Reply to Drancourt

To the Editor—Professor Drancourt correctly highlights the need for improved sampling methods to establish a microbiologic diagnosis of pulmonary tuberculosis (PTB) in children, but questions the rationale for gastric aspirates being used in young children with suspected PTB [1].

He asserts, based on his work on isolation of Mycobacterium tuberculosis (Mtb) from the stools of (predominantly) adult tuberculosis patients using a chlorhexidine-decontamination protocol (referenced in [1]), for a shift away from gastric aspirate sampling to testing of stool for Mtb, although such evidence is sparse in children <5 years of age, the age group that was the subject of the Pneumonia Etiology Research for Child Health (PERCH) study. In fact, stool culture for Mtb has been shown to be inferior in yield compared to culture of gastric aspirates in children in countries with a high burden of tuberculosis [2, 3].

In the era of molecular diagnostic assays such as Xpert® MTB/RIF (Cepheid, Sunnyvale, CA, USA), which evolved largely after the design and implementation of the PERCH study, there has been increased evaluation of stool samples for diagnosing PTB in children. Of the published studies evaluating the Xpert® MTB/RIF assay on stool specimens in children suspected of having PTB, the sensitivity of stool samples ranges from 31.9% [4] to 68.4% [5] overall, compared to culture of respiratory specimens. Although a single study by Banada et al reported a higher yield for identification of Mtb from stool samples compared to gastric aspirate or induced sputum culture using Xpert® MTB/RIF [6], the study did not include a formal comparison against culture of Mtb from respiratory samples, as the case definition of “confirmed tuberculosis” required gastric aspirates or induced sputum samples that were positive on Xpert® MTB/RIF, rather than the traditional gold standard of a culture-positive specimen.

Pioneered in France in the late 19th century [7], gastric aspirates are currently recommended as an appropriate and suitable diagnostic modality in young children [8], and are widely used. In a systematic review of national and international childhood tuberculosis guidelines, gastric aspiration was the favored sampling method for the investigation of PTB in children too young or unable to provide expectorated specimens [9]. The most recent guidance document of the World Health Organization regarding the management of childhood tuberculosis similarly reinforces the use of gastric aspirates in these patients [10]. None of these recommendations view patient discomfort to be a suitable reason to dispense with the use of gastric aspiration in an attempt to achieve a microbiological diagnosis of PTB in young children.

Noninvasive sampling techniques are to be rigorously sought after, so as to attain a reliable diagnosis at the cost of least risk and discomfort to the patient. In the arena of tuberculosis diagnostics, stool samples are a prototype for this more humane approach. At the current juncture, however, we believe that there is insufficient evidence to abandon gastric aspirate sampling as a modality for identifying Mtb in young children with suspected PTB in favor of stool specimens.

Note

Potential conflicts of interest. K.O. has received a grant from the Bill & Melinda Gates Foundation (BMGF). S. M. has received grants from BMGF, GSK, Pfizer, and Minervax; has consulted for BMGF and Pfizer; and has received payment for development of educational presentations from Medscape. All other authors report no potential conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References