The occurrence of tuberculosis (TB) in patients treated with immunosuppressive drugs (ISD) is an old problem that has been highlighted by cases occurring in patients using anti-TNFalpha drugs. After a brief review of anti-tuberculosis immunity to highlight the immunosuppressant targets, we show how an epidemiological approach allows the control of risk in patients with drug-induced immunosuppression. The assessment and the control of this risk are usually complicated by underlying immunosuppressant disease, co-morbidities, associated drugs and local disease prevalence. Steroid therapy in systemic diseases and ISD protocols in graft rejection prevention illustrate this problem particularly well. The management strategy adopted to combat anti-TNF related tuberculosis in rheumatoid arthritis (RA) has allowed these biases to be avoided. The incidence of TB in RA has been recorded in some national databases (USA, Spain). A four fold increase was registered after the introduction of anti-TNF agents in 2001 which could be considered as the true risk of the drug. Several national health agencies proposed guidelines to screen and prevent TB risk in these patients. Their effectiveness was confirmed by the incidence of TB returning the level prior to the introduction of these agents. Recommendations could be improved by using interferon-gamma screening tests and a better benefit/risk prophylaxis. They should be observed and if possible extended to new and other ISD.