It is recommended to evaluate the presence of latent tuberculosis infection (LTBI) prior to the use of antitumor necrosis factor α. The aim of this study is to assess the presence of LTBI in patients with rheumatic diseases undergoing treatment with infliximab in an endemic area for tuberculosis (TB). LTBI was searched through the contact history, chest X-ray and tuberculin skin test with purified protein derivative (PPD) ≥5 mm. We studied 157 patients in the period from May 2005 to October 2008, 99 (63.1%) were women with average age of 49 years and 58 (36.9%) were men with average age of 41 years. The group comprising 90 patients (57.3%) with rheumatoid arthritis (RA), 54 (34.4%) with ankylosing spondylitis (AS) and 13 (8.3%) with psoriatic arthritis (PsA) had PPD reactor 13.4% (21/157), being prevented by isoniazid (INH) in these patients. There are dissimilar responsiveness to the PPD between the three pathologies, and the reactivity was lower in RA (RA × AS: χ(2) = 12; P = 0.0004; and RA × PsA: χ(2) with Yates' correction = 3.6; P = 0.05). No significant difference between the reactivity of the PPD and the use of immunosuppressive drugs (P = 0.81) is observed. The immunoprophylaxis with INH showed an efficacy of 95% (20/21); three (1.9%) patients developed active TB (spondylodiscitis, meningitis and lymphadenopathy) after the use of infliximab, reaffirming extrapulmonary involvement. These results suggest that PPD has a low sensitivity for detection of LTBI in RA and that the previous use of immunosuppressive drugs does not affect the response to PPD.