We have previously demonstrated that Mycobacterium bovis BCG-specific immunoglobulin G antibodies in lymphocyte secretions (ALS) can be employed as a marker for active tuberculosis (TB). We aimed to determine whether the ALS method allows detection of subclinical TB infection in asymptomatic individuals. A prospective study of family contacts (FCs) of patients with active TB and healthy controls was performed. Thirteen of 42 FCs had high ALS responses, including 6 FCs who subsequently developed active TB. No correlation was observed between the tuberculin skin test and the ALS responses in the FCs (r ؍ 0.1, P ؍ 0.23). Among patients with active TB, BCG-specific ALS responses steadily declined from the time of diagnosis through 6 months following antimycobacterial chemotherapy (P ؍ 0.001). The ALS assay enabled detection of infection in exposed symptom-free contacts, who are at greater risk for developing active TB. The method may also allow discrimination between effective treatment of active infection and suboptimal response to therapy.With the emergence of multidrug-resistant Mycobacterium tuberculosis and the pandemic of human immunodeficiency virus, tuberculosis (TB) is a reemerging infectious disease concern worldwide. About one-third of the world's population is believed to be latently infected with M. tuberculosis (7). Immunocompetent individuals with M. tuberculosis infection have a 10% risk during their lifetimes of developing active disease, becoming new foci for transmission, with half of this risk being in the first 2 years after exposure (5, 25). It is therefore important to identify people at risk for recent infection with M. tuberculosis, i.e., close contacts of patients with infectious pulmonary TB, persons having clinical conditions associated with an increased risk for progression to active TB (for example, human immunodeficiency virus infection, injection drug use, and fibrotic lesions on chest radiograms), children, and those who are in an early active phase of their infection (1, 4).Determination of the prevalence of latent infection is helpful for better understanding the epidemiology of TB and for designing and evaluating TB control strategies (1). The accurate diagnosis of latent or subclinical TB infection is an important component of many TB control programs and depends largely on tuberculin skin testing (TST). In developed countries, in the absence of a better method, the TST using M. tuberculosis purified protein derivative (PPD) is currently used nearly exclusively. However, in developing countries with high rates of M. bovis BCG vaccine use, TST is unreliable because of the broad antigenic cross-reactivity of PPD with environmental nontuberculous strains and the BCG vaccine. About 10 to 20% of patients with proven TB and without any apparent immunosuppression have negative TST results (13,20). The sensitivity of TST is low, especially in critically ill and immunosuppressed patients with disseminated TB (14). Therefore, there is substantial need for methods to accurately discri...