HAV

HAV is a single-stranded RNA virus in the family Picornaviridae. For decades it has been associated with outbreaks of "infectious hepatitis" stemming from fecally contaminated food and water, in particular, in countries with poor sanitary conditions (34). However, infections in industrialized nations are not infrequent, and often occur in outbreaks related to unsanitary food handling in restaurants (5). The clinical course of HAV infection ranges from asymptomatic infection, particularly in children, to fulminant hepatic failure necessitating liver transplantation. Most adults are symptomatic, with jaundice occurring in 70%, but infections are usually mild and self-limited; no chronic carrier state has been identified (61). However, the elderly (5, 61) and those with preexisting liver disease, particularly chronic HBV and HCV infection (58, 133), apparently are at an increased risk for severe hepatitis and death if superinfected with HAV. No effective treatment, other than supportive measures, is available for HAV infection. Vaccines are commercially available, effective, and well tolerated (59). International travelers, military personnel, and certain high-risk populations, including homosexual males and patients with chronic liver disease, have been targeted for vaccination (5).

HBV

HBV is a relaxed, circular, partially double-stranded DNA virus from the family Hepadnaviridae (26). Since its discovery nearly 35 years ago as the "Australia antigen" by Blumberg et al. (16), its virologic and clinical features have been investigated extensively. HBV DNA has four partially overlapping, open reading frames that encode: (i) the viral envelope proteins [including hepatitis B surface antigen (HBsAg)]; (ii) the nucleocapsid including the hepatitis Be (HBeAg) and core antigens (HBcAg) from the precore and core regions; (iii) the polymerase that has reverse transcriptase, DNA polymerase, and RNase activities involved in HBV replication; and (iv) the X protein that is a potent transactivator and may play a role in hepatocarcinogenesis (76). Like HAV, acute HBV infection may be entirely asymptomatic, but 30% of patients have icteric hepatitis and some develop acute liver failure (26). The clinical features of chronic HBV infection include an asymptomatic carrier state, chronic hepatitis, cirrhosis, and HCC (26). The natural history of chronic infection varies with the time course of infection and host factors including age, gender, and race (36). When acquired perinatally or during infancy and childhood, either from an infected mother or close contact who is infected, as occurs in high-prevalence areas such as Southeast Asia and Africa, chronic infection ensues in more than 90% of individuals (26). These patients have a high risk of progression to cirrhosis and complications such as liver failure and HCC (36). When acquired in adulthood, usually by way of injection drug use or high-risk sexual practices, as occurs most often in North America, less than 5% of patients progress to chronicity, the remainder clearing HBsAg and remaining free of complications (26). Patients with chronic hepatitis B, and markers of active viral replication (including HBeAg and HBV-DNA positivity) and hepatic damage (including elevated liver enzymes and active necroinflammatory lesions on liver biopsy) are candidates for antiviral therapy. Currently, two forms of therapy are available: interferon alpha and the nucleoside analog lamivudine. Unfortunately, neither treatment is perfect; long-term response rates are only in the range of 20 to 30% (76). Again, as in HAV infection, effective vaccines are available and have proven effective in reducing the incidence of acute and chronic hepatitis (76, 80), and the incidence of childhood HCC in Taiwan (27), a major medical advance.

HCV

HCV is a single-stranded RNA virus in the family Flaviviridae, transmitted via the parenteral route. The virus has six major genotypes (1 to 6) with heterogeneous genetic sequences, and major differences in responsiveness to antiviral therapy. In North America and Western Europe, for example, HCV genotype 1 is found in approximately 70% of individuals and is notoriously resistant to interferon therapy (68). HCV causes acute icteric hepatitis in roughly 25% of individuals. Its major feature, however, is its propensity to cause chronic hepatitis in more than 80% of those infected (4). The natural history of chronic hepatitis C is generally characterized by slow histologic progression (98). Most patients remain asymptomatic or have nonspecific symptoms, and ultimately die of an unrelated condition (109, 110). Approximately 10 to 20% of individuals, however, progress to cirrhosis during a 20-year time span, and are at risk for complications including end-stage liver disease and HCC (98). This progressive fibrotic process is accelerated by factors including older age at infection, male gender, and alcohol consumption (96, 99). Thus far, because of the lack of small animal models and an effective cell culture system for HCV, as well as the genetic heterogeneity of the virus, development of a vaccine has been unsuccessful (69). Effective treatment, however, is available; long-acting, pegylated interferon in combination with ribavirin led to sustained clearance of HCV RNA from serum in approximately 60% of individuals (46, 87, 97, 146).

HDV

HDV is a defective RNA-containing passenger virus of the family Deltaviridae that requires the helper functions of HBV, including provision of the HBsAg coat, for virion assembly and penetration into hepatocytes. Transmission is by the same routes as HBV, mainly parenteral contact. The prevalence of HDV varies worldwide in correlation with that of HBV (101). With declining rates of HBV infection related to vaccination, the prevalence of HDV has also declined (41). The complex interaction of these two viruses is discussed below. The natural history of chronic HDV infection varies widely, but it has been linked to rapidly progressive liver disease (35, 38, 100, 106). In endemic areas, most individuals contract HDV in their teenage years or early adulthood. Although most develop cirrhosis within a few years, the majority remain stable for two to three decades until they present with features of decompensation including esophageal variceal bleeding, ascites, and HCC (38, 101). The only approved therapy for chronic HDV is interferon alpha, but this leads to sustained clearance of the virus in less than 20% of non-cirrhotic individuals (77, 103).

HEV

Hepatitis E virus (HEV) is a nonenveloped RNA virus of the family Caliciviridae. It is transmitted via the fecal-oral route, usually by contaminated water, and has been implicated in outbreaks and occasional sporadic cases of acute hepatitis in Southeast and Central Asia, Africa, the Middle East, and Mexico (34). In most cases, the clinical illness is similar to that of other forms of acute viral hepatitis. In pregnant women, however, acute HEV infection has been linked to serious illness with a casefatality rate of nearly 25%. No chronic carrier state exists (34). The impact of HEV superinfection in patients with other hepatotropic viruses has rarely been studied, but it is not known to cause severe illness in these patients. In an Israeli study of 188 hemophiliacs, the prevalence of antibodies to HEV was 9% (7). The anti-HEV-seropositive hemophiliacs had the same seroprevalence of antibodies to HBV, HCV, and HIV, and the same number of cases with chronic hepatitis, as among those negative for anti-HEV. Because HEV is rare in North America and Europe, and it seems to have little impact on the course of other hepatotropic infections, HEV infection is not discussed further.

HGV

Hepatitis G virus (HGV) is a recently discovered, positive-stranded RNA virus that is a member of the family Flaviviridae (2, 75). HGV is transmitted parenterally (3), sexually (57), and perinatally (134). The prevalence of HGV infection is 1.4% among American blood donors, and approximately 10% among recipients of blood transfusions from the seventies to nineties (3). Since its discovery, numerous reports have attempted to link HGV to acute and chronic hepatitis (3), cryptogenic cirrhosis (28), and even HCC (74), but none have been convincing. Furthermore, as discussed, HGV infection apparently does not alter the clinical features or course of other hepatotropic infections (22, 86, 115, 124). HGV RNA levels decrease in many individuals receiving interferon treatment for coexistent HBV or HCV infection, but the results are transient (22, 67, 86). Ribavirin appears to have only a minimal antiviral effect (67).

HBV-HCV Coinfection

Since the identification of HCV in 1989 and development of an accurate serologic screening test (29, 64), interactions between HBV and HCV in coinfected individuals have been investigated extensively. Coexistent HCV infection is estimated to occur in 10 to 15% of patients with chronic hepatitis B, depending on the risk factors of the population studied (40, 93). Conversely, the prevalence of chronic hepatitis B in patients with hepatitis C is lower because of the greater propensity of HBV to clear spontaneously. In our own series of 3,546 French HCV-infected patients, only 119 (3.4%) were HBsAg-positive. Among these 119 patients, 34 (29%) were HIV-positive (T. Poynard, unpublished observations).

HBV-HCV coinfection is characterized by "viral interference," whereby the replication of one virus is suppressed by another (56, 93–95, 104). Early on, Brotman and colleagues reported this phenomenon in an animal study of 19 chimpanzees administered infected serum (20). In this study, all the chimpanzees exposed to HBV alone developed HBsAg positivity and biochemical hepatitis within 4 to 9 weeks after inoculation. On the contrary, only three of the seven animals inoculated with NANB and HBV simultaneously developed liver enzyme abnormalities, and only five became positive for HBsAg. The onset of antigenemia was significantly delayed, and the duration of antigenemia and the clinical illness was significantly shorter in coinfected chimpanzees. The authors concluded that concurrent NANB infection had likely interfered with HBV replication and accounted for the apparent amelioration of the serologic and clinical features of HBV infection (20). These studies have been duplicated in other laboratories (21, 53, 132).

In most cases observed clinically, one virus remains dormant while the other replicates actively (72, 112). Most often, HCV replication is dominant, such that HBV DNA levels in serum (31, 56, 62, 72, 93, 104, 107) and liver (71, 142) are lower in coinfected individuals than in those with HBV alone. HCV genotype 1 may be particularly efficient in this regard (93, 94), but results have been inconsistent (56, 81). Several studies have also shown a lower frequency of HBV precore mutant strains (those unable to express HBeAg) in patients with dual (and triple) infection (56, 143). This may relate to an HCV-induced reduction in HBV synthesis because mutagenesis is directly related to the rate of HBV replication. The failure, however, to find a lower prevalence of precore mutants in HDV coinfection (56), which is also associated with a reduction in HBV replication (56, 81, 141), does not support this hypothesis. An alternative hypothesis, supported by an immunologic study assessing the proliferative responses of peripheral blood mononuclear cells to viral antigens (131), suggests that host immune responses react more vigorously to HCV antigens than those of HBV (56). The hitherto reduction in "selective pressure" exerted on HBV may explain the lower prevalence of precore mutant strains.

With respect to the replication of HCV, RNA levels tend to be lower in coinfected patients than in those with isolated HCV infection, suggesting that HBV replication exerts a counteracting inhibitory effect on the replication of HCV (56, 62, 81, 93, 104, 145). In some cases, HCV RNA is at very low levels or undetectable, and HBV is the dominant replicating virus (62). Furthermore, in other cases, patients may show alternating appearance of HBV DNA or HCV RNA with concomitant disappearance of the other virus (72). The role of host immunocompetence (e.g., organ transplantation or HIV infection) in affecting these interactions is unclear, although a recent study showed that coinfection with HIV had no impact on nucleic acid levels in HBV-HCV or HBV-HDV-HCV-coinfected patients (56).

The precise mechanism for these interactions is unknown, although it appears that several mechanisms involved in viral processes and host humoral and cellular immune responses can be implicated. For example, possible alterations in the mechanisms responsible for virus absorption, penetration and/or replication have been suggested (21). The induction of soluble mediators, including interferon, has been considered, but this cytokine was not detected in the serial serum samples of chimpanzees infected with NANB hepatitis (23). Other potential mediators include tumor necrosis factor alpha and interleukin 6, which appear to activate certain intracellular pathways that can down-regulate HBV gene expression (44). Another hypothesis relates to the subcellular localization of the HBV core protein, which is regulated by the cell cycle. It has been suggested that liver injury and the subsequent cell renewal induced by coexistent HCV infection may increase the expression of HBV epitopes on the hepatocyte surface (72). Subsequent T-lymphocyte-mediated elimination of HBV-infected hepatocytes may thus account for the reduced HBV DNA levels observed in serum and liver tissue. Finally, in a cotransfection study using a human hepatoma cell line (HuH-7), Shih and colleagues implicated the HCV core protein specifically in the down-regulation of the processes of transcription and encapsidation of HBV pregenomic RNA (113). In this study, the secretion of HBV viral particles, including the nucleocapsid and mature virion, was suppressed 20-fold in the presence of HCV structural genes. The authors suggested that the HCV core protein mediates these effects via gene-regulatory functions (113). Pontisso and colleagues showed that the core protein of HCV genotypes 1 and 3 (not genotypes 2 and 4) share sequence homology with the HBV core protein (95). This potentially explains reciprocal inhibition of these viruses and the observation in some studies that HCV genotype 1 is a more potent inhibitor of HBV replication than the other HCV genotypes (93, 94). In light of all the available evidence, it seems that no single process can explain viral interference.

The impact of coinfection has been examined in the acute and chronic settings, and from the perspectives of both HBV and HCV in isolation. As in the chimpanzee experiments, in patients with chronic hepatitis B, HCV superinfection apparently increases the rate of HBsAg seroconversion by inhibiting HBV replication (71, 111). In doing so, HCV usurps HBV as the predominant cause of liver disease in these patients, as evidenced by liver biopsies showing HCV RNA, not HBV DNA, and histologic features compatible with chronic hepatitis C (71). In the setting of acute HBV-HCV coinfection, which has been reported in injection drug users and transfused patients, results have been discordant. An increased risk of acute liver failure has been reported (39, 140); however, others have described an attenuating effect of acute HCV on HBV-mediated hepatic damage (88). In one study, acute HBV-HCV coinfection was associated with a delay in the onset and a shortened duration of HBsAg seropositivity, and with lower peak aminotransferase levels (88).

In the chronic setting, HBV-HCV-coinfected patients seem to have more severe liver disease than those with isolated HBV infection (31, 40, 93, 104). For example, in a study of 148 patients with chronic hepatitis B, of whom 16 (11%) had coexistent HCV infection, the incidence of cirrhosis (44 versus 21%) and hepatic decompensation (24 versus 6%) were significantly higher in coinfected patients (40).

Similar results have been reported in studies using controls with isolated HCV infection (81, 104, 136, 145). In one study, 23 patients with chronic HBV-HCV coinfection were compared with 69 age- and sex-matched patients with hepatitis C, but negative for HBsAg (145). Although epidemiological, biochemical, and virologic parameters were the same between the groups, the prevalence of cirrhosis was significantly higher in those with dual infection. Among coinfected patients, HCV RNA levels were significantly lower in HBV DNA-positive patients, and histologic lesions (including piecemeal necrosis, fibrosis, and presence of cirrhosis) were more severe (145). We found similar results in a case-control study of 34 HBV-HCV-infected patients and 34 patients with isolated HCV infection matched on the basis of age, gender, alcohol consumption, and duration of HCV infection (81). In this study, serum HCV RNA levels were lower and more frequently undetectable [particularly in patients with serological markers of active HBV infection (HBeAg and/or HBV DNA)], fibrosis was greater, and a trend toward an increased prevalence of cirrhosis occurred in coinfected patients. Thus, it appears that HBV exerts an inhibitory effect on HCV replication, but contributes to accelerated liver injury in coinfected patients.

Because both viruses have been implicated in hepatocarcinogenesis, the influence of dual infection on the development of HCC has also been studied. HBV and HCV apparently act synergistically in this regard (13, 19, 144), in particular, with respect to the infiltrating variant of HCC (14). In a prospective study of 290 consecutive Italian patients with cirrhosis of varying etiologies, the cumulative incidence of HCC was 11% in HBsAg-positive cases, 9% in HCV-infected patients, and 36% in those with HBV-HCV coinfection (13). The significant effect of coinfection on the progression to HCC persisted in multivariate analysis in addition to the consistently identified risk factors of age and male gender. In another study from South Africa involving 231 black patients with HCC and matched controls, HBsAg positivity was associated with a 23-fold increase in the risk of HCC, whereas anti-HCV positivity was associated with a 6.6-fold risk (60). When both markers were present, the relative risk rose to 82.5; dual infection was estimated to be the cause of 20% of HCC in this population. The mechanisms behind this synergism are unclear as are those underlying the hepatocarcinogenic potential of these viruses in isolation. It has been hypothesized that the oncogenic effects of HBV and HCV are related to chronic liver injury, which sets up a cascade of events including the increased rates of DNA synthesis necessary for cellular repair. This may set the scene for acquired DNA mutations that predispose to the development of HCC. If this is the case, HBV-HCV coinfection may exert its carcinogenic effect simply through enhanced hepatocellular injury as evidenced by the histologic studies. We also know that HCC is much more common in individuals with cirrhosis than those without cirrhosis. The synergism of HBV and HCV in causing HCC may occur via the propensity of coinfection to accelerate progression of fibrosis. Alternatively, the "hit" may occur after the development of cirrhosis, in the transition from cirrhosis to HCC. Whether the temporal course of HBVHCV coinfection (i.e., simultaneous acquisition of both viruses, or superinfection of HBV in a patient with chronic hepatitis C, or vice versa) is a factor is also unclear.

HBV-HDV Coinfection

As described previously, HDV infection occurs only in the setting of coexistent HBV infection, acquired either simultaneously (coinfection) or through superinfection of a HBsAg-positive patient. The clinical course and virologic features differ markedly between these scenarios, but both are characterized by a complicated interplay between the two viruses. In HBV-HDV coinfection, acute hepatitis is usually icteric and often fulminant, likely more so than in HDV superinfection. Both viruses seem to replicate actively, but often in a transient manner (101). In fact, acute coinfection may have a biphasic nature with peaks of serum aminotransferases separated by several weeks; the first peak is typically related to active HBV replication, and the second to HDV replication (6). In milder cases of acute coinfection, HDV replication is lower and replication of HBV is suppressed by HDV (101). In fact, early suppression of HBV replication may be associated with a reduction in the synthesis of HBsAg. In such cases, the acute illness may be diagnosed as "non-A-E hepatitis" unless serologic markers for HDV are tested specifically (24). Most cases of coinfection resolve spontaneously; in fact, only 2% of coinfections evolve to chronic disease (102).

Cases of HDV superinfection are usually symptomatic and often lead to liver failure in patients with significant preexisting chronic hepatitis B (101). In general, HBV replication is suppressed markedly by superinfection with HDV, often to the point that HBV DNA is undetectable even by sensitive PCR-based assays. In a few fortunate patients, HBV replication is suppressed to the point of HBsAg clearance; that is, the HBV infection (and, therefore, the HDV infection) is aborted. In some, this inhibition is transient and accompanied by rising titers of antibody to HBsAg (anti-HBs), simulating resolution of HBV infection, only to `relapse' when HDV replication diminishes after the acute illness. Unlike HDV coinfection, progression to chronicity occurs in approximately 90% of patients with HDV superinfection (101).

Most of the available literature suggests that chronic hepatitis D is a rapidly progressive disease, more so than isolated HBV infection (35, 38). Approximately 80% of individuals develop cirrhosis within 5 to 10 years of infection (35, 100, 101). Patients in whom both viruses are actively replicating seem to be at a particularly high risk of progression (106). Most patients with cirrhosis, however, remain stable until their fourth or fifth decade at which time they may show evidence of hepatic decompensation—one to two decades earlier than the typical patient with isolated HBV or HCV infection (35, 100, 101). Wu and colleagues investigated the virologic correlates of disease progression in 185 HBV-HDV-coinfected patient with use of sensitive PCR-based assays (141). As described above, acute HDV superinfection was generally characterized by active HDV replication and HBV suppression with high alanine aminotransferase levels; during the chronic phase, HDV replication decreased and HBV synthesis reactivated with moderate alanine aminotransferase levels; and in the late phase, development of cirrhosis or HCC was associated with the replication of either virus, and disease remission with quiescence of both viruses (141).

HBV-HCV-HDV Coinfection

The relationship between the hepatotropic viruses in cases of triple infection is interesting; as in other hepatotropic coinfections, viral interference is prominent. We addressed this issue in a study of 16 patients with triple infection and 16 matched controls with isolated HCV infection (81). In patients with triple coinfection, HDV emerged as the dominant virus; 88% had markers of active HDV infection (HDV RNA, hepatitis D antigen, or IgM anti-HDV positivity). HCV RNA, on the other hand, was detectable in only 2 of 16 (12.5%) coinfected patients versus all the HCV-infected controls. Consequently, mean HCV RNA levels were significantly lower in the former group. Likewise, HBV replication, as indicated by the presence of HBeAg or detectable HBV DNA, was observed in only 27% of coinfected patients. A similar study from Spain confirmed the suppressive effect of HDV on HCV replication (56). In addition, using a control group of patients with isolated HBV infection, these authors documented suppression of HBV replication in patients with triple infection. In this study, the inhibitory influence of HDV was stronger on the replication of HCV than that of HBV (56). In terms of liver histology, necroinflammatory activity and fibrosis scores were significantly higher in the triple infection group, and the prevalence of cirrhosis was higher (58 versus 6%; P = 0.004) in our study (81).

Similar results have been reported in the post-liver transplantation setting (125). In a retrospective analysis of 13 HBV-HCV-coinfected patients undergoing transplantation for end-stage liver disease, 5 patients were also infected with HDV. After liver transplantation, HCV RNA was positive in all of the dually infected patients, but none of the patients with triple infection. With respect to liver damage, however, serum aminotransferases (admittedly a surrogate marker) were elevated in 88% of dually infected patients, but only 20% of those with triple infection. The authors concluded that HDV suppresses HCV replication and modulates the liver inflammatory process after liver transplantation in patients with HBV-HDV-HCV coinfection (125).

These results disagree with those of a Taiwanese group who reported that HCV, not HDV, dominates in patients with triple infection (73). In this study of 60 coinfected patients, HCV RNA was detectable in 70% of patients, HDV RNA in 23%, and HBV DNA in 13%. The reason for the discrepancy between these studies may relate to the study populations in terms of the timing of infection. Whereas the Taiwanese patients presumably acquired their infections perinatally or during early childhood, most of the patients in our study were infected as adults after blood transfusion or injection drug use. Differences in host immune responses and viral strains must also be considered. Whereas HDV genotype 1 predominates in Western Europe, HDV genotype 2 is more common in East Asia (56).

Coinfection with HGV in Chronic Hepatitis B and C

Coinfection with HGV apparently exists in 10 to 20% of patients with chronic hepatitis C (22, 86, 115, 124). Both agents are transfusion-transmissable, accounting for their high rate of association (3). Most reports suggest that HGV has a minimal impact, if any, on the virologic, biochemical, and histologic features, as well as clinical course, of HCV infection (22, 67, 115). In the largest such study thus far reported examining 671 patients with chronic hepatitis C, 65 patients (9.7%) were coinfected with HGV (115). In this study, no significant differences occurred in liver biochemistry, hepatic necroinflammation or fibrosis, or HCV genotype genotype and viral load in HGV-infected versus uninfected controls. In this and other studies (22), patients with HGV had a shorter mean duration of infection with HCV, suggesting that HGV is a relatively recently introduced infection, at least in North American patients. Because most studies have shown no overall effect of HGV on liver fibrosis, this latter finding suggests that the rate of fibrosis in chronic hepatitis C may actually be increased by concurrent HGV infection. No studies, however, have carefully examined this issue by controlling for other factors, such as age, gender, and alcohol consumption, which influence the rate of fibrosis progression in this disease (96, 99).

The response to treatment, including antiviral therapy and liver transplantation, has also been examined in HCV-HGV-coinfected patients. Again, the impact of HGV coinfection apparently is minimal. The response to interferon therapy is unchanged (22, 67), and HGV status has no impact on patient or graft survival, or the recurrence of histological hepatitis in HCV-infected patients after liver transplantation (15, 22). In one study (22), HGV RNA levels decreased in all patients after transplantation. One possible explanation for this is that HCV may inhibit HGV replication, as observed in other forms of coinfection, because HCV RNA levels are know to climb substantially after transplantion (128). The failure to find any significant correlation between HGV and HCV RNA levels before transplantation, however, argues against this hypothesis and suggests that perhaps immunosuppressive therapy may play a role (22).

Not surprisingly, given their common modes of transmission, HGV infection is also frequent in patients with chronic hepatitis B, particularly those with HCV and/or HDV coinfection (mainly injection drug users) (37). In a study of 125 patients with chronic hepatitis B, 82 asymptomatic HBsAg carriers, and 103 healthy controls, the seroprevalence of HGV was 17, 16, and 17% (P = not significant) respectively (37). HGV coinfection apparently had no effect on the clinical course of the patients with chronic hepatitis B.

Superinfection with HAV in Chronic Hepatitis B and C

As discussed previously, the spectrum of illness caused by acute HAV is highly variable. Acute liver failure caused by HAV is uncommon, but often fatal, particularly in the elderly (61) and patients with a history of chronic hepatitis B or C (58, 133). In general, these individuals do not seem to be at an increased risk of acquiring HAV infection, although outbreaks of HAV have been reported in populations of injection drug users, a large proportion of whom are infected with HBV and HCV (49). Such epidemics presumably are related to the use of contaminated drug paraphernalia onto other behaviors such as high-risk sexual practices (49). The risk of HAV superinfection in patients with chronic hepatitis B is highlighted by the Shanghai epidemic of 1988 during which more than 300,000 individuals acquired HAV through the consumption of contaminated, raw shellfish (58). A retrospective analysis of this epidemic reported that the risk of death was 5.6-fold higher in patients with chronic HBV infection than in the general population (58). In the same study, 115,551 cases of HAV infection reported to the Centers for Disease Control between 1983 and 1988 were reviewed (58). In this analysis, the casefatality rate was 58.5-fold greater in those with chronic HBV infection than in those without preexisting liver dysfunction.

Evidence linking acute HAV to more severe clinical outcomes in those with chronic hepatitis C is more controversial. In an alarming study from Italy, Vento and colleagues reported the course of 17 patients with chronic hepatitis C who acquired acute HAV infection (133). In this study, 41% of those infected developed acute liver failure and 86% of these individuals died, representing a case-fatality rate of 35%, approximately 45 times greater than the general population. These findings, however, have not been duplicated in other studies including several describing epidemics of HAV infection among intravenous drug abusers (49). Furthermore, retrospective analyses of cases of severe HAV infection have not revealed revealed an increased prevalence of hepatitis C in affected patients (51).

In view of the potential severity of acute HAV in patients with preexisting liver disease, it is now recommended that all persons with chronic liver disease be vaccinated against HAV (5). Several HAV vaccines are commercially available and have been effective and well tolerated in patients with chronic hepatitis B and C (59). Indeed, vaccination has become the standard of care in these patients. In light of the relative rarity of acute HAV and the frequency of chronic hepatitis C in North America, and the high cost of available vaccines, we examined the cost-effectiveness of this recommendation in patients with chronic hepatitis C by using a decision analysis model (91). In this relatively low-prevalence area, routine vaccination of HCV-infected patients cost more than $21,000,000 (in 1998 Canadian dollars) per HAV-related death prevented. This expense is clearly in excess of traditional measures of cost-effectiveness particularly in today's era of finite healthcare resources. The feasibility of this approach in high-prevalence areas remains to be determined.

Coinfection with HIV in Chronic Hepatitis B and C

Although HIV is not a hepatotropic virus per se, coinfection with HIV and the hepatitis B and C viruses deserves mention for several reasons. First, coinfection occurs frequently because of common routes of transmission. In a French cohort study examining 1,935 HIV-positive patients, the prevalence of antibodies to HCV was 42.5%; this rose to 86.1% in patients with parenterally acquired HIV infection (predominantly injection drug users) (105). In the same cohort, the prevalence of antibodies to HBcAg, signifying prior exposure to HBV, was 56.4%; 6.9% had residual, chronic hepatitis B as evidenced by HBsAg positivity. Conversely, in our own cohort of 3,546 consecutive patients with chronic hepatitis C, 396 (11%) were positive for anti-HIV antibodies (Poynard, unpublished). Second, with the introduction of highly active antiretroviral therapy (HAART) in the nineties, the prognosis for HIV infection has improved significantly (90), such that chronic liver disease has emerged as an important cause of morbidity and mortality in these patients (89, 121). For example, in a retrospective analysis from an HIV/AIDS institution in Madrid, 9% of 1,670 consecutive admissions were attributable to end-stage liver disease (121). HCV, alone or in combination with other hepatotropic viruses, was responsible for 89% of these admissions; and liver-related deaths occurred in 15 individuals, representing 5% of all cases of in-hospital mortality. Finally, emerging evidence suggests that HIV and the hepatotropic viruses exert reciprocal effects that have negative implications for the respective diseases (45, 48, 92, 120, 123). The impact of the hepatotropic viruses on the progression of HIV-related disease is beyond the scope of this chapter, but the interested reader is referred elsewhere (45, 48, 92).

The Impact of HIV on HCV Infection

Most studies linking HIV infection to more severe disease in patients with hepatitis C have assessed virologic and histologic parameters. With respect to HCV virology, although results are conflicting, several studies suggest that HCV viremia is increased in the presence of HIV coinfection (8, 32, 83, 123, 127). Furthermore, HCV viral load seems to increase as HIV progresses (43, 83). It has been hypothesized that these findings relate to faulty cytotoxic (CD8⁺) lymphocyte responses directed against HCV-infected hepatocytes and the HCV itself in patients with HIV coinfection (122). Paradoxically, although HAART leads to substantial reductions in HIV viral load and reciprocal increases in CD4⁺ lymphocyte counts, this therapy apparently does not have an impact on HCV viremia (82). The explanation for this discordance in associations has yet to be elucidated.

Somewhat less controversial is the impact of HIV coinfection on hepatic histology in HCV-coinfected patients. HIV infection clearly intensifies the degree of hepatic necrosis and inflammation and accelerates the fibrotic process that ultimately leads to cirrhosis and complications such as HCC, liver failure, and death (1, 9, 42, 123). We have studied the effect of HIV coinfection on hepatic pathology in a cohort of patients with chronic hepatitis C monitored at our institution (1, 9). HIV-HCV-coinfected patients were matched with HCV-infected controls for risk factors known to have an impact on liver fibrosis progression in hepatitis C (including age at infection, duration of infection, alcohol consumption, and gender) (96, 99). In a study of 58 HIV-HCV-coinfected patients and 58 HCV-infected controls, we found significantly higher scores for overall necroinflammatory activity, piecemeal necrosis, and a trend toward an increased frequency of cirrhosis in coinfected patients (1). When patients with advanced HIV disease (as defined by a CD4⁺ count below 200 cells/μl) were compared with HIV-negative patients, the prevalence of cirrhosis was significantly higher in coinfected patients (10 versus 45%; P = 0.003). Other authors have duplicated these findings; in a Spanish series of injection drug users, the incidence of cirrhosis after 10 years of HCV infection was 14.9% in coinfected patients versus only 2.6% in those who were HIV negative (123).

In another study, we examined more rigorously hepatic fibrosis in a cohort of 244 coinfected patients in an attempt to identify risk factors for fibrosis progression in this population (9). In this study, the mean rate of fibrosis progression was significantly greater in coinfected patients than in well-matched, HCV-infected controls (0.181 versus 0.135 Metavir fibrosis units/year; P < 0.0001). At these rates of fibrosis progression, the median duration from HCV infection to cirrhosis was 26 years in HIV-infected patients versus 38 years in non-HIV-infected patients. With use of multivariate analysis, we identified alcohol consumption (>50 g/day), CD4⁺ lymphocyte count (≤200 cells/μl), and age at infection (>25 years) as being independently associated with fibrosis progression in this population. As an example, an HIV-infected patient who has less than 200 CD4⁺ cells/μl at the time of liver biopsy and drinks more than 50 g of alcohol daily could be expected to develop cirrhosis in only 16 years, whereas an HIV-infected patient with more than 200 CD4⁺ cells/μl and drinking less alcohol would have an expected time to cirrhosis of 36 years (9). These results obviously have important implications for HIV-HCV-coinfected patients; alcohol consumption should be strongly discouraged, and antiviral therapy against HIV should be maximized in an attempt to maintain, and hopefully restore, the immune system.

With respect to the impact of antiviral therapy, per se, we recently investigated the role of protease inhibitors (PI) in the prevention of fibrosis progression in HIV-HCV-coinfected patients (12). In this study, PI therapy was associated with reduced necroinflammatory activity and a 4.7-fold reduction in the rate of progression to cirrhosis. After 15 years of follow-up, the actuarial rate of cirrhosis was 18% in PI-naïve patients versus only 5% in those who had received a PI. This protective effect was independent of CD4⁺ lymphocyte count and HIV viral load, and not seen with other agents known to be effective against HIV, such as the nucleoside reverse transcriptase inhibitors (12).

Several explanations have been proposed for the damaging effects that HIV infection has in HCV-coinfected patients. First, as alluded to previously, HCV viremia may be increased in HIV-infected patients, particularly those with an advanced stage of HIV disease (8, 32, 83, 123, 127). Greater HCV replication may enhance the cytopathic effect of this virus on hepatocytes. It is most frequently suggested that the immunodeficiency associated with HIV infection is responsible for this phenomenon because of blunted T-lymphocyte responses (122). However, the failure to document a consistent relationship between HCV and HIV viral loads and CD4⁺ lymphocyte counts, the absence of an association between HCV viral load and hepatic damage in the non-HIV-infected population, and our own observations with respect to HAART, would argue against this hypothesis. Alternatively, HIV and HCV may interact directly within lymphocytes or hepatocytes, perhaps through the expression of various cytokines or adhesion molecules, which serve to alter the pathogenicity of HCV (1). Finally, it is known that HIV infects Kupffer cells and sinusoidal endothelial cells, but the importance of this interaction is unclear (54, 108). Kupffer cells are involved in the activation of hepatic stellate cells which are responsible for hepatic fibrosis. Likewise, an interaction between HIV and endothelial cells lining the sinusoids may be involved in perisinusoidal fibrosis, which is more frequent in coinfected patients (1).

Studies examining the impact of HIV coinfection on clinical outcomes in patients with chronic hepatitis C have revealed conclusions similar to those described above for virologic and histologic endpoints; HIV apparently has a negative impact on the morbidity and mortality of coinfected patients (33, 47, 126). In the largest reported study to our knowledge, which examined a cohort of 4,865 British hemophiliac males exposed to blood products at high risk of HCV contamination monitored between 1969 and 1993, the risk of death from liver-related causes was nearly fivefold higher in HIV-infected versus noninfected patients (33). Compared with the general population, liver-related deaths were nearly 94 times higher among male hemophiliacs with HIV (33). In another study of 255 hemophiliacs, liver failure was reported in 10.8% of HCV-infected patients 20 years after first exposure; the risk in HIV-HCV-coinfected patients was 21-fold higher than in those with HCV alone (126). Finally, a recent meta-analysis confirmed the detrimental effect of HIV infection on the clinical course of hepatitis C (47). In this review, the relative risks of progression to cirrhosis and hepatic decompensation were 2.07 (95% confidence interval, 1.40 to 3.07) and 6.14 (95% confidence interval, 2.86 to 13.20), respectively, in HIV-infected versus uninfected patients with chronic hepatitis C.

The Impact of HIV on HBV Infection

The outcome of chronic hepatitis B depends on the interaction between HBV, which is not directly cytopathic, and the immunologic response of the host to this virus (26). Not surprisingly, then, coinfection with HIV, which has a profound effect on host immunity, modifies the phenotype of chronic HBV infection. Most studies suggest that HIV has a permissive effect on HBV replication, that is, chronic hepatitis B patients with HIV have higher levels of HBV DNA (30, 45) and a greater proportion of hepatocyte nuclei which stain positive for HBcAg (30). Enhanced HBV replication may relate to decreased cytotoxic T (CD8⁺) lymphocyte activity because of impairment of helper (CD4⁺) T lymphocytes and monocyte function (30). Reports of exacerbations of chronic hepatitis B in HIV patients with immune reconstitution after the introduction of HAART lend further support to the importance of the immune system in this disease (25). Furthermore, patients with serologic markers of remote HBV infection may reactivate their disease as their immune system deteriorates because of advancing HIV infection (102).

Other features observed in HIV-HBV-coinfected patients deserve mention. In patients acutely infected with HBV, prior infection with HIV predisposes them to the development of the chronic HBV-carrier state (18, 114). These patients are more likely to express HBeAg (17), which has important implications for infectivity, and are less likely to have spontaneous seroconversion of HBeAg to anti-HBe (45, 63), an event that usually heralds clearance of the virus.

In the presence of HIV coinfection, serum aminotransferase concentrations tend to be lower, an effect that may increase with the progression of immunosuppression (17, 30, 45). Paradoxically, despite similar liver biochemistry, and liver biopsies that generally show no difference in the intensity of necroinflammation, those with HIV-HBV coinfection appear more likely to progress to cirrhosis (30). In fact, in one study of 132 homosexual men with chronic hepatitis B, the risk of cirrhosis was increased 4.2-fold in the presence of HIV infection, after controlling for confounders such as age, duration of HBV infection, HBeAg status, and alcohol consumption (30). It is possible that some patients progress to advanced fibrosis or cirrhosis with only minimal hepatic inflammation, as seen in other immunosuppressed populations such as transplant recipients (128).

HBV-HCV Coinfection

As discussed earlier, in most cases of HBV-HCV coinfection, one virus is dominant and the other is dormant with low or undetectable serum nucleic acid levels. In these situations, it is recommended that treatment be directed against the dominant virus with standard antiviral regimens (112). That being said, reports describing the efficacy of treatment in dually infected patients are limited. In one study of eight coinfected patients treated with interferon alpha (3 million units [MU] thrice weekly [TIW] for 6 months), liver enzymes normalized in only two patients, and HBsAg was lost in one (137). Long-term responses, including clearance of HBV DNA and HCV RNA, were not reported. In another study, 14 patients with detectable HBV DNA and HCV RNA were treated with interferon alpha 6 MU TIW for 6 months (50). Six months after the end of treatment, ten patients (71%) lost HBV DNA and four patients (29%) lost HCV RNA. In addition, all three HBeAg-positive patients cleared HBeAg after a flare at an average of 45 days into treatment. This report suggests that interferon therapy is effective in coinfected patients, particularly with respect to HBV clearance. On the contrary, Liaw retrospectively analyzed the responses to interferon of 15 patients with chronic HBV who were subsequently shown by serologic testing of stored serum to have dual infection with HCV (72). Only one of these patients cleared HBeAg and HBV DNA. Interferon therapy in patients with dual infection may successfully clear one virus, only to lead to symptomatic reactivation of the persistent virus at a later date (72, 135). Clearly, on the basis of the divergent results reported, randomized controlled trials of antiviral therapy are needed in this patient population.

HBV-HDV Coinfection

Numerous therapies have been tested in the treatment of chronic HDV infection, but most have been ineffective or downright detrimental. Interferon alpha is most widely used, but results have generally been disappointing (101). In a meta-analysis of five randomized, controlled trials in a total of 87 treated patients, Malaguarnera et al. reported a complete response (defined as sustained normalization of alanine aminotransferase, loss of HDV viremia, and improved histology during short-term follow-up) in 18.4% of interferon-treated patients versus only 1.5% of controls (77). Unfortunately, at standard doses and durations of therapy, prolonged responses are rare because relapses commonly occur after cessation of treatment. As a result, long-term therapy with high doses of interferon remains the recommended approach (66). Unfortunately, although useful in isolated HBV infection, the addition of lamivudine has not proven helpful for the eradication of HDV (138). In the future, long-acting, pegylated interferons, which are administered once weekly, may prove useful, particularly with respect to patient compliance during longer courses of therapy.

HBV-HCV-HDV Coinfection

Reports describing treatment in patients with triple hepatitis virus infection are sparse. In a single study of seven HBV-HDV-HCV-coinfected patients, interferon therapy at a dosage of 3 MU TIW for 6 months led to sustained liver enzyme normalization in only one patient (137). Virologic responses were not reported. On the basis of this single study, no firm recommendations can be made for the treatment of these coinfected patients.

HCV-HIV Coinfection

No randomized controlled trials of anti-HCV therapy have been reported in HIV-infected populations. Several small, uncontrolled studies of interferon alpha monotherapy have been published (20, 79, 84, 118), however. In the largest of these trials, 90 HIV-HCV-coinfected patients with CD4⁺ lymphocyte counts above 200 cells/μl were treated with 5 MU of interferon TIW for 3 months (118). Responders were then treated for an additional 9 months with 3 MU TIW. Twelve months after treatment, a sustained response (defined as normalization of alanine aminotransferase and disappearance of HCV RNA from serum) was seen in 18 of the 80 patients (22.5%) who completed the study. This rate of response was no different from that observed in a control population of 27 HIV-negative patients (25.9%, P = not significant), and is consistent with those reported in HIV-negative patients (130). Two factors were independently associated with a response in HIV-positive patients: a CD4⁺ lymphocyte count above 500 cells/μl, and a baseline HCV viral load below 10⁷ copies/ml (118). The impact of baseline fibrosis on liver biopsy, which has an important impact on antiviral therapy in HIV-negative patients (87, 97), was not assessed in this trial. Because HIV-HCV-coinfected patients tend to have more advanced hepatic fibrosis, this may prove to be an important factor in determining the efficacy of antiviral therapy in this population.

More recently, an uncontrolled pilot study of interferon and ribavirin combination therapy, currently the standard of care in hepatitis C treatment, was reported in 20 HIV-HCV-coinfected patients (65). Treatment consisted of interferon alpha-2b 3 MU TIW and ribavirin 1,000 to 1,200 mg daily for 6 months. At the end of treatment, serum HCV RNA was no longer detectable in 50% of patients (65). This rate is nearly identical with that observed in HIV-negative subjects treated for a similar duration in two large, randomized, controlled trials (87, 97). Although the rate of sustained virologic response was not reported, a recent report has suggested that the rate of relapse after cessation of treatment is similar in HIV-positive and -negative patients (119). Currently, several randomized controlled trials of combination therapy are ongoing in this patient population (Poynard, unpublished).

Treatment in both of the cited studies was well tolerated and seemed to have no effect on the progression of HIV infection (65, 117). This is an important point because previous reports have highlighted a rapid decline in CD4⁺ lymphocyte counts following the introduction of interferon therapy in 10 to 15% of HIV-infected patients (116). Furthermore, HIV-positive patients frequently have cytopenias related to HIV and its therapies which may be exacerbated by both interferon and ribavirin. Finally, recent in vitro studies have highlighted important interactions between ribavirin and anti-HIV medications including zidovudine and stavudine (52). These interactions may predispose to adverse drug reactions and/or an insufficient anti-HIV effect. The significance of these effects in vivo has yet to be realized (55).

We recommend that HIV-HCV-coinfected patients be treated in expert centers, preferably in the setting of randomized, controlled trials. On the basis of the existing literature, combination therapy with interferon and ribavirin apparently is the treatment of choice. Pegylated interferons, which are being evaluated in ongoing studies and have shown positive preliminary results, may prove to be more efficacious in the future. Such therapy requires the close monitoring of CD4⁺ lymphocyte counts and HIV viral loads, and the occurrence of treatment-related side effects. Alcohol consumption should be restricted in these patients because of its potent effect on the progression of hepatic fibrosis. Furthermore, because of the risk for severe hepatitis caused by superinfection with other hepatotropic viruses, and the frequent presence of risk factors that predispose to superinfection, these patients should receive HAV and HBV vaccination if they are susceptible. This vaccination preferably should be completed early in the course of HIV infection when immune responses to vaccination are maximal.

HBV-HIV Coinfection

The response to interferon therapy in HBV-HIV-coinfected patients is somewhat less promising than that reported in patients with HIV and HCV. Thus far, five uncontrolled studies of interferon therapy in dually infected patients have been published (78, 85, 129, 139, 147). In all, response rates have been modest with loss of HBeAg reported in fewer than 10% of patients. This rate is considerably lower than that reported in the HIV-negative patients enrolled in the same studies whose average response rate was 31% (76). Therefore, interferon seems to be a poor choice for therapy of HBV in the HIV-infected individual.

Another antiviral drug, lamivudine, has potent inhibitory effects on both HIV and HBV and has been tested in coinfected patients. In HIV-negative patients, lamivudine leads to rapid HBV DNA suppression to undetectable levels within a few weeks in most patients, and to HBeAg loss in 15 to 20% of patients after 1 year and nearly 40% after 3 years (70). The responses seen in HIV-positive patients seem lower, but prognostic factors and the duration of follow-up have not been taken into account. In a study of 57 patients treated with lamivudine for 1 to 5 years at our institution, HBV DNA suppression was observed in 47% of patients at 2 years, but this response dropped off dramatically to only 9% at 4 years with a 91% actuarial incidence of lamivudine resistance (10).

Trials of other nucleoside analogs, including adefovir dipivoxil in isolation and in combination, are currently underway and may hold promise for these patients who are at serious risk of morbidity and mortality (11). We evaluated the safety and efficacy of adefovir dipivoxil (10 mg daily) in an open-label trial for the treatment of lamivudine-resistant HBV infection in 35 HIV-infected patients (11). All patients had detectable serum HBV DNA despite lamivudine therapy, and mutations (M550V or M550I) in the HBV polymerase gene. Patients were treated for 48 weeks while continuing their existing anti-HIV therapy, including lamivudine. HIV disease in all patients was well controlled at screening, as assessed by HIV RNA levels (≤2.60 log₁₀ copies/ml). Mean decreases in HBV DNA serum levels from baseline (8.64 ± 0.08 [standard error] log₁₀ copies/ml) were 3.40 ± 0.12 log₁₀ copies/ml at week 24 (n = 31) and 4.0 ± 0.17 log₁₀ copies/ml at week 48 (n = 29) (P < 0.0001), respectively. Two patients underwent HBeAg seroconversion at weeks 32 and 36, respectively. Adefovir dipivoxil was generally well tolerated; no significant changes in either HIV viral load or CD4⁺ cell count were observed.