| |
Occult Hepatitis B in HIV-Infected Patients
|
| |
| |
JAIDS Journal of Acquired Immune Deficiency Syndromes: Volume 36(3) 1 July 2004
Shire, Norah J. MPH; Rouster, Susan D. BS; Rajicic, Natasa MS*; Sherman, Kenneth E. MD, PhD
From University of Cincinnati College of Medicine, Cincinnati, OH, and *Harvard School of Public Health, Boston, MA.
Received for publication September 11, 2003; accepted January 23, 2004.
Supported by grants from National Institute of Allergy and Infectious Diseases (NIAID) AIDS Clinical Trials Unit (grant AI25897), Adult AIDS Clinical Trials Group NWCS 205 (grant AI38858), and NIAID AI49501.
"...The present study analyzed 240 randomly selected serum samples from study AACTG groups 320 and 343...Our analysis suggests that approximately 65% (155 of 240) of HIV-infected patients in this study population demonstrated some marker of current or past HBV infection (including the 2.5% [6 of 240] who may have received prior vaccination). This is much higher than the overall US prevalence estimate of 5.4% but is in line with previous studies of HIV-infected populations. Other studies also suggest that HB vaccination rates in HIV-infected patients are low. 29 Only 6 of the 240 samples tested positive for IgM antibody, suggesting that most patients in this study group had past or chronic infection rather than acute. Approximately 7% of subjects in this study (17 of 240) demonstrated a marker pattern indicative of acute or chronic infection; about 38% (91 of 240) showed evidence of prior infection and clearance; and nearly 16% (38 of 240) demonstrated the anti-HBc alone pattern, indicating possible occult HBV infection. This proportion of anti-HBc alone is lower than has been described in previous studies of HIV-positive patients yet is higher than would be expected in the general population, even in areas of high HBV endemicity...
"...the prevalence of occult HBV in this HIV-positive study population is higher than that of non-HIV-infected cohorts 2,48 but lower than would have been predicted based on previous studies. The low DNA titers found in occult HBV samples were not found to be predictive of ALT levels, but the sample population was HAART naive, and immune reconstitution may alter this finding. However, an occult HBV prevalence of 10% out of the anti-HBc alone subjects raises the question of whether patients, blood, or liver donors that are anti-HBc alone should undergo further screening for HBV replication. It was recommended in a recent workshop report that the screening for anti-HBc alone in the transplant setting should be universal and mandatory because the HBV transmission evidence is so compelling. 49 It would further appear that patients treated with HAART who exhibit serum transaminase abnormalities, hyperbilirubinemia, or liver failure should be tested for HBV markers, and if found to be positive for anti-HBc alone, they should be further tested for HBV DNA. It is less clear whether decisions regarding choice of antiretroviral agents should be guided by anticore positivity. Prospective evaluation of this subgroup following initiation of HAART is clearly indicated..."
Hepatitis B virus (HBV) remains endemic in many parts of the world and continues to pose a serious threat to public health. As of the year 2000, approximately 2 billion people globally had been infected with the virus; an estimated 350 million of these people are chronic carriers of the disease. A classic carrier state characterized by chronic hepatitis B surface antigenemia (HBsAg) is most prevalent in developing nations, such as those in sub-Saharan Africa, Asia, and the Pacific region. In contrast, occult HBV infection is characterized by absence of surface antigenemia with potentially infectious HBV DNA present in the liver, serum, or both. The presence of anti-HBc alone has been identified as a predictive marker of occult HBV infection.
Occult HBV has been identified in several cohorts, including blood donors, liver transplant donors, and HIV-infected patients. The prevalence and clinical significance of occult HBV in HIV-infected patients remain controversial. Hofer et al 4 investigated a 57-patient subset of the Swiss HIV cohort who tested positive for anti-HBc alone. They found this pattern in 54% of drug abusers who were infected with HIV. Sequential polymerase chain reaction (PCR) testing of serum from these subjects determined that 29.8% had consistently detectable HBV DNA for the core gene, and 89.5% had at least 1 sample test positive for HBV DNA. However, Núñez et al 5 studied serum samples from 85 HIV-infected patients who were anti-HBc positive, anti-HBs negative, and HBsAg negative. HBV DNA was not identified by PCR amplification in any subjects.
While the prevalence of typical or occult HBV and HIV co-infection has been investigated in discrete populations and individual high-risk cohorts, including injection drug users, homosexual men, or both, only 1 study has addressed US prevalence of chronic HBV in a large HIV-positive prospective cohort. This study relied upon chart reviews of routine clinical care to determine HBV status. Furthermore, it failed to address putative markers of occult HBV. A comprehensive cross-sectional analysis of HBV prevalence in a HIV-positive, pre-highly active antiretroviral therapy (HAART) cohort representing a mix of risk groups has not previously been reported. Our objectives in the study were 2-fold: to determine the prevalence of current and past HBV infection in such a cohort and to determine the prevalence and significance of occult HBV.
Abstract
Prevalence of hepatitis B virus (HBV) markers, including occult HBV, has not been described in diverse cohorts among HIV-infected patients. The objective of this study was to assess prevalence and significance of active and occult HBV infection in an HIV-positive US cohort.
A random sample was taken from 2 prospective multicenter treatment intervention cohorts. The sample population (n = 240) was HIV-1 infected and highly active antiretroviral therapy-naive. Prevalence of HBV serologic markers and quantitative HBV DNA were determined. Serum alanine aminotransferase (ALT) levels were measured to evaluate correlates of hepatocyte injury.
A total of 64.6% of subjects demonstrated reactivity for any marker of current or past HBV infection or prior vaccination. Chronic HBV infection characterized by hepatitis B surface antigen (HBsAg) reactivity was present in 7.1% while 15.8% exhibited HB anticore IgG only. Approximately 10% of the latter group was HBV DNA positive by a polymerase chain reaction-based assay. Only patients with a serologic pattern of HBsAg or HB anticore alone reactivity had HBV DNA. Occult HBV was observed in approximately 10% of HIV-infected patients with HB anticore IgG antibody in a geographically representative national cohort. Though viral titers and serum ALT levels were low, screening of this subset of HIV-infected patients may have implications in terms of antiretroviral therapy and risk of immune reconstitution-associated flares.
DISCUSSION
Although a high concordance between HBV and HIV infections is well established, the prevalence of HBV infection in a nationally representative, geographically diverse, HIV-positive US study population has only recently been described in published literature. Kellerman et al determined the prevalence of chronic HBV to be 7.6% in the Adult/Adolescent Spectrum of HIV Disease Project, and the incidence of acute HBV to be 12.2 cases per 1000 person-years. However, this study was limited to chart review of testing performed in the context of routine clinical care. Serologic patterns were not reported and centralized testing for HBV DNA was not performed.
The present study analyzed 240 randomly selected serum samples from study AACTG groups 320 and 343. The samples were collected in 1996 and 1997, allowing for HBV marker reactivity unaltered by HAART. Immune reconstitution may lead to alteration of serologic patterns. 21,22 The study sample is demographically representative of the full 320/343 AACTG study groups and appears to mirror the AACTG study population (n = 57,064 registered adults) in key demographic characteristics. CD4+ T-cell counts and HIV RNA titers are not reported for the AACTG in total; however, these were not determined to be factors in either HBV serology or in HBV DNA titers. Our analysis suggests that approximately 65% (155 of 240) of HIV-infected patients in this study population demonstrated some marker of current or past HBV infection (including the 2.5% [6 of 240] who may have received prior vaccination). This is much higher than the overall US prevalence estimate of 5.4% but is in line with previous studies of HIV-infected populations. Other studies also suggest that HB vaccination rates in HIV-infected patients are low. Only 6 of the 240 samples tested positive for IgM antibody, suggesting that most patients in this study group had past or chronic infection rather than acute. Approximately 7% of subjects in this study (17 of 240) demonstrated a marker pattern indicative of acute or chronic infection; about 38% (91 of 240) showed evidence of prior infection and clearance; and nearly 16% (38 of 240) demonstrated the anti-HBc alone pattern, indicating possible occult HBV infection. This proportion of anti-HBc alone is lower than has been described in previous studies of HIV-positive patients yet is higher than would be expected in the general population, even in areas of high HBV endemicity.
Serum DNA was detectable in only 2 of the groups: group 1 (representing acute or chronic infection), in which 14 of the 17 (82.35%) were positive, and group 3 (the anti-HBc alone group), in which 4 of the 38 (10.53%) were positive. It is not uncommon for a small proportion of HBsAg-positive samples to test negative for HBV DNA, as surface antigen may be produced in the absence of replicating virus. The difference in median HBV DNA titers between the 2 groups strongly trended toward statistical significance (P = 0.06). Three of the 4 DNA-positive samples in the anti-HBc alone group had very low HBV DNA titers of 200 copies/mL. This is consistent with the low-level replication reported in other studies. Longitudinal studies will be necessary to determine whether low-level replication is persistent or punctuated by flares and remissions.
Overall, nearly 10% of the sample population (23 of 240) had elevated aminotransferases, defined as ALT or AST levels >2 x ULN. ALT levels were independent of serologic marker group and HBV DNA positivity or titer, although median AST levels did differ significantly between groups (P = 0.02), with the highest levels present in group 2 (prior infection and clearance). This is in contradiction to the Swiss HIV Cohort Study, in which HBV DNA detection was associated with elevated ALT levels, with or without HCV co-infection, and it is also in contrast with studies that have found higher proportions of occult HBV in patient groups selected by elevated aminotransferases vs. normal-ALT control groups. Other studies, however, have not detected such an association. The difficulty in cross-study comparisons of occult HBV prevalence has been one of the challenges in assessing the clinical significance of the anticore alone pattern, even after the advent of highly sensitive PCR methodology for DNA detection. Hofer et al used a proprietary nested PCR to determine DNA sequences for HBV core and surface genes, rather than a commercially available methodology. They also tested their samples multiple times and found that DNA detectability was not constant over the testing period. Consistent use of a standardized, commercially available PCR methodology, such as that used in this study, would be very useful for accurate prevalence comparisons.
Another factor adding to the difficulty of cross-study comparisons is the prevalence of hepatitis C virus (HCV) co-infection in the study populations. It is well established that HCV-positive patients exhibit the anti-HBc alone pattern more often that HCV-negative patients, and it has been hypothesized that HCV suppresses HBV replication, leading to a lack of circulating HBsAg. We were unable to assess the possible effect of HCV co-infection in this study cohort, including any possible effect on ALT/AST levels, because of limited sample availability for further testing, though prior testing from the same AACTG population revealed HCV reactivity in 16.1%.
A critical question that has yet to be answered definitively is the clinical significance of occult HBV infection in the presence of HIV co-infection. Occult HBV has been implicated in significant liver pathology leading to hepatocellular carcinoma and has been identified as the causative agent in HBV transmission via liver transplantation and blood donation. The primary mechanism of HBV-associated liver injury is the cell-mediated immune response to the virus. CD4+ T lymphocytes play a central role in this response by recognizing viral antigen presented by Kupffer cells and by secreting cytokines that regulate the activities of B cells and CD8+ T cells. The immune dysfunction in HIV-infected patients can abrogate the impact of HBV pathogenesis, although this diminished immune response appears to permit increased HBV replication, possibly leading to increased infectivity. It has also been found that HIV infection can increase the likelihood for patients acutely infected with HBV to become chronic carriers.
In conclusion, the prevalence of occult HBV in this HIV-positive study population is higher than that of non-HIV-infected cohorts 2,48 but lower than would have been predicted based on previous studies. The low DNA titers found in occult HBV samples were not found to be predictive of ALT levels, but the sample population was HAART naive, and immune reconstitution may alter this finding. However, an occult HBV prevalence of 10% out of the anti-HBc alone subjects raises the question of whether patients, blood, or liver donors that are anti-HBc alone should undergo further screening for HBV replication. It was recommended in a recent workshop report that the screening for anti-HBc alone in the transplant setting should be universal and mandatory because the HBV transmission evidence is so compelling. It would further appear that patients treated with HAART who exhibit serum transaminase abnormalities, hyperbilirubinemia, or liver failure should be tested for HBV markers, and if found to be positive for anti-HBc alone, they should be further tested for HBV DNA. It is less clear whether decisions regarding choice of antiretroviral agents should be guided by anticore positivity. Prospective evaluation of this subgroup following initiation of HAART is clearly indicated.
PATIENTS AND METHODS
Patient Samples
Stored plasma specimens obtained from patients enrolled in 2 Adult AIDS Clinical Trials Group (AACTG) trials, 320 and 343, were used. The randomly selected samples were taken from HAART-naive patients in 1996 and 1997, allowing for HBV marker reactivity unchanged by HAART activity. Both trials were primary treatment intervention trials that used HAART; they were selected for this study because they included diverse and geographically representative patient populations with all stages of HIV-associated disease. AACTG 320 included HIV-infected subjects with CD4+ lymphocyte counts of <=200 cells and a history of treatment with zidovudine of >=3 months. AACTG 343 included HIV-infected subjects with HIV RNA levels of >1000 copies/mL and CD4+ lymphocyte counts of >200 cells who had not previously received therapy with a protease inhibitor. The study sample is demographically representative of the full 320/343 AACTG study groups. All subjects provided informed consent at time of entry into the AACTG and the studies were approved by the institutional review boards at each individual site. Vaccination records were not available for the sample population; therefore, prior vaccination was suggested by serology. Because this is a cross-sectional analysis, patient outcome data are not provided. All samples were obtained in acid citrate dextrose collection tubes or in serum separator tubes and were stored in accordance with principles of viral DNA preservation, including rapid separation from cells, followed by rapid freezing and storing at -70°C or lower. HIV RNA testing by use of Roche Amplicor (Roche Diagnostics, Indianapolis, IN) was performed as part of the AACTG study protocols by the Johns Hopkins HIV Specialty Laboratories (Baltimore, MD).
HBV Test Methods
Serologic evaluation for HBsAg, anti-HBs, and total anti-HBc was performed using enzyme immunoassay test kits from Abbott Laboratories (Abbott Park, IL), in accordance with manufacturer's specifications. Those samples that tested positive for total anti-HBc were subsequently tested by Abbott enzyme immunoassay for the presence of the specific antibody IgM anti-HBc.
Samples were tested for HBV DNA using the Roche COBAS Amplicor HBV Monitor assay, which is a PCR amplification-based assay. HBV DNA from the specimen was isolated by polyethylene glycol precipitation followed by centrifugation and recovered from the pellet with a lysis buffer. An internal standard of known HBV concentration was added to each specimen to permit accurate quantitation. PCR amplification of DNA from the pre-core/core region of the HBV genome followed by colorimetric detection was performed on the COBAS Amplicor Analyzer. Results were reported as HBV DNA copies/mL with an assay range of 200-200,000 copies/mL. Sample results above the linear range of the assay were serially diluted from 1:10 to 1:1,000,000 and retested to determine the viral load.
Serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels were determined on all specimens by batch testing using the Hitachi 917 autoanalyzer. Local normals were 5-46 U/L or 5-40 U/L for ALT and AST, respectively. Results were considered elevated if they were >=2 times greater than the upper limit of normal (ULN).
RESULTS
A total of 240 representative patient samples from 2 AACTG study groups (320 and 343) were studied. The median age was 37 (range, 20-75) years. Fifty percent (120 of 240) were white non-Hispanic and 83% (199 of 240) were male. Median CD4+ count was 137 (range, 0-856) cells/mm3. Forty percent (96 of 240) of the subjects had HIV RNA titers of >100,000 copies/mL. An analysis was performed to compare demographic characteristics of the tested study sample (n = 240) with the entire study population (n = 1687). No statistically significant demographic differences were detected between the sample and the population
Subjects were divided into 6 groups representing discrete patterns of antibody to HBV. These include:
* group 1: anti-HBs negative, anti-HBc positive, HBsAg positive (acute or chronic infection)
* group 2: anti-HBs positive, anti-HBc positive, HBsAg negative (infection and clearance)
* group 3: anti-HBs negative, anti-HBc positive, HBsAg negative (anticore alone, marker for possible occult if an-ticore IgM negative)
* group 4: anti-HBs positive, anti-HBc negative, HBsAg negative (generally indicative of prior vaccination, although this marker pattern may appear in the setting of remote infection with loss of anti-HBc)
* group 5: all negative
* group 6: all positive, indicating a possible window phase of infection
Overall, 64.6% of subjects (155 of 240) demonstrated reactivity for any marker of current or past HBV infection (95% CI, 58.5-70.7%), although 2.5% (6 of 240) of those had possibly received prior vaccination. Six samples tested positive for IgM antibody, indicating acute infection; three were found in group 1, two in group 2, and one in group 3. The sample testing IgM positive in group 3 was negative for HBV DNA, indicating the possible window phase of acute infection.
Testing for HBV DNA by PCR detected a positive result in 7.5% (18 of 240) of all the study samples (95% CI, 4.14-10.86). Fourteen of these were in group 1 (acute or chronic infection) such that 82.35% (14 of 17) of group 1 was HBV DNA positive (95% CI, 61.15-102.56). Group 3, the anticore alone group, yielded 4 of 38 (10.53%) HBV DNA-positive subjects (95% CI, 0.30-20.75). There were no samples with detectable HBV DNA among subjects without prior serologic markers of HBV infection, or among those with a pattern of infection and clearance (group 2).
The median HBV DNA titer for all DNA-positive subjects was 1.01 x 109 copies/ml. The median DNA titer for subjects in the chronic/acute group (group 1) was 2.16 x 109 copies/ml. The median DNA titer for subjects in the anti-HBc alone group (group 3) was 200 copies/mL. The Wilcoxon rank sum test indicated a strong trend toward difference in median DNA titers between the 2 marker groups (P = 0.06).
Overall, 9.6% of the sample population (23 of 240) had ALT or AST levels >=2 x ULN. Nearly 50% of those with abnormal levels were in the infection and clearance group (group 2). The median ALT level was 30 U/L (range, 7-258 U/L) and the median AST was 33 U/L (range, 10-252 U/L). The Kruskal-Wallis test detected a statistically significant difference in median AST levels among marker groups, with the highest in group 2 (P = 0.02). This difference was not detected for ALT (P = 0.1). No statistical differences in ALT levels by positive or negative PCR results were found via the Wilcoxon rank sum test (P = 0.57). Spearman rank test for correlation did not detect a significant correlation between HBV DNA titers and ALT or AST levels (P = 0.55, P = 0.20, respectively), indicating that ALT and AST levels are statistically independent of HBV DNA titers in this cohort.
|
|
| |
| |
|
|
|
