Investigators reported high discrimination for both C-reactive protein (CRP) and serum amyloid A (SAA) when distinguishing bacteriology-confirmed tuberculosis from a mixed set of probable TB cases and healthy controls. The assays used were near point-of-care fluorescence immunoassays (NPOCT-FIA) for CRP and SAA. Short and direct: both markers tracked with confirmed disease. Longer explanation: receiver operating characteristic (ROC) analysis returned AUCs consistent with strong diagnostic performance, indicating these biomarkers may function as effective triage tests in the sampled population.
Data indicated SAA produced an AUC of 0.961 and CRP produced an AUC of 0.953 for identifying confirmed TB versus the combined comparator groups. The investigators reported an SAA cut-off of >31.0 mg/L and a CRP cut-off of 21.3 mg/L. At those thresholds the cohort showed 96.5% sensitivity for each marker. Specificity varied: SAA 88.6% and CRP 86.0%. When the two tests were combined, the study team observed 100.0% sensitivity and 83.3% specificity in these archived samples. The source did not report downstream operational characteristics such as time-to-result in field conditions or cost-per-test in programmatic settings.
Interferon gamma-induced protein 10 (IP-10) was measured by enzyme-linked immunosorbent assay (ELISA). The study team observed poor discrimination for IP-10, with an AUC of 0.545. That AUC indicates near-random classification performance in this sample set. The authors therefore did not include IP-10 among the biomarkers meeting WHO criteria in their analysis.
The analysis used archived serum from three predefined groups: bacteriology-confirmed symptomatic TB (n=57), probable TB (n=56), and healthy controls (n=60). The report described a cross-sectional diagnostic evaluation of stored samples rather than a prospective screening study. Samples originated from prior studies (including the EVIDENT study and a COVID-19 vaccine study) and were measured using NPOCT-FIA for CRP and SAA, and ELISA for IP-10. The source did not provide additional operational details such as sample storage duration, freeze–thaw cycles, or lot-specific assay performance in the public abstract. Those methodological specifics were not reported in the abstract and would require consulting the full manuscript or corresponding authors for confirmation.
The authors noted ethical approvals from the Health Research Ethics Committee of Dr. Hasan Sadikin General Hospital, Bandung, Indonesia, and from Universitas Padjadjaran for the parent studies supplying archived samples. They stated that all necessary participant consent procedures were followed and that data are available upon reasonable request. The article is posted as a preprint; the authors and the medRxiv notice explicitly indicate it has not been certified by peer review and therefore should not yet be used to guide clinical practice.
The report has important limitations anchored in the source text. The dataset comprised archived serum rather than samples collected during prospective screening, so real-world triage performance (in clinics or community screening programs) was not assessed in this analysis. The abstract did not report operational metrics such as assay stability under field conditions, time-to-result at point of care, or cost-effectiveness—details needed for programmatic adoption. The investigators did not report those implementation parameters in the abstract. Another open question: how the reported cut-offs perform across different populations, including people living with HIV, children, or those with non-tuberculous inflammatory conditions; the source did not present subgroup analyses in the abstract. Finally, the preprint status means peer review could alter interpretation or highlight additional constraints.
Next steps implied by the reported findings are empirical: prospective validation of CRP and SAA measured with near-point-of-care platforms in programmatic screening settings, assessment of operational feasibility, and demonstration of consistent thresholds across diverse patient groups. The source limited its claims to the diagnostic results from this archived-serum cohort and declared that SAA and CRP met the WHO target product profile criteria for triage screening in this sample set, but did not claim implementation readiness or regulatory approvals.