|
GB Virus C Reduced HCV-Liver Disease in HIV/HCV Coinfected
|
|
|
A New Variable Influencing HCV-Related Liver Disease in HIV-HCV Coinfected Individuals? EDITORIAL
Gastroenterology Dec 2007
Jack Stapleton
Division of Infectious Diseases, Department of Internal Medicine, The University of Iowa and Iowa City VA Medical Center, Iowa City, Iowa
See "Reduction in hepatitis C-related liver disease associated with GB virus C in human immunodeficiency virus coinfection" by Berzsenyi MD, Bowden DS, Kelly HA, et al, on page 1821. SEE BELOW
![liver-1.gif](../images/121807/liver-1.gif)
Figure 1. Interactions between HCV, HIV, GBV-C, and liver disease. HIV infection is associated with increased HCV plasma viral RNA concentrations (viral load), and also increased severity and rate of progression of liver disease; thus, HIV interacts with HCV. The converse-HCV altering HIV disease progression-has not been clearly demonstrated (dashed line and question mark). GBV-C infection and HIV interact in a bidirectional manner, and GBV-C is associated with prolonged survival, increased CD4 counts, lower HIV viral load, improved response to HIV therapy, and decreased maternal-fetal transmission of HIV, whereas HIV viral load is inversely related to GBV-C viral load. Although previous studies did not suggest a direct interaction between GBV-C and HCV (dashed line and question mark), Berzsenyi et al found that GBV-C was associated with diminished hepatic fibrosis, liver-related death, and cirrhosis in HIV-HCV coinfected individuals. Direct interactions between HIV and hepatic function have been suggested, but are difficult to document. (Color version of this figure available online at www.gastrojournal.org)
Mortality resulting from human immunodeficiency virus (HIV) infection plummeted following widespread use of highly active antiretroviral therapy (HAART), which prevents HIV-related immune dysfunction. Nonetheless, mortality in HIV-infected people has not fallen to rates observed in the general population, and the etiology of HIV-related mortality has shifted; as a result end-stage liver disease is now the leading cause of death in this population.1 Due to shared modes of transmission, hepatitis C virus (HCV) infection is common among those who are infected with HIV, particularly those who acquire HIV through intravenous drug use.2 As a consequence, HCV has emerged as the main cause of morbidity and mortality in the HIV- infected population.2, 3
Current knowledge of the mechanism(s) by which HCV is able to persist and cause hepatic injury is incomplete. Approximately 20% of individuals who acquire HCV spontaneously clear viremia, presumably as a result of host immunologic control of viral replication.4 By contrast, HCV persists in the majority of infected individuals, and persistence involves viral and host interactions, including direct interference with host innate antiviral pathways and escape from humoral and adaptive cellular immune responses.5 Progression of HCV-related liver disease is slow, and the majority of infected individuals never develop clinically apparent liver disease.4 Factors that increase the likelihood and rate of hepatic fibrosis involve immune suppression, including concurrent HIV infection.5 Identification of viral and host factors that increase or slow HCV disease progression in HIV-HCV coinfected individuals is critically important, because liver injury and fibrosis are the most common cause of morbidity and mortality in HIV-infected population receiving HAART.
GB virus type C (GBV-C; also called hepatitis G virus) was discovered in 1995 in serum obtained from individuals with non-A, non-B, non-C hepatitis (reviewed by Stapleton6). Transmission of GBV-C occurs by parenteral, sexual, and maternal-fetal exposure to the virus, and infection is extremely common in humans, present in ~2% and up to 17% of healthy blood donors from industrialized and developing countries, respectively.6 Among those with other blood borne or sexually transmitted infections, the prevalence of GBV-C is much higher, reaching 42% in 1 HIV-infected cohort.7 Based on RNA sequence and genome organization comparisons, GBV-C is the most closely related human virus to HCV.6 Although initially thought to cause hepatitis, extensive epidemiologic studies did not identify an association between GBV-C and either acute or chronic liver disease, nor for any other disease entity studied to date (reviewed in Stapleton6 and Alter8). Presumably, the reason GBV-C does not cause hepatitis relates to specific cellular tropism; GBV-C primarily replicates in lymphocytes, unlike HCV, which replicates primarily in hepatocytes.9, 10, 11, 12 Surprisingly, persistent GBV-C infection was found to be associated with prolonged survival in numerous studies performed before HAART and in studies in which GBV-C viremia was detected late in the course of HIV infection14, 15, 16 (and reviewed in Zhang et al13). Viremia appears to be required for the association, as transient GBV-C infection is not associated with improved survival.14, 15, 16 Most, although not all, studies conducted during the HAART era do not identify prolonged survival in association with GBV-C, possibly due to the confounding effect of HAART.
In this issue of Gastroenterology, Berzsenyi et al. examined the effect of GBV-C infection on HCV-related liver disease and survival among 158 HIV-HCV co-infected individuals studied during the HAART era.17 Subjects with evidence of alternative causes of liver disease were excluded to ensure that hepatic morbidity and mortality was related primarily to HCV infection. GBV-C viremia prevalence was high (36%), and as shown in other studies,14, 15, 16 clearance (23%; 13/57) and acquisition (21%; 12/57) of GBV-C viremia was observed frequently. Viremia was associated with significantly less compensated and decompensated cirrhosis, and with cirrhosis-free survival. Consistent with these data, hepatic fibrosis and hepatic encephalopathy were significantly lower in the group with GBV-C viremia, and there was a trend toward decreased ascites among the GBV-C infected group (P = .06).17 GBV-C infected subjects also had a trend towards higher CD4+ T-cell counts, and the low HIV RNA concentration observed in both groups demonstrate that HAART was prescribed and effective in this cohort.
These results raise questions as to how GBV-C might influence HCV-related liver disease, and what role does HIV infection play in this interaction? GBV-C modulates lymphocyte function(s), which may explain the beneficial associations between GBV-C viremia and HCV-related hepatic disease, and HIV-related survival. HCV infection causes hepatic inflammation, and liver injury appears to be the result of immunologic destruction of infected hepatocytes (reviewed in Dustin et al5). Paradoxically, liver disease is accelerated among immunocompromised individuals including those with HIV infection or recipients of solid-organ transplants, suggesting either a direct viral mechanism of liver injury or altered immunologic function leading to accelerated HCV-related hepatic injury.5
The pro-inflammatory Th2 cytokine IL-10 is associated with establishment of persistent HCV infection,18 and IL-10 induces the expression of the profibrotic Th2 cytokines IL-4 and IL-13.19 In contrast, infusion of exogenous IL-10 into HCV-infected individuals resulted in increased plasma HCV RNA levels and decreased numbers of HCV-specific CD4+ and CD8+ T cells.20 Consequently, Th2 cytokines may also be involved in HCV pathogenesis and hepatic injury, although the characterization of this hypothesis is incomplete, GBV-C replicates in both T and B lymphocytes,12 a primary source of Th cytokines, and GBV-C infection is associated with modulation of cytokines in vivo and in vitro. Specifically, GBV-C viremia was associated with low plasma levels of IL-4 and IL-10 (Th2 cytokines) and high levels of IL-2 and IL-12 (Th1 cytokines) during ~ 9 years of follow up in an cohort of HIV-infected Sicilian subjects, whereas those without GBV-C viremia demonstrated an increase in IL-4 and IL-10 and reduced IL-2 and IL-12 plasma concentrations over time.21 Ongoing studies demonstrate that Th2 cytokines are down-regulated in fresh human lymphocytes infected with GBV-C in vitro compared with mock-infected control cells.22 This effect appears to be mediated at least in part by the GBV-C nonstructural protein 5A, as expression of this protein in a CD4+ T cell line significantly down-regulated Th2 cytokines (IL4, IL10, and IL13).22 This polarization towards a Th1 cytokine profile may conceivably contribute to a protective effect of GBV-C on HCV-related fibrosis. A switch from a Th1 to a Th2 cytokine profile is also associated with HIV disease progression,23 and the effect of GBV-C on Th cytokines may contribute to the beneficial effect on survival associated with GBV-C. The propensity of HIV-infected people to develop Th2 cytokines could also promote HCV-related liver injury later, and this may explain why GBV-C infection of HCV-infected, HIV-uninfected cohorts have not observed a beneficial effect on liver disease.8 It is worth noting that the effect of GBV-C on the polarization of T cell cytokines towards a Th1 profile might also influence diseases other than HIV and HCV, an area that deserves further study.
GBV-C has additional effects on T cell function that may represent potential mechanisms by which GBV-C could influence HCV-related liver injury. GBV-C infection induces a variety of chemokines and down-regulates the HIV co-receptors CCR5 and CXCR4 on the surface of lymphocytes (reviewed in Stapleton et al24). GBV-C was recently reported to be associated with reduced CD4+ T cell activation in HIV-infected subjects, as determined by reduced CD38 surface expression,25 and the GBV-C NS5A protein down-regulates CD4 mRNA levels and CD4 surface density in vitro.26 These effects on T cells may modify antigen presentation, slow T cell recruitment, and potentially reduce the effectiveness of HCV-specific cellular immune responses, which may subsequently influence the rate and severity of HCV-related disease.
GBV-C infection is associated with improved responses to HIV therapy in several studies (reviewed in Stapleton et al24), and this may represent another explanation for the reduction in hepatic disease in this co-infected cohort as effective antiretroviral therapy and higher CD4 counts are associated with a reduction in HCV-related fibrosis.3, 27 GBV-C infection also inhibits HIV replication in vitro,24 and it is possible that the beneficial effect of GBV-C is not on HCV or HCV-immune responses directly, but rather the beneficial effect of GBV-C on HIV replication and immune function which results in decreased HCV-related liver injury.
The study by Berzsenyi et al17 provide an opportunity to examine and better understand microbial copathogenesis. Unlike the laboratory setting in which microbes are studied in isolation, humans with HCV infection are multiply infected. The resulting microbial and host interactions are complex and interdependent, and these may influence the rate and extent of clinical disease.28 The extent of the interactions will undoubtedly be influenced by the cellular tropism of the organisms, and the lymphotropic GBV-C is capable of widespread alteration in cytokine and T-cell function. Future studies on the effect of GBV-C on HCV-related pathogenesis should include examination of the effect of GBV-C on cohorts with HCV-related and other forms of liver disease, and in HIV-infected and uninfected cohorts. This will allow determination as to whether the effect is restricted to HIV-infected people, and if this result is specific for HCV-induced hepatic injury.
One additional issue to consider is the effect of HCV therapy on GBV-C. As noted by Berzsenyi et al, interferon therapy (with or without ribavirin) is active against some isolates of GBV-C, and the 2 viruses share considerable amino acid identity and predicted structural similarities in their protease and polymerase proteins (reviewed in Stapleton6). The effects of investigational drugs that target the HCV protease and polymerase have not yet been tested for activity against GBV-C. Studies to determine the effects of new HCV medications on GBV-C and characterization of the effect of GBV-C clearance on HCV treatment response are thus warranted. (Figure 1).
Reduction in Hepatitis C-Related Liver Disease Associated With GB Virus C in HIV Coinfection
Gastroenterology Dec 2007
Mark D. Berzsenyi, D. Scott Bowden, Heath A. Kelly, Kerrie M. Watson, Anne M. Mijch, Rachel A. Hammond, Suzanne M. Crowe, Stuart K. Roberts
Department of Gastroenterology, Alfred Hospital, Prahran, Victoria, Australia
Molecular Microbiology Laboratory, Victorian Infectious Diseases Reference Laboratory, North Melbourne, Victoria, Australia
Epidemiology Unit, Victorian Infectious Diseases Reference Laboratory and School of Population Health, University of Melbourne, Melbourne, Victoria, Australia
Infectious Diseases Unit and Victorian HIV Service, Alfred Hospital, Macfarlane Burnet Institute for Medical Research and Public Health, Melbourne, Victoria, Australia
AIDS Pathogenesis and Clinical Research Program, Macfarlane Burnet Institute for Medical Research and Public Health, Melbourne, Victoria, Australia
This work was presented in part at the 57th Annual Meeting of the American Association for the Study of Liver Diseases (AASLD) held in Boston, Massachusetts, October 27-31, 2006
Background & Aims: It has been reported that GB virus C infection (GBV-C) leads to improved morbidity and mortality in patients with human immunodeficiency virus (HIV) infection. However, GBV-C has no effect on the course of liver disease in hepatitis C virus (HCV) monoinfection. The aim of the study was to determine the influence of GBV-C infection on liver disease in patients with HCV/HIV coinfection.
Methods: Data on 158 HCV/HIV patients were collected from January 1996 to October 2005. Two plasma specimens, collected at least 18 months apart, were tested for GBV-C RNA by reverse transcription-polymerase chain reaction with primers to the NS5B gene and confirmed using E2 gene primers and sequencing. Antibodies to GBV-C E2 protein were also determined. Liver-related morbidity and mortality were assessed from patient records.
Results: Fifty-seven of 158 (36%) patients had GBV-C RNA and 94 (59%) had evidence of exposure to GBV-C based on combined polymerase chain reaction and antibody results. Thirty-four (21%) patients had features of cirrhosis, with 20 having compensated and 14 having decompensated cirrhosis.
Active GBV-C RNA was significantly associated with a reduction in cirrhosis, both compensated and decompensated in multivariate analysis (hazard ratio, 0.27; 95% confidence interval, 0.08-0.88; P = .03), as well as in analysis for cirrhosis-free survival vs duration of HCV infection (P = .006). No significant effect on liver-related or overall survival was observed.
Conclusions: In these HCV/HIV-coinfected patients, GBV-C RNA was associated with a significant reduction in the severity of HCV-related liver disease.
GB virus C (GBV-C) is the most closely related human virus to hepatitis C virus (HCV), another member of the Flaviviridae.1, 2 Unlike HCV, GBV-C is not hepatotrophic and neither replicates in hepatocytes nor causes acute or chronic hepatitis.3 Rather, GBV-C has been found to have tropism for cells of the hemopoietic lineage with evidence of replication in a number of different mononuclear cells, including CD4-positive T cells.4, 5 Similar to the human immunodeficiency virus (HIV), GBV-C can be transmitted by sexual and percutaneous routes and is frequently found in populations at risk for blood-borne or sexually transmitted viruses.6, 7, 8, 9
GBV-C infection on its own has not been associated with the development of a specific disease.10 However, studies performed in HIV-infected individuals from the pre- and post-highly active antiretroviral treatment (HAART) eras have suggested that GBV-C infection is associated with an improvement in morbidity and mortality, with slower progression to acquired immunodeficiency syndrome (AIDS).11, 12, 13, 14, 15, 16 Nevertheless, controversy remains regarding this viral interaction because not all studies clearly support a beneficial effect.17, 18, 19, 20
GBV-C prevalence among patients with HCV monoinfection varies from 11% to 24% depending on the population studied.3 GBV-C has been reported to have no impact on the outcome of HCV infection, HCV-related chronic liver disease, hepatic fibrosis, or transaminase levels.21, 22 However, in HCV/HIV-coinfected patients, there is clear evidence that HCV-related liver disease is accelerated, with a significant increase in progression of hepatic fibrosis.22, 23 The latter can be slowed by antiretroviral therapy,24 although hepatoxicity remains a concern in the setting of advanced liver disease.3
Information is limited about the role GBV-C plays in HCV/HIV coinfection. No study to date has addressed the effect of GBV-C infection on HCV-related liver disease in the setting of HIV. This study provides evidence that the presence of GBV-C RNA is associated with a significant reduction in the severity of HCV-related liver disease in HCV/HIV coinfection.
Discussion
In 158 patients coinfected with HCV/HIV, the presence of GBV-C RNA was associated with a significant reduction in HCV-related liver disease. Cirrhosis, hepatic encephalopathy, and advanced liver fibrosis on liver biopsy were significantly reduced with active GBV-C infection, whereas ascites approached significance. No difference was seen for bilirubin, international normalized ratio (INR), or creatinine, and hence MELD score, whereas encephalopathy was the only factor relevant to the assessment of the Child-Pugh score to show a significant difference between groups. In multivariate modeling, development of cirrhosis was reduced in the presence of GBV-C RNA. However, this association was not apparent following clearance of GBV-C RNA. Similarly, others have shown that loss of GBV-C RNA was not associated with improved outcomes among HIV monoinfected men.16 GBV-C viremia had no effect on liver-related survival or overall survival despite categorical univariate analysis suggesting an association. The findings are principally due to a reduction in HCV-related liver disease outcomes associated with GBV-C viremia seen among male patients, who constituted the majority of the study group. The sex bias seen in this study reflects the epidemiology of HIV infection in Australia. Despite the mean age of the active GBV-C group being less than other groups, there was no statistical relationship found between age and GBV-C infection. Previous studies have noted GBV-C infection to be more common in younger patients.27, 31, 32, 33
All patients in this study were well characterized, with HCV infection the only identifiable cause of cirrhosis. Various host and viral factors have previously been suggested as influencing HCV disease progression in HIV coinfection. Levels of CD4 immunosuppression, older age at infection, and excess alcohol intake have been implicated, and HCV viral load and genotype have been suggested, although convincing evidence is lacking.34 Other than the degree of CD4 immunosuppression and increasing age, we found no other association with these factors. Significant alcohol consumption was not examined because it was excluded in patient selection to avoid confounding bias. Because analysis was based on only 2 time points, it was not possible to determine the relative effect of early or late loss of GBV-C RNA in the group clearing or acquiring viremia.
A potential influence of GBV-C status on HIV/AIDS was detected, with a trend toward higher CD4 counts in the presence of GBV-C RNA as well as cirrhosis being significantly less common with maintenance of higher CD4 counts. No other influence on other HIV/AIDS-related parameters was apparent. The findings of observed improvement in CD4 counts with active GBV-C are consistent with a number of studies suggesting that GBV-C may modify parameters that may in turn slow the progression to HIV/AIDS.11, 12, 13, 14, 15, 16 The only other study that examined the effect of GBV-C on HIV in the setting of HCV coinfection failed to show any association with the prevalence of liver disease; however, that study did not include patients with hepatic decompensation.31 Another study has suggested that GBV-C could be a phenomenon secondary to HIV progression rather than an independent prognostic factor.17 Others have shown an improved outcome for HIV/AIDS with persistent GBV-C RNA but concluded that RNA persistence may depend on sufficient CD4 T cells and that loss of these T cells in HIV progression may be the cause rather than consequence of GBV-C RNA loss.18 In addition to this, there are also a number of studies performed in the post-HAART era showing no clear association between GBV-C and HIV/AIDS.19, 20 A number of factors of particular relevance to this study may potentially contribute to the observed differences in the literature. A significant proportion of patients in our study had liver-related morbidity, which may mask further beneficial effect of GBV-C on HIV/AIDS. Other factors may include use of HAART, duration of HIV infection, and time-dependent change in CD4 count as well as the possible effect of different GBV-C genotypes on immune function.16, 18, 31
The majority of patients who cleared GBV-C viremia during the study failed to develop detectable levels of E2 antibodies. Others have reported similar findings following interferon-induced clearance of GBV-C in HCV/HIV-coinfected patients31 and following spontaneous clearance of GBV-C in HIV-infected patients.16, 18 Alterations in immune function because of HIV infection and changes in HAART regimens as well as drug holidays may contribute to lack of seroconversion. This suggests that E2 antibodies are not a reliable marker of past GBV-C infection in populations with impaired immune function, particularly because the E2 antibody test was validated in healthy blood donors.27 Therefore, results from E2 antibody assays were not used in the final analysis of this study.
An association of GBV-C with cellular immunity seems clear, with trophism for lymphocytes demonstrated and replication occurring in CD4-positive T cells.4, 5 However, GBV-C replication has not been demonstrated in hepatocytes.3 That may help explain why no role for GBV-C has been seen in immunocompetent patients with HCV monoinfection.21 In contrast, there are a number of studies suggesting that GBV-C exerts an influence on HIV monoinfection by effecting T-cell function. GBV-C is thought to lead to up-regulation of T-helper 1 cytokines and down-regulation of T-helper 2 cytokines, leading to an immunologic profile that is associated with an improved outcome for HIV/AIDS. GBV-C induces chemokines such as regulated on activation, normal T-cell expressed and secreted (RANTES), which leads to decreased expression of CCR5 or CXCR4 that are coreceptors for HIV entry to the T cell. Furthermore, GBV-C E2 envelope protein has been found to interact with CD81 to increase RANTES and decrease CCR5 surface expression thus also affecting entry of HIV to the T cell. Also, different genotypes of GBV-C may confer varying degrees of "protective effect," and recent work has shown that GBV-C NS5A phosphoprotein can inhibit HIV-1 replication.3, 31, 35 Should GBV-C be shown to have a causal relationship in limiting the progression of liver disease in HCV/HIV coinfection, perhaps similar mechanisms involving alteration in T-cell function can be implicated. It is already clear that liver injury in HCV monoinfection results from the cellular immune response rather than via a direct viral effect,36 and our findings and that of others34 show that CD4 immunosuppression has a role in liver disease progression in HCV/HIV coinfection. Further evidence to support this comes from molecular studies of liver biopsy specimens performed on a subgroup of patients participating in our study (data not shown). Our preliminary data indicate that GBV-C viremia alters cellular messenger RNA expression of genes involved in signal transduction from the T-cell receptor complex on intrahepatic T cells. This is currently under further investigation and may represent a potential mechanism for the effect of GBV-C on liver disease in HCV/HIV coinfection.
Ultimately, this study was limited by the relatively small study population and the partial retrospective analysis. The sample size was a direct consequence of the stringent selection criteria used in this study to identify a population that was actively infected with only HCV and HIV without other known hepatitis viruses or confounding factors such as alcohol abuse. In this way, all liver-related outcomes could be ascribed to HCV. Prospective validation of this work is required with a multicentered study recruiting a larger population of patients. Perhaps a larger study population with greater follow-up time may show differences with respect to liver-related mortality that was noted in our univariate categorical analysis but not with further survival analysis. Changes in overall mortality would be unlikely because of the current widespread use of HAART. In conclusion, this study identified a significant reduction in HCV-related liver morbidity associated with GBV-C viremia in HCV/HIV-coinfected patients. GBV-C infection may contribute to the unpredictable clinical outcomes in HCV/HIV-coinfected patients and could have a role as a prognostic marker in this setting.
Results
A total of 158 patients, of whom 137 (87%) were male, met the study inclusion criteria (Figure 1). GBV-C RNA was detected in 57 of 158 (36%) patients using NS5B primers, with confirmation in 56 of 57 using E2 primers. Sequence analysis of the E2 region showed a high level of homology following comparison with published E2 GBV-C sequences. Consecutive Genbank accession numbers were obtained for the 56 E2 gene sequences from EF612208 to EF612263. Of the patients who were GBV-C RNA positive, 13 cleared RNA during the course of the study. Within the active GBV-C group, 12 acquired RNA during the study period, and 32 were persistently infected. Overall exposure to GBV-C was 94 of 158 (59%) patients based on 38 (24%) patients who were GBV-C E2 antibody positive combined with positive RT-PCR results. Only 1 patient was both GBV-C RNA and antibody positive from the same sample, whereas 3 other patients had equivocal E2 antibody results on repeated testing. Of note, 10 of 13 (77%) patients who cleared GBV-C viremia during the course of the study failed to develop antibodies.
Demographic characteristics of patients across the study groups were similar (Table 1). The mean age of patients with active GBV-C was younger than in other groups, but this difference was not statistically significant (P = .10). The minimum duration of known HIV infection was longer than known HCV infection because reliable testing for HCV using second-generation enzyme immunoassay only became widely available in Australia in 1991. HIV testing has been available since 1985. The recorded time from baseline sample to follow-up sample was similar across all groups with a minimum period of 3.9 years. The median time from follow-up sample to assessment of clinical data was 1 day thus eliminating potential for change of clinical or virologic status. HCV genotype had been determined for 98 patients: 66 (67%) were genotype 1 and 26 (26%) were genotype 3 (Table 1), which reflects the genotype distribution seen in Australia. HCV viral load had been determined for 79 patients with no differences seen on the basis of GBV-C viremia. No differences were found for sex, racial origin, or mode of acquisition of HIV/HCV. CD4 counts were noted to be higher in the active GBV-C group but failed to reach statistical significance in univariate analysis. No differences were found in HIV viral load, frequency of AIDS diagnosis, or time from HIV diagnosis to AIDS between groups. Similar levels of HAART were used. No association was found between GBV-C E2 antibodies and the study's primary or secondary outcome measures.
Cirrhosis was detected in 34 of 158 (21%) patients, with 20 having compensated and 14 having decompensated cirrhosis. There was an agreement of 37 of 39 (95%) patients in the findings of the clinical audit and the study with respect to the diagnosis of cirrhosis. Only 2 patients with borderline clinical features of cirrhosis were classified differently. Both were GBV-C RNA negative, and the audit classification did not affect the outcome classification by GBV-C exposure in the study. GBV-C RNA was detected in 28% of the 39 audited patients compared with 36% of the 158 study patients (P = .45), whereas 41% of audit patients were classified with cirrhosis compared with 21% in the study (P = .02).
Active GBV-C viremia was associated with a significant reduction in cirrhosis (both compensated and decompensated) (P = .003), and decompensated cirrhosis alone was also reduced in association with viremia (P = .008) in univariate analysis (Table 2). Subgroup analysis indicated that those persistently infected with GBV-C RNA, among the active GBV-C group, also had a significant reduction in compensated/decompensated cirrhosis (P = .02) and decompensated cirrhosis alone (P = .03). In Kaplan-Meier analysis, active GBV-C was associated with a significant reduction in compensated/decompensated cirrhosis as a function of days from diagnosis of HCV infection (P = .006) and days from diagnosis of HIV infection (P = .006) (Figure 2). A reduction in the stage of hepatic fibrosis was associated with active GBV-C (P = .02) and hepatic encephalopathy (P = .04) with a trend noted for ascites (P = .06). There was a trend for cirrhosis to be more common among patients older than 35 years of age (P = .09). No reduction of liver-disease related outcomes was observed following clearance of GBV-C RNA. Potential factors associated with cirrhosis-free survival were identified in univariate analysis (Table 3). These factors were then incorporated into Cox proportional-hazards regression analysis, which indicated 2 variables significantly associated with compensated/decompensated cirrhosis-free survival: active GBV-C RNA (HR, 0.27; 95% CI: 0.08-0.88; P = .03) and increasing CD4 count at follow-up bleed (HR, 0.99; 95% CI: 0.99-1.00; P = .004).
In univariate comparison of cause of death, an increased risk of mortality was associated with cirrhosis, both compensated and decompensated (OR, 7.19; 95% CI: 2.86-18.10; P < .001), and the diagnosis of AIDS (OR, 5.15; 95% CI: 2.10-12.60; P < .001), but no association for active GBV-C RNA was observed (OR, 1.58; 95% CI: 0.64-3.89, P = .32). No association was found between active GBV-C and liver-related death vs days from diagnosis of HCV (P = .13) or days from diagnosis of HIV (P = .22) in Kaplan-Meier analysis. Furthermore, no effect on overall survival according to GBV-C status as a function of days from diagnosis of HCV infection (P = .38) and days from diagnosis of HIV infection (P = .21) was observed (Figure 3). However, 10 of 25 (40%) deaths during the study period were liver related, with 9 of 10 of these deaths occurring in the GBV-C-negative group (Table 2).
Materials and Methods
Study Population and Design
Patients with HCV/HIV infections in the post-HAART era were identified from outpatient attendance at the Alfred Hospital, Melbourne, from January 1996 onward. This was an exploratory study with retrospective identification of patients and prospective follow-up of living subjects. The study protocol, including methods of patient identification and data extraction, were approved by the Alfred Hospital Ethics Committee. In addition, with Ethics Committee approval, patient consent was not sought for data extraction, and all data were anonymous. Patient clinical and laboratory data were obtained from the time that each potential study patient was diagnosed with HIV and had documented clinical and laboratory data relevant to the investigation of liver disease. Eligible patients were followed from the time they were identified as satisfying inclusion criteria until the study censure date of October 2005.
To be eligible for this study, it was necessary to identify for each patient stored plasma samples that were both HIV antibody and HCV RNA positive collected a minimum of 18 months apart. All specimens were obtained prior to commencement of interferon-based therapies for chronic HCV because this treatment can be associated with clearance of GBV-C RNA.25, 26 Specimens were assigned coded identifiers. Exclusion criteria were: patients without clinical information or plasma suitable for analysis; active hepatitis B virus (HBV) infection (hepatitis B surface antigen and/or HBV DNA positive); significant alcohol consumption (>40 g or 4 standard drinks/day) and/or alcoholic liver disease; nonalcoholic fatty liver disease; hemochromatosis; Wilson's disease; α-1-antitrypsin deficiency; and neoplastic conditions including cholangiocarcinoma, lymphoma and metastatic carcinoma.
GBV-C RNA was detected using reverse-transcription polymerase chain reaction (RT-PCR). This was determined for both baseline and follow-up plasma specimens. GBV-C negative was defined as absence of RNA at baseline and follow-up specimens. Cleared GBV-C was defined as RNA positive at baseline and negative at follow-up. Active GBV-C was defined as those patients who were RNA positive at follow-up. This included patients who were RNA positive at both baseline and follow-up and those who were RNA negative at baseline but positive at follow-up. The presence of GBV-C RNA in the follow-up plasma samples was used to determine both primary and secondary outcomes. Antibodies to GBV-C, reported as a marker of past infection in healthy blood donors, was determined in the follow-up specimens.27
Evidence of HCV-related liver disease was sought from medical records and biochemistry, radiology, histology, and endoscopy results as previously described.28 Cirrhosis was defined by at least one of the following in the absence of a liver biopsy: radiologic evidence of cirrhosis/portal hypertension on more than one occasion; endoscopic evidence of varices/portal gastropathy; or clinical features of chronic liver disease including hepatic decompensation such as abdominal ascites, hepatic encephalopathy, spontaneous bacterial peritonitis, bleeding varices, and hepatocelluar carcinoma. Appropriate investigations were reviewed when determining whether cirrhosis was present to rule out differential diagnoses. Abnormal liver function tests, prolonged clotting profiles, and/or decreased platelet counts were not considered sufficient alone to establish a diagnosis of cirrhosis. Deaths were classified as liver related if autopsy or clinical data described significant complications of hepatic decompensation, end stage liver disease, or hepatic failure leading to the patient's death in the absence of other identifiable causes and competing risks.
Clinical data closest to the date of the follow-up plasma specimen were used for the final assessment of liver disease-related outcomes. Also recorded were HIV/AIDS-related factors such as CD4-positive T-cell count at baseline and follow-up, HIV viral load, use of HAART, and occurrence of AIDS-defining illnesses. Clinical information was randomly audited from the hospital records of 39 of 158 (25%) study patients by a physician (S.K.R.) blinded to the results of GBV-C exposure. Audit and study classification of liver disease were then compared. Primary end points were advanced liver disease defined as either compensated or decompensated cirrhosis and liver-related mortality. Secondary end points included Model for End Stage Liver Disease (MELD) score and Child-Pugh score.
Virologic Analysis
Viral nucleic acid was extracted from plasma using the MagNA pure LC instrument and Total Nucleic Acid isolation kit (Roche Diagnostics, Penzberg, Germany) with amplification performed using SuperScript III One-Step Reverse-Transcription PCR system with Platinum Taq DNA Polymerase (Invitrogen, Carlsbad, CA). GBV-C was detected as previously described29 using primers from the nonstructural gene NS5B as an initial screen. The sensitivity of the GBV-C RNA assay using NS5B primers was conservatively estimated to be 105 RNA copies/mL based on comparisons with similar in-house RT-PCR assays. Plasma positive for GBV-C RNA with the NS5B primers was then confirmed using a 2 round RT-PCR assay with primers to the structural envelope gene E2 as previously described.30 E2-positive samples underwent nucleic acid sequencing using the ABI PRISM Dye Terminator cycle sequencing ready reaction kit (PE Applied Biosystems, Foster City, CA), and homologies were compared with other GBV-C sequences using an NCBI blast search (www.ncbi.nlm.nih.gov). Antibodies to GBV-C E2 protein were identified by the enzyme immunoassay kit, ĘPlate anti-HGenv test (Roche Diagnostics, Mannheim, Germany). Equivocal results for GBV-C E2 antibodies were retested and, if still equivocal, were deemed negative for the purpose of this study.
HCV RNA was detected by qualitative RT-PCR with Cobas AMPLICOR HCV Test, version 2.0 (Roche Diagnostics, Branchberg, NJ) and genotyping by the Versant HCV genotype assay (LIPA) (Bayer HeathCare, Tarrytown, NY). HCV viral load was determined by Versant HCV RNA 3.0 assay (bDNA) (Bayer HeathCare), HBV surface antigen by Abbott/Imx HBsAg microparticle immunoassay (Abbott Laboratories, Chicago, IL), and HBV DNA by Versant HBV DNA 3.0 assay (bDNA) (Bayer HeathCare). Antibodies to HIV were detected by Abbott Murex HIV-1.2.0 EIA assay (Abbott) and confirmed by HIV BLOT 2.2 Western blot assay (Genelabs Diagnostics, Geneva, Switzerland). HIV viral load was assayed by Cobas AMPLICOR HIV-1 MONITOR Test Version 1.5 (Roche) and CD4 count by flow cytometry using Cyto-Stat triCROME reagent (Beckman Coulter, Fullerton, CA).
Statistical Analysis
Separate univariate analyses of categorical variables with cirrhosis or death as the outcome variable were done using the χ2 test or Fisher exact test. Univariate analyses of continuous normally distributed variables were done by logistic regression and of continuous nonparametric variables were done using the Kruskal-Wallis test. Kaplan-Meier curves were constructed comparing study end points in GBV-C-active and GBV-C-negative groups. The starting point for survival analysis was time from diagnosis of HCV or HIV. The P values for Kaplan-Meier curves were performed using the log-rank test for equality of survival function. Multivariate analysis for survival time was performed using Cox proportional hazards regression and reported using hazard ratios (HR) with 95% confidence intervals (CI). Variables included in the regression analysis were presented with associated odds ratios and 95% CI. Entry criteria for multivariate modeling were as follows: variables of potential biologic significance (age and sex), variables previously associated with improved outcome in HIV/GBV-C coinfection (HIV viral load, CD4 count, diagnosis of AIDS), and variables with a P value of <.05 in the univariate analyses. The exit criterion was a P value of >.1. Multivariate models were constructed using a stepwise selection technique and validated using a backwards elimination technique before undergoing a final assessment for clinical and biologic plausibility. A 2-sided P value of <.05 was considered to be statistically significant. All analyses were performed using SAS version 8.2 (SAS Institute, Cary, NC).
|
|
|
|
|
|
|