Authors' reply

摘要

We thank Marcel Behr and colleagues for their comments. We share their view on the importance, of endpoint specificity in tuberculosis vaccine efficacy trials but disagree with their interpretation of our data.1 Since the incidence rate ratios (IRR) for endpoints 0-4 presented in Behr and colleagues' table do not differ statistically, no differences in vaccine efficacy might be inferred from the IRR point estimates presented. Pulmonary tuberculosis in young children is known to be pauci-bacillary, and only a minority is positive for Mycobacterium tuberculosis. However, endpoint 1 contained more microbiologically positive cases than negative- cases, suggesting that our composite clinical case definition was rigorous.