Stool-based Xpert MTB/RIF assay for the diagnosis of pulmonary tuberculosis in children at a teaching and referral hospital in Southwest Ethiopia

BACKGROUND: Diagnosis of tuberculosis (TB) in children is challenging mainly due to the difficulty of obtaining respiratory specimen and lack of sensitive diagnostic tests. The objective of this study was to evaluate the diagnostic performance of Xpert MTB/RIF (Xpert here after) for the diagnosis of pulmonary TB (PTB) from stool specimen in children. METHODS: A cross-sectional study was conducted among consecutively recruited children (less than 15 years old) with presumptive PTB at Jimma Medical Center, Ethiopia. One pulmonary specimen (expectorated sputum or gastric aspirate) was collected from each participant and tested for TB by Xpert and Lowenstein-Jensen (LJ) culture. In addition, one stool specimen per child was collected and tested by Xpert after a single step, centrifuge-free stool processing method adapted from KNCV TB Foundation. Diagnostic performance of Xpert was calculated with reference to LJ culture and to a composite reference standards (CRS) comprising of confirmed TB (positive by Xpert and/or culture) and unconfirmed TB (clinical diagnosis with improvement after anti-TB treatment). RESULTS: A total of 178 children were enrolled; 152 of whom had complete microbiological results. Overall, TB was diagnosed in 13.2% (20/152) of the children with presumptive TB. Of these, only ten had microbiologically confirmed TB (positive Xpert and/or culture) and the remaining ten were clinically diagnosed with positive response to anti-TB treatment and were classified as unconfirmed TB. Stool Xpert had sensitivity of 100% (95%CI: 66.4–100) and specificity of 99.3% (95%CI: 96.2–100) compared to culture; however, the sensitivity was decreased to 50% (95%CI: 27.2–72.8) when compared to CRS. The Xpert on gastric aspirate had sensitivity of 77.8% (95%CI: 40–97.2) compared to culture and 40% (95%CI: 19.1–64) compared to CRS. CONCLUSIONS: The sensitivity of Xpert for stool sample is comparable to that for gastric aspirate. Stool sample is a potential alternative to pulmonary specimen in the diagnosis of pulmonary TB in children using Xpert.