Biomarkers and diagnostics for tuberculosis: progress, needs, and translation into practice.

Human infection with Mycobacterium tuberculosis can progress to active disease, be contained as latent infection, or be eradicated by the host response. Tuberculosis diagnostics classify a patient into one of these categories. These are not fixed distinct states, but rather are continua along which patients can move, and are affected by HIV infection, immunosuppressive therapies, antituberculosis treatments, and other poorly understood factors. Tuberculosis biomarkers-host or pathogen-specific-provide prognostic information, either for individual patients or study cohorts, about these outcomes. Tuberculosis case detection remains difficult, partly because of inaccurate diagnostic methods. Investments have yielded some progress in development of new diagnostics, although the existing pipeline is limited for tests for sputum-smear-negative cases, childhood tuberculosis, and accurate prediction of reactivation of latent tuberculosis. Despite new, sensitive, automated molecular platforms for detection of tuberculosis and drug resistance, a simple, inexpensive point-of-care test is still not available. The effect of any new tests will depend on the method and extent of their introduction, the strength of the laboratories, and the degree to which access to appropriate therapy follows access to diagnosis. Translation of scientific progress in biomarkers and diagnostics into clinical and public health programmes is possible-with political commitment, increased funding, and engagement of all stakeholders.