Hepatitis B is a global health problem. According to the World Health Organization report, about 350 million individuals are infected with hepatitis B virus (HBV) worldwide and 1 million patients die due to hepatitis B related complications, annually.^1,2 Till date, serological assays that detected hepatitis B surface antigen (HBsAg) have been used for the diagnosis of hepatitis B infection while level of HBV deoxyribonucleic acid (DNA), alanine aminotransferase (ALT) and histological findings play a role in planning treatment and monitoring natural course of HBV. Some studies in the recent years have demonstrated that serum HBsAg level could possibly reflect the amount and transcriptional activity of covalently closed circular (ccc) DNA inside the hepatocytes which  is a non-integrated minichromosome, acting as a template for transcription of viral genes.^3-5 Assay of cccDNA in liver tissue is considered to be the most accurate index of infected hepatocytes but the complexity involved with the process of  testing, makes it an unsuitable option for routine use in clinical practice. So quantitative HBsAg (qHBsAg), for which simple and reproducible tests are available could be used as a surrogate marker for detecting transcriptional activity of cccDNA. However its correlation with other available markers of replication like HBV DNA, ALT is still debatable. Few studies tried comparing qHBsAg and HBV DNA during natural history of the disease, but outcomes were conflicting.^4,6,7 Thus there are still a lot of  unanswered questions that set hurdles and prevents generalized usage of serum HBsAg quantification in clinical practice.Therefore in this study we aimed to investigate the correlation between qHBsAg and HBV DNA during the dynamic natural history of chronic hepatitis B(CHB).

This was a prospective study evaluating treatment naive CHB patients visiting department of gastroenterology, SMS Medical College and Hospitals, Jaipur between June 2017- May 2018. Patients positive for HIV or anti-HCV, having  history of significant ethanol intake (more than 20 gm alcohol/day) or hepatotoxic drugs within last 6 months prior to enrolment, cirrhotic patients, patients with chronic hepatitis due to causes other than hepatitis B  and patients receiving immunosuppressive agents were excluded. Informed consent in writing was obtained from each patient and the study protocol confirmed to the ethical guidelines of the 1975 declaration of Helsinki as reflected in priori approved by the appropriate institutional ethics committeeon 25th May 2017 (Ethics committee SMS Medical College and Hospitals). Consecutive patients of CHB (both from outdoor and inpatient department) who fulfilled inclusion and exclusion criteria during the study period were enrolled. Demographic details and clinical parameters were recorded. Quantitative HBV DNA, qHBsAg, HBeAg, anti HBe, aspartate aminotransferase (AST), ALT levels were measured and recorded.HBV DNA was analysed using real-time polymerase chain reaction (RT-PCR) by COBAS TaqMan; Roche diagnostics, Indianapolis, IN, USA; the lower limit of detection was 6 IU/mL and higher limit was 1.6 x 10⁸ IU/ml (range- 6 - 1.6 x 10⁸ IU/ml). Due to this test limitation, any reading of HBV DNA of  <6 IU/ml was considered  as 0 while reading of  >1.6 x 10⁸ IU/ml was noted as 1.6 x 10⁸ IU/ml. Quantification of HBsAg was done using the AbbottArchitectTMHBsAg assay,  based on the use of chemiluminescent microparticles immunoassay.  HBeAg and anti-HBe testing was done using the mini-VIDAS kit (BioMerieux, Inc, USA). Based on the results patients were divided into two groups namely HBeAg positive and  HBeAg negative. The study variables of both the groups were compared. Beckman kit and auto-analyser (Beckman Coulter, Inc, USA) was used for measuring  ALT, AST, bilirubin levels. 

Numerical variables are represented in the form of mean as well as median or inter-quartile range (IQR) wherever applicable. Qualitative data are represented by means of frequency and percentage. Correlation was analysed using the non-parametric two-tailed Spearman rank correlation test where variables were not normally distributed. All p values less than 0.05 were considered significant. Numerical data was calculated using Microsoft Excel and analysed using SPSS trial version 23 software package (SPSS, Inc., Chicago, IL, USA).

A total of 90 treatment-naive HBsAg-positive patients were evaluated during the study period, out of which 73 were male and 17 were female with a median age of  23, IQR 21.75-31.75 years. Of them 24 (26.67%) patients were HBeAg positive CHB and 66 (73.33%) were HBe Agnegative. Overall  median HBV DNA was 1.1 x 10³ (IQR 2.9 x 10² -5.5 x 10³) IU/ml and  median qHBsAg  was  9.7 x 10³ ( IQR 4.5 x 10³ – 1.4 x 10⁴)IU/ml. Median HBV DNA was significantly higher in HBeAg positive CHB patients compared to HBeAg  negative, 5.5 x 10⁷ (IQR 8.0 x 10² -1.1x 10⁸) IU/ml vs 4.78 x 10² ( IQR 6.8 x 101 -3.3 x 10³) IU/ml, p value <0.001. Median qHBsAg, in HBeAgpositive group, 1.2 x 10⁴ ( IQR 7.5 x 10³ – 1.7 x 10⁴)  IU/ml was higher than HBeAg negative 9.2 x 10³ (IQR 4.3 x 10³ - 1.3 x 10⁴) IU/ml,  although  the difference was not statistically significant, p value = 0.062. Median  ALT and AST levels were significantly higher in HBeAg positive compared toHBeAg negative group, 261 (IQR 28-398)  IU/L  vs33 ( IQR 26 - 55) IU/L, p value = 0.001 and 229.5 (IQR 30-476) vs 63 ( IQR 30.75-76), p value = 0.001, respectively. Similarly median serum bilirubin was significantly higher in HBeAg positive compared to HBeAg negative group, 3mg/dl (IQR 0.9-14) vs 0.9 mg/dl (IQR 0.8-1.0), p value <0.001 but there was no statistically significant difference in level of median serum albumin between the two groups 4 gm/dl (IQR 3.9-4.1) vs 4.2 gm/dl (IQR 4.1-4.2), p value = 0.58 (Table 1). There was a strong correlation between serum qHBsAg levels and serum HBV DNA  inHBeAg positive group (r 0.961, p <0.001s) while in HBeAg negative patients there was no correlation (r -0.062 , p 0.622).  Overall there was no correlation between serum qHBsAg level and serum HBV DNA level(r -0.048, p=0.65) as shown in table 2. The distribution of chronic hepatitis B patients and HBV DNA level in different age groups in HBeAg positive and HBeAg negative patientsis shown in table 3.

Results from recent studies have suggested that qHBsAg values correlate with transcriptionally active ccc DNA and its quantification may thus indirectly reflect the number of infected hepatocytes and thus can act as a surrogate marker of viral replication.^5,8 Based on these observations, difference in serum qHBsAg level depending on the replicative phase of hepatitis B virus is expected. In the present study, HBeAg-positive CHB patients had higher HBV viral load than HBeAg negative CHB cases. Similar result has been observed inprevious studies.^9,10 Similarly, qHBsAg was noted to be higher in HBeAg positive (higher viral load) group as compared to  HBeAg negative (lower viral load) groupin concordance with a previous study, although the difference was not statistically significant.10 Similar results were observed in another study in which serum HBsAg levels were significantly higher in the group with HBV DNA > 1x10³ copies/ml compare to group with HBV DNA < 1x10³ copies/ml.¹¹ However in few studies results are contradictory to our results, and showed low qHBsAg in the presence of significant viral replication.^12,13 In the present study we found that qHBsAg correlates significantly with HBV DNA in HBeAg positive patients (r=0.96, p<0.001s) but not in HBeAg negative patients (r -0.062, p=0.622), in agreement with results from previously published studies.^14,15 Possibly suggesting that qHBsAg is more related with HBV DNA levels in initial replicative phase than in later non-replicative phases. A previous study also observed low correlation between qHBsAg and HBV DNA (r=0.597, p<0.001), however the correlation increased gradually with increase in HBV DNA level and was more significant with HBV DNA log10 >7 (r=0.597, p<0.001). In addition qHBsAg and HBV DNA levels showed significant positive correlation in HBeAg positive group (r=0.321,p<0.001) but not in HBeAg negative group (0.016, p=0.825). In this study when the assay was done with PreS1-Ag the combined predictive correlation improved.¹⁵ But contradictory results were noted in previous  studies  by the  investigators, one study asserting correlation between HBV DNA and qHBsAg  in both HBeAg positive as well as negative patients, (although correlation in HBeAg negative patients was reported to be weak), while another study refuting any correlation between HBsAg level and HBV DNA level even after subdivision into HBeAg positive and negative group.^13,16 In the present study, overall we found no correlation between serum qHBsAg and serum HBV DNA levels. Previous studies comparing overall correlation of HBV DNA with qHBsAg  have contradictory results. Few studies demonstrated significant but weak correlation while a study from Iran demonstrated no correlation between the two variable.^13,17 Lack of correlation between  the two variables in our study could likely be because of comparatively more number of HBeAg negative patients, a low replicative phase, which  had no significant correlation between serum qHBsAg  and serum HBV DNA levels on subgroup analysis. The difference in the results between the mentioned studies could probably be because of factors like heterogeneous study population, different replicative phase or regional variation in viral genotype. In our study, ALT, AST and serum bilirubin level was significantly higher in HBeAg positive compared to HBeAg negative group. Similar to our data, a study from China noted higher ALT levels in HBeAg positive compared to HBeAg negative patients.¹⁸ But likely due to heterogeneity in study population, one previous study noted higher levels of ALT, AST and bilirubin in HBeAg negative CHB patients compared to HBeAg positive CHB patients.¹⁰ Higher ALT, AST  levels  amongst HBeAg positive group noted  in our study, could possibly  be because,  most of our patients in HBeAg positive group were probably  in immunoclearance phase. The difference in results between HBeAg positive and negative patient can either be because difference in immunopathogenesis of virus itself or inadequacy of methods currently used for HBsAg quantification. In CHB, viral integration seems to occur early in infection, but at very low levels during the HBeAg positive phase of the disease. Recent studies in chronically infected chimpanzees demonstrate that there is a dramatic increase in the number of integration events in animals in the HBeAg negative phase compared to the HBeAg positive phase, so much so that >90% of the mRNA in liver of HBeAg negative chimpanzees was derived from integrated HBV sequences, with only 10% derived from the HBV minichromosome.19 This  results in positive correlation  between HBsAg titres, serum HBV DNA and liver cccDNA in HBeAg-positive patients  and lack of correlation between HBsAg levels in the serum and intrahepatic cccDNA levels in patients with HBeAg negative CHB.^14,20 In a study on CHB patients intrahepatic ccc DNA  positively correlated with serum qHBsAg (r=0.392, p=0.005), with HBV DNA (r=0.301, p=0.036) in immune-clearance phase and in HBeAg negative hepatitis phase (r=0.588, p=0.013).²¹ Existent qHBsAg assays available to detect all three forms of circulating HBsAg  (HBV, spherical and filamentous SVP)  synthesized from the viral minichromosome as well as from integrated HBV DNA.16 As antibodies used in these enzyme immunoassays target particular conserved epitopes in the S protein usually within the ‘‘a” determinant they are unable to differentiate between the three hepatitis B surface protein subtypes and their source. Additionally, if escape mutants are  predominant HBsAg quantification may become inaccurate or misleading.²² Also HBsAg may be partially masked in immune complexes, thereby influencing the results.²³ All these factors and limitations need to addressed for improving the results and exploiting complete potential of qHBsAg estimation in monitoring disease course and treatment response. There are few limitations in the present study. First, there were less number of HBeAg positive patients fulfilling the inclusion criteria compared to HBeAg negative. Second, viral genotype and escape mutant analysis was not done, however the predominant genotype in our region has been reported to be genotype D.²⁴ Third, HBV DNA quantification was not possible for values below 6 IU/ml and > 1.6 x 10⁸ IU/ml. Fourth, We did not assess the fibrosis in study groups and small sample size may also limit the conclusion of the study.Based on the results of the present study, we concluded that qHBsAg correlates with HBV DNA levels in initial replicative phase than in later non-replicative phases and is less likely to replace serum HBV DNA test in the low replicative stage of chronic hepatitis B which forms our majority of adult CHB population.