Reactivation of Hepatitis B Virus (HBV) Associated with Rituximab Use in Lymphoma: Under Reporting and Lack of Data Quality Reported to the FDA.

HBV is a major public health problem with infection that can lead to cirrhosis, liver failure, and death. Immunosuppressive therapy, such as steroids and/or chemotherapy, is known to cause a flare or “reactivation” of HBV (HBV-react). Rituximab was approved in 1997 for the treatment of B-cell lymphoma. In 2004, based on 3 case reports, the FDA warned healthcare professionals of rituximab-associated HBV-react. Since that time, multiple cases and retrospective series of rituximab-induced HBV-react have been reported in the literature. However, the characteristics and scope of this association still remains largely unknown. We evaluated the characteristics of patients with lymphoma who developed HBV-react after exposure to rituximab, and the quality of case reports available in the medical literature and at the FDA.

Data sources included 2 observations at Northwestern Memorial Hospital, 83 reports from the medical literature, and 10 reports obtained from the FDA MedWatch database (n=97). Two reports of rituximab-related HBV-react not associated with lymphoma (vasculitis and gloumerulonephritis) were excluded. HBV-react was defined as ≥ 2-fold increase in serum HBV DNA with an increase in serum ALT compared with baseline. A completeness analysis was performed comparing cases submitted to the FDA MedWatch database vs. the medical literature and active surveillance.

Of 83 unique cases of HBV-react associated with rituximab reported in the literature, 28 were published as case reports, while 55 were included in case series. Of these 83 cases, 46 occurred in patients (pts) with anti-HBV core antibody (HBcAb+) in the absence of HBV surface antigen (HBsAg-), while 37 cases involved pts with chronic HBV hepatitis (i.e, HBsAg+). Among the 28 case reports and 2 Northwestern cases (n=30; HBcAb+ n=16, HBsAg+ n=14), the median age at HBV-react was 55 years (23M/9F). Histology of these cases were diffuse large B-cell lymphoma (n=19), indolent lymphoma (n=8), CLL (n=2), and mantle-cell lymphoma (n=1). In terms of concomitant treatment at time of HBV-react, 25 pts were receiving concurrent immunosuppressive therapy (chemotherapy +/- steroids n=22 and steroids n=3), while only 5 cases involved single-agent rituximab treatment. Each of these 5 latter pts had received chemotherapy prior to rituximab treatment (2, 3, 12, 24, and 34 months). The median number of rituximab doses received prior to HBV-react was 6 (range 3-10). The median time from last rituximab dose to HBV-react was 5 months (0-21 months); of note, 30% of cases occurred >6 months from last rituximab dose. In terms of outcome, 60% of pts experienced fulminant liver failure, while the remaining had HBV-related hepatitis. Furthermore, 40% (12/30) of pts died due to HBV-react. Quality comparison of source data of the literature and surveillance reports vs. the FDA is contained in Table 1. Overall completeness ratio for literature and observed reports vs. the FDA was 2.18. Of rituximab-related HBV-react occurrences (n=55) reported in 8 case series, five studies included a control group; there was a suggestion of increased risk of HBV-react with rituximab-based therapy, especially with concurrent steroids, although the absolute risk was not consistent (e.g., among HBcAb+ cases, reported rate of HBV-react: 2.7% to 23.8%).

We have found 95 total cases of rituximab-associated HBV-react. A paucity of safety reports regarding rituximab-associated HBV-react have been reported to the FDA MedWatch. Furthermore, published cases in the medical literature and through active surveillance were superior in data quality compared with FDA reports. However, the absolute risks of rituximab-related HBV-react in HBsAg+ or HBcAb+ pts are still not known. Further examination of the relationship of HBV-react with single-agent rituximab and rituximab combined with immunosuppressive therapy, with and without steroids, using an active surveillance strategy is warranted.

No relevant conflicts of interest to declare.