Preventing Hepatitis B reactivation in immunosuppressed patients: Is it time to revisit the guidelines?


Hepatitis B virus (HBV) infection remains a serious global health concern. Approximately one-third of the world's population has evidence of previous HBV infection, and 350 million people are chronic HBV carriers (1). In the US, the rate of new HBV infections has declined by approximately 82% since 1991, when a national strategy to eliminate HBV infection was implemented (2). The greatest decline has been among those born since 1991, when universal vaccination of children was first recommended (2). Despite these improvements, HBV remains prevalent in the US, with 800,000 to 1.4 million individuals estimated to be chronic carriers (2). Although intravenous drug users and homosexual men are at a high risk for chronic HBV, most cases in the US are in people emigrating from high-prevalence areas (e.g., Asia or Africa), where early-life horizontal and vertical transmission are common (3).

Reactivation of HBV in immunosuppressed individuals has been well documented in the literature for several decades. Reactivation can occur either at the cessation of therapy when immune reconstitution occurs, or with prolonged immunosuppression that can result in an accelerated course of HBV infection (4–6). Most cases occur in individuals with cancer undergoing chemotherapy, where reactivation or exacerbation among HBV-positive patients is common (e.g., >50% in some lymphoma series) particularly in regimens containing rituximab, and sometimes fatal (7–18). In rheumatology, HBV reactivation has been reported with a number of disease-modifying antirheumatic drugs (DMARDs), although the lack of large studies makes it hard to ascertain the exact risk for most drugs. However, it is increasingly apparent that newer biologic therapies, such as tumor necrosis factor α (TNFα) inhibitors and rituximab, may pose a significant risk of HBV reactivation. Therefore, it is especially timely to take stock of how rheumatologists approach HBV in clinical practice.

In this issue of Arthritis Care & Research, Stine et al report the results of a national survey of rheumatologists regarding screening for and management of HBV infection in patients initiating immunosuppressive therapies (19). Although only 15% of the 1,000 rheumatologists approached completed the survey, the authors' findings are nonetheless interesting and provide a useful starting place for evaluating rheumatologists' awareness of HBV and their practice patterns regarding screening and management. The study highlights areas of significant practice variation among rheumatologists, especially pertaining to screening (e.g., who is screened and which laboratory tests are used for screening), prophylaxis against reactivation, and monitoring intervals for HBV laboratories in chronic carriers who are receiving immunosuppressive therapies. This observed variation in physician practice patterns likely reflects a number of factors, including the lack of a strong evidence base in some of these areas, heterogeneity among patient populations, and the absence of clear and specific guidelines regarding HBV screening and management in patients with rheumatic disease. The study also highlights several areas where quality improvement efforts should be targeted.


With regard to screening, 42% of the respondents routinely screen for HBV before beginning nonbiologic DMARDs and 69% screen before beginning biologic DMARDs. The former number is somewhat difficult to interpret because specific DMARDs were not defined, and therefore medications that carry little or no risk of causing HBV reactivation, such as hydroxychloroquine, may have figured into the physicians' responses. Nevertheless, given the known risk of reactivation with most nonbiologic DMARDs, these results may be a cause for concern. Almost all physicians reported screening in the presence of certain HBV risk factors (e.g., human immunodeficiency virus, intravenous drug use, and men who have sex with men). However, respondents were less likely to universally screen other well-established high-risk populations, such as individuals deriving from endemic regions or health care workers. These variations in screening highlight a limitation of identifying individuals with HBV by assessing risk factors in clinical practice: screening for risk factors may not be performed or may be incomplete. Another important limitation is that screening strategies to identify individuals at high risk may have poor predictive value, since a large percentage of infected individuals may not have easily identifiable risk factors (20, 21). Taking these realities into consideration, we agree with the recent Centers for Disease Control and Prevention (CDC) guidelines that recommend universal screening among persons with rheumatologic disorders receiving immunosuppressive therapies (3).

Risk awareness

The authors' findings regarding rheumatologists' awareness of drug package inserts are notable given the high percentage of respondents who were unaware or unsure of HBV screening recommendations; ranging from a low of 19% for anakinra, to a high of 53% for rituximab. If we assume that awareness of drug package inserts is a reasonable proxy for awareness of risk, then these findings may also be a cause for concern. Evidence is very clear that rituximab can induce HBV reactivation, even in patients with remote HBV infection (hepatitis B surface antigen [HBsAg]–negative but anti-hepatitis B surface antibody [anti-HBs]–positive) (8, 9, 11–18, 22) (Table 1), and there have been several deaths reported from fulminant liver failure in these patients. These reports are extremely compelling, and strongly suggest that all patients initiating therapy with rituximab should be screened, regardless of the absence of risk factors.

Table 1. Case reports of HBV reactivation among rheumatic patients treated with TNFα inhibitors or rituximab*
Drug (ref.)IndicationAge, yearsSexPretreatment serologyDurationOther drugsAntiviral prophylaxisOutcome
  • *

    HBV = hepatitis B virus; TNFα = tumor necrosis factor α; RA = rheumatoid arthritis; HBsAg = hepatitis B surface antigen; HBe = hepatitis B e antibody; HBc = hepatitis B core antibody; MTX = methotrexate; AST = aspartate aminotransferase; ALT = alanine aminotransferase; MP = methylprednisolone; HCQ = hydroxychloroquine; SSZ = sulfasalazine; AS = ankylosing spondylitis; HBeAg = hepatitis B e antigen; ANCA = antineutrophil cytoplasmic antibody.

Infliximab (29)RA49MHBsAg+, anti-HBe+, anti-HBc+18 monthsMTX 10 mg/wk, prednisone 8 mgNoHepatitis (AST 336 and ALT 573, HBV DNA 1,492 pg/ml); treated with lamivudine
Etanercept (30)RA62FHBsAg+, anti-HBe+, anti-HBc+2 yearsMP 8 mg, MTX 10–15 mg/wk, HCQ 200 mgNoHepatitis (AST 112 and ALT 234, HBV DNA 1,590 copies/ml); etanercept later restarted with lamivudine prophylaxis
Etanercept (31)RA48FHBsAg+, HBV DNA−13 monthsMTX, SSZ, HCQNoSlight increase in transaminases; HBV DNA 514 copies/ml
Infliximab (32)AS43MHBsAg+, HBeAg−, anti-HBe+, anti-HBc+14 weeks (5 mg/kg)NoneNoALT increased to 49, HBV DNA became positive; lamivudine started and infliximab successfully continued
Etanercept (33)AS73MHBsAg−, anti-HBs+, anti-HBc+, anti-HBe+14 monthsPrednisone 5 mgNoAST rose to 141, ALT to 65, HBV DNA 1,507 IU/ml; treated with lamivudine; etanercept later restarted successfully with concomitant lamivudine
Infliximab (34)AS35FHBsAg+, HBe−, anti-HBe+, HBV DNA−3 infusionsMTX 15 mg/weekNoHepatitis (ALT 10 × normal; HBV DNA positive); treated with lamivudine and infliximab successfully continued
Infliximab (35)AS31MHBsAg+, HBe−, anti-HBe+, HBV DNA−3 infusions (5 mg/kg)NoHepatitis (AST 457, ALT 1,054, HBV DNA 3.13 × 106 IU/ml); resolved with entecavir; this was only the patient to reactivate out of 8 HBV carriers in a population of 103 patients treated with anti-TNFα therapy
Rituximab (36)RA56FHBsAg+, HBe−, anti-HBe+, HBV DNA 9.6 × 102 IU/ml2 infusions (1,000 mg)YesHepatitis (AST 110, ALT 150, HBV DNA >1.1 × 108IU/ml); tenofivir added to lamivudine; transaminases normalized
Rituximab (37)ANCA-associated vasculitis73MHBsAg−, anti-HBc+4 infusions (375 mg/m2)Prednisolone 40 mgNoHBsAg+, HBV DNA >110 × 106 IU/ml 11 months later; patient died of renal failure; unclear if HBV reactivation contributed to decline

Experience with TNFα inhibitors in rheumatic patients with chronic HBV infections remains limited, but a growing number of case reports suggest that reactivation is a concern (Table 1). Corroborating the reports in the literature, Stine et al found that 7.4% of the rheumatologists surveyed had also witnessed HBV reactivation with biologic DMARDs (4 cases with infliximab and 1 case with etanercept were reported) (19). Reflecting the growing concern about HBV reactivation with these agents, Health Canada issued a class warning in 2006 regarding HBV reactivation with TNFα inhibitors. The Food and Drug Administration followed suit soon thereafter, and package inserts for TNFα inhibitors now carry similar warnings.

Laboratory testing

The authors queried rheumatologists regarding the laboratory tests routinely used to assess HBV. They did not assess which combination of tests rheumatologists use; instead, they reported the percentage of rheumatologists that used each of the 6 assays (liver function tests, HBsAg, anti-HBs, hepatitis B core antibody [anti-HBc], HBV DNA, and hepatitis panel). A vast majority of rheumatologists (92%) use HBsAg, the classic assay used to determine HBV infection. Somewhat surprising is the lower percentage of respondents who routinely assess anti-HBc. In recent years, the importance of this test in defining a carrier state in some individuals has become clear, with worrisome reports, such as those cited above, of individuals who were HBsAg-negative having reactivation of their disease. Based on these reports, many experts, as well as the recent CDC recommendations (3), suggest that screening for HBV in patients initiating immunosuppressive therapy should routinely include HBsAg, anti-HBs, and anti-HBc. Assessment of anti-HBs is important to determine the need for vaccination in patients initiating immunosuppressive therapy. In addition, although HBV DNA was reported by 7% of respondents, this assay is not appropriate for screening. Instead, HBV DNA should be used to evaluate viral replication or response to antiviral therapy in patients with established chronic HBV infections, or to determine whether the patient with anti-HBc has occult infection with active viral replication.

Prophylaxis and monitoring

Lastly, the survey questioned rheumatologists about prophylaxis and monitoring of individuals with HBV infection. Although avoiding immunosuppression in the setting of HBV may be preferable, poorly-controlled rheumatic disease that results in reduced function, decreased quality of life, or significant morbidity and mortality, may lead physicians and patients to consider therapy. Unfortunately, experience in the form of randomized controlled trials in this area among patients with rheumatic disease is not available. However, there is growing evidence in the literature that antiviral prophylaxis, along with careful monitoring for HBV reactivation, may be a reasonable strategy if the patient understands the risks of this approach (23–25). Still, the fact that most rheumatologists feel uncomfortable in this scenario is evidenced by the finding that most respondents (81%) would prefer to defer to a gastroenterologist or hepatologist to determine prophylactic therapy. Most respondents report that patients were given lamivudine, a medication that is associated with a high degree of viral resistance; many experts now recommend the use of newer antiviral agents such as entecavir, especially if the anticipated duration of treatment is long (>12 months) (26, 27).

Revisiting the American College of Rheumatology (ACR) Guidelines

Although assessing the effectiveness of dissemination strategies for ACR guidelines was not an explicit goal of the study by Stine et al, their findings on this topic are interesting nonetheless (19). The authors report that many rheumatologists in their sample were unaware of any ACR guidelines regarding screening for HBV prior to starting either biologic therapy (30%) or nonbiologic DMARD therapy (47%). These results are interesting given that the study was conducted several months after the publication and widespread dissemination of the ACR 2008 Recommendations for the Use of Nonbiologic and Biologic DMARDs in Rheumatoid Arthritis (28). Contained in the guidelines, which were developed using a validated method for combining scientific evidence and expert consensus, are several recommendations regarding HBV screening and management.

Given the lack of robust scientific evidence regarding hepatitis B screening strategies in patients with rheumatoid arthritis, the recent ACR recommendations are based largely on expert consensus. For screening, the recommendations state that high-risk patients (e.g., individuals using intravenous drugs, those with multiple sex partners in the previous 6 months, and health care personnel) receiving leflunomide or methotrexate should be screened for HBV. The guidelines go on to say that an appropriate evaluation might include tests for HBsAg, anti-HBs, and anti-HBc. Noticeably absent is a recommendation for HBV screening prior to starting biologic DMARD therapy. Rather than representing an omission, it is likely that the process used to establish the recommendations yielded neither enough scientific evidence nor a high enough degree of expert consensus to result in advocating specific screening practices. However, the fact that the authors grappled with the issue is apparent in another section of the guidelines that lists contraindications for starting or resuming immunosuppressive therapy in patients with rheumatoid arthritis. In that section, the authors state that acute infection with HBV is a contraindication to starting all DMARDS except hydroxychloroquine. For chronic HBV infection, the recommendations are somewhat more complicated and are based on the patients' Child-Pugh classification. In patients treated with antiviral therapy for HBV, leflunomide and methotrexate were contraindicated for all Child-Pugh classifications, and minocycline and sulfasalazine were contraindicated for Child-Pugh class C only. For untreated chronic HBV, leflunomide, methotrexate, minocycline, and sulfasalazine were contraindicated for all Child-Pugh classifications, and hydroxychloroquine was contraindicated for Child-Pugh class C. Biologic agents were contraindicated in all HBV patients classified as Child-Pugh B or C.

In the next iteration of the ACR guidelines, recommendations regarding HBV should be strengthened and simplified. We believe a major impediment to a widely understood and accepted screening strategy is the use of wording requiring an assessment of personal risk before testing. Such language (incorporated in both the ACR guidelines and most package inserts) suggests that HBV risk assessment is reliable and routinely obtained. Although we agree that all patients at high risk for HBV infection should be screened, it would serve the rheumatology community far better to state that all patients initiating high-risk drugs (i.e., those contraindicated in acute and chronic forms of HBV infection) should be screened with an appropriate panel regardless of predetermined risk. Screening for HBV should include HBsAg to establish the presence of chronic infection, anti-HBs to establish the need for vaccination, and anti-HBc to determine whether the patient may be an occult carrier. Consultation with a gastroenterologist or hepatologist to specify a monitoring regimen and prophylactic therapy for those with chronic HBV prior to initiating immunosuppressive therapy is also prudent.


Drs. Yazdany and Calabrese drafted the article, revised it critically for important intellectual content, and approved the final version to be submitted for publication.


The authors wish to thank Jennifer Guy, MD for her review of the manuscript.