Safety of Long-Term Biologic Therapy in Rheumatologic Patients With a Previously Resolved Hepatitis B Viral Infection

European and Asian studies report conflicting data on the risk of hepatitis B virus (HBV) reactivation in rheumatologic patients with a previously resolved HBV (prHBV) infection undergoing long-term biologic therapies. In this patient category, the safety of different immunosuppressive biologic therapies, including rituximab, was assessed. A total of 1218 Caucasian rheumatologic patients, admitted consecutively as outpatients between 2001 and 2012 and taking biologic therapies, underwent evaluation of anti-HCV and HBV markers as well as liver amino transferases every 3 months. Starting from January 2009, HBV DNA monitoring was performed in patients with a prHBV infection who had started immunosuppressive biologic therapy both before and after 2009. Patients were considered to have elevated aminotransferase levels if values were >1x upper normal limit at least once during follow-up. We found 179 patients with a prHBV infection (14 treated with rituximab, 146 with anti-tumor necrosis factor-alpha, and 19 with other biologic therapies) and 959 patients without a prHBV infection or other liver disease (controls). The mean age in the former group was significantly higher than the controls. Patients with a prHBV infection never showed detectable HBV DNA serum levels or antibody to hepatitis B surface antigen/hepatitis B surface antigen seroreversion. However, when the prevalence of elevated amino transferases in patients with prHBV infection was compared to controls, it was significantly higher in the former group only for aminotransferase levels >1x upper normal limit but not when aminotransferase levels >2x upper normal limit were considered. Conclusion: Among patients with a prHBV infection and rheumatologic indications for long-term biologic therapies, HBV reactivation was not seen; this suggests that universal prophylaxis is not justified and is not cost-effective in this clinical setting. (Hepatology 2015;62:40-46)