Rheumatoid arthritis (RA) is a chronic, debilitating polyarthritis affecting ∼1% of the adult population. Current treatment recommendations (1) include early use of disease-modifying antirheumatic drugs (DMARDs), singly or in combination, aimed at improving symptoms and retarding joint erosion. The initiation of treatment with biologic agents should, according to the existing guidelines, be delayed until 1 or more of the conventional DMARDs has been shown to fail to control the disease. Nevertheless, the development of new agents designed to target key molecules such as TNFα, a central cytokine in the inflammatory cascade (2), has provided new hope for the efficacious management of this potentially devastating disease. Furthermore, despite their substantial cost, TNFα antagonists have been shown in a number of clinical studies to be a cost-effective therapeutic modality.

Although generally considered safe, the use of TNFα inhibitors has been associated with an increased risk of opportunistic infections, particularly tuberculosis (TB) (3), mainly involving reactivation of latent infection during the early phase of treatment. After the need for strict screening criteria and initiation of chemoprophylaxis in potentially susceptible individuals was established, new reported cases of TB were impressively decreased, but not eliminated (4).

From a cohort of 180 RA patients (∼600 patient-years) receiving anti-TNFα therapy at our center, 5 patients were hospitalized for prolonged fever, with clinical symptoms and radiographic evidence of respiratory tract infection. Onset of symptoms coincided with a positive purified protein derivative (PPD) skin test result for TB in 4 of the patients, all of whom had exhibited a negative PPD result (<5 mm induration) during the screening prior to initiation of anti-TNFα treatment. The diagnosis of TB was established by lymph node biopsy, lung biopsy, and sputum culture in 1 patient each; in the remaining 2 patients it was based on clinical and radiographic criteria as well as ex juvantibus, i.e., response to treatment. Three of the patients developed active TB within 6 months of treatment initiation; this was considered to be reactivation of latent TB with a false-negative skin test result during screening, since no additional risk factor besides anti-TNF treatment could be identified and no case of active TB infection was reported among their close associates. The other 2 patients developed TB 30 months and 46 months, respectively, after initiation of anti-TNF treatment, and their cases were considered to more likely represent newly acquired TB. The prevalence of latent TB is relatively low among the native Greek population; new cases of TB are, however, increasing in number, partly due to rising immigration/repatriation rates.

All 5 of our patients received anti-TB treatment (ethambutol, isoniazid, or a combination of isoniazid and rifampin [3 also received streptomycin]) until complete recovery (6–12 months) (Table 1). Anti-TNFα treatment (infliximab in 4 of the patients and adalimumab in 1 patient) was discontinued according to the current treatment guidelines (1), and RA treatment during the TB treatment phase consisted solely of low-dose corticosteroids with subsequent reinitiation of DMARDs (methotrexate).

Table 1. RA patients in whom active TB developed during anti-TNFα treatment*
Patient/age/sexRA duration, yearsAnti-TNFα treatment (months)TB localizationBasis of TB diagnosisTB treatment (months)Anti-TNFα agent readministered (months after TB flare)Duration of anti-TNFα readministration, months
  • *

    RA = rheumatoid arthritis; TB = tuberculosis; anti-TNFα = anti–tumor necrosis factor α; inflix. = infliximab; ETB = ethambutol; RIF = rifampin; INH isoniazid, STR = streptomycin; ada. = adalimumab.

1/73/F20Inflix. (6)PulmonaryRadiographic findings, ex juvantibusETB (3), RIF/INH (12)
2/63/M7Inflix. (46)PulmonaryEx juvantibusSTR (1), ETB (3), RIF/INH (12)Inflix. (12)17
3/63/F23Inflix. (2)Cervical lymph-adenopathyLymph node biopsyETB (3), INH/RIF (10)Ada. (48)19
4/79/F15Inflix. (5)PulmonarySputum cultureSTR (1.5), ETB (3), INH (12)
5/64/F10Inflix. (4), ada. (26)PulmonaryBronchoscopy/ lung biopsySTR (1.3), ETB (3), RIF/INH (6)Ada. (6)27

Three of the 5 patients experienced disease flare after the initial discontinuation of anti-TNF treatment (6 months, 1 year, and 4 years, respectively, after discontinuation). Since the primary disease could not be controlled with corticosteroids and/or DMARDs only, we elected to readminister anti-TNFα agents (adalimumab in 2 patients and infliximab in 1). The patients were informed in detail and gave their consent. In 2 of the patients, anti-TNF treatment was reinitiated immediately after completion of anti-TB treatment; in the third patient, it was reinstituted 3 years after completion of the anti-TB treatment. In 2 of the 3 patients, the previously used anti-TNF agent was readministered, mainly because of good clinical response during the initial treatment phase. The third patient was switched to another TNFα antagonist due to a mild allergic reaction to the anti-TNFα agent originally administered. Reinitiation of treatment was followed by rapid remission of RA symptoms and laboratory abnormalities. All 3 patients were still receiving anti-TNF therapy at the time of submission of this manuscript (17–27 months after reinstitution) and, despite the fact that they are not receiving any prophylactic anti-TB treatment, they show no clinical, chest radiographic, or molecular evidence of relapse of TB infection (negative findings on gene probe polymerase chain reaction detection assay for mycobacterial genes in peripheral blood and urine).

Two clear issues are 1) whether it is safe to reintroduce anti-TNFα therapy in patients who have exhibited tuberculosis during the course of their initial treatment, and 2) if so, what is the optimum time point for reintroduction? Tumor necrosis factor is an essential cytokine for host defenses against tuberculosis. It is expressed by monocytes after phagocytosis of mycobacteria or after stimulation by mycobacterial proteins or glycolipids, subsequently inducing the release of cytokines, endothelial adhesion molecules, and chemokines. Animal studies as well as genetic analysis of polymorphisms have shown that the functional TNF gene complex is essential for the formation and maintenance of granulomas (5), rendering affected individuals susceptible to granulomatous infections. TNFα antagonism–associated TB is typically characterized by the absence of orderly formed granulomas, although this finding remains mostly anecdotal than officially reported (6). Granulomas provide the microenvironment in which mycobacteria become sequestered; however, they may also serve as the sanctuary for surviving dormant mycobacteria responsible for future reactivation of TB infection.

Data regarding the safety of TNFα antagonist readministration following TB flare are sparse. The interference of anti-TNFα antibodies with the formation of granulomas in combination with appropriate anti-TB therapeutic regimens may in fact improve sterilization in affected individuals. From this point of view, TNFα antagonism has even been proposed as potential adjuvant immunotherapy for the treatment of TB (7).

Based on the findings in our 3 patients in whom reinitiation of anti-TNFα therapy proved to be effective and safe, at least within the time frame of our followup, we propose that readministration of anti-TNF treatment should not be ruled out, and that it could represent a safe option for patients with otherwise uncontrollable RA flares. Since it is difficult to distinguish whether TB in anti-TNFα–treated patients represents a flare of latent disease or reacquisition of new TB infection (8), especially in areas of high prevalence, close, longer-term followup of re-treated patients is mandatory in order to establish the safety of anti-TNF readministration in analogous cases.


  1. Top of page

Dr. Aslanidis 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 design. Aslanidis, Pyrpasopoulou.

Acquisition of data. Aslanidis, Pyrpasopoulou, Douma.

Analysis and interpretation of data. Aslanidis, Pyrpasopoulou, Douma, Petidis.

Manuscript preparation. Aslanidis, Pyrpasopoulou, Douma, Petidis.