High Risk of Hepatitis B Reactivation among Patients with Acute Myeloid Leukemia

Background Hepatitis B virus (HBV) infections are common and associated with significant morbidity and mortality in cancer patients. However, the incidence and risk factors of HBV reactivation in patients with acute myeloid leukemia (AML) are rarely investigated. Methods AML patients followed-up at the National Taiwan University Hospital between 2006 and 2012 were analyzed. The clinical characteristics and laboratory data were retrospectively reviewed. Results Four hundred and ninety patients comprising 265 men and 225 women were studied. The median age was 52 years (range, 18 - 94). Chronic HBV carriage was documented at the time of leukemia diagnosis in 57 (11.6%) patients. Forty-six (80.7%) of the 57 HBV carriers received prophylaxis with anti-HBV agents. Sixteen HBV carriers (28.1%) developed hepatitis B reactivation during or after chemotherapy, including 7 patients who had discontinued antiviral therapy. The incidence of hepatitis B reactivation among AML patients with HBV carriage was 9.5 per 100 person-years. Prophylaxis with anti-HBV agents significantly decreased the risk of hepatitis B reactivation among HBV carriers (13% vs. 61%, p<0.001). Four (2.8%) of 142 patients with initial positive anti-HBsAb and anti-HBcAb experienced hepatitis B reactivation and lost their protective anti-HBsAb. Multivariate analysis revealed that diabetes mellitus (p=0.008, odds ratio (OR) = 2.841, 95% confident interval (CI): 0.985-8.193) and carriage of HBsAg (p<0.001, OR=36.878, 95% CI: 11.770-115.547) were independent risk factors for hepatitis B reactivation in AML patients. Conclusions Hepatitis B reactivation is not uncommon in the HBsAg positive AML patients. Prophylaxis with anti-HBV agent significantly decreased the risk of hepatitis B reactivation.


Results
Four hundred and ninety patients comprising 265 men and 225 women were studied. The median age was 52 years (range, 18 -94). Chronic HBV carriage was documented at the time of leukemia diagnosis in 57 (11.6%) patients. Forty-six (80.7%) of the 57 HBV carriers received prophylaxis with anti-HBV agents. Sixteen HBV carriers (28.1%) developed hepatitis B reactivation during or after chemotherapy, including 7 patients who had discontinued antiviral therapy. The incidence of hepatitis B reactivation among AML patients with HBV carriage was 9.5 per 100 person-years. Prophylaxis with anti-HBV agents significantly decreased the risk of hepatitis B reactivation among HBV carriers (13% vs. 61%, p<0.001). Four (2.8%) of 142 patients with initial positive anti-HBsAb and anti-HBcAb experienced hepatitis B reactivation and lost their protective anti-HBsAb. Multivariate analysis revealed that diabetes mellitus (p=0.008, odds ratio (OR) = 2.841, 95% confident interval (CI): 0.985- 8.193) and carriage of HBsAg (p<0.001, OR=36.878, 95% CI: 11.770-115.547) were independent risk factors for hepatitis B reactivation in AML patients.

Introduction
Acute myeloid leukemia (AML) is a heterogeneous disorder with regards to the morphology and chromosome aberrations detected in the leukemic cells [1]. In spite of this, most AML patients receive standard induction chemotherapy containing anthracycline and cytarabine, and consolidation with high dose cytarabine [1]. The chemotherapy agents used for AML are similar to those used for lymphoma [2][3][4]. High dose cytarabine infusion can cause skin rash and fever [5], and thus, steroid is usually added as prophylaxis. Abnormal liver function tests are frequently found among AML patients receiving chemotherapy, possibly related to the chemotherapy itself, the concomitant use of antifungal agents and total parenteral nutrition, or to sepsis related organ dysfunction, leukemic infiltration, or graft versus host disease. Reactivation of viral hepatitis, including hepatitis B, is also a potential cause for abnormal liver function tests in leukemic patients receiving chemotherapy [6,7].
Hepatitis B virus (HBV) affects 350 to 400 million persons worldwide and constitutes a major global health burden [8]. HBV is a DNA virus transmitted parenterally, sexually, and perinatally [8]. HBV infection can cause acute and chronic liver disease including cirrhosis and hepatocellular carcinoma. Following immunosuppression, HBV replication along with signs of hepatocellular injury in a silent hepatitis B surface antigen (HBsAg) carrier may occur [9]. The clinical presentation of hepatitis B reactivation is variable, ranging from asymptomatic to fulminant hepatitis, liver failure, and death. Previous reports of HBV reactivation have been observed in patients with lymphoma [10,11], treated with corticosteroids [12,13] and rituximab [14,15] as well as in patients undergoing stem cell and bone marrow transplantation [16,17]. However, the epidemiology and clinical manifestations of hepatitis B reactivation among AML patients are rarely described [10]. Taiwan has long been an area endemic for HBV, and, previously, the seroprevalences of anti-HBc approached 80-90% and of HBsAg 15-20% prior to the nationwide hepatitis B vaccination program [18]. In this study, we retrospectively reviewed the epidemiology, clinical and laboratory data of hepatitis B in AML patients at a medical center to understand the epidemiology and clinical outcomes of HBV reactivation among AML patients. hepatitis B serology, hepatitis B virus DNA and outcome of all adult AML patients during the period January 2006 to December 2012 at NTUH were analyzed retrospectively by chart review. Liver function tests, hepatitis B serology titers, and HBV DNA levels were performed as clinically indicated.

Definitions
The serum biochemistry (alanine transaminase, ALT) was measured by the Beckman Coulter AU5800 platform (Beckman Coulter Inc, Brea, CA, USA); HBsAg, anti-HBsAb, Anti-HBc, HBeAg, and anti-HCV were performed with Abbott ARCHITECT i2000SR (Abbott Laboratories, Abbott Park, North Chicago, IL, USA); HBV DNA was analyzed with COBAS AmpliPrep /COBAS TaqMan HBV Test, v2.0 (Roche, Basel, Switzerland) according to the manufacturer's instruction. The serological results were defined by the following cut-off values: positive hepatitis B surface antigen (HBsAg) if≧ 0.05 IU/mL; positive anti-hepatitis B surface antibody (anti-HBsAb) with a threshold of 10 mIU / mL [19]; positive anti-hepatitis B core antibody (anti-HBcAb) if ≧ 1.0. Chronic hepatitis B carrier status was defined by the detection of positive hepatitis surface antigen (HBsAg) for more than 6 months. Hepatitis B reactivation was defined as a greater than 10-fold increase, compared with previous nadir levels, of HBV DNA or by the reappearance of hepatitis B "e" antigen in the serum for patients whose baseline HBeAg was negative. HBV-related hepatitis was defined as a greater than 3-fold increase of ALT level (the upper normal limit is 41 U/L at NTUH) accompanying or following HBV reactivation. In July 1984, the Taiwan government launched a nationwide universal HBV vaccination program [20]. The following catch-up program of the nationwide HBV vaccination had a coverage rate of 86.9 to 98.0% [21].
The diagnosis of cirrhosis was based on clinical symptoms / signs and imaging study (ultrasound, computed tomography). The diagnostic criteria of hepatocellular carcinoma(HCC) in this study was based on histologic and/or clinical findings and on the presence of all of the following criteria: chronic viral hepatitis infection and liver cirrhosis, hepatic tumor with imaging (ultrasound, computed tomography) characteristics compatible with a diagnosis of HCC and without evidence of gastrointestinal or other primary tumor, and a persistent elevation of the serum level of alpha-fetoprotein (AFP) to 400 ng / mL or higher.

Chemotherapy of acute myeloid leukemia
Standard induction chemotherapy consisting of cytarabine and anthracycline were used for patients with AML at NTUH. Standard dose chemotherapy was defined as cytarabine 100mg / m 2 body surface area administered as a continuous infusion for seven days. Low dose chemotherapy was defined as cytarabine 10 to 20 mg / m 2 body surface area for 7 to 14 days. Consolidation chemotherapy consisted of high-dose cytarabine-based regimens. Steroid was used routinely for prevention of fever and skin rash in AML patients receiving high dose cytarabine chemotherapy. Methylprednisolone 40 mg twice daily by intravenous injection was administered before cytarabine for 6 to 8 doses. Patients with acute promyelocytic leukemia (APL) were treated with all trans retinoic acid (ATRA) in combination with anthracycline-based chemotherapy. Hypomethylating agents such as azacitidine were given at a dose of 75 mg / m 2 body surface area for 7 days every 4 weeks.

Statistical analysis
Survival was estimated by the Kaplan-Meier analysis and compared using the log-rank test. Categorical variables were compared using the Chi-square test. Univariate and multivariate analyses of factors associated with time to HBV reactivation were performed using the Cox proportional hazards model. Factors with p-value of 0.10 in the univariate analysis were considered in the multivariate model. All statistical analyses were performed using the statistical package SPSS for Windows v.18 (SPSS Inc., Chicago, IL). A p-value of 0.05 was considered significant and all statistical tests were two-tailed.

Epidemiology
There were 593 AML patients diagnosed and regularly followed at NTUH during the study period. Four hundred and ninety patients with at least one hepatitis B serology test were enrolled in this study, of which 265 were men and 225 were women. The median age was 52 years and ranged from 18 to 94 years. The median observation period was 596 days (ranged from one to 3065 days). The cytogenetic of this population compromised 234 patients with normal karyotype, 27 patients with t(8;21), 35 patients with t(15;17), 21 patients with inv(16), 13 patients with chromosomal 11q23 changes, 15 patients with del(5q)/monosomy5 or del(7q)/monoso-my7, 57 patients with complex (i.e. more than three) chromosomal changes, and 80 patients with other simple chromosomal changes. There were 6 patients whose cytogenetic study revealed no mitosis and 2 patients without cytogenetic analysis. There were 23 patients with preceding history of myelodysplastic syndrome (MDS), 3 patients with therapy-related leukemia. There were 35 patients diagnosed as acute promyelocytic leukemia. The factors of cytogenetic change, leukemic subtype, preceding myelodysplastic syndrome and therapy-related were not significantly correlated with HBV reactivation and carrier status.
The clinical characteristics of these 490 AML patients are shown in Table 1. At the time of leukemia diagnosis, there were 57 patients with positive HBsAg giving an estimated prevalence of 11.6% and 433 patients with negative HBsAg (Fig 1). The anti-HBsAb serology was positive in 324 of 433 patients, negative in 50 patients, and not tested in 59 patients. Fifteen (3%) patients had chronic hepatitis C at diagnosis of leukemia and two patients had both hepatitis B and C. The incidence of hepatitis B reactivation and HBV-related hepatitis were 9.5 and 8.3 per 100 person-years among chronic HBV carriers of AML patients. Three of 32 HBV vaccinees who were born after July 1984 were HBV carriers.

Prophylaxis of Hepatitis B reactivation in the AML patients
Currently, there is no consensus for antiviral prophylaxis for AML patients with HBV carrier status in Taiwan. In clinical practice, antiviral prophylaxis generally starts at diagnosis of leukemia, and continues for at least one year after chemotherapy according to the clinical condition. Accordingly, forty-six of 57 positive HBsAg patients received HBV prophylaxis with antiviral agents after diagnosis of leukemia; 28 patients received lamivudine, 13 patients received entecavir, two patients received adefovir, two patients received telbivudine, and one patient received tenofovir. Ten patients were not prescribed any prophylaxis after diagnosis of leukemia. Overall, sixteen of 57 positive HBsAg patients suffered from recrudescence of hepatitis B virus during or after chemotherapy, including four patients who were antiviral-naive, seven patients who had discontinued initial HBV prophylaxis after a median of 118 days (range, 30 to 447 days), and five patients who were taking antiviral prophylaxis but still had hepatitis B reactivation. The latter five were under antiviral (lamivudine n = 2, adefovir n = 1, telbivudine n = 1, entecavir n = 1,) prophylaxis at the time of HBV reactivation, documented by a 10-fold or more elevation of hepatitis B viral load and accompanied by a median rise of ALT to 116 U/L, range 84-378 U/L). These five patients had good antiviral compliance. The durations from antiviral therapy to hepatitis B reactivation were 4, 12, 7, 14, and 3 months, respectively. Case 2 and case 4 had received hematopoietic stem cell transplantation (HSCT) and hepatitis B reactivation developed after 2 and 7 months after HSCT. The two patients under lamivudine prophylaxis with HBV reactivation were subsequently shown to harbor HBV virus with the YMDD mutation. Nevertheless, AML patients with positive HBsAg under antiviral prophylaxis had significantly decreased risks of hepatitis B reactivation than AML patients with positive HBsAg without prophylaxis (5 of 39 (13%), vs. 11 of 18 (61%), p<0.001).

Risk Factors For hepatitis B reactivation
We compared the 20 patients who had hepatitis B reactivation and the other 419 patients who did not have hepatitis B reactivation to evaluate the risk factors for hepatitis B reactivation among patients previously exposed to HBV (Table 2). Fifty-one patients with negative anti-HBcAb were excluded. By univariate analysis, diabetes mellitus (p = 0.006), hematopoietic stem cell transplant (p = 0.014), liver cirrhosis (p = 0.002), positive HBsAg (p<0.001) and age less than 65 years (p = 0.020) appeared to be associated with HBV reactivation. By multivariate analysis, diabetes mellitus (p = 0.008, odds ratio (OR) = 2.841, 95% confident interval (CI):  Table 3. Univariate analysis showed diabetes mellitus was associated with higher risk of HBV reactivation (p = 0.005) and antiviral therapy with less risk of hepatitis B reactivation (p<0.001). Multivariate analysis showed that HBsAg positive patients who had received antiviral therapy had significantly lower risk of hepatitis B reactivation (p = 0.033, odds ratio (OR) = 0.094, 95% confident interval (CI): 0.025-0.355).
Four of 142 negative HBsAg AML patients had hepatitis B reactivation. Their unusual clinical courses are shown in Table 4. All four patients had seroconverted with positive anti-HBsAb

Discussion
There is still controversy regarding the role of screening for hepatitis B reactivation in cancer patients. The American Society of Clinical Oncology Provisional Clinical Opinion reported that the evidence is insufficient to determine the net benefits and harms of routine screening for chronic HBV infection in individuals with cancer who are about to receive cytotoxic or immunosuppressive therapy or who are already receiving such therapy [22]. To the best of our knowledge, this is the first large cohort study to investigate hepatitis B reactivation in AML patients. In this study that was conducted in a country of HBV hyperendemicity, we showed that AML patients had a similar risk for HBsAg sero-positivity (11.6%), because most exposure to HBV that results in either development of anti-HBs antibody or chronic carriers of HBV in Taiwan before the implementation of the nationwide HBV vaccination program, occurred during the perinatal period or in early childhood. Our results showed that AML patients, especially those with diabetes and with HBsAg carriage, are also at high risk for hepatitis B reactivation. Prophylaxis with antiviral therapy significantly decreased the risk of hepatitis B reactivation. However, development of lamivudine resistance during antiviral therapy should be suspected if one observes a rebound of HBV plasma viral loads. The incidence of hepatitis B reactivation and HBV-related hepatitis in our study were 9.5 and 8.3 per 100 person-years in AML patients who are also chronic hepatitis B carriers, which is similar to the incidence of hepatitis B reactivation in lymphoma patients (10.4 per 100 person-years) [11,23]. While Nakamura et al. [24] reported that 10 of 13 (76.9%) AML patients had severe hepatitis due to HBV in 85 patients with hematological malignancy; Yeo et al. [10]   found that patients with either leukemia or myeloma were at comparatively less risk of hepatitis B reactivation than patients with lymphoma. Since fulminant hepatitis B is a catastrophic event for HBsAg positive AML patients [6,25,26], periodic assessment of liver function and follow up of HBV serological status is important during chemotherapy. Further prospective cohort studies of patients with AML especially in the endemic hepatitis B area should be undertaken to capture the true incidence of HBV flare-ups.
In this study, AML patients with positive HBsAg under antiviral agents were significantly protected from hepatitis B reactivation. We therefore recommend that all HBsAg positive AML patients should be initiated on antiviral prophylaxis prior to chemotherapy. Although, there is no current consensus for antiviral prophylaxis for AML patients with HBV carrier status in Taiwan, the National Comprehensive Cancer Network (NCCN) guidelines indeed have some recommendation with regards to patients with high risk for HBV [27]. Moreover, the major HBV treatment guidelines proposals originating from America, Europe, and the Asia-Pacific region, recommended that all cancer patients should check HBV markers, including HBsAg and anti-HBc, prior to initiation of chemotherapy, and to use routine prophylactic antiviral therapy for individuals who are positive for HBsAg before the start of cancer chemotherapy [28][29][30]. The optimal timing, duration and choice of anti-HBV agent for prophylaxis remain unknown [30][31][32]. In our study cohort, despite antiviral prophylaxis, seven AML patients experienced hepatitis B reactivation after discontinuation of prophylaxis. Hence, close monitoring of the liver functional assays as well as hepatitis B plasma viral loads may be indicated in the first year after discontinuation of antiviral prophylaxis in patients with ongoing immune suppression. Of the patients with reactivation under ongoing anti-HBV prophylaxis, two of the four patients underwent genotypic resistance testing. Both patients had proven YMDD (rtM204I/V) mutation under lamivudine prophylaxis. Hence, the risk for emerging resistance should be considered in the event of virologic breakthrough during lamivudine therapy. Resistance to other anti-HBV viral agent in the first year is less frequently encountered. In this study, four patients with positive anti-HBsAb lost seroprotection with reappearance of HBsAg. Of cancer patients receiving chemotherapy or transplantation, anti-HBsAb levels might decline below the threshold for protection against new HBV infection or reactivation [33][34][35][36]. Although HBV vaccination is highly immunogenic and efficacious in inducing protective antibody against HBV [37], a gradual decline in titers decades after vaccination have been reported [38]. Moreover, serologic evidence of recovery from hepatitis B infection does not preclude its reactivation after immunosuppression. Thus, screening for serologic evidence of hepatitis B and booster HBV vaccination for those without protective anti-HBsAb should be considered in individuals in whom immunosuppressive therapy is planned.
The finding that diabetes mellitus is associated with HBV reactivation is biologically plausible since the hyperglycemic environment increases the virulence of viral pathogens. The underlying mechanisms may involve reduced production of interleukins in response to infection, reduced chemotaxis and phagocytic activity, and reduced humoral immunity [39]. Wang et al. reported diabetes was associated with serum alanine aminotransferase activity elevation in patients with hepatitis B infection [40]. Diabetes mellitus has also been documented as an independent risk factor for HBV related hepatocellular carcinoma [41] and occult HBV infection [42]. In general, diabetic patients are considered to be relatively immunocompromised. However the exact role of hyperglycemia in hepatitis B reactivation should be further clarified.
This study suffered from several limitations. Since the study is a retrospective review, the clinical data was derived from medical records in accordance to clinical practice. Therefore, data were limited or unavailable for several variables. Hence, subclinical hepatitis or asymptomatic reactivations, newly acquired HBV infection or seroconversion of HBsAg may be potentially missed. In addition, patients with early mortality during AML treatment may have contributed to underestimated risks of HBV reactivation.
In conclusion, hepatitis B virus on reactivation is not uncommon in HBsAg positive AML patients receiving or undergoing chemotherapy. Prophylaxis or early preemption with an anti-HBV agent significantly reduced the risk of hepatitis B reactivation and HBV related hepatitis. Diabetes mellitus and chronic hepatitis B carrier status were independent risk factors of hepatitis B reactivation in AML patients.