To determine the incidence and risk of tuberculosis in rheumatoid arthritis (RA) patients exposed or unexposed to TNF⍺ antagonists, the impact of recommendations about managing latent tuberculosis, the time to diagnosis of active tuberculosis, and the proportion of extrapulmonary forms.

Systematic review of articles retrieved using Medline. From each article, we abstracted the incidence and risk of tuberculosis in RA patients exposed or unexposed to TNF⍺ antagonists, the duration of TNF⍺ antagonist exposure at the diagnosis of tuberculosis, and the distribution of the tuberculosis foci.

We selected 14 articles. The risk of tuberculosis was increased 2- to 10-fold in RA patients unexposed to TNF⍺ antagonists and 2- to 4-fold in those exposed to TNF⍺ antagonists, compared to the general population. The incidence of tuberculosis in TNF⍺ antagonist-treated patients varied across studies (9.3 to 449/100,000) according to the country, observation period, and TNF⍺ antagonist used. The risk was greater with monoclonal antibodies than with the soluble receptor. Official recommendations have decreased the risk of tuberculosis in TNF⍺ antagonist-treated patients. Over half the cases of active tuberculosis were diagnosed during the first treatment year. Among TNF⍺ antagonist-treated patients with tuberculosis, 60% had extrapulmonary lesions. Disseminated tuberculosis was more common with monoclonal antibodies.

The risk of tuberculosis is increased during TNF⍺ antagonist therapy, and the increase is larger with the monoclonal antibodies than with the soluble receptor. Tuberculosis during TNF⍺ antagonist therapy is a rare event that occurs early after treatment initiation. Extrapulmonary involvement is common and potentially severe. Therefore, clinicians should direct careful attention to the risk of tuberculosis associated with TNF⍺ antagonist therapy.