HBVr can be associated with a significant risk of morbidity and mortality or it can be very mild and clinically inapparent. Practical experience has shown that many mild cases are not detected because changes in serum aminotransferase levels are attributable to the chemotherapy and the patient is not known to have HBV infection. Many observational studies include a number of cases in which no alteration of liver chemistries are appreciated. The overall frequency with which biochemical abnormalities are seen is often in excess of 50 % in observational studies that report data on all patients who meet the virologic criterion [11].

Mortality from HBV reactivation is reported in most clinical series. It is difficult to determine the absolute risk of death from HBVr. Some studies have reported a high mortality risk (greater than 50 %) [17, 18]. However, in most studies liver-related mortality from HBVr has been considerably lower varying from 0 to 10 % [11]. Mortality has generally been observed to be a less frequent outcome in patients who reactivate during treatment of nonmalignant conditions with tumor necrosis factor (TNF) alpha inhibitors [11, 19]. The relatively low mortality rate is unlikely to be due to treatment with antiviral therapy because studies that predated antiviral therapy have also revealed relatively low mortality. Even though fulminant liver failure is uncommon when HBVr is precipitated by cancer chemotherapy, it is important to emphasize that these patients lack the usual option of liver transplantation due to the underlying malignancy [20].

Observational studies tend to give less emphasis and inconsistent reporting on outcomes other than severity of hepatitis and case fatalities. One of the important clinical concerns, however, is the relatively high frequency with which cancer chemotherapy needs to be interrupted, either temporarily halted or prematurely discontinued. This has been shown to occur in as many as 20–40 % of patients with breast cancer [21]. There has also been little concerted effort to report late complications of chemotherapy interruption such as cancer-related fatalities due to the progression of underlying malignancy. Another underreported outcome is disease morbidity. An international survey of the American Association for the Study of Liver Disease (AASLD) membership described 188 cases of HBVr induced by cancer chemotherapy in which nearly half of the patients had chemotherapy interruption and more than 50 % required hospitalization with a large number requiring intensive care management. In addition, liver related mortality was reported in 23 % of patients (Fig. 18.1) [12].

HBVr occurs worldwide and there is reason to suspect that the incidence may be increasing due to the convergence of a number of factors: (1) the rapid development of immunologically potent biologic agents that are used across medical specialties; (2) relatively poor provider awareness of the benefits of HBV screening and antiviral prophylaxis of at risk patients; and (3) ambiguous and weak specialty practice guidelines [11, 13, 22]. In a recent international survey distributed to liver disease specialists, 40 % of whom practiced outside of the United States, a disproportionate number of cases (131 of 188, or 70 %) were patients born in areas of intermediate to high HBV endemicity (Fig. 18.2) [12]. A major concern in the findings from this survey was that only 40 % of the 188 HBVr cases were screened for HBsAg and anti-HBc before initiation of chemotherapy and an additional 13 % had HBsAg screening alone. Further disappointing, only 10 % received prophylactic antiviral therapy.

The widespread use of immunosuppressive therapy for nonmalignant conditions is a major contributing factor to a rising incidence of HBVr from ISDT. There is broad availability of TNF alpha inhibitors and broad licensing of potent biologic agents with inhibitory effects on T cell signaling, cell adhesion molecules, tyrosine kinase activity, and cytokines including interleukin (IL)-12 and IL-23. Within this landscape of therapeutic options have been increasing reports of HBVr induced by anti-CD20 B cell depleting agents used in chemotherapeutic regimens for lymphoma as well as in the management of rheumatologic diseases [11, 14, 23]. The routine use of some of these agents in the treatment of various chronic inflammatory disorders has greatly widened the spectrum of patients potentially at risk of HBVr beyond those receiving cancer chemotherapy or traditional immunosuppressive therapy such as glucocorticoids, methotrexate, or azathioprine. A diverse range of medical specialties such as gastroenterology, dermatology, and rheumatology now utilize these biologic agents on a regular basis.

It has been difficult to directly study the immunopathogenetic events linked to HBVr due to limited biologic samples, which by necessity are harvested after the disorder is clinically apparent [24]. A sequence of events has been depicted for HBV infected persons during and after cancer chemotherapy administration, which is supported by a limited number of in vitro and numerous observational studies (Fig. 18.3). Longitudinal cohort studies evaluating the kinetics of viral replication in patients with resolved HBV infection undergoing cytotoxic chemotherapy have provided insight into the timeline of evolving HBVr in individuals at risk. During the course of immunosuppressive drug therapy, an initial phase of enhanced viral replication appears to take place, in which up to a 100-fold or greater increase in serum HBV DNA levels can be observed as early as 3–6 months prior to the onset of overt clinical HBVr [25]. During this period of increased viral replication, serum aminotransferases remain normal (Fig. 18.3b) A second phase ushered in by immunologic restitution often follows after discontinuation of the immunosuppressive agent. During this phase, onset of clinical hepatitis may occur with biochemical evidence of hepatocellular injury and risk of hepatic failure (Fig. 18.3c). Preliminary evidence using biologic samples from patients with HBVr due to cancer chemotherapy suggests that an acute resurgence of HBV-specific cell mediated immunity plays a primary role in liver cell injury during the second phase of reactivation. In these patients increased HBV-specific CD8+ T cells have been detected along with diminished regulatory T cells; similar findings were observed in patients with active chronic hepatitis B but not in inactive HBsAg carriers [26]. A similar pattern may occur when HBVr occurs in association with other immunosuppressive agents; however, immunologic changes occurring in other forms of immunosuppression have yet to be described.

Reconstitution of the host immunity appears to be a critical factor leading to liver injury after discontinuation of immunosuppressive agents, in the setting of cancer chemotherapy, or abrupt glucocorticoid withdrawal [4]. As the immunologic events leading up to HBVr are thought to occur several weeks before the development of overt clinical signs or elevated aminotransferases, a strategy of HBV DNA monitoring and deferred antiviral therapy is unlikely to avoid liver injury or potential liver failure in many patients [27–29]. This is well illustrated in the case depicted in Fig. 18.4.

As mentioned above, HBVr induced by ISDT occurs in patients with resolved as well as active infection and reactivation is precipitated by disruption of the balance between the host’s immune control over viral replication and the inherent fitness of HBV to replicate. However, there is a key difference in these two clinical states in that very stringent immunosuppression is usually needed when patients with resolved infection undergo HBVr whereas this is not the case with HBsAg-positive individuals. This discrepancy is better understood by an awareness of the biological gradient of covalently closed circular DNA (cccDNA) that exists in hepatic tissue during the various phases of infection. Covalently closed HBV DNA is a highly stable molecule that acts as the genomic template for viral transcription and a concentration gradient that encompasses three orders of magnitude has been detected in hepatic tissue [30, 31]. The highest concentrations of cccDNA have been demonstrated in patients with active chronic hepatitis B and high serum levels of HBV DNA, followed in descending order by the inactive HBsAg carrier state, and those with resolved infection [30, 31]. As patients with resolved infection have only minute amounts of cccDNA such individuals would be anticipated to undergo HBVr only under conditions in which highly aggressive ISDT is given. By contrast, inactive HBsAg carriers and those with chronic hepatitis B with high levels of serum HBV DNA have progressively higher amounts of cccDNA and are in a more ready state for uncontrolled viral replication with moderate and minimal immunosuppression, respectively. Much of this is inferential, but the hypothetical importance of different levels of cccDNA in determining the risk for HBV under varying conditions of immunosuppression is consistent with many of the clinical observations that have been made in this area.

HBV reactivation occurs most commonly during chemotherapy for leukemia or lymphoma . Rates of 50 % or greater have been routinely reported in patients not given antiviral prophylaxis [11]. However, HBVr has been observed in a wide array of solid organ malignancies including breast, colon, lung, stromal tumors of the gastrointestinal tract, head and neck cancer, retinoblastoma, sarcoma, and teratoma. Among solid organ malignancies, HBVr occurs most commonly with breast cancer where rates of 20–40 % have been reported. The frequent occurrence of HBVr has been attributed to the use of anthracycline (doxorubicin or epirubicin)-based regimens [32]. Anthracyclines are also used for ovarian, uterine, and lung cancer, and are frequently incorporated into the treatment of hepatocellular carcinoma with transarterial chemoembolization (TACE). Reactivated hepatitis B may occur at any time during chemotherapy and may even occur several months after discontinuation of treatment. While the timing of HBVr has not been linked to any particular set of features, this most likely is determined by a number of interactive variables including the intensity of immunosuppression and the baseline virologic and serologic status of the host.

The inordinate frequency of HBVr during chemotherapy for hematologic and solid organ malignancies tends to parallel the degree of potency of the immunosuppressive regimen and in particular the incorporation of drugs like rituximab and high-dose glucocorticoids [33, 34]. It is unclear whether it also reflects a greater degree of baseline immunodeficiency in these populations due to the lymphotrophic properties of HBV. The extensive immunologic conditioning in conventional bone marrow or hematopoietic stem cell transplant recipients has been associated with reactivation in as many as 50 % or more of HBsAg-positive persons and in 10–20 % of HBsAg negative, anti-HBc positive patients further affirming that an important determining relationship exists between HBVr and the degree of immune suppression provided by drug therapy [35, 36].

The risk of HBVr during cancer chemotherapy has been recognized for several decades but the past 10–15 years has witnessed increasing attention to the risk that exists when ISDT is used for nonmalignant disorders. Table 18.3 lists some of the nonmalignant disorders commonly treated with ISDT that have been associated with HBVr. The most common settings have been with the treatment of rheumatologic, dermatologic, and inflammatory bowel diseases. The range of agents used for these disorders includes antimetabolites, glucocorticoids, biologic agents, monoclonal antibodies, and calcineurin inhibitors. The most commonly used biologic agents are disease-modifying antirheumatic drugs (DMARDs ) , which as a broad category, includes azathioprine, methotrexate, cyclosporine, and a range of biologics such as TNF alpha inhibitors. Newer agents which block costimulation of lymphocytes, tyrosine kinase inhibitors, and integrin inhibitors have been developed for a variety of indications but have not had sufficient use to allow determination of the magnitude of risk for HBVr. B cell depleting agents such as rituximab or ofatumumab have potent and long lived effects on B cell function and have come under close scrutiny recently. The US Food and Drug Administration has recently mandated a box warning on these agents and strongly recommends HBV screening for HBsAg and anti-HBc and antiviral therapy when appropriate due to numerous reports of severe HBVr when used for rheumatic as well as malignant indications.

The determination of the magnitude of risk for HBVr with any biologic agent is generally difficult until there is broad clinical experience. However, there are several circumstances other than limited clinical experience that may confound estimate of risk. Patients who are started on newer biologic therapies for conditions such as ankylosing spondylosis, plaque psoriasis, and ulcerative colitis have often failed previous therapy with other disease modifying agents, suggesting that there may be a lengthier period of immunologic suppression prior to the initiation of the new agent. This may explain why in one large case series of TNF alpha inhibitor therapy, HBVr occurred significantly more commonly in patients given prior immunosuppressive therapy when compared to individuals who lacked prior exposure (96 % vs. 70 %, respectively, p < 0.03) [19]. It is also clear from observational studies that many patients continue with more than one immunosuppressive agent for control of refractory inflammatory disorders, thus making it difficult to assign direct causality to any one particular agent. In one review of the literature, two thirds of patients who developed HBVr while taking TNF alpha inhibitors were reported to be taking other immunosuppressive agents such as glucocorticoids, methotrexate, or calcineurin inhibitors. These considerations may explain why there are often wide estimates of the magnitude of risk with TNF alpha inhibitors, a biologic drug class in which HBVr has been reported to occur in 0–40 % of HBsAg positive patients [11, 19, 60, 61].