Abstract
BACKGROUND: Microbiological diagnosis of pediatric pulmonary and intestinal tuberculosis remains difficult due to difficulty in specimen collection such as with sputum, gastric aspirates or intestinal biopsies. Stool Xpert MTB/Rif may be useful as specimen collection is non-invasive. OBJECTIVES: To determine the sensitivity and specificity of Stool Xpert for the diagnosis of pulmonary or intestinal tuberculosis in Indian children. METHODS: We retrospectively compared the performance of stool Xpert, pulmonary sample Xpert (gastric lavage, sputum, bronchoalveolar lavage), and Mycobacteria Growth Indicator Tube (MGIT) for 52 children with pulmonary, intestinal or disseminated tuberculosis. RESULTS: Thirty (57.7%) children had positive pulmonary Xpert results, while 17(32.7%) had positive stool Xpert results. The sensitivity and specificity of stool Xpert compared to pulmonary Xpert was 43.3% and 81.8%, respectively. On using MGIT as a reference, pulmonary Xpert had a higher sensitivity compared to stool Xpert (81.8% vs. 63.6%, respectively) but lower specificity (48.8%vs.75.6%, respectively). Kappa analysis showed fair agreement between stool and pulmonary Xpert results (0.23), and moderate agreement between stool and gastric lavage Xpert results (0.43). Stool Xpert was found to be equivalent to pulmonary Xpert in detecting tuberculosis in children younger than 5 years (36.4% in both; 4 children each tested positive out of 11). No patients with isolated intestinal tuberculosis had positive stool Xpert. CONCLUSION: Stool Xpert may serve as a supplementary diagnostic tool in paediatric tuberculosis, particularly in young children where specimen collection is difficult. However, its lower sensitivity compared to pulmonary samples limits its potential as a standalone test.