Diagnosing pulmonary tuberculosis in children is challenging because obtaining sputum is often difficult and bacteriologic confirmation is less frequent than in adults. Urine-based assays offer a non-sputum alternative that could be more feasible in paediatric populations. The current commercially available urine lipoarabinomannan (LAM) test has limited sensitivity in children and is primarily used among those with HIV. Co-detection of Mtb proteins with LAM in urine could provide complementary pathogen-derived biomarkers to improve diagnostic performance.
Researchers developed a multiplex electrochemiluminescence (ECL immunoassay) to quantify five Mycobacterium tuberculosis proteins (Ag85B, CFP-10, ESAT-6, MPT32, MPT64) and LAM in urine. Analytical limits of detection were established, and diagnostic performance was evaluated using urine samples from 630 children enrolled across five high-burden countries (The Gambia, India, Peru, South Africa, and Uganda). The cohort had a median age of 4 years (IQR 2–8), 44% were female, 15% were living with HIV, 19% were underweight, and 24% had Confirmed pulmonary tuberculosis. Children were categorized as Confirmed, Unconfirmed, or Unlikely pulmonary TB for performance analyses.
The multiplex ECL assay achieved femtomolar analytic limits of detection, reported as 1.5 to 4.0 fM for the measured analytes. These analytical sensitivities enabled measurement of low-abundance pathogen-derived proteins and LAM in pediatric urine specimens.
Individual Mtb proteins measured in urine demonstrated very high specificity but modest sensitivity for Confirmed TB. Across the proteins, sensitivity ranged from 12% to 33% while specificity ranged from 98% to 100%. Of the individual markers, Ag85B had the highest sensitivity at 33% (95% CI 26–41) for Confirmed tuberculosis and showed similar sensitivity to LAM when considered alone.
Combining biomarkers improved sensitivity relative to individual markers. A four-antigen signature comprising Ag85B, MPT64, MPT32, and LAM yielded 50% sensitivity (95% CI 42–58) and 94% specificity (95% CI 90–96) for detecting Confirmed pulmonary TB. This multi-antigen signature was significantly more sensitive than LAM alone, with the greatest incremental sensitivity observed among children without HIV.
An additional finding was that 16 children (representing 10% of those classified as Unconfirmed TB) had at least one Mtb protein or LAM detected in urine, suggesting the assay may identify pathogen-derived biomarkers in some cases lacking bacteriologic confirmation.
Performance was assessed overall and stratified by HIV and nutritional status. The four-antigen signature improved sensitivity particularly among children without HIV. The source reports subgroup analyses were performed, but detailed numeric breakdowns by each subgroup beyond the summary statements above were not provided in the abstract.
The study demonstrates that multiple Mtb proteins are detectable in pediatric urine with high specificity using an ultrasensitive ECL immunoassay. Multi-antigen urine signatures can augment sensitivity compared with LAM alone while maintaining high specificity, and may detect pathogen biomarkers in a subset of children without bacteriologic confirmation. These results define analytical targets (femtomolar detection) and support further development of multiplex urine-based assays as potential point-of-care diagnostics for childhood TB.
The study was funded by the National Institutes of Health. Several authors (SM, MW, JW, GBS) are employees of Meso Scale Diagnostics, LLC, and one author (JW) has an ownership interest in the company; the company and funders did not have a role in the decision to publish. Ethical approvals were obtained from multiple institutional review boards and ethics committees across participating sites, including committees at Mulago Hospital, The Gambia/MRC, London School of Hygiene and Tropical Medicine, the National Institute of Health – Peru, University of Cape Town, BJ Government Medical College, Johns Hopkins University, and University of California, San Francisco.
De-identified individual participant data and a data dictionary are available after publication to researchers who submit a methodologically sound proposal approved by the study investigators and sign a data sharing agreement. Contact the corresponding author for proposals.