Diagnosing pulmonary tuberculosis (TB) commonly relies on respiratory specimens such as sputum, but some patients are unable to expectorate. The study assessed whether testing stool with the Xpert MTB/RIF Ultra (Xpert-Ultra) molecular assay can reliably detect Mycobacterium tuberculosis (MTB) in adult presumptive TB patients in a setting with a generalized HIV epidemic. The authors positioned stool testing as a potential alternative diagnostic approach when respiratory sample collection is difficult.
This was a prospective, cross-sectional diagnostic accuracy study conducted at outpatient and inpatient departments of the Shoklo Malaria Research Unit (SMRU) clinics and the Mae Tao Clinic (MTC) on the Thailand–Myanmar border. The study enrolled presumptive TB patients of any age who presented between 14 November 2022 and 23 May 2023 based on reported signs, symptoms, and/or radiological findings. A total of 113 participants were included in the analysis. Nine participants (7.96%) had human immunodeficiency virus (HIV) infection.
The index test under evaluation was the Xpert-Ultra assay performed on stool specimens. The diagnostic performance of stool Xpert-Ultra was assessed against a reference standard of liquid culture for MTB performed on sputum. For comparison, the study also reported the performance of smear microscopy and the Xpert-Ultra assay when performed on sputum specimens.
Among the 31 culture-confirmed TB cases (31/113; 27.43%), the sensitivity of Xpert-Ultra on stool was reported as 90.32% (95% confidence interval [CI], 74.25% to 97.96%). Against culture-negative cases, the specificity of stool Xpert-Ultra was 98.78% (95% CI, 93.39% to 99.97%). These point estimates indicate high accuracy of the assay on stool in this cohort.
When comparing sample types, the absolute difference in sensitivity for stool Xpert-Ultra was 3.23% lower than sputum (95% CI of the difference: −9.46% to 3.00%). This difference was not statistically significant according to the report. The specificity difference between stool and sputum Xpert-Ultra was 1.22% (95% CI: −1.16% to 3.59%). The authors also evaluated smear microscopy and sputum Xpert-Ultra as comparator diagnostics, but detailed comparator figures beyond the sensitivity and specificity differences reported were not provided in the abstract.
The authors state that the diagnostic performance of Xpert-Ultra on stool in this study is consistent with World Health Organization (WHO) target product profiles for low-complexity assays intended to detect Mycobacterium tuberculosis. No further detail on which exact WHO targets or thresholds were used is provided in the source abstract.
Ethical oversight: the study received approval from the Oxford Tropical Research Ethics Committee (Ref: 570-22) and the local Community Ethics Advisory Board of Tak province (Ref: CEAB-2022-016). Written informed consent was obtained from adult participants or from parents/legal guardians for children. The authors declared no competing interests.
Data availability: the underlying data supporting the findings are available on request under the data-sharing policy of the Mahidol Oxford Tropical Medicine Research Unit (MORU); requests are evaluated by the MORU Data Access Committee.
Limitations reported or implicit from the source: the available text is an abstract of a preprint. The report emphasizes that the manuscript has not been peer-reviewed and should not be used yet to guide clinical practice. The abstract does not provide granular subgroup analyses (for example, sensitivity by HIV status), detailed methods for stool processing, or full breakdowns of comparator test results; those details were not reported in the source abstract.
In this cohort of adult presumptive TB patients on the Thailand–Myanmar border, the Xpert MTB/RIF Ultra assay on stool demonstrated high sensitivity (90.32%) and high specificity (98.78%) against sputum liquid culture. Differences in sensitivity and specificity between stool and sputum Xpert-Ultra were small and not statistically significant in this dataset. The authors conclude that stool testing with Xpert-Ultra can be considered a potential alternative diagnostic approach for presumptive pulmonary TB in adults when respiratory specimens are difficult to collect.
Clinicians and program planners should note that the findings derive from a preprint that has not undergone peer review; further peer-reviewed evidence and operational data (including details of stool processing, cost, and feasibility) would be needed before routine programmatic adoption.