Clinical characteristics and diagnostic test performance among children evaluated for tuberculosis in Kenya

Introduction Tuberculosis (TB) is a leading cause of illness and death in children globally, but bacteriological confirmation remains challenging. There is an urgent need to develop improved diagnostic approaches in children. Methods Three hundred children <5 years of age with symptoms suggestive of TB disease were evaluated at baseline and over six months (October 2013 – February 2016) in Kenya. Up to fourteen specimens per child were tested for Mycobacterium tuberculosis complex by fluorescence microscopy, Xpert MTB/RIF assay (Xpert), and mycobacterial culture. The sensitivity and specificity of different specimen types and microbiological tests were determined using standardized research clinical case definitions. Findings Among 266 children with sufficient information to be classified according to the research clinical case definition, 36% (96/266) had TB disease and of these, 32% (31/96) had bacteriologically confirmed intrathoracic TB. Compared to culture, the sensitivity of a single Xpert ranged from 60-67% and specificity from 97·5%-100% for different specimen types. Interpretation Despite extensive specimen collection and laboratory testing, in almost two-thirds of children with intrathoracic TB according to research clinical case definitions, TB could not be bacteriologically confirmed. Improved diagnostic tests are needed to identify children with clinical TB disease and to exclude other potential causes of illness. Funding United States Agency for International Development, Centers for Disease Control and Prevention, President's Emergency Plan for AIDS Relief (PEPFAR), National Institutes of Health.