The reactivation of latent tuberculosis (TB) is a major complication of tumor necrosis factor (TNF)-alpha inhibitors. Screening for TB infection is recommended before anti-TNF therapy is initiated; however, the use of tuberculin skin testing (TST) is controversial, due to the high rate of false-negative results in patients receiving immunosuppressive treatment. To compare the performance of two commercial interferon (IFN)-gamma release assays (IGRA), T-SPOT.TB (TS-TB) and QuantiFERON-TB Gold "In-tube" (QFT-GIT), with TST for the detection of TB infection in patients due to start anti-TNF therapy, 69 human immunodeficiency virus (HIV)-negative Italian patients (mean age: 45.2 +/- 12.6 years; male=39) were enrolled between September 2005 to August 2006. Patients affected by rheumatoid arthritis (n = 18), psoriatic arthritis (n = 26), ulcerous rectocolitis (n = 6), and Crohn's disease (n = 19) were tested simultaneously with TST, TS-TB, and QFT-GIT. Overall, 26% of patients were positive by TST, 30.4% by TS-TB, and 31.8% by QFT-GIT. Agreement with TST was similar (kappa = 0.21, p = 0.0002 and kappa = 0.26, p < 0.001, respectively). In 11 TST-negative cases, IFN-gamma release assays were positive. In addition, in seven Mantoux-positive cases with no TB risk factors, TST result agreement was achieved with at least one blood test. Indeterminate results were detected in 5.8% and 2.8% of cases, respectively, with TS-TB and with QFT-GIT (p = not significant [ns]). In conclusion, our results suggest that IGRAs may be helpful for screening purposes in patient candidates for anti-TNF therapy to confirm positive TST results and in selected cases when false-negative results are suspected. The utility of blood tests in patients with low or no TB risk remains to be assessed.