HIV and the Liver

Introduction

Liver disease is an important source of morbidity and mortality in those with human immunodeficiency virus (HIV) infection. Indeed, in Western countries, liver disease is ranked among the top three causes of attributable death in those with HIV. The etiologies of liver disease are diverse and include chronic viral infections, nonalcoholic fatty liver/steatohepatitis (NAFLD/NASH), and drug toxicities ( Table 37-1 ). Hepatic fibrosis appears to be accelerated due to the presence of a persistent state of immune activation, and perhaps due in part to direct HIV infection of stellate cells that leads to activation and excess collagen deposition. In addition, combination antiretroviral therapy (cART) used to suppress HIV replication can directly and indirectly lead to liver injury. Healthcare providers, including hepatologists who manage HIV-infected patients, need to be familiar with the subtleties of managing these patients, as drug-drug interactions and their effects on the liver can impact management and interpretation of common laboratory tests.

Epidemiology of Liver Disease in Those With HIV Infection

The importance of liver disease as a factor in the morbidity and mortality of those with HIV infection was not appreciated until the mid- to late-1990s. The introduction of highly effective antiretroviral agents changed the epidemiology of HIV-associated death previously attributable to opportunistic infections. Multiple large cohort studies have demonstrated the significant impact of liver-related processes on patient survival. In the French Mortavic study, 26,000 HIV-infected patients were followed between 1995 and 2010. AIDS-related morbidity accounted for 20% of the recorded deaths, followed by 13% attributable to chronic liver diseases. The Data collection on Adverse events of Anti-HIV Drugs (D : A : D) study cohort followed HIV-infected patients between 1999 and 2011 in the United States, Europe, and Australia. The study enrolled 49,731 patients representing 308,719 person-years of follow-up. The leading cause of death was AIDS-related diseases followed by non-AIDS malignancies, with 13% of deaths attributed to liver disease, making it the second leading cause of non-AIDS death in the cohort. With combination antiretroviral therapies, patients are living longer with HIV and liver disease has emerged as a leading etiology of serious disease and death. These observations opened the door to our deeper understanding that chronic viral infections of the liver, drug toxicity, and fatty liver disease often led to accelerated disease outcomes among HIV-infected patients. The epidemiology of liver disease has not been static but has evolved as new treatment modalities for HIV, as well as prevention strategies, have modified the disease processes ( Fig. 37-1 ). The blood supply, which was highly contaminated with both hepatitis C virus (HCV) and HIV in the early 1980s, has been effectively eliminated as a source of these infections. In the late 1990s and early 2000s, severe and sometimes fulminant liver disease was found to be attributable to hepatitis B infection during either immune reconstitution or lamivudine breakthrough. Recognition of this issue and the widespread use of antiretroviral agents such as tenofovir as part of the cART backbone incidentally led to reductions in severe liver disease attributable to chronic hepatitis B infection. Pioneering work by Benhamou et al. established that HCV infection was associated with more rapid progression of hepatic fibrosis in the setting of HIV/HCV coinfection. However, it has taken more than a decade to appreciate that HIV suppression played a key role in slowing HCV-associated liver disease progression. Many early antiretroviral agents were highly hepatotoxic but were used because of a lack of safer options. Overall, the evolution of treatment for HIV has led to the use of agents with little or no intrinsic drug hepatotoxicity.

New disease processes are being increasingly recognized as key players in liver disease. Hepatitis E was not thought to play an important role in disease in Western countries, but is increasingly recognized as a common pathogen in those areas. Relatively recently came the observation that HIV-infected persons and other immunosuppressed populations can develop chronic HEV infection and that this can be a source of progressive hepatic fibrosis leading to cirrhosis. Hepatitis D is well recognized, but has been poorly characterized in the HIV population. Long-term use of antiretroviral agents has played a role in fat redistribution and development of nonalcoholic fatty liver disease/nonalcoholic steatohepatitis (NAFLD/NASH), which is increasingly recognized as a factor in progressive liver disease among those with HIV.

Changes in population behavior can also significantly influence changes in overall epidemiology. There has been a significant uptick in new acute HCV infections among HIV-infected men who have sex with men (MSMs) which is in part attributed to increased high-risk sexual behavior. This is thought to be linked to serosorting in which HIV-positive men, effectively managed on cART, believe that they can engage in high-risk behaviors without consequence. A significant change in drug-use patterns with emergence of injection drug use of heroin has appeared in the Eastern United States throughout the past several years. This outbreak, which in some areas includes HIV/HCV coinfections, is a nonurban epidemic in young adults.

Mechanisms of Liver Disease Unique to HIV

It is reasonable to assume that known mechanisms of liver injury that have been elucidated for a variety of etiologies in those without HIV infection also apply when HIV infection is present. There are notable exceptions, however. For example, early studies of HBV infection in persons with HIV documented decreased risk of liver injury in patients with low CD4 counts. This reflects the immunologic nature of liver injury associated with HBV and can be significantly modulated by immune dysfunction. Overall, the patient with HIV represents a unique environment with extrinsic factors that modulate normal responses leading to either accentuation or diminution of phenotypic outcomes ( Fig. 37-2 ).

After HIV infection, there is an early and rapid loss of mucosal tight junctions in the gut, associated with profound loss of gut-associated lymphoid tissue (GALT). This alteration leads to gut leakiness associated with translocation of bacteria and bacterial remnants into the splanchnic circulation. Most typically, these are defined by the presence of endotoxin, though other less well-defined bacterial components are probably involved. These substances enter the liver and lead to significant Kupffer cell depletion. Failure to clear these products completely leads to a generalized state of immune activation. Simultaneously, infected individuals have both decreased T cell–mediated immune responses and activation of innate immune pathways which have the potential to lead to hepatic fibrosis associated with an altered cytokine microenvironment. Despite strong suppression of HIV and increases in peripheral CD4 counts with antiretroviral agents, there appears to be little restoration of gut immune function and HIV infection is characterized by persistent immune activation.

HIV infects and replicates mainly in CD4+ T cells. It binds using C-C chemokine receptor type 5 (CCR5), and in later infection the CXC-receptor 4 (CXCR4) which are highly expressed on CD4+ cells. However, other cell types, including hepatocytes and hepatic stellate cells, also express these receptors. Infected T cells produce not only virions but also large amounts of gp120 protein that can be found in the circulation in HIV-infected individuals. Both gp120 and HIV virions can bind to their cognate receptors on hepatocytes and stellate cells where they activate various innate immune pathways. Interaction of gp120 and HIV virions with CXCR4 on hepatocytes up-regulates TRAIL (TNF-related apoptosis-inducing ligand) leading to apoptosis mediated by JNK II. Increased reactive oxygen species may facilitate nuclear factor kappa B (NF-κB)-mediated fibrosis progression. Engagement of surface chemokine receptors also increases transforming growth factor beta-1 (TGF-β1) expression. This may be a driver of HCV replication.

Initiation of appropriate cART rapidly decreases HIV viral loads in most patients. Flares of alanine aminotransferase (ALT) activity, particularly in those with HCV infection are well described. Whereas a portion of these may be attributable to drug toxicity (see below), recent efforts to elucidate this process have revealed that there is a complex interplay between HIV suppression and down-regulation of interferon-stimulated genes associated with innate immune responses. This results in increased HCV viral replication, which in the setting of immune reconstitution is accompanied by increased apoptosis, and transient ALT increases. Liver biopsy findings are not consistent with drug hepatotoxicity and in fact show decreases in hepatic inflammation. ALT and HCV flares tend to resolve within 48 weeks with what appears to be establishment of a new set point, leading to lower levels of circulating HCV RNA.

Chronic use of cART may lead to ongoing liver injury. Both didanosine and stavudine demonstrate increased risk of mitochondrial toxicity, relative to other nucleoside/nucleotide analogs. Protease inhibitors may lead to liver injury by affecting retinoic acid–binding proteins. Drug hypersensitivity reactions are seen with abacavir and nevirapine. Except for the immune-mediated injury processes, hepatocyte injury tends to be dose-dependent. Thus, high doses of ritonavir are intrinsically hepatotoxic, but boosting doses are generally safe. Drug-related hepatotoxicity is examined in greater detail below.

Liver Biopsy and Noninvasive Markers of Liver Disease Severity

Accurate determination of the presence and degree of hepatic fibrosis is essential for prognostication and guiding treatment decisions in patients with chronic liver diseases. Percutaneous liver biopsy remains the gold standard for assessing hepatic fibrosis, and has traditionally been considered the key diagnostic tool in determining the degree of inflammation and fibrosis in chronic liver disease. Liver biopsy also provides additional information regarding factors that might affect disease progression such as steatosis. Liver biopsy, however, is invasive and carries a complication rate ranging from 1% to 5%, with the risk of mortality ranging from 1 in 1000 to 1 in 10,000. Risks may be higher in certain patients with HIV/AIDS, as lesions at high risk of bleeding (e.g., peliosis hepatitis) are more likely to be present and alternate sampling modalities such as transjugular liver biopsy may be employed. Because of the risks associated with liver biopsy, there have been attempts to identify noninvasive tests or surrogate markers that could accurately predict liver histology. These include simple laboratory tests, such as serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT), prothrombin time (PT) or international normalized ratio (INR), albumin, and bilirubin either alone or in combination. However, because a high proportion of coinfected patients can have a normal ALT at the time of liver biopsy despite having significant histologic abnormalities, especially in those with HCV compared with those with HCV monoinfected individuals, serum liver enzymes alone may not be that useful. Because most data on noninvasive assessment of liver disease severity and fibrosis are from patients with HIV/HCV coinfection with far fewer studies of patients with HIV/HBV and HIV/NAFLD, we will focus on the HIV/HCV-coinfected population.

Given the risks of liver biopsy, paired with the need to assess liver fibrosis in those with HIV and abnormal liver enzymes, particularly in the HIV/HCV-coinfected population, there has been increased evaluation of noninvasive markers of fibrosis in the coinfected population. Various biochemical markers and imaging modalities have been validated in the HCV monoinfected population, but require further understanding of limitations in those with HIV.