Keywords

1 Introduction

Hepatitis B virus (HBV) infection is typically classified into acute infection with full recovery and chronic infection with life-long carriage state. It is generally believed that the former has no long-term clinical sequelae, whereas the latter is associated with a considerable risk of development of long-term liver-related complications. These include disease reactivation with hepatitis flares, cirrhosis leading to liver decompensation and/or hepatocellular carcinoma (HCC). To distinguish these two prognostically distinct conditions, the presence or absence of HBsAg of more than 6 months after the initial infection is the criterion. In other words, disappearance of HBsAg measured by conventional assays within 6 months signifies acute HBV infection, whereas persistence of HBsAg beyond this duration signifies chronic HBV infection. However, challenges were brought out by studies showing that HBV DNA were still present in serum and/or liver of subjects after decades of documented acute HBV infection (Raimondo et al. 2008a, 2013; Michalak et al. 1994; Yotsuyanagi et al. 1998; Yuki et al. 2003; Rehermann et al. 1996; Bläckberg and Kidd-Ljunggren 2000). One of these studies also documented the presence of HBV DNA in the peripheral blood mononuclear cells of subjects up to 70 months after recovery of acute hepatitis B infection (Michalak et al. 1994). This special condition is under the category of “occult hepatitis B infection” (OBI) which was consensually defined in a meeting held in Taormina 2008 as HBsAg-negative subjects with detectable HBV DNA either in serum or the liver (Raimondo et al. 2008b). In fact, this entity of OBI has been firstly described back in 1979 (Tabor and Gerety 1979).

It is noteworthy that majority of patients with OBI are not subjects with acute HBV with persistent viraemia. Patients with chronic hepatitis B (CHB) who lost HBsAg (HBsAg seroclearance) either naturally or through drug treatment indeed attribute for the majority of this population. There is also a small proportion of patients with CHB in which the HBV has hepatitis B surface gene mutations altering the epitopes of HBsAg leading to non-reactivity to the conventional HBsAg assays (Yamamoto et al. 1994; Hou et al. 1995; Carman et al. 1997). This is considered to be the minor group because studies have shown that the HBV in most of the OBI subjects do not have corresponding mutations in the HBsAg genome (Pollicino et al. 2004, 2007; Huang et al. 2014;).

2 Classifications of OBI Patients

OBI patients can be classified by two different ways. First, patients can be classified according to the serology of HBV. Second, patients can be classified according to their medical histories of HBV infection.

2.1 HBV Serology

Sero-positive OBI refers to the presence of anti-HBs and/or anti-HBc in the serum. On the contrary, absence of detectable serum anti-HBs and anti-HBc indicates sero-negative OBI . The hypothesis to explain the total absence of antibodies in sero-negative OBI is the insufficient specific immune response in the presence of a very minute level of hepatitis B viraemia. According to several studies, up to 20% of OBI subjects are sero-negative (Morales-Romero et al. 2014; Raimondo et al. 2007). At present, there are no detailed studies to compare sero-positive with sero-negative OBI in terms of the disease underlying mechanisms and the clinical disease profiles.

2.2 Medical History of HBV Infection

2.2.1 Subjects with Documented History of Acute HBV Infection

More than 95% adult subjects contracting HBV is associated with “clinical” recovery without HBV chronicity. They would have anti-HBs and anti-HBc positivity with or without clinical symptoms of malaise, loss of appetite and jaundice. Although it is believed that these subjects are free of developing long-term clinical sequelae, it has been shown that HBV DNA was detectable in some subjects even after 70 months of acute HBV infection (Michalak et al. 1994). The findings were subsequently confirmed by two other studies in which HBV DNA were still detectable 6 months to 13 years after the acute HBV infection (Yotsuyanagi et al. 1998; Penna et al. 1996). In spite of the presence of low-grade viraemia in these subjects, their livers are usually with minimal necroinflammation (Yuki et al. 2003; Bläckberg and Kidd-Ljunggren 2000). As such, nearly all these subjects would have normal liver biochemistry. These are compatible with the findings in studies of woodchucks recovering from acute HBV infection with detectable viraemia in that the degree of liver inflammation was low (Mulrooney-Cousins and Michalak 2007).

The clinical significance of OBI subjects after acute HBV infection remains obscure. It is however expected that the chance of developing long-term liver-related complications is low basing on the fact that these subjects have very low viraemic levels, normal histology and alanine aminotransferase levels. However, the clinical implications become important in the aspect of HBV transmission to uninfected subjects through blood product donations. In addition, with the more profound immunosuppressive effects exerted by B-cell depleting monoclonal antibodies for various malignant and immunological diseases, the problem of HBV reactivation as a result of immune reconstitution in these subjects has been increasingly recognized (Yuen 2016). These topics will be discussed below.

2.2.2 Subjects with Chronic Hepatitis B Infection

As mentioned above, majority of subjects with OBI are CHB patients who undergo an uncommon event of HBsAg seroclearance. This event happens either spontaneously or with the assistance of antiviral therapy. In the former situation, patients undergo the different phase of CHB infection from immunotolerance through immune clearance before entering into the residual phase where HBsAg seroclearance may occur (Fig. 15.1). It has been shown that HBsAg seroclearance occurs in these subjects at an annual rate of 0.1–2% (Lai and Yuen 2009; Liu et al. 2010). In patients receiving antiviral therapy, the rate of HBsAg seroclearance is around 10% in long-term follow-up studies (Yuen et al. 2016).

Fig. 15.1
figure 1

Different phases of chronic hepatitis B infection

Full size image

Around 60% of CHB patients with HBsAg seroclearance would develop anti-HBs at the time of HBsAg seroclearance or at the subsequent follow-up after HBsAg seroclearance (Yuen et al. 2008). CHB patients with HBsAg seroclearance usually have undetectable HBV DNA in the serum. The serum HBV DNA detection rate inversely correlates with the time of measurement after HBsAg seroclearance. For instance, serum HBV DNA was detectable in 13%, 6% and 3.7% when the measurements were carried out at 1, 5–10 and >10 years after HBsAg seroclearance (Yuen et al. 2008). It is worth to note that the intrahepatic HBV DNA was detectable (although at low levels) practically in all subjects with HBsAg seroclearance at all times (Yuen et al. 2008; Ahn et al. 2005). The HBV transcriptional activities however are minimal as indicated by negligible mRNA expression of the core, surface and X genes (Yuen et al. 2008; Kuhns et al. 1992). It is therefore an expected observation among 29 patients with liver biopsies that two-third of them had no necroinflammation and fibrosis and the remaining one-third only had mild necroinflammation or fibrosis (Yuen et al. 2008). It has further been shown that for those with early HBsAg seroclearance, e.g. <40 years old, none of these subjects had fibrosis in the liver (Yuen et al. 2004).

With regard to the chance of the development of long-term liver-related complications in CHB patients with HBsAg seroclearance, it largely depends on the clinical status at the time of HBsAg seroclearance. According to an early study (Huo et al. 1998), HBsAg seroclearance did not reduce the risk of development of complications as the cumulative probability of developing them was as high as 30% after 4 years of HBsAg seroclearance. Nevertheless, a subsequent study showed that this risk would be lower if there were no other viral co-infection (e.g. HCV infection) and no pre-existing cirrhosis (Chen et al. 2002). A study conducted in Hong Kong showed that those who achieved HBsAg seroclearance before the age of 50 years had a significantly lower risk of development of complications compared with patients who had the HBsAg seroclearance after the age of 50 years (Yuen et al. 2008). None of the former group developed HCC compared to 10% (p = 0.004) of patients developing HCC in the latter group at the median follow-up of 3 years after HBsAg seroclearance. It is likely related to the lower proportion of patients having severe fibrosis or cirrhosis when the HBsAg seroclearance occurring at a younger age.

2.2.3 Subjects with Unknown HBV History

There are subjects in whom OBI is diagnosed by routine screening especially during the blood donation surveillance . There are many blood donation centres worldwide routinely performing nuclear acid testing either in pooled or individual samples which are firstly tested negative for HBsAg. The absence of the medical history of HBV infection in these subjects makes it impossible to differentiate unnoticed clinically silent acute HBV subjects from CHB patients with HBsAg seroclearance who are never known to themselves or their physicians.

In the context of OBI subjects diagnosed incidentally in the blood donation services, a study with 40 OBI blood donors showed that the liver biochemistry was normal in all subjects. They also had negligible liver necroinflammation (median necroinflammatory score: 1 out of 18) and fibrosis (median Ishak’s fibrosis score: 1 out of 6) (Wong et al. 2016). The viral replicative status was extremely low. Only 13% of OBI subjects had detectable pre-genomic RNA, and the median intrahepatic HBV DNA was 0.22 copies/cell compared to 260 copies/cell (p < 0.0001) in age- and sex-matched HBsAg-positive CHB patients. Only 45% of OBI subjects had serum HBV DNA > 1.1 IU/mL.

3 Prevalence of OBI

The true prevalence of OBI worldwide has not been well studied because of several unresolved issues. First, the HBV DNA level in serum is notoriously low for it to be measurable by standard commercial assays. Although the HBV DNA level was described to be usually less than 200 IU/mL (Morales-Romero et al. 2014), studies have shown that the HBV DNA levels were typically less than 20 IU/mL (Yuen et al. 2010, 2011). Second, the fluctuations of low HBV DNA levels in OBI subjects often cast doubts on the reliability of the sensitive HBV DNA assays. These two difficulties may be partly overcome by using large-volume samples for testing. Finally, although the yield of detectable HBV DNA can be greatly enhanced by measuring it in the liver tissues, invasive nature of liver biopsy makes it an unpopular methodology in many studies. With all these issues to be considered, prevalence of OBI was estimated to be <1 to 18% (Yuen et al. 2008, 2010; Song et al. 2009; Bhatti et al. 2007; Werle-Lapostolle et al. 2004; Reesink et al. 2008; Georgiadou et al. 2004; Fang et al. 2009; Minuk et al. 2005; Kim et al. 2007; Svicher et al. 2012). According to a large-scale study in Hong Kong with 9990 blood donors, the prevalence of OBI was found to be 0.11% (Yuen et al. 2010). It suggests that the prevalence of OBI may not be as high as one would expect even in region with high CHB endemicity.

4 Pathogenesis of OBI

There are several mechanistic postulations for the existence of OBI. These include the viral and host factors. It has long been proposed that mutations over the major surface gene, e.g. G145R, may lead to detection escape from the standard HBsAg assays (Yamamoto et al. 1994; Hou et al. 1995; Carman et al. 1997). It is however not supported by three studies in which corresponding mutations as suggested were not found in OBI subjects (Pollicino et al. 2004, 2007; Huang et al. 2014). In addition, a study also showed that most of the standard assays were able to detect various mutated HBsAg epitopes (Yuen and Lai 2006). Another hypothesis is the persistently low viral replication in OBI subjects because of two possible but interlinking factors. The HBV may be intrinsically replicate at an extremely low level because of multiple genomic mutations or rearrangements in splice donor sites of pre-S1, pre-S2, S genes as well as their regulatory genomic regions (Candotti et al. 2012; El Chaar et al. 2010). This is supported by studies showing greater sequence diversities in both nucleoside and amino acid levels in subjects with OBI compared with HBsAg-positive CHB subjects (Huang et al. 2014; Pollicino et al. 2004; Lai and Yuen 2009). The viral replication capacity may be much restrained with all these changes which may be additive or synergistic. At the post-transcriptional level, it was found that Pre-S2/S RNA splicing may also reduce the pre-S2/S and HBsAg transcripts (Candotti et al. 2012; Hass et al. 2005). Epigenetic factor exerting on the virus has also been postulated. It has been shown that the cccDNA-bound histones in OBI subjects were highly hypoacetylated which may result in low HBV replication (Pollicino et al. 2006).

The other factor interlinking with viral replication is the more intense immune control of the HBV in OBI subjects. There is an indirect but very strong evidence from the phenomenon of HBV reactivation in HBsAg-negative, anti-HBc-positive subjects (potential OBI subjects) under immunosuppressive therapy especially with potent anti-B-cell depletive effect, e.g. anti-CD 20 (Larrubia 2011; Bes et al. 2012). It suggests that OBI status is successfully kept by adequate immune control and the virus reactivates upon withdrawal of this controlling factor under immunosuppression. To explore this, studies to compare the HBV-specific T-cell responses between OBI and CHB subjects are deemed to be important. In fact, intense immune response has been shown to be detectable many years after the recovery of acute HBV infection (Rehermann et al. 1996; Penna et al. 1996).

Finally, there may be external factors inhibiting HBV replication, e.g. co-infection with other viruses, leading to the OBI state. In HCV-HBV co-infection, HBV replication is often diminished because HCV core protein has been shown to exert inhibitory effects on HBV replication (Chen et al. 2003; Raimondo et al. 2005; Schüttler et al. 2002; Shih et al. 1995). It seems to be in accordance to the observation that OBI is apparently more commonly found in HCV patients than other populations. Therefore, HCV may play a role in inducing OBI status (Pollicino and Raimondo 2014).

5 Clinical Significance of OBI

There are three major clinical areas where OBI plays an important role (Fig. 15.2). They include (1) the possibility of transmission of HBV disease through blood and organ donations, (2) the pathologic role on chronic liver disease and liver-related complications and (3) the HBV reactivation during immunosuppressive therapy.

Fig. 15.2
figure 2

Clinical scenarios of occult hepatitis B infection

Full size image

5.1 Transmission of HBV

OBI donors, although usually having very low HBV viraemia, may transmit the HBV to other individuals in the setting of blood donation and organ transplantation. It is because any products containing full HBV viral particles are considered to be potentially infectious. According to a chimpanzee study, HBV DNA of only ten copies can already achieve the minimum 50% infectious dose of HBV (Komiya et al. 2008). In another study using severe immunodeficient chimeric mice with human hepatocytes, serum HBV DNA, intrahepatic HBV DNA as well as covalently closed circular (ccc) DNA (the viral replication template) were detectable in one of the four mice inoculated with the sera from OBI subjects (Yuen et al. 2011). The mouse hepatocytes were also stained positive for HBcAg. Whether these findings are applicable to human would require actual human studies to verify as the immune components in animals are different and the mice are basically immunodeficient. In addition, there are other confounding factors in human situations influencing the chance of HBV transmission from OBI subjects. These include anti-HBs status of both the donors and the recipients, the kind and the volume of the blood products being transfused to the recipients and the immunological status of the recipients (Raimondo et al. 2013; Mosley et al. 1995; Satake et al. 2007). Finally, the actual circulating viral load (which may fluctuate with times) of the OBI donors at the time of donation may also affect the outcome (Chemin et al. 2009).

Although there are already many human studies examining the transmissibility rate of HBV from OBI blood donors, most of these studies are retrospective in nature. This greatly limits the effectiveness of contact tracing to the recipients. In areas where HBV is of high endemicity, meticulous phylogenetic analysis of the viral sequence between the donors and recipients which are especially essential to establish the HBV transmission is required. This is however difficult to achieve in context of the retrospective nature of the studies. In addition, the difficulty in obtaining good quality of viral sequence (from amplification of the low HBV DNA levels) also impinges the study accuracy.

Nevertheless, even with all these problematic issues, there are several studies performed to determine the HBV transmission rate from OBI donors. It was found that the transmission rate was low at around 1–3% (Yuen et al. 2011; Mosley et al. 1995; Candotti and Allain 2009). One the studies further showed that the risk was greatly reduced if the donor serum was anti-HBs positive (Mosley et al. 1995). From a study conducted in Japan, of the identified 12 subjects with transfusion-related hepatitis B, all received the blood products from donors who were negative for both anti-HBs and anti-HBc (Satake et al. 2007). It is likely that these donors were at the window period of the acute HBV infection when all the antibodies were yet to be detectable. This group of donors may also be sero-negative OBI donors if their sera remain to be negative for both anti-HBs and anti-HBc.

There are studies showing that the HBV transmission is possible from anti-HBc-positive donors (Hoofnagle et al. 1978; Lander et al. 1978; Koziol et al. 1986). The transmissible rate is around 2.4–3.0% according to one study (Lai and Yuen 2009). It is compatible with another study conducted in Hong Kong using detailed phylogenetic relatedness methodology. This study showed that the transmission rate of 2.2% (Yuen et al. 2011).

In the context of organ transplantation in which the recipients receive organs from OBI donors, HBV transmission may also be of concern. Unlike the situation in blood product donations given to usually immunocompetent recipients. Patients receiving organ transplantation have an additional risk factor of acquiring the HBV from OBI donors because of the immunosuppressive status rendered by the immunosuppressive therapy for prevention of organ rejection. In the setting of liver transplantation, it has been estimated that 22–100% of liver-transplanted recipients would have de novo HBV infection of the liver grafts from HBsAg-negative, anti-HBc-positive donors (Dickson et al. 1997; Muñoz 2002). It seems that recipients positive for both anti-HBs and anti-HBc are at a relatively low risk for de novo HBV infection. It should also be noted that HBV reactivation from pre-transplanted, “hidden” HBV of the OBI recipients is also possible especially after the use of intensive immunosuppressive therapy for the liver transplantation.

Finally, unlike the situation in liver transplantation, the risk of HBV transmission from OBI donors to recipients of kidneys, hearts and bone marrows seems to be lower (Wachs et al. 1995; De Feo et al. 2005). It is now a routine procedure in many centres that prophylactic antiviral therapy would be given to patients who are going to receive organs from OBI donors. Some centres are also extending this prophylactic measure to recipients receiving anti-HBc-positive donors with undetectable HBV DNA.

5.2 Role on Chronic Liver Disease and Complications

For chronic liver diseases, the initial activities can be assessed by the liver biochemistry and the liver histology, the more reliable parameter for assessment of liver damage. According to various studies (Yuen et al. 2010; Fung et al. 2010; Wong et al. 2014), OBI subjects usually have normal liver biochemistry as well as very minimal histological inflammation and fibrosis. However, it has been shown in early 1980s that HBV DNA were detectable in HBsAg-negative patients with chronic liver disease implying a possible pathologic role of OBI (Brechot et al. 1981, 1985). It is therefore not surprising to observe that around 5–50% of cirrhotic patients without identifiable cause have documented presence of OBI (Pollicino et al. 2004; Hou et al. 2001; Bréchot et al. 2001; Alavian et al. 2012).

The pathologic roles of OBI are firstly described in HBsAg-negative alcoholic patients with HCC (Bréchot et al. 1982). This suggests a possible synergism between these two factors leading to the development of HCC. According to a recent study, 56% of HBsAg-negative alcoholic patients with HCC had the presence of HBV DNA in the liver (Wong et al. 2011). There are even more studies examining the role of OBI in patients with chronic hepatitis C virus (HCV) infection. They have shown the pathological component of OBI in causing cirrhosis and HCC in patients co-infected with HCV (Bréchot et al. 1985; Cacciola et al. 1999; Fukuda et al. 1999; Tanaka et al. 2004; Squadrito et al. 2006; Shetty et al. 2008; Adachi et al. 2008; Shi et al. 2012). For instance, HCV patients with OBI had more advanced fibrosis compared to HCV patients without OBI (Matsuoka et al. 2008). Liver cirrhosis was found in 33% of HCV patients with OBI compared to 19% of HCV mono-infected HCV patients (Cacciola et al. 1999). In addition, HBV DNA and anti-HBc positivity were found to be two independent risk factors for the development of HCC in HCV patients (Adachi et al. 2008; Matsuoka et al. 2008; Ikeda et al. 2007). All these findings suggest that HCV patients with OBI carry a more progressive course for the development of cirrhosis and HCC. A recent meta-analysis study showed that HCV patients with OBI had a 2.8-folds higher risk of development of HCC compared to mono-infected HCV patients (Shi et al. 2012). However, there are studies which did not show an additively or synergistically increased risk for the development of liver complication by OBI (Sagnelli et al. 2008; Kao et al. 2002; Tsubouchi et al. 2013). As these studies are all cross-sectional and also of small scale, the best way to confirm whether OBI has additive or synergistic effects on development of liver complications in HCV patients is to have large-scale longitudinal studies.

Apart from the possible role of OBI in patients with alcoholic liver disease and patients with HCV infection, OBI may also be the primary aetiological cause for liver disease patients with apparently no identifiable cause, i.e. cryptogenic cirrhosis or HCC. It has been shown by very early studies that HBV DNA was detected in HBsAg-negative patients with HCC (Bréchot et al. 1981; Shafritz et al. 1981). The proportion of patients with cryptogenic HCC having OBI is in a range between 48% and 73% according to two studies (Wong et al. 2011; Yotsuyanagi et al. 2000). In fact, OBI was shown to be an independent risk factor for liver carcinogenesis with the hazard ratio of 8.3 (Ikeda et al. 2007). According to a longitudinal study performed in Japan, there was an association between OBI and HCC. In a meta-analysis study including 16 retrospective and prospective studies, OBI was confirmed to be a risk factor for the development of HCC with the hazard ratio of three when the analyses were limited to the five prospective studies (Shi et al. 2012). Another meta-analysis revealed a ninefold increased risk of development of HCC in OBI subjects (Covolo et al. 2013).

Concerning the best methodology for detection, different studies use different experimental protocols targeting for different genomic regions of HBV. In general, the detection rate is higher in non-tumorous parts compared to tumorous parts of the liver. In addition, sequencing the X gene compared to other genomic regions also yields a higher positive rate (Wong et al. 2011).

Similar to the oncologic mechanisms of overt CHB disease, mechanisms of OBI causing HCC include the following:

  1. 1.

    Persistently low-grade necroinflammation leading to cirrhosis of the liver (it may be possible that there are unrecognized high-grade necroinflammation during the overt stage of CHB disease before entering into the OBI state) (Yuki et al. 2003; Bläckberg and Kidd-Ljunggren 2000).

  2. 2.

    Integration of viral sequence from OBI into the human genome triggering the oncogenic cascade (Pollicino et al. 2004; Bréchot et al. 1981, 2000; Shafritz et al. 1981; Tamori et al. 1999; Paterlini et al. 1995). Integration of the HBV gene into the human genomes is also commonly found in OBI patients (Brechot et al. 1981). It has also been shown that the process of HBV DNA integration can happen even at the time of acute HBV infection (Murakami et al. 2004; Kimbi et al. 2005).

  3. 3.

    OBI may have HBV transcriptional activities producing oncogenic proteins, namely, protein X and truncated pre-S-S protein (Cougot et al. 2005).

5.3 HBV Reactivation

HBV reactivation can be associated with significant liver-related morbidity and mortality. Under normal circumstance, spontaneous HBV reactivation from OBI is rare. In most of the OBI subjects, the serum HBV DNA level remains undetectable, and if it is detectable at baseline, it usually remains very low throughout subsequent follow-up. However, there are circumstances where HBV may reactivate from this low viraemic state. It is well known that HBsAg-positive CHB patients undergoing chemotherapeutic treatment and immunosuppressive therapy, in particular steroid containing regimes, have a considerable risk of HBV reactivation (Yuen 2016; Cheung et al. 2016). As such, it is already firmly established that all HBsAg-positive CHB patients before receiving immunosuppressive therapy are recommended to receive antiviral prophylaxis to prevent HBV reactivation. However, this recommendation has not been universally implemented in OBI subjects yet.

As a result of the increasing use of biologics and monoclonal antibodies with potent immunosuppressive effects for autoimmune diseases and haematological malignancies, HBV reactivation from OBI subjects have emerged to be a more serious problem nowadays. It was found that there was no HBV reactivation from HBsAg-negative, anti-HBc-negative subjects with lymphoma receiving chemotherapy (Yeo et al. 2009). However, approximately 24% of anti-HBc-positive subjects receiving cyclophosphamide, doxorubicin, vincristine and prednisolone (CHOP) plus anti-CD 20 (rituximab) had HBV reactivation resulting in one death. On the other hand, there was no HBV reactivation in anti-HBc-positive subjects if they received rituximab-free treatment regime. It strongly suggests that rituximab significantly increases the risk of HBV reactivation in OBI subjects. This issue was studied in detail by a prospective study recently (Seto et al. 2014). Of the 63 HBsAg-negative, anti-HBc-positive subjects (all with undetectable serum HBV DNA) receiving rituximab, 42% of subjects develop HBV reactivation (Seto et al. 2014). It was also found that the rate of HBV reactivation was lower in anti-HBs-positive subjects compared with anti-HBs-negative subjects (34% vs. 68%, respectively, p = 0.012). Prompt treatment using entecavir at the time of HBV reactivation was able to control the HBV DNA level to undetectable level in all patients, and therefore no hepatitis flares were observed.

Concerning the biologics, the rate of HBV reactivation from OBI subjects was lower than that of using rituximab. According to a large-scale study with HBsAg-negative and anti-HBc-positive, the use of anti-tumour necrosis factor was associated with the HBV reactivation rate of 1.7% (Lee et al. 2013). The rate of HBV reactivation is however, considerably higher (15–35%) in HBsAg-positive subjects receiving anti-tumour necrosis factor (Perez-Alvarez et al. 2011).

Patients with haematological malignancy are usually at a higher risk of HBV reactivation because of the high degree of bone marrow suppression by intense immunosuppressive therapy. Among those, patients receiving haematopoietic stem cell transplantation (HSCT) are having the highest rate of HBV reactivation. The rate of HBV reactivation in HSCT patients who were HBsAg-negative and anti-HBc-positive is in a range of 3 and 43% as reported by different studies (Vigano et al. 2011; Hammond et al. 2009; Nakamoto et al. 2014). A recent prospective study with 62 HSCT patients showed that the HBV reactivation was as high as 40% with 2 years of follow-up (Seto et al. 2017). In addition, patients with chronic graft-versus-host disease (requiring more intense and long-term immunosuppressive therapy) and patients older than 50 years old would have a higher rate of HBV reactivation compared to subjects without these factors.

The underlying mechanisms involved in the HBV reactivation in OBI subjects are believed to be the same as for CHB subjects. These include direct increase in HBV replication related to the agents of the therapy. It has been well known that steroid can increase HBV replication as it can act through the glucocorticoid-responsive element in the HBV genome. As the immune control may be further jeopardized by the B-cell depleting agent, e.g. rituximab, HBV may reactivate whenever under a less intensive immune control environment. Finally, at the later stage upon cessation of the immunosuppressive therapy, the immune reconstitution may exert an even more overwhelming immune-mediated attack to the hepatocytes where the HBV load is high.

Because of the potential mortality from HBV reactivation, prevention using prophylactic antiviral agents is generally recommended. It is now a universally accepted protocol to prescribe antiviral agents for HBsAg-positive patients undergoing chemotherapy or immunotherapy. For OBI subjects, this practice is also widely implemented for those with detectable HBV DNA. A controversy, however, exists in HBsAg-negative, anti-HBc-positive patients with undetectable HBV DNA. Some clinicians advocate the prophylactic approach, whereas others adopt a close surveillance strategy. The latter approach recommends HBV DNA monitoring at a 4–12 weekly interval. According to a study of lymphoma patients receiving rituximab, prompt administration of entecavir at the time of detectable HBV DNA checked at 4 weekly intervals is able to render HBV DNA undetectable without hepatitis flare in all patients (Seto et al. 2014). This strategy is also recommended by treatment guidelines (European Association for the Study of the Liver 2012). At present, there are no studies to determine the best monitoring intervals for HBV DNA. Practical logistics, manpower and costs are the major factors for the consideration. Another issue is whether the “interval” of monitoring would be the same for all kinds of chemotherapy and malignancy. It is because the risk of HBV reactivation would largely be determined by the type of chemotherapy and to a lesser extent the specific malignancy.

In summary, OBI is now a well-recognized disease entity with increasing attention in various aspects of liver diseases. The prevalence is yet to be studied in greater detail. The diagnosis of OBI is greatly dependent on the clinicians’ awareness and the availability of standard and accurate HBV DNA testing. The presence of anti-HBc would trigger off the investigative steps. Subjects with OBI are potentially infectious with a low transmissibility rate in which the level of viraemia and the immune status of the recipients are the two major determinants. OBI can lead to end-stage liver disease and HCC, and this potential aetiology should be actively sought for in patients with cryptogenic cirrhosis and/or HCC. HBV reactivation from OBI is now being increasingly recognized in the present era of using potent B-cell depleting therapy for various malignant and immunological conditions. While effective antiviral therapy reducing the morbidity and mortality from the HBV reactivation is already in place, the best strategy of prevention of HBV reactivation in these subjects remains to be defined.