TNF-⍺ antagonist therapy is associated with a risk of severe, extrapulmonary, disseminated tuberculosis, which is fatal in 10% of cases. The risk of tuberculosis is increased four-fold in patients on TNF-⍺ antagonist therapy. The main risk factors are a history of untreated or inadequately treated primary tuberculosis, recent contact with a tuberculosis patient, and residence in or travel to a high-endemicity region. Infection surveillance agencies throughout the world have issued recommendations to ensure the detection and treatment of latent tuberculosis before TNF-⍺ antagonist initiation. These recommendations have returned the incidence of tuberculosis to the level seen before the introduction of TNF-⍺ antagonists. Nevertheless, there is still room for improvement. Recommendations about latent tuberculosis screening include the use of tuberculin skin tests. However, these tests are positive in individuals vaccinated with the BCG vaccine, which leads to overuse of tuberculosis chemoprophylaxis and, therefore, to unnecessary patient exposure to hepatotoxic effects. Furthermore, tuberculin skin tests may be falsely negative in immunosuppressed patients, leading to underuse of tuberculosis prophylaxis. These shortcomings of tuberculin skin tests have generated interest in interferon-gamma release assays (IGRAs). In patients with overt tuberculosis, IGRAs are more sensitive and more specific than tuberculin skin tests. However, the accuracy of IGRAs for diagnosing latent tuberculosis remains unknown, because no reference standard is available. In addition, patients taking immunosuppressant agents to treat systemic disease may exhibit anergia, which complicates the interpretation of IGRAs. Until additional data become available, caution requires that IGRAs be used only when a positive or negative result, as assessed on a case-by-case basis, will help to decide whether tuberculosis chemoprophylaxis is in order.