| |
Occult HBV May Be Clinically Significant
|
| |
| |
The Prevalence and Significance of Occult Hepatitis B Virus in a Prospective Cohort of HIV-Infected Patients
JAIDS Journal of Acquired Immune Deficiency Syndromes: Volume 44(3) 1 March 2007 pp 309-314
Shire, Norah J MPH*; Rouster, Susan D BS*; Stanford, Sandra D BS*; Blackard, Jason T PhD*; Martin, Christina M BS*; Fichtenbaum, Carl J MD; Sherman, Kenneth E MD, PhD*
From the *Division of Digestive Diseases, University of Cincinnati, Cincinnati, OH; Division of Epidemiology and Biostatistics, University of Cincinnati, Cincinnati, OH; and Division of Infectious Diseases, University of Cincinnati, Cincinnati, OH.
"....It is critical to determine if progression to end-stage liver disease (ESLD) or HCC with occult HBV is hastened by HIV coinfection, as is the case with HCV. We anticipate that occult infection is likely to present an ongoing challenge in the management of HBV until such studies are completed in the years to come...."
Abstract
Background: Occult hepatitis B virus (HBV) is defined as low-level HBV DNA without hepatitis B surface antigen (HBsAg). Prevalence estimates vary widely. We determined the prevalence of occult HBV at the University of Cincinnati Infectious Diseases Center (IDC).
Methods: Patients in the IDC HIV database (n = 3867) were randomly selected using a 25% sampling fraction. Samples were pooled for HBV nucleic acid extraction. Pools were tested for HBV DNA by a real-time polymerase chain reaction (PCR) assay to coamplify core/surface protein regions. The PCR assay was run on all individual samples from each DNA+ pool. DNA+ samples were tested for HBV serologic markers.
Results: A total of 909 patients without known HBV were selected. The mean CD4 count was 384 cells/mm3. Forty-three patients were HBV DNA+. Twelve of 43 were DNA+/HBsAg- (95% confidence interval for database: 0.58% to 1.90%). Five of 12 were negative for all serologic markers. Alanine aminotransferase, aspartate aminotransferase, and HBV DNA titers were elevated in HBsAg+ patients versus occult patients and versus HIV-monoinfected patients. No other significant differences were detected. No occult HBV patient was on treatment with anti-HBV activity.
Conclusions: Forty-three percent of those with HBV were not previously identified as HBV+, indicating the need for ongoing screening in high-risk populations. Occult HBV may occur in persons with all negative serologic markers, representing a challenge for identification.
Coinfection with HIV and hepatitis B virus (HBV) is common because of shared blood-borne transmission routes such as injection drug use (IDU). Among those infected with HIV, rates of chronic HBV infection range from 7% to as high as 70%1-5 and markers of prior exposure may be found in 90%.5 Combined, these 2 viruses continue to constitute major public health threats, causing 3.7 million deaths each year. In the era of highly active antiretroviral therapy (HAART), prolonged survival of patients with HIV has facilitated the emergence of chronic liver disease, including viral hepatitis, as a leading cause of morbidity and mortality in several HIV-positive cohorts.6-8
The classic chronic carrier state of HBV is characterized by persistent surface antigenemia (hepatitis B surface antigen [HBsAg]) and hepatitis B envelope antigen (HBeAg) for >6 months after infection and by antibodies directed against surface (anti-HBs) and core (anti-HBc) hepatitis B. In contrast, occult HBV has been defined as chronic low-level HBV replication in the absence of detectable HBsAg. The advent of sensitive polymerase chain reaction (PCR)-based assays has enabled detection of this low-level replication. Seropositivity for anti-HBc alone may serve as a surrogate for some cases of occult HBV infection. Other marker patterns have been observed in occult HBV infection, however.
Although the existence of occult HBV and the transmissibility of HBV via occult infection have been established in the literature,9 its prevalence and clinical significance remain unclear. It has been implicated in significant liver pathologic findings,10-15 raising the question of whether this phenomenon underlies some of the liver-related morbidity seen in HIV-positive cohorts. Successful antiretroviral therapy (ART) for HIV creates immune reconstitution that can result in immune-mediated liver injury in the setting of HBV infection. Occult HBV infection may be associated with elevation of liver enzymes that could be misclassified by physicians caring for these patients as HAART-associated liver toxicity. Furthermore, it is unknown if HBV treatment is indicated in these patients, which could have an impact on HIV clinical management.
Our objectives were to determine the prevalence of occult HBV in the HIV-infected population at the University of Cincinnati HIV clinical center and to characterize this cohort of occult HBV/HIV-coinfected subjects for associations between occult infection, transaminase elevations, and ART.
DISCUSSION
Occult HBV is a controversial phenomenon, with conflicting data on its prevalence and clinical significance. In the current analysis, 12 occult patients were detected from a total of 909 subjects, representing a prevalence of 1.3% (95% confidence interval [CI]: 0.6% to 1.9% for the entire cohort of 3867 patients), with a range of serologic markers. This is in contrast to prior studies. For example, an analysis of a random sample of 240 treatment-naive subjects from the Adult AIDS Clinical Trials Group (AACTG) detected a 16% prevalence of anti-HBc alone,4 of which 10% had detectable HBV DNA. Another recent analysis detected occult HBV infection in 4% of 93 anti-HBc-alone patients in an HIV-positive HCV- cohort.18 A study of the French Aquitaine cohort detected HBV DNA in only 1 of 160 HIV-positive anti-HBc-alone individuals, however.19 It has been postulated that occult HBV strains harbor mutations that prevent generation of excess surface antigen, and some studies have found such sequence variations.20-22 This does not adequately address the higher prevalence of occult HBV seen in patients with HIV, however.
Rather than identifying those with the anti-HBc-alone marker pattern, we identified those with HBV DNA positivity first and then performed serologic testing. This allowed us to detect 5 subjects harboring occult infection with no serologic evidence of infection, who were negative for HBsAg, anti-HBc, and anti-HBs. In these 5 patients, HBV replication was low (<4.3 log IU/mL), 1 patient was on ART, another had elevated ALT (63 U/L), a third had a detectable HIV viral load, and CD4+ counts ranged from 106 to 559 cells/mm3. There were no distinguishing factors that would differentiate these patients from the other occult patients. This phenomenon has not been previously described to our knowledge, likely because it is rare to test for HBV DNA before serologic status. To address the theoretic possibility of PCR contamination, we retested each of these 5 serologically negative patients separately and in triplicate by real-time PCR assay. As with the first round, PCR products were again detected in the core or surface region (or both) of all 5 samples. In a separate qualitative assay, we amplified the core and presurface regions of the genome from each sample (data not shown). The retesting of serologically negative individuals by real-time PCR assay and amplification using a qualitative PCR assay strongly suggest that our findings are not attributable to contamination.
The finding of 5 all negative patients raises the question of clinical relevance. Because occult HBV has been shown to have hepatopathogenic potential, it may eventually prove necessary to identify and monitor these patients. If occult infection is possible with all-negative serologic marker patterns and there are no other clinical signs or symptoms of infection, identification of these patients is not possible unless a sensitive PCR assay is performed, which is an impractical option for most clinical settings. These patients may remain unidentified for years until symptoms develop, and even then, elevated ALT/AST levels may be misattributed to ART-related hepatotoxicity if no other evidence of viral hepatitis is present. Indeed, a recent prospective study detected a significantly higher rate of liver enzyme elevations in HIV-positive patients with occult HBV than in those without HBV DNA.23
Occult HBV has been implicated in HBV transmission through blood donation9,24 and liver transplantation;25 in HBV reactivation;22,26 and in significant liver pathologic findings, including hepatocellular carcinoma (HCC).11-13 In vitro and in vivo studies have shown that the hepatitis B X gene (HBx) proteins bind to p53, a tumor suppressor gene, modifying its function.27-30 Although many activities of HBx are proapoptotic,31,32 mutations in the HBx encoding sequences from HCC cells can inhibit this activity.29,33 It is unclear if occult HBV patients have an altered prevalence of HBx mutations compared with other chronic HBV patients or if coinfection with HIV could mediate this effect. Virologic characterization of occult HBV and longitudinal studies of infected patients are required to understand the clinical implications of occult infection.
Several limitations to this study should be noted. The cross-sectional nature of the study does not permit conclusions regarding the clinical significance of occult HBV, although 1 occult HBV patient with ALT >2 times the ULN had no other risk factors for liver enzyme elevations, suggesting that adverse clinical findings may occur in at least some cases. The single-center nature of this analysis could limit its generalizability to settings with significantly different demographic characteristics. We were unable to conduct separate testing for IgM and IgG antibodies. It is possible, therefore, that some of the occult patients actually represent a window phase of infection, in which the patient has been acutely infected and lost core antibody but has not yet developed anti-HBs. Our prior analysis in the AIDS Clinical Trials Group (ACTG) did include IgM testing, however, and no patient in a window phase was identified.4 Thus, this possibility seems unlikely. Finally, the number of occult patients limits power for in-depth analysis. Expanding the characterization of this cohort should facilitate further comparisons and contrasts.
Questions and controversy surrounding occult HBV remain. Until clinical occult HBV is better characterized by prospective studies, it would be impractical to recommend DNA testing on serologically negative patients. Conversely, patients with HIV who are at risk for HBV (occult or otherwise) should be screened for surface antigen and core antibody, and screening should be repeated every 2 to 3 years if negative. In the event of unexplained liver enzyme abnormalities and negative serology, occult HBV may be suspected and DNA testing may be warranted, even in the case of negative serology. It is also important to emphasize the low vaccination rate in this patient population and to note that 2 patients with prior vaccine histories were actually HBV DNA+. Patients with HIV and a history of prior vaccination should be repeatedly screened for anti-HBs titers. It is critical to determine if progression to end-stage liver disease (ESLD) or HCC with occult HBV is hastened by HIV coinfection, as is the case with HCV. We anticipate that occult infection is likely to present an ongoing challenge in the management of HBV until such studies are completed in the years to come.
RESULTS
Patient Population
Using sample survey techniques as described, a 25% sampling fraction was used to calculate a required sample size of 967 patients from the entire sample database for testing. Thus, every fourth patient was randomly selected from the database to yield the required number of samples for testing. Fifty-eight patients were documented to have chronic HBV; these individuals were excluded from further analysis, yielding 909 patients for further analysis. The mean patient age was 35 ± 8.7 years, 77% of patients were male, and 50% were white. Six hundred eighty-eight patients had a recorded HIV viral load; of those, 62% had a detectable HIV viral load. The mean CD4+ count was 385 ± 305 cells/mm3. Twenty-three percent (210 of 909) of patients were on ART. Of these, 128 were on ART with activity against HBV (lamivudine, adefovir, tenofovir, or combinations including these treatments). Serum samples were typically obtained from individuals on entry to the IDC, explaining, in part, the lower level of use of ART.
Hepatitis B Virus DNA
Using a pooled testing approach, 32 sample pools of 100 were positive for HBV DNA. Each individual in the positive pools was then tested separately. Forty-three individual patients (4.7%) were HBV DNA+. Twelve of these 43 patients (27.9% [1.3% of the cohort]) were HBsAg-negative, indicating occult HBV infection. Serologic marker patterns for all HBV DNA+ patients are shown in Table 1. Although the typical marker pattern of anti-HBc alone was seen in 2 patients, 5 patients were negative for all serologic markers, and 4 patients were positive for anti-HBc and anti-HBs. Analysis of demographic and/or clinical differences between the 4 serology groups for the occult patients was not feasible because of the sample size. In the 3 main groups of patients (HBV-, HBsAg+, and occult), gender proportion was the only significant between-group difference, with HBsAg+ patients being 100% male. The most prevalent risk factor for HBV infection was homosexual sex. Of the patients on ART with anti-HBV activity, 4 were chronic HBsAg+. None were occult. The demographic and clinical characteristics of these patient groups are summarized in Tables 2A and 2B. There was a significant difference in HBV DNA titers between occult and HBsAg+ patients (3.71 ± 1.00 vs. 6.09 ± 2.23 log IU/mL; P < 0.001, t test with Satterthwaite correction), as shown in Figure 1. Interestingly, 1 patient was repeatedly and strongly HBsAg+ yet HBV DNA-. The PCR assay was performed in triplicate and manually on this subject with reproducibly negative results, raising the possibility that sequence mutations precluded amplification with our primer sets. This patient was not included as occult or HBV+ in analysis. Only 24 patients had a record of prior vaccination; none were chronic HBsAg+, but 2 of these were occult HBV+. One patient was anti-HBs+ and anti-HBc+, and the other was anti-HBs+ only. Both had low-level replication (4.29 and 3.11 log IU/mL, respectively).
Liver Enzyme Elevations
Overall, 121 patients had alanine aminotransferase (ALT) elevations greater than 2 times the upper limit of normal (ULN). Of these, 109 were HBV DNA-, 10 were HBsAg+, and 2 had occult HBV infections. Mean levels of ALT and aspartate aminotransferase (AST) for the HBV- group were 34.5 ± 30 U/L. For the 2 antigen-positive and occult HBV+ groups, mean ALT and AST levels were 64 ± 30 and 33 ± 22 U/L, respectively. Because neither ALT nor AST was normally distributed, values were log-transformed for further analysis. The differences in transformed ALT and AST between the HBsAg+ patients and the other 2 groups were statistically significant (P < 0.001, analysis of variance [ANOVA] with the Scheffe adjustment) as shown in Figure 2. ALT and AST did not differ between occult versus HBV- patients, however. Of those 121 patients with elevated ALT, 99 had information available on ART regimens. Seventy-two (73%) of the 99 patients were not on any ART. Eight-four of the 121 patients had HCV ELISA data available, and 20 (24%) were positive by ELISA. Two occult patients had an ALT >2 times the ULN. One was HCV-coinfected and was not on ART; the second was neither HCV+ nor on ART. Of the 31 HBsAg+ patients, 10 had an ALT >2 times the ULN, 2 of whom were on ART; 18 had HCV ELISA records; and all were negative. Log ALT and log AST correlated positively and significantly with the log HBV viral titer (r = 0.30, P = 0.05; r = 0.33, P = 0.04, respectively).
CD4 and CD8 Cell Counts
Overall, the mean CD4+ count was 385 ± 305 cells/mm3. Although counts were highest in the HBV- group and lowest in the occult group (shown in Table 2), this trend was not statistically significant (P = 0.20, ANOVA with log-transformed CD4+ cell count). Similarly, CD8+ counts were lowest in the occult group but not statistically different from the other 2 groups. These trends remained after removing patients on ART from the analysis. CD4+ and CD8+ cells counts were not significantly associated with ALT, AST, or HBV viral titer.
METHODS
Patient Population
The University of Cincinnati Infectious Diseases Center (IDC) maintains a serum bank and database of HIV-positive patients seen in the HIV clinic from 1988 to the present, a total of 3867 patients. All patients have provided informed consent under an institutional review board (IRB)-approved protocol for entry into this bank and database.
Patients with an available serum sample were included in the study. Exclusion criteria included previously documented HBV infection, interferon-based treatment at the time of serology, and nonviral liver disease (eg, known autoimmune hepatitis, nonalcoholic steatohepatitis). Prior vaccination for HBV was not an exclusion criterion, because patients with occult HBV infection may include those who did not respond to HBV vaccination.16
Patients with serum samples from 1988 through 2004 were randomly selected for HBV DNA testing. Patient information extracted from the database for this study included age, race, gender, risk factor (eg, IDU, men who have sex with men [MSM], health care worker, transfusion, hemophilia) if available, CD4+ and CD8+ cell counts, any ART, hepatitis C virus (HCV) status on the basis of an enzyme-linked immunosorbent assay (ELISA), detectable/undetectable HIV viral load, and HBV serostatus if known (including any record of prior HBV vaccination).
Laboratory Methods
Sample Collection
Samples consisted of serum or plasma separated from ethylenediaminetetraacetic acid (EDTA) anticoagulated blood. Samples were collected, processed within 4 hours of venipuncture, and stored at -70C until testing.
Hepatitis B Virus DNA Extraction
Initially, using the 909 total patient samples, 91 sample pools were created using 100-μL aliquots of each of 10 samples. HBV nucleic acid extraction was performed on the pools using the Qiagen QIAamp UltraSens Virus kit (Valencia, CA), which is designed to concentrate viral DNA from serum or plasma for increased sensitivity. Briefly, 1-mL sample pools were mixed with a lysis buffer to sediment nucleic acids into a pellet using low-speed centrifugation. After removal of the supernatant, preheated resuspension buffer and proteinase K were added to dissolve the pellet, which was then heated for 10 minutes at 40C. The lysate was then loaded onto a QIAamp spin column. HBV DNA was eluted with 60 μL of low-salt buffer and stored at 4C. Each isolation process included a known positive control.
Hepatitis B Virus DNA Amplification
Pooled DNA samples were tested for HBV DNA using a real-time PCR assay to simultaneously coamplify 2 separate regions of the HBV genome. The primers Chen forward (5'-AGT GTG GAT TCG CAC TCC T-3') and Chen reverse (5'-GAG TTC TTC TTC TAG GGG ACC TG-3') amplified a 119-base pair (bp) fragment of HBV core (positions 2269-2387).17 The primers surface forward (5'-GGA CTT CTC TCA ATT TTC TAG GG-3') and surface reverse (5'-CAA ATG GCA CTA GTA AAC TGA GC-3') were developed in-house to amplify a 432-bp fragment of the surface protein (positions 261-692). PCR assays were performed using both primer sets, and a sample was considered positive for HBV DNA if a product was found in either region. Ten microliters of eluted DNA was used in a 50-μL reaction with the Stratagene (Cedar Creek, TX) Brilliant SYBR Green QPCR Master Mix kit and 150 nM of each of the 4 primers. Amplification was performed in triplicate on a Stratagene Mx3000P Real-Time PCR system with the following thermal cycling profile: an initial 10 minutes at 95 for 1 cycle; 65 cycles of 30 seconds at 95, 30 seconds at 56, and 40 seconds at 72; and then a final 5-minute extension at 72. A dissociation profile was generated at the end of each run plotting the fluorescence negative derivative against the melting temperature to confirm the correct sizes of the PCR products generated.
If any product was amplified in the 10-specimen sample pool, the DNA isolation and subsequent real-time PCR assays were repeated on the 10 individual specimens to determine the positive sample(s). Each amplification process included a known positive control from a previous PCR run and a no-template control (PCR reaction mix and water).
Hepatitis B Virus DNA Quantification
HBV DNA viral titers were determined by running in triplicate a standard panel (OptiQuant HBV Viral DNA Panel; Acrometrix, Benicia, CA) consisting of 7 samples with HBV DNA concentrations ranging from 0 to 20,000,000 IU/mL. This panel was calibrated against the World Health Organization (WHO) International Standard for HBV DNA for use in assessing and standardizing nucleic acid test procedures. A linear relation between the sample threshold cycle (cT) and the log DNA standard concentration was observed across the range of the panel members (r = 0.95). The detection limit of the real-time HBV DNA assay was determined to be 100 IU/mL.
Hepatitis B Virus Serology
Serologic testing was performed in triplicate on those individual specimens that were HBV DNA+. Results were obtained using standard commercially available ELISA test kits for anti-HBc, anti-HBs, and HBsAg (BioChain, Hayward, CA). Some patients had prior serologic test results recorded in the database. If retest results did not correlate with prior results, or if results were close to the limit of detection, the samples were tested again. Most (2 of 3) test results were considered accurate.
Sample Size and Statistical Analysis
Sample survey techniques were implemented to facilitate DNA testing. For sample size determination within the full 3867-person database, we conservatively assumed an occult HBV prevalence of 1.5%. To detect this proportion of occult patients in a whole sample with a variance of 0.08% (50% of the prevalence) with 95% confidence, a 25% sampling fraction was sufficient, corresponding to 967 patients.
Differences in continuous outcomes between patient groups were determined by the Student t test, the Wilcoxon rank-sum test, analysis of variance with the Scheffe adjustment for multiple comparisons, or the Kruskal-Wallis test. Normality was assessed by the Shapiro-Wilk test, and data were transformed if necessary. Dichotomous outcomes were assessed by the χ2 test or Fisher exact test as appropriate. In all cases, 2-tailed P values ≦0.05 were considered statistically significant.
|
|
| |
| |
|
|
|
