Abstract
BACKGROUND: Despite major public health initiatives and the existence of efficacious treatment regimes, tuberculosis (TB) remains a threat, particularly in resource-limited settings. A significant part of the problem is the difficulty of rapidly identifying infected individuals, and as a result, there has been renewed interest in developing better diagnostics for infection or disease caused by Mycobacterium tuberculosis. Many of the existing tools, however, have limitations such as poor sensitivity or specificity, or the need for well-equipped laboratories to function effectively. Serodiagnostic approaches in particular have long drawn attention, due to their potential utility in large field studies, particularly in resource-poor settings. Unfortunately none of the serodiagnostic approaches have so far proven useful under field conditions. RESULTS: We screened a large panel of antigens with serodiagnostic potential by ELISA and selected a subpanel that was strongly and broadly recognised by TB patients, but not by controls. These antigens were then formulated into a simple immuno-chromatographic lateral flow assay format, suitable for field use, and tested against panels of plasma and blood samples from individuals with different clinical status (confirmed TB patients, household contacts, and apparently healthy community controls), recruited from Ethiopia (a highly TB-endemic country) and Turkey (a TB meso-endemic country). While specificity was good (97-100% in non TB-endemic controls), the sensitivity was not as high as expected (46-54% in pulmonary TB, 25-29% in extra-pulmonary TB). CONCLUSIONS: Though below the level of sensitivity the consortium had set for commercial development, the assay specifically identified M. tuberculosis-infected individuals, and provides a valuable proof of concept.