ABSTRACT Pediatric pulmonary tuberculosis (TB) remains a diagnostic challenge. Children swallow their sputum rather than coughing it up, making analysis with standard TB tests difficult to perform. We led the development and testing of the Xpert MTB/RIF (Xpert) assay, which simultaneously detects the presence of M. tuberculosis and its susceptibility to the important first line drug rifampin in less than two hours. Hailed as a major advance for detecting adult pulmonary TB, the test is suboptimal for detecting pediatric TB since sputum samples may be difficult to obtain, and repeated testing is necessary for acceptable sensitivity. Here, we propose to adapt the Xpert sputum assay so that it can be performed on the stool of pediatric TB suspects, developing a simple- to-use kit that can rapidly detect TB in children. We will then test this assay using a newer assay cartridge (Ultra) that is approximately 10 times more sensitive and almost twice as rapid as current Xpert MTB/RIF cartridge. Our hypothesis is that swallowed M. tuberculosis currently detected in induced sputum samples or gastric aspirates of pediatric patients should be concentrated and detectable in stool. Stool is the final repository of all swallowed respiratory secretions and it is easily obtainable from infants and children, making it an excellent test matrix. We will develop and test a simple sample processing bottle that can pre-process large (> 1 gram) volumes of stool, removing particulate matter and PCR inhibitors, without compromising PCR TB detection. The contents of the bottle will be able to be squeezed into the Xpert assay cartridge with little difficulty or unpleasantness. The “Ultra stool” assay will be as easy to perform as the “Xpert sputum” assay, and will have the same or better diagnostic sensitivity and specificity. The successful completion of this project will literally transform pediatric pulmonary tuberculosis diagnosis. This proposal builds on an existing US – South African collaboration focused on testing novel TB assays, enabling an established clinical and laboratory team to work on this new problem. This project has the following four aims: 1) develop and test “out of cartridge” sample processing protocols that maximize the sensitivity of the Ultra assay to detect M. tuberculosis in stool. 2) Perform extensive analytic tests of the Ultra stool assay focusing on limit of detection, inclusivity, specificity, effect of different stool samples, effect of different M. tuberculosis strains and susceptibility to interfering substances. 3) Develop and test a simple stool pre-processing bottle. 4) Perform iterative evaluation of the assay using stool samples from a cohort of children presenting with symptoms suggestive of pulmonary TB .

PUBLIC HEALTH BENEFIT It can be very difficult to diagnose tuberculosis in children since they do not easily cough up sputum where tuberculosis is usually detected. Instead, we think that Mycobacterium tuberculosis bacteria end up in children's stool, because they swallow their sputum after coughing. This project will develop a simple method to rapidly detect tuberculosis in children by examining their stool samples with a simple and rapid test. Rapid detection of tuberculosis will save many lives by enabling more rapid treatment for this disease.