HIV-1, hepatitis B virus, and risk of liver-related mortality in the Multicenter Cohort Study (MACS)
"...results show an increased risk of death (8 times greater risk) from liver disease...in men coinfected with both viruses (HIV and HBV)...than men infected with HIV alone." Individuals with HBV and HIV had a 7 times greater risk of death before 1996 (HAART) and a trend towards a greater risk for death after 1996 than before 1996; the risk of death after 1996 was 11 times greater for coinfected individuals in this study than individuals with HIV alone. This data suggests HIV accelerates HBV progression, and this appears similar to HCV. A number of studies find that HIV accelerates HCV progression 2-4 times faster. Several studies suggest HIV impairs the immune response to HCV and this may also be true for HBV. In this study there is a trend that lower CD4 counts (<100) or alcohol consumption greater than 210 grams per week may accelerate progression. The authors suggest below several possible explanations for the trend of increased death rates for coinfected individuals after 1996 including longer duration of HBV, toxicity from HAART, restoration of immune responses by HAART, and stopping lamivudine (3TC) therapy. The study data did not indicate stopping lamivudine was associated with increased risk for death but a small number of individuals were available to evaluate this. There is controversy over whether HAART accelerates HBV or HCV progression. This question has not been adequately studied to offer a clear answer.|
Since HIV-infected individuals with HBV or HCV are at increased risk for liver related mortality, it is crucial to identify these individuals by testing. It is also important to show special attention to care, treatment, and selection of HIV HAART regimens for coinfected individuals. We need better testing and prevention programs for HBV and HCV in HIV communities.
HCV is more prevalent than HBV among HIV-infected individuals. It's estimated that 10% of HIV-infected individuals also have HBV. A high percentage of individuals exposed to HBV spontaneously clear it (80-90%). It's estimated 30% of HIV-infected individuals also have HCV and 60-90% of individuals who contracted HIV by IVDU have HCV. Both HBV and HCV can be transmitted by IVDU. IVDU and sex are important risk factors for transmission of HBV. The risk for sexual transmission is greater for HBV than for HCV. The risk for sexual transmission of HCV is low but studies suggest risk can be increased when certain risk factors which may facilitate blood-to-blood contact are present such as the presence of STDs and open sores, anal sex, rough sex, sex during menstruation. If a pregnant woman is chronically infected with HBV e antigen, 80-90% of the newborns become chronically infected. The risk for transmission of HCV from mother to child is about 5% but the risk increases to 17% if the mother has HIV. There is a vaccine to prevent HBV transmission but there is not a vaccine to prevent HCV transmission. Current drug treatment available for HBV includes lamivudine (3TC), adefovir, and interferon. Entecavir is a potent new drug for HBV in phase III studies. Several additional drugs for HBV are in earlier stages of development.
Summary of study: Although coinfection with HIV-1 and hepatitis B virus (HBV) is common, few long-term studies on liver-disease mortality in coinfected people have been undertaken. Our aim was to examine liver-related mortality among people at risk for HIV-1 and HBV infections. We used data from a multicentre, prospective cohort study to classify 5293 men who had sex with men, according to their HIV-1 antibody status, ascertained semiannually, and their hepatitis-B surface antigen status (HBsAg), which we ascertained at baseline. Mortality rates were estimated in terms of person-years and Poisson regression methods were used to test for significance of relative risks. 326 (6%) men were HBsAg positive, of whom 213 (65%) were HIV-1 positive. Of the 4967 HBsAg negative men, 2346 (47%) were infected with HIV-1. 3.8% of deaths were attributable to liver disease. Patients were 32 years of age at study entry and were followed for an average of 10 years. The liver-related mortality rate was 1.1/1000 person years, and was higher in men with HIV-1 and HbsAg (14.2/1000) than in those with only HIV-1 infection (1.7/1000, p<0⋅001) or only HBsAg (0.8/1000, p<0⋅001). In coinfected individuals, the liver-related mortality rate was highest with lower nadir CD4+ cell counts and was twice as high after 1996, when highly active antiretroviral therapy (HAART) was introduced.
Individuals coinfected with HIV-1 and HBV, especially those with low CD4+ nadir counts, are at increased risk for liver-related mortality, underscoring the importance of prevention, identification, and comprehensive management of hepatitis B in people infected with HIV-1. Lancet 2002; 360: 1921-26. Chloe L Thio, Eric C Seaberg, Richard Skolasky Jr, John Phair, Barbara Visscher, Alvaro Munoz, David L Thomas.
More than 95% of adults spontaneously recover from an acute HBV infection, an outcome that is defined by clearance of the hepatitis B surface antigen (HBsAg) from blood. Those persistently infected with HBV are at increased risk of development of cirrhosis and hepatocellular carcinoma over the ensuing decades. HIV-1 causes multidimensional immunosuppression, associated with
reduced frequency of spontaneous recovery from HBV infection. The effect of
HIV-1 infection on those infected with HBV is not well understood. Results of some studies suggest that the progression of liver disease is diminished, which might be expected since the pathogenesis of hepatitis B is believed to be immunologically mediated. Findings of other studies, however, suggest that progression of liver disease is increased, as is the case for hepatitis C virus (HCV).
The effect of HIV-1 infection on the natural history of hepatitis B could also be modified by highly active antiretroviral therapy (HAART), which has increased the life expectancy of people infected with HIV-1 and decreased the incidence of AIDS. Since the introduction of HAART, the proportion of deaths attributable to liver disease has risen. A longer life expectancy might allow greater time for cirrhosis to develop, and both HAART-associated liver toxicity and immune restoration could accelerate liver damage. Studies on people coinfected with HIV-1 and HBV have been limited by small size, cross-sectional design, confounding with HCV, or selection bias of a referral clinic population, which sees more severe cases. Our aim was to investigate the inter-relation of HIV-1 and HBV infection with liver-related mortality.
We prospectively recruited men who reported having had sex with men into the
Multicenter AIDS Cohort Study (MACS) from four metropolitan areas in the USA
(Baltimore, MD; Chicago, IL; Pittsburgh, PA; and Los Angeles, CA) between April, 1984, and March, 1985, and during a second recruitment period designed to increase minority representation in the cohort between 1987 and 1991. We included all men who had been tested for HBsAg and hepatitis B core antibody, and HIV-1 by enzyme immunoassay (Abbott Laboratories, Abbott Park, IL, USA, and ELISA, Du Pont, Wilmington, DE, respectively) at study entry.
Our primary outcome was rate of liver-related mortality in HBsAg positive-HIV-1 seropositive individuals compared with HBsAg positive-HIV-1 seronegative, HbsAg negative-HIV-1 seropositive, and HBsAg negative-HIV-1 seronegative individuals. We ascertained primary and secondary causes of death from death certificates.
We recruited 5622 men, of whom 5293 (94%) had appropriate HBV and HIV-1
testing to be included in these analyses. 3356 (64%) men had been exposed to
HBV and were seropositive for hepatitis B core antibody. The groups were similar with respect to age, ethnic origin, number of sexual partners, and amount of alcohol consumed per week. However, HIV-1 seropositive men were more likely to have ever used injection drugs and had lower nadir CD4 counts than HIV-1 seronegative men. The median age at study entry was 32.6 years, and the median follow-up time was 10.5 years.
During 55,123 person years of follow-up, there were 1648 deaths, resulting in an all-cause mortality rate of 29.9/1000 person years. This rate was much higher in men with HIV-1 than in those without (relative risk 33.8, p<0⋅001). 62 (3.8%) deaths were attributable to liver disease; more of which arose in HBsAg positive than in HBsAg negative men (12.7, p<0⋅0001). There were 61 liver-related deaths among the HIV-1 seropositive men and one among the HIV-1 seronegative men (88.9, p<0⋅0001).
In individuals with HIV-1, deaths related to liver disease rose from 2.5/1000 person years before 1996 to 4.0/1000 after 1996, the introduction of HAART (1.6, p=0.16). By contrast, there was a large decrease in AIDS-related mortality rates, which fell from 68.9/1000 before 1996 to 35.6/1000 after 1996 (0.5, p<0⋅001).
No liver-related deaths were observed in those who were not infected with either HIV-1 or HBV. The rates increased significantly in individuals infected with either virus. However, the highest rate was noted in coinfected men. Coinfected men were more than eight times as likely to die from liver disease than those infected with HIV-1 alone (8.3, p<0⋅001), and almost 19 times as likely to die from liver disease than those infected with HBV alone (18.7, p<0⋅001).
Because all but one liver-related death were in men with HIV-1, we examined other possible correlates of liver-related death exclusively in the HIV-1 seropositive men, according to HBsAg status. In the HIV-1 seropositive HBsAg negative men, the two factors that were significantly associated with liver-related mortality were quantity of alcohol consumed and CD4 count. Men who drank 210 g or more of alcohol per week had a much higher rate of liver-related mortality than those who drank less than 210 g (7.6, p<0⋅001). Rate of liver-related death also increased as the nadir CD4 count fell from more than 250 cells/ÁL to less than 100 cells/ÁL (7.6, p<0⋅001). In HIV seropositive-HBsAg positive men, higher liver-related mortality was observed for each risk factor but none was significantly associated with liver death.
In men with HIV-1 and nadir CD4 counts below 100 cells/ÁL, HBsAg positivity carried a relative risk of 11.6 compared with 6.8 when the nadir CD4 count remained greater than 250 cells/ÁL. HBsAg positivity carried a lower relative risk before 1996 than after. Most of the increased liver-related mortality after 1996, therefore, probably occurred among individuals with both HIV-1 and HBsAg.
Before 1996 (advent of HAART) HBV/HIV coinfected individuals had a 7.6 times greater risk of liver death than patients with HIV alone. Before 1996, the mortality rate for HBV negative individuals was 1.6 per 1000 patient years versus 12.3 per 1000 patient years for HBV positive individuals. After 1996, coinfected individuals had an 11.4 times greater risk for liver death than individuals with HIV alone. Before 1996 the mortality rate was 2.2 per 1000 patient years HBV negative individuals versus 24.7 per 1000 patient years for individuals with HIV alone.
No association was detected between liver-related mortality and lamivudine use, which almost exclusively was reported after 1996. The median percentage of lamivudine use was 69% of these, compared with 80% in the 63 men who died from other causes and acknowledged lamivudine use (p=0.60). Likewise, no association was detected between lamivudine discontinuation and liver-related death, but data are sparse. Sufficient data to assess discontinuation of lamivudine was available for ten men. One of the four (25%) men with a liver-related death and three of six (50%) men with a non-liver-related death discontinued lamivudine within the year before death (odds ratio 0.33, p=0.90).
Discussion By Authors
Our results show an increased risk of death from liver disease in persons infected with either HIV-1 or HBV, but this risk was highest in men coinfected with both viruses. Rates of liver-related deaths were higher in men with the lowest CD4 nadir and seemed to increase after 1996, when HAART was introduced. These findings suggest that although HAART is effective against HIV-1, identification and comprehensive management of individuals coinfected with HIV-1 and HBV is important, especially in the era of HAART.
The increased risk of liver-related mortality in HBsAg positive men infected with HIV-1 was notable, but the explanation for this observation is not clear. HIV-1 may destroy CD4+ lymphocytes that are activated to combat HBV infection. However, it is not clear how depletion of CD4+ lymphocytes would directly enhance hepatic fibrosis, which has been linked with a more vigorous cytotoxic T-cell response. CD4 depletion could alter the intrahepatic cytokine milieu, contributing to increased fibrosis. Depletion of CD4 cells is also associated with liver injury from other causes, such as opportunistic infections or hepatotoxicity of drugs used for prophylaxis or treatment of such infections. HIV-1 infection also affects the immune system in many other ways that could directly or indirectly enhance liver fibrosis from HBV, as has been previously reported for HCV infection.
The trend toward increased risk of death related to liver disease in HIV-1-HBV
coinfected men noted after 1996 has several possible explanations. First, toxicity from antiretroviral medications could explain these findings; an increased frequency of hepatitis from HAART has been reported in individuals with HBV or HCV infections. Second, stopping treatment with lamivudine can exacerbate HBV-related liver disease. Our analysis did not indicate that this factor contributed to mortality, but only a small number of men had sufficient data to assess this possibility. Third, the increase could be due to the longer duration of HBV infection by the advent of HAART. Fourth, immune responses restored by HAART might actually exacerbate progression of HBV infection, as has been reported for tuberculosis and cytomegalovirus infections. Such exacerbations of HBV infections also arise after chemotherapy or after cessation of treatment with steroid. Although episodes of hepatitis have been reported in association with enhanced immune responses to HBV (clearance of hepatitis B e antigen [HBeAg]), it remains unexplained (and somewhat paradoxical) that both infection with HIV-1 and successful antiretroviral therapy should be associated with greater progression of hepatitis B.
In those with CD4 counts between 100-250 cells/mm3 and in those with CD4 counts of greater than 250 cells/mm3, the rate of liver-related death was about twice as high after 1996. Unfortunately, there was an insufficient amount (45 person years) of follow-up in those with CD4 counts of less than 100 cells/mm3 to draw any definite conclusions. Longer follow-up is, therefore,
needed to understand the competing effects of HAART and chronological time on liver-related deaths.
Because infection with HCV also causes liver-related mortality that seems to be
enhanced in some individuals with HIV-1, it is important that the members of this cohort were not at high risk for HCV. In fact, the prevalence of antibodies against HCV in a subset of this cohort has previously been shown to be about 2%, which is equivalent to the national average. Although only 8% of people infected with both HBV and HCV died of a liver-related illness, almost 30% of those who had liver-related deaths in the HBsAg negative group were positive for antibodies against HCV. Thus, the increased risk of liver-related death noted in coinfected individuals compared with individuals with only HIV-1 would probably be even greater if HCV-related liver disease were excluded. Unfortunately, presence of antibodies against HCV was not uniformly assessed in the MACS cohort, so an analysis excluding HCV infected persons could not be done.
It is also noteworthy that few deaths from liver disease were seen in men positive for HBsAg who were not infected with HIV-1. Since it is likely that these men acquired HBV infection as adults, it is possible that too little time had passed for death from liver failure to occur. Little liver-related mortality has also been reported in other studies of people who were likely infected as adults.
These results expand on existing studies of HBV infection in people with HIV.
Results of several cross-sectional studies, have shown that individuals who are infected with HIV-1 and HBV are more likely to be HBeAg positive, have higher amounts of HBV DNA in serum, and have lower serum alanine transferase concentrations than individuals without HIV-1, suggesting that the immune response to HBV is less vigorous. Although the lower liver enzymes recorded in people infected with HIV-1 and hepatitis B might indicate less disease these are indirect measures, and greater disease would be expected because a higher proportion are HBeAg positive.
By use of a more direct measure of liver disease (liver histology), Colin and
colleagues noted more cirrhosis in coinfected individuals than in those without
HIV-1. Likewise, Puoti and co-workers examined 308 individuals with HIV-1 who
died while in hospital, 35 of whom died of liver failure, and noted that being HbsAg positive was independently associated with death due to liver failure. However, 33 of the 35 individuals were also positive for antibodies against HCV, diminishing the extent to which HIV-1 on HBV alone could be examined.
We did not test for HBeAg or hepatitis D virus (HDV) in the chronically infected individuals. In individuals with chronic HBV infection, the presence of HBsAg tends to correlate with higher concentrations of HBV DNA and more severe
liver disease. Likewise, HDV coinfection could lead to more aggressive liver disease in individuals infected with HIV-1. Different courses of the liver disease might be seen dependent on the HBeAg or HDV status of coinfected individuals. Our findings suggest that further examination of these other viral cofactors would be informative.
Although mortality is arguably the most important HBV-related outcome, it is
a fairly insensitive measure of liver disease. Prospective examination of changes in liver histology is a more sensitive indicator of liver disease and might have enabled us to see greater liver disease in individuals with hepatitis B but not infected with HIV-1.
Further studies are needed to better understand the mechanisms by which HIV-1 increases liver-related mortality, including the effects of HAART. Our data do, however, emphasise the importance of prevention, identification, and
management of hepatitis B in individuals infected with HIV-1.