Tuberculosis control relies on the identification and preventive treatment of individuals who are latently infected with Mycobacterium tuberculosis. However, direct identification of latent tuberculosis infection is not possible. The diagnostic tests used to identify individuals latently infected with M. tuberculosis, the in vivo tuberculin skin test and the ex vivo interferon-gamma release assays (IGRAs), are designed to identify an adaptive immune response against, but not necessarily a latent infection with, M. tuberculosis. The proportion of individuals who truly remain infected with M. tuberculosis after tuberculin skin test or IGRA conversion is unknown. It is also uncertain how long adaptive immune responses towards mycobacterial antigens persist in the absence of live mycobacteria. Clinical management and public healthcare policies for preventive chemotherapy against tuberculosis could be improved, if we were to gain a better understanding on M. tuberculosis latency and reactivation. This statement by the TBNET summarises knowledge and limitations of the currently available tests used in adults and children for the diagnosis of latent tuberculosis infection. In summary, the main issue regarding testing is to restrict it to those who are known to be at higher risk of developing tuberculosis and who are willing to accept preventive chemotherapy.