Evaluation of New Strategies for the Diagnosis of Tuberculosis Among Pediatric Contacts of Tuberculosis Patients

Background: In young children, underdiagnosis and diagnostic delay have an adverse effect on morbidity and mortality of tuberculosis (TB). This study evaluated new strategies for early TB diagnosis using an outpatient protocol in children between 0 and 5 years of age, with a recent household TB contact. Methods: Case recruitment was performed in Manaus, Amazonas, Brazil, from 2008 to 2009. Epidemiologic and clinical data, tuberculin test, chest radiograph and 2 induced sputum respiratory samples from each participant were obtained. Laboratory diagnosis was based on Lowenstein-Jensen (LJ) culture, mycobacteria growth indicator tube (MGIT) and polymerase chain reaction. We conducted a study of comparison of diagnostic tests and a study of cases and controls to identify the clinical characteristics of the population with positive culture and polymerase chain reaction results. Results: A total of 102 children were evaluated. Thirty-two fulfilled criteria of suspicion of TB. MGIT was more sensitive (P = 0.035) and faster (P < 0.001) than LJ. Clinical score, MGIT, LJ and polymerase chain reaction presented no concordance or slight concordance. A positive MGIT culture was only associated with a strong tuberculin test reaction (P = 0.026). The combination of MGIT with the clinical score allowed the diagnosis of 33% more cases with little or no symptomatology compared with the exclusive use of the clinical classification. Conclusions: The sensitivity and speed of MGIT demonstrate the utility of liquid cultures for the diagnosis in children. Furthermore, these results suggest that the use of MGIT in children presenting recent household TB contact and a strong tuberculin test reaction may be a strategy to improve early TB diagnosis.