BACKGROUND Screening and treatment of latent tuberculosis infection (LTBI) prior to anti-tumour necrosis factor alpha (anti-TNF-alpha) therapy has been shown to decrease the incidence of active tuberculosis (TB) by more than 80%, and is recommended by the British Thoracic Society. In the absence of a gold standard test for LTBI, conventional screening currently involves taking a clinical history of risk factors, a chest X-ray and a tuberculin skin test (TST) which can be difficult to interpret in immunosuppressed patients. Alternative cellular immune-based screening tests have been developed to detect Mycobacterium tuberculosis. OBJECTIVE To examine, evaluate and summarize the quality of evidence on the use of interferon gamma release assay (the ELISpot test) in the diagnosis of latent tuberculosis prior to initiation of anti-TNF-alpha and examine the agreement with the tuberculin skin test. METHODS Ovid Medline, Embase and the Cochrane library were searched using search terms that included tuberculosis, each of the current anti-TNF-alpha biological agents, TST and interferon-gamma release assay. Terms were searched using MeSH (medical subject headings) terms and/or free text where relevant. RESULTS Discordance between tuberculin skin test and ELISpot is greater in individuals who have had the bacillus Calmette-Guérin (BCG) vaccination and are taking corticosteroids. ELISpot technique using CFP-10 and ESAT-6 antigens is more sensitive than TST in detecting M. tuberculosis infection in patients taking corticosteroids. ELISpot avoids cross-reaction with BCG, making it a more specific test in this group of patients. Agreement between the tests was found to be fair (72.8% kappa value=0.38). CONCLUSIONS Tuberculosis resulting from reactivation of latent tuberculosis following treatment with anti-TNF is a continuing problem. Screening reduces the risk but does not eliminate it. Further studies are needed into the cost-effectiveness and sensitivity of ELISpot and the tuberculin skin test in routine clinical practice.