HIV Articles  
Back 
 
 
Rapid fibrosis progression among HIV/hepatitis C virus-co-infected adults
 
 
  AIDS:Volume 21(16)October 2007p 2209-2216
 
Sulkowski, Mark Sa; Mehta, Shruti Hb; Torbenson, Michael Sb; Higgins, Yvonnea; Brinkley, Sherilyn Ca; de Oca, Ruben Montesb; Moore, Richard Da,b; Afdhal, Nezam Hc; Thomas, David La,b From the aJohns Hopkins University School of Medicine, Baltimore, Maryland, USA bJohns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, USA cBeth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts, USA.
 
Abstract
Objectives: To define the incidence of fibrosis progression among hepatitis C virus (HCV)/HIV-co-infected adults, to assess whether HCV or HIV treatment alters the risk of progression, and to determine the utility of liver biopsy to predict future disease.
 
Design: This prospective cohort evaluated 184 HIV/HCV-co-infected individuals who had at least two liver biopsies (median interval 2.9 years).
 
Methods: Biopsies were scored according to the Ishak modified histological activity index scoring system by a single pathologist blind to biopsy sequence. Significant fibrosis progression was defined as an increase of at least two Ishak fibrosis units between the first and second liver biopsy. Logistic regression analysis was used to assess determinants of fibrosis progression.
 
Results: A total of 174 non-cirrhotic patients were eligible; the majority were African-American men undergoing HIV treatment. On initial biopsy, no or minimal fibrosis was identified in 136 patients (77%). Significant fibrosis progression occurred in 41patients (24%). Measures of HIV disease and its treatment before and after initial biopsy were not significantly different in progressors and non-progressors. Fibrosis progression was not associated with HCV treatment, which was received by 37 patients (21%) but only three sustained HCV-RNA suppression. In adjusted analysis, only an elevated serum aspartate aminotransferase level between biopsies was associated with progression (odd ratio 3.4, 95% confidence interval 1.4-7.9).
 
Conclusion: Over a 3-year interval, significant fibrosis progression can occur in co-infected individuals even if minimal disease was detected on initial biopsy. In this context, factors other than treatment for HIV or HCV modify the risk of fibrosis progression.
 
Introduction TOP
 
As a result of shared transmission modes, hepatitis C virus (HCV) infects 15-30% of HIV-infected individuals [1]. HIV co-infection has been associated with more rapid progression of liver fibrosis, leading to cirrhosis, end-stage liver disease and hepatocellular carcinoma in some patients [2,3]. In the context of reduced HIV-related mortality, HCV-related liver disease is a leading cause of death among HIV-infected individuals.
 
Conflicting data have emerged regarding the effect of antiretroviral therapy (ART) on HCV disease. In some, but not all, retrospective studies, the prevalence of cirrhosis and mortality was lower in patients treated with ART compared with those who were untreated [4,5]. In other studies, ART has been associated with acute and chronic liver injury, including inflammation and steatosis [6,7]. Similarly, limited data exist regarding the effect of HCV treatment on the progression of hepatitis C. Effective treatment, leading to sustained virological response (SVR), has been associated with disease improvement in co-infected patients. The impact of HCV treatment in individuals who do not achieve SVR is, however, controversial. Recent reports suggest that HCV treatment may be associated with attenuated disease progression even among individuals with virological failure [8]. In light of the relatively low SVR rate observed in co-infected patients, data in this regard are particularly relevant [9].
 
Despite this uncertainty, expert guidelines regarding HCV treatment are based on assumptions regarding the natural history of HCV disease and the impact of HIV and HCV treatment. For example, some experts recommend the deferral of HCV treatment in individuals with mild liver disease based on the assumption of HCV disease stability and a beneficial effect of ART, whereas other experts recommend HCV treatment independent of liver disease stage based on the assumption of rapidly progressive HCV disease and a beneficial effect of HCV treatment [10-13]. On the basis of these conflicting recommendations, the role of liver biopsy in the management of co-infected patients is debated.
 
The objectives of this prospective study were: (i) to define the incidence of fibrosis progression among co-infected adults receiving HIV care; (ii) to test the hypothesis that HCV and HIV treatment alters the risk of liver disease progression; and (iii) to determine the utility of liver biopsy as a diagnostic test to predict future liver disease.
 
Patients and methods
Study population and sample selection

The study population derives from co-infected members of the Johns Hopkins University HIV clinic cohort, located in an urban setting (Baltimore, Maryland, USA). We prospectively evaluated 184 individuals who had at least two liver biopsies performed between January 1998 and July 2006. Of these individuals, 10 were excluded because of the finding of cirrhosis on the initial biopsy.
 
For all subjects, information on prescribed medications and laboratory parameters was obtained from the clinical and laboratory database. As described previously, data on patient demographics, social practices, clinical and laboratory parameters, and prescribed antiretroviral and other medications are abstracted from charts by trained personnel and transferred electronically from the laboratory database at enrollment and subsequent 6-month intervals [14]. The designation of injection drug use and alcohol abuse was based on physician diagnosis and self-reports via audio computer-assisted self interview. Validity checks are performed periodically on a 5% sample of abstracted fields in the database; to date, errors have been found in 0.2% of variable fields.
 
Laboratory evaluations
Patients had standard laboratory assessments performed by licensed clinical laboratories, including a complete blood cell count, serum chemistry panels, alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels, CD4 cell count, and plasma HIV-RNA level (reverse-transcriptase polymerase chain reaction). HCV testing was performed using third-generation enzyme immunoassay (Ortho Diagnostics, Raritan, New Jersey, USA). HCV RNA and genotype testing was performed using reverse-transcriptase polymerase chain reaction. A transcutaneous liver biopsy was performed using an 18-gauge needle. Liver tissue was then fixed in 10% formalin, and paraffin-embedded sections were stained with hematoxylin-eosin and trichrome stains. Slides, deemed adequate based on specimen size and number of portal tracts, were evaluated by a single pathologist (M.T.), who was blinded to the biopsy sequence. Tissues were scored according to the Ishak modified histological activity index scoring system [15]. The median biopsy length was 12 mm [interquartile range (IQR) 9-14 mm].
 
Statistical analysis
Significant fibrosis progression was defined as an increase of at least two Ishak fibrosis units between biopsies because this change was judged to exceed sampling error and to be clinically significant. In addition, an observed fibrosis progression rate was calculated by subtracting the fibrosis score of the first biopsy from the second and dividing by the years between biopsies. Univariate and multiple logistic regression analysis were used to assess determinants of fibrosis progression. For multivariate models, variables that were associated with fibrosis progression in the univariate analysis with a P value less than 0.15 were considered in multivariate models after assessment of multicollinearity by variance inflation factors and tolerance. Further variables that had been previously identified as predictors of progression were forced into models regardless of statistical significance. Variables were entered into multivariate models in a backward stepwise fashion. Those that were significant in the multivariate models with a P value less than 0.05 were retained in the final model.
 
Factors considered in analysis included demographics, alcohol abuse, duration of HCV infection (estimated as time since first injection drug use), past and current ART, serum ALT and AST level, HIV-RNA level, CD4 cell count, histological necroinflammatory activity, and liver steatosis. Furthermore, to capture the longitudinal exposure, we analysed ALT and AST levels, HIV-RNA levels and CD4 cell count both at the time of biopsy and also as cumulative variables of the longitudinal exposure data. ALT and AST were examined as the cumulative proportion of ALT and AST levels more than 2.5 times the upper limit of normal reference range (AST 37 U/l; ALT 40 U/l).
 
For analyses of the predictive accuracy of initial liver biopsy, the AST and aspartate aminotransferase-to-platelet-ratio index (APRI) at initial biopsy and results of the second biopsy, outcomes were defined as: (i) fibrosis progression (≥ 2 Ishak stages between biopsies); (ii) fibrosis exceeding the HCV treatment threshold (Ishak stage ≥ 2 at second biopsy); (iii) advanced fibrosis (Ishak fibrosis stage ≥ 3 at second biopsy).
 
Analyses were performed using SAS version 9.1 software (SAS Institute, Cary, North Carolina, USA).
 
Approval
The study was approved by the Johns Hopkins Medicine Institutional Review Board and written informed consent was obtained for all participants.
 
Results
 
Clinical and demographic characteristics at initial biopsy
Of 184 HCV/HIV-co-infected patients who underwent two liver biopsies, 10 were ineligible because of cirrhosis at first biopsy. The demographic and clinical characteristics of subjects at the time of the first liver biopsy are displayed in Table 1. The majority of patients were African-American men with a median age of 44 years. Past or active injection drug use (78%) and alcohol abuse (38%) were frequently reported. Most patients were prescribed potent ART (62.1%) and had an HIV-RNA level below the limit of detection (60%). Before the first biopsy, the median duration of ART was 3.6 years (IQR 0.9-6.2 years); only 17% had never received ART. The median CD4 cell count was 379 cells/μl (IQR 248-538 cells/μl) and 17% had CD4 cell counts less than 200 cells/μl. The median ALT and AST were 44 U/l (IQR 31-82 U/l) and 46 U/l (IQR 32-69 U/l), respectively. HCV genotype 1 was identified in the majority of individuals tested (147 subjects, 95%) whereas HCV genotypes 2 and 3 were found in only 2% (two subjects each). The median HCV-RNA level was 573 000 IU/ml (IQR 220 000-1 110 000).
 
On initial biopsy, no (fibrosis stage 0, F0) or minimal fibrosis (F1) was identified in 136 patients (77%), whereas 17 had F2 fibrosis and 21 patients had bridging fibrosis (F3 or F4) (Table 2). The median hepatic inflammatory activity index was 3 (IQR 2-5). Hepatic steatosis was observed in 38.5% of patients; of whom the majority had steatosis involving less than 5% of hepatocytes; active steatohepatitis was rarely observed (< 1%). No previously undiagnosed liver disorders were identified by biopsy.
 
Change in hepatic fibrosis between first and second biopsy
The median time between biopsies was 2.9 years (IQR 2.3-3.4 years). When considered as a function of the time between biopsies, the median observed fibrosis progression rate for all individuals was zero Ishak modified fibrosis units/year. There were, however, marked differences among subjects: range 0-1.24 units/year. The Ishak fibrosis stages at first and second biopsies are summarized in Table 2. On second biopsy, most co-infected patients experienced no change (84 patients, 48%) or insignificant progression, defined as an increase in one Ishak stage (37 patients, 22%) in fibrosis stage. Significant fibrosis progression (≥ 2 Ishak unit increase) was, however, observed in 41 patients (24%), of whom 15 patients (8.6%) had an increase of three Ishak units or greater. Conversely, a decrease in Ishak fibrosis stage two units or more was observed in only one patient (0.6%). Among the 136 individuals (78%) with no or minimal fibrosis on initial biopsy, 93 patients (68%) had similar histological findings on second biopsy and 43 patients (32%) had Ishak stage two or greater fibrosis on second biopsy, including 19 patients (14%) with evidence of bridging fibrosis (F3 or F4) or cirrhosis (F5 or F6). Among the progressors, the median fibrosis progression rate (FPR) was 0.83 Ishak fibrosis units/year (IQR 0.58-1.24), whereas for non-progressors the median FPR was zero Ishak fibrosis units/year (IQR 0-0.22).
 
Correlates of progression
Overall, 41 patients (24%) had significant fibrosis progression of two Ishak units or greater between biopsies and are referred to as 'progressors'. Measures of HIV disease (CD4 cell count, HIV-RNA level and ART exposure) before and after initial biopsy were not significantly different in progressors and non-progressors (Table 3). Before the initial biopsy, the median exposure to ART was 3.2 years (IQR 0.8-5.4 years) and 3.8 (IQR 1.1-6.4 years) among progressors and non-progressors, respectively. Between biopsies, potent ART, which included protease inhibitors or non-nucleoside reverse transcriptase inhibitors was used by 51% (median exposure 0.6 years, IQR 0.0-1.9 years) and 65% (median exposure 1.3 years, IQR 0.0-2.5 years) of progressors and non-progressors, respectively (P = 0.12 for ART use; P = 0.22 for ART exposure). We also examined the duration of exposure to individual antiretroviral drugs between biopsies, but did not detect as significant association with significant fibrosis progression. Similarly, no significant difference was observed in HIV suppression or in the median CD4 cell count nadir between biopsies among progressors and non-progressors (P > 0.05).
 
HCV treatment with interferon (standard or pegylated) plus ribavirin was prescribed to 37 patients (21.3%) between biopsies. Overall, incident fibrosis progression was observed in 13 of 37 treated patients (35.1%) and 28 of 137 untreated patients (20.4%; P = 0.06). Compared with untreated patients, however, those who were treated were more likely to have significant fibrosis (≥ F2) on first biopsy (treated patients, 40.5%; untreated patients, 17%; P = 0.01) and the median individual serum ALT and AST levels between biopsies were significantly higher in treated patients (ALT 63 U/l, IQR 40-89 U/l; AST 64 U/l, IQR 48-101 U/l) compared with those who were untreated (ALT 40 U/l, IQR 29-67 U/l; AST 40 U/l, IQR 31-65 U/l; P = 0.001). None of the 13 treated progressors achieved SVR, whereas SVR was observed in three out of 24 treated non-progressors (12.1%; P = 0.21).
 
Measures of hepatic inflammation (e.g. serum AST and ALT) before and after the first biopsy were slightly greater in progressors compared with non-progressors (Table 3). Between biopsies, individuals with fibrosis progression had a significantly greater proportion of AST values higher than 2.5 times the upper limit of normal (37 U/l) compared with those without progression (P = 0.001). In addition, the median individual AST level between biopsies was significantly higher among patients with fibrosis progression compared with those without fibrosis progression (P = 0.001; Fig. 1). In addition, compared with non-progressors, progressors were more likely to have a modified histological activity index greater than 5 at the first biopsy (progressors, 26.8% > non-progressors, 12.9%; P = 0.03) but no difference was observed with respect to the finding of steatosis at first biopsy (progressors, 41%; non-progressors, 38%, P = 0.87).
 
After multivariate adjustment, only the serum AST level between biopsies was significantly associated with fibrosis progression (odds ratio 3.4, 95% confidence intervals 1.4-7.9). Other covariates including age, sex, race, HCV-RNA level, HCV genotype, ART, HCV therapy, CD4 cell count and HIV-RNA level were not significantly associated with fibrosis progression. Analysis of correlates of the observed fibrosis progression rate was not substantially different.
 
Progression algorithms
As the primary reason liver biopsy is performed is to determine whether the risk of future fibrosis progression is sufficient to merit treatment, we examined whether the initial biopsy findings or laboratory results predicted progression (Table 4). For the detection of individuals with subsequent fibrosis progression, the positive predictive value of fibrosis stage (> 1) at first biopsy, elevated AST level (> 40 U/l) and elevated APRI level (> 1.5) were relatively low (24-38%), whereas the negative predictive values of these tests were higher, ranging from 77% for initial biopsy stage to 81% for AST level and APRI score. We also examined the predictive value of these indices for the related outcome of having sufficient fibrosis on second biopsy to warrant treatment (≥ F2 on second biopsy) and nearly identical results were obtained. For each analysis, the platelet count was added the serum AST level as the APRI did not significantly alter the performance of the AST level alone.
 
Discussion
In the era of effective ART, HCV treatment recommendations are guided by conflicting assumptions about HCV disease progression. Some experts recommend HCV treatment for all eligible individuals based on the assumption that liver disease will be rapidly progressive and HCV treatment beneficial; whereas others experts recommend deferral of HCV treatment in individuals with mild disease, based on the assumption that liver disease will be stable and ART beneficial [10-12,16]. To test these hypotheses, we prospectively assessed the incidence of fibrosis progression in 174 co-infected patients who underwent two liver biopsies approximately 3 years apart. We observed significant fibrosis progression in 24% of patients, leading to bridging fibrosis or cirrhosis in 15%. The finding of progressive liver disease in co-infected patients, including those with mild disease on initial biopsy, has several important implications for the management of such patients.
 
First, the observed incidence of fibrosis progression over a relatively short time interval is substantially greater than has been reported in cohorts of HCV-mono-infected individuals. Ghany and coworkers reported that 12% of 123 HCV-mono-infected patients experienced an increase in Ishak fibrosis stage two or greater over a median interval of 44 months [17]. Similarly, Ryder et al. [18] observed fibrosis progression of two Ishak fibrosis stages or more in 8% of 219 HCV-mono-infected patients assessed at a median interval of 30 months. Although more rapid progression of liver disease has previously been detected in co-infected individuals, it was unexpected in this cohort of patients with minimal liver disease on initial biopsy.
 
Second, there are conflicting data regarding the effect of ART on HCV disease. In several retrospective studies, ART exposure, particularly protease inhibitors, has been with the absence of advanced fibrosis in co-infected individuals who underwent a single liver biopsy [19,20]. For example, compared with individuals not taking effective ART, Brau and colleagues [5] reported that the estimated FPR was slower among individuals with high CD4 cell counts and undetectable HIV-RNA levels at the time of biopsy. In contrast, we failed to detect a significant association between prospectively assessed fibrosis progression and specific drugs, HIV-RNA level and CD4 cell count. Equally important, we did not detect evidence that long-term HIV treatment was associated with an increased risk of histological liver injury (e.g. steatosis, inflammation or fibrosis. Therefore, in this cohort of relatively immunocompetent patients, our data suggest that the effect of ART on hepatic fibrogenesis is likely to be modest over a 3-year period. Observation over a longer duration in a larger number of patients will be needed to detect smaller and longer-term effects.
 
Third, in this cohort of predominately genotype 1-infected, African-American patients, we did not detect an effect of HCV treatment on fibrosis progression. HCV treatment has been associated with significant improvements in liver disease in HIV-infected and uninfected individuals who achieve SVR [21-24]. We observed no fibrosis progression in individuals who achieved SVR; however, only 37 patients received HCV treatment, of whom just three achieved SVR. Unfortunately, the low rate of HCV treatment observed in our cohort is similar to that reported in many HIV care settings in which few patients receive therapy and even fewer achieve SVR [9,25-29]. This limited effectiveness underscores the need to understand the impact of HCV treatment in individuals who do not respond. Studies have reported improved liver histology and slower FPR among treated non-responders compared with those not treated [8,21-23,26]. Among HIV-infected, non-responders, Rodriguez-Torres and coworkers [8] reported that HCV therapy was associated with stable or improved hepatic fibrosis. In contrast, compared with those not treated, we failed to detect a beneficial effect of HCV treatment on fibrosis progression in the absence of SVR. The non-random selection of patients to receive HCV treatment may, however, introduce bias and randomized controlled trials are needed. While such trials are underway, our data highlight the need for more effective HCV treatment.
 
Finally, we assessed the utility of pretreatment liver biopsy and other putative progression markers in the management of co-infected patients. First, no causes of liver disease other than hepatitis C were detected, suggesting that histological evaluation is not an effective way to screen for alternative causes of disease [30,31]. Furthermore, biopsy was not a sensitive method of identifying individuals at risk of subsequent fibrosis progression; in fact, the finding of an elevated serum AST level at initial biopsy was approximately twofold more sensitive, although substantially less specific than biopsy. Furthermore, for the identification of non-progressors, the positive predictive value of a normal serum AST level (81%) was similar to the finding of no or minimal fibrosis (77%). These findings are similar to another study conducted among community-based Baltimore injection drug users [32]. Similar to previous studies, we found that the level of AST, which is found predominantly in hepatic mitochondria (80%), was more predictive of liver disease outcome than ALT, which is found only in cytoplasm [33,34]. Although we did not assess the potential mechanism, AST elevations may reflect mitochondrial injury or ongoing alcohol use, both of which may lead to fibrogenesis [35-37]. Taken together, these data suggest that serum AST level, an inexpensive, non-invasive and dynamic measure of liver inflammation, may be useful to assess the risk of disease progression. Nonetheless, better markers of liver disease progression are clearly needed [32,38-40].
 
Several characteristics of the study may impact our findings and their generalizability to other settings. First, many subjects were enrolled after referral for medical evaluation and may differ from co-infected patients who were not referred for care. For example, individuals not referred may be more likely to use alcohol and less likely to use ART. Data on patients referred for care are, however, most germane to clinicians managing these patients. Second, factors that impact liver disease progression may be unknown or incompletely assessed in our study. For example, alcohol use may be underreported despite the use of audio computer-assisted self interview, reflecting the reluctance of participants to report such exposures. Third, for the majority of subjects, liver biopsy was not obtained before the initiation of ART, and relatively few patients with advanced AIDS were enrolled. It is possible that we failed to detect a significant effect of ART because many subjects had already been treated before entry. We accept this limitation in light of the ethical imperative to treat HIV disease. Fourth, inaccurate fibrosis staging because of sampling error may lead to misrepresentation of progression [41]. To limit misclassification, slides were read at one sitting in blinded pairs by a single, experienced hepatopathologist who excluded tissues that were insufficient. Moreover, we a classified progression as a change in two Ishak stages, and observed only one subject (0.6%) with a 2-stage decrease, suggesting that sampling error was not a major factor. Finally, the HCV treatment effect observed is limited by low SVR rates in our population; higher SVR rates would be expected in patients with genotypes 2 or 3.
 
In summary, nearly a quarter of co-infected patients had significant fibrosis progression over a 3-year interval, including those with minimal disease. In our cohort, ART and HIV disease measures were not associated with liver disease progression. Importantly, we did not detect evidence of long-term liver injury as a result of ART. Similarly, in the absence of SVR, HCV treatment was not associated with a lower risk of progression. Although our study can not exclude a modest effect of HIV and HCV treatment, further research is needed to identify novel factors that influence fibrogenesis. Our data also suggest that the level of serum AST plays an important role in predicting future liver disease, lending support to the use of this prognostic marker in HCV treatment decision algorithms. Finally, novel therapies with improved antiviral efficacy are needed to manage co-infected individuals.
 
This work was presented, in part, at the 13th Conference on Retroviruses and Opportunistic Infections, Denver, Colorado, 5-8 February 2006.
 
Support: Financial support for this study came from DA-11602, DA-16065 and DA-13806 from the National Institute on Injection Drug Abuse, grant HS 07-809 from the Agency for Healthcare Policy and Research and the General Clinical Research Center at the Johns Hopkins Medical Institutions, MO1-RR00052.
 
Conflicts of interest: None.
 
 
 
 
  icon paper stack View older Articles   Back to Top   www.natap.org