Hepatitis B surface antigen (HBsAg) levels in the natural history of hepatitis B virus (HBV)-infection: A European perspective
"In conclusion, this data demonstrates that (i) HBsAg levels are associated with the phase of HBV-infection (ii) HBsAg shows a strong correlation with HBV replication only in the early phases of infection. (iii) HBsAg levels and correlations with HBV replication differ between patients infected with HBV-genotypes A and D. (iv) HBsAg levels might be a marker predicting HBV-reactivation in low viremic HBeAg(-) carriers. Overall, these data may therefore have important implications for monitoring HBsAg during antiviral therapies as already performed for IFN treatment, and prospective studies will be important to validate these observations......The quantitative monitoring of HBsAg has been suggested as a predictor of treatment response, especially for IFN-based therapies in chronic HBV-infection , . The response to treatment with PEG-IFN shows considerable disparities between HBV-genotypes, with highest response rates in infection with HBV-genotype A and lower sustained HBV-DNA suppression and HBsAg clearance rates in patients infected with HBV-D , , . Moreover, using HBsAg as a predictor of early response, there are differences between HBeAg(+) and HBeAg(-) patients , , , . Since HBV-infection is a highly dynamic disease, this study aimed to determine HBsAg levels in different phases of HBV-infection in a cohort of European HBsAg-positive patients mainly infected with HBV-genotypes A and D. HBsAg levels and HBsAg/HBV-DNA ratios showed significant differences both during the natural course of HBV-infection and between HBV-genotypes A and D."
Jnl of Hepatology Volume 52, Issue 4, Pages 514-522 (April 2010)
Jerzy Jaroszewicz12, Beatriz Calle Serrano1, Karsten Wursthorn1, Katja Deterding1, Jerome Schlue3, Regina Raupach1, Robert Flisiak2, C.-Thomas Bock45, Michael P. Manns1, Heiner Wedemeyer1, Markus Cornberg1
1 Department of Gastroenterology, Hepatology and Endocrinology, Hannover Medical School, Carl-Neuberg-Str.1, 30625 Hannover, Germany
2 Department of Infectious Diseases and Hepatology, Medical University in Bialystok, Bialystok, Poland
3 Department of Pathology, Hannover Medical School, Hannover, Germany
4 Robert Koch Institute, Department of Infectious Diseases, Berlin, Germany
5 Department of Molecular Pathology, University Hospital of Tubingen, Tubingen, Germany
Refers to article:
A new role for an old marker, HBsAg , 01 February 2010
Maurizia Rossana Brunetto
Journal of Hepatology
April 2010 (Vol. 52, Issue 4, Pages 475-477)
Background & Aims
The quantifiable level of HBsAg has been suggested as a predictor of treatment response in chronic hepatitis B. However, there is limited information on HBsAg levels considering the dynamic natural course of HBV-infection. This study aimed to determine HBsAg levels in the different phases of HBV-infection in European HBsAg-positive patients.
226 HBV-monoinfected patients, not undergoing antiviral therapy, were analyzed in a cross-sectional study. Patients were categorized according to the phase of HBV-infection: HBeAg(+) immune tolerance phase (IT, n=30), immune clearance phase (IC, n=48), HBeAg(-) low-replicative phase (LR, n=68), HBeAg(-) hepatitis (ENH, n=68), and acute hepatitis B (n=12). HBsAg was quantified and correlated with HBV-DNA, HBV-genotypes and clinical parameters. In addition, 30 LR-patients were followed longitudinally.
HBsAg levels were higher in IT-patients and IC-patients compared to LR-patients and ENH-patients (4.96/4.37/3.09/3.87-log10IU/ml, p<0.001). HBsAg showed a strong correlation with HBV-DNA during acute hepatitis B (R=0.79, p<0.01). Correlation of HBsAg and HBV-DNA was weak or missing when analyzing different phases of persistent HBV-infection separately. However, associations between HBsAg and HBV-DNA were observed in patients infected with HBV-genotype D but not with HBV-genotype A. LR-patients with HBV-reactivation during follow-up (increase of HBV-DNA >2000IU/ml) showed >3-fold higher baseline HBsAg levels with a NPV of 95% for an HBsAg cut-off of 3500IU/ml.
HBsAg levels show significant differences during the natural course of HBV-infection and between HBV-genotypes. These findings may have important implications for understanding the natural history of HBV-infection and for using quantitative HBsAg as a diagnostic tool, i.e. as a marker for predicting HBV-reactivation.
Abbreviations: HBsAg, hepatitis B surface antigen, HBV, hepatitis B virus, HBeAg, hepatitis B early antigen, PEG-IFN, pegylated interferon alpha, cccDNA, covalently closed circular DNA, IT, immune tolerance phase, IC, immune clearance phase, LR, low-replicative phase, ENH, HBeAg(-) hepatitis, HDV, hepatitis D virus, HCV, hepatitis C virus, HIV, human immunodeficiency virus, HBcAg, hepatitis B core antigen, CI, confidence intervals
Hepatitis B surface antigen (HBsAg) was the first hepatitis B virus (HBV) protein to be discovered . Detection of HBsAg in serum is a fundamental diagnostic marker of HBV-infection. HBsAg is secreted by infected cells as subviral particles to a larger extent than infectious virons . This may serve as a possible mechanism for evading the host immune responses  while anti-HBs (anti-hepatitis B surface antigen) antibodies provide protective immunity. The loss of HBsAg and the development of anti-HBs antibodies (HBsAg-seroconversion) is the ultimate goal of anti-HBV therapy , , and thus quantitative HBsAg may be a promising prognostic marker during the natural history of HBV-infection and during antiviral therapy.
The first quantitative assays measuring HBsAg levels by use of enhanced chemiluminescence were introduced over 20years ago . However, the main drawback was a lack of appropriate standardization. Initially, HBsAg levels were expressed as serial dilutions of a reference sample from the Paul Ehrlich Institute (Langen, Germany) . During recent years new HBsAg quantitative assays have since been developed fulfilling the prerequisites of a biomarker: reproducibility, automated quantification, and a relatively low cost and standardization (IU/ml) , , . Indeed, the latest pieces of evidence support the role of HBsAg as a predictive marker for the anti-HBV treatment response. In a cohort of 386 HBeAg(-) hepatitis B patients Brunetto et al.  have shown that on-treatment HBsAg reduction of >1 log10IU/ml and levels below 10IU/ml at the end of treatment with pegylated interferon alpha-2a (PEG-IFN) were strongly associated with a sustained HBsAg clearance 3years after treatment cessation. A further study by Moucari et al.  analyzed the dynamics of HBsAg decrease during PEG-IFN therapy in HBeAg(-) hepatitis. A decrease of HBsAg level by 0.5 log10IU/ml at week 12 and of 1 log10IU/ml at week 24 of therapy was associated with positive predictive values of 89% and 92% for HBV-DNA negativity 24weeks after the end treatment. For HBeAg(+) patients, a recent study suggested that a good predictive value for treatment response was a fixed HBsAg level of 1500IU/ml after 12weeks of therapy instead of a logarithmic decline. Patients who achieved a decline below 1500IU/ml after 12weeks of PEG-IFN therapy showed HBeAg seroconversion in half of the cases, in contrast to only 16% in patients with HBsAg levels above 20,000IU/ml after 12weeks . For long-term treatment with nucleos(t)ide analogues the quantification of HBsAg may also be valuable to monitor treatment responses in patients with successful HBV-DNA suppression in order to predict HBsAg loss , .
In spite of the usefulness of HBsAg quantification as a predictor for treatment response, its clinical significance is not entirely elucidated. Previous reports have suggested that HBsAg may reflect the content of intrahepatic HBV-DNA and/or covalently closed circular (cccDNA), which serves as a intracellular template for viral RNA transcription , , . However, other studies using the same HBsAg quantification platform, could not confirm the association between HBsAg levels and cccDNA  or intrahepatic HBV-DNA  in HBeAg(-) hepatitis B.
The dissimilarity of different HBsAg-values as predictors and the discrepancy in studies analyzing associations of HBsAg and cccDNA or serum HBV-DNA  may be due to the highly dynamic nature of HBV-infection as well as the influence of HBV-genotypes on HBsAg levels. This is in agreement with variations in baseline HBsAg levels among HBV-infected patients observed in previous studies . Indeed, persistent HBV-infection has a complex and dynamic natural history, as reviewed in , , , and can be divided into four phases based on the evolution of the virus and the host immune responses: immune tolerance phase (IT), immune clearance phase (IC), low-replicative phase (LR) and HBeAg(-) hepatitis (ENH). The impact of the natural course of HBV-infection is underappreciated in many studies investigating prognostic markers of anti-HBV treatment response.
Hence, the aim of this study was to evaluate (i) HBsAg levels in consecutive phases of natural history of HBV-infection without the influence of ongoing antiviral treatment, (ii) the association between HBsAg levels and HBV-genotypes, and (iii) other virological, biochemical and clinical parameters in European patients with persistent HBV-infection. For controls, we further analyzed HBsAg levels longitudinally in patients with acute hepatitis B and in patients with low replicative HBV-infection who had no indication for antiviral therapy during the follow-up.
Patients and methods
Two hundred and fourteen European patients with persistent HBV-infection, not undergoing antiviral therapy, were included in this cross-sectional study. Patients were recruited in two reference university hospitals: Hannover Medical School, Germany (n=168) and Medical University of Bialystok, Poland (n=46). Seventy-eight patients were HBeAg(+) and 136 HBeAg(-). There were 140 male and 74 female with a median age of 37years (min. 2, max. 85 yrs). None of the individuals included had hepatitis delta virus (HDV), hepatitis C virus (HCV) or human immunodeficiency virus (HIV) co-infection. Further exclusion criteria were end-stage liver insufficiency, autoimmune disorders, immunosuppressive treatment and malignancies. Patients were categorized according to the phase of persistent HBV-infection: HBeAg(+) immune tolerance phase (IT, n=30), immune clearance phase (IC, n=48), as well as HBeAg(-) low-replicative phase (LR, n=68) and HBeAg(-) hepatitis (ENH, n=68). The criteria were based on the German guidelines for the management of hepatitis B virus infection  (Table 1). In order to assure proper classification 30 HBeAg(-) patients in the LR-phase were followed for a median period of 14months (min. 6months, max. 60months). Furthermore, 12 patients with acute hepatitis B (9 male, 3 female; median age 40years) were included with a median follow-up of 28days after diagnosis.
Serum HBsAg quantification
Serum HBsAg levels were quantified using the Abbott ARCHITECT assay (Abbott Diagnostics, Abbott Park, IL). The test has a dynamic range of 0.05-250IU/ml. Samples were diluted 1:100 in horse serum and if >250IU/ml, samples were retested at a dilution of 1:500 and 1:1000, respectively. Samples with HBsAg levels <0.05IU/ml have been retested without prior dilution. Results are given in IU/ml.
Serum HBV-DNA was measured by use of COBAS AmpliPrep/COBAS TaqMan (Roche Diagnostics, Mannheim, Germany) and TaqMan Universal Master Mix (Applied Biosystems, Foster City, CA) with detection limit of 12IU/ml and 50IU/ml, respectively. HBV-DNA levels are expressed in IU/ml.
HBV genotyping was performed in 142 patients. DNA was extracted from 200µl of serum by High Pure Viral Nucleic Acid Kit (Roche Diagnostics, Mannheim, Germany) according to manufacturer's protocol. HBV-genotype was determined by a modified RFLP PCR assay with amplified HBV-DNA by use of nested PCR as previous described . The sequences obtained were matched with the National Center for Biotechnology Information GenBank and compared with described HBV-prototypes (Accession No. for: HBV-A Z72478; HBV-B D00329; HBV-C X01587; HBV-D V01460; HBV-E X75657; HBV-F X75658; HBV-G AF160501). The HBV S sequences and related references sequences were aligned with CLUSTALW Multiple alignment tool (The BioEdit Sequence Alignment Editor software, Department of Microbiology, North California State University). Bootstrap resampling and reconstruction were carried out 1000 times. Genetic distance was estimated using the DNADIST distance matrix software (University of Washington).
HBcAg and HBsAg expression in the liver
In a subset of 26 patients (8 HBeAg(+) and 18 HBeAg(-)), HBcAg and HBsAg expression was assessed in liver biopsies by immunohistological staining. Slides from paraffin embedded specimens were processed with the ZytoChem-Plus HRP Polymer-Kit (Zytomed Systems GmbH, Berlin, Germany) and a DAB Substrate High Contrast kit (Zytomed Systems GmbH, Berlin, Germany). Sections (2µm) were deparaffinised, treated with 3% H2O2 for 10min and incubated for 1h with antibodies against HBsAg (monoclonal mouse, ready to use, clone 3E7, PDM079, Diagnostic BioSystems, CA, USA) and HBcAg (polyclonal rabbit, 1:200 dilution, B0586, DakoCytomation, Hamburg, Germany). For the measure of antigens expression semi-quantitative, 6-grade scale based on percentage of HBcAg or HBsAg positive hepatocytes was applied (0: none, 1: <5%, 2: 6-25%, 3: 26-50%, 4: 50-75%, 5: 76-100%). The immunohistological procedure was performed by pathologist blinded to clinical data (J.S.).
Data are presented as medians (10-90% Confidence Intervals, CI), unless indicated. The following non-parametric (distribution-free) tests were applied: Mann-Whitney U and Fisher's exact test for univariate, Kruskall-Wallis ANOVA for multivariate comparisons and Spearman rank test for correlations analyses. Multiple regression with HBsAg concentration as an independent factor was carried out. Statistical analyses were performed by use of Statistica 8.0 (Statsoft, Tulsa, USA).
Two hundred and fourteen HBsAg-positive patients were divided into four consecutive phases of persistent HBV-infection. HBeAg(+) immune tolerant patients and HBeAg(+) patients in the IC-phase were significantly younger than patients in LR-phase or patients with HBeAg(-) hepatitis. Gender and HBV-genotype distribution was comparable between the phases of persistent HBV-infection. Detailed description of the studied population is given in Table 2.
Distribution of HBsAg levels
Serum HBsAg levels varied significantly between patients in different phases of persistent HBV-infection in both univariate and multivariate analyses (p<0.001). The particularly high levels of HBsAg (median 90,881IU/ml) were observed in young, immune tolerant individuals. In the IC-group median HBsAg was 23,371IU/ml, while in LR-patients HBsAg was moderate and almost 20-fold lower (1230IU/ml), and 3-fold lower in ENH (7545IU/ml) (Table 2, Fig. 1A). Seventy percent of individuals in the IT-group had serum HBsAg levels over 50,000IU/ml, and 66% of IC-patients had HBsAg between 15,000 and 50,000IU/ml. In 80% of individuals in the LR-phase, HBsAg was below 5000IU/ml, while in HBeAg(-) hepatitis HBsAg levels were widely distributed (Fig. 1B). Interestingly, some of the individuals in the LR-phase showed relatively high HBsAg levels exceeding 10,000IU/ml.
HBsAg quantification and HBV-genotype
HBV genotyping was performed in 66% of patients (n=142). The most prevalent was HBV-genotype D (HBV-D) (n=85, 59.8%) and HBV-genotype A (HBV-A) (n=36, 25.3%). Consistent with previous data , HBV-A was more prevalent in Poland than in Germany (58% vs. 18%, p<0.001). The distribution of HBV-genotypes was comparable in IC, LR and ENH-phases, while none of patients in the IT-phase was infected with HBV-A (Table 2). There were no significant differences in serum HBsAg and HBV-DNA levels between individuals infected with HBV-D and HBV-A in the overall cohort (4.21 vs. 3.91 log10IU/ml, p=0.31 and 6.81 vs. 5.16 log10IU/ml, p=0.10, respectively). However, analyzing patients in the IC-phase, HBsAg level were slightly higher in patients infected with HBV-D than HBV-A (4.46 vs. 4.3 log10IU/ml, p=0.18). In patients with ENH, this difference was reverted, although not statistically significant (3.87 vs. 4.17 log10IU/ml, p=0.07) (Fig. 2).
Correlation between serum HBsAg and HBV-DNA
In patients with persistent HBV-infection, serum HBsAg was correlated with HBV-DNA levels (R=0.75, p<0.001); although this strong association was no longer observed in consecutive phases of HBV-infection (Fig. 3A, Table 3). However, when analyzing subgroups of patients with different genotypes, HBsAg and HBV-DNA levels showed a positive correlation in patients infected with HBV-D during IC-, LR, and ENH-phases, but was not the case for patients infected with HBV-A (Fig. 3B). For all patients with persistent HBV-infection, the correlation of HBsAg and HBV-DNA was highly significant for HBV-D (R=0.82, p<0.001), but was not in patients infected with HBV-A (R=0.28, p=0.13).
Interestingly, the ratio of HBsAg/HBV-DNA, which reflects the association between HBsAg production and HBV replication, was significantly higher in LR-patients in comparison to other phases of persistent HBV-infection (1.17 vs. 0.57 in IT-group, 0.60 in IC and 0.64 in ENH, p<0.001) (Fig. 3C). Again, this parameter displayed a wide distribution in LR-patients, thus suggesting that production of HBsAg is well preserved in selected HBeAg(-) patients with low HBV replication. Moreover, the ratio of HBsAg/HBV-DNA was higher in HBV-A in comparison to patients infected with HBV-D in the ENH-phase (0.76 vs. 0.58, p=0.05) and in LR-patients (1.14 vs. 0.91, p=0.03), which was not the case when analyzing patients in the IC-phase (0.61 vs. 0.56, p=0.85).
Correlation between serum HBsAg and clinical parameters
In none of the phases of persistent HBV-infection, serum HBsAg was associated with ALT activity (Table 3). We did not observe significant correlations of serum HBsAg levels with gender nor with INR, serum bilirubin, albumins, and sodium or cholinesterase levels. Serum HBsAg quantification showed a negative association with age (R=-0.48, p<0.001) (Table 3). Interestingly, HBsAg levels were positively associated with platelet count in the IC-phase (R=0.34, p=0.03) and ENH-phase (R=0.30, p=0.02) i.e. only in phases with active inflammatory activity. Overall, the multiple logistic regression analysis revealed that the only independent factors associated with serum HBsAg quantification were the patients' age and platelet count (ß=-0.44 and 0.28, respectively, p<0.001), while ALT and HBV-DNA were not significant.
HBsAg levels in acute hepatitis B
In order to investigate the kinetics and correlations of HBsAg in the early phase of HBV-infection, twelve patients with acute hepatitis B were included and followed-up for a median time of 28days. At baseline (diagnosis of acute HBV), serum HBsAg showed a strong, positive correlation with HBV-DNA (R=0.79, p=0.002, Fig. 4D), while no association with age (R=0.06, p=0.84) or ALT activity (R=0.18, p=0.57) was observed. In contrast to persistent HBV-infection, HBsAg dynamics were associated with the decrease of HBV-DNA in the early phase of HBV-infection (Fig. 4A-C). The median log10 decrease of serum HBsAg and HBV-DNA levels at the first follow-up (median 14days) were: 0.59 (0.13-1.63) and 1.15 (0.40-2.39) IU/ml, respectively and at the second time-point (median 28days): 1.86 (0.78-4.80) and 1.75 (0.83-4.16) IU/ml, respectively.
HBsAg during the follow-up of HBeAg(-) patients in the low-replicative phase
The relatively wide distribution of HBsAg levels in HBeAg(-) patients with a low HBV replication (Fig. 1) inspired us to perform a longitudinal analysis in those patients. Follow-up data were available in 30 LR-individuals for a median duration of 14months (min. 6, max. 60months). During this period, two patients became seronegative for HBsAg, 10 (33%) showed stable values of HBV-DNA (HBV-DNA change <0.5 log10), while 13 (43%) had fluctuations in HBV-DNA (0.5-0.9 log10 change) and 5 had a reactivation of the HBV with an increase in HBV-DNA of more than 1 log10 and exceeding 2000IU/ml; four out of 5 patients with a reactivation had HBsAg levels over 3500IU/ml. Interestingly, individuals with HBV-reactivation had over 3-fold higher baseline HBsAg quantity (4116IU/ml) than those with a fluctuating HBV-DNA of less than 1 log10 change (1138IU/ml, p=0.03).
HBsAg and HBcAg expression in the liver
Liver HBcAg expression showed an association with the phase of persistent HBV-infection (IC: 2.3, LR: 0.0, RE: 0.9, p=0.01). Accordingly, median liver HBsAg expression was lowest in LR-patients versus IC and ENH-patients (0.5 vs. 1.4 and 1.6, respectively), although this difference was not significant (Fig. 5). In this relatively small cohort of patients analyzed, no direct association between serum HBsAg levels and liver HBsAg expression was observed.
The quantitative monitoring of HBsAg has been suggested as a predictor of treatment response, especially for IFN-based therapies in chronic HBV-infection , . The response to treatment with PEG-IFN shows considerable disparities between HBV-genotypes, with highest response rates in infection with HBV-genotype A and lower sustained HBV-DNA suppression and HBsAg clearance rates in patients infected with HBV-D , , . Moreover, using HBsAg as a predictor of early response, there are differences between HBeAg(+) and HBeAg(-) patients , , , . Since HBV-infection is a highly dynamic disease, this study aimed to determine HBsAg levels in different phases of HBV-infection in a cohort of European HBsAg-positive patients mainly infected with HBV-genotypes A and D. HBsAg levels and HBsAg/HBV-DNA ratios showed significant differences both during the natural course of HBV-infection and between HBV-genotypes A and D.
HBsAg levels were higher in HBeAg(+) patients compared to HBeAg(-) patients and HBsAg correlated with HBV-DNA in the entire study cohort. However, the correlation of HBsAg and HBV-DNA was weak or missing when analyzing the different phases of persistent HBV-infection separately. This dissociation of HBV-DNA and HBsAg in a narrow window of time for each HBV-phase may reflect a disconnection between HBV replication and HBsAg production during persistent HBV-infection. These differences then accumulate over time during the natural history of persistent HBV-infection; they are most likely due to various reasons such as the integration of HBV into the host genome ,  that potentially provides a separate template for the production of HBsAg, or the cytokine dependent modification of viral replication pathways  leading to the disruption in stability of cytoplasmic viral capsids . Overall this dissociation of HBsAg production and HBV replication does not seem evident in the early phase of HBV-infection as our data of patients with acute hepatitis B show a strong correlation of HBsAg and HBV-DNA (Fig. 4). In agreement with this idea, the accompanying paper by Nguyen et al., investigating Asian patients with HBV-genotype B and C, also showed a strong correlation of HBsAg and HBV-DNA during the immune clearance phase. Surprisingly, we could not reproduce these results in our cohort of European patients within the IC-phase. However, when patients were analyzed separately for HBV-genotypes A and D, we could document a correlation between HBsAg and HBV-DNA in patients infected with HBV-D but not in patients infected with HBV-A (Fig. 3B). Hence, HBsAg seems to correlate with HBV-DNA during the immune clearance phase in patients infected with HBV-genotypes B, C and D but not with HBV-A.
HBeAg(-) hepatitis B is usually associated with lower intrahepatic cccDNA levels , , . It is therefore not surprising that HBsAg levels are significantly lower in HBeAg(-) patients than in HBeAg(+) patients, as also documented by other investigators , , . Nevertheless, HBsAg/HBV-DNA ratios did not differ between HBeAg(-) hepatitis and HBeAg(+) patients (IT and IC) (Fig. 3C). This finding is in agreement with data showing that the production of subviral particles is not impaired in HBeAg(-) hepatitis .
In line with the concept that HBsAg production and HBV replication are disconnected in later phases of HBV-infection, there was no correlation of HBsAg and HBV-DNA in the ENH cohort. However, we observed a moderate correlation of HBV-DNA and HBsAg in HBV-D patients but not in HBV-A patients, similar to our findings in IC-patients (Fig. 3B). Our data suggest important differences in HBsAg production between patients infected with HBV-genotypes A and D. A higher transcription efficiency of preS/S mRNA by subgenotypes A has been described which may account for higher synthesis of HBsAg in relation to HBV-DNA in HBV-A patients . Accordingly, ENH-patients with HBV-A tended to have higher HBsAg levels, and a higher HBsAg/HBV-DNA ratio than in ENH-patients with HBV-D (Fig. 2). Another reason for the differences between genotypes may be that there is a higher susceptibility of HBV-D to acquire pre-core or basal core promoter variants ; whether these variants have any influence on HBsAg production warrants further investigation. A third explanation for the differences between genotype A and other genotypes may be that there are differences in innate or adaptive immune responses. Patients infected with HBV-A may show stronger immune response as reflected by a better response to IFN-based therapies , and this may in turn explain the different dynamics of HBsAg levels and the lack of correlation with HBV-DNA, especially in the IC-phase.
The levels of HBsAg were lowest in HBeAg(-) patients in the LR-phase, thus representing the immune control of HBV-infection. However, 20% of LR-patients showed HBsAg levels above 5000IU/ml and the 10 patients who had documented HBV-DNA below the detection limit of the TaqMan PCR still showed HBsAg levels of up to 3600IU/ml (median 1358IU/ml, data not shown). Subsequently, the HBsAg/HBV-DNA ratio was highest in the LR-phase among all HBV-phases. This suggests that immune control of HBV replication does not necessarily impair HBsAg production, possibly due to HBV integration into the host genome . These findings are supported by the immunohistochemical analysis of HBsAg and HBcAg in the liver. While HBcAg as a marker of ongoing HBV replication was only absent in LR-patients, HBsAg could be detected in liver biopsies of IC, ENH and LR-patients (Fig. 5).
HBsAg monitoring of LR-patents may also have clinical implications. Our preliminary follow-up data of a subgroup of 30 LR-patients identified quantitative HBsAg as a potential predictive marker, identifying individuals with an increased risk for HBV-DNA reactivation. The reactivation rate has recently been determined to be 4.3% per year in asymptomatic HBeAg(-) patients . A cut-off for HBsAg, in order to identify those who are at risk or without risk for reactivation, would be practical and more valuable than an HBV-DNA cut-off. Here we propose a level of 3500IU/ml HBsAg with negative predictive value for reactivation of 95% in our small case series. Thus, we strongly suggest including HBsAg monitoring in future prospective trials determining the risk of hepatitis B reactivation in LR-patients.
In multiple logistic regression analyses, HBsAg levels were independently associated only with age and platelets (Table 3). While age may reflect the natural course of HBV-infection over the different phases of HBV-infection, an association between HBsAg and platelets has not been shown before. Importantly, this positive correlation between platelets and HBsAg levels was only evident in phases with inflammatory activity (IC and ENH) suggesting a role of platelets for the immune response in HBV-infection. A link between platelets and immune control of HBV-infection has recently been documented in animal models , . This first description of an association of platelets and HBsAg levels in humans may therefore be an interesting observation for future studies.
Some limitations of our study need to be considered. Firstly, the cross-sectional design of the study requires long-term follow-up of individuals since HBV-infection is a highly dynamic disease, and an appropriate follow-up is of highest importance especially in low replicative HBeAg(-) patients because of the possible fluctuating profile . In this study appropriate follow-up data was available in 30 out of 68 LR-patients and therefore some of the apparently LR-patients may have been misclassified; however, even a follow-up over a period of 12months may oversee fluctuating profiles. This emphasizes the need for HBsAg cut-offs and validation in prospective studies with long-term longitudinal follow-up designs. Secondly, this study included only European patients, mainly infected with genotypes A and D and obviously similar data on other HBV-genotypes would be highly desirable. Thirdly, intrahepatic virologic profiles including cccDNA, pregenomic RNA and replicative RNA intermediates of HBV were not studied here. Thus, no definite conclusions on virion productivity in the different phases of HBV-infection can be drawn from this study.
In conclusion, this data demonstrates that (i) HBsAg levels are associated with the phase of HBV-infection (ii) HBsAg shows a strong correlation with HBV replication only in the early phases of infection. (iii) HBsAg levels and correlations with HBV replication differ between patients infected with HBV-genotypes A and D. (iv) HBsAg levels might be a marker predicting HBV-reactivation in low viremic HBeAg(-) carriers. Overall, these data may therefore have important implications for monitoring HBsAg during antiviral therapies as already performed for IFN treatment, and prospective studies will be important to validate these observations.