HIV and liver disease forum: Conference proceedings

Abstract

The management of HIV/AIDS has undergone a revolution in recent years. Following the introduction of highly active antiretroviral therapies (HAARTs) in 1996, there was a rapid and dramatic decrease in mortality associated with opportunistic infections that complicate advanced HIV infection. The drop in the incidence and prevalence of formerly common opportunistic disease processes such as atypical mycobacterial infections, Pneumocystis jiroveci (formerly Pneumocystis carinii) pneumonia, and Kaposi's sarcoma was accompanied by a concomitant increase in the proportion of persons with morbidity and mortality associated with liver disease.1, 2 The reasons for this are multifactorial and include longer life expectancy in those with HIV infection, accelerated injury associated with hepatitis C virus (HCV) and hepatitis B virus (HBV) infections, and idiosyncratic hepatotoxic reactions. Before the widespread use of HAART, hepatologists seldom had a direct role in the care and management of patients, and when complex hepatic issues, such as the treatment of coinfected patients, assessment and management of complex HIV medications, and care of HIV-infected patients with end-stage liver disease, became more common, they were poorly prepared to deal with them. To address these issues, an international forum was convened in Jackson Hole, WY, in September 2006. The forum was supported by grants from 3 institutes of the National Institutes of Health (National Institute of Allergy and Infectious Diseases, National Institute of Diabetes and Digestive and Kidney Diseases, and National Institute on Drug Abuse) and by unrestricted grants from the pharmaceutical industry. The objective of this meeting was to bring together laboratory and clinical researchers, drug developers, and representatives of key government agencies in an effort to define the current state of the field, determine research needs, and encourage collaborations between investigators representing a number of different disciplines. This document summarizes key presentations and areas of discussion. ACTG, AIDS Clinical Trial Group; ART, antiretroviral therapy; FDA, US Food and Drug Administration; HAART, highly active antiretroviral therapy; HBsAg, hepatitis B surface antigen; HBV, hepatitis B virus; HCV, hepatitis C virus; IDU, intravenous drug use; MELD, model for end-stage liver disease; NNRTI, nonnucleoside reverse transcriptase inhibitor; NRTI, nucleoside/nucleotide reverse transcriptase inhibitor; OLT, orthotopic liver transplantation; PEG IFN, pegylated interferon; PI, protease inhibitor. Hepatitis viral infections remain important causes of morbidity and mortality worldwide and may exist as isolated infections or as coinfections either with HIV or with each other. Worldwide, it is estimated that 370 million persons are infected with HBV, 130-177 million are infected with HCV, and 40 million are infected with HIV. In 2005, there were over 11,000 new HIV infections per day, of which 1000 were in children less than age 15, 50% were in women, and 40% were among those from 15-24 years of age. Approximately 3 million deaths per year are due to HIV.3 This highest prevalence of HIV is in sub-Saharan Africa, although south-central and southeast Asia and the former Soviet republics also have significant numbers of persons with HIV infection. The prevalence of chronic viral hepatitis in those with HIV varies by region and even by country within a region. For example, in sub-Saharan Africa, approximately 11% of those with HIV have chronic HBV infection, and 60% of adults have evidence of past infection; the prevalence of chronic HCV infection is 2%-3% in this population but also varies widely by country.4 Therefore, of the estimated 40 million people currently living with HIV, approximately 3 million to 4 million of those are infected with HBV, and 4 million to 5 million have chronic hepatitis C. Factors influencing the rate of coinfection include the natural history of disease, the primary mode of transmission, and the geographic distribution. This is particularly true for HBV. When the prevalence of chronic infection is high in a given region, primary infection occurs mainly among infants and young children because of perinatal and horizontal transmission. Unsafe injection practices play a role in viral hepatitis spread in other areas. Where the overall prevalence of chronic viral hepatitis is low, the acquisition of HBV infection is mostly among adolescents and adults, primarily through sexual transmission and intravenous drug use (IDU). HCV is transmitted primarily through injection drug use, although in areas of high endemicity, unscreened transfusions and unsafe therapeutic injections are also major risk factors. Although it is clear that HIV can increase the likelihood of developing chronic HBV and HCV infection and the likelihood of HCV transmission from mother to infant, there are still unanswered questions regarding the seroprevalence rates and the impact of coinfection in sub-Saharan countries most affected by HIV. It is also unclear whether the reactivation of HBV might be common with progressing immunodeficiency. Although the overall incidence of acute HCV has declined in the US, several recent outbreaks in men who have sex with men have been reported.5, 6 Phylogenetic analyses of these outbreaks have confirmed clusters of infection consistent with common sources. Analyses of risk factors have shown associations with multiple partners; sexual practices associated with mucosal trauma, such as “fisting”; or sexually transmitted diseases, such as lymphogranuloma venereum or syphilitic proctitis, which might similarly disrupt the mucosal barrier. Major unanswered questions for the patients who are HIV-infected include the optimal timing of treatment, the duration of therapy for acute HCV in the setting of a patient who is infected with HIV, and whether ribavirin is required. Some investigators have suggested the early treatment of acute HCV without ribavirin use may lead to excellent response rates, and this observation is particularly important in the setting of HIV because of the higher toxicity and risk of interactions with HIV medications. The natural history of coinfection, particularly for HCV/HIV in the era of HAART, is still a matter of debate and is important as it influences intervention strategies. Clearly, viral hepatitis is the most common cause of liver disease in HIV, and recent cohort studies have shown increasing rates of liver disease and death among those with HIV.2 Patients with fewer than 200 CD4 cells/mm3 have a significant risk for increased liver disease progression; however, it is not clear whether those with a CD4 cell count that never decreases below 200/mm3 or those with effective immune reconstitution have the same rate of liver disease progression as those that are HIV-negative. A recent study showed that the presence of detectable HIV viremia (>400 copies) was independently associated with a more rapid progression rate in HCV.7 Although some studies have demonstrated a decline in the fibrosis progression rate on HAART, the use of HAART could theoretically accelerate disease progression by increasing pathogenic T cell responses or through drug toxicity. This area needs further clarification. Liver disease in the setting of coinfection is a multifactorial process (Fig. 1). Studies on the pathogenesis of liver injury often focus on 1 potential mechanism of liver injury, such as hepatitis C, without acknowledging the impact of other cofactors on that process, such as the interaction of alcohol and HCV and its effects on the immune response.8 Although acknowledging this limitation, the discussion has focused on the role of immune responses in the pathogenesis of chronic viral hepatitis and drug-induced mitochondrial injury. Diagram showing factors associated with hepatic injury in HIV-infected patients. Immune responses against HCV are clearly important for the acute spontaneous resolution of HCV.9 In both acutely infected chimpanzees and humans, the initiation and maintenance of broadly directed CD4 and CD8 responses are necessary and critical to the resolution of acute infection; however, why these fail in most exposed individuals is a matter of considerable debate, and there may be multifactorial processes that include viral interference with innate interferon signaling,10 viral variation leading to escape from adaptive T cell responses,11-15 failure to produce critical cytokines,16-18 and possibly failure of dendritic cells.19 There is a relative paucity of literature regarding the innate immune response and the intrahepatic immune response in acute disease. The role of immune responses once chronic disease is established is more controversial. On the basis of classic adoptive transfer experiments in transgenic mice, it would appear that HCV-specific CD8+ T cells are necessary for inflammation in the liver. In clinical cohorts, however, the association of CD8 responses with liver fibrosis has not been shown. Although it would be logical that patients with depressed cellular immunity would have slower HCV disease progression, this is clearly not true. More recent experiments in several cohorts, including those HIV-coinfected, have demonstrated that HCV-specific CD4 cells are associated with protection from liver disease progression.20, 21 In multivariable modeling in 1 cohort of HIV-positive subjects, HCV-specific immune responses were the strongest predictor of liver fibrosis in comparison with other clinical factors.20 Similarly, HCV-specific CD4 responses are strongly associated with sustained virologic response to interferon-based treatment.22 However, CD4 function is not always associated with the CD4 count. One item of discussion was whether or not the earlier initiation of HAART would be beneficial to patients with coinfection, both in preventing liver disease progression and in enhancing the likelihood of response to HCV treatment. Most research on the pathogenesis of viral hepatitis in HIV has focused on HCV rather than HBV, and this has also been identified as an area needing additional research. Recent data were presented demonstrating an increase in HBV precore/core mutations in those with HBV/HIV coinfection that might affect T cell recognition.23 There was also a discussion of the interaction of the 2 viruses in disease pathogenesis. Most of the literature has focused on the influence that HIV has on HCV-specific immune responses. Until recently, it was not possible to culture HCV, and it was impossible to determine whether or not the 2 viruses might interact. Recent studies have demonstrated the induction of hepatocyte apoptosis by both HIV and HCV envelope proteins24 through the induction of signal transducer and activator of transcription factors 1.25 Recent data discussed at this meeting also included the induction of apoptosis in hepatocytes through HIV binding to the HIV coreceptor Cys-X-Cys receptor 4, which is expressed on hepatocytes26; this might sensitize hepatocytes to tumor necrosis factor apoptosis-inducing ligand–mediated apoptosis. None of these studies required HIV entry or infection of hepatocytes, however. Data presented at the meeting also demonstrated that HCV-permissive cell lines express the necessary coreceptors for HIV entry and that HIV can enter these permissive cell lines. This will be an important area of future research in understanding the interaction of HCV and HIV in promoting liver disease. An alternative mechanism of liver injury is related to the potential for antiretroviral agents to cause hepatotoxicity through several different mechanistic pathways. Common class effects are discussed in more detail later. However, a number of broad mechanisms have been implicated. In animal models, the administration of stavudine led to pronounced mitochondrial DNA depletion in the liver but not in skeletal muscle.27 In fact, there are multiple effects of both HIV and antiretrovirals on mitochondria, including effects on DNA polymerase gamma, uncoupling functions, and oxidative stress.28, 29 Reactive oxygen species play a key role in the induction of apoptosis, inflammation, and activation of the stellate cells.30 The intersection of HCV infection and altered lipid metabolism in the setting of HIV was identified as an area needing additional research, but it is a difficult one, given the lack of appropriate models. For example, the interaction of HCV and steatosis is poorly understood, but it is clear that HCV infection may directly alter lipid metabolism,31, 32 and there are changes in the peroxisome proliferator-activated receptors in both HCV and HIV.33, 34 Recent data reported at the meeting included the effect of HCV on mitochondrial DNA depletion35 and whether the treatment of HCV modifies mitochondrial DNA levels back to pretreatment baselines. Numerous guidelines now recommend formal screening for HBV and HCV among patients with HIV infections.36, 37 This recommendation is independent of the presence or absence of liver enzyme abnormalities and is linked to the recognition that treatment decisions may be affected by the presence of 1 or more viral coinfections. Although the recommendation for HCV screening is the initial use of a third-generation enzyme-linked immunosorbent assay, with confirmation provided by HCV RNA testing, this strategy may not be optimal in the coinfected population. False-negative enzyme-linked immunosorbent assay testing has been reported, and screening with HCV RNA may be considered when CD4+ counts are low (<200) or when otherwise unexplained liver enzyme abnormalities are present.38-40 Seroreversion or the return of either a positive or negative antibody state has been described in patients undergoing immune reconstitution or those with dropping CD4+ levels.41 Patients with active IDU may not have positive HCV serologies following acute infection. Optimal timing for repeat screening among IDU and other patients with continued risk exposures has not been determined and represents a key epidemiological research question. High HCV viral loads are present in a significant proportion of patients with HIV coinfection, and some assays provide limited ability for quantitative measurement at the upper end of the scale.42-44 Several studies now document a paradoxical increase in HCV following the initiation of HAART.45-47 This effect is most pronounced among those with low CD4 counts at the initiation of antiretroviral therapy (ART). The mechanism for this rise is unclear, but it may be associated with transient increases in serum aminotransferases that may be misinterpreted as direct drug hepatotoxicity. HCV genotyping before the initiation of therapy for HCV remains critically important, although clear genotype-specific treatment paradigms are less well developed for HCV/HIV-coinfected patients than for the HCV-monoinfected group. The adequate assessment of HBV status is important in this patient population because of the dual HBV activity of several agents used for the treatment of HIV. Hepatitis B testing should include serologic evaluation with hepatitis B surface antigen (HBsAg), hepatitis B e antigen, anti–hepatitis B core antigen immunoglobulin G, and anti–hepatitis B surface antigen. All patients with HBsAg should undergo subsequent HBV DNA testing to determine the need for treatment intervention and as part of the HAART decision process. Patients who are HBsAg-negative and who do not have anti–hepatitis B surface antigen antibodies should be vaccinated for HBV. Standard vaccination protocols are less effective among those with HIV than in the general population, particularly among patients with low CD4 counts.48, 49 There is active research interest in altered regimens of vaccine dose and timing and in the use of adjuvant vaccine agents to enhance vaccine efficacy. Some patients who fail to respond to the vaccine may have anti–hepatitis B core antigen immunoglobulin G alone and have been shown in some cases to have detectable HBV DNA in serum and in the liver. These occult HBV infections appear to be more common in HIV-infected patients than in the general population.50-54 The clinical significance, however, remains unknown, and research in this area was encouraged. Liver biopsy remains an important tool in the assessment of liver fibrosis and in the evaluation of inflammatory activity. Recent studies have demonstrated significant changes in liver fibrosis among HCV/HIV-coinfected patients in timeframes of under 3 years.55 A highly active area of research is the use of noninvasive markers to assess the degree of fibrosis. There was considerable discussion regarding various evaluative algorithms for the assessment of the fibrosis stage. Although encouraging results have been reported, few appear to withstand external validation in cohorts with HCV/HIV-coinfected subjects. Furthermore, there is poor discrimination between fibrosis stages, limiting interpretation to the description of the extremes (e.g., minimal fibrosis versus cirrhosis). However, data were presented suggesting that transient elastometry is a promising modality that warrants further evaluation.56-58 There has been a sea of change in thinking about the management of HIV since 1996, from a terminal disease at presentation to a chronic medical illness with a good prognosis. As a result of improvements in management, the US Department of Health and Human Services recently issued guidelines calling for opt-out testing rather than opt-in testing for HIV. The goal of this strategy is to identify the estimated 300,000 individuals in the United States infected with HIV who are unaware of their status.59, 60 These guidelines would make HIV testing part of routine care for all individuals between the ages of 13 and 64, with the consent for such testing part of the general consent to treatment (similar to cholesterol testing), without the requirement for specific consent. However, legal statutes regarding consent for HIV testing vary by state, and this issue has not been fully resolved. Certainly, for hepatologists seeing patients with viral hepatitis, awareness of the shared modes of transmission should prompt HIV testing in most patients. The initiation of treatment for HIV with ART is currently based principally on the CD4 count, along with the presence of any AIDS-defining illness or symptoms attributable to HIV.61, 62 These guidelines are based on current knowledge of the long-term benefits of treatment balanced against the potential side effects, including an apparent increase in cardiovascular disease and dyslipidemia. All individuals with CD4 counts of less than 200/mm3 should be treated, as well as those with AIDS-defining illnesses or symptoms, regardless of the viral load. Individuals with CD4 counts between 200 and 350/mm3 may be offered treatment; the decision to treat in this range can be individualized on the basis of the rate of decline of the CD4, the presence of any symptoms, and the patient's willingness to initiate treatment. Most individuals with a CD4 count greater than 350/mm3 do not currently start ART unless there are symptoms felt to be due to HIV or very high sustained levels of viremia. However, guidelines concerning the timing of ART may change as more data emerge suggesting that initiation at a higher CD4 count may result in improved immune reconstitution.63 ART regimens are complex, but the basic principal remains that combination therapy is required with drugs active against multiple steps in the viral life cycle or with binding predilection toward different sites of virally encoded genes. The choice of regimens varies somewhat according to patient preference, the toxicities of the agents, other coexisting medical problems, and the potential for interactions with other medications, with guidance provided by various expert consensus groups. Regimens for persons naive to ART include either a protease inhibitor (PI) boosted with low doses of ritonavir to enhance its pharmacokinetic properties or a nonnucleoside reverse transcriptase inhibitor (NNRTI) such as efavirenz; either is given in combination with 2 nucleoside/nucleotide reverse transcriptase inhibitors (NRTIs; see Table 1). The choices of such agents include an assessment of potential tolerability by the individual patient balanced against convenience. For example, efavirenz-based regimens can be administered as a single daily dose and are often well tolerated, but persons with underlying psychiatric disease may have significant neuropsychiatric toxicity, and efavirenz should not be administered to women who may become pregnant because of its teratogenicity. The goal of all treatments is to suppress HIV to levels below the limits of detection by conventional assays and to achieve stabilization or improvement in immune function, usually measured by an increasing CD4 cell count. If the treatment can be maintained, there is a 95% chance of maintaining HIV suppression.64 There have been 3 multicenter randomized clinical trials and several other single-center studies that have evaluated the role of pegylated interferon (PEG IFN) alphas for their ability to achieve sustained clearance of the HCV virus in HCV/HIV-coinfected patients. These studies clearly demonstrate that PEG IFN with ribavirin is superior to older standard interferon regimens. However, the response rates have been consistently lower than those for HCV-monoinfected subjects.65-67 Although several factors may contribute to the low response rates, 2 factors appear to stand out. First, patients with HCV/HIV coinfection have very high viral loads compared with HCV-monoinfected subjects.68 Additionally, in the key pivotal trials, lower doses of ribavirin were generally used (800 mg/day or a 600-1000 mg/day escalating dosage). The overall response rates ranged from 27%-40% for a sustained viral response. A complex metaregression model that was presented suggested that differences between the study outcomes could not be attributed to differences in age, race, CD4 count, degree of fibrosis, or other identifiable factors. Therefore, it is unclear what specific selection criteria other than genotype could be used to enhance the treatment response. Data from the Peginterferon Ribavirina España Confección [PRESCO] trial were presented and suggested that better responses might be achieved with higher doses of weight-based ribavirin, but because active controls are lacking, only historical response comparisons are available.69 In terms of treatment tolerability and side effects, the general consensus is that treatment with interferon-based therapies remains difficult but achievable among those with HCV/HIV coinfection. The avoidance of azido-thymidine (zidovudine) is desirable because it reduces the severity of ribavirin-associated anemia.70 Growth factors are often required to maintain adequate levels of hemoglobin and limit leucopenia in this population. Specific issues related to drug interactions and hepatotoxicity are noted later. There is growing evidence that intravenous drug users, which represent a key reservoir for HCV and HIV infection, can be identified, evaluated, and managed in terms of HCV treatment intervention. There was considerable discussion regarding recurrent disease following successful therapy among patients at risk for continued IDU. Early data suggest a low but clearly important subset of patients who do become reinfected.71 The role of multidisciplinary care programs for patients with active IDU was recognized.72 The role of maintenance therapy remains unclear. AIDS Clinical Trial Group (ACTG) 5071 demonstrated histological improvement among HCV/HIV subjects who failed to clear HCV by 24 weeks of treatment.65 A large, multicenter trial of maintenance therapy (SLAM-C [Suppressive Long-term Antiviral Management of Hepatitis C]) is currently underway in the ACTG at more than 30 US sites, and large trials in HCV-monoinfected patients should be completed in the near future. There are a number of treatment options and strategies relevant to HBV management in HIV-infected patients. These include agents licensed and approved by regulatory bodies for the treatment of HBV (lamivudine, adefovir, entecavir, telbivudine, interferon, and PEG IFN) and those with HBV activity that are approved only for HIV treatment (tenofovir and emtricitabine). Among coinfected patients, however, there are limited data available in comparison with HBV-monoinfected populations. A metaanalysis of interferon utility in HBV/HIV-coinfected patients showed limited efficacy in comparison with a placebo.73 There are no published data regarding PEG IFN use in coinfected subjects, and many agreed that trials in HBV/HIV-coinfected subjects with preserved CD4 counts are indicated, particularly for patients with high CD4 counts who would not typically be placed on ART. Among the nucleoside/nucleotide analogues, data were presented for lamivudine, adefovir, entecavir, and tenofovir. All appear to have viral suppressive activity on HBV DNA with variations in the potency and treatment response. Among lamivudine-resistant patients, both adefovir and tenofovir appear to be effective treatment regimens, although 1 study from Peters et al.74 suggests a trend toward tenofovir superiority versus adefovir in an ACTG treatment cohort. Combination therapy is now routine in coinfected patients (e.g., using Truvada, which in a single pill combines tenofovir plus emtricitabine), although the optimal combination to maintain long-term virologic suppression is not known. Viral mutation and drug resistance are key themes of scientific interest. Intermittent and frequent exposure to lamivudine and infection with preexisting lamivudine mutants leads to significant concerns regarding drug choice and the use of single-agent therapy among those with HIV infection. Unique HBV core deletion mutants were described, and the question of whether such mutations might be associated with increased liver fibrosis and disease severity was entertained. It is clear that detailed sequence analysis of HBV in those with HIV may lead to important findings related to disease progression, and more research work in this area was encouraged. The development of new agents for the treatment of viral hepatitis in the setting of HIV infection was a key topic for discussion. Although all agree that the significance of HBV and HCV was high in this patient population, multiple problems exist for those involved with drug development. First, patients with HIV have multiple sources for hepatic injury, which complicates the evaluation of a putative drug treating a single disease process. Drug tolerability tends to be lower, and this limits compliance with the treatment regimen and creates difficulty in adverse event attribution. This is well demonstrated in several pivotal HCV treatment trials in coinfected subjects in which high patient dropout rates were observed.66, 67 Another example is the attribution of renal toxicity in subjects who are also at risk of primary HIV-associated nephropathy. The potential for drug-drug interactions is significantly increased in a setting in which patients may be on multiple agents, thus requiring significant evaluation of pharmacokinetics before drug approval. For example, a new agent that interacts with cytochrome P450 may either increase or decrease HIV PIs or NNRTI levels, resulting in HIV treatment failure or toxicity. Finally, there is significant risk of dichotomous outcomes based on the stage of HIV disease. Those with early HIV have a different cytokine milieu than those with advanced HIV disease. On the other hand, patients with advanced liver disease will metabolize medications differently than those with minimal fibrosis. An important tool to evaluate many of these issues in limited subcohorts of patients and over shorter time periods is viral kinetic modeling. Within just a few weeks, complex viral decline models, coupled with pharmacokinetic sampling, may yield important information that will guide the development of larger trials. Talal et al.75 described the relationship between the interferon concentration and HCV decline during the first 2 weeks of therapy. This approach appears to be informative and has wide application as protease and polymerase inhibitors are developed for HCV treatment. The group achieved consensus in terms of the need to begin trials of new agents among HCV/HIV-infected and HBV/HIV-infected patients early in the drug development process. Drug-associated hepatotoxicity is a major primary factor in the development of liver injury among those with HIV infection and a significant cofactor among those with other concomitant sources of liver injury (e.g., HCV, HBV, and alcohol). There is a wide spectrum of disease presentation. The most dramatic but rarest form of injury is manifest as fulminant hepatic failure. In contrast, subacute toxicity associated with liver enzyme abnormalities is extremely common. An analysis of data from a community-based AIDS treatment program, the Community Programs for Clinical Research on AIDS, demonstrated that the incidence of grade 4 liver aminotransferase abnormalities (values greater than 10 × ULN [upper limit of normal]) was significantly more common than other forms of drug-associated toxicities.76 There are 4 primary mechanisms of ART-associated hepatotoxicity. Broadly, these may be viewed as class effects, although significant variability exists within classes when specific agents are evaluated. Table 1 lists the classes of antiretroviral agents and the agents within that class. In general, NRTIs may cause liver injury through the inhibition of polymerase gamma within the mitochondria of hepatocytes. Polymerase gamma inhibition leads to the development of steatosis, lactic acidosis, and, in extreme cases, liver failure.77 This process may be exace