Objectives: To identify diagnostic features associated with culture of Mycobacterium tuberculosis (MTB), the standard for tuberculosis (TB) diagnosis, to inform clinical end point definitions for new TB vaccine trials.
Methods: Children <2 years of age (n = 1445) were screened and investigated for TB during a Bacille Calmette Guerin vaccine trial in South Africa. Standardized clinical, radiologic, and microbiologic data were collected, including paired gastric lavage and induced sputum for MTB liquid culture. Adjusted odds ratios (AORs) were calculated using a multivariate logistic regression model.
Results: Adjusted odds of positive MTB culture increased by 90% with history of wheezing (AOR, 1.9) and by 4% with each 1-mm increase in Mantoux diameter (AOR, 1.04). Odds of positive MTB culture doubled if the chest radiograph was suggestive of pulmonary TB (AOR, 2.16) and more than tripled if lower chest retraction was observed clinically (AOR, 3.37). Fever, night sweats, and presence of lymphadenopathy were negatively associated with MTB culture (AOR: 0.5, 0.62, and 0.2, respectively). Persistent cough, weight loss, and failure to thrive were not significantly associated with MTB culture in this study population.
Conclusions: Wheezing and lower chest retraction, consistent with intrathoracic airway obstruction; chest radiography suggestive of pulmonary tuberculosis; and Mantoux diameter were predictive of positive MTB culture. These variables should be considered for inclusion in composite clinical end point definitions for infant TB vaccine trials. Several clinical features, commonly used for TB diagnosis in older children, were not associated with positive MTB culture among children younger than 2 years.