Tumor Necrosis Factor Blockers and Reactivation of Latent Tuberculosis

Patients with rheumatoid arthritis (RA) may be at increased risk of reactivation of infectious diseases that normally remain latent in granuloma (e.g., tuberculosis and histoplasmosis). This may be because of an immunosuppressed state that accompanies autoimmune disease or because of immunosuppressive treatments, such as steroids or monoclonal anti-TNF antibodies (mAb). Among immunosuppressed patients, we expect the pattern of tuberculosis disease to be unusual, with a higher incidence of extrapulmonary and disseminated disease than among immunocompetent patients. Unlike fungal disease, the prescribing physician thankfully can perform tests and administer anti-tuberculosis prophylaxis to prevent tuberculous complications associated with autoimmune disease and treatment. In the United States, an increased risk of reactivation tuberculosis in RA has been suggested but not proven, but it has been demonstrated for patients with RA taking infliximab, a monoclonal TNF blocker [1]. In this issue of Clinical Infectious Diseases, Mohan et al. [2] describe 25 cases of reactivation tuberculosis, the data for which were obtained by passive reporting, that occurred in patients after they received etanercept, a soluble TNF-receptor blocker (sTNFR) [2]. The evidence linking sTNFR use and tuberculosis is strengthened by animal model data, which show that TNF plays a key role in maintaining tuberculous infection in the latent state. Mohan et al. [2] point out, however, that a comparison of data for these 25 patients with available background data does not demonstrate an increased risk of reactivation tuberculosis for patients with RA who are receiving etanercept. TNF blockers have been in use for >5 years, and a comparison of data in the report by Mohan et al. [2] with reported mAb rates suggests that tuberculosis reactivation may occur less often with sTNFR than with mAb TNF blockers. This comparison is significantly limited by methodological differences in surveys and by the limitations of passive reporting. This notion is supported by the temporal association of tuberculosis reactivation and infliximab, which is associated with a shorter treatment period (median duration, 11 weeks) [3] and less variability than etanercept (median duration, 11 months). This difference between these 2 classes of TNF blockers in the propensity to cause reactivation tuberculosis is reflected in the tuberculosis screening advice in the drug package inserts and the conclusions of the US Food and Drug Association advisory committee. There are theoretical reasons why mAbs might have a stronger association with reactivation tuberculosis than do sTNFRs. The difference could simply be dose related, yet both classes of agents treat the TNF-mediated diseases with similar efficacy. The difference in rates of tuberculosis could be a an artefact of population sampling, with more infliximab-associated patient reports from Europe, but assessment of American patients only also shows an increased risk in the mAb-treated group [1]. Drug synergy might be an issue, because mAbs are prescribed concomitant with methotrexate for patients with RA. However, many patients receiving sTNFRs also receive methotrexate, and methotrexate by itself has not been associated with tuberculosis reactivation [1]. The pharmacologies of the 2 classes of agents differ significantly, with mAbs causing a longer, more sustained neutralization of more TNF species than do sTNFRs [4]. Animal studies have suggested that a partial, rather than complete, TNF blockade may allow preservation of the beneficial and anti-inflammatory functions of this cytokine in cases of tuberculosis immunity [5]. More studies are required to compare the effects of these different classes of TNF blockers on tuberculosis immunity. Some studies of patients treated with TNF blockers suggest that infliximab may affect important protective aspects of the host defence, such as monocyte and T cell viability [6], Th1 skewing of T cells [7], and IFN-γ activity [8]. The results of these studies involving TNF blocker—treated patients will require more careful control for underlying immune defects that are present in patients because of their coexisting autoimmune condition. Mohan et al. [2] suggest that caution should be used with TNF blockers and that appropriate measures should be taken to avoid tuberculosis reactivation. The best option for treating latent tuberculosis infection is a 9-month regimen of isoniazid, but individual countries will have to generate their own guidelines that take into account local conditions. In Europe, the European Agency for the Evaluation of Medicinal Products has developed a useful patient alert card for patients receiving infliximab, which will help in managing latent tuberculosis (http://www.emea.eu.int/pdfs/human/press/pus/003202.pdf). In my opinion, prescribers of these agents should be aware of a number of issues. TNF blockers should not be used until active tuberculosis has been ruled out or completely treated if present. Tests for detection of latent tuberculosis infection (LTBI) should be performed and LTBI diagnosed if the patient has a positive tuberculin skin test (TST) result and normal chest radiograph findings. Receipt of bacille Calmette-Guérin vaccination during childhood should be ignored. Patients with latent tuberculosis should receive prophylactic therapy, if possible. Treatment with TNF blockers should be delayed, if possible, until prophylaxis is complete. Isoniazid is potentially hepatotoxic (affecting 10%–15% of recipients), which would argue for avoiding treatment in countries in which there is a low prevalence of tuberculosis. I think that this would be a mistake, because of the potentially dire consequences of failing to treat LTBI in this patient group. These include disseminated disease that may be initially refractory to treatment because of lingering TNF blockers in the system [9]. Tuberculosis is a public health issue, and failed treatment could lead to an outbreak of tuberculous involving a family or indeed other immunosuppressed individuals with whom the TNF-treated patient may be in contact. This advice—which involves screening for latent tuberculosis and treating it—appears to be working. In this regard, Gomez et al. [10] reported that the number of tuberculosis cases decreased after the publication of guidelines that promote prophylactic treatment. Administration of anti-tuberculosis medication to these patients is often complex, and it would be beneficial to seek the advice of a tuberculosis specialist when treating them. The effects of immunosuppressive treatment and autoimmunity are likely to result in a high rate of false-negative TST results. Although there is currently no data to support the administration of prophylactic isoniazid to such patients, a negative TST result should be considered suspect, and these patients should be carefully monitored for an atypical presentation of tuberculosis. This issue is especially important in countries in which there is a high prevalence of latent tuberculosis. Physicians in these areas also have to contend with the possibility of increased susceptibility to primary tuberculosis or exogenous tuberculosis reinfection after receipt of TNF blockers. In this regard, continued gathering of data will be required to see whether the TNF blockers are associated with primary tuberculosis disease. By focusing on LTBI and the immunosuppressed patient, I would hope that the article by Mohan et al. [2] leads to improved awareness among prescribers of preventable reactivation tuberculosis in patients with chronic inflammatory disease who are receiving old and new immunosuppressive agents. This is especially important, because TNF blockers are being licensed for treating more and more diseases and because new TNF-blocker treatments are being developed, such as TNF vaccines and anti-TNF proteases. Conflict of interest. J.K. is on the speakers' bureaus of Wyeth and Centocor.