Chronic hepatitis C virus (HCV) results in slowly progressive hepatic fibrosis, which leads to cirrhosis and its complications in a subset of individuals. Several studies have shown that coinfection with the HIV-1 is associated with more rapidly progressive liver disease.1,2 After the introduction of highly active antiretroviral therapy (HAART) and as a result of prolonged HIV-related survival, HCV infection has now emerged as a leading cause of morbidity and mortality in coinfected patients.3-6 Because the complications of HCV infection are primarily related to the development of advanced liver fibrosis (cirrhosis), noninvasive tests to monitor progressive fibrosis and the impact of therapies on liver fibrogenesis would have great clinical utility.7 Liver biopsy remains the “gold standard” to stage liver fibrosis but is expensive, carries significant risks, and is associated with substantial sampling error and observer error.8,9 Furthermore, repeated liver biopsy is seldom feasible, such that monitoring disease progression remains a major clinical challenge.
Over the last few years, many serologic tests and indices of liver fibrosis have been assessed as potential noninvasive markers of liver fibrosis. These can be broadly divided into nonspecific markers and those that reflect extracellular matrix (ECM) metabolism.7 Nonspecific markers of fibrosis include a number of simple biochemical and hematologic parameters, such as aspartate aminotransferase (AST)/alanine aminotransferase (ALT) ratio, platelet count, and international normalized ratio (INR)10-12 as well as more sophisticated indices such as Fibrotest,13,14 the Forns index,15 and the AST platelet ratio index (APRI).16 Each of these detects alterations in liver function associated with progressive liver disease as well as the development of portal hypertension. Markers of ECM metabolism include the procollagens, hyaluronic acid (HA), laminin, and YKL-40 as well as the tissue metalloproteinases and their inhibitors (TIMPs).17
Although many of these noninvasive tests have been evaluated in HIV-negative populations, only Fibrotest has been validated in an HCV/HIV-coinfected population.18 HIV infection and the use of HIV medications have the potential to affect the diagnostic performance of these assays significantly. For example, HIV infection and HAART can directly affect the platelet count and serum transaminases.19,20 Furthermore, HIV can alter serum cytokines important for the regulation of ECM metabolism, and the effects of HIV on extrahepatic tissues could change the performance of these markers in HCV/HIV-coinfected populations.21-23 The aim of this study was therefore to assess the applicability of several noninvasive markers of fibrosis in an HCV/HIV-coinfected population and to compare their diagnostic characteristics with those of a population infected only with HCV.
All subjects are part of an ongoing prospective natural history study on the impact of HCV infection on HIV disease (CHARM cohort). The CHARM cohort consists of 370 HCV-infected injection drug users with and without HIV coinfection. The subjects were recruited from the Infectious Disease and Liver Clinics at Boston University Medical Center and from the Boston Veterans Affairs Health System Infectious Disease Clinic. Subjects had given written informed consent as part of the larger cohort study, and a proportion underwent liver biopsy as part of their clinical evaluation. Serum was obtained at the time of the liver biopsy or within 6 months of the biopsy and was stored in aliquots at −80°C until assayed. Other biochemical, hematologic, and serologic studies were analyzed at their respective hospital laboratories as part of the routine preliver biopsy assessment. The AST/ALT ratio, Forns index, and APRI were calculated as previously reported.10,15,16 The APRI is a normalized ratio of AST to platelet count, and the Forns index is calculated from measurements of γ-glutamyl transferase, total cholesterol concentration, platelet count, and the subject's age. Biopsies were graded using the Ishak score by an experienced pathologist (MO'B), who was blinded to prior readings and to the patients' clinical information.24,25
Markers of ECM metabolism included YKL-40, HA, and procollagen III N peptide (PIIINP). YKL-40 and HA levels were measured by enzyme-linked immunosorbent assay (ELISA) according to the manufacturer's instructions (YKL-40: Chondrex, Metra Biosystems, Mountain View, CA; HA, Corgenix, Denver, CO). PIIINP levels were measured by radioimmunoassay (RIA-gnost P III P; Cis Bio International, France).
HCV-monoinfected subjects were compared with HCV/HIV-coinfected subjects on categoric characteristics using the χ2 test and on measurement characteristics and assay results using the Wilcoxon rank-sum test. Associations between assay results and Ishak stage (range: 0-6, ordinal scale) were described using the Spearman rank correlation coefficient. The performance of individual assays in predicting cirrhosis (Ishak stage 5 or 6) and in predicting fibrosis (Ishak stage 3 or greater) was described in the area under the receiver operating characteristic (AUROC) curve. The Hanley and McNeill method was used to compare AUROC curves. Analysis was performed using Sigmastat 2.0 statistical software (SPSS, Chicago, IL); receiver operating characteristic (ROC) curve analysis was performed using Analyze-It software (Analyze-It Software, Ltd., England, UK).
Study Patient Demographics
The patient population comprised 97 subjects with HCV: 57 with HCV and 40 with HCV/HIV coinfection (Table 1). The 2 groups were of similar age and gender. The racial mix was different (P = 0.03), however, with a higher proportion of white subjects and fewer Hispanics in the HCV group. In the HIV-infected subgroup, 33 subjects were on HAART, of whom 24 (75%) had an undetectable HIV viral load. Those with a detectable HIV viral load tended to have more advanced liver fibrosis (median fibrosis score = 5) than those with an undetectable HIV viral load (median fibrosis score = 2), but this did not reach statistical significance (P = 0.18).
HCV RNA levels were similar in the HIV-positive and-negative groups, with a trend to increased prevalence of genotype 1 patients in the HCV group. The distribution of liver fibrosis stage was similar in both groups; approximately one third of each group had cirrhosis (Ishak stages 5 and 6). Levels of the liver transaminases, platelet count, and INR as well as the indices of fibrosis and markers of ECM metabolism of fibrosis were similar, although a higher AST/ALT ratio and a trend to a higher Forns index score and lower platelet count were noted in the HCV/HIV-coinfected group (see Table 1).
Relationship to Fibrosis Stage
The correlation between each of the parameters studied and fibrosis stage is shown in Table 2. Age, platelet count, and INR all showed a significant correlation to the stage of fibrosis. Interestingly, whereas CD4 cell count showed a significant correlation to the stage of disease, CD4 percentage did not. The AST/ALT ratio did not show a correlation with the stage of disease, perhaps related to the 10 subjects with a normal ALT level. The APRI and Forns index correlated well with the fibrosis stage, although the analysis for the Forns index was limited by a smaller sample size (n = 68). Similarly, YKL-40 and PIIINP levels correlated significantly with the fibrosis stage. HA showed a poor correlation in the HCV-infected group but a strong correlation in the HCV/HIV-coinfected population. Almost all fibrosis markers showed a stronger correlation in the HCV/HIV-coinfected group than in the HCV-infected group. A comparison of correlation coefficients between the HCV/HIV-coinfected and HCV groups revealed significantly higher coefficients in the HCV/HIV-coinfected group (P = 0.006).
ROC curve analysis was performed to compare the diagnostic performance of each of these assays and indices (Table 3). Data were analyzed for the 2 levels of hepatic fibrosis: significant hepatic fibrosis (Ishak stage ≥3) and cirrhosis (Ishak stages 5 and 6). The AUROC curves were generally greater for the diagnosis of cirrhosis as compared with significant hepatic fibrosis. When the results for the 2 groups were compared, the AUROC curves in the HCV/HIV-coinfected group were generally higher but within the confidence limits of those achieved in the HCV group. These differences were most apparent for the markers of ECM metabolism (HA, YKL-40, and PIIIINP). A similar but less marked trend was observed for the nonspecific markers of fibrosis (APRI, Forns index, platelet count, and INR) for the diagnosis of cirrhosis, but this trend was not evident for the diagnosis of significant hepatic fibrosis.
Table 4 shows the same data expressed as the performance of each of the assays in terms of their sensitivity and specificity for the diagnosis of cirrhosis. Overall, the performance of each of these assays was fair to good, with some assays attaining sensitivities and specificities in the 80% to 90% range, with more assays attaining this level of performance in the HCV/HIV-coinfected population. Similar but less good diagnostic performance was obtained for the diagnosis of significant hepatic fibrosis (data not shown).
The performance characteristics of the assays were different depending on HIV status. Levels of almost all the markers tended to be more abnormal in the HCV/HIV-coinfected group at the later stages of fibrosis (Fig. 1). For example, the median level of HA in cirrhotic patients was 122 ng/mL in those with HCV infection and 279 ng/mL in those with HCV/HIV coinfection (P = 0.02). Similarly, these differences reached statistical significance for platelets (HCV: 180 × 109/L, HCV/HIV: 130 × 109/L; P = 0.04) and APRI (HCV = 46, HCV/HIV = 93; P = 0.04). There was also a trend to higher levels of YKL-40 and PIIINP and a higher Forns index score in the HCV/HIV-coinfected group (P < 0.1). These trends were also reflected in the cutoff values selected for the diagnosis of cirrhosis and were apparent for the markers of ECM metabolism (YKL-40, HA, and PIIINP) at the earlier stages of fibrosis but were not observed for the nonspecific markers of fibrosis (see Fig. 1). As shown in Figure 1, YKL-40 and PIIINP levels were substantially higher in the HCV/HIV-coinfected group as compared with the HCV group at stage 3 or greater fibrosis, but this difference was not observed for the APRI and INR. Similar results were obtained for HA, Forns index, and platelet count (data not shown). The presence of HIV infection did not affect the levels of any of these markers at the earliest stages of fibrosis (Ishak stages 0-2; see Fig. 1).
To assess the possible impact of liver biopsy length and observer error in the staging of the liver fibrosis, we performed 2 further ROC curve analyses. In the first series of analyses, biopsies of less than 10 mm, 15 mm, and 20 mm were excluded and the analyzes were repeated. Whereas there was a trend to larger AUROC curves for some assays, the improvements were not large and the AUROC curves remained within the confidence intervals (data not shown). In the second series of analyses, biopsies with fibrosis scores adjacent to the cutoffs used to classify the biopsies as significant fibrosis (Ishak stage 2/3) or cirrhosis (Ishak stage 4/5) were excluded. In some cases, the AUROC curves showed substantial improvements (eg, when stage 2/3 was excluded for the diagnosis of significant fibrosis, the AUROC curve for PIIINP improved from 0.80 [0.71 to 0.90] to 0.92 and that for YKL-40 improved from 0.78 [0.68 to 0.87] to 0.85). Although these trends held true for the other markers of fibrosis, they were generally less marked (data not shown).
Over the last few years, there has been increased interest in the development of new markers and indices of liver fibrosis. This interest has been fueled by the importance of liver fibrosis as the major prognostic marker for most chronic liver diseases. Several of these markers have been gaining acceptance as research tools as well as in clinical practice. The use of noninvasive markers of fibrosis has not been well studied in HIV-positive populations. Myers et al18 reported good results with Fibrotest/Fibrosure, a now commercially available marker of fibrosis that combines 5 biochemical parameters, including haptoglobin, α2-macroglobulin, γ-glutamyl transferase, total bilirubin, and apolipoprotein A1, with age and gender. More recently, Kelleher et al26 described the SHASTA index, a score derived from HA, AST, and albumin levels, but the index was derived from a largely African American population and was not validated in separate cohorts. In this study, the APRI showed reasonable diagnostic performance, although YKL-40 performed less well. There has been only 1 small direct comparison between the performance of these assays in HIV-positive and -negative populations.27 The aim of the current study was therefore to compare the diagnostic performance of several noninvasive markers of liver fibrosis in a population of HIV-positive and -negative patients with HCV and to evaluate for differences in the characteristics of these assays in the 2 populations.
We have shown that a variety of nonspecific markers and markers of ECM metabolism perform similarly in HIV-positive and -negative populations and are probably valid in coinfected populations. Indeed, there was a trend to improved performance of these assays in the HCV/HIV-coinfected population as compared with those patients with only HCV. This was demonstrated by the higher correlation coefficients between stage of fibrosis and the fibrosis markers as well as by the larger AUROC curve when applied as diagnostic tests for significant hepatic fibrosis and cirrhosis. The strength of this observation is, however, limited by the small sample size and large confidence intervals and needs to be confirmed in larger studies. For example, the performance of HA measurement as a noninvasive test of fibrosis in the HCV group was less than that reported in previous larger studies.28-31
Most of these assays performed better for the diagnosis of cirrhosis than for that of significant hepatic fibrosis, particularly in the HCV/HIV-coinfected group, suggesting that these tests may be of particular value for the noninvasive diagnosis of cirrhosis. Simple measures such as platelet count and INR as well as the more complex but readily available APRI all showed that they had reasonable discriminant value to be clinically useful in HIV-positive populations. Although some of the markers of ECM metabolism showed moderately good diagnostic value when used alone, they were not markedly superior to these simple indices and tests. More complex indices combining the markers of ECM metabolism with the other parameters might provide an improved index as has been suggested by the SHASTA index.26
The characteristics of the diagnostic performance of these assays did seem to be different in the 2 cohorts studied. For example, the selected cutoff values for the diagnosis of the varying levels of fibrosis tended to be more abnormal in the HCV/HIV-coinfected population as compared with those with in population with HCV infection only. When we apply the APRI cutoff value of 1.5 proposed by Wai et al16 for the exclusion of significant hepatic fibrosis in the HIV-negative population, this gave a specificity of 96% compared with only 80% in the HCV/HIV-coinfected group. These data suggest that application of these tests to HIV-positive cohorts may require adjustment of the cutoff values used and that these need to be determined in larger studies.
There are several reasons why the diagnostic performance of all these assays may be altered in HIV-positive cohorts. These include the accelerated natural history of HCV, alterations in bone marrow and immune function, and potential effects of HAART-associated hepatotoxicity. For example, HIV-induced thrombocytopenia would significantly affect the performance of the APRI.19,20 The hepatotoxic effects of HAART and accelerated fibrosis progression observed in HIV-positive individuals may reflect alterations in serum transaminase levels as well as markers of ECM metabolism. HIV infection causes tissue injury at extrahepatic sites, including the bone marrow and lymphatic tissues, as well as increasing serum levels of profibrogenic cytokines such as transforming growth factor (TGF)-βg.21-23 Each of these factors could act alone or in concert to change the performance of these markers in this population.
There are several important limitations to this study. First, the sample size is small, limiting statistical power and requiring that these findings be confirmed in larger studies. Second, the use of needle liver biopsies as the gold standard remains a source of controversy because of problems of sampling and observer error. For example, studies have shown that liver needle biopsy misses cirrhosis in 10% to 30% of cases and that there is discordance of at least 1 stage in up to 30% of paired needle biopsies. Furthermore, postmortem and laparoscopic studies have shown that 2 to 3 biopsies are necessary to stage liver fibrosis adequately.8,32-35 In more recent studies, Bedossa et al9 have shown that the optimal length of a liver biopsy is 25 mm, a length not frequently achieved in routine clinical practice, as seen in 1 study in which only 14% of liver biopsies met this criteria.36 Because of the problems associated with needle biopsy, Poynard et al14 have argued that Fibrotest and perhaps other noninvasive markers of fibrosis may more accurately reflect the stage of disease than needle biopsy. In support of this argument is the observation that the diagnostic performance of Fibrotest improves when measured only against larger biopsies.14 Second, in a detailed study, Poynard et al36 found that incorrect fibrosis staging was more often attributable to liver biopsy than to Fibrotest. Although we did not use strict size criteria for the evaluation of our biopsies, the mean length was 1.5 cm, but all biopsies were deemed to be “adequate” by an experienced pathologist with an interest in this area.25 Assessment of the impact of liver biopsy length on the performance of the assays in this study showed minimal or no improvement in assay performance.
A third important limitation is the ability to generalize these findings to other populations. For instance, most of the HCV/HIV-coinfected subjects had well-controlled HIV disease with good immune function and no recent alcohol or drug use. Whether these assays would perform equally well in more immunosuppressed populations or those with ongoing drug or alcohol use is unclear and requires further study.
In summary, this study shows that many of the available noninvasive markers of liver fibrosis have a diagnostic accuracy in an HCV/HIV-positive population comparable to that observed in a population with only HCV infection. Larger studies are needed to define the performance characteristics of these assays in HCV/HIV-coinfected populations and in those with more advanced immune suppression. Several of these simple tests, including platelet count, APRI, Fibrotest/FibroSure, HA and the Forns index, have adequate diagnostic value to be of clinical utility when assessing fibrosis in a broader clinical context, however. When combined with other clinical, biochemical, and radiologic data, they could be useful tools to help guide clinical decision making regarding the introduction of interferon therapy as well as the initiation of screening for hepatocellular carcinoma and esophageal varices.
1. Bica I, McGovern B, Dhar R, et al. Increasing mortality due to end-stage liver disease in patients with human immunodeficiency virus infection. Clin Infect Dis
2. Puoti M, Spinetti A, Ghezzi A, et al. Mortality for liver disease in patients with HIV
infection: a cohort study. J Acquir Immune Defic Syndr
3. Mohsen AH, Easterbrook PJ, Taylor C, et al. Impact of human immunodeficiency virus (HIV
) infection on the progression of liver fibrosis in hepatitis C virus
infected patients. Gut
4. Benhamou Y, Bochet M, Di Martino V, et al. Liver fibrosis progression in human immunodeficiency virus and hepatitis C virus
coinfected patients. The Multivirc Group. Hepatology
5. Powderly WG. Antiretroviral therapy in patients with hepatitis and HIV
: weighing risks and benefits. Clin Infect Dis
. 2004;38(Suppl 2):S109-S113.
6. Monga HK, Rodriguez-Barradas MC, Breaux K, et al. Hepatitis C virus
infection-related morbidity and mortality among patients with human immunodeficiency virus infection. Clin Infect Dis
7. Afdhal NH, Nunes D. Evaluation of liver fibrosis: a concise review. Am J Gastroenterol
8. Regev A, Berho M, Jeffers LJ, et al. Sampling error and intraobserver variation in liver biopsy in patients with chronic HCV infection. Am J Gastroenterol
9. Bedossa P, Dargere D, Paradis V. Sampling variability of liver fibrosis in chronic hepatitis C. Hepatology
10. Giannini E, Risso D, Botta F, et al. Validity and clinical utility of the aspartate aminotransferase-alanine aminotransferase ratio in assessing disease severity and prognosis in patients with hepatitis C virus
-related chronic liver disease. Arch Intern Med
11. Imperiale TF, Said AT, Cummings OW, et al. Need for validation of clinical decision aids: use of the AST/ALT ratio in predicting cirrhosis in chronic hepatitis C. Am J Gastroenterol
12. Pohl A, Behling C, Oliver D, et al. Serum aminotransferase levels and platelet counts as predictors of degree of fibrosis in chronic hepatitis C virus
infection. Am J Gastroenterol
13. Imbert-Bismut F, Ratziu V, Pieroni L, et al. Biochemical markers of liver fibrosis
in patients with hepatitis C virus
infection: a prospective study. Lancet
14. Poynard T, McHutchison J, Manns M, et al. Biochemical surrogate markers of liver fibrosis
and activity in a randomized trial of peginterferon alfa-2b and ribavirin. Hepatology
15. Forns X, Ampurdanes S, Llovet JM, et al. Identification of chronic hepatitis C patients without hepatic fibrosis by a simple predictive model. Hepatology
16. Wai CT, Greenson JK, Fontana RJ, et al. A simple noninvasive index can predict both significant fibrosis and cirrhosis in patients with chronic hepatitis C. Hepatology
17. Oh S, Afdhal NH. Hepatic fibrosis: are any of the serum markers useful? Curr Gastroenterol Rep
18. Myers RP, Benhamou Y, Imbert-Bismut F, et al. Serum biochemical markers accurately predict liver fibrosis in HIV
and hepatitis C virus
co-infected patients. AIDS
19. Scaradavou A. HIV
-related thrombocytopenia. Blood Rev
20. Walker UA, Bauerle J, Laguno M, et al. Depletion of mitochondrial DNA in liver under antiretroviral therapy with didanosine, stavudine, or zalcitabine. Hepatology
21. Wiercinska-Drapalo A, Flisiak R, Jaroszewicz J, et al. Increased plasma transforming growth factor-beta1 is associated with disease progression in HIV
-1-infected patients. Viral Immunol
22. Alonso K, Pontiggia P, Medenica R, et al. Cytokine patterns in adults with AIDS. Immunol Invest
23. Gressner AM, Weiskirchen R, Breitkopf K, et al. Roles of TGF-beta in hepatic fibrosis. Front Biosci
24. Ishak K, Baptista A, Bianchi L, et al. Histological grading and staging of chronic hepatitis. J Hepatol
25. O'Brien MJ, Keating NM, Elderiny S, et al. An assessment of digital image analysis to measure fibrosis in liver biopsy specimens of patients with chronic hepatitis C. Am J Clin Pathol
26. Kelleher TB, Mehta SH, Bhaskar R, et al. Prediction of hepatic fibrosis in HIV
/HCV co-infected patients using serum fibrosis markers: the SHASTA index. J Hepatol
27. Mastroianni CM, Liuzzi GM, D'Ettorre G, et al. Matrix metalloproteinase-9 and tissue inhibitors of matrix metalloproteinase-1 in plasma of patients co-infected with HCV and HIV
. HIV Clin Trials
28. Oberti F, Valsesia E, Pilette C, et al. Noninvasive diagnosis of hepatic fibrosis or cirrhosis. Gastroenterology
29. McHutchison JG, Blatt LM, de Medina M, et al. Measurement of serum hyaluronic acid
in patients with chronic hepatitis C and its relationship to liver histology. Consensus Interferon Study Group. J Gastroenterol Hepatol
30. Patel K, Gordon SC, Jacobson I, et al. Evaluation of a panel of non-invasive serum markers to differentiate mild from moderate-to-advanced liver fibrosis in chronic hepatitis C patients. J Hepatol
31. Patel K, Lajoie A, Heaton S, et al. Clinical use of hyaluronic acid
as a predictor of fibrosis change in hepatitis C. J Gastroenterol Hepatol
32. Schlichting P, Holund B, Poulsen H. Liver biopsy in chronic aggressive hepatitis. Diagnostic reproducibility in relation to size of specimen. Scand J Gastroenterol
33. Colloredo G, Guido M, Sonzogni A, et al. Impact of liver biopsy size on histological evaluation of chronic viral hepatitis: the smaller the sample, the milder the disease. J Hepatol
34. Maharaj B, Maharaj RJ, Leary WP, et al. Sampling variability and its influence on the diagnostic yield of percutaneous needle biopsy of the liver. Lancet
35. Abdi W, Millan JC, Mezey E. Sampling variability on percutaneous liver biopsy. Arch Intern Med
36. Poynard T, Munteanu M, Imbert-Bismut F, et al. Prospective analysis of discordant results between biochemical markers and biopsy in patients with chronic hepatitis C. Clin Chem