BACKGROUND: An accurate, rapid, non-sputum-based triage test for diagnosing tuberculosis (TB) is needed. METHODS: A prospective evaluation of the Cepheid GeneXpert Mycobacterium tuberculosis Host Response cartridge (Xpert-MTB-HR), a prototype blood-based host response mRNA signature assay, among individuals presenting with TB-like symptoms was performed in Pakistan and results were compared to 3 reference standards: Xpert MTB/RIF Ultra, bacteriological confirmation (Xpert MTB/RIF Ultra and/or culture positivity), and composite clinical diagnosis (clinician diagnosis, treatment initiation, Xpert MTB/RIF Ultra, and/or culture positivity). Analyses were conducted both for the entire study cohort and separately in the adolescent and young adult cohort (aged 10-24 years). RESULTS: A total of 497 participants, aged 6-83 years, returned valid Xpert-MTB-HR results. When a diagnostic threshold was set for a sensitivity of >90%, specificity was 32% (95% confidence interval [CI], 28%-37%) compared to Xpert MTB/RIF Ultra, 29% (95% CI, 25%-34%) compared to a bacteriological confirmation, and 22% (95% CI, 18%-26%) compared to a composite clinical diagnosis. However, when evaluating only the adolescent and young adult cohort with a diagnostic threshold set for sensitivity of >90%, specificity was 82% (95% CI, 74%-89%) compared to Xpert MTB/RIF Ultra, 84% (95% CI, 75%-90%) compared to a bacteriological confirmation, and 54% (95% CI, 44%-64%) compared to a composite clinical diagnosis. CONCLUSIONS: While the Xpert-MTB-HR does not meet World Health Organization minimum criteria in the general population, in our study it does meet the minimum sensitivity and specificity requirements for a non-sputum-based triage test among adolescents and young adults when compared to Xpert MTB/RIF Ultra or bacteriological confirmation.