Safety of tumor necrosis factor α blockers in hepatitis B virus occult carriers (hepatitis B surface antigen negative/anti–hepatitis B core antigen positive) with rheumatic diseases




To assess the safety of anti–tumor necrosis factor α (anti-TNFα) therapy on the course of hepatitis B virus (HBV) infection in carriers of antibodies to hepatitis B core antigen (anti-HBc) affected by chronic inflammatory arthropathies.


From January 2001 to December 2008, HBV markers were determined before the first administration of anti-TNFα agents in all 732 patients affected by inflammatory arthropathies treated with anti-TNFα at 2 outpatient rheumatologic clinics in Northern Italy. Anti-HBc–positive patients were prospectively evaluated and HBV markers and HBV DNA were assessed every 6 months, in case of aminotransferase elevation, and at the end of the study.


At the time of recruitment, 72 patients were anti-HBc carriers, 5 of whom were positive for hepatitis B surface antigen (HBsAg) and not included in the study. The ratio of men:women was 26:41 and the mean ± SD followup was 42.52 ± 21.33 months. Of the patients, 25 were treated with infliximab, 23 with etanercept, and 19 with adalimumab. Fifty-one patients were treated also with methotrexate, 52 with nonsteroidal antiinflammatory drugs, and 43 with prednisone (3 with a dosage >7.5 mg/day). All anti-HBc patients were HBV DNA negative at the first observation. During followup, no patient presented HBV reactivation with viral load increase and no patient became HBsAg positive.


Anti-HBc positivity in HBsAg-negative patients is a sign of previous HBV infection and does not indicate chronic hepatitis. In these patients, anti-TNFα therapy appears to be quite safe, as no HBV reactivation was found in our study. Nevertheless, careful monitoring is necessary.


It is well known that therapy for rheumatic diseases with tumor necrosis factor α (TNFα) blockers can be associated with an increased rate of serious infections (1). Patients have to be screened for the presence of infections before the initiation of treatment, and markers for viral hepatitis B and C should be a part of this assessment because the long-term safety of these drugs in case of chronic hepatitis is not known (2). In the general population, hepatitis B virus (HBV) infection is frequently detected, and ∼350 million people worldwide are estimated to be chronic carriers of the virus. HBV is regarded as a leading cause of acute hepatitis, cirrhosis, and hepatocellular carcinoma (3).

Persistent HBV infection is defined as overt when hepatitis B surface antigen (HBsAg) is detectable and as occult in HBsAg-negative subjects in whom HBV DNA can be revealed in the serum or liver. Patients who present antibodies to hepatitis B core antigen (anti-HBc) with concurrent HBsAg negativity do not have chronic hepatitis, but only experienced HBV infection and were able to clear it. Nevertheless, some of these patients may be occult carriers, in whom intrahepatic HBV replication can be detected despite HBsAg and HBV DNA negativity in the serum and despite the absence of signs of active liver disease (4, 5).

The reactivation of HBV replication in patients undergoing immunosuppressive therapy is a frequently reported complication in oncology and hematology. In this field, the frequency of viral reactivation in HBsAg-positive patients under immunosuppressive treatment ranges from 14% to 50%, while mortality ranges from 5% to 12% (6–8). Reactivation can also occur in HBsAg-negative patients with serologic markers of HBV infection, such as anti-HBc and antibodies to HBsAg (anti-HBs), in the case of chemotherapy for hematologic diseases. This complication is reported in <5% of cases (9).

Among HBsAg-positive patients treated with infliximab for Crohn's disease, several cases of HBV reactivation have been described (10–14), including a case of reactivation in an HBsAg-negative, anti-HBc–positive patient (15). Data from patients treated with TNFα blockers for rheumatic diseases show some case reports of HBV reactivation in HBsAg-positive patients as well (16–22), and a single case of reactivation in an HBsAg-negative and anti-HBc–positive patient (23). A recent retrospective study showed 1 case of viral reactivation among 8 HBsAg carriers treated with anti-TNFα therapy (24). On the other hand, many cases have been described in which TNF blockade, either in association with lamivudine prophylactic therapy or not, did not lead to HBV reactivation (6, 25–30), and recently a retrospective study showed no reactivation during anti-TNFα therapy (31). The use of lamivudine to prevent HBV reactivation is recommended in HBsAg-positive patients undergoing immunosuppressive therapies, including anti-TNFα drugs (4, 32), but its use can lead to viral resistance (33). There are no prospective studies regarding HBV reactivation and TNFα blocker therapy, and in particular few data are available about HBsAg-negative, anti-HBc–positive patients. The aim of this study was to assess the safety of TNFα blockade in these patients affected by chronic inflammatory arthropathies.


From January 2001 to December 2008, 732 consecutive patients affected by inflammatory arthropathies started treatment with anti-TNFα at 2 outpatient rheumatologic clinics in Northern Italy. Serologic HBV markers were performed as a part of the baseline assessment prior to the beginning of therapy. Anti-HBc carriers were prospectively followed up for HBV reactivation. HBV reactivation was defined as HBsAg appearance and/or detectable HBV DNA in the serum. Inclusion criteria for the study were diagnosis of rheumatoid arthritis (RA) according to the American College of Rheumatology (formerly the American Rheumatism Association) criteria (34), or ankylosing spondylitis (AS) according to the modified New York criteria (35), or psoriatic arthritis (PsA) according to the criteria by Moll and Wright (36); documented anti-HBc positivity; no evidence of active HBV replication before anti-TNFα treatment; and treatment with anti-TNFα (infliximab, etanercept, or adalimumab, following the current indications).

All patients included in the study gave informed consent according to local guidelines. Disease activity was evaluated at baseline by the Disease Activity Score in 28 joints (DAS28) (37) in patients with RA and by the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) (38) in patients with PsA and AS.

Serum alanine aminotransferase (ALT) and serum aspartate aminotransferase (AST) were used as markers of hepatic injury, and HBsAg, anti-HBs, anti-HBc, hepatitis B e antigen (HBeAg), antibodies to HBeAg (anti-HBe), and viral load were used as markers of disease progression. ALT and AST were determined by standard laboratory methods. Markers for HBV were identified using the following techniques. HBsAg, anti-HBs, anti-HBc, HBeAg, and anti-HBe were detected by Architect/Aeroset (Abbott Diagnostics). HBV DNA was evaluated at different laboratories by AMPLICOR HBV MONITOR (Roche Molecular Diagnostics Global), RealART HBV LC PCR kit (Artus), and Molecular Beacons Assay (Molecular Diagnostics, Abbott Laboratories). Serum ALT and AST were followed up every 2 months, HBV markers and HBV DNA were monitored every 6 months and in the case of aminotransferase elevation, ALT and AST were monitored again 3 weeks after aminotransferase elevations, and all 4 markers were monitored at the end of followup (Figure 1). Patients left the study in December 2008 or in case of discontinuation of anti-TNFα therapy; the last assessment of liver function, HBV markers, and viral load was included in the analysis.

Figure 1.

Design of the study. HBsAg = hepatitis B surface antigen; anti-TNF-α = anti–tumor necrosis factor α; anti-HBc = antigen to hepatitis B core antigen; HBV = hepatitis B virus; pts = patients.

A descriptive statistical analysis was performed by SPSS statistical software, version 10.00 for Windows (SPSS). Data were expressed as the mean and SD or median and interquartile range (IQR) when appropriate.


Of 732 patients, 72 were identified as anti-HBc carriers, but 5 of them were HBsAg positive and therefore were not enrolled in the study; the latter were treated with lamivudine prophylaxis before the onset of anti-TNFα therapy, and are still under treatment in accordance with the current guidelines (4, 32). Sixty-seven patients were included in the study. The ratio of men:women was 26:41, the mean ± SD age was 57.36 ± 12.63 years, the median disease duration was 8 years (IQR 4–15 years), and the mean ± SD followup was 42.52 ± 21.33 months. A single patient had a followup of 6 months only and 5 patients had a followup of 12 months. Sixty-one patients were followed for >18 months, among whom 52 were followed for >24 months and 35 were followed for >36 months.

Of the patients, 59 were affected by RA, 4 by PsA, and 4 by AS. All patients presented at the beginning of the study with active disease: in patients with RA, the mean ± SD DAS28 was 5.65 ± 0.96, and in patients with PsA and AS the mean ± SD BASDAI score was 5.1 ± 1.57. Of the patients, 25 were treated with infliximab, 23 with etanercept, and 19 with adalimumab. Of the 67 patients, 51 were also treated with weekly methotrexate (MTX), 52 with nonsteroidal antiinflammatory drugs (NSAIDs), 40 with prednisone ≤7.5 mg/day, and 3 with prednisone >7.5 mg/day. The demographic and clinical features of the patients are reported in Table 1.

Table 1. Baseline characteristics of patients included in the study (n = 67)*
Characteristics of patientsValue
  • *

    Values are the number of patients unless otherwise indicated. IQR = interquartile range; DAS28 = Disease Activity Score in 28 joints; BASDAI = Bath Ankylosing Spondylitis Disease Activity Index.

Age, mean ± SD years57.36 ± 12.63
 Rheumatoid arthritis59
 Psoriatic arthritis4
 Ankylosing spondylitis4
Disease duration, median (IQR) years8 (4–15)
DAS28, mean ± SD5.65 ± 0.96
BASDAI score, mean ± SD5.1 ± 1.57
Followup, mean ± SD months42.52 ± 21.33
Tumor necrosis factor α inhibitor 
Nonsteroidal antiinflammatory drugs52
Prednisone dosage 
 ≤7.5 mg/day40
 >7.5 mg/day3

The baseline liver assessment is described in Table 2. In particular, the median values of ALT and AST were within the normal range, no patients presented detectable HBV DNA (including the 21 patients positive for anti-HBe), and 2 patients presented concurrent hepatitis C virus (HCV) infection. After 6 months of followup, anti-HBc positivity was not confirmed in 2 patients. No case of HBV reactivation was noted, in particular no patients presented the appearance of HBsAg, and no increase in viral load was observed. At least 1 spike (2 times the upper normal limit) of aminotransferase was noted in 34 patients, but the concurrent monitoring of markers of HBV reactivation revealed no changes. The median time of occurrence was 12 months (IQR 8–16 months) after the onset of biologic agent therapy. Of these patients, 28 were taking MTX and 30 were taking NSAIDs. In 28 patients, a subsequent AST and ALT determination 3 weeks later showed the normalization of aminotransferase levels without the need of any modification in therapy. For 3 patients in whom aminotransferase levels were still exceeding the upper normal limit after 3 weeks, MTX was withdrawn, and in 2 patients NSAIDs were discontinued. In these patients, normalization of aminotransferase levels was seen 1 month after the modification of therapy. In 1 patient with concurrent HCV infection, aminotransferase levels did not decrease, the subsequent investigations showed an increase in HCV viremia, and liver biopsy demonstrated active hepatitis. HCV reactivation was diagnosed, anti-TNFα treatment was stopped, and the patient left the study.

Table 2. Serologic hepatitis markers and liver tests at baseline and at the end of followup*
Liver assessmentBaselineEnd of followup
  • *

    Values are the number of patients unless otherwise indicated. HBsAg = hepatitis B surface antigen; HBeAg = hepatitis B e antigen; Anti-HBc = antibodies to hepatitis B core antigen; Anti-HBe = antibodies to HBeAg; Anti-HBs = antibodies to HBsAg; HBV = hepatitis B virus; AST = aspartate aminotransferase; IQR = interquartile range; ALT = alanine aminotransferase.

  • December 2008 or discontinuation of tumor necrosis factor α blockers.

  • Anti-HBc positivity was not confirmed in 2 patients.

  • §

    Normal value 11–39 mU/ml.

  • Normal value 11–34 mU/ml.

HBsAg positive00
HBeAg positive00
Anti-HBc positive6765
Anti-HBe positive2123
Anti-HBs positive2832
HBV DNA positive00
Hepatitis C virus positive22
AST, median (IQR) mU/ml§21 (17–25.75)19 (17–25)
ALT, median (IQR) mU/ml20 (12.25–28.75)20 (15.75–24.25)

Six other patients left the study before December 2008. Two patients discontinued TNFα blockers for lack of efficacy and were treated with rituximab, without HBV reactivation until December 2008. One patient stopped anti-TNF because of infusion reaction, 1 for the development of heart failure, and 2 for inefficacy; these patients were treated afterward with traditional disease-modifying antirheumatic drugs and none of them experienced HBV reactivation. One patient, a 60-year-old woman with RA and dyslipidemia treated with adalimumab, died after 8 months of followup because of myocardial infarction, but the liver assessment at 6 months did not show HBV reactivation. There were no patients lost at followup; all 7 patients who left the study had to discontinue anti-TNFα for inefficacy or adverse events.

The liver assessment at the end of followup is reported in Table 2. The only changes observed in the final liver profiles were the appearance of anti-HBs in 4 patients and of anti-HBe in 2 patients.


HBV reactivation is a commonly reported complication in patients undergoing chemotherapy for solid organ or hematologic malignancies, and in patients after bone marrow transplantation (6, 7). Less data are available on the safety of TNFα blockers in the course of HBV infection (2, 32).

In experimental models, TNFα seems to play different roles in the course of HBV infection: some studies show the importance of proinflammatory cytokines, and of TNFα in particular, in the clearance of HBV (39) and the suppression of HBV replication (40). Elevated serum and hepatic levels of TNFα (41, 42) and TNFα receptor p75 (43, 44) can be found in patients with acute or chronic hepatitis B. Studies of HBV infection in animal models show that TNFα produced by HBV-specific cytotoxic T lymphocytes (CTLs) down-regulates HBV replication in hepatocytes by noncytopathic mechanisms (45). In TNFα-knockout mice, lack of TNFα induces impaired proliferation of HBV-specific CTLs (46), which are thought to be responsible for viral clearance and liver damage during HBV infection. This antiviral function could therefore be insufficient in the case of a lack of TNFα. On the other hand, there is evidence that TNFα induces hepatocyte apoptosis, and this phenomenon is enhanced by the presence of HBV infection (47).

Data deriving from clinical experience in treating Crohn's disease and chronic inflammatory arthropathies include several case reports of HBV reactivation in HBsAg carriers (11, 13, 16, 18–21, 24, 48, 49). Recently, Kaur and colleagues reported the result of a liver biopsy performed on a HBsAg carrier treated with adalimumab for RA, in whom the viral load had increased after 3 months of therapy. Liver histology revealed minimally active hepatitis and therapy was continued for 2 years without the institution of any prophylaxis, and viral load decreased after 9 months of adalimumab therapy and remained stable afterward. Two years later the patient was switched to etanercept for loss of efficacy, and lamivudine prophylaxis was initiated. Since then, HBV DNA has not been detectable (27). A recent review of the literature revealed that all HBsAg carriers receiving a prophylactic lamivudine regimen did not develop active liver disease (49); however, no prospective studies are available on this subject.

Even less data are available on HBsAg-negative anti-HBc carriers, who represent a cluster of individuals who just experienced HBV infection that the organism was able to fight against. This serologic status corresponds in the majority of cases with complete viral clearance, but a subgroup of patients may present detectable liver HBV DNA, demonstrating asymptomatic HBV replication and defining the status of occult carrier. In accordance with this status, 2 cases of reactivation in HBsAg-negative, anti-HBc–positive patients have been described during anti-TNFα therapy (15, 23). For this reason, the occurrence of HBV reactivation during anti-TNFα treatment in anti-HBc carriers is probably unlikely, but it represents a risk that needs to be defined.

Recent guidelines from the Italian Association for the Study of the Liver (AISF) on the management of HBV infection during immunosuppressive therapy suggest the use of lamivudine, a nucleoside analog, in HBsAg carriers undergoing immunosuppressive therapies, including TNFα blockers (4). However, the use of lamivudine in patients receiving an extended course of immunosuppressive therapy has an uncertain benefit versus risk rate, as its prolonged employ is associated with the emergence of lamivudine-resistant mutants (30, 33, 50). Therefore, lamivudine administration as universal prophylactic therapy in all patients with previous contact with HBV should be considered with caution.

To our knowledge, the present study is the first one carried out prospectively on a large series of HBsAg-negative, anti-HBc–positive patients followed for >3 years during anti-TNFα treatment. This cohort was followed from the onset of biologic agent therapy in order to assess the risk rate of HBV reactivation in the absence of lamivudine administration. The most important result of our study is that no patient showed any sign of viral reactivation.

A limitation of this study could be the fact that our cohort comprised individuals who had already experienced a long course of immunosuppressive therapy before beginning TNFα blockers without presenting HBV reactivation, as the median disease duration ranged from 4 to 15 years. For this reason, indirect serologic HBV marker variations, including the appearance of anti-HBs and of anti-HBe in a few patients, should be interpreted with caution. The high percentage of anti-HBs negativity in our series may be related to previous immunosuppressive treatment, but it could also indicate that the antibodies were lost due to the long amount of time that had passed since infection had occurred. The disappearance of anti-HBc in 2 patients during the followup could also be related to fluctuations in antibody titers, even though we cannot exclude testing failure or unspecific reactions. These considerations led us to give particular importance, for study purposes, to direct indicators of viral replication, including both HBsAg and HBV DNA, as is also suggested by AISF guidelines (4).

In all 34 patients presenting an increase in aminotransferase levels of 2 times the upper normal limit, we did not observe HBsAg seroconversion or increase in viral load, even if no prophylactic therapy was administered. In 3 patients, the discontinuation of MTX therapy led to the normalization of functional liver tests, and in 2 patients discontinuation of NSAIDs was necessary. It is well known that NSAIDs and MTX can be responsible for drug-related hepatotoxicity, and MTX is listed among the high-risk therapies with concern to HBV reactivation, because some cases have been described even in occult carriers (6). One patient presented HCV reactivation, whereas the remaining 28 patients showed only a slight and transitory increase of aminotransferase.

In conclusion, our results support the safety of TNFα blockade in anti-HBc carriers, even when prophylaxis with lamivudine is not administered. Therefore, patients should only be monitored for rises in aminotransferase and for viral reactivation or HBsAg seroconversion, adopting an early intervention approach (4).

Our data, showing a very low risk of viral reactivation in a large cohort of patients, may be helpful in defining the optimal monitoring strategy in potential occult HBV carriers treated with long-term anti-TNFα therapy for rheumatic diseases.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Caporali had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Caporali, Bobbio-Pallavicini, Montecucco.

Acquisition of data. Caporali, Bobbio-Pallavicini, Atzeni, Sakellariou, Caprioli, Sarzi-Puttini.

Analysis and interpretation of data. Caporali, Bobbio-Pallavicini, Atzeni, Sakellariou, Caprioli.