Reactivation of tuberculosis (TB) is a major concern during treatment with TNF inhibitors.¹ Different guidelines to detect active and latent TB have been recommended in various countries before starting treatment with these drugs. There is evidence that their application has led to a significant reduction in the number of cases of TB,² but we do not know which is the most cost effective strategy.

In our department 69 consecutive patients with rheumatoid arthritis (n = 53), ankylosing spondylitis (n = 10), and psoriatic arthritis (n = 6) considered for treatment with TNF inhibitors have recently been screened for TB infection according to the Italian guidelines. All underwent a careful history, tuberculin skin testing by intradermal injection of 0.2 ml 10 TU PPD (Mantoux method), and chest radiography. In order to enhance the sensitivity of tuberculin testing we had stopped steroid treatment in all patients at least 1 week before performing the test. Patients were considered to be affected by latent TB if they had any of the following conditions: (1) unequivocal history of previous TB; (2) positive tuberculin reaction (at least 5 mm skin induration at 72 hours); (3) radiographic lesions consistent with old TB (calcified nodular lesions, apical fibrosis, pleural scarring). According to our guidelines, patients with latent TB undergoing treatment with TNF inhibitors receive preventive chemotherapy. Our patients were predominantly women (63.8%) with a mean age of 55.8 years (range 21–81). We found a history of previous TB in 2.9% of patients, tuberculin positivity in 8.7%, and radiographic lesions consistent with latent TB in 20.3%. Globally, a diagnosis of latent TB was made in 24.6% of our patients, six of whom underwent treatment with TNF inhibitors (notably, five of the six had a negative tuberculin test). We started preventive chemotherapy with isoniazid in all patients but this drug was discontinued in four because of liver toxicity.

Our data suggest that tuberculin skin testing is not sufficiently sensitive to detect latent TB in patients with rheumatoid arthritis and other spondyloarthropathies or in those with inflammatory bowel diseases.³ In these patients chest radiography is essential if we do not want to miss a significant proportion of cases. The Italian guidelines for TB screening before starting treatment with TNF inhibitors allow recognition of these cases, increasing the indications for preventive chemotherapy. However, liver toxicity caused by isoniazid may be enhanced in these patients, probably due to concomitant treatment with other drugs such as methotrexate and NSAIDs. This suggests that the risk of chemoprophylaxis should be compared with the chance of contracting TB in the individual patient, and that a cost effectiveness evaluation of the different strategies used to minimise the risk of TB reactivation during treatment with TNF inhibitors is indicated.