation, and fibrosis progression.

Keywords: hepatitis C; fibrosis; co-infection