Liver failure from delayed hepatitis B reactivation in anti-HBc-positive patient following rituximab for B-cell lymphoma

Abstract

A 93-year-old man was admitted with 1 week of frank jaundice and abdominal pain. His medical history included diffuse large B-cell lymphoma treated with rituximab and cyclophosphamide, hydroxydaunomycin, oncovin and prednisolone (R-CHOP) chemotherapy 10 months prior. His investigations revealed marked hyperbilirubinemia with a total bilirubin of 355 μmol/L, along with a 17-fold elevation in alanine transaminase and impaired hepatic synthetic function. He tested hepatitis B surface antigen (HBsAg) and hepatitis B surface antibody (HBsAb) negative, hepatitis B core antibody (HBcAb) positive and had elevated hepatitis B virus DNA level at 13 691 IU/L. This was in the setting of radiological evidence of suspected cirrhosis. He was later found to have tested positive for HBcAb and negative for HBsAg and HBsAb prior to chemotherapy, but had not received antiviral prophylaxis. He was diagnosed with fulminant hepatitis secondary to delayed hepatitis B reactivation in the setting of rituximab. Hepatitis B reactivation and the role of screening and antiviral prophylaxis in isolated HBcAb-positive patients is reviewed.

Background

Hepatitis B virus reactivation (HBR) is broadly characterised by increasing hepatitis B virus (HBV) viral load in a patient with chronic HBV infection or the reappearance of HBV particles in those with resolved infection.¹

In the setting of cytotoxic or immunosuppressive therapy in the oncological population, HBR can be associated with significant morbidity and a mortality rate as high as 30%.² To date HBR has been documented following immunosuppressive treatment among patients with haematological and solid organ malignancies, rheumatoid arthritis, psoriatic arthritis, inflammatory bowel disease and renal transplant.³ Haematological malignancies and, in particular, lymphoproliferative disorders are associated with higher risk of HBR.^4–6 The degree of immunosuppression is an important determinant in HBR, with the addition of B-cell depleting therapies like rituximab to chemotherapy regimens greatly augmenting this risk.^7–10

HBR can occur both in patients with chronic HBV infection, as defined by the presence of hepatitis B surface antigen (HBsAg) positive status for at least 6 months and detectable HBV DNA, and those with likely resolved infection, as defined by HBsAg negative and hepatitis B core antibody (HBcAb) positive status.¹ There are a number of interpretations for a positive HBsAg status, including the hepatitis B e-antigen (HBeAg)-positive chronic infection, HBeAg-positive chronic hepatitis, HBeAg-negative chronic infection and HBeAg-negative chronic hepatitis, which are differentiated based on the HBsAg level, HBeAg status, HBV DNA load, alanine transaminase level and the presence of liver disease.¹¹ Similarly, isolated HBcAb positive status can also reflect a number of possibilities, from resolved prior infection to infection by HBV with HBsAg mutations.¹² The exact rate of HBR among isolated HBcAb-positive patients is unclear.⁵ Historically, decision-making regarding screening, monitoring and antiviral treatment has been controversial in this particular population.¹³ In the last decade, hepatitis B screening and prophylactic antiviral therapy for all HBsAg-positive patients and HBcAb-positive patients receiving high-risk immunosuppression, has become standard of care.^{11 14 15} Despite this, screening practices often differ from the recommendations and identification of an isolated HBcAb positive status does not always translate into antiviral prescribing.^16–18

We describe a case of fatal liver failure secondary to HBR in a patient with isolated untreated anti-HBc, post exposure to rituximab-containing chemotherapy for diffuse large B-cell lymphoma 10 months prior to presentation.

Case presentation

A 93-year-old man was admitted with 1 week of increasing jaundice, fatigue, nausea and anorexia. He described increasing dark urine and intermittent epigastric and right upper quadrant pain 3 days preceding admission. He did not have fever or confusion. This was on a background of stage I diffuse large B-cell lymphoma treated with 6 cycles of rituximab and cyclophosphamide, hydroxydaunomycin, oncovin and prednisolone (R-CHOP) to a complete metabolic response 10 months prior to admission. His other history included hypertension, osteoarthritis and multinodular goitre, medicated with amlodipine 10 mg daily and irbesartan–hydrochlorothiazide 150 mg/12.5 mg daily. He also took over-the-counter normotherapeutic doses of ibuprofen and paracetamol as required, did not consume alcohol and had migrated to Australia in 1952 from Southeastern Europe.

On physical examination, he was haemodynamically stable, afebrile and did not have evidence of encephalopathy. His body mass index (BMI) was 24.9. He had frank jaundice and his abdominal examination demonstrated right upper quadrant tenderness without hepatomegaly or peritonism. He did not have any ascites or lymphadenopathy.

Investigations

On laboratory investigation, the full blood count showed a haemoglobin of 146 g/L, white cell count of 5.7×10⁹/L with normal differential and platelet count of 179×10⁹/L. The urea, creatinine and electrolytes revealed baseline renal function with serum creatinine of 65 μmol/L, estimated glomerular filtration rate (eGFR) of 79 mL/min/1.73 m² and normal electrolytes. The liver function tests demonstrated a total bilirubin of 355 μmol/L of which 170 μmol/L was conjugated bilirubin, alanine transaminase 684 U/L, aspartate transaminase 873 U/L, gamma-glutamyl transferase 173 U/L and alkaline phosphatase 149 U/L. The prothrombin time was 29.7 s, the international normalized ratio (INR) 2.5, albumin 27 g/L and the venous lactate 3.2 mmol/L. On imaging evaluation, sonography of the abdomen was unremarkable, while MRI cholangiopancreatography (MRCP) ruled out biliary obstruction and identified early cirrhotic liver changes with possible portal hypertension. The patient was not known to have had cirrhosis prior to admission on previous clinical, biochemical or radiological investigations.

Serological testing revealed HBcAb positive and HBsAg negative status, with hepatitis B surface antibody (HBsAb) <3.00 mIU/mL. HBeAg was not tested. HBV DNA was detected and the viral load was 13 691 IU/L. The viral serologies and molecular tests for HIV, hepatitis A, hepatitis C, cytomegalovirus, Epstein-Barr virus and herpes simplex virus did not reveal infection, while the hepatic autoantibody and metabolic panels were negative. On retrospective review, the patient tested HBcAb positive with HBsAg, HBsAb and hepatitis B DNA negative status, and normal alanine transaminase prior to chemotherapy initiation 12 months earlier, but did not receive prophylactic antiviral therapy at the time. The hepatitis B genotype was not tested and the HBsAg and HBV DNA levels were not repeated again during his admission.

Differential diagnosis

The initial differential diagnosis included both medical and surgical causes of hyperbilirubinemia. On sonographic evaluation of the abdomen and on consideration of the pattern of liver function derangement, surgical causes were felt less likely. Notably, MRCP ruled out choledocholithiasis and pancreatic malignancy. A Doppler ultrasound ruled out hepatic vein thrombosis, and there was no history of haemodynamic instability to suggest ischaemic hepatitis. The MRCP revealed evidence of cirrhosis and suggested chronicity of hepatic impairment. However, there was no clinical, biochemical or imaging evidence of cirrhosis on prior investigations. There was no history of metabolic syndrome or excessive alcohol intake to suggest non-alcoholic steatohepatitis and alcoholic hepatitis, while a negative full liver panel reduced the likelihood of other metabolic and autoimmune causes of hepatic impairment. Of the medical causes, chemotherapy-induced liver injury and viral hepatitis were thought to be the most likely, with past lymphoproliferative disorder and rituximab exposure identified as significant red flags on admission. The presence of HBV DNA confirmed HBR.

Treatment

The patient was initially commenced on antibiotic coverage for ascending cholangitis with ceftriaxone and metronidazole, which was discontinued when biliary obstruction was excluded. Supportive care, including 3 days of intravenous vitamin K and nutritional support was provided. On day 2, entecavir 0.5 mg daily was commenced and continued for the duration of the patient’s admission.

Outcome and follow-up

The HBsAg status and HBV DNA levels were not followed during the admission, so ‘seroreversion’ and the response of the viral load to treatment was not documented, respectively. The patient’s liver function continued to deteriorate, such that on day 14, his total bilirubin was 776 μmol/L, INR 3.6 and venous lactate 3.7 mmol/L. This was associated with grade II hepatic encephalopathy. He developed hepatorenal syndrome, with a serum creatinine peak of 352 μmol/L and eGFR 12 mL/min/1.73 cm² with an associated urine albumin to creatinine ratio 36.2 on day 9 of admission. He was started on terlipressin infusion and two times a day concentrated albumin without recovery of renal function. After a 5-day trial of terlipressin, the patient’s goals of care were reoriented to palliative care. The patient died on day 15 of admission.

Discussion

In the case report, the patient had tested HBcAb positive, and HBsAg and HBsAb negative prior to initiation of chemotherapy. His HBsAb status is regarded negative or ‘non-responsive’ as it is <10 mIU/mL.¹⁹ On presentation to hospital, he remained isolated HBcAb positive, but had detectable HBV DNA, meeting criteria for HBR. There are a number of possible interpretations for an isolated anti-HBc. Among patients with intermediate-to-high epidemiological risk, resolved infection is most likely, in which the majority of individuals had recovered from a prior acute HBV infection and their anti-HBs had waned to undetectable levels.^{12 20} In low endemicity regions, false positive tests are a consideration, but this is less common.¹² Other explanations include early hepatitis B infection, in which only HBc IgM is detectable and also coinfection with HIV and HBV.^{12 20} Finally, occult hepatitis B infection can occur, where infection is from certain genotypes with HBsAg mutations that cause false negative HBsAg results.^{12 20 21} In this patient group, despite acute hepatis B infection, the HBsAg would remain undetectable. Of these explanations, resolved hepatitis B infection or occult hepatitis B infection seem most likely for the patient in the case report.

The prevalence of HBR varies geographically, with endemic areas having a proportionally greater burden of cases.²² The overall risk of HBR among patients receiving chemotherapy ranges between 20% and 57%.²³ Specifically, in patients with haematological malignancies, 18%–73% of HBsAg-positive patients and 34%–68% of HBsAg-negative/HBcAb-positive patients develop reactivation with immunosuppression.³ The exact rate of HBR in HBsAg-negative/HBcAb-positive patients exposed to rituximab is uncertain. A meta-analysis of isolated HBcAb-positive lymphoma patients undergoing rituximab treatment found great heterogeneity in HBR rates.²⁴ HBR occurred in 9% of exposed patients, with subgroup analysis revealing rates of 17% (95% CI 0.08 to 0.39, p<0.001) in prospective and 7% (95% CI 0.05 to 0.11, p=0.07) in retrospective studies.²⁴ Further prospective studies are needed to better characterise the risk of HBR in patients exposed to rituximab.

A number of risk factors for HBR have been identified. These include HBsAg positivity, high HBV DNA viral load, male gender, older age and exposure to high-risk chemotherapy, such as B-cell depleting therapies (eg, rituximab), high-dose steroids, anthracyclines, tyrosine kinase inhibitors and protease inhibitors.²⁵ The highest reactivation rates are observed among patients treated with rituximab combined with CHOP chemotherapy.²⁶ It has also been demonstrated that patients receiving rituximab alone experience more HBR than patients with non-rituximab chemotherapy.^{5 27 28} HBR peaks immediately after withdrawal of chemotherapy, with case series data demonstrating that reactivation usually occurs within 3 months of rituximab exposure, while 29% have reactivation after 6 months.^{27 29} Cases of delayed HBR after 12 months from exposure have been reported.^{30 31} HBR has a wide spectrum of clinical severity, ranging from asymptomatic biochemical recurrence to fulminant liver failure. Severe hepatitis occurs in 25%–50% of HBR cases, while the overall mortality rate in the setting of chemotherapy is 30%.^{2 32–34} Specifically for rituximab-related HBR, the case mortality rate among lymphoma patients is as high as 58% according to Food and Drug Administration registry data.²⁷

There are inconsistencies in national and international guidelines with regards to hepatitis B screening prior to chemotherapy and prophylactic antiviral use. These differences are a reflection of the regional variations in hepatitis B endemicity, healthcare system factors and cost–benefit considerations.¹⁴ In general, for serological screening, the Australian, Asia-Pacific and European gastroenterology society guidelines recommend universal screening, while the American gastroenterology and oncology guidelines suggest a risk stratification approach, in which only high-risk patients are tested.12 14–16 35 36 In particular, for patients with lymphoma receiving R-CHOP, universal prechemotherapy screening has been shown to reduce HBR 10-fold and lead to a higher 1-year survival rate than a risk stratification approach to screening.³⁷ Controversies regarding non-haematological malignancy screening remain.³⁸ In regard to prophylactic antiviral therapy, for isolated HBcAb patients most guidelines suggest treatment if the patient is receiving high-risk immunosuppressants.12 14–16 35 36 This is well supported by the evidence, and generally entecavir and tenofovir are the preferred first-line agents.^39–41 However, there are discrepancies in the recommendations for the duration of prophylaxis, ranging from 6 to 12 months for low-risk chemotherapy to 12–24 months for higher risk regimens and haematopoietic stem cell transplant.¹⁴ It is important to note that antiviral prophylaxis does not confer risk-free status, with HBR occurring once antiviral therapy is ceased.^42–44 There is a paucity of evidence surrounding the efficacy of antiviral therapy once HBR occurs, with case reports suggesting that early initiation of treatment may reduce morbidity and mortality.^45–48 However, many patients deteriorate despite antiviral therapy and the evidence surrounding prognosis is ultimately lacking.^45–48

In summary, HBR is an important consideration for any patient undergoing chemotherapy. Despite growing evidence surrounding the epidemiology and management of isolated HBcAb-positive populations undergoing chemotherapy, a lot of questions remain. The exact prevalence and risk of HBR, the role of universal screening in non-haematological malignancies, as well as the duration of prophylactic antiviral therapy and treatment efficacy for HBR flare are yet to be fully characterised.

Learning points.

• Hepatitis B reactivation is a predictable complication in patients receiving high-risk immunosuppressive therapy with isolated hepatitis B core antibody positive status.

• Patients with haematological malignancy and those exposed to B-cell-depleting therapies, such as rituximab are at greatest risk of hepatitis B reactivation.

• Hepatitis B reactivation can be delayed by many months post exposure to rituximab.

• Universal hepatitis B screening of patients receiving chemotherapy and prophylactic antiviral treatment with entecavir or tenofovir are endorsed by most gastroenterology and oncology society guidelines.

Ethics statements

Patient consent for publication

Next of kin consent obtained.