Testing Multiple Noninvasive Specimen Types May Improve Diagnosis of Pediatric Tuberculosis

Pediatric exam, chest, cough, back tuberculosis
Pediatric exam, chest, cough, back tuberculosis
Investigators sought to determine what combination of minimally invasive specimens is most effective for the diagnosis of tuberculosis in children.

The far less-invasive collection and testing of nasopharyngeal aspirate (NPA) samples in combination with stool and urine sample testing may improve the diagnosis of tuberculosis (TB) in children in limited-resource settings. This finding from a prospective, cross-sectional diagnostic study, was published in JAMA Pediatrics.

Children (N=300) aged younger than 5 years who presented with persistent cough, fever, or malnutrition from 2013 to 2015 in Kisumu County, Kenya, were recruited for this study. Two samples each of NPA, gastric aspirate, suctioning following sputum induction (IS), string tests, urine, stool; in some cases cervical lymph node fine-needle aspirate (FNA); and a single blood sample were collected and assessed for TB infection by Xpert MTB/RIF or mycobacteria growth indicator tube (MGIT) assays.

The median age of participants was 2 years (interquartile range [IQR], 1.0-3.6), 50.3% were girls, and 24.3% were HIV positive. A total of 31 children had confirmed TB infection. Testing up to 2 specimens by Xpert MTB/RIF or MGIT assays identified TB among 24 gastric aspirate samples (sensitivity, 77%; interdecile range [IDR], 68-87), 23 NPA specimens (sensitivity, 74%; IDR, 64-84), 20 IS (sensitivity, 64%; IDR, 53-76), 15 string test samples (sensitivity, 48%; IDR, 37-60), 14 stool specimens (sensitivity, 45%; IDR, 33-57), and 4 urine specimens (sensitivity, 13%; IDR, 5-21).

Combining results from 2 specimen types identified 3 additional children who were positive for TB when adding gastric aspirate or IS specimens (sensitivity, 10%; IDR, 3-17) and 2.5 for NPA samples (sensitivity, 8%; IDR, 3-15).

Combining results from gastric aspirate and NPA identified 25 children who were positive for TB (sensitivity, 81%; IDR, 71-90), 2 gastric aspirate and 2 NPA found 28 positives (sensitivity, 90%; IDR, 83-97), and 2 gastric aspirate, 2 NPA, 1 IS, and 1 urine sample identified TB among 30 children (sensitivity, 97%; IDR, 92-100).

In order to successfully detect TB among all 31 infected children, testing of 8 specimens was required. Of the 2 diagnostic tests, the MGIT assay outperformed Xpert MTB/RIF in all specimen types except urine and stool.

This study may have been biased by the choice to include MGIT assays because their use is rare in clinical settings without access to laboratory equipment. Because MGIT outperformed Xpert MTB/RIF in most specimens, these values may be overestimations for care settings in which only Xpert MTB/RIF testing is available.

The study authors concluded that compared with standard diagnosis strategies for TB in young children, testing 2 NPA or 1 NPA and 1 stool sample are equally as sensitive as current diagnostic standards but are less-invasive specimens to collect.


Song R, Click ES, McCarthy KD, et al. Sensitive and feasible specimen collection and testing strategies for diagnosing tuberculosis in young children. JAMA Pediatr. Published online February 22, 2021. doi:10.1001/jamapediatrics.2020.6069